Alternatives For Environmental Evaluation

Alternatives for Environmental
Valuation
Environmental cost–benefit analysis was developed by economists in the belief

that monetary valuation of the environmental repercussions of economic activity
is essential if the ‘environment’ stands any chance of being included in
governmental and business decisions.
This volume examines the limitations of this monetary approach, and considers
the alternatives. Three broad angles from which to view environmental values are
presented: applying social psychology concepts which challenge standard
approaches; introducing multidimensional and non-monetary techniques; and
examining vested interest group and citizen participation in processes of
environmental valuation. Combinations of these approaches are also covered.
The contributions presented are a valuable resource for both environmental
and ecological economists. This authoritative book will also prove useful for those
with a general interest in the environment, including policy-makers and non-
governmental organizations.
Michael Getzner is Associate Professor of Economics at the University of

Klagenfurt, Austria.
Professor Clive L. Spash holds the Research Chair in Environmental and Rural
Economics at the University of Aberdeen, UK. He is head of the Socio-Economic
Research Programme at the Macaulay Institute and President of the European
Society for Ecological Economics.
Sigrid Stagl is Lecturer and coordinator of the Ecological Economics MA

Programme at the School of the Environment, the University of Leeds, UK. She
is Vice-President of the European Society for Ecological Economics.
Routledge Explorations in Environmental Economics
Edited by Nick Hanley
University of Glasgow
1 Greenhouse Economics
Values and ethics
Clive L. Spash
2 Oil Wealth and the Fate of Tropical Rainforests

Sven Wunder
3 The Economics of Climate Change

Edited by Anthony D. Owen and Nick Hanley
4 Alternatives for Environmental Valuation

Edited by Michael Getzner, Clive L. Spash and Sigrid Stagl
Alternatives for
Environmental Valuation
Edited by
Michael Getzner, Clive L. Spash and
Sigrid Stagl
First published 2005
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
Simultaneously published in the USA and Canada
by Routledge
270 Madison Ave, New York, NY 10016
Routledge is an imprint of the Taylor & Francis Group
This edition published in the Taylor & Francis e-Library, 2005.
“To purchase your own copy of this or any of Taylor & Francis or Routledge’s
collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.”
© 2005 Michael Getzner, Clive L. Spash and Sigrid Stagl for selection
and editorial matter; the contributors for individual chapters
All rights reserved. No part of this book may be reprinted or reproduced
or utilised in any form or by any electronic, mechanical, or other means,
now known or hereafter invented, including photocopying and recording,
or in any information storage or retrieval system, without permission in
writing from the publishers.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging in Publication Data
A catalog record for this book has been requested
ISBN 0-203-41287-7 Master e-book ISBN
ISBN 0-203-34079-5 (Adobe eReader Format)

ISBN 0–415–31012–1 (Print Edition)
Contents v
Contents
List of figures vii

List of tables ix
List of contributors xi
Preface xiii
Acknowledgements xiv
List of abbreviations xv
1 Exploring alternatives for environmental valuation 1

C L I V E L . S PA S H , S I G R I D S TAG L A N D M I C H A E L G E T Z N E R
PART I
Extending the environmental valuation approach 21
2 A framework for valuing nature: regional biodiversity 23

MICHAEL GETZNER
3 Non-use values and attitudes: wetlands threatened by climate

change 51
J Ü RG E N M E Y E R H O F F
4 Modelling environmental behaviour: socio-psychological

simulation 69
H A N S - J OAC H I M M O S L ER
PART II
Taking multiple criteria into account 97
5 Assessing the quality of different MCDA methods 99

A N D R E A D E M O N T I S, P A S QUA L E D E T O RO, B E RT D RO S T E - F R A N K E ,
I N E S O M A N N AND S I G R I D S TAGL
vi Contents
6 MCDA and stakeholder participation: valuing forest resources 134
WENDY P ROCTOR
7 Confronting risk with precaution: a multi-criteria mapping of

genetically modified crops 159
A N DY S T I R L I N G A N D S U E M AY E R
PART III
Deliberation, participation and value expression 185
8 Consumer valuation and citizen deliberation: towards a

comparison 187
JONATHAN ALDRE D
9 Three approaches to valuing nature: forest floodplain restoration 209

W E N DY K E N YO N A N D N I C K H A N L E Y
10 Deliberation and economic valuation: national park management 225

R O S E M A RY F. J A M E S A N D R U S S E L L K . B L A M E Y
11 The challenges of stakeholder participation: agri-environmental

policy 244
K AT JA A R Z T
12 Preference transformation through deliberation: protecting world

heritage 263
S I M O N N I E M E Y ER
Index 290
List of figures vii
Figures
2.1 The ‘standard’ and ‘extended’ model of valuation 25

3.1 A conceptual model of WTP determinants 57
3.2 Tree diagram for non-users 62
3.3 Tree diagram for users 62
4.1 Framework model of behaviour toward the environment 73
4.2 Representation of the ‘processing of group influences’ 77
4.3 Change in average attitute due to various group-oriented
interventions 79
4.4 Representation of ‘processing of observed behaviour towards the
environment’ 80
4.5 Average population behaviour with differing interventions based
on observational learning 82
4.6 Representation of ‘processing of information on the collective
use of a resource’ 84
4.7 Use of an environmental resource 86
4.8 Representation of ‘processing of communicative influencing
under conditions of knowledge and environmental concern’ 88
4.9 Information campaign experiments without environmentally
relevant communication 90
4.10 Experiments with multiplicators 92
6.1 Hierarchy of objectives, criteria and indicators for the Southern
Region RFA 145
6.2 Preferences of criteria 146
6.3 Preferences of environmental criteria 147
6.4 Preferences of economic criteria 147
6.5 Preferences of social criteria 148
6.6 Overall ranking of options 153
6.7 Priorities of broad criteria 153
6.8 Range of individual priorities 154
6.9 Environmental preference 154
6.10 Economic preference 155
7.1 Risk, uncertainty, ambiguity and ignorance defined 163
viii List of figures
7.2 Ambiguity of ordering in risk assessments 166
7.3 A model of the relationships between risk, science and precaution 168
7.4 The multi-criteria mapping process 175
7.5 Final individual rankings for the basic options 179
7.6 Diverse options mixes favoured by individual participants 181
10.1 CJs applied to non-market valuation 230
10.2 Spending on parks excluding salaries 232
10.3 Process for decision making under ‘charge 1’ 234
10.4 Process for decision making under ‘charge 2’ 235
11.1 Participants at the first six meetings of the AEF 253
12.1 The Bloomfield Track 267
12.2 Juror loadings on factor 1 ‘preservation’ 274
12.3 Juror loadings on faction 2 ‘optimism’ 275
12.4 Juror loadings on factor 3 ‘pragmatism’ 277
12.5 Juror loadings on factor 4 ‘symbolism’ 277
12.6 Changes to preference factor loadings before and after
deliberation 280
12.7 Changes to loadings on ‘preservationist’ and ‘close road’ factors 281
12.8 Changes to loadings on ‘pragmatist’ and ‘status quo’ factors 282
12.9 Changes to loadings on ‘optimist’ and ‘status quo’ factors 283
List of tables ix
Tables
2.1 Consumer vs. citizen and students’ perspectives 36

2.2 Epistemological problems of WTP surveys 37
2.3 WTP regression model variables 38
2.4 Logit model estimated coefficients for WTP1 39
2.5 Log-normal OLS estimated coefficients for WTP2 41
3.1 Distribution of interviews and in-principle WTP 55
3.2 Summary statistics of the WTP (in DM) for users and non-users 56
3.3 Attitudes towards nature 58
3.4 Attitudes towards climate change 58
3.5 Attitudes towards nature and WTP 59
3.6 Attitudes towards climate change and WTP 59
3.7 Relationship between attitudes towards money and mean WTP 60
3.8 Potential explanatory variables 61
3.9 OLS results of non-user group 64
3.10 OLS results of user group 64
5.1 List of quality criteria for MCDA methods 101
5.2 Summary of method comparison 126
6.1 The AHP intensity of importance scale 141
6.2 Example pairwise comparison 143
6.3 Normalized matrix example 143
6.4 Pairwise comparison of environmental criteria 143
6.5 Overall ranking example 143
6.6 Impact table for forest use options 149
6.7 Pairwise comparison of options under environmental criteria 150
6.8 Priority of options in terms of environmental criteria 150
6.9 Pairwise comparison of options under economic criteria 151
6.10 Priority of options in terms of economic criteria 151
6.11 Pairwise comparison of options under social criteria 152
6.12 Priority of options in terms of social criteria 152
7.1 The definitions of the ‘basic options’ appraised by all participants 173
7.2 The participants 174
7.3 Additional options defined by participants 176
x List of tables
7.4 Broad groupings of criteria defined by participants 177
8.1 Possible advantages of discussion 191
9.1 Descriptive statistics for WTP for the Ettrick project 214
9.2 Positive issues identified with the forest floodplain project 216
9.3 Jury suggestions measuring environmental project success 217
9.4 Original and revised bids with reasons for change 218
9.5 WTP descriptive statistics for initial valuation 219
9.6 WTP descriptive statistics for final valuation 219
9.7 Good aspects of the Ettrick project 219
9.8 Project problems, solutions and the importance of bad aspects 220
10.1 Charge 1: CJ group preference among three options 231
10.2 Charge 2: introducing Option 4 231
11.1 Variables feeding into trust, reciprocity and reputation 251
11.2 Proceedings at the agri-environmental forum 254
12.1 Statement scores for Q attitudinal factors 272
12.2 Rank ordering of preferences for management options 278
12.3 Factor scores for preference ranking 279
12.4 Correlation between preference and Q factors loading 279
List of contributors xi
Contributors
Jonathan Aldred, Department of Land Economy, Cambridge University, UK,

email: jsa1001@cam.ac.uk
Katja Arzt, Department of Agricultural Economics and Social Science, Chair of
Resource Economics, Humboldt University of Berlin, Germany, email:
katja.arzt@agrar.hu-berlin.de
Russell K. Blamey, Programme Visitor, Research School of Social Sciences,
Australian National University, Canberra, Australia, email: russell.blamey@
anu.edu.au
Andrea De Montis, Dipartimento di Ingegneria del Territorio, Sezione di
Urbanistica, Università di Cagliari, Cagliari, Italy, email: ademonti@
vaxca1.unica.it
Pasquale De Toro, Dipartimento di Conservazione dei Beni Architettonici ed
Ambientali, Università di Napoli Federico II, Naples, Italy, email:
detoro@unina.it
Bert Droste-Franke, Institute of Energy Economics and the Rational Use of
Energy, University of Stuttgart, Stuttgart, Germany, email: bd@ier.uni-
stuttgart.de
Michael Getzner, Department of Economics, University of Klagenfurt,
Klagenfurt, Austria, email: michael.getzner@uni-klu.ac.at
Nick Hanley, Department of Economics, Adam Smith Building, The University
of Glasgow, Glasgow, UK, email: n.d.hanley@ed.ac.uk
Rosemary F. James, Centre for Resource and Environmental Studies, Australian
National University and CRC for Weed Management Systems, Campbell,
Australia, email: rfjames@cres.anu.edu.au
Wendy Kenyon, Socio-Economic Research Programme, Macaulay Institute,
Aberdeen, UK, email: w.kenyon@macaulay.ac.uk
Sue Mayer, GeneWatch UK, Tideswell, Buxton, Derbyshire, UK, email:
sue.mayer@genewatch.org
xii List of contributors
Jürgen Meyerhoff, Institute of Management in Environmental Planning,
Technical University Berlin, Berlin, Germany, email: Meyerhoff@imup.tu-
berlin.de
Hans-Joachim Mosler, Department of Psychology, Division of Social
Psychology, University of Zürich, Zürich, Switzerland, email: mosler@
sozpsy.unizh.ch
Simon Niemeyer, Social and Political Theory Group, Research School of Social
Sciences, Austalian National University, Canberra, Australia, email:
simon@coombs.anu.edu.au
Ines Omann, Department of Economics, Karl-Franzens University, Graz,
Austria, email: ines.omann@kfunigraz.ac.at
Wendy Proctor, Centre for Resource and Environmental Studies, Australian
National University, Canberra, Australia, email: Wendy.Proctor@csiro.au
Clive L. Spash, Socio-Economic Research Programme, Macaulay Institute,
Aberdeen, UK, and Geography and Environment Department, University
of Aberdeen UK, email: c.spash@abdn.ac.uk
Sigrid Stagl, School of the Environment, University of Leeds, Leeds, UK, email:
sts@env.leeds.ac.uk
Andy Stirling, Science Policy Research Unit (SPRU), University of Sussex,
Brighton, UK, email: a.c.stirling@sussex.ac.uk
Preface xiii
Preface
This book arose from on-going debates occurring amongst European ecological
economists who have been reflecting upon extensions of the ‘standard’ economic
model of environmental valuation. Such extensions try to account for citizens’
attitudes, institutional framing and social context as they affect any given
environmental change being valued. Three broad approaches to valuation can
then be identified and are pursued here. First, economists have been looking to
learn from social psychology. Thus, for example, the concept of reasoned action
has been one way in which qualitative and quantitative factors have begun to be
linked in some contingent valuation studies. At the same time the process of learning
from social psychology challenges standard approaches and requires that new
alternatives be developed. A major set of alternatives to environmental cost–benefit
analysis, which form the second broad approach, focus upon multiple criteria for
understanding and improving decisions. The work reported here explores both
the potential, as well as the pitfalls of multidimensional and non-monetary
techniques. The final approach deals with innovation in the area of vested interest
group and citizen participation in processes of environmental valuation. One focus
of research in this area has been in terms of combining economic valuation and
participatory deliberative approaches arising from political science. Work is
presented comparing citizens’ juries and contingent valuation and reflecting upon
potential combinations termed deliberative monetary valuation. Stakeholder or
vested interest group participation is more common than trying to represent the
general public but, as discussed, can also prove difficult. Overall, this book shows
a range of approaches being explored by ecological economists and how they are
open to interdisciplinary critiques, often dismissed by narrowly focused disciplines.
In particular, European ecological economists have been engaging with traditional
economics approaches while also learning from applied philosophy, geography,
political theory, social psychology and sociology.
xiv Acknowledgements
Acknowledgements
Michael Getzner would like to pay tribute to Professor Egon Matzner (Vienna),
who sadly passed away prematurely in October 2003, and who he will miss as
both an outstanding teacher and a colleague.
We all acknowledge our debt to the European Society for Ecological Economics
which has provided an active forum for discussion on such matters as are covered
in this volume. There is now a great deal of networking that takes place across
Europe amongst those addressing the socio-economic aspects of ecological
problems. This includes both those involved in policy processes and academics.
The area of science and society interactions is a general and key concern with
specific realisations in new technologies. Friends and colleagues who have stimulated
our thoughts in these areas with respect to the topics in this volume include: Anders
Biel, Jacquie Burgess, Claudia Carter, Mick Common, Silvio Funtowicz, Colin
Green, Alan Holland, Sybille van den Hove, Joan Martinez-Alier, Martin
O’Connor, John O’Neill, Ortwin Renn, Peter Söderbaum and Arild Vatn. There
are also several people outside Europe with whom we interacted and acknowledge
as active in these debates including: Edmundo Claro, John Gowdy, Jack Knetsch,
Michael Lockwood, Dick Norgaard, Bryan Norton and Sabine O’Hara. Of course
as editors we also thank all the contributors to this volume.
Abbreviations
AEF agri-environmental forum

AHP analytic hierarchy process
CBA cost–benefit analysis
CJs citizens’ juries
CVM contingent valuation method
DMV deliberative monetary valuation
EU European Union
Evamix evaluation matrix
GM genetically modified
GP goal programming
MAUT multiple attribute utility theory
MCA multiple criteria analysis
MCDA multiple criteria decision analysis
MCM multi-criteria mapping
MOP multi-objective programming
NAIADE novel approach to imprecise assessment and decision
environments
NGO non-governmental organisation
NOAA National Oceanic and Atmospheric Administration
UK United Kingdom
USA United States of America
WTA willingness to accept
WTP willingness to pay
xvi List of abbreviations
Exploring alternatives for environmental valuation 1
1 Exploring alternatives for

environmental valuation
Clive L. Spash, Sigrid Stagl and
Michael Getzner
Introduction
The method of inclusion of environmental resources and ecosystem services in
decision processes determines how far they are taken into account with results
affecting the quality of our lives and those of future generations. A persistent
argument has been that monetary valuation is essential if the ‘environment’ is to
have any chance of being included in government and business decisions.
Environmental cost–benefit analysis (CBA) was developed by environmental
economists in order to achieve this monetisation of environmental entities so that
the prices in market economies might be adjusted. A range of methods were
developed including travel cost, hedonic pricing, production function analysis,
contingent valuation and choice modelling (see Hanley and Spash 1993; Spash
and Carter 2002). The overall aim has been to select project options on the basis
of their welfare impacts and to support government taxes which reflect the social
costs of environmental degradation. This approach has met with some success in
that various national and international agencies have been interested in performing
monetary valuation exercises as part of their overall assessment of projects. The
idea of environmental taxation has also risen on the political agenda if remaining
limited in practice.1
However, there has also been criticism of environmental CBA, or more
specifically of some of the studies conducted under that guise. Critiques can be
broadly grouped into those concerned with the theoretical foundations of economic
values, and those looking at the validity of specific numbers being produced and
the tools employed. In the former case, Kapp (1950) provides an isolated early
example which shows the limits to monetary as opposed to other values in society.
More recently, environmental philosophers have produced a whole body of
literature showing the narrowness of value theory typically expressed by economists
and the failure of economic training to place this theory in the context of wider
social values (see O’Neill 1993). Reflecting upon this literature requires realising
that most economics concerns consequentialism using a form of preference-based
utilitarianism, which is a very specific philosophy of value rather than a generally
accepted meta ethic which can be universally applied. In trying to address the
environmental problems of the 20th century, economic value theory has shown
2 Clive L. Spash, Sigrid Stagl and Michael Getzner
itself lacking in several areas, such as the treatment of time, complexity, strong
uncertainty, political systems, rights and social norms. The enhanced Greenhouse
Effect raises all these issues and exposes the failures of discounting, partial
equilibrium analysis, risk assessment and assuming that efficiency is the key social
goal (Spash 2002b).
The other area of critical analysis abstracts from the first set of problems and
tends to be a debate on validity which has been largely internal to the economics
profession (although implications of the value debate are also related via theoretical
foundations). Validity concerns result in specific applications being criticised for
their failure to respect microeconomic and welfare theoretic constraints. Those
working within decision making government institutions, such as environmental
protection agencies or the treasury, may see such theoretical restrictions as academic
and going beyond what is required for practical policy making (for example, see
comments by Burney 2000). However, once valuation becomes divorced from its
theoretical roots, numbers can be produced which have little content or meaning,
and are defensible only in terms of their political role rather than their theoretical
basis. In this respect the popularisation of environmental CBA has been a problem
as studies proliferate, numbers are transferred out of context and applications
seek to meet policy desires. Excessive aggregation is exemplified by attempts to
value regional, national and even the world’s ecosystems, although what is meant
by a trade price in these contexts is mystifying, and asking who might be the ‘seller’
and ‘purchaser’ exposes the fallacious use of economic value theory. Furthermore,
the range of reasonable numbers stemming from monetisation exercises is often
very broad, which facilitates the arbitrary use of valuation results.
The contingent valuation method (CVM) has become the most widely conducted
CBA tool. The main advantage attracting this attention is the ability of CVM to
estimate what are termed option, ‘existence’ and bequest values in addition to
direct use values. These are amounts reflecting: the preservation of an option to
use a resource in the future, the maintained existence of a resource or entity
regardless of personal use, and the desire to pass the resource to one’s descendents.
The sum of these indirect or passive use values can be large compared with the
direct use values associated with non-market goods to which the other CBA methods
are solely restricted. The popularity of the CVM is also based upon the apparent
simplicity of asking members of the general public the maximum they are willing
to pay (WTP) for an environmental improvement or, less commonly, the least they
would be willing to accept (WTA) in compensation for environmental degradation.
Considerable research energy has gone into refining CVM and related stated
preference approaches. The result has been to increase the complexity and cost of
conducting a CVM if it is to attain the highest standards of validity. In the United
States of America this is the legally defensible standard which can cost millions of
dollars for a nationally representative random sample including a range of pre-
tests, focus groups and internal consistency and validity checks in the survey, as
well as several variations on survey design.
The litigation surrounding the oil spill from the tanker Exxon Valdez led to a
set of such standards being produced by a panel of experts commissioned by the
National Oceanic and Atmospheric Administration (NOAA). This panel consisted
of a number of famous economists, such as Nobel Prize winners Kenneth Arrow
and Robert Solow. The panel was also asked to investigate the concept of ‘existence’
value. The panel’s report gave a qualified support for the CVM (NOAA 1993).
The panel stated that ‘existence’ values are a theoretically meaningful aspect of
value. As regards the ability of the CVM to estimate such values, the panel noted
five main problem areas: inconsistency of responses with economic models of
rational choice; mental account bias, especially when respondents are inadequately
informed as to substitutes; aggregating benefits; providing information; and warm
glow effects. However, the panel felt that so long as certain guidelines were followed,
CVM results could be judged as both meaningful and useful indicators of natural
resource damages. These guidelines are that (i) WTP should be used in preference
to WTA; (ii) mail surveys should be avoided; (iii) respondents should be given full
information on the resource change (including information on substitutes), and be
asked how well they understand this information; (iv) open-ended responses should
be rejected in favour of closed-ended referendum formats; (v) a random population
sample is required; (vi) respondents should be reminded about the need to reduce
expenditure on some item of their budget in order to be able to pay their stated
bid; and (vii) careful pre-testing should be carried out.
Such rules tend to be excessively prescriptive and ignore the specific context in
which a valuation exercise must take place. There are certainly general design
features in any contingent valuation study which are desirable. These include clear
description of the institutional context, explaining the consequences and expected
benefits of payment, being aware of various biases which can be introduced in
design, such as use of a contentious bid vehicle, and generally producing a realistic
scenario. Without addressing such issues any research which claims to be relevant
to the economic debate will be seen as inadequate and may also be misleading.
However, employing rules which prevent hypothesis testing and experimental design
is more worrisome. The NOAA Panel holds an implicit belief in a single universal
standard of best practice which ignores cultural, legal and institutional variation.
People trade differently in different countries (e.g. haggling vs. accepting the stated
price), property rights vary (WTP vs. WTA), areas of trade which are taboo differ,
and truly random sampling is difficult at best and impossible in industrially
developing countries. A serious concern here is the extent to which research designs
which fail to conform to the rules are branded as poor practice when they may be
valid research aiming to question the underlying tenets of the model. Blind use of
highly specific guidelines can remove the room for original research, especially in
the areas of economic psychology and human behaviour.
CBA, along with mainstream microeconomics, has been increasingly criticised
for building upon an unsound theoretical base in terms of human behaviour and
the CVM has made this even more evident. That is, the microeconomic axioms of
choice make assumptions which are at best inconsistent with theories of modern
psychology and empirical evidence. In the valuation area, Kahneman and Tversky’s
(1979) prospect theory revealed that people value gains and losses asymmetrically
and this can explain the observed gap between WTP and WTA measures. Knetsch
(1994, 1995, 1999; also Knetsch and Sinden 1984) has shown the frequent
occurrence of behaviours which are inconsistent with accepted economic norms
but commonly dismissed by economic models. The refusal to make trade-offs has
been shown to arise in CVM studies both amongst those who protest against the
use of monetary valuation of the environment and also those prepared to make a
contribution (Spash 2000b, 2002a, 2002c). The results can be linked to rights-
based beliefs which contrast with those of the economic utilitarian (Spash and
Simpson 1994; Spash 2000c). In terms of the economic model, the only way in
which such behaviour can be characterised is as lexicographic preferences which
are deemed, at best, to be held only by a strange minority (Spash 1998, 2000a;
Spash et al. 2000). A range of techniques is therefore required to understand human
psychology in terms of the attitudes, ethical beliefs and social norms which motivate
behavioural responses.
Concerns raised about the practice of attaching monetary values to all
dimensions of socio-economic and biophysical systems have led to the call for
alternative valuation methods (Martinez-Alier et al. 1998). There has been a
powerful argument for shifting away from the sole focus on outcome towards the
quality of decision processes (Funtowicz and Ravetz 1990; O’Connor et al. 1998).
In particular, multiple criteria assessment has become a more widely used method,
offering the potential for the integrated assessment of local, national and inter-
national policies and as a means for combining different perspectives associated
with sustainability goals. Sustainability raises a set of issues based on civil rights of
current and future generations as well as respect for ecological systems. Under-
standing that biophysical as well as human systems are complex, and will never be
fully understood, has led to the development of approaches which favour adaptive
behaviour and learning processes over optimal solutions. Closely linked to this
view is the acknowledgement of uncertainty of future events and their impacts on
human and non-human systems. All these factors redefine the role of experts, the
meaning of knowledge and how decision processes need to be designed to make
more effective policy.
In the environmental policy arena, and elsewhere, there has been a push for
greater public participation, e.g. in Europe the Aarhus Convention (European
Commission 1998), and the inclusion of non-governmental stakeholders in project
appraisal. Researchers are meeting this challenge both through interdisciplinary
discourse and also by contributing to the formation of novel democratic insti-
tutions. There is both innovation in traditional environmental CBA and in terms
of new political institutions for addressing valuation issues. In the former case are
the range of techniques termed deliberative monetary valuation (Spash 2001),
and in the latter such tools as the citizens’ jury. Research into the operational
criteria for effective public participation is at an early stage of development.
The search for validity in applying CBA methods has lead to a variety of appeals
to interest groups and/or members of the public in an attempt to supplement the
normal information content of prices placed on the environment. Thus, the travel
cost method may be combined with an interview approach in order to sustain
assumptions of how individuals behave, value time and relate to the environment.
Hedonic pricing has sought corroboration from estate agent valuations as
representing ‘informed’ preferences. Focus groups have been used in conjunction
with the CVM to test survey design on the basis that group deliberation could
validate the information content and help identify design biases. This last approach
is most clearly where deliberative practices have begun to enter. For example, the
largest CVM study in the UK was conducted on environmental impacts associated
with aggregates (Department of the Environment, Transport and the Regions
1999). An interesting feature of this work in the current context was the informal
use of vested industrial interests (stakeholders) in the first part of the study and the
use of public focus groups in the design stage of the second, although the feedback
from the public proved problematic by diverging from economic assumptions (e.g.
the expressed desire for community compensation unrelated to the individual),
and neither process was formally reported.
Thus, two broad approaches to combining deliberation and monetary valuation
can be identified. The first regards monetary valuation as basically sound but
being able to benefit from supplementary, and often informal, processes borrowing
elements of deliberation. The second sees the use of deliberative approaches as a
new method allowing the (collective or individual) production of a monetary
valuation for environmental goods and services. Under the first approach a variety
of alternatives exists, and monetary valuation may be either followed or preceded
by some element of deliberation. Stakeholder participation, as mentioned above,
may be employed to validate outcomes. The implication being that ex post
deliberations can be used in some way to adjust valuation results or their
presentation. Deliberative processes and environmental valuation may also be
sequential, e.g. selecting a sub-sample of participants from a CVM survey for a
citizens’ jury on the same environmental issue. Ex ante deliberation has been
employed in designing CVM surveys with the use of focus groups to test the wording
and respondents’ understanding of survey questions. Deliberation is then regarded
as useful in providing insight into the processes by which respondents produce
their WTA or WTP bid. This may be extended to allowing a deliberative process
to determine the options or institutional context to be valued in the survey.
The second approach is what Spash (2001) has termed deliberative monetary
valuation (DMV) as advocated by, for example, Brown et al. (1995), Jacobs (1997),
Ward (1999) and Kenyon et al. (2001). DMV is the use of formal deliberation
concerning an environmental impact in order to express value in monetary terms
for policy purposes, and more specifically as an input to CBA. For example, consider
a proposal to build a new road through a wilderness area and so destroy the habitat
of a number of rare or threatened species. A group of citizens would be selected
and meet to discuss information about these environmental damages associated
with the development. Known costs and benefits (discounted) would be presented,
while those pertaining to environmental damages would be deliberated. The citizens
would form a jury aiming to provide a monetary value for environmental damages
which might be in terms of an individual WTA to allow the project to proceed.
The result would then be incorporated into a net present value calculation to
determine the viability of the project.
The meaning of such values remains contentious, as they are mediated individual
values. In order to address the increasingly evident fact that preferences are formed
and that there is a lack of arbitrage in valuing environmental goods and services,
economists are appealing to methods from political science. However, approaches
such as citizens juries and contingent valuation, or more generally deliberative
forums and monetary valuation, differ in fundamental ways. This has been pointed
out by Niemeyer and Spash (2001) as involving the approach taken to theoretical
factors (individual and social ontology, preference basis, rationality theory), practical
factors (justification, framing, value representation, institutional setting), and
political factors (manipulation, representation, social impact). The variation
between approaches to each of these factors means the very concept of DMV is
brought into question. In simplified terms, can DMV take the best of both monetary
and deliberative methods as advocates hope or does it merely create a messy
confusion as to the values being expressed?
Clearly, research into valuation needs to span different disciplines such as social
psychology, applied philosophy, political theory and economics. While interest in
doing so seems to have increased, the quality of the applications can be questionable
(Spash 2000b). Part of the problem is for economists to understand the requirements
of other disciplines, while practitioners in those other disciplines must similarly
have a good practical knowledge of what economists are trying to do. Thus,
applying attitudinal measures from social psychology may prove insightful with
respect to WTP, but only if researchers also understand welfare economics and
CBA tools so that their work can address the economics literature. This is a
challenging research agenda but one which, as several contributions in this volume
show, is now being taken on at several different levels.
Contributions to the debate in this book

The book is organised around three major drivers influencing the field of
environmental valuation, namely attempts at: understanding the results from
environmental CBA and the foundations of economic behaviour, taking multiple
values into account, and exploring the role for inclusive deliberation in valuation
processes. The chapters reflect the debates over method development which are
on-going in the valuation community. Clearly, all methods have advantages and
disadvantages and there is no pretence on the part of the editors that any one
approach offers a panacea. What the reader will find is the presentation of a range
of attempts to learn from different disciplines in order to achieve practical
approaches which can improve our understanding and expression of environmental
values.
Economic values and the psychology of behaviour

Getzner begins the book by exploring epistemological and other theoretical
problems in understanding economic values with a specific focus upon the CVM.
What might be regarded as specific validity issues are addressed (i.e. hypothetical
bias, framing of questions), but a different perspective is gained by placing these
issues within the context which the social role respondents adopt when answering
CVM surveys. The refusal to trade money for environmental change is raised as
an important behaviour and related to the literature on lexicographic preferences
(see for example Spash and Hanley 1995; Spash 1998; Spash et al. 2000). In addition,
some of the literature on the supposed dichotomy of individuals as political citizens
and economic consumers is discussed. This then provides the background to a
range of variables for inclusion in bid curve analysis. The empirical test presented
uses WTP for a nature preservation programme in Austria stated to be able to
secure the survival of certain species and ecosystems. A convenience sample of
189 university students form the database. The results on answers consistent with
lexicographic preferences are mixed. While 94 per cent of respondents claimed
species protection should be undertaken regardless of the cost, only just over 7 per
cent rejected WTP questions because they regarded species could not be valued in
monetary terms. This implies the need for careful probing of interviewees in order
to understand their responses (Spash 2000b). The explanatory power of the model
is greatly enhanced by what can be regarded as non-economic variables or at least
non-standard ones. These factors are interpreted by Getzner as showing the
importance of social and institutional context. A strong influence of variables
related to charitable giving is revealed, which supports claims by others that CVM
fails to measure a trade price as is assumed by practitioners (see Spash 2000b). A
clear implication of this type of work is the need to probe more deeply into the
motives behind stated preference behaviour.
Meyerhoff follows up on this challenge by focusing upon the importance of
attitudes as explanatory motives for WTP to improve environmental quality. A key
contention here has been the claim by some social psychologists that the results of
CVM are merely poor measures of attitudes. One method of testing this hypothesis
is to conduct bid curve analysis and probe which variables best explain WTP or
WTA. Meyerhoff focuses on the difference between non-users and users with the
expectation that non-users only express indirect or passive use values (sometimes
erroneously termed non-use values) and that these are related to non-economic
factors. Indeed, Westra is cited as showing the relationship of such values to
attitudes. Meyerhoff uses WTP for protection of the nature and landscape of the
German Wadden, a coastal wetlands ecosystem, against threats from climate
change. A total of 1412 semi-random interviews were conducted face to face. The
data analysis proceeds by excluding ‘protest bids’ which are zero bids for reasons
regarded as inconsistent with placing a zero value on the environmental
improvement, e.g. protesting against the use of taxes to achieve this. The practice
is highly questionable as Meyerhoff notes but he proceeds to use follow-up questions
to exclude 56 per cent of the sample. How this may have affected the results is
unclear. Among the remaining sample a higher WTP is found amongst those
accepting money as a measure of environmental value, although interestingly
rejection of money as a measure does not mean all such respondents give a zero
WTP. This is consistent with other findings where those rejecting an economic
rationale for valuing the environment (e.g. rights based) have been found to have a
positive WTP; although, in apparent contrast with the results here, WTP has also
been found to be higher amongst such groups than amongst those favouring
economic trade-off justifications for their WTP (Spash 2000a). A key finding by
Meyerhoff is that non-users’ WTP is explained by attitudinal and economic
variables while only attitudinal variables prove significant for the user group. This
would seem to imply that users have a stronger non-economic value orientation.
In discussing attitudes, Meyerhoff notes the difference between measuring general
attitudes which are related to ethical beliefs, and basic values, and those attitudes
which are specific to a behaviour. The latter category and correspondence between
the level of attitude framing and behaviour are recommended if attitudes are to
be expected to predict behavioural intentions. His measures of attitudes towards
Nature and climate change fall closer to the former category in terms of their
generality, i.e. they are related to but unspecific in terms of the WTP scenario
being valued. The results show there are clearly strong statistical relationships
between CVM results and underlying attitudes. What remains unclear is the extent
to which these motives are driven by social norms and ethical beliefs and how far
the type of environmental entities currently being subject to CVM surveys fall
outside the economic realm of value.
Mosler is also concerned with the mix of economic and non-economic motives
for human behaviour. He pursues a model of behaviour which includes social and
personal factors which he differentiates from those which are purely economic.
However, his emphasis is on how the psychology of the individual can be modelled
to explain social decisions rather than narrowly defined values. The agent based
social simulation model he develops is explained carefully along with a series of
definitions and feedback loops which aim to characterise human resource use and
environmental impacts. Amongst the factors included as explanatory of human
behaviour are attitudes, subjective norms, behavioural control, weighing costs and
benefits, and sustainability motives (combining a value and norm orientation).
The CBA component is regarded as encapsulating economic considerations via
personal returns from a given action. The sustainability motive is strongest where
an individual values the environment highly and the gap between actual and
sustainable patterns of use is small (on the basis that a large gap is demotivating
due to the difficulty of closing it). Mosler examines the conditions necessary for a
collective reorientation towards environmentally sustainable behaviour. His
simulation modelling shows the role which can be played by the environmentally
friendly behaviour of some ‘pioneer’ individuals and how policy might be developed
to increase the number of persons joining ranks with such pioneers. Sustainability
requires such an increase to build its own momentum leading to a large-scale
‘turn-around’ of previous environmentally harmful patterns of behaviour. The
aim of the simulation approach is to indicate means by which policy instruments,
such as environmental campaigns, can be implemented most effectively. One aspect
of the model is how people process information and whether they become engaged
and find an argument persuasive. This has also been raised as an issue for CVM
because central processing means deep evaluation while peripheral processing
means being led by comparatively insignificant factors. Indeed, for example, ethical
beliefs have been cited as such a peripheral factor (Ajzen et al. 1996), but empirical
evidence shows those entering central processing mode tend to rely on their ethical
beliefs as well as scientific information (Spash 2002c). As Mosler notes, there is
scope for the interaction of simulation modelling and empirical studies which would
seem essential for increasing understanding of complex systems.
Taking multiple values into account

Multiple criteria analysis (MCA) was initially developed for production planning
and methods were also applied in the field of financial management. The field has
since expanded dramatically and a great number of algorithms are now available
for weighting, comparing and combining information. Since they were developed
with different problems in mind, researchers and policy-makers in the
environmental field need some way to distinguish between the different algorithms.
De Montis, De Toro, Droste, Omann and Stagl develop a framework for the
comparison of MCA methods. They set out to identify a list of criteria referring to
issues of particular relevance for the analysis of questions concerning sustainable
development such as uncertainty, stakeholder involvement and non-substitutability
between criteria. This framework is then employed to analyse and compare seven
MCA methods frequently used for integrated assessment, namely: multiple attribute
value theory (MAUT), analytic hierarchy process (AHP), evaluation matrix
(Evamix), Electre III, Regime, Novel Approach to Imprecise Assessment and
Decision Environments (NAIADE), and Multi-Objective-Programming (MOP)/
Goal Programming (GP). The results highlight the types of issues for which the
respective methods are most suitable as well as problems that cannot be addressed
by any MCA method. Thus, MAUT requires adherence to welfare theory, NAIADE
and AHP are useful where there are value conflicts, MAUT and AHP are useful as
learning tools; where constraints are important Electre III and MOP/GP are best,
the latter also serve to address cases lacking discrete alternatives, and for ranking
alternatives MAUT, AHP, Evamix and Regime are suitable. Each approach has its
own context within which it can work best but the analyst is then responsible for
selecting the most appropriate tool and justifying that choice.
When first developed, MCA methods were meant as tools for optimising
production processes and for identifying Pareto optimal outcomes of planning
tasks. Since identifying an optimal solution is impossible for problems in complex
systems, MCA methods have more recently been combined with processes of public
and/or vested interest (stakeholder) participation. This allows the analyst to gain
legitimacy for the analysis but also adds to MCA the aspect of a learning tool, i.e.
allowing participants to explore the values and qualities of the problem.
Proctor addresses the Comprehensive Regional Assessments of Australia’s forests
which aimed to ensure sustainable management. Planners faced the task of
integrating different forest values in order to designate areas as reserves. Proctor
uses a case study, outside the official assessment, which shows how MCA can be
applied to the problem. She makes use of official assessments, criteria and priorities
placed on forest values by a stakeholder group which was involved in the assessment
of the Southern Forest Region of New South Wales. Both the preferred option
and the disparity/agreement between group members’ evaluations are explored.
Polarised opinions can thus be identified and discussions about compromises can
be supported within the MCA framework. Proctor points also to new requirements
for research processes of this kind, such as the need for a trained facilitator to aid
interactive group decision-making. The outcome shows how values come to the
fore in such processes. The two most favoured options reveal a choice which can
be summarised as between environmental quality and employment. Sensitivity
analysis can be used to show the robustness of the outcome and the failure of what
might appear to be compromise options. This shows the way in which MCA can
reveal value differences which inform policy.
Stirling and Mayer are also concerned by the approach governments take to
form regulations and implement legislation. Their aim is to explicitly address issues
of uncertainty related to the regulation of environmental and health risks. The
chapter explains the need for a set of precautionary approaches, and as a
consequence, reveals the problems associated with narrow risk assessment
techniques based on rational choice and probability theory. Stirling and Mayer
develop a set of criteria by which any ‘regulatory appraisal’ (i.e. the way in which
regulations are established) can be evaluated. The general approach is to select a
single criteria for assessment and exclude much information relevant to, say,
technological risks. Hence the criteria here concern the need for: greater humility,
completeness, consideration of net benefits in different contexts, transparency, full
engagement of interested and affected parties, mapping out different value
judgements and framing assumptions, and allowing for divergent views by
considering diverse option mixes. These requirements might sound like something
of a wish list which is itself hard to implement but Stirling and Mayer then go on
to show how such an appraisal can be performed.
An area of considerable conflict in the UK has been the genetic modification
of agricultural crops. Stirling and Mayer use this as a case study to apply a novel
approach which they call multi-criteria mapping (MCM). Twelve individual high
profile protagonists in the GM food debate were selected as participants. These
individuals came from a variety of backgrounds, including academics and
government advisers, environmental, consumer and religious organisations and
representatives from the farming, food and biotechnology industries. Results are
reported on each individual and under the categories of ‘academic’, ‘NGO’,
‘industry’ and ‘government’. Interviews were used to get each individual to add
options, appraisal criteria, apply and scale the criteria against the options, and
weight the criteria. In addition to the six ‘basic options’ a further 18 alternative
agricultural strategies were identified; 117 appraisal criteria were defined by
different participants. There are lessons for standard MCA as the authors state:
‘Participants adopt a variety of different “framing assumptions”, resulting in
significant differences in the scores assigned by different participants under the
same criteria. This has implications for conventional multi-criteria analysis, in
which scoring is often conducted by a separate body of experts, with an assumption
that different value judgements can be captured simply in the weightings.’ In
addition, the information gained on vested interest is often surprising: for example,
the biotechnology industry representatives underemphasised the social,
environmental and safety considerations which are prominent under all other
perspectives. The approach provides specific information on the motives behind
positions as well as allowing an overall appraisal of options. In the latter regard,
the organic farming and integrated pest management options tend to perform
significantly better overall.
Participation, deliberation and valuation

Aldred starts out with a comparison of the CVM with the CJ. The former is an
individualistic preference based method while the latter is a community based
form of social deliberation. The CVM and CJ answer to different institutional
needs, cultural roles and social contexts and are based upon different explanations
of rationality. This raises the question as to what extent they can in fact be compared.
Aldred argues that an assessment of which consultation process is superior for
policy formation can be undertaken on the basis of considering the added value
of discussion. He specifies the characteristics of the policy process in terms of
open and closed inputs (or task set) and outputs (or expected result). The former is
principally classified according to control, while the latter according to agreement.
The CVM has a closed input while the CJ is normally open but has the option of
considering a closed input. Aldred argues that comparison requires he match the
requirements so a CJ with closed input must be considered. In order to achieve the
same for outputs he argues the complete decision process must be considered
through to the final recommendation, i.e. a full CBA compared with the CJ best
option. However, he does allow that these requirements only apply for direct
comparison and that indirect comparison can also be made, although that is not
the aim in his chapter. Aldred draws a distinction between the role of discussion
versus deliberation in decision processes and prefers to use the former term. So his
aim is to compare closed forms of CVM and CJs on the basis that they differ in
term of the latter adding discussion prior to an aggregative decision, and his aim
is to judge whether the content of decisions is better with that discussion. He goes
on to show that arguments by participants in a CJ must appeal to widely shared
principles and he discredits the oft-cited self-interested/selfish and strategic
explanations for human behaviour in such contexts. He cites O’Neill (1993) on the
Aristotelian view which encourages self-awareness on behalf of listeners so they
recognise their inability to evaluate certain claims and instead concentrate on
assessing the character of the speaker. In general, such an approach is applied
when we judge a person’s motivation, their trustworthiness and their good
judgement. CJs are then seen as having the potential to allow listeners to assess the
credibility of speakers as they share arguments. Having negated the criticisms
arising mainly from a rational choice theory perspective, Aldred finishes by
exploring the implications of imposing a closed output on a CJ and hence moves
into the area of DMV (discussed earlier in this chapter), which claims to combine
the advantages of discussion with the merits of a valuation output. The concerns
raised here cover forcing a decision on the CJ, losing the trust of the jurors in the
process, and whether quantitative or qualitative outputs are appropriate. In
summary, where bounded rationality is likely to be severe, principles of public
interest arise, or important private information is present, CJs appear more
attractive, and while both CJs and the CVM have problems Aldred does not see
DMV as the solution. He does, however, appeal for ‘a careful, empirical, case-by-
case comparison’ to add to his abstract epistemic arguments.
Kenyon and Hanley are concerned with the empirical application and
comparison of all three methods (CJs, CVM, DMV) to the Ettrick Forest Floodplain
Restoration Project in Scotland. The arguments are firmly based upon an economic
model with, for example, CJs described as addressing ‘the information problem
better than the CVM’, i.e. addressing the fact that preferences are formed and not
merely informed during a valuation process and any information set is never
‘neutral’ (see Spash 2002c). The reason Kenyon and Hanley pursue DMV is because
CJs ‘do not provide an economic estimate of the value of any particular project,
nor whether it constitutes an efficient use of resources’. For the CVM a stratified
sample of nine towns in the Borders Region of Scotland resulted in 336 responses
to the survey including the request for a charitable donation. Of these, 29 per cent
were classified as protest bids and removed from the data set and any further
analysis. CJ participants were selected from the CVM questionnaire respondents.
Kenyon and Hanley claim that the eleven jurors were selected ‘to be representative
of the Borders population’, although what they represent is left unspecified. The
jurors were concerned that they might need to set up a trust fund to cover under-
funding (as implied by the CVM) but ‘after speaking with a member of the local
community’ were reassured that the local community was fully involved in the
project and the site would be well managed into the future. The CJ considered
environmental and social elements important in judging the success of the project
and in making management recommendations, but ‘did not seem to consider
economic criteria important’. Selection for the DMV was on the basis of responses
to a letter sent to 500 households. Two workshops were carried out in each of two
towns in the Borders, giving a total of four workshops and 44 participants. The
DMV started by administering the CVM, then took participants into small group
(four to seven people) discussions on problems and management options, and finally
asked them as individuals to answer some questions including whether they now
wanted to revise their WTP and only 14 per cent did so. DMV initial bids included
5 per cent protest bids and 34 per cent ‘don’t know’ responses, and only two people
moved from the latter after the discussion stage. An advantage of the DMV over
the CVM was seen to be the information on positive and negative views of the
project. When an aggregated WTP figure was presented to DMV participants
there were three responses: the impossibility of putting a value on such a project,
the poor economic situation meaning the need to spend money elsewhere, and
wanting some other public fund to pay. Overall the authors conclude in favour of
DMV as a middle path.
James and Blamey concentrate on a DMV approach, which they describe as a
CJ with the added task of determining societal WTP for a specified programme
involving environmental improvement. They open by addressing problems and
issues relating to the theory and application of both the CVM and deliberative
approaches. This aims to show the potential of a DMV approach. The empirical
case study is the management of national parks in New South Wales, Australia. As
in the chapter by Kenyon and Hanley, there was a concern that the participants be
representative and in this case the criteria are given (gender, age, place of residence,
ranking of the environment in relation to other social issues, occupation, income,
income source and education). Such representation is a little strange in these studies
as the only aim can be statistical significance but the sample sizes are so small as to
make this irrelevant (for elaboration on the difference between political and
statistical representation, see O’Neill 2001). There were thirteen participants (and
one no show). The DMV panel was given three options developed by the researchers
and a fourth developed by the researchers on the basis of answers given to a ‘straw
preference poll’ (although what this involved remains unclear). Participants were
limited to considering only these options. This approach can be seen as Aldred’s
closed input and closed output case, making the DMV comparable (in those terms)
to the CVM. After individual consideration of three options the panel was convened
and the choices discussed. The aim of the researchers seems to have been for a
consensus report and the panel gave a preferred option with qualifications covering
the concerns of those who were initially against that option. Next, the valuation
question was introduced as part of the fourth option. The task for the panel was to
determine how high a park ‘levy’ would have to be before the NSW public would
be no better off than under the status quo. That is, they were being asked to set the
average annual WTP, rather than their own maximum WTP as in a CVM survey.
Perhaps surprisingly, the panel was able to come up with two amounts and voted
to decide which amount was to be recommended. The use of majority voting
occurred at several stages in the discussions in order to close down dissent. In their
conclusions the authors discuss a question which remains open: how to interpret
the value obtained. The amount fails to relate to economic welfare theory and
would seem hard to compare with other microeconomic welfare theoretic measures.
At the same time, the thirteen panellists cannot claim to be statistically representa-
tive, which only leaves them with stating a political position as might be done
more directly. The authors also raise several other issues including consensus
formation, decision rules for polling or voting, equality of juror impact, and
provision of information.
Arzt picks up some of the same themes in addressing whether researchers should
aim to reveal existing preferences or initiate and foster learning processes and
thereby help form preferences. The contention is that environmental valuation is
a subjective, cultural and contextual phenomenon. The approach for addressing
policy needs here is to appeal to vested interest group representation in a framework
termed an ‘interactive valuation process’, where individuals, in groups, construct
values rather than express prior preferences. This approach is equated to a
roundtable and cited as deriving from Ostrom’s (1998) ‘second generation models
of rationality for social dilemma situations’ to understand when co-operation in
small groups can take place. Much is made of the need to build trust over time
through repeated meetings, although this would seem to contrast significantly with
the CJs used in other chapters of this volume. In this respect, the motives for
behaviour described by Arzt are worth comparing with those discussed by Aldred.
She also describes the roundtable as a consensus seeking exercise in which decisions
are made without voting or other mechanisms, and this can be contrasted with the
way in which James and Blamey ran their DMV panel. In this roundtable consensus
seeking, some tools and techniques are advocated including MCA, visualisation
techniques and professional moderation.
The case study concerned a roundtable run in the Uckermark, Brandenburg
county, region of northeast Germany with the topic of decentralising decisions
about agri-environmental measures. Agri-environmental schemes are already
run from the local government level (as opposed to, say, those in the UK) but
decisions on targets are made by experts without consultation with farmers or
conservationists. A community meeting was used to select the roundtable with
43 people out of 100 attending the meeting and 23 volunteering. Six meetings
were held on a monthly basis with varying participation rates. The inter-
disciplinary research team from different institutions found their own problems
before even starting the roundtable: ‘The different backgrounds of the project
team made it difficult to find a general understanding of the definition of “value”
and “learning processes” beforehand. The team could not create a common
understanding about the institutional arrangement’. The aim had been to cover
a range of five issues in sequence: reasons for the problem and expected long-
term consequences, solutions including justifications, measures to achieve these
solutions including monitoring, financing and action planning. However, after
three meetings, the main interest of the stakeholders was still focused on measures
and their financing. One farmer is reported as stating: ‘I would do anything, if I
got enough cash.’ Attempts to discuss scientific uncertainties over a specific type
of water body (Sölle) management, at the third meeting, also failed as the farmers/
land managers expected the expert scientists to tell them what was best to do.
Only when the next (fourth) meeting followed a very structured approach to
management of these water bodies was some headway achieved. The final two
meetings switched to soil erosion: the fifth meeting was dominated by a soil expert
and the sixth, where only three farmers attended, by the agricultural adminis-
tration. The difficulty of running this process compared with the high expectations
of the theoretical literature raises a range of good lessons for research and
practice. Power is an important element and affected representation with small
farmers unwilling to attend knowing that ‘consensus’ processes in the past were
dominated by large cooperatives, and women being a minority and silent. This
has obvious parallels with the ‘willingness to say’ problems as raised by O’Neill
and discussed by Aldred, and this issue also arose directly with experts dominating
some meetings. Representation problems also arose in a different form with one
organisation alternating their representatives but with the two individuals holding
very different opinions. The need for greater structure in the process might be
met by use of an MCA type of process such as the MCM approach developed
by Stirling and Mayer. However, there also may be some cultural differences
arising in this study (e.g. desire for structure and being told the best option, high
esteem of scientists), so that how far the design features can be determined in
any one country-specific study becomes a relevant research question. Indeed
more generally the extent to which the benefits of participation and deliberation
can be regarded as universal remains unanswered.
Niemeyer shows how we might begin to address this last point with respect to
CJs. Political science advocates of CJs have seen them as deliberative forums in
which people can find truths and transform themselves away from more mundane
daily preoccupations. Whether CJs are a process which can lead to the change of
fundamental attitudes and preferences in relation to environmental policies is then
a testable hypothesis. Niemeyer contends that preferences are shaped during
deliberation by information, attitude and process. The chapter shows how to
measure both changes to (rank ordered) preferences and the underlying subjectivity
for those preferences (using Q method). The case study concerns a controversial
‘road’, the Bloomfield Track, constructed by politically influential developers
through a rainforest in north Queensland, Australia, against the protests of
conservationists. The area was later, after a change of regional government,
designated the Wet Tropics World Heritage Area. What action should be taken
with respect to the route remains a problem. Local and Aboriginal communities
now use the track as a convenient means of communication while environmen-
talists regard it as a scar across a World Heritage site causing erosion and damage
to forest and coastal reef ecosystems. A random-stratified sample of twelve people
from the region was given the task of considering management options for the
Track over a four-day period. Attitude information was obtained by surveying
participants at the outset of deliberations, at the midway point, and immediately
after proceedings. The combined data for all three Q sorts by the jurors produced
four distinct factors: Preservation, Optimism, Pragmatism and Symbolism. Factorial
analysis revealed four preference options: Close Road, Stabilize, Bituminize and
Status Quo. This approach allows analysis of motive, and examples of correlations
are Preservation with Close Road and Pragmatism with Status Quo. Preferences
with regard to these options changed substantially during deliberation. Two jurors
experienced a near complete reversal of their original orderings, and preferences
among jurors converged, although they never reached consensus. Jurors were
encouraged by the process to take a community perspective, rather than that of
individual self interest. By the end of deliberation road closure had become the
most preferred option. As Niemeyer notes, this CJ was a success in terms of the
expectations from political theory but it is also only one example and others have
had mixed results, as did Arzt in this volume. He concludes that the outcome of
deliberative forums is highly sensitive to context and design, and a wider
transformation of society towards deliberative democracy is required but that such
institutional designs can help with that transformation.
Drawing out some lessons
There are significant gains to be made by further use of participatory approaches
in environmental management, but maintaining the interest, involvement and
commitment of civil society may prove a major challenge. Enthusiasm on behalf
of researchers for new approaches needs to be qualified by awareness of the
difficulties in applying techniques in different contexts and their potential for both
failure and manipulation. At the same time consumer boycotts and support for
non-violent direct action (e.g. anti-road demonstrations) do indicate public
dissatisfaction with current institutions and a demand for greater say in policy, as
opposed to an infrequent vote for a possible representative. The concern for the
political legitimacy of government policy is particularly relevant in societies with
low voter turn-out or where systems fail to represent minority parties. Both the
UK and the USA, for example, have systems which increasingly seem inadequate
as far as voters are concerned. This general problem takes on specific significance
in the formation of environmental policy where the directly concerned political
parties are small and lack representation via established institutions. Thus, for
example, under the UK system, which has no proportional representation at the
national level, a substantial minority of votes for the Green Party has in the past
been unable to gain even one member in parliament. At the same time redressing
the political imbalance through new institutions is no easy task.
In this volume contributors show the difficulty in getting politically weak groups
to voice their opinions in forums which are hoping to empower them. Various
features of representative democracy have been raised by O’Neill (2001). In
particular, environmental policy brings to the fore the question of representation
for the silent voices of non-humans and future generations. Another aspect of
representation is how information is delivered through the interaction of science
and policy. The role of expert and ‘objective’ scientific information to inform
policy on uncertain and complex environmental problems has been brought into
question on issues ranging from nuclear power to genetically modified crops. That
there is a problem with the traditional approach to environmental policy processes
is apparent from the greater attention given in recent years to a range of alternative
institutional approaches (Royal Commission on Environmental Pollution 1998).
However, while several techniques have been increasingly deployed they also have
their own limitations. The use of focus groups in the UK was seen as providing
some additional legitimacy for policy, although now such an approach often receives
popular derision as a marketing technique which has been overused. Consensus
conferences have been employed in several European countries but consensus is
itself a questionable goal when facing fundamental differences in perspective and
values. There are problems with closing down a debate in an effort to achieve an
artificially stable outcome described as consensus. Instead there is a need for
mapping out the contours of the debate via an approach such as that suggested by
Stirling and Mayer in this volume.
The use of ‘stakeholders’ to represent vested interests directly in the decision-
process has entered into both policy and research with the emphasis on addressing
the needs of ‘end-users’. Both CJs and the appeal to vested interests raise similar
questions as to who should be represented and how they should be represented.
CJs have become more common as well as being an increasing focus of research.
Some environmental economists have also become increasingly concerned to
employ such approaches to increase the legitimacy of their valuation work.
The rise of DMV is a significant development as an indicator of the desire for
greater legitimacy for the values being derived in economic assessments. The reasons
driving this move are recognised inadequacies in the economic model of human
behaviour. There is now far greater acceptance by economists that psychological
motives are important and preferences are often constructed in response to research
aiming to discover how people value the environment. The problem which then
arises is how far the values being derived from DMV remain valid having been
removed from their theoretical basis in economic welfare theory. If values as derived
by Blamey and James are to be given weight in decision processes, there is a need
for a new theoretical underpinning. At the same time, social psychology raises
many insights which economists are only just beginning to discover. The work on
CVM has exposed problems with standard approaches and data interpretation.
Yet there is still a surprising readiness amongst researchers to throw away large
parts of their samples for analytical and/or practical convenience. There are two
examples in the current volume where protest bids are excluded, removing 29 per
cent of the sample in one case and 56 per cent of the sample in another.
Rather than excluding such information, researchers need to investigate and
understand what individuals are telling them. Perhaps the technique is failing but
with CVM what seems to be happening is that the economic model is being revealed
as only a limited frame in which humans are prepared to operate. Some of the
people may be homo œconomicus all of the time, all of the people may be homo
œconomicus some of the time, but all of the people are never homo œconomicus all of
the time. In environmental valuation, rejection of the economic motive for
behaviour is common and this cannot be disregarded purely on pragmatic grounds.
In the past, the counter to critiques of CBA has often been that ‘there is no
option’ and that ‘because society uses a money metric, so must environmentalists’.
The treasury departments of government are certainly often the strongest and lie
behind many decisions, but there are also other branches of government and civil
society, and different forums in which decision processes operate. This volume
shows there are alternatives and these can be most rewarding in terms of gaining
insight into value conflicts, while aiding and improving policy. There is a need to
evaluate the opinions of experts as one perspective on the implications of any
project, and set their role within a broader social context where a range of motives
operate and different values are expressed with equal legitimacy. However, there is
no pretence that alternatives to traditional environmental valuation lack their own
problems and can be difficult to implement successfully. Clearly, more work is
required which breaks down the boundaries between disciplines and more
researchers are required who are prepared to venture out of the comfort zones of
mono-disciplinarity into the challenging areas of interdisciplinarity. Ecological
economics is one such attempt at bringing together diverse perspectives and
approaches.
Note
1 Revenue raising is a key objective of taxation rather than designing environmentally
efficient and effective taxes. This means taxes which may be described as environmental
in fact fail to address the pollution problem to which they are supposedly related. For
example, a tax justified as reducing the Greenhouse Effect may select carbon dioxide
emissions regardless of the other gases needing control and the tax may then relate
only to industrial emissions from point sources, ignoring a whole range of non-point,
household and public sources, or select only one specific fuel. Such decisions are
taken on the basis of ease of administration and revenue-raising potential rather
than their economic efficiency at controlling pollution and matching marginal social
costs of environmental use.
References
Ajzen, I., T.C. Brown and L.H. Rosenthal (1996) ‘Information bias in contingent valuation:
effects of personal relevance, quality of information and motivational orientation’,
Journal of Environmental Economics and Management 30(1): 43–57.
Brown, T.C., G.L. Peterson et al. (1995) ‘The values jury to aid natural resource decisions’,
Land Economics 71(2): 250–60.
Burney, J. (2000) ‘Is valuing Nature contributing to policy development’, Environmental Values
9(4): 511–20.
Department of the Environment, Transport and the Regions (1999) The Environmental Costs
and Benefits of the Supply of Aggregates: Phase 2, London, Department of the Environment
Transport and the Regions: 208.
European Commission (1998) Aarhus Convention on Access to Information, Public Participation in
Decision Making and Access to Justice in Environmental Matters, Brussels.
Funtowicz, S.O. and J.R. Ravetz (1990) Uncertainty and Quality in Science for Policy, Dordrecht,
Kluwer Academic Publishers.
Hanley, N. and C.L. Spash (1993) Cost–Benefit Analysis and the Environment, Aldershot, Edward
Elgar.
Jacobs, M. (1997) ‘Environmental valuation, deliberative democracy and public decision-
making institutions’, in J. Foster (ed.) Valuing Nature? Economics, Ethics and Environment,
London, Routledge: 211–31.
Kahneman, D. and A. Tversky (1979) ‘Prospect theory: an analysis of decision under
risk’, Econometrica 47(2): 263–91.
Kapp, K.W. (1950) The Social Costs of Private Enterprise, New York, Shocken.
Kenyon, W., C. Nevin and N. Hanley (2001) ‘Citizens’ juries: an aid to environmental
valuation?’ Environment and Planning C 19: 557–66.
Knetsch, J.L. (1994) ‘Environmental valuation: some problems of wrong questions and
misleading answers’, Environmental Values 3(4): 351–68.
Knetsch, J.L. (1995) ‘Assumptions, behavioral findings, and policy analysis’, Journal of Policy
Analysis and Management 14(1): 68–78.
Knetsch, J.L. (1999) ‘Environmental valuations and standard theory: behavioural findings,
context dependence, and implications’, in T. Tietenberg and H. Folmer (eds) The
International Yearbook of Environmental and Resource Economics 2000/2001, Cheltenham,
Edward Elgar: 41.
Knetsch, J.L. and J.A. Sinden (1984) ‘Willingness to pay and compensation demanded:
experimental evidence of an unexpected disparity in measures of value’, Quarterly Journal
of Economics 99(3): 507–21.
Martinez-Alier, J., G. Munda and J. O’Neill (1998) ‘Weak comparability of values as a
foundation for ecological economics’, Ecological Economics 26(3): 277–86.
Niemeyer, S. and C.L. Spash (2001) ‘Environmental valuation analysis, public deliberation
and their pragmatic syntheses: a critical appraisal’, Environment & Planning C: Government
& Policy 19(4): 567–86.
NOAA (1993) ‘Natural resource damage assessment under the Oil Pollution Act of 1990’,
Federal Register 58(10): 4601–14.
O’Connor, M., S. Funtowicz, F. Agliera-Klink, C.L. Spash and A. Holland (1998) Valuation
for Sustainable Environments: The VALSE Project Full Final Report. Ispra, European
Commission, Joint Research Centre: 395.
O’Neill, J. (1993) Ecology, Policy and Politics: Human Well-Being and the Natural World, London,
Routledge.
O’Neill, J. (2001) ‘Representation’, Government and Policy 9(4).
Ostrom, E. (1998) ‘A behavioral approach to the rational choice theory of collective action’,
presidential address, American Political Science Association. American Political Science
Review 92(1): 1–22.
Royal Commission on Environmental Pollution (1998) Setting Environmental Standards,
London, Her Majesty’s Stationery Office: 232.
Spash, C.L. (1998) ‘Investigating individual motives for environmental action: lexicographic
preferences, beliefs and attitudes’, in J. Lemons, L. Westra and R. Goodland (eds)
Ecological Sustainability and Integrity: Concepts and Approaches, Dordrecht, Kluwer Academic.
Spash, C.L. (2000a) ‘Ecosystems, contingent valuation and ethics: the case of wetlands re-
creation’, Ecological Economics 34(2): 195–215.
Spash, C.L. (2000b) ‘Ethical motives and charitable contributions in contingent valuation:
empirical evidence from social psychology and economics’, Environmental Values 9(4):
453–79.
Spash, C.L. (2000c) ‘Multiple value expression in contingent valuation: economics and
ethics’, Environmental Science & Technology 34(8): 1433–8.
Spash, C.L. (2001) ‘Deliberative Monetary Valuation’. 5th Nordic Environmental Research
Conference, University of Aarhus, Denmark.
Spash, C.L. (2002a) ‘Empirical signs of ethical concern in economic valuation of the
environment’, in D.W. Bromley and J. Paavola (eds) Economics, Ethics, and Environmental
Policy: Contested Choices, Oxford, Blackwell: 205–21.
Spash, C.L. (2002b) Greenhouse Economics: Value and Ethics, London, Routledge.
Spash, C.L. (2002c) ‘Informing and forming preferences in environmental valuation: coral
reef biodiversity’, Journal of Economic Psychology 23(5): 665–87.
Spash, C.L. and C. Carter (2002) ‘Environmental valuation methods in rural resource
management’, in F. Brouwer and J. van der Straaten (eds) Nature and Agricultural Policy in
the European Union: New Perspectives on Policies that Shape the European Countryside, Cheltenham,
Edward Elgar: 88–114.
Spash, C.L. and N. Hanley (1995) ‘Preferences, information and biodiversity preservation’,
Ecological Economics 12(3): 191–208.
Spash, C.L. and I.A. Simpson (1994) ‘Utilitarian and rights-based alternatives for protecting
Sites of Special Scientific Interest’, Journal of Agricultural Economics 45(1): 15–26.
Spash, C.L., J. van der Werff ten Bosch, S. Westmacott and J. Ruitenbeek (2000)
‘Lexicographic preferences and the contingent valuation of coral reef biodiversity in
Curaçao and Jamaica’, in K. Gustavson, R.M. Huber and J. Ruitenbeek (eds) Integrated
Coastal Zone Management of Coral Reefs: Decision Support Modeling, Washington, DC, World
Bank: 97–118.
Ward, H. (1999) ‘Citizens’ juries and valuing the environment: a proposal’, Environmental
Politics 8(2): 75–96.
A framework for valuing nature 21
Part I
Extending the
environmental valuation
approach
22 Michael Getzner
2 A framework for valuing

nature
Regional biodiversity
Michael Getzner
Introduction
The contingent valuation method (CVM) seems to divide the economics profession.
While there might be higher agreement among a core of (neoclassical)
environmental economists on the theoretical foundations of valuing public goods
with that specific method, applying the CVM is an important point of scientific
discourse. Ecological economists not only doubt empirical applications of the CVM
in many respects, but also question its welfare theoretic foundations and more
generally those of cost–benefit analysis (CBA). A crucial point in that debate is the
understanding of markets as social institutions, and the behaviour of individuals
as embedded in a specific social and institutional context.
An important element in that debate is also the critique of the concept of Homo
oeconomicus who is assumed to have well-behaved, stable preferences, which are
‘home-grown’ and can be elicited by means of CVM surveys – provided several
guidelines are followed (e.g. Arrow et al. 1993; Department of the Interior 1994).
An aggregation of these individual values is said to point to the (societal) ‘value’ of
a natural good, and compared with the costs of protecting the respective natural
good, it is assumed that society’s welfare is in principle not much more than the
sum of individual welfare aggregated in a social welfare function.
Underlying preconditions for this approach are the notion that individuals are
in principle willing to trade money for natural goods. That means that even in the
case when individuals hold ethical beliefs or commitments often associated with
natural goods, they can be ‘bribed’ to give up their beliefs for money. It has to be
assumed that individuals who state a willingness to pay (WTP) base their decisions
on similar foundations and considerations regardless of the nature and the
characteristics of the good in question. Thus, using the CVM in a CBA for societal
decisions assumes that individuals do not take different viewpoints when valuing
private or public (natural) goods because contingent valuation puts individuals in
a situation where they decide on public goods in a market-like setting. Another
element in this debate is whether the hypothetical bias can have such an influence
that results of CVM studies are fundamentally flawed. The main argument is that
people behave differently when actually forced to ‘realize’ their stated WTP.
24 Michael Getzner
The ‘standard’ and ‘extended’ model of valuation
Figure 2.1 shows the ‘standard’ model of valuation and an ‘extended’ model that
includes a number of additionally influential factors regarding an individual’s stated
behaviour (stated WTP). Normally, studies explaining WTP in a bid function try
to relate a number of individual factors (socio-economic characteristics like age
and income, knowledge, perception of environmental quality) to respondents’ WTP
bids.1 The explanatory power of such models is often very low (Mitchell and Carson
1989, argue that models should at least explain 15 per cent of the dependent
variable’s variation). A low explanatory power might indicate omitted variables
bias indicating that those individual variables chosen to explain stated behaviour
are not the variables which fully influence a respondent’s stated WTP. Following
that argument, other variables might be able to reasonably explain stated behaviour
with a higher degree of accuracy. Recently, a number of authors have stressed the
importance of the social and institutional context of decisions, especially when
these decisions refer to public goods (e.g. Hodgson 1997; Hayden 1993; see with
reference to bathing water quality: Georgiou et al. 1998).
An extension of the ‘standard’ model of valuation2 to account for additional
factors is presented in the lower part of Figure 2.1. First, extending individual
factors may play a vital part in a valuation survey when respondents are not willing
to trade natural goods for money (lexicographic preferences, see this section). The
institutional context of the survey is on the one hand, of course, set up by the
payment vehicle and the valuation questions themselves.3 On the other hand,
respondents’ beliefs about and perceptions of the CVM instrument may alter their
behaviour (next section). The social context comprises both beliefs about the referen-
dum and about assumptions on others’ contributions, and the consumer–citizen
dichotomy (third section). Figure 2.1 also shows the labels of variables that
operationalize the different additional factors (to be used in the econometric model
in the empirical part of this chapter).
The central questions for which answers are sought by means of the empirical
survey in the third section are: how important is the notion of lexicographic prefer-
ences and what is the effect of the respondents’ beliefs about institutional and
social context variables in explaining WTP bids for a regional biodiversity
programme in Carinthia (Austria)?
The structure of this chapter is as follows. First, the elements of the ‘extended’
model of valuation are briefly discussed, presenting a focused literature review of
the specific issues which build the basis of the empirical study presented in the
second part of the chapter. Second, results of a valuation study designed to include
‘extended’ factors into the bid function are presented.
The extended context of valuation: some important issues
Lexicographic preferences and the willingness to exchange

When it comes to valuing biodiversity4 in economic (monetary) terms, the issue of
substitutability is an important aspect: economic valuation becomes much more
Individual factors
• Socioeconomic characteristics
(e.g. ‘Age’, ‘Gender’, ‘Income’, Reasoning and processing Stated behaviour
‘Children’) • Bounded rationality • WTP in contingent markets
• Knowledge, perceptions and • Risk aversion • WTA
behaviour • Other stated preferences
(e.g. ‘Biodiversity’, ‘Member’,
‘Donations’)
The ‘standard ’ model of explaining WTP bids

Extending the model
Additional individual factors
• Lexicographic preferences
Elements of the institutional Elements of the social

context context
• Individual beliefs on setting and • Individual beliefs about
instrument and epistemological the context of the referendum,
issues personal contribution above/
(e.g. ‘Protest’, ‘Influence’) below average (‘Referendum’,
• Hypothetical referendum ‘Average’, ‘Majority’)
(‘Cheap’) • Attitudes to public goods
(‘Public’, ‘Minority’)
• Consumer vs citizen
(‘Con_cit’)
A framework for valuing nature
Figure 2.1 The ‘standard’ and ‘extended’ model of valuation

Note: For a detailed description of variables see Table 2.3 on page 38.
25
26 Michael Getzner
simple if the natural good ‘biodiversity’ can simply be exchanged for man-made
capital (private goods, money). Up to now there has been much discussion of the
possibilities of substituting environmental goods or single characteristics of these
goods such as biodiversity.5 On the one hand, ecologists argue that ‘[b]iodiversity
is a resource for which there is absolutely no substitute; its loss is irreversible on
any time scale of interest to society’ (Ehrlich and Ehrlich 1992: 225). In contrast to
this argument, Solow (1993: 181) finds that
history tells us an important fact, namely, that goods and services can be
substituted for one another. … There is no specific object that the goal of
sustainability, the obligation of sustainability, requires us to leave untouched.
… Sustainability doesn’t require … that any particular species of fish or any
particular tract of forest be preserved’ (emphasis in the original).
While there might be different approaches to the substitutability issue in natural

science, it is crucially important to separate findings of natural science from
viewpoints (perceptions) and preferences held by respondents in a CVM survey.6
The CVM is part of a social science research method focusing on people’s (as
economic subjects) preferences and perceptions in a very specific manner (Luhmann
1986).
Elicitation of monetary values in contingent valuation surveys requires well-
defined utility functions on the good in question as well as fulfilment of the basic
welfare postulates of CBA. According to that foundation there must exist a utility
function Ui = U(Xj, E) which includes Xj (a vector of marketed, private goods)
and E (a vector of public goods, e.g. biodiversity). The WTP of an individual i to
prevent a decrease from Ek1 to Ek2 is given by WTPi = e(pi, Ek1, U1) – e(pi, Ek2, U1)
with e(.) being the respective expenditure function (e.g. Spash and Hanley 1995:
193). U1 denotes the utility of i before the quality change of Ek took place.
However, in the context of environmental valuation, respondents often exhibit
behaviour corresponding to the opinion ‘that some preferences are superior to
others’ (Edwards 1992: 121). Such preferences are often formed on the basis of
ethical judgements or commitments where a ‘right’ or a ‘wrong’ exists, but where
a willingness to trade at the margin does not exist. Such preferences are labelled
‘non-compensatory’ preferences (Lockwood 1997; Gowdy 1996; Hanley et al.
1995). Such preferences can be subsumed under the broader concept of
lexicographic preferences, which have quite a long history (e.g. Jeffrey 1974; Sen
1977; Menger 1871; Peukert 1998; de Dios 1995; Encarnación 1990; Stork and
Viaene 1992). The neoclassical model of exchange rests on the assumption of
the willingness to trade (willingness to exchange, ‘model of pure exchange’;
Gowdy and Olson 1994: 162). A conceivable trade-off, which is nothing more
than the well-known indifference curve between goods, must exist in the mind
of respondents being asked a CVM question.7 While the ‘axiomatic development
of economic utility theory does not a priori exclude any motive or any good from
an agent’s utility function’ (Carson 1998: 18; Johansson 1992; Aldred 1994),
the willingness to trade is at the core of contingent valuation. Lexicographic
preferences imply a hierarchical ordering of wants and therefore unwillingness
to trade a specific good at the margin (e.g. environmental goods such as
biodiversity) for private consumer goods.
There is some empirical evidence from psychological research indicating that
decisions are often made by lexicographic ordering (ranking) rather than by
matching costs and benefits as usually assumed by economists. While observed
preferences are actually constructed in the elicitation process in a context-sensitive
way, the prominence hypothesis is often strengthened by empirical research (Tversky et
al. 1988). That hypothesis holds that choice is expected to be more lexicographic
than trading (matching). Individual decisions on biodiversity or on the survival of
single species can be considered as a typical form of choice (no matching or trade-
offs will be made) if preferences are non-compensatory as outlined above (e.g.
Spash and Hanley 1995).
Interestingly, the critique of CBA made by Boulding (1969) in part comprises
what nowadays could be called lexicographic preferences. While Boulding in
principle acknowledges the role and importance of CBA in the evaluation of social
choice and even social institutions, he admits that two fundamental problems arise
out of the concept of CBA: one which he rejects is the inclusion of ‘benevolence’
and ‘malevolence’ in CBA (selfishness, altruism) which can in principle be included
in interdependent utility functions. What might be ‘fundamental and harder to
repulse’ is the notion of a ‘heroic ethic’ (Boulding 1969: 9), which cannot be included
in CBA because a typical trade-off (and willingness to trade) does not exist. While
Boulding only sees three main fields of this ‘heroic ethic’ (the military, religion,
sports), he admits that the picture of economic man painted by economists is a
caricature because humans also hold ethical beliefs. In those cases, trade is thought
to be ‘somehow dirty’, and the ‘principle of prostitution’ would apply ‘to virtually
all areas of human life’ (Boulding 1969: 10).
Ethical commitments and rights-based beliefs towards the environment and
other species are often seen as origins of lexicographic preferences (Common et al.
1997; Boyce et al. 1992).8 Sagoff puts such preferences in the light of Kantian
ethics (Sagoff 1998). Moreover, Vatn and Bromley write: ‘Commitment and moral
judgements are concepts often attached to those domains where life, quality of
life, and personal integrity are at stake. These are areas where social norms restrict
or reject the commodity fiction’ (Vatn and Bromley 1994: 135).
Recently, some empirical evidence has been collected on the extent to which
respondents in contingent valuation studies show lexicographic preferences.9
Stevens et al. (1991: 398) concluded that, when valuing wildlife, a significant
proportion of respondents gave answers either consistent with lexicographic
preference orderings or inconsistent with both neoclassical and lexicographic
models of behaviour. Spash and Hanley (1995) find – depending on the base
line and definition of lexicographic preferences – a large proportion of
respondents (about one quarter) exhibiting rights-based beliefs on which
lexicographic preferences may be founded. In contrast, Lockwood (1998: 79),
constructing preference maps, shows that only a small percentage of participants
had lexicographic preferences.10
28 Michael Getzner
The institutional context of CVM surveys
Hypothetical bias is considered as a major threat to the reliability and validity of
the concept of stated preferences in constructed hypothetical markets in which
respondents are invited to express their values for different states of the environment
(see recent reviews of Schulze et al. 1996: 97ff.; Jakobsson and Dragun 1996: 84ff.;
Getzner 2000). However, there is mixed evidence on the question to what extent
people behave differently in hypothetical situations compared with real economic
commitments. Critics argue that even if referenda are made as realistic as possible
(e.g. by conservative wording of the questionnaire or incentive-compatible auction
designs), ‘there does not appear to be a close substitute for The Real Thing’
(Cummings et al. 1995a; see also the discussion in Blackburn et al. 1994, who deliver
methods for ‘correcting’ results from surveys to account for hypothetical bias if
the distribution of error due to the bias is systematic).
Cummings et al. (1995b) and Harrison (1996) report a number of approaches
to reduce the hypothetical bias in CVM surveys by directly addressing this problem
in the entire survey situation. They designed experiments that compared a real
referendum involving subjects making real economic commitments with two
hypothetical referenda in which hypothetical bias was taken into account by
presenting adequate information on hypothetical bias to respondents. They called
their experiments ‘light cheap talk’ and ‘heavy cheap talk’ which presented
information to respondents in various densities, e.g. by presenting written and oral
text as well as charts (‘cheap talk’ refers to additional information which can be
provided to respondents without high cost and effort). The aim of including cheap
talk information in the survey situation is to make respondents directly aware of
the hypothetical situation in which they act and in which the whole valuation
procedures takes place. Cummings et al. (1995b) report that – compared with a
‘normal’ CVM questionnaire without any additional information on the possible
hypothetical bias of such surveys – cheap talk can reduce WTP (thus hypothetical
bias) by 21 per cent compared with a reduction of WTP of 23 per cent in an
actual (real) referendum. They conclude that ‘there exists one set of words which
will result in a hypothetical survey generating the same results as a comparable
survey eliciting real economic commitments’ (Cummings et al. 1995b: 13). These
results indicate that the embedding of CVM questions in a realistic context is
crucial to the reliability of CVM results. In the empirical part of this chapter (next
section), the possible differences of a ‘cheap talk’ sample compared with an ordinary
sample are statistically explored (Harrison 1996).
As Boulding (1969: 3) noted a while ago, economics as a science tries to
obtain knowledge about a system by changing its inputs and outputs of

information, but these inputs and outputs will change the system itself, and
under some circumstances they may change it radically. […] [The]
generalised Heisenberg principle predominates because knowledge of the
social sciences is an essential part of the social sciences itself. […] As science
develops it no longer merely investigates the world; it creates the world which
it is investigating.
Boulding clearly makes this point when he describes the epistemological
problems arising out of a science that examines those facts entirely produced by it.
In our context, respondents in a CVM survey might never have thought about
valuing natural goods in a monetary way. Economics in this sense extends the
fields of research to social relations that have not been considered as economically
significant by those being asked a valuation question. CVM questions may produce
values which were not formed before or which the respondent was not aware of.
Such questions might thus provide incentives for formulating preferences rather
than merely eliciting existing and stable preferences that were formed in the past.
Applying Boulding’s critical comments on CVM surveys, one may feel curious
about the attempts to elicit ‘home-grown values’ in CVM surveys. It is often assumed
that respondents a priori have ‘home-grown values’ for biodiversity in a monetary
form. It can be doubted whether respondents taking part in a CVM survey have
generally ever thought about valuing biodiversity in a monetary form (Bishop et al.
1983). Respondents therefore construct their value statements during the CVM
interview (Goodman et al. 1999). The very form of valuation questions – which
are worded by the researcher with the theory of Homo oeconomicus in the
background – pushes respondents into a certain role (see next section). These
theoretical problems have often been addressed. For instance, in the context of
strategic behaviour which has been considered as crucial for valuing public goods
due to the supposed free-rider behaviour, it turns out that many assumptions made
by economists and expressed by valuation questions do not meet respondents’
actual attitudes, perceptions and behaviour (see e.g. Burgess et al. 1998; Marwell
and Ames 1981).
There are a number of attempts to overcome possibly serious problems with
value construction in CVM surveys. Lazo et al. (1992) present a methodology for
CVM questionnaires in order to design a shortened version, which can efficiently
be administered to a large sample without loss of information. Thus, value
construction is not considered as a major problem of the CVM methodology by
some researchers but something that needs to be taken into account when designing
surveys (McDaniels and Roessler 1998).
In the case of CVM surveys eliciting monetary values for natural goods like
biodiversity, it is of great interest to determine to what extent respondents are used
to, and influenced by, such questions. Besides the fact that some respondents merely
dislike valuation questions, the empirical results presented below concentrate on the
influence of CVM questions (surveys) on respondents’ perceptions of natural goods
as goods to be further protected. If the influence of the valuation question and
information presented is overwhelming (e.g. if reality is solely constructed in the
valuation process), this can nevertheless be a major problem for CVM studies (see
discussions in Hanley et al. 1995; Holm-Müller et al. 1991; Bishop and Welsh 1992).
Closely related to the epistemological questions dealt with above, one can be
curious if economists or economics students are ‘different’, i.e. if they perceive
valuation questions in ways different to those of the general public, or if they
express more or less solidarity or altruism than students of other disciplines (see
the discussion in the Journal of Economic Perspectives, starting with Carter and Irons
30 Michael Getzner
1991). In the study described in this chapter, economics students and students of
other disciplines are compared in order to examine the influence of studying
economics (and business administration) on the amount of WTP bids. Eventual
differences are taken as an auxiliary measure for the possibly influential effect of
learning the economic view.
The social context: ‘consumer’ vs. ‘citizen’

As mentioned above, lexicographic preferences can constitute an important
argument when it comes to a monetary valuation of natural goods like
biodiversity. One reason for such preferences may lie in the individual view that
there is only one ‘right’ or ‘wrong’. However, the divergence between two roles
individuals play when valuing public goods is more a social than an individualistic
argument. The utilitarian viewpoint that – in the case of biodiversity – the result
of an action is the adequate measure might be weakened by the social context11
in which such a valuation takes place (see below). The social role individuals
play when asked for their WTP refers to their role as ‘consumer’ exchanging
consumption for a higher environmental quality. The hypothesis put forward
here is that when deciding the appropriate amounts of public goods (like bio-
diversity), individuals adopt a different social role, namely the one of a ‘citizen’.
In other words, it is hypothesized that private households (respondents in a survey)
act differently in the realm of the market than in a situation where societal affairs
(e.g. public goods) are at stake (one of the first authors trying to clarify this
argument was Sagoff 1988).12 That does not mean that economic considerations
are completely left out in societal decisions.
Such behaviour of respondents as ‘citizens’ might not meet Becker’s claim that
‘all human behaviour can be viewed as involving participants who maximize their
utility from a stable set of preferences and accumulate an optimal amount of
information’ (Becker 1976: 14). If we take Becker’s statement, it can be considered
as being tautological. There is, of course, no human behaviour that is not
maximizing the individual’s utility as defined by neoclassical theory. Even suicide
can be thought of being the ‘best’ alternative for the person committing suicide.
Altruism can also fit perfectly into the model of economic behaviour as one’s
utility increases when another individual’s utility increases (for a critical view and
arguments for a multiple-utility conception, see Etzioni 1986). Furthermore, rational
behaviour can be considered as a constituting element of humans. However, it is
the opinion of the author that such argumentation does not contribute very much
to the understanding of economic and social behaviour. Economic reasoning and
rationality might play an important role in nearly all decisions, be they ‘private’ or
‘public’, ‘individualistic/egoistic’ or ‘altruistic’. Economic arguments are often only
implicit in individual decisions. It is the task of economics to clarify those preferences
particularly in the context of environmental valuation. It is not put forward here
that ‘altruistic citizens’ do not behave rationally (they can be thought of having a
societal goal that should be achieved with least cost). It is the author’s understanding
that discussing different approaches or explanations can be more fruitful than trying
to find a common theory to all kinds of human behaviour.13 The main argument
of the dichotomy is that ‘citizens’ do not maximize their individual utility in a
narrow sense (seeking personal advantages) but include broader societal arguments
in their decision (see also Frey 1997).14
Besides different roles of respondents, some problems with respondents’ role
differences might also be attributable to limited abilities to process information
(‘bounded rationality’; Conlisk 1996). Bounded rationality in the sense discussed
here could explain why respondents do not answer WTP questions in a way that
would maximize their individual utility. Answers or behaviour departing from the
individual utility maximization assumption can nevertheless be rational to the extent
that citizens answer differently compared with consumers given a different context
and different points of view and reference.
Schkade and Payne (1994) offer an interesting focus on the social context. They
let respondents ‘think aloud’ when making their WTP statement. Respondents
primarily considered their ability to pay, referring to the respondent’s economic
situation or to the payment vehicle. Some additional considerations can be
interpreted as including elements of respondents acting as citizens, but no
connection has been made to the sole benefits of society or its future development.15
Some evidence from psychological research stresses the argument that CVM
questions mainly focus on the consumer role of the respondent. Surveys might
include a tendency ‘to emphasize hedonic (self-interest) rather than social values
by asking respondents for their personal opinions’ (Fischhoff 1991: 843). ‘If one
wants to predict how people will behave in situations presenting a particular
perspective, then one should elicit their values in ways evoking that perspective’
(Fischhoff 1991: 843). Boyce et al. (1992: 1367) make a similar argument, when
they write that ‘WTP only measures intrinsic value to the extent that consumers
accept moral responsibility in a WTP context’ (emphasis added).
A number of publications appearing since Sagoff ’s book have discussed this
issue from several viewpoints. Bingham et al. (1995: 76) argue that it is important
‘to understand the various roles that individuals play and how these roles affect
the notion of “valuing” something. … [Individuals] value ecosystems differently
when expressing personal values and when serving in some advisory or decision-
making role for a public institution. People also reason differently when acting as
decision makers or advisors … than they do as private agents’ (also Blamey 1996;
Kosz 1997; Cameron 1997; Stevens et al. 1994; Common et al. 1993; Fischhoff
and Furby 1988; Common and Perrings 1992).
The central point of the divergence between the social roles individuals take is
that contingent valuation is a concept, which derives out of a (normative) political
theory founded on a distinct model of private choice in markets. As Jacobs (1997:
213f.) puts it: ‘Using these valuation institutions is essentially to treat choices about
public goods as if they were choices about private ones. … The principal difference
concerns the frame of reference people employ when making choices about public
goods. … They are seen as having value to society over and above the value they
give to individuals.’ It is this distinctive framework, which might not be appropriate
because in reality decisions about environmental quality are indeed made by
32 Michael Getzner
individuals (in elections or referenda within a democratic process; Bingham et al.
1995: 77).
This argument additionally involves the idea that not only the mere outcome
of a decision regarding societal problems matters but also how these results are
achieved. In this sense, environmental valuation via CVM is – besides the utilitarian
foundations – based on a teleological view of decision-making. Public goods require
public discourse (the discourse ethics approach proposed by O’Hara 1996). It can
be very important for those affected by a public project to learn how a decision is
being made.16 There is some recent literature on procedural rationality. Citizens
might in the first instance not be interested in the result of a public decision as
such, but rather concentrate on the process of public decision making (e.g. Toman
1999; Holland 1997; Burgess et al. 1998; O’Hara 1996; Getzner 1999).
Empirical results17
Explaining WTP-bids: incorporating the social context

The empirical models presented below are a logit model explaining respondent’s
principal WTP, and a regression model on the respondents’ WTP amounts
estimated by OLS. Both include several dummy variables to take context variables
(see Figure 2.1) into account. Thus the model does not merely consider the usual
socio-economic variables, but is extended by additional individual factors and by
the institutional and social context. The models estimated can be described in
analytical terms as
Pr (WTPi > 0 ) = f (Si , Pi , I i , Ci ) and (3.1a)

WTPi = f (Si , Pi , I i , Ci ), (3.1b)
where Pr(WTPi > 0) is the probability that respondent i exhibits some positive
WTP. WTPi refers to respondent i’s WTP bid (in Austrian schillings).
Four groups of variables are included in the analysis. The first group S describes
the socio-economic characteristics of respondent i (e.g. income, gender, age, job);
P denotes respondent i ’s preferences and behaviour regarding natural goods in
general (e.g. knowledge concerning biodiversity, donations to environmental
organizations, memberships). I is a set of institutional context variables, which
describe not only ‘objective’ arguments (e.g. whether hypothetical bias has been
explicitly mentioned in the survey by ‘cheap talk’) but also respondent’s beliefs
about the CVM setting. Finally, C comprises arguments regarding ‘subjective’
perceptions and citizen’s values about the social context of the survey. These
variables include the respondent’s opinions about the outcome of the referendum
and the consumer–citizen dichotomy. The WTP function increases or decreases
in arguments18 (for a complete list of variables, see Table 2.3).
Design of the survey
The survey for testing the relevance of extending the ‘standard’ valuation model
took place in December 1997 and January 1998 at the University of Klagenfurt
(Austria). Subjects were mainly undergraduate students of business administration
as well as sociology, English and French language. The surveys were held in the
classes; about 20 to 25 minutes were necessary to complete the questionnaire
(sample size 189).
The ‘design guidelines’ of Arrow et al. (1993) regarding the appropriate design
and setting of the CVM survey were followed as closely as possible. After a number
of pre-tests (in classes and with focus groups), the survey was finally structured as
follows. After filling out a number of questions regarding socio-economic variables
(age, gender, studies/courses enrolled, income, children), a set of indicators dealt
with the respondent’s attitudes towards the environment (membership of
environmental organizations, preferred activities in recreational areas, e.g. sports)
and with their knowledge of and familiarity with the concept of biodiversity (using
open-ended as well as closed questions). After these introductory ‘warming-up’
questions, the sample was divided into two groups: one group went straight on
answering the following valuation questions without any further explanation from
the instructor. The other group was asked to take a short break and listen to the
instructor’s short presentation of the problems associated with the hypothetical
bias of such CVM surveys.
The next set of questions asked for the respondent’s WTP for a nature
protection programme in Carinthia.19 These open-ended questions were worded
as conservatively as possible,20 including references to the respondent’s budget
constraint. The payment vehicle was an earmarked contribution to a nature
protection fund. The payment motives as well as reasons for protest or zero bids
were asked for. Additionally, a question on the respondent’s vote on a public
referendum regarding the introduction of an earmarked nature protection tax
in Carinthia was included.
A final set of questions tried to reveal respondents’ attitudes comprising the
‘extended’ WTP model, asking for lexicographic preferences as well as the above-
discussed distinction between social role and epistemological questions (results for
the most important questions of these survey sections, including the wording of
the questions, are presented below).
Descriptive overview of the empirical results

The first questions dealt with respondents’ familiarity with the concept of
biodiversity. It is often asserted in the literature that respondents should be familiar
with and informed about the good to be valued (Mitchell and Carson 1989). It
turned out that the vast majority of respondents had never heard about the concept
of biodiversity. Over 75 per cent of respondents had never heard of this concept,
only 5 per cent of respondents were at least familiar with it.21 Many interpreted
the word ‘biodiversity’ in open-ended questions in the correct sense, meaning
34 Michael Getzner
diversity of species and ecosystems.22 However, the meaning of ‘biodiversity’ was
explained thereafter in the questionnaire.
Quite a large number of respondents were in principle prepared to contribute
to a Carinthian nature protection programme securing the survival of certain
species and ecosystems (written information on the content of this programme
was provided in the survey). Out of 189 respondents, 148 (78.3 per cent) were
willing to pay a personal contribution in the form of an earmarked tax that
would flow into a fund solely committed to financing the programme (16 per
cent of those refusing to pay rejected the entire valuation question). Interestingly,
62.4 per cent of all respondents stated that such a programme should not merely
be financed out of public budgets, but also by private contributions. Those who
refused to pay mainly thought that inefficiencies in the public sector (e.g.
bureaucracy) should first be cut significantly before citizens should contribute
with personal donations.
Respondents’ mean WTP bids in the survey amounted to 509 ATS (Austrian
schillings) per person as an annual contribution to the above-mentioned fund
providing financial resources for implementing the Carinthian nature protection
programme. The payment is limited to a time span of 5 years (this is the time span
when the implementation costs of the whole programme have to be financed).
The standard deviation was 964 ATS, and the median was 250 ATS (all numbers
were calculated without truncation as some of the high bids proved to be ‘reason-
able’ in the sense that respondents bidding so high indeed had the purchasing
power to do so).
Respondents’ main motive to state a WTP was the existence motive which has
been described as comprising a rights-based protection motive in the questionnaire.
A share of 47.3 per cent of respondents marked the existence motive as the main
motive to state a WTP; 45.9 per cent voted for the bequest motive, the rest thought
about the option value of natural resources.
A majority of respondents (about 60 per cent) would vote for a (mandatory)
Carinthian nature protection tax amounting to 160 ATS, which is lower than the
open-ended mean and median WTP elicited in prior questions. Forty per cent of
respondents would vote against the proposition. Furthermore, 32.5 per cent of
respondents thought that their WTP might lie above the sample mean, while 67
per cent were of the opinion that their WTP was at or below the average (a similar
question regarding the median was left out of the questionnaire due to space
restrictions).
A large proportion, 66 per cent of respondents, showed motives that can be
interpreted as being altruistic. They would donate a personal contribution to the
Carinthian nature protection programme regardless of the outcome of the
referendum; only 7.5 per cent would donate nothing, 5.7 per cent would lower
their original donation, and around 20 per cent would await the forming of a new
majority. That means that a majority of respondents would be willing to contribute
to a nature protection fund even if not forced to do so.
Students as subjects seem to be a group where nature and in particular species
protection is a high priority. Regarding lexicographic preferences, nearly all
respondents stated that species should be protected regardless of the cost of
protection; apparently they hold rights-based beliefs, as the wording of the questions
and the appropriate answers could suggest (around 94 per cent of respondents
believed that species should be protected regardless of the economic costs associated
with it). The wording of this question is consistent with other surveys (e.g. Stevens
et al. 1991; Spash and Hanley 1995). Unfortunately, this question could not
discriminate between groups of respondents (that circumstance did not show up
in the pre-tests).
The consumer–citizen differences in respondents’ roles were explored in a simple
setting similar to Sagoff ’s (1988) historic example. Respondents were simply asked
for their preferences regarding skiing in the Nockberge national park, which is an
important regional natural area of about 18,500 hectares. The national park lies
50 kilometres to the north-west of Klagenfurt (the capital city of Carinthia) near
the Hohe Tauern national park. While one-quarter of respondents would like to
go skiing in the ‘Nockberge’ national park, about 95 per cent of all respondents
would vote for the protection of the national park in a referendum (see Table 2.1).
Only two respondents would not go skiing there, but would vote for a ski resort in
the Nockberge.
A number of epistemological questions were raised in the survey. Only 32 per
cent of respondents stated that they were not influenced by the survey and the
specific CVM questions, while nearly two-thirds of the respondents had not thought
about their WTP before answering the survey questions. Fourteen per cent were
explicitly influenced by the ‘cost–benefit’ approach, and again about 7 per cent
rejected that kind of questioning (see Table 2.2).
Explaining WTP bids: the importance of the ‘extended’

valuation framework
The estimated equations for explaining respondent’s WTP on the basis of the
explanatory variables (for a detailed description, see Table 2.3) are presented for
the logit model in Table 2.4 and for the OLS regression model in Table 2.5.
Generally, all included variables have the expected signs (e.g. higher income,
knowledge or a positive vote in a public referendum all increase the respondent’s
acceptance of a personal contribution and the WTP bid). Regarding the different
variables, the respondent’s WTP should increase with higher income, education
and knowledge about biodiversity. Furthermore, it is expected that members of
an environmental organization are, ceteris paribus, willing to pay more while those
opposing the question format exhibit a lower WTP. Those with a probably positive
vote in a public referendum regarding a nature protection tax should again be
willing to pay more – like those willing to pay a voluntary contribution, even if
they would be in a minority in a respective referendum. These expectations
regarding the direction of the influence of the variables are generally satisfied
with the models estimated.
The respondent’s principal WTP is estimated with a logit model. The dependent
variable is WTP1 (=1, if respondent in principle is willing to contribute some
36 Michael Getzner
Table 2.1 Consumer vs. citizen and students’ perspectives
Respondent as consumer
‘The Nockberge national park is not (yet) an internationally acknowledged national
park, but is protected by provincial law. Let us assume that the Nockberge would be
changed into an attractive and also reasonably priced ski and recreation resort with
numerous lifts and sports facilities, and the protection of nature would be abandoned.
Would you personally like to go skiing in the Nockberge national park or enjoy other
attractive sports and recreation facilities?’
Frequency Per cent
No 138 73.02
Yes 51 26.98
Total 189 100.00
Respondent as citizen
‘Regardless of your personal possibilities or wishes to use the Nockberge national park
for your sports and recreation enjoyment: would you as a citizen taking all arguments
into account vote in a referendum for the use of the Nockberge as an attractive ski
resort, or do you think that the Nockberge should in all cases remain a national park?’
Frequency Per cent
I think that the national park should
strictly be protected regardless of the
economic benefits of a ski resort. 178 94.18
I would vote for the ski resort because
of the economic benefits. 7 3.70
Sum 185 97.88
Missing/no answer 4 2.12
Sum 189 100.00
Note: All questions are translated from the original German questionnaire.
amount). Explaining the respondent’s yes-vote is done stepwise. Four different

models are estimated to show the influence of several extensions to the ‘standard’
WTP model, which is shown in the first two columns in Table 2.4. In the ‘standard
model’, the explanatory variables ‘Age’, ‘Economics’, ‘Biodiversity’, ‘Member’,
‘Donations’ and ‘Recreation’ have a significant influence on the respondent’s
accepting a personal contribution to a nature protection fund. The log-likelihood
of the model is –64.18 with a corresponding McFadden’s R² of about 26 per cent
(which is well above the usually applied thresholds for CVM survey validity). About
85 per cent of the answers can be correctly predicted.
If the model is extended to account for an individual’s beliefs about the CVM
setting, the variables ‘Protest’ and ‘Influence’ are additionally significant in
explaining the respondent’s WTP. While the coefficients of the other explanatory
variables stay roughly in the same order of magnitude, these additional
(‘epistemological’) variables improve the explanatory power of the model
(McFadden’s R²) to 34 per cent.
Extending the model even further to account for the social context, variables
‘Referendum’ and ‘Public’ turn out to be significantly correlated with WTP1,
Table 2.2 Epistemological problems of WTP surveys
‘Surveys asking respondents about their personal willingness to pay for environmental
improvements are relatively rare. Answering the questions and this specific kind of
questioning is perceived by some as strange or is rejected. Which of the following
statements would you agree to?’ [Multiple answers]
Frequency Respondents (%)
This sort of questioning has neither
influenced me personally nor my opinions
about natural protection. 61 32.28
Before this survey, I had not thought about
my willingness to pay for a nature
protection programme. 124 65.61
This kind of question has influenced me.
I will in the future pay more attention to
whether nature protection yields economic
benefits compared to the costs. 26 13.76
I reject these kinds of questions on the
willingness to pay for nature protection,
because species protection cannot be valued
in monetary terms. 14 7.41
N 189
leading to an R² of 42 per cent. Finally, the inclusion of the variable ‘Con_cit’ only
marginally improves the quality of the model. About 87 per cent of all answers
can correctly be predicted.
After having explained the respondent’s principal WTP, the second part of the
‘nested’ model was estimated by OLS. Extending the valuation framework to
explain the respondent’s WTP bids by means of a simple OLS regression model
improves the model, too. As Table 2.5 shows, the usually significant socio-economic
variables in regression models are significant in this estimation as well. Most
influential are ‘Income’ and ‘Children’, and (interestingly) ‘Gender’. Some
expectedly influential variables regarding environmental behaviour and
environmental consciousness also play a significant role in explaining the
respondent’s WTP bid. Variables describing the institutional and social context in
which decisions on nature protection programmes are made are of crucial
importance. A highly significant explanatory power is revealed by those variables
that show the respondent’s decisions (behaviour) in public referenda as well as
their willingness to donate a personal contribution even if they are in a minority
position. That means that the higher the supposedly altruistic motives
operationalized with this concept, the higher the respondent’s WTP bid is.
Beginning with the usual socio-economic characteristics of respondents,
variables ‘Income’, ‘Gender’, ‘Biodiversity’, ‘Member’, and ‘Recreation’ play a
significant role. The explanatory power of the OLS model for the respondent’s
WTP bids is 19 per cent. If we include institutional context variables (respondent’s
perception of the CVM setting), about 24 per cent of the variance of the dependent
38 Michael Getzner
Table 2.3 WTP regression model variables
Variable name Dimension Description

Dependent variable
WTP1 1=yes Principal WTP for nature protection programme in
Carinthia
WTP2 ATS (natural WTP including zero bids
log)
Individual factors: socio-economic variables
Income in 1,000 ATS Income of respondent
(natural log)
Profession 1=yes At least in part-time employment
Age in years Age of respondent
Gender 1=female Gender of respondent
Children 1=yes Has children or children in the household
Economics 1=yes Studies business admin. and has introductory economics
Individual factors: environmental consciousness and behaviour
Biodiversity 1=yes Partially familiar with the concept of biodiversity
Member 1=yes Member of an environmental organisation
Donations ATS (natural Monthly average donation to environmental groups
log)
Recreation 1=yes Likes recreation in the open countryside
Institutional context
Protest 1=yes Explicitly disliked questions valuing natural goods
Influence 1=yes Admits influence by the kind of valuation question
Thoughts 1=yes Never thought about economic valuations of biodiversity
before
Cheap 1=yes The valuation question was framed by verbal explanations
regarding hypothetical bias (‘cheap talk’)
Social context
Majority 1=yes Thinks that the majority of citizens would vote for the
introduction of a nature protection tax in Carinthia
Minority 1=yes WTP regardless of the outcome of the referendum
Referendum 1=yes Votes yes for the introduction of nature protection tax in
Carinthia
Average 1=yes Believes contribution is above the average
Public 1=yes Believes that nature protection programmes should be
financed by public budget
Con_cit 1=yes Respondent exhibited different preferences as consumer
and citizen (questions as in Table 2.1)
Table 2.4 Logit model estimated coefficients for WTP1
Model 1 Model 2 Model 3 Model 4

Variables Coefficient z-Statistic Coefficient z-Statistic Coefficient z–Statistic Coefficient z–Statistic
Std. Error Std. Error Std. Error Std. Error
Constant –3.8269 –1.1517 –3.6743 –0.9982 –2.4258 –0.6303 –1.9814 –0.4443
3.3227 3.6810 3.8487 4.4592
Income 0.5326 1.2260 0.3908 0.8216 0.1895 0.3767 0.1995 0.3371
0.4344 0.4756 0.5030 0.5919
Profession –0.1196 –0.2380 –0.4198 –0.7505 –0.2037 –0.3503 0.0570 0.0856
0.5024 0.5594 0.5813 0.6661
Age –0.0680 –1.6955* –0.0245 –0.5187 –0.0201 –0.4056 –0.0117 –0.2095
0.0401 0.0473 0.0495 0.0557
Gender –0.0929 –0.1727 –0.2845 –0.4931 –0.2645 –0.4327 –0.7243 –1.0350
0.5381 0.5770 0.6113 0.6998
Children 0.0970 0.1627 –0.1270 –0.1963 –0.3359 –0.5044 –0.3869 –0.4902
0.5960 0.6469 0.6659 0.7891
Economics 0.8620 1.6232(*) 0.9594 1.6698* 1.1795 1.9147* 1.1912 1.6263(*)
0.5311 0.5746 0.6160 0.7324
Biodiversity 1.5275 2.1372** 1.9659 2.3544** 1.7591 2.0018** 2.2522 1.9677**
0.7147 0.8350 0.8788 1.1446
Member 2.3797 2.0127** 3.2502 2.4445** 3.0327 2.3707** 3.0285 2.2762**
1.1823 1.3296 1.2793 1.3305
Donations 0.4708 1.7181* 0.3961 1.4663 0.3627 1.3175 0.3136 1.0986
0.2740 0.2702 0.2753 0.2854
Recreation 1.6453 3.3224*** 1.7444 3.1943*** 1.8223 3.2355*** 1.9002 2.7542***
0.4952 0.5461 0.5632 0.6899
Protest –2.6699 –2.0462** –2.7654 –1.9380* –2.8254 –1.7794*
1.3048 1.4270 1.5878
Influence –1.3761 –2.0176** –1.6679 –2.3378** –2.3451 –2.8739***

0.6820 0.7135 0.8160
39
continued …
Table 2.4 Continued

Variables Coefficient z-Statistic Coefficient z-Statistic Coefficient z–Statistic Coefficient z–Statistic
Constant –3.8269 –1.1517 –3.6743 –0.9982 –2.4258 –0.6303 –1.9814 –0.4443
40 Michael Getzner
Thoughts 0.7090 1.3168 0.2858 0.4951 0.6531 0.9298

0.5384 0.5772 0.7024
Referendum 1.0153 1.9468* 1.0553 1.7244*
0.5215 0.6120
Public –1.3579 –2.2454**
0.6047
Con_cit –0.8726 –1.2391
0.7042
n 171 171 170 159
df 10 13 14 16
S.E. of regr. 0.3509 0.3353 0.3248 0.3067
Log likelihood –64.1778 –57.2930 –53.8363 –43.2056
McFadden R² 25.95% 33.89% 36.71% 44.92%
Notes
*** p<0.01
** p<0.05
* p<0.1
(*) 0.1<=p<0.12
Table 2.5 Log-normal OLS estimated coefficients for WTP2

Variables Coefficient t-Statistic Coefficient t-Statistic Coefficient t-Statistic Coefficient t-Statistic
Constant –0.8344 –0.3313 –2.4679 –0.9364 1.1999 0.6039 2.3239 1.1388
2.5186 2.6356 1.9869 2.0407
Income 0.5458 1.6144(*) 0.6477 1.8992* 0.2054 0.7789 0.0233 0.0877
0.3381 0.3410 0.2637 0.2655
Profession 0.3333 0.8294 0.2783 0.7088 –0.0667 –0.2264 0.0463 0.1548
0.4019 0.3927 0.2949 0.2991
Age –0.0486 –1.4058 –0.0444 –1.2650 –0.0259 –0.9231 0.0023 0.0836
0.0346 0.0351 0.0280 0.0280
Gender –0.7283 –1.7754* –0.7238 –1.8147* –0.5672 –1.8753* –0.5906 –1.9984**
0.4102 0.3988 0.3025 0.2955
Children 0.7330 1.4903 0.7270 1.4984 0.7278 1.9513* 0.8268 2.2392**
0.4918 0.4852 0.3730 0.3692
Economics 0.4958 1.1410 0.5573 1.3111 0.2805 0.8485 0.4593 1.3659
0.4345 0.4251 0.3306 0.3363
Biodiversity 1.5229 3.4049*** 1.5703 3.5827*** 0.4179 1.2368 0.4625 1.3963
0.4472 0.4383 0.3379 0.3312
Member 1.3432 2.4941** 1.6912 3.1619*** 0.7102 1.6980* 0.7955 1.9267*
0.5385 0.5349 0.4182 0.4129
Donations 0.1766 1.5144 0.1210 1.0451 0.0520 0.6108 0.0059 0.0721
0.1166 0.1158 0.0851 0.0821
Recreation 1.1439 2.6680*** 1.1361 2.6972*** 0.8460 2.6339*** 0.7025 2.1364**
0.4287 0.4212 0.3212 0.3288
Protest –1.4219 –1.7735* –1.3956 –1.8278* –1.6480 –2.2465**
0.8018 0.7635 0.7336
Influence –0.6708 –1.2698 0.2236 0.5317 0.2430 0.5889

0.5282 0.4206 0.4127
41
continued …
Table 2.5 Continued

Variables Coefficient t-Statistic Coefficient t-Statistic Coefficient t-Statistic Coefficient t-Statistic
Thoughts 0.5516 1.3532 –0.0500 –0.1547 0.3301 0.9629
42 Michael Getzner
0.4076 0.3232 0.3428

Cheap 0.8214 2.2484** 0.6309 2.2944** 0.5194 1.8641*
0.3653 0.2750 0.2786
Majority –0.0621 –0.1979 0.1655 0.5252
0.3139 0.3150
Minority 1.1957 4.0001*** 1.2641 4.1934***
0.2989 0.3014
Referendum 0.8607 2.7771*** 0.7873 2.5404**
0.3099 0.3099
Average 1.1955 3.6743*** 0.9334 2.7823***
0.3254 0.3355
Public –1.1655 –3.6273***
0.3213
Con_cit –0.5637 –1.7969*
0.3137
n 172 172 144 135
df 10 14 18 20
Adj. R² 19.11% 23.79% 39.04% 45.20%
S.E. of regr. 2.3338 2.2654 1.5449 1.464
F–statistic 5.0410*** 4.8131*** 6.0877*** 6.5255***
Notes
***p<0.01 **p<0.05 *p<0.1 (*)0.1<=p<0.12
a
Results of Tobit estimation are available on request; however, the significance of variables stays roughly the same.
variable can be explained. Variables ‘Protest’ and ‘Cheap’ are additionally
significant.
Extending the model even further to include individual beliefs about the social
context, variables ‘Minority’, ‘Referendum’, and ‘Average’ increase the power of
the model to 39 per cent. If we try to account for the consumer–citizen dichotomy
(‘Con_cit’) and beliefs about the appropriate financing mechanism of public goods
(‘Public’), an explanation of 45 per cent of the variance can be achieved.
The high explanatory power of the model shows that not only are the usual
individual characteristics of respondents as individuals and consumers crucially
influential, but knowing about respondents’ attitudes and behaviour in the
institutional and social context improves the quality of the model and adds much
explanatory power to give a richer picture of the framework in which respondents
decide upon their contribution to a regional biodiversity programme.
Discussion and conclusions

Criticism of the CVM is as old as the method itself, and its attackers come from
a number of different corners. One major strand of critical work focuses on the
assumption of welfare economics and CBA, which holds that individual costs
and benefits can be aggregated in a way to judge the economic reasonability of
a project or policy on the level of society. However, society, as the term suggests,
comprises more than the narrow economic (or market-based) context. Ultimately,
the institutional and social context is manifest in individual perceptions, attitudes,
beliefs and behaviour. Ethical commitments and rights-based beliefs towards
other species constitute problems with the narrow economic notion of efficiency
and the concept of contingent valuation and CBA, as individuals might not be
willing to make the usual trade-offs between income (or consumer goods) and
biodiversity.
These so-called lexicographic preferences are more ‘individualistic’ than the
much-discussed divergence between social role as consumers and citizens. The
latter dichotomy refers to a broader model of human behaviour taking not only
individual utility maximization into account, but also referring to social values
held by individuals. However, epistemological problems with the assumption of
the Homo oeconomicus and the consequent use of methods appropriate to a narrow
economic theory rather than methods appropriate to more broader approaches
targeted at the perception and behaviour of individuals in an environmental context,
threaten the validity of CVM-based valuation.
There has been much controversy over these topics in the literature; but
numerical or statistical analysis is rather rare compared with more qualitative or
explorative approaches. In this chapter, empirical evidence is presented on the
extent and importance of some of the most crucial elements of the institutional
and social context in which the valuation of regional biodiversity takes place. While
the sample chosen for the survey is not representative for the whole population, it
can be assumed that the significant variables found in regression models will in
principle also prove influential for the whole population.
44 Michael Getzner
A large proportion of respondents exhibited lexicographic preferences, and
the results of the survey suggest that the consumer–citizen difference in perceiving
a context-inherent question is also relevant. Although these two crucial elements
of the valuation procedure are of great importance, some of the theoretical
considerations regarding the amount of WTP bids cannot be proven empirically.
Theory would suggest that – due to the priority of biodiversity in the case of
lexicographic preferences – WTP bids could be zero or infinite. The WTP bids in
this study are quite reasonable in the sense that the respondents considered their
budget constraint when stating a WTP bid. One possible explanation could be
that respondents want their preferences to be respected in the decision process.
They take those instruments offered to them, and a monetary expression of their
preferences is a way to influence the outcome even if they distrust CVM-based
research.
Respondents answered a number of epistemological questions. A majority of
respondents was not used to answering such valuation questions, and thought about
it for the first time when dealing with the questionnaire. Variables comprising
these elements of instrument-based influence were statistically significant in
explaining WTP bids in the regression model. WTP bids are influenced by
respondents’ perception about the CVM instrument as well as the framing of the
WTP question.
Variables encoding the respondents’ perceptions about the outcome of a similar
referendum additionally exhibited significant influence (supposedly altruistic
motives). It seems that respondents who would vote publicly for a nature protection
tax are prepared to state higher private WTP bids. Furthermore, the perception
of one stating higher WTP bids as well as the WTP in a situation when a general
nature tax is not approved contribute significantly to the respondent’s WTP amount.
The social (and political) context of decisions thus turns out to be crucially important
in explaining respondents’ WTP bids. Beliefs about the respondent’s behaviour in
a minority position as well as supposedly altruistic motives influence the respondent’s
answers in a CVM survey.23
The empirical results indicate that, besides socio-economic variables, the
institutional and social context in which respondents answer valuation questions
can be of crucial importance. The explanatory power of the WTP bid models
improves as more context variables are entered into the equation. Respondents of
CVM surveys behave more like citizens in a political and societal debate – with
their personal beliefs and perceptions of the institutional and social context of the
valuation decision – rather than fulfilling the economic model of pure exchange
by trading money vs. natural goods. CVM surveys without reference to institutional
and social context variables thus may lack important arguments in explaining
respondents’ WTP bids because environmental decisions apparently are more
political than economic because they include morality (rights-based beliefs) and
community rather than individual consumer interests.
As a final conclusion regarding the use of contingent valuation results, it seems
that contingent valuation exhibits economic values useful for evaluating policy
programmes. However, environmental policy decisions cannot be based solely on
such results, as the valuation of environmental goods regularly includes a broad
range of different values, perceptions and assumptions. CVM results might be an
important element of decision making from an economic perspective (e.g. Carson
2000). Multi-criteria valuation frameworks and participatory decision processes
may thus account for such elements and integrate them in a broader and more
inclusive decision-making framework.
Acknowledgements
The author would like to thank all participating students of the University of
Klagenfurt as well as the instructors of those courses who ‘donated’ 25 minutes of
their course time. Particular thanks go to G. Haber, W. Hediger, S. Niemeyer, M.
Riedel and N. Wohlgemuth for commenting on earlier drafts of this chapter. The
author accepts the responsibility for all remaining errors.
Notes
1 Based on the recommendation of Arrow et al. (1993), the concept of respondents’
WTP is used. However, many results of this chapter might also be transferred to
respondents’ willingness to accept.
2 The ‘standard’ model referred to in this chapter is a sketch of the usually applied
approach to explain the respondent’s WTP (see also the theory of reasoned action by
Fishbein and Ajzen 1975).
3 The term ‘institutional’ is taken from the auction literature where the ‘institution’
refers to the elicitation instrument and auction setup.
4 The concept of biodiversity has become ‘popular’, particularly after the United Nations
Conference on Environment and Development in Rio de Janeiro (1992) which
produced a convention on the protection of biodiversity and biological resources.
This convention defines biodiversity in natural science terms as the diversity within
species (genetic diversity), species diversity (number of different species living in an
ecosystem), as well as ecosystem and landscape diversity (different ecosystems and
their protection).
5 Diversity is a characteristic of a species as well as of an ecosystem. Thus, valuation
becomes much more complicated. Respondents in CVM studies are usually asked to
assess policy-relevant proposals that often involve a number of biodiversity issues.
Thus, respondents must consider their sympathy towards single species as well as
taking into account their perceptions towards the importance of a (frequently
unknown) number of species, their interrelations to secure the stability of ecosystems
as a whole, including risk and (fundamental) uncertainties, as well as expected needs
of future generations. In the light of some ecological fundamentals including
uncertainty, the task of appropriately valuing biodiversity seems particularly difficult.
6 That means that respondents in a CVM survey might consider several elements of
an ecosystem to be substitutable or not vital for the function of that ecosystem while
ecologists and biologists would strongly oppose such beliefs.
7 For further critique, see Boulding (1994: 8): ‘It is curious that economists, almost
without exception, take human valuations for granted and assume that everybody
has a set of preferences that can be expressed as an indifference curve.’ Preferences
are often found to be imprecise, as there is ‘plentiful evidence of intervals within
which respondents found it difficult to be sure whether they would or would not pay
or accept compensation for marginal changes in risks of injury’ in road accidents
(Dubourg et al. 1994: 127).
46 Michael Getzner
8 It can be argued that preferences based on existence values are more likely to be of
lexicographic nature since existence values are often founded on rights-based beliefs
(Stevens et al. 1991). Furthermore, a WTP bid based on existence values can
additionally be founded on a number of other reasons, e.g. ‘warm glow’ giving
(Kahneman and Knetsch 1992), paying for ‘good causes’ (Stevens et al. 1994) or impure
altruism (Andreoni 1990).
9 For more examples on the ‘unwillingness to accept’, see O’Connor (1997: 463).
10 Some differences of results of these studies can, of course, be attributed to differences
in the institutional settings and contexts of these experiments.
11 Hanemann (1995: 88) argues that statements of respondents and their preferences
are ‘inherently context-laden’. Furthermore, he reviews in detail the problems involved
with CVM and voting, and offers a number of methods to overcome problems with
inadequately defined context variables.
12 Note, the diverging roles assumed here are not generally accepted in mainstream
economics ‘since citizens can be expected to contribute, in one way or another, toward
the costs of policy, the respondent trades off the prospect of an amenity improvement
against the prospect of a reduction in disposable income’ (Hoehn and Randall 1987:
229). However, the strict distinction between ‘citizen’ and ‘consumer’ is made for
illustrative purposes. Those two concepts are better thought of as two poles of a
continuum from self-interested utility maximization to purely altruistic social
behaviour. Individuals can, of course, behave according to both patterns (i.e. they are
not substitutes).
13 System theory particularly stresses the importance of differences between systems in
order to be able to strictly divide between circumstances that are inside or outside a
particular system (e.g. Luhmann 1986).
14 ‘Individual behavior is always mediated by social relations’ (Arrow 1994: 5).
15 In discussing their results, Schkade and Payne (1994: 107) explicitly mention the ‘citizen
input (values) into policy decision’. Their study shows that within CVM questionnaires
citizen’s considerations are generally excluded rather than acknowledged.
16 Public projects may run into problems not only because of project characteristics but
also due to missing or inadequate involvement of citizens. In this sense CVM as a
hypothetical market transaction is ‘blind to reasons’ (Keat 1997: 36). Nevertheless, a
CVM survey is one way of involving affected citizens.
17 The questionnaire, additional written material as well as all data, are available from
the author on request.
18 There are a number of hypotheses regarding the sign of the coefficients in the model.
For instance, WTPi / Yi 0 , with Yi being the income of respondent i as part of the
group of Sm variables in equations (3.1a) and (3.1b).
19 An actually existing nature programme of the regional government of Carinthia to
protect and enhance biodiversity was valued. The programme was described in terms
of what was to be valued (¾ page of description of the programme), how it would be
provided and at what level (regional government and nature protection fund), who
would pay (partly regional government budget and the fund), and how long it would
last (5 years).
20 Due to budgetary constraints (small sample) and lack of previous valuation estimates,
a question format including dichotomous choice was deemed infeasible. In any case,
the significance of institutional and social context variables seems independent of
the chosen question format.
21 The figures presented in this section illustrate preferences of the entire sample
consisting mainly of students. The values do not provide a basis for aggregation to
the whole population. Rather than producing WTP figures for Carinthia, the aim
was to show the influence of extended context variables on WTP bids, which can be
assumed to exist in the whole population.
22 Respondents’ problems with understanding the concept of biodiversity can also be in
part interpreted as being a semantic difficulty: In German, the word Biodiversität is
apparently poorly known. The words Artenschutz and Artenvielfalt would be understood
more clearly, but these concepts only comprise species protection and diversity, leaving
out genetic as well as ecosystems diversity.
23 Of course, the credibility of the proposed programme can also be of crucial
importance. Respondents thus might particularly value policy programmes and
environmental goods as ‘public’ goods rather than context-free discrete ‘bits’ of
environmental goods (Brouwer et al. 1999).
References
Aldred, J. (1994) ‘Existence value, welfare and altruism’, Environmental Values, 3(4): 381–
402.
Andreoni, J. (1990) ‘Impure altruism and donations to public goods: a theory of warm-
glow giving’, The Economic Journal, 100 (June): 464–77.
Arrow, K.J. (1994) ‘Methodological individualism and social knowledge’, American Economic
Review, 84(2): 1–9.
Arrow, K.J., Solow, R.M., Leamer, E., Portney, P., Randner, R. and Schuman, H. (1993)
‘Report of the NOAA panel on contingent valuation’, Federal Register, 58(10): 4602–14.
Becker, G.S. (1976) The Economic Approach to Human Behavior, Chicago: University of Chicago
Press.
Bingham, G., Bishop, R., Brody, M, Bromley, D., Clark, E., Cooper, W., Costanza, R.,
Hale, T., Hayden, G., Kellert, S., Norgaard, R., Norton, B., Payne, J., Russell, C. and
Suter, G. (1995) ‘Issues in ecosystem valuation: improving information for decision
making’, Ecological Economics, 14(2): 73–90.
Bishop, R.C. and Welsh, M.P. (1992) ‘Existence values in benefit-cost analysis and damage
assessment’, Land Economics, 68(4): 405–17.
Bishop, R.C., Heberlein, T.A. and Kealy, M.J. (1983) ‘Contingent valuation of
environmental assets: comparison with a simulated market’, Natural Resources Journal,
23(3): 619–33.
Blackburn, M., Harrison, G.W. and Rutström, E.E. (1994) ‘Statistical bias functions and
informative hypothetical surveys’, American Journal of Agricultural Economics, 76(5):
1048–88.
Blamey, R.K. (1996) ‘Citizens, consumers and contingent valuation: clarification and the
expression of citizen values and issue options’, in V. Adamovicz (ed.) Forestry, Economics
and the Environment, Wallingford: CAB International.
Boulding, K.E. (1969) ‘Economics as a moral science’, American Economic Review, 59(1):
1–12.
Boulding, K.E. (1994) ‘Economic ethics and the environmental crisis’, in E. Hickey and
L.A. Longmire (eds) The Environment: global problems, local solutions, Westport, CT/London:
Greenwood Press.
Boyce, R., Brown, T.C., McClelland, G.H., Peterson, G.L. and Schulze, W.D. (1992) ‘An
experimental examination of intrinsic values as a source of the WTA-WTP disparity’,
American Economic Review, 82(5): 1366–73.
Brouwer, R., Powe, N., Turner, R.K., Bateman, I.J. and Langford, I.H. (1999) ‘Public
attitudes to contingent valuation and public consultation’, Environmental Values, 8:
325–47.
48 Michael Getzner
Burgess, J., Clark, J. and Harrison, C.M. (1998), ‘Respondents’ evaluations of a contingent
valuation survey: case study based on an economic valuation of the wildlife enhancement
scheme, Pevens Levels in East Sussex’, Area 30(1): 19–27.
Cameron, J.I. (1997) ‘Applying socio-ecological economics: a case study of contingent
valuation and integrated catchment management’, Ecological Economics, 23(2): 155–65.
Carson, R.T. (1998) ‘Valuation of tropical rainforests: philosophical and practical issues
in the use of contingent valuation’, Ecological Economics, 24(1): 15–29.
Carson, R.T. (2000) ‘Contingent valuation: a user’s guide’, Environmental Science and Technology,
34(8): 1413–18.
Carter, J.R. and Irons, M.D. (1991) ‘Are economists different, and if so, why?’, Journal of
Economic Perspectives, 5(2): 171–7.
Common, M. and Perrings, C. (1992) ‘Towards an ecological economics of sustainability’,
Ecological Economics, 6(1): 7–34.
Common, M., Blamey, R. and Norton, T. (1993) ‘Sustainability and environmental
valuation’, Environmental Values, 2(4): 299–334.
Common, M., Reid, I. and Blamey, R. (1997) ‘Do existence values for cost benefit analysis
exist?’, Environmental and Resource Economics, 9(2): 225–38.
Conlisk, J. (1996) ‘Why bounded rationality?’, Journal of Economic Literature, 34 (2): 669–700.
Cummings, R.G., Harrison, G.W. and Osborne, L.L. (1995a) ‘Are realistic referenda real?’,
Economics Working Paper B-95-06, Division of Research, College of Business
Administration, Columbia: University of South Carolina.
Cummings, R.G., Harrison, G.W. and Osborne, L.L. (1995b) ‘Can the bias of contingent
valuation surveys be reduced? Evidence from the laboratory’, Economics Working Paper
B-95-03, Division of Research, College of Business Administration, Columbia:
University of South Carolina.
de Dios, E.S. (1995) ‘Re-reading Menger’s table’, European Journal of Political Economy, 11(2):
317–34.
Department of the Interior (1994) ‘Natural resource damage assessment’, Federal Register,
59(85): 23098–111.
Dubourg, W.R., Jones-Lee, M.W. and Loomes, G. (1994) ‘Imprecise preferences and the
WTP-WTA disparity’, Journal of Risk and Uncertainty, 9(2): 115–33.
Edwards, S.F. (1992) ‘Rethinking existence values’, Land Economics, 68(1): 120–2.
Ehrlich, P. and Ehrlich, A. (1992) ‘The value of biodiversity’, Ambio, 21(3): 219–26.
Encarnación, J. (1990) ‘Consumer choice of qualities’, Economica, 57(1): 63–72.
Etzioni, A. (1986) ‘The case for a multiple utility conception’, Economics and Philosophy, 2(2):
159–83.
Fischhoff, B. (1991) ‘Value elicitation: is there anything in there?’, American Psychologist,
46(8), 835–47.
Fischhoff, B. and Furby, L. (1988) ‘Measuring values: a conceptual framework for
interpreting transactions with special reference to contingent valuation of visibility’,
Journal of Risk and Uncertainty, 1(1), 147–84.
Fishbein, M. and Ajzen, I. (1975) Belief, Attitude, Intention and Behavior: An Introduction to Theory
and Research, Reading: Addison-Wesley.
Frey, B.S. (1997) Not Just for the Money, Cheltenham: Edward Elgar.
Georgiou, S., Langford, I.H., Bateman, I.J. and Turner, R.K. (1998) ‘Determinants of
individuals’ willingness to pay for perceived reductions on environmental health risks:
a case study of bathing water quality’, Environment and Planning A, 30 (3): 577–94.
Getzner, M. (1999) ‘Risk, uncertainty and discounting in practical environmental decision
making’, in K. Steininger and F. Prettenthaler (eds) Proceedings of the Workshop ‘Risk and
Uncertainty in HD-Research’, Austrian Programme on Human Dimensions of Global
Environmental Change, Graz: Department of Economics of the University of Graz.
Getzner, M. (2000) ‘Hypothetical and real economic commitments, and social status in
valuing a species protection program’, Journal of Environmental Planning and Management,
43(4): 541–59.
Goodman, S., Jaffry, S. and Seabrooke, B. (1999) ‘Assessing public preferences for the
conservation quality of the British coast, in M. O’Connor and C.L. Spash (eds) Valuation
and the Environment: Theory, Method and Practice, Cheltenham: Edward Elgar: 165–82.
Gowdy, J. (1996) ‘Discounting, hierarchies, and the social aspects of biodiversity protection’,
International Journal of Social Economics, 23(4,5,6): 49–63.
Gowdy J. and Olsen P.R. (1994) ‘Further problems with neoclassical environmental
economics’, Environmental Ethics, 16(2): 161–71.
Hanemann, W.M. (1995) ‘Contingent valuation and economics’, in K.G. Willis and J.T.
Corkindale (eds) Environmental Valuation: New Perspectives, Wallingford: CAB International:
79–117.
Hanley, N., Spash, C. and Walker, L. (1995) ‘Problems in valuing the benefits of biodiversity
protection’, Environmental and Resource Economics, 5(3): 249–72.
Harrison, G.W. (1996) ‘Experimental economics and contingent valuation’, Economics
Working Paper B-96-10, Division of Research, College of Business Administration,
Columbia: University of South Carolina.
Hayden, F.G. (1993) ‘Ecosystem valuation: combining economics, philosophy, and ecology’,
Journal of Economic Issues, 27(2): 409–20.
Hodgson, G.M. (1997) ‘Economics, environmental policy and the transcendence of
utilitarianism’, in J. Foster (ed.) Valuing Nature? Ethics, Economics and the Environment,
London/New York: Routledge: 48–66.
Hoehn, J.P. and Randall, A. (1987) ‘A satisfactory cost indicator from contingent valuation’,
Journal of Environmental Economics and Management, 14(3): 226–47.
Holland, A. (1997) ‘The foundations of environmental decision-making’, International Journal
of Environment and Pollution, 7(4): 483–96.
Holm-Müller, K., Hansen, H., Klockmann, M. and Luther, P. (1991) Die Nachfrage nach
Umweltqualität in der Bundesrepublik Deutschland, Berlin: Erich Schmidt Verlag.
Jacobs, M. (1997) ‘Environmental valuation, deliberative democracy and public decision
making’, in J. Foster (ed.) Valuing Nature? Ethics, Economics and the Environment, London/
New York: Routledge: 211–31.
Jakobsson, K.M. and Dragun, A.K. (1996) Contingent Valuation and Endangered Species,
Cheltenham: Edward Elgar.
Jeffrey, R.C. (1974) ‘Preferences among preferences’, Journal of Philosophy, 71(39): 377–91.
Johansson, P.-O. (1992) ‘Altruism in cost-benefit analysis’, Environmental and Resource Economics,
2(4): 605–13.
Kahneman, D. and Knetsch, J.L. (1992) ‘Valuing public goods: the purchase of moral
satisfaction’, Journal of Environmental Economics and Management, 22(1): 57–70.
Keat, R. (1997) ‘Values and preferences in neo-classical environmental economics’, in J.
Foster (ed.) Valuing Nature? Ethics, Economics and the Environment, London/New York:
Routledge: 32–47.
Kosz, M. (1997) ‘Probleme der monetären Bewertung von Biodiversität’, Zeitschrift für
Umweltpolitik und Umweltrecht, 20(4): 531–50.
Lazo, J.K., Schulze, W.D., McClelland, G.H. and Doyle, J.K. (1992) ‘Can contingent
valuation measure nonuse values?’, American Journal of Agricultural Economics, 74 (6): 1127–
32.
50 Michael Getzner
Lockwood, M. (1997) ‘Integrated value theory for natural areas’, Ecological Economics, 20(1):
83–93.
Lockwood, M. (1998) ‘Integrated value assessment using paired comparisons’, Ecological
Economics, 25(1): 73–87.
Luhmann, N. (1986) Ökologische Kommunikation, Opladen: Westdeutscher Verlag.
Marwell, G. and Ames, R.E. (1981) ‘Economists free-ride, does anyone else?’, Journal of
Public Economics, 15(3): 295–310.
McDaniels, T.L. and Roessler, C. (1998) ‘Multiattribute elicitation of wilderness preservation
benefits: a constructive approach’, Ecological Economics, 27 (3): 299–312.
Menger, C. (1871) Einführung in die Volkswirthschaftslehre: erster, allgemeiner Theil, Vienna:
Braumüller.
Mitchell, R.T. and Carson, R.C. (1989) Using Surveys to Value Public Goods: The contingent
valuation method, Washington DC: Resources for the Future.
O’Connor, M. (1997) ‘The internalization of environmental costs: implementing the
polluter pays principle in the European Union’, International Journal of Environment and
Pollution, 7(4): 450–82.
O’Hara, S. (1996) ‘Discursive ethics in ecosystem valuation and environmental policy’,
Peukert, H. (1998) Das Handlungsparadigma in der Nationalökonomie, Marburg: Metropolis Verlag.
Sagoff, M. (1988) The Economy of the Earth, New York: Cambridge University Press.
Sagoff, M. (1998) ‘Aggregation and deliberation in valuing environmental public goods: a
look beyond contingent pricing’, Ecological Economics, 24(2/3): 213–30.
Schkade, D.A. and Payne, J.W. (1994) ‘How people respond to contingent valuation
questions: a verbal protocol analysis of willingness to pay for an environmental
regulation’, Journal of Environmental Economics and Management, 26(1): 88–109.
Schulze, W., McClelland, G., Waldman, D. and Lazo, J. (1996) ‘Sources of bias in contingent
valuation’, in D.J. Bjornstad and J.R. Kahn (eds) The Contingent Valuation of Environmental
Resources, Cheltenham: Edward Elgar: 97–116.
Sen, A.K. (1977) ‘Rational fools: a critique of the behavioral foundations of economic
theory’, Phil. and Public Affairs, 6(4): 317–44.
Solow, R.M. (1993) ‘Sustainability: an economist’s perspective’, in R. Dorfman and N.S.
Dorfman (eds) Economics of the Environment, 3rd edn, New York: W.W. Norton.
Spash, C. and Hanley, N. (1995) ‘Preferences, information, and biodiversity preservation’,
Stevens, T.H., Echeverria, J., Glass, R.J., Hager, T. and More, T.A. (1991) ‘Measuring the
existence value of wildlife: what do CVM estimates really show?’, Land Economics, 67(4):
390–400.
Stevens, T.H., More, T.A. and Glass, R.J. (1994) ‘Interpretation and temporal stability of
CV bids for wildlife existence: a panel study’, Land Economics, 70 (3): 355–63.
Stork, P. and Viaene, J.-M. (1992) ‘Policy optimization by lexicographic preference ordering’,
Journal of Policy Modeling, 14 (5): 655–73.
Toman, M.A. (1999) ‘Sustainable decision making: the state of the art from an economics
perspective’, In M. O’Connor and C.L. Spash (eds) Valuation and the Environment: Theory,
Method and Practice, Cheltenham: Edward Elgar: 59–72.
Tversky, A., Sattath, S. and Slovic, P. (1988) ‘Contingent weighting in judgement and
choice’, Psychological Review, 95 (3): 371–84.
Vatn, A. and Bromley, D.W. (1994) ‘Choices without prices without apologies’, Journal of
Environmental Economics and Management, 26 (2): 129–48.
Non-use values and attitudes 51
3 Non-use values and

attitudes
Wetlands threatened by climate
change
Jürgen Meyerhoff
Introduction
Climate change is expected to cause extensive damage in coastal areas due to a
rise in sea level and increasingly frequent violent storms. Scientists expect that, in
particular, the nature and landscape of the German Wadden Sea, a coastal wetland
ecosystem, will be negatively influenced. The adjacent salt marshes, a habitat for
endangered species, have already been adversely affected in quality and quantity
by human interference, such as by dam construction and waste disposal into the
North Sea, and are expected to deteriorate even faster. However, it is possible to
mitigate the consequences of climate change by taking suitable measures such as
removing dams (Reise 1993, 1996). To obtain information about the population’s
support for the protection of the Wadden Sea against the consequences of climate
change, the willingness to pay (WTP) of German households was determined by
using the contingent valuation method (CVM).1 Several studies have shown that
non-use values can form an important part of the total economic value of
environmental assets (Garrod and Willis 1996; Bateman and Langford 1997),
particularly for highly unique resources. Therefore, non-use values were also
thought to form an important part of the total economic value of the Wadden
Sea. This area is a very valuable natural refuge, ranking third out of thirteen
national parks in Germany.
However, despite an ongoing debate among environmental economists about
the compatibility of non-use values with neoclassical theory, it is far from obvious
whether all bids from non-users express economic values and should therefore be
used in cost-benefit analyses. As a guideline, some authors require that it should at
least be possible to explain stated non-use values partly by predictors suggested by
economic theory. ‘If it is argued that non-use values are theoretically included in
economic benefits, these values should therefore be suitable for economic models
of choice, this would imply, among other things, that the value for a good with
non-use value should be partly explained by income, and possibly by other economic
determinants of value’ (Wierstra 1996: 118; also Blamey et al. 1995: 272). One
aim of this study was to clarify the relationship between attitudes and non-use
values, testing the results by Wierstra. In his study On the Domain of Contingent Valuation,
he found that the variation of stated non-use values could be explained solely
based on attitudes.
52 Jürgen Meyerhoff
This chapter is organized in the following way. In the second section, a brief
overview of the discussion about the relationship of attitudes and WTP is given.
The third section presents the design of the contingent valuation and its results in
relation to the use and non-use of the Wadden Sea area. Section four presents
findings on the relationship between measured attitudes and WTP, while in section
five the results of multiple analyses by SPSS-CHAID and OLS regression are
discussed. Further conclusions are drawn in section six.
Attitudes as predictors of WTP within the CVM

There is increasing debate about how knowledge of social psychology can and
should be incorporated into contingent valuation studies. In general, two different
approaches exist. On the one hand, economists are looking for ways to extend
their basic models in order to be able to explain stated WTP more comprehensively.
For example, Georgiou et al. (1998: 592) conclude: ‘These diverse influences upon
WTP suggest that simplistic models of preference formation are inadequate bases
for CVM studies. Rather an extended model of how individuals form preferences
is required’. This approach is also supported by the NOAA panel, as Kotchen and
Reiling (2000) point out. On the other hand, attitudes are used as a determinant
competing with economic preferences. Particularly when non-use values form an
important part of the economic value, there is some disagreement about whether
the stated amounts are due to preferences or whether they are just another measure
of attitude on an arbitrary scale (Kahneman et al. 1993; 1999).
A general definition of attitudes is given by Eagly and Chaiken (1993: 1). They
define attitudes as being ‘a psychological tendency that is expressed by evaluating
a particular entity with some degree of favour or disfavour’. In order to be able to
use attitudes to clarify the question of whether preferences are at work or not, it is
necessary to distinguish between different approaches of attitudes. Kaiser et al.
(1999: 2) differentiate between three attitude approaches in connection with
ecological behaviour: first, the Fishbein and Ajzen tradition with attitudes towards
a specified behaviour; second, attitudes towards the environment commonly
referred to as environmental concern; and, third, the New Environmental Paradigm
(NEP) as an alternative, single component measure of environmental attitude.
The theory of reasoned action by Fishbein and Ajzen (1975) and its successors as
well as the NEP by Dunlap and van Liere (1978) are at the centre of attention
within the debate. Although sometimes connected in CVM studies, both approaches
represent in the end different concepts of environmental attitudes.
The theory of reasoned action and its successors try to explain the relationship
between attitudes and behaviour. Based on this relationship, it should be possible
to predict actual behaviour. The main component of these theories is the
behavioural intention, which is influenced by attitudes towards a specified behaviour
and a subjective norm towards that behaviour. Under certain circumstances the
behavioural intention accurately predicts real behaviour. Among others, one
requirement is that the principle of compatibility is fulfilled (Ajzen 1996: 387; also
Bamberg 1996). This principle demands that attitudes and behaviour should be
measured at the same level of generality. In this case the likelihood increases that
the same set of salient beliefs will be activated and strong correlations between
attitudes and behaviour could be observed. Within the theory of reasoned action
the expression of WTP can be seen as an intention to behave in a certain way; that
is to actually pay the stated amount of money. Bishop and Heberlein were among
the first authors to point out the potential for CVM studies to incorporate attitudes
towards WTP. In accordance with their work the theory of reasoned action can
help to validate measured values and, therefore, supports the assumption of
articulated values. ‘Still, if conditions for high attitude–behaviour correspondence
are fulfilled, some grounds would exist for arguing that legitimate economic values
are being established at least to a rough approximation’ (Bishop and Heberlein
1986: 146).
An example of a combination of both CVM and theory of reasoned action is
the work by Moisseinen (1999) which examines the WTP of residents of an area
in Finland to prevent the extinction of the Saimaa seal. As the results of a former
CVM study on the protection of the Saimaa seal were contradictory in itself,
Moisseinen used, among other things, the theory of reasoned action to explain the
variations of WTP. Following this theory she asked interviewees not only about
their WTP, but also about their attitude towards it and about their subjective norm.2
With an extended behavioural model based on the theory of reasoned action and
the standard economic approach, it was possible to explain a significantly larger
amount of variation (adjusted R-square at a value of 0.43) than usual. Thus, as
these attitudes correspond to the ‘principle of compatibility’ they should be a good
predictor of the actual WTP.
The NEP is an alternative method used to measure environmental attitudes. In
its original version the scale consists of 12 items. According to Kaiser et al. (1999),
it represents a shift towards a more evaluative conception of attitude because its
proponents regard one’s moral values as the core concept of environmental attitude.
Therefore, NEP is an instrument to measure value orientations rather than attitudes.
Critics have pointed out that NEP and ecological behaviour correlate only weakly
at best. An example of a combination of both CVM and NEP is the study conducted
by Kotchen and Reiling (2000). They used an expanded and updated version of
the NEP measure to elicit the WTP of 1200 Maine residents for the protection of
two endangered species. From this study it could be shown that attitudes, measured
with the NEP scale, are a significant explanatory variable both of the general
WTP and also of the amount people were willing to pay. But although they are a
good explanatory variable of WTP it is necessary to keep in mind that these attitudes
are not a good predictor of behaviour. Rather it is the use of the NEP scale that
provides the possibility to measure what Fischhoff (1991) called basic values.
Using attitudes to explain WTP, Wierstra (1996) explored under what conditions
the CVM produces valid and reliable measures of value for environmental goods.
Wierstra had hypothesized that a reliable measure of value for environmental
goods by the CVM is dependent on the positioning of the good to be valued in the
dimensions of ‘use/non-use value’ and ‘concrete/abstract’. More specifically, it
was tested whether the quality of CVM results declined with increasing degree of
non-use value in the good and with increasing degree of abstraction in the good.
To test this hypothesis, he carried out three CVM surveys with different
environmental goods. In the case of the protection of the island of Rottumeroog
within the Wadden Sea off the Dutch coast, only non-use values could arise because
the island is uninhabited and visiting it is not allowed.
In addition to economic factors such as household size, net income, and fixed
housing expenses, Wierstra also determined attitudes from his interviewees.
However, although the theory of reasoned action is referred to, attitudes towards
a specific behaviour, such as WTP, were not determined. For example, in the case
of Rottumeroog, attitudes recorded were concerned more with subjects such as
the government’s areas of responsibility, cultural or historical meaning of the island,
and how respondents saw nature as being an important aspect of the island. For
this reason, however, the ‘principle of compatibility’ wasn’t fulfilled as these attitudes
are not suitable as a basis for statements about the relationship of attitude–
behavioural correspondence. Rather the elicited attitudes also represent some kind
of general attitudes which measure basic values. In the end, Wierstra found that
especially for the non-use good some attitudinal factors were sufficient to explain
the variation in WTP, while none of the economic factors were significant in this
case. Therefore, conditions for the validity of CVM results were not met because
the value of a good should at least be partly explained by income, and possibly by
other economic determinants of value.
Study design and main results

The aim of the contingent valuation survey was to place a monetary value on
measures which would allow the protection of the German Wadden Sea against
the consequences of climate change. Therefore, a random sample was drawn from
the German population, and 2390 households in Germany were contacted by
professional interviewers in April and May 1999. The sample was drawn in a
three-stage process: first, 140 sample points representing the size of German towns
were determined; second, random walk was used to select the household; and,
third, the kish-selection-grid was used to address the target person within the
household. In order to test for distance-decay effects, the sample was divided into
four groups, defined by concentric zones around the Wadden Sea (each with a
radius-increment of 220 km). For various reasons, such as ‘nobody at home’, or
‘household or target person not willing to participate in an interview’, 1412 face-
to-face interviews were completed.3 This corresponds to a response rate of 59 per
cent. In order to inform interviewees about climate change and its possible
consequences for the Wadden Sea, they were presented with various short texts, a
map of the Wadden Sea area, and a card with photos of parts of the Wadden Sea.
The card also contained a map illustrating the distribution of important biotypes
in the area and a short description of the main consequences of climate change
for the Wadden Sea.
As Bateman and Langford (1997: 572) point out, studies which try to measure
non-use values are presented with a definitional problem. Generally, non-use values
are associated with benefits from the knowledge that an environmental asset exists
or is maintained independent of any actual or potential use. Moreover, they are to
a large extent expected to be elicited by people who live far away from the site
involved and who never have and probably never will visit it (Brouwer and Slangen
1998). However, non-use values can also be attached to environmental assets by
people who have visited the site. Despite an ongoing debate, there is no generally
accepted approach to decompose values held by users into use and non-use values.
On the other hand, those who were, at the time of the survey, non-users may have
valued the protection of the Wadden Sea area even if they have no intention or
desire to visit it. Therefore, in this study non-use values are defined as the WTP
solely of non-users. Accordingly, use values are defined as the WTP of users. This
way of proceeding gives at least a well-founded lower bound of the non-use benefits
which arise from the environmental good (Silberman and Gerlowski 1992: 226).
Moreover, to inform politicians about the economic value of an environmental
good such as the Wadden Sea area it is not necessary to know which part of the
users’ WTP arises from non-use values. Crucial is, rather, the total economic value
which comprises user and non-users’ total WTP.
Users were, in contrast to non-users, defined as persons who have visited the
North Sea coast or any of the coastal islands at least once. In accordance with this
definition 49 per cent (695 people) out of the entire sample were users. Out of this
group more than 60 per cent had visited the North Sea coast within the last five
years, more than 80 per cent within the last 10 years. In addition, 72 per cent had
visited the coast more than once, implying that the North Sea coast is an important
area for recreation in Germany.4 On the other hand, 51 per cent (717 people) had
never visited this area before and were therefore taken as non-users. The spatial
distribution of users and non-users was as expected. The portion of users decreased
and that of non-users increased proportionally to the distance from the coast,
between 30 per cent non-users in zone 1 (mainly covered by Schleswig-Holstein
and Hamburg) to 65 per cent in zone 4 (covered by Bavaria and Baden-
Württemberg). As Table 3.1 shows, out of the whole sample 336 people or 25 per
cent were in principle willing to pay for the offered programme. This group consisted
of 225 users and 111 non-users.
Before the data were examined further, protest bids were excluded from the
following analysis. Protest responses are often determinated within CVM studies
on an ad hoc basis and in almost every CVM study different approaches are used to
determine these responses (Jorgensen et al. 1999: 135–7). However, all approaches
Table 3.1 Distribution of interviews and in-principle WTP

All User Non-user
WTP Yes No Yes No

Zone 1 300 61 148 3 88
Zone 2 407 74 153 23 157
Zone 3 403 46 108 32 217
Zone 4 302 44 61 53 144
Total 1412 225 470 111 606
share the aim to differentiate real zero bidders from people who stated zero bids
but for reasons other than not valuing the good in question. Protest responses are
normally excluded from further analysis because they do not represent true
economic values. Despite the fact that a lot of problems are associated with this
practice of censoring protest responses, a simple follow-up question was also used
in this study, asking the respondent to give a reason for not being willing to pay.
People who answered ‘current taxes deemed already too high’, ‘I do not trust
payments addressee’ or ‘not I but the state is responsible’ were determined as
protest voters and excluded from further analysis. Taken together, this results in 56
per cent (797 people) protest responses. This is, compared with other studies, a
relatively high amount of protest bids. For example, Spaninks et al. (1996) report
98 protest bids or 11 per cent out of 841 responses. One explanation for the high
rate of protest bids could be that shortly before the main survey an eco-tax on fuel
was introduced in Germany (1 April 1999). Out of the protest group, 62 per cent
(497 people) stated, at least as one of the two most important reasons for not being
willing to pay, that current taxes are already too high. Compared with the study by
Spaninks et al. it is also necessary to keep in mind that they used a mail survey
instead of personal interviews. As the 841 responses represent only a response rate
of slightly more then 30 per cent, many people who did not respond were perhaps
also ‘protesting’.
Within the remaining group of users, 67 per cent (225 people) were in general
willing to pay for the protection of the Wadden Sea while within the group of
remaining non-users only 37 per cent (108 people) were in general willing to pay.
As expected, the portion within non-users is substantially lower than the percentage
of users who are willing to pay. Also the mean amounts show significant differences.
While users are willing to pay DM 11.51 per household per month, non-users
were willing to pay DM 3.57. To test whether there is a statistically significant
difference in WTP between the two groups, the non-parametric Mann–Whitney
test was used. It requires no assumptions about the distribution and tests whether
the observations in two samples have the same distribution. The result (z = –8.38;
p < 0.001) allows the rejection of the null hypothesis of equal WTP of users and
non-users at a 1 per cent significance level. Table 3.2 summarizes the main results
of the WTP statistics for users and non-users pooled and for each group alone.
Table 3.2 Summary statistics of the WTP (in DM) for users and non-users
All User Non-user
n 1412 695 717
Protest votes 790 362 428
Used sample 622 333 289
Willing to pay 336 225 111
Mean WTP 7.8 11.1 3.6
Standard deviation 17.2 20.8 10.3
Median WTP 1.0 5 0
Mode WTP 0 0 0
Maximum WTP 170 170 100
Influence of attitudes on WTP
As mentioned in section one, it was the aim of this study to test Wierstra’s results
about the relationship between attitudes and WTP. According to him, it was
hypothesized that the conditions for economic values were not met if the WTP of
non-users could be sufficiently explained by attitudinal factors; in this case the
survey would support the results by Wierstra. Also following Wierstra, attitudes
towards more general objects such as nature and climate change were gathered.
The reason for this is that only people with enough information about the good
are in a position to hold site-specific attitudes. Less informed people, such as non-
users, will possibly construct attitudes (Slaby 1996). Similar to the problem of
constructed preferences, these attitudes will supply no sound basis for any analysis.
Therefore, the gathering of attitudes was designed more generally, allowing for
uninformed non-users to be considered. Figure 3.1 shows the conceptual model
which is assumed to determine the WTP. Attitudes were mainly elicited towards
three attitude objects: climate change, nature, and valuation of nature with money.
The following subsections present the items and results regarding attitudes towards
climate change and nature, and the results regarding attitudes towards money.
Attitude towards nature and climate change

The items which were used to measure the attitude towards climate change and
nature are presented in Tables 3.3 and 3.4. Attitudes were measured on a five-
point scale, ranging from ‘completely agree’ to ‘completely disagree’. The scales
were developed in accordance with items used in previous studies involving
contingent valuation (Goodman et al. 1996; Wierstra 1996), and in psychological
studies (Hofer 1987; Preisendörfer 1996). As a measure of reliability, Cronbach’s
• Nature
Attitudinal factors • Climate change
• Valuation with money
• User/non-user
-
WTP for protection
• Planned visits
of the Wadden Sea
Landscape-related • Knowledge about the
against
factors Wadden Sea
consequences of
• Interest in conservation
climate change
• Distance to Wadden Sea
• Household net income

Economic factors • Household size
• Age
Figure 3.1 A conceptual model of WTP determinants

Table 3.3 Attitudes towards nature
1. A nature conservation area is only valuable for me if it is allowed to be entered.
2. I am very confident that our children and grandchildren will experience nature in a
similar way as we do today.
3. We are allowed to form nature regarding only our own needs.
4. Animal and plant species threatened by extinction should be protected without
regard of the cost.*
5. Nature and landscape in Germany did not change during the last 20 years.
6. Spending time amidst nature is a repeatedly wonderful experience.*
7. More consumption goods can compensate the losses of nature and landscape.
8. To protect nature for its own sake should not be a goal of environmental politics.
9. For the protection of nature and landscape I would give up a part of my present
income.*
In the scale all except item 2 were taken into account; Cronbach’s alpha is 0.67.
Notes
Items marked with an asterisk were transformed before adding scores. The transformation of a
negative formulated item with scores 1 to 5 is: ‘new score’ = 6 – ‘original score’.
Table 3.4 Attitudes towards climate change

1. I myself can contribute no further to protect the climate.
2. Within the discussion on environmental problems the meaning of climate change is
exaggerated.
3. For us today the consequences of climate change are without meaning because they
occur only in a few decades.
4. Unless causes of climate change are not completely known, measures to protect the
climate should not be carried out.
5. For me it doesn’t matter if the following generations will suffer any disadvantages
from climate change.
6. I am very confident that policy will be successful in protecting our climate.
All items were taken into account; Cronbach’s alpha is 0.75.
alpha was calculated. It is the current standard statistic assessing the reliability of
a scale composed of multiple items and it is the most appropriate reliability measure
to use for Likert scales. Because it considers the degree to which items on a scale
intercorrelate, it is often referred to as a measure of internal consistency (Eagly
and Chaiken 1993: 67). The associated alpha is 0.67 for the scale ‘nature’, and
0.75 for the scale ‘climate change’. These low values of Cronbach’s alpha illustrate
the problem arising from the use of a scale which is not as well developed as the
NEP scale, for example. Compared with a generally accepted value of 0.80
(Schumann 1999: 42), the study fails to gather reliable attitudes. However, if
compared with the range of values for Cronbach’s alpha obtained in other studies,
our data prove reasonably reliable. Wierstra reported values of Cronbach’s alpha
of between 0.45 and 0.85. Diekmann (1997: 220–3) mentioned that in the General
Population Survey (Allgemeine Bevölkerungsumfrage (ZUMA-Allbus)) short scales
of four to six items were used and Cronbach’s alpha values of approximately 0.60
were obtained. Therefore, the gathered attitudes towards nature and towards
climate changes were subjected to further evaluation. Tables 3.5 and 3.6 illustrate
the relationship between the attitude score and the mean WTP.
Table 3.5 Attitudes towards nature and WTP
Score User and non-user User Non-user
Mean Std. N Mean Std. N Mean Std. N

up to 24 1.54 3.98 92 2.63 4.73 34 0.87 3.30 58
25 to 30 4.92 11.25 231 8.13 15.15 104 2.29 5.29 127
31 to 36 11.22 19.36 260 14.12 21.52 170 5.74 12.86 90
37 to 42 19.54 37.09 31 22.22 40.94 20 14.68 30.04 11
X2 93.85 p < 0.001 32.15 p < 0.001 42.25 p < 0.001
Table 3.6 Attitudes towards climate change and WTP

Score User and non-user User Non-user
Mean Std. N Mean Std. N Mean Std. N

up to 12 0.03 0.12 18 0.08 0.20 7 0.00 0.00 11
13 to 18 3.24 7.98 156 6.17 11.73 56 1.60 3.96 100
19 to 24 7.86 16.09 288 11.30 19.86 160 3.56 7.59 128
25 to 30 13.10 23.83 160 15.27 25.10 110 8.31 20.16 50
X2 61.9 p < 0.001 18.62 p < 0.001 27.59 p < 0.001
The relationship between attitudes towards nature and mean WTP is positive
for all three groups. The results confirm the expectation that people with stronger
attitudes towards nature are on average willing to pay more than people with
weaker attitudes. The mean WTP of users is for each level of the attitude score
higher than for the corresponding group of non-users. The last line of Table 3.5
presents results of the Kruskal–Wallis H-test. This non-parametric test is a test for
the null-hypothesis that the distribution of some variable (WTP in this case) is the
same in all samples defined by a certain grouping variable (attitudes towards nature
in this case). The results allow the rejection of the null-hypothesis of equal
distribution within each attitude class at a high significance level.
The figures in Table 3.6 show the same pattern as in Table 3.5. The mean WTP
is at each level of attitude score higher than at the preceding level. Again, mean
WTP of users is in each case higher than WTP of non-users. The results of the
Kruskal–Wallis H-test allow once more the rejection of the null-hypothesis of equal
distribution of WTP at a high significance level. Thus, attitudes towards climate
change also have, like attitudes towards nature, a significant influence on WTP. In
both cases, a higher score on attitude corresponds to a higher WTP on average.
In addition, the analysis of the relationship between attitudes and WTP confirms
two more findings presented by Kotchen and Reiling (2000). First, respondents
with stronger pro-environmental attitudes are more likely to participate in the
valuation procedure. People who were in-principle willing to pay show a higher
mean score on both the nature and the climate change scale. Second, people who
were defined as protest voters attain lower mean scores on both scales. The Mann–
Whitney test indicates that both differences are significant at the 0.1 per cent level.
Attitude towards money as a measure of value
As another expression of attitude, the degree of agreement with the statement ‘I
reject money as a measure to express the value of nature and landscape’ was used.
Nearly 45 per cent out of the whole sample ‘completely agree’ or ‘agree’ with this
statement, while 25 per cent ‘disagree’ or ‘completely disagree’. Slightly more
than 29 per cent are in between both of these groups. The relationship between
the stated answers and mean WTP for both users and non-users is given in Table
3.7. In both cases the results of the Kruskal–Wallis H-test allow the rejection of
the null-hypothesis of equal distribution of WTP within the defined samples. At a
first glance, it is interesting that also people who ‘completely agree’ or ‘agree’ with
the presented statement were in-principle willing to pay, and, especially in the
group of users, that their mean WTP is relatively high compared with the amounts
stated by people belonging to the other categories. However, a post hoc test (Sachs
1997: 395–7) qualifies these findings. In the group of users, their mean WTP does
not differ between those who ‘completely agree or agree’ or ‘neither agree nor
disagree’ at a 5 per cent significant level. In the group of non-users, only the mean
WTP of those who ‘disagree or completely disagree’ differs from the means of the
other samples. Nevertheless, the expected result remains unchanged. People who
accept money as a measure to express the value of nature and landscape show a
significantly higher mean WTP than those who reject money as such a measure.
Multiple analysis of WTP determinants

For a multiple analysis of factors influencing WTP, two tools were used. First,
CHAID (Chi-squared Automated Interaction Detector) was used to analyse the
relationship between the in-principle WTP and the potential explanatory variables.
Second, multiple regression equations were estimated using ordinary least squares.
This allows us to examine the relationship between the stated amount of WTP
and the independent variables. In both cases, equations for users and non-users
are presented separately. The aim of this analysis was to test whether use and non-
use values were influenced by totally different factors as Wierstra (1996) found.
Table 3.7 Relationship between attitudes towards money and mean WTP
‘I reject money as a measure to User Non-user
express the value of nature
and landscape’ n % per Mean n % per Mean
column column
Completely agree 41 12.3 11.7 59 20.4 3.4
Agree 91 27.2 6.6 95 32.9 2.2
Neither agree nor disagree 101 30.2 9.7 81 28.0 3.7
Disagree or 100 30.2 17.8 54 18.7 6.0
Completely disagree
χ2 (p) 30.4; df. 3; p < 0.001 24.77; df. 3; p < 0.001
Note: Because of a low number of observations in the group ‘completely disagree’, it was merged
with the group of people who ‘disagree’.
Unlike his study, both groups of users have to be analysed separately because only
one good with both characteristics was the subject of the CVM. Table 3.8 reports
the potential explanatory variables.
SPSS-CHAID analysis
SPSS-CHAID is a market segmentation technique that is designed for use with
categorical variables and continuous variables which have been divided into
categories (e.g. income, age). Based on the chi-square test, CHAID divides a
population at each level of the analysis into two or more segments that differ with
respect to a designated criterion. The analysis stops if the program is not able to
detect further significant relationships between the dependent and any remaining
variable or if a defined minimum of observations is reached. The first line of each
box in the tree diagrams in Figures 3.2 and 3.3 shows the corresponding variable.
The second line, except for the root of the tree, shows the category of the respective
variable by which CHAID divides the population. In the third line, the percentage
of people who are willing to pay within the subgroup is presented, while the fourth
line shows the total number of people within that group.
In both the user and the non-user group only a very limited number of variables
have influence on the dependent variable ‘in-principle WTP’. Among these
variables none is predicted by economic theory as an important predictor of WTP.
Rather, the attitudinal factors dominate. The attitude towards nature is in both
cases the most influential variable. People who show stronger attitudes towards
Table 3.8 Potential explanatory variables

Variable Description
AGE Respondent’s age (continuous variable)
INCOME Total monthly net household income (income was set equal to the
midpoint of the interval of the 9 intervals in which the respondent’s
income fell)
HSIZE Household size including children (continuous variable)
DISWAD Distance to Wadden Sea in kilometres (continuous variable)
KNOWS Previous knowledge of the Wadden Sea area as a natural landscape;
ranges from 1 (no knowledge) to 4 (thorough knowledge)
KNOWCL 1 if respondent had previous knowledge of negative impact of climate
change on Wadden Sea
TRAVEL Intention of respondents to travel to North Sea coast: 1 if respondent will
travel, 2 if respondent wishes to travel, 3 if respondent neither plans nor
wishes to travel
WADHI 1 if respondent has participated in a hike through the Wadden Sea area
CONSER Degree to which respondent is interested in conservation of Wadden Sea
area; ranges from 1 if respondent is very interested to 5 if respondent is
not interested at all
ATTNAT Attitudes towards nature (continuous variable)
ATTCLI Attitudes towards climate change (continuous variable)
ATTMON Attitudes towards valuation with money (reject money as a measure);
ranges from 1 if respondent completely agrees to 5 if respondent
completely disagrees
In-Principle
Willingness to Pay
yes: 37%
n = 289
ATTNA ATTNA ATTNA

up to 24 24–30 31–42
yes: 10% yes: 31% yes: 61%
n = 58 n = 127 n = 104
TRAVEL TRAVEL TRAVEL TRAVEL

1–2 3 1–2 3
yes: 48% yes: 22% yes: 10% yes: 50%
n = 41 n = 86 n = 50 n = 88
Figure 3.2 Tree diagram for non-users
In-Principle
Willingness to Pay
yes: 67%
n = 334
ATTNAT ATTNAT ATTNAT

19–24 25–30 31–54
yes: 38% yes: 58% yes: 77%
n = 34 n = 104 n = 196
ATTMON ATTMON
tot-agree – nn disag – tot-disag
yes: 71% yes: 88%
n = 123 n = 73
Figure 3.3 Tree diagram for users
nature are more likely to pay for the proposed programme to protect the Wadden
Sea. On the second level, variables differ between both models. In the case of the
non-user group, TRAVEL is the second most important variable. People who plan
a visit or wish to travel to the North Sea are more likely to pay than people who
neither plan nor wish to travel. In the case of the user group, the variable ATTMON
appears with two new segments merged out of the five-point scale. The more
non-users agree to value nature in monetary terms, the higher is the percentage of
people who are in-principle willing to pay.
However, the third attitudinal variable, that is attitudes towards climate change,
did not show any influence on the in-principal WTP answer either in the group of
users or in the group of non-users. One explanation could be that people perceived
the offered programme to protect the Wadden Sea as not being something that is
connected to climate change. It is supported by the finding that most of the people
who were willing to pay stated that they understood that their payment was not
solely to finance measures against consequences of climate change but to finance
a programme which would also protect the landscape against other negative
interventions.
Multiple regression analysis

Tables 3.9 and 3.10 report the results from the regression analysis for both the
group of non-users and the group of users. Various functional forms were fitted to
the data. The preferred models, which show the best statistical fit, are based on the
double-log form.
As the Breusch–Pagan test indicated presence of heteroscedasticity, the White
estimator was used and instead of the F-test the non-parametric Wald test was
conducted (Greene 2000: 506–11). At a value of 44.61, the Wald test shows that
all variables together have a significant influence on the dependent variable.
Furthermore, all variables in the non-user model show the expected sign. INCOME
is highly significant and shows a positive influence on the dependent variable.
AGE is also significant, but only at the 10 per cent level. In contrast, the third
variable out of the group of economic variables, HSIZE, is even at the 10 per cent
level not significant and was therefore removed from the equation. Out of the
group of landscape-related variables CONSERV (‘very interested’ and ‘interested’)
shows a positive influence and proves significant at the 1 per cent level. Finally,
ATTMON has a negative influence on the stated WTP. Non-users who completely
agree with the statement that they reject money as a term to value nature are less
willing to pay than people who agree with this statement at a lower degree or
disagree with it. Also ATTNA is highly significant and shows the expected positive
influence. In view of these results, the requirements set up by Wierstra seem to be
fulfilled. Stated bids by non-users are influenced at a significant level by explanatory
variables which economic theory predicts. As shown in Table 3.10, this is not the
case when bids stated by users are analysed.
As opposed to the non-user group, none of the predictors out of the group of
economic variables was significant, either at the 5 per cent level or at the 10 per
cent level. Therefore, these variables were removed completely from the equation.
Slightly more influential were some of the landscape-related variables. WADHI
and DISWAD prove significant at the 1 per cent level and at the 5 per cent level
and show the expected sign. Users who had participated in a hike through the
Wadden Sea area have a higher mean WTP than users who had not. Distance to
the Wadden Sea, in contrast to the non-user model, proves significant and supports
findings by Sutherland and Walsh (1993) or Pate and Loomis (1997) about the
negative relationship between distance and WTP. Similar to the non-user group,
both ATTNA and ATTMON show a high and significant influence. As the
Table 3.9 OLS results of non-user group
Coefficient Std. error t-ratio
CONSTANT –7.62 1.62 –4.71 ****
AGE –0.24 0.14 –1.75 *
INCOME 0.26 0.09 3.41 ****
CONSERV (very interested) 0.48 0.16 3.08 ***
CONSERV (interested) 0.39 0.12 3.21 ***
ATTNAT 2.12 0.41 5.21 ****
ATTMON (completely agree) –0.55 0.15 –3.81 ****
Notes:
**** significant at p < 0.001, *** significant at p < 0.01, ** significant at p < 0.05, * significant at
p < 0.1
Table 3.10 OLS results of user group

Coefficient Std. error t-ratio
CONSTANT –6.5261 1.4299 –4.5638 ****
WADHI 0.4127 0.1511 2.7311 ***
DISWAD –0.1717 0.0871 –1.9721 **
ATTNAT 2.6759 0.4268 6.2694 ****
ATTMON (completely agree) –0.6208 0.2343 –2.6492 ***
ATTMON (agree) –0.4727 0.1508 –3.1339 ***
Notes
**** significant at p < 0.001, *** significant at p < 0.01, ** significant at p < 0.05, * significant at
p < 0.1
R2 = 0.17; adjusted R2 = 0.16; Breusch–Pagan = 7.04 (5df); Wald = 17.29 (p <0.001)
Breusch–Pagan test indicated presence of heteroscedasticity also for this model,

the White estimator was used once again. The Wald test with a value of 17.29
allowed the rejection of the global hypothesis of no joint influence of the
explanatory variables. Compared with the non-user group, the coefficient of
determination is 9 per cent lower than in the coefficient in the user group.
In conclusion, one can say that the presented results support findings of other
studies such as those by Green et al. (1993), Spash (1998), or Kotchen and Reiling
(2000) about the influence of attitudes on WTP. The measured attitudes are
significant predictors of both in-principle WTP and mean WTP. Moreover, they
are also predictors in both user and non-user groups.
Discussion
The aim of this chapter was to test whether non-use values are only influenced by
attitudes or whether economic predictors influence these values as well. According
to Wierstra (1996), goods which have a high degree of non-use value could be out
of the domain of contingent valuation because stated WTP for these goods may
only be influenced by attitudes. His results about the WTP for the protection of
the uninhabited island of Rottumeroog support that hypothesis. As opposed to
this, contingent valuation of public benefits of protecting the Wadden Sea against
consequences of climate change shows exactly the opposite results. While the bids
stated by non-users were influenced among other things by economic variables, in
the case of users they show no influence at all. However, attitudes and selected
landscape-related variables are significant in both models.
Confronted only with the results of the presented analysis, one could draw the
conclusion that non-use values are influenced more by economic variables than
use values. According to Randall (1998) there is no crucial experiment which allows
us to answer the question ‘Does the CVM work?’ Therefore, the findings presented
in this chapter do not allow us to draw the conclusion that non-use values are
always influenced by economic variables. They are just another contribution to
the map of performance characteristics of contingent valuation. In this sense,
they support findings of several other studies that environmental attitudes are
significant explanatory variables of both principle WTP and of the amount of
WTP. Furthermore, they also raise the question of whether it is meaningful to
draw a dividing line between use and non-use values. Instead, it seems to be more
promising to study further which factors determine people’s WTP for collective
goods (Green 1997; Green and Tunstall 1999).
It is also important to examine in which case the Fishbein–Ajzen attitude-model
could be applied to non-users. Fishbein and Ajzen’s work has recently been
recognized as being important in environmental valuation studies. One example
of a study which was similar to their approach is that conducted by Moisseinen
(1999). However, her study aimed at measuring use values. To be in a position to
apply this approach to non-users, several aspects have to be considered. For example,
the theory of reasoned action demands that some requirements such as the ‘prin-
ciple of compatibility’ be fulfilled. But in the case of non-users it is not possible to
completely fulfil this requirement. As already mentioned, non-users are, normally,
people with less information about the good in question. Therefore, it could be
misleading to use the theory of reasoned action when non-users’ WTP is of interest.
The predictability of actual behaviour which follows the behavioural intention
is much weaker if instead of the theory of reasoned action a more general attitude
approach such as the NEP, the one used by Wierstra or in this study is applied.
Therefore, Bishop and Heberlein’s (1986) point that some grounds exist for arguing
that legitimate economic values are being established, at least to a rough approx-
imation, is perhaps no longer valid. A next step in the analysis of the relationship
between attitudes and WTP could be to develop and test models which incorporate
general attitudes (e.g. environmental concern) and attitudes towards specific
behaviour. Within these models, general attitudes are seen as determinants of
attitudes towards specific behaviour (Bamberg et al. 1999; Fransson and Gärling
1999). Furthermore, the amount of knowledge people have about the environmental
goods should be incorporated in more detail.
Notes
1 This was part of a study on costs of climate change under the project ‘Effects of climatic change
on the Island of Sylt: An interdisciplinary research project’. Further information is available at
http://soel.geographie.uni-kiel.de/sylt/english/Welcome.htm.
2 At this stage of the sampling procedure interviewees were not informed about the type of survey.
Therefore, people who refused to participate did not refuse because they don’t agree with
monetary valuation of environmental resources. They refused generally to participate in any
kind of survey.
3 The wording of the corresponding items is: ‘What is your attitude towards the extra payment to
protect the Saimaa seal? The idea of collecting money from the households is: good ´ bad’
(attitude towards the behaviour) and ‘Majority of people I know would participate by paying
the asked amount: likely ´ unlikely’ (subjective norm), (Moisseinen 1999: 202).
4 The definition of ‘users’ as people who have visited the North Sea coast at least once before the
interview is very broad because people who have ‘actually used’ the coast many years ago are
also defined as users. Further analysis showed no significant difference between people who
used the coast many years ago and people who have visited it only one or two years ago. Therefore,
this broader definition was applied during the whole evaluation.
References
Ajzen, I. (1996) ‘The directive influence of attitudes on behavior’ in P.M. Gollwitzer and
J.A. Bargh (eds) The Psychology of Action. Linking Cognition and Motivation to Behavior, New
York/London: The Guilford Press: 385–403.
Bamberg, S. (1996) ‘Allgemeine oder spezifische Einstellungen bei der Erklärung umwelt-
schonenden Verhaltens? Eine Erweiterung der Theorie des geplanten Handelns um
Einstellungen gegenüber Objekten?’, Zeitschrift für Sozialpsychologie, 27(1): 47–60.
Bamberg, S., Kühnel, S.M. and Schmidt, P. (1999) ‘The impact of general attitudes on
decisions’, Rationality and Society, 11(1): 5–25.
Bateman, I.J. and Langford, I.H. (1997) ‘Non-users’ willingness to pay for a national park:
an application and critique of the contingent valuation method’, Regional Studies, 31(6):
571–82.
Bishop, R.C. and Heberlein, T.A. (1986) ‘Does contingent valuation work?’, in R.C.
Cummings, D.S. Brookshire and W.D. Schulze (eds) Valuing Environmental Goods. An
Assessment of the Contingent Valuation Method, Totowa, NJ: Rowman & Allanheld: 123–47.
Blamey, R, Common, M. and Quiggin, J. (1995) ‘Respondents to contingent valuation
surveys: consumers or citizens?’, Australian Journal of Agricultural Economics, 39(3):
262–88.
Brouwer, R. and Slangen, L.H.G. (1998) ‘Contingent valuation of the public benefits of
agricultural wildlife management: the case of the Dutch peat meadow land’, European
Review of Agricultural Economics, 25(1): 53–72.
Diekmann, A. (1997) Empirische Sozialforschung, Reinbeck: Rowohlt.
Dunlap, R.E. and van Liere, K.D. (1978) ‘The “New Environmental Paradigm” ’, Journal
of Environmental Education, 9(4): 10–19.
Eagly, A.H. and Chaiken, S. (1993) The Psychology of Attitudes, Orlando: Harcourt Brace
Jovanovich College Publishers.
Fischhoff, B. (1991) ‘Value elicitation: is there anything there?’ American Psychologist, 46(8):
835–47.
Fishbein, M. and Ajzen, I. (1975) Belief, Attitude and Behaviour. An Introduction to Theory and
Research, Reading, MA: Addison-Wesley.
Fransson, N. and Gärling, T. (1999) ‘Environmental concern: conceptual definitions,
measurement methods, and research findings’, Journal of Environmental Psychology, 19(4):
369–82.
Garrod, G.D. and Willis, K.G. (1996) ‘Estimating the benefits of environmental
enhancement: a case study of the River Darent’, Journal of Environmental Planning and
Management, 39(2): 189–203.
Georgiou, S., Langford, I.H., Bateman, I.J. and Turner, R.K. (1998) ‘Determinants of
individuals’ willingness to pay for perceived reductions in environmental health risks: a
case study of bathing water quality’, Environment and Planning A, 30: 577–94.
Goodman, S., Seabrooke, W., Daniel, H. Jaffry, S. and James, H. (1996) Determination of
Non-use Values in Respect of Environmental and Other Benefits in the Coastal Zone, Portsmouth:
Centre for Coastal Zone Management.
Green, C. (1997) ‘Are blue whales really simply very large cups of coffee?’, paper presented
at Stockholm Water Symposium, Stockholm, January 1997.
Green, C. and Tunstall, S. (1999) ‘A psychological perspective’, in I.J. Bateman and K.G.
Willis (eds) Valuing Environmental Preferences. Theory and Practice of the Contingent Valuation
Method in the US, EU, and Developing Countries, Oxford: Oxford University Press: 207–58.
Green, C., Tunstall, S., Herring, M. and Sawyer, J. (1993) ‘Customer preferences and
willingness to pay for selected water and sewerage services’, a summary report to the
Office of Water Services, Middlesex: Flood Hazard Research Centre.
Greene, W.H. (2000) Econometric Analysis, New York: Macmillan.
Hofer, D. (1987) ‘Naturschutz als Wertobjekt. Eine exemplarische Studie über Einstellungen
zu Schutzgebieten’, PhD dissertation, Munich: Ludwig-Maximilians-Universität.
Jorgensen, B.S., Syme, G.J., Bishop, B.J. and Nancarrow, B.E. (1999) ‘Protest responses in
contingent valuation’, Environmental and Resource Economics, 14(1): 131–50.
Kahneman, D., Ritov, I., Jacowitz, K.E. and Grant, P. (1993) ‘Stated willingness to pay for
public goods. A psychological perspective’, Psychological Science, 4: 310–15.
Kahneman, D., Ritov, I. and Schkade, D. (1999) ‘Economic preferences or attitude
expressions? An analysis of dollar responses to public issues’, Journal of Risk and Uncertainty,
19(1–3): 203–35.
Kaiser, G.F., Wölfing, S. and Fuhrer, U. (1999) ‘Environmental attitude and ecological
behaviour’, Journal of Environmental Psychology, 19(1): 1–19.
Kotchen, M.J. and Reiling, S.D. (2000) ‘Environmental attitudes, motivations, and
contingent valuation of nonuse values: a case study involving endangered species’,
Moisseinen, E.M. (1999) ‘On behavioural intentions in the case of the Saimaa Seal.
Comparing the contingent valuation approach and the attitude-behaviour research’,
in M. O’Connor and C.L. Spash (eds) Valuation and the Environment. Theory, Method and
Practice, Cheltenham: Edward Elgar: 183–204.
Pate, J. and Loomis, J.B. (1997) ‘The effect of distance on willingness to pay values: a case
study of wetlands and salmon in California’, Ecological Economics, 20(3): 199–207.
Preisendörfer, P. (1996) ‘Umweltbewußtsein in Deutschland. Ergebnisse einer
repräsentativen Bevölkerungsbefragung 1996’, report for the Umweltbundesamt, Berlin:
Umweltbundesamt.
Randall, A. (1998) ‘Beyond the crucial experiment: mapping the performance characteristics
of contingent valuation’, Resource and Energy Economics, 20(2): 197–206.
Reise, K. (1993) ‘Verschwommene Zukunft der Nordseewatten’, in H.-J. Schellnhuber
and H. Sterr (eds) Klimaänderung und Küste, Berlin: Springer: 223–9.
Reise, K. (1996) ‘Wattökologische Folgen bei Änderung von Klima und Küste’, Schriftenreihe
der Schutzgemeinschaft Deutsche Nordseeküste, 1: 31–45.
Sachs, L. (1997) Angewandte Statistik, Berlin: Springer.
Schumann, S. (1999) Repräsentative Umfrage, München: Oldenbourg.
Silberman, J. and Gerlowski, D.A. (1992) ‘Estimating existence value for users and nonusers
of New Jersey beaches’, Land Economics, 68(2): 225–36.
Slaby, M. (1996) ‘Einstellungsmessung oder Einstellungsgenerierung? Die Bedeutung der
informationellen Basis bei Befragten für die empirische Rekonstruktion von
Einstellungen zu gentechnischen Anwendungen (Measuring or generating attitudes?
Subjective knowledge base for the empirical reconstruction of attitudes towards new
applications of genetic engineering)’, Stuttgart: Institut für Sozialforschung of the
Universität Stuttgart.
Spaninks, F.A., Kuik, O. and Hoogeven, J.G.M. (1996) ‘Willingness to pay of Dutch
households for a natural Wadden Sea. An application of the contingent valuation
method’, Amsterdam: Instituut voor Milieuvraagstukken of the Free University
Amsterdam.
Spash, C.L. (1998) ‘Environmental values and wetland ecosystems: CVM, ethics and
attitudes’, in M. O’Connor (ed.) The VALSE Project: Social Processes for Environmental
Valuation, full final report to the DG-XII, European Commission, chapter 6–B.
Sutherland, R.J. and Walsh, R.G. (1993) ‘Effect of distance on the preservation value of
water quality’, Land Economics, 61: 281–91.
Wierstra, E. (1996) On the Domain of Contingent Valuation, Enschede: Twente University Press.
Modelling environmental behaviour 69
4 Modelling environmental
behaviour
Socio-psychological simulation
Hans-Joachim Mosler
Introduction
In economically highly developed countries, many of the conditions that would
allow people to behave in environmentally responsible ways are already in place.
We have a lot of knowledge; for years surveys have shown that people give top
priority to the need to act on environmental issues. The necessary technical and
economic resources are also available. But there is little sign of a real about-face
except in limited areas. We believe that the much-cited discrepancy between
cognition and behaviour, between lip service and a person’s own contribution to
conserving the environment, can be better understood if we also take people’s
perceptions of the social surround into consideration. To view human beings as
Homo economicus (Diekmann 1996; Harsanyi 1977) falls short, because the economic
man approach does not explicitly take into account inner psychological factors,
such as motives, values, norms and attitudes, nor does it consider social influences
on a person’s behaviour. The present research conceives a model of behaviour
that in addition to economic factors includes personal and social factors. Environ-
mental consciousness is determined to a significant degree by social systems, or in
other words, by people’s corresponding social representations. An individual’s
personal contribution seems insignificant in the face of massive destruction of the
environment caused by many. This perception – that there is nothing we can do
personally, that each one of us is powerless – as well as a reluctance to be the
‘sucker’, are important causal factors in behaviours that overuse environmental
resources. It does not seem rational to exercise personal restraint (for example, by
not driving), because not only will we suffer from the harm caused by the general
public’s overuse (the consequences of air pollution), but from a reduction in our
own direct return as well (time saved, comfort). However, as this state of affairs
applies equally to all individuals in a society overusing environmental resources,
people mutually trap each other in patterns of actions that harm the environment.
It is for this reason that we are particularly interested in examining the psychological
conditions that would form the basis of a collective reorientation towards environ-
mentally sustainable behaviour. Starting out from new environmentally friendly
behaviours of some ‘pioneer’ individuals, we wish to discover the social psycholog-
ical conditions that would ensure that the number of persons joining ranks with
70 Hans-Joachim Mosler
such pioneers would continue to automatically increase and result in a true, large-
scale ‘turn-around’ of previous, environmentally harmful patterns of behaviour.
The focal question of our research can thus be framed as follows: what are the
conditions that foster widespread, effective inner dynamics that change collective
environmental overuse (in thinking and action) to collective, environmentally
responsible thinking and acting? Findings generated by this research will lend
scientific support to the planning of environmental protection campaigns.
Method
Procedure
If we start from the assumption that environmental problems originate in the
overusing behaviours of very many individuals, we need to consider how new
solutions might be tested in a problem area of this magnitude. The instrument of
the questionnaire, based upon imaginary situations or conditions, seems ill-suited
(‘How expensive would gasoline have to be for you to change to public
transportation?’). Massive field experiments that translate the issue directly in real
social systems can also be eliminated; given our present state of scientific knowledge,
such experiments would be both financially and ethically irresponsible. Laboratory
experiments, which would not require intervention in existing social systems, cannot
be carried out with large groups of persons (1000 and more). Computer simulation
provides a possible solution. Simulation aims to reconstruct the relevant cause-
and-effect relationships in a problem area in the form of a model. With the aid of
empirical data, the relationships can then be validated. In this way, we can test
‘experimentally’ the most various and unconventional ideas of ways to spread
environmentally responsible thinking without incurring the risk of intervention in
real social systems.
It is important to remember, however, that simulation results are not empirical
results found in reality. Simulation results can generate fruitful hypotheses that
must be tested in reality. The great advantage is that the hypotheses are derived
consistently using a clearly defined procedure. Another important point to consider
is that findings gained through the simulation method allow only comparative
conclusions, for example that under the conditions studied, Measure A produced
better results than Measure B. The simulation method, furthermore, does not
allow statements to be made about the means to be used to implement certain
measures. For this, real experiments with real people are required. Despite all of
this, computer simulation is the only method available for investigating, by means
of many systematically varied experiments, how environmentally responsible
changes can be instituted in large populations.
Our simulation procedure consisted of the following steps.
Following a preliminary selection of the most important, empirically well-
founded theoretical approaches within the field of social influence, the theories’
most relevant and significant core statements were formulated, according to content
criteria.
In the ensuing modelling, the core statements of the theories were described
with the existing variables and set in relation to one another according to certain
systems-theoretical rules (see ‘Modelling of the theoretical concepts’). For lack of
space, we here dispense with discussion of programming/technical implement-
ations of the theories (see Mosler 2000 for details).
The design of the simulation model was validated through the aid of experts’
evaluations and replication of findings from the fields of environmental and social
psychology. Again, due to space considerations, the validation process is not reported
here (cf. Mosler 2000 for details).
Experiments with various strategies for the spread of behaviours were then
conducted. The following will report on the most important and meaningful
experiments.
From the findings of the simulations, conclusions were drawn pertinent to both
environmental practice and basic research. The most well known forms of inter-
vention stemming from environmental research receive a new interpretation; well-
founded recommendations for the field of intervention are derived.
The reader may wish to note that the following deals intensively with psycho-
logical concepts (to psychologists, however, the descriptions will appear somewhat
superficial). The author sees the present chapter as a contribution that takes a
perspective that differs greatly from the economic viewpoint, although some relevant
connections are pointed out. This becomes clear in the next section. The reader
interested in the exact derivation of the findings should read all sections, while
others may wish to read only the sections on the simulation experiments and results.
The most important findings are reported and discussed in the concluding section.
The framework model

Very many simulations work from the start at the level of collective variables
(macroanalytic simulation, such as of, for example, the influence of prices on total
consumption by the public). However, if we aim to tap into the dynamics of the
spread, or dissemination (or non-spread, respectively), of environmentally
responsible thinking and behaving, we must first of all begin at the level of
individuals. We then allow them to interact with each other within the framework
of the simulation (microanalytic aggregative simulation). The difficult task is to
model the relevant processes of the individual in this area of conflict on the
computer. We assume that humans are in principle beings that are free to make
their own decisions, and we hold that a computer program cannot represent the
richness of human individuality, as it is primitive in comparison. On the other
hand, in many areas we do find empirically well-founded uniformities in human
behaviour. Our way of proceeding springs from the will and hope to develop,
from such empirically proven knowledge, a useful working model of the processes
taking place within the individual. Usefulness is measured according to whether
or not the model can be validated by the behaviour of real collectives and in terms
of whether the simulation based upon the model widens and furthers our
understanding of the dynamics of these processes.
For this reason, we first design a basic model of an individual (see Mosler 2000
for details), which – as 10,000 identically structured copies (equipped, however,
with individually different characteristics) – serves as the basis for the simulated
influencing and resource-use processes (see Figure 4.1). These persons differ
individually only in their values of the variables. They all function according to
the same social psychological principles. These principles are based on a few central
and well-founded essences of theories, which are presented in detail in the
presentation of results. The framework model first specifies the input and output
variables of the theory-based processing by individuals. The model yields inform-
ation about the inner psychological processes that take place when people use
environmental resources (for example, use the resource of air when they drive or
heat their homes) or influence each other mutually, whether deliberately or
unintentionally. Inner psychological processes are triggered as people communicate
with one another in daily life and observe themselves and others. These processes
change, depending on further internal and external conditions, the ways in which
we feel, think, and argue about the environment and the way we act toward the
environment. These simulated individuals have at their disposal differently
structured social contact nets (number of friends, acquaintances, neighbours, and
strangers they observe).
In its basic form, the model is an expanded ‘rational choice’ model (Diekmann
1996; Becker 1976). This means that the variables and processes specified in the
model more or less consciously enter into the decisions relevant to using behaviour.
The variables and processes thus having an effect are not, however, purely cognitive
components. They include, in addition, motives, emotions, norms, and so on. This
is what distinguishes our model from the usual ‘rational choice’ models: with a
psychological conception, it is not a problem that non-rational components enter
into rational decision-making. The following explains the external and internal
input variables as well as the inner processes.
External input variables are the influences that exert upon a person from the
outside and are perceived by the person in some form. Possible distorted perceptions
of individuals on the basis of biases are not taken into consideration in this model.
For each individual some output variables are calculated that function as external
input variables for other persons with whom the individual has contact (‘contacts’)
(see Figure 4.1). Also, some non-social input variables, such as specific situational
parameters and the momentary state of the resource being used, enter into the
calculation of an individual’s output value.
• Use, contact: A summary of all environmentally related behaviours shown by a

contact. These behaviours towards the environment are conceived as resource-
using behaviours. For resource utilization, a particular resource can be entered
into the model with its characteristic parameters (resources: for example, water,
air, wild game populations; parameters: for example, rate of regeneration).
• Attitude, contact: A summary of a contact’s opinions and evaluations of
environmental issues, objects, and behaviour towards the environment, as
expressed in various ways.
Figure 4.1 Framework model of behaviour toward the environment
Modelling environmental behaviour
73
• Status, contact: This variable represents the sum of a contact’s relevant personal
resources (for example, social competence, trust, prestige, knowledge, power,
possessions).
• Persuasiveness, contact: The intensity and quality with which a contact makes a
case for various attitudes related to the environment. While direction is given
for an attitude (in other words, ‘for’ or ‘against’ environmental protection),
persuasiveness indicates the intensity and quality with which these attitudes
are presented.
• Situational factors and incentives: These include all influences from the societal-
institutional surround (rules and prohibitions, positive and negative incentives,
and the like). From the entire spectrum of the possible scope of a behaviour,
situational factors ‘filter’ a partial spectrum.
• State of resource: State of the environment, or state of a particular environmental
resource, as ‘noticed’ and established by the person. Internal input variables
have an effect on the simulated psychological processes ‘from the inside’. The
emergence of individual degrees of markedness of these variables due to the
individual’s learning history is not examined in this simulation.
• Values: Stable orientation with regard to environmental facts, objects, and
behaviours.
• Knowledge: Extent and quality of information about the environment, such as
knowledge of the regeneration parameters of a specific resource.
• Self-responsibility: Describes the extent to which people attribute responsibility
to themselves for the state of the environment (as opposed to holding other
persons, organizations, or institutions responsible).
• Motives: A summary of various motives, such as curiosity and laziness, that
enter into a person’s readiness to act in certain ways. Dependent upon these
motives, an act – independently of its consequences – will be viewed as ‘easy’
or ‘desirable’.
External and internal input variables are processed in different sub-models (see
Figure 4.1) according to the theory being applied. The output values of the sub-
models have either a direct outward effect on other persons with whom the
individual has contact (attitude, persuasiveness), or they affect variables leading
up to behaviour and decisions on the use of resources and consequent actions.
The following sub-models were designed and simulated (simulations of sub-models
II and IV will not be presented):
• Processing of group influences upon attitudes towards the environment:

application of the theory of Social Comparison Processes.
• Processing of discrepancy between environmental behaviour and
environmental attitude: application of Dissonance Theory.
• Processing of observation of others’ behaviour towards the environment:
application of Bandura’s Social Learning Theory.
• Processing of resistance to environmental protection measures: application
of Reactance Theory.
• Processing of information on the collective utilization of an environmental
resource: application of Commons Dilemma research.
• Processing of communicative influence in view of the individual’s feeling of
concerned consternation about the state of the environment, knowledge about
the environment, and biases: application of the Elaboration Likelihood Model.
Variables leading to behaviour are direct preliminary steps towards behaviour. On

the basis of the Theory of Resource Mobilization (Klandermans 1984) and of the
Theory of Planned Behaviour (Frey et al. 1993b; Ajzen 1991; Ajzen and Madden
1986), we start from the assumption that the following five components play an
important role in behavioural intentions relating to the environment:
• Attitude towards the environment as the attitude towards environmental

protection and individual behaviour regarding the environment.
• The subjective norm as ‘perception of pressure from the social surround’.
This expresses the expectations that confront the individual and the degree to
which the individual is willing to fulfil these expectations.
• Behavioural control as the subjective conviction that one can in fact carry out
the behaviour.
• Cost–benefit analysis as a motive for behaviours, which results from weighing
the direct costs and direct benefits of actions.
• Sustainability as a motive, leading to willingness to restrict personal use of a
resource. This readiness depends upon both the absolute value that a person
places on an environmental resource and on the current discrepancy between
collective, sustainable use and the actual pattern of use shown by others. The
higher the subjective value of the environment, or the smaller the discrepancy
between actual and sustainable patterns of utilization, the greater the effect
of the sustainability motive in the sense of a person’s own restraint in use.
All of these variables are conceptualized at a general level, thus according with
Ajzen and Fishbein’s (1977) correspondence hypothesis, which states that there is
a high correlation between behaviour and the determining variables if they have
been operationalized with the same degree of specification.
As a fundamental extension of the theory of planned behaviour (Ajzen and
Madden 1986), we have added an additional factor, cost–benefit analysis. Although
the theory of planned behaviour does contain a kind of cost–benefit analysis, in
that its components are broken down in ‘expectation evaluation products,’ we
conceptualise this factor as an independent component. With this approach we
expect to achieve better explanations of the variance for intention, but more
importantly, to gain added starting points for intervention measures. Cost–benefit
analysis models people’s ‘economic’ considerations, as they determine whether it
will pay for them to act in certain ways. The model is thus the first to integrate
economic and social psychological components (see also Mosler and Tobias 2000).
These components all influence behavioural intention, which leads to behaviour
towards the environment. Additionally, we introduce volition into the model, which
is closer to actual behaviour than behavioural intention. Psychological research
(Frey et al. 1993b) has shown that the act of making a decision commits a person to
the when and where of attempting to realise an intention. As the entire model is
deterministic in construction, random variables play absolutely no role in deter-
mining individual behaviour towards the environment.
Simulation experiments and results

In the following, after a brief introduction of the various sub-models, the conditions
for the simulation experiments are described, and the results are presented. In
describing the sub-models, we focus on experiments with populations of 10,000
persons and omit the experiments that we conducted with 1 to 10 persons in order
to reach a basic understanding of the processes and to validate the model with the
aid of existing empirical findings. All variables range within a dimension from 0 to
100. For variables related to environmental orientation, 100 signifies maximum
environmental friendliness, and 0 indicates maximum lack of environmental
friendliness; 50 represents a point of ‘neither/nor’. In the graphics presenting the
findings, average behaviour towards the environment, or the average attitude
towards the environment by the population, is always put on the ordinate. If a
curve rises, this means that the population is changing in the direction of
environmental responsibility. On the abscissa, we find the simulation steps. A step
means that in a complete calculation cycle, all individuals exert influence on their
contacts, or are themselves influenced, and – together with the others – utilize a
resource. The linear progression of the steps should be interpreted as the progression
of time; more exact temporal pinpointing is not possible.
For the experiments, we usually started out from populations that were rather
eco-unfriendly (mean of 40), whereby the values of this variable in the individuals
strew around this mean. In order to avoid the reproach that we may have built the
results of the simulation right into the program, all experiments were conducted
according to a baseline/control group design. The results of a number of steps
with and without manipulations (control group) form the basis for comparison;
populations were always identical (which is only possible in computer simulation).
In the following, we will examine important issues in social intervention and
campaign planning with the aid of the simulation model. The following sections
on the sub-models are each organized according to a research issue and a reference
to the theories being applied.
Group influences upon attitudes

How must a minority of people behaving in an environmentally friendly way be
distributed and networked within a population in order that the environmentally
unfriendly majority comes to change its attitudes and behaviour? Frey (1993a:
114ff.) developed an integrative concept of social comparison processes that
encompassed Festinger’s theory (1954), Tajfel’s theory of social identity (1978,
1979; cited in Frey et al. 1993a), and Thibaut and Kelley’s theoretical concept of
the comparison level for alternatives (1959; cited in Frey et al. 1993a). According
to the integrative conception of social comparison processes (Frey et al. 1993a:
111ff.), persons in groups change their own attitudes in dependency upon the
existing pressure to conform, the attractiveness of the group, and the perception
of threat to their self-concepts should they change their own positions. A change
in attitude results from the pressure to conform, weighted by the attractiveness of
the group. If, however, the person’s self-concept is threatened, there will be no
attitude change (see Figure 4.2).
Threats to a person’s self-concept arise when their values differ to a certain
degree (a value of 30) from the average attitudes of other members of the group.
Individuals experience too great a discrepancy between their own values and the
group’s attitude as threatening: in order to adapt to the group, the individual would
have to deviate too far from personal values. Under social pressure, they would
become untrue to themselves. People are more likely to change an attitude the
more attractive the group is and the greater the group pressure to conform. Attitudes
will change in the direction of the average attitude of the group. Attitude change
occurs, however, only if the change does not pose a threat to self-concept. The
attractiveness of the group results from two components: (1) the difference between
a person’s own attitude and the average attitude of the group – a group is all the
more attractive the more that person and group are on the same ‘wavelength’; and
(2) the average status of the group – its prestige, power, social resources. The
conformity pressure of a group can be calculated according to Tanford and Penrod’s
(1984) well-documented formula. Pressure to conform, in reference to the total
INDIVIDUAL
PROCESSING OF GROUP INFLUENCES
Values Threat to self-concept
Attitudes of contact Pressure to

persons conform Attitude
Status of contact
persons Attractiveness of group
Figure 4.2 Representation of the ‘processing of group influences’

number of group members, is calculated in an exponential function separately for
a minority and a majority member.
The population is structured in groups of 10 persons, whereby each group is in
contact with two other groups. For the intervention, we assume that an
environmental campaign can bring 10 per cent of the population to clearly embrace
more friendly attitudes towards the environment than previously (an increase in
25 points), for a limited period of time (for example, for one week = one step in the
simulation). These persons we call (attitude) ‘pioneers’. Pioneers are, moreover,
more strongly sensitized to deviations from their own positions. Through this, they
acquire a particularly low susceptibility to social influencing attempts: even minor
deviations of other individuals in the group from their own, pioneer positions
imply a potential endangering of their sense of self-esteem. This leads to the less
susceptible attitudes of the ‘pioneers’ in the group.
As a further form of intervention, pioneers are variously organized. Following
the short environmental campaign, they remain in their groups (isolated) with no
wider contacts to the outside; they remain in their groups, but are in contact with
other pioneers (networks); or they form their own groups (core groups), which may
have few or many contacts with other groups. Figure 4.3 documents the experiments
conducted as well as their results. The intervention applies only to the fifth step;
the dynamics described below result from this short-term intervention.
It is extraordinarily important for pioneers’ susceptibility to be low, in order
that they do not too soon once again adapt to the surrounding majority. If
susceptibility is not low, there will be no positive effects from being networked with
other pioneers and from no longer experiencing themselves as a minority. Core
groups (also with low susceptibility) only then have an effect on the surrounding
majorities in the group if they have numerous outer contacts (curve F in Figure
4.3 continues to climb in further steps, but not as high as curves B and D). In sum,
pioneers have to hold to a stronger environmentally responsible attitude, and their
susceptibility to influence must be low. In effect, they feel quickly threatened in
their self-concept. Pioneers should be activated or active in as many groups as
possible and should not be concentrated in just a few groups. Further measures
are not required.
Social learning of behaviour

What factors lead the population to follow the examples of behavioural role models?
According to Bandura (1979) people learn not only through direct personal
experience, but also through observing others. In this way they gain an idea of
how an action is performed. The probability that a learned behaviour will be
carried out depends on motivation, and motivation is dependent upon expectations
of efficacy (self-efficacy, Bandura 1986: 390ff.) and determined by anticipated self-
regulation (Stalder 1985; Bandura 1986: 390ff.). Bandura does not, however, give
any indication of the effect on a person’s behavioural repertory when there are
several, contradictory behavioural models. For example, someone might observe
environmentally responsible behaviour in some people and behaviour that is
Environmentally
responsible
55
50
Attitude of the population
45
40
35
Not 0 5 10 15 20 25
environmentally
Runs
responsible
Intervention
Control population D: Networked ‘pioneers’ with

low susceptibility
A: Isolated ‘pioneers’, E: ‘Pioneer’ core groups, few
normal susceptibility contacts, low susceptibility
B: Isolated ‘pioneers’, F: ‘Pioneer’ core groups, many
low susceptibility contacts, low susceptibility
C: Networked ‘pioneers’ with
normal susceptibility
A. In each group there is an isolated pioneer whose susceptibility to influence is the same in degree
as that of the members of the group majority.
B. In each group there is an isolated pioneer whose susceptibility is lower than that of the members
of the group majority.
C. In each group there is a pioneer whose susceptibiity is the same in degree to that of group
members; in addition, the pioneer is in contact with 10 other ‘pioneers’ in other groups. They
thus no longer experience themselves to be a minority.
D. As in C above, but here pioneers have a lower degree of susceptibility.
E. All pioneers are concentrated within their own core groups and show low susceptibility to
influence. Each pioneer is in contact with another person outside of the core group.
F. As in E above, whereby here pioneer group members are in contact with 10 other persons
outside of the core group.
Figure 4.3 Change in average attitude due to various group-oriented interventions
damaging to the environment in others. In order to model contradictory influences

on self-efficacy, we turned to Latané’s theory of social impact (Latané 1981; Latané
and Wolf 1981; further specified in Nowak and Latané 1994; Nowak et al. 1990).
In this theory, social influence bases upon the following factors, which stand in a
multiplicative relationship:
• Strength: power, importance, intensity, unusual quality or features of the source

person for the target person.
• Immediacy: directness, immediacy in space and time, absence of barriers or
filters.
• Number of sources: number of group members, number of persons present.
The multiplicative relation of the three variables expresses the fact that the effect
of one of the variables is greater, the greater the value of the other variables.
There is no effect at all if one of the variables equals zero. According to the theory,
moreover, the effect of the variable N (number) is not linear, but rather is an
exponential function: I = sN t, where I is impact, s is a constant, and the exponent
t is a value less than 1. The parameters s and t are different for each situation and
have to be determined empirically. The factor ‘number’ thus has the effect that the
first person has the greatest impact and each person thereafter has ever less of an
impact. With an increasing number of influencing source persons, the social impact
on a person rapidly decreases.
Expectations of efficacy can be approximately represented by behaviour control,
because we define behaviour control as a person’s conviction that they can indeed
execute the behaviour. In accordance with the framework model, behaviour control
affects the behavioural intention and, thus, behaviour. According to Latané’s theory
of social impact (1981), behaviour arises from observation of the using behaviour
of contact persons: social influence is a multiplicative function of the number,
strength, and immediacy of contact persons. In our model (see Figure 4.4), the
variable ‘number’ is determined by the number of the individual’s contact persons
who behave in an environmentally responsible, or irresponsible, way. The ‘strength’
of contact persons’ influence is defined by their status. ‘Immediacy’ is conceived
INDIVIDUAL
PROCESSING OF OBSERVED BEHAVIOUR
Using behaviour
Using
behaviour of Behaviour control
contact persons Intention
Status of contact persons
Values Self-confirmation
Volition
Self-responsibility
Using
behaviour
Figure 4.4 Representation of ‘processing of observed behaviour towards the environment’

of as psychological immediacy. This means that friends have an influence of 1/1,
acquaintances of 3/4, neighbours of 1/2, and strangers of 1/4. The social influence
of contact persons who behave in an environmentally responsible way and that of
those who behave irresponsibly are calculated separately according to the formula
and then compared by dividing them (compare Nowak and Latané 1994; Nowak
et al. 1990). This ratio determines the behaviour control of the person: a person’s
belief that they are able to perform the concrete environmentally friendly behaviour
(that is, their behaviour control is the more environmentally responsible) is stronger
the greater the number and the higher the status of individuals behaving environ-
mentally responsibly in comparison to persons not behaving environmentally
responsibly that they observe.
Self-confirmation arises when people compare their own values (personal choice
of behavioural standards) with their behaviour. Self-confirmation is also weighted
by self-responsibility in order to incorporate the element of feeling responsible. A
particular behaviour is all the more important to a person the better it accords
with the person’s values and the more strongly the person feels responsible. Self-
confirmation affects volition and, thus, behaviour.
Within the framework of an intervention, incentives can move persons having
above-average status, for example, to change their behaviours during the intervention
phase in the direction of environmental acceptability. On the other hand, it is possible
to raise the status of persons displaying above-average environmentally friendly
patterns of behaviour through various measures (for example, by means of public
commendation awards, coverage in the media, and so on). The experiments were
conducted with 300 or 500 persons and included varying degrees of ‘visibility’, in
that selected persons displayed their environmental behaviours to many (15) or a few
(5) contacts (Figure 4.5). Visibility of resource use has already been demonstrated to
be an effective factor (Jorgenson and Papciak 1981; Mosler 1993).
A. In this experiment, 500 persons of high status are selected and caused to
adopt a behaviour that is more eco-friendly for the duration of the intervention.
They are to demonstrate this behaviour to many others (15 contacts).
B. As in A above, but here only 300 persons are selected.
C. As in A, but here the behaviour is demonstrated to only 5 contacts.
D. Here we selected 500 persons showing very environmentally sustaining
behaviour patterns, which they demonstrate to 15 contacts. In this case, their
status is raised for the duration of the intervention.
It appears to be more effective to select persons having high status as role models
and to induce them to behave in a more environmentally friendly way for the
duration of the intervention than to temporarily raise the status of persons already
showing such behaviours. An increase in the number of contacts, that is, an increase
in the visibility of the environmentally sound behaviour, achieves relevant effects.
An increased number of role models also has a strong effect. The dissemination
process continues to progress automatically for some time after the intervention,
as the entire social system must again adapt to the changes.
Environmentally
responsible 56
A. 15 CP/500 P/Act+
54
B. 15 CP/300 P/Act +
52
C. 5 CP/500 P/Act +
50
D. 15 CP/500 P/Status +
48
46
44
Control situation
42
Not 40
environmentally 0 5 10 15 20 25 30 35 40 45 50 Runs
responsible
Intervention
CP: contacts per person; P: number of persons addressed (role models); Act+, Status+: more responsible acts towards environment.
Figure 4.5 Average population behaviour with differing interventions based on observational learning
In developing concepts for environmental campaigns, simulation can contribute
support with regards to the implementation of efficient means. It can answer
questions as to the number of role models required, how these should be selected,
and what changes they should show, the degree of visibility necessary, and so on.
In particular, simulation allows us to estimate the effects that can be achieved by
means of compensation in other values, which makes a direct contribution to an
increase in efficiency (for example, an increase in visibility at the expense of the
number of role models). However, the simulation method does not allow conclusions
about implementing the results, such as statements on how realistic it is to induce
high status people to behave in a more environmentally friendly way. As discussed
above, the simulation method permits only comparative conclusions.
The commons dilemma

How must the social surround be perceived for people to become willing to use an
environmental resource sustainably? When people act upon the environment, they
usually utilize an environmental resource. Utilization of a resource available to all
(common property) involves an interpersonal conflict (Liebrand et al. 1992; Mosler
1995). It is in each individual’s interest to keep the personal benefit as large as
possible, while the harm incurred or the depletion of the resource must be borne
by all (Spada and Ernst 1990; Mosler and Gutscher 1996).
The sustainability motive expresses the degree to which a person is willing to
protect common property, the environment. This motive depends upon both the
absolute value a person places on environmental goods and the current discrepancy
between collective, sustainable use and actual patterns of use shown by others
(expectation of sustainable utilization). The more highly the individual values the
common property of the environment have, or the smaller the discrepancy between
actual and sustainable patterns of use is, the greater the effect a sustainability
motive will have. Under these conditions, the individual is motivated to make a
personal contribution to sustainable patterns of use.
The sub-model of the commons dilemma (compare Figure 4.6) yields a sustain-
ability motive as an expression of the desire to use the common property in a
resource-sustaining way through cautious use. The sustainability motive results
from a multiplicative connection between the expectation of sustainable utilization
and valuation of the resource ‘environment’.
Valuation of the resource ‘environment’ is determined by the perceived state
of the resource (state of the environment) and a person’s environmental values.
Environmentally-oriented people use exhausted resources less. Various studies have
shown the legitimacy of equating social orientation with environmentally-oriented
values (Liebrand 1986; Liebrand et al. 1986; Mosler 1990). The degree of exhaust-
ion of a resource is calculated on the basis of its state, as a percentage of deviation
from its maximal state, and weighted by values. This means that a person’s eval-
uation of the resource is higher the more depleted a resource is and the more the
person is environmentally oriented. Our model presently conceives of the percep-
tion of the state of the resource as direct perception, that is, without any distortion
INDIVIDUAL
PROCESSING OF INFORMATION ON COLLECTIVE USE
Estimated proportion
Using behaviour of resource- Expectation of
of contact sustainable Sustainability
sustaining users
persons utilization motive
Required proportion Intention

of resource-
Knowledge about sustaining users
the resource
Perception of the Volition

state of the Valuation of
resource the resource
Values Using
behaviour
Figure 4.6 Representation of ‘processing of information on the collective use of a resource’
stemming from inner psychological biases or from external influences in the form
of media reports, disputes among the experts, and so on.
The expectation that one’s own way of using the environmental resource will
make a contribution towards preservation of the resource is formed from the required
proportion of resource-sustaining users in the population and the estimated proportion
of resource-sustaining users. In other words, the greater the proportion of non-
sustaining users of a resource as estimated by the person in comparison with the
required proportion of resource-sustaining users in the population, the lower the
person’s expectation that they themselves can make an effective contribution to
sustainable use of the resource. This expectation rises in the degree to which the
proportion of estimated resource-sustaining contact persons exceeds the required
proportion of resource-sustaining users. Knowledge of the number of resource-
sustaining users required in order to preserve the resource comes from resource
knowledge and the perception of the state of the resource. The model presumes
people to have – at present not a very realistic assumption! – knowledge about the
regeneration rate, regeneration characteristics, and the maximal state of the resource.
From the discrepancy between the maximal state of the resource and the current
state of the resource, the person derives the required proportion of resource-sustaining
users. The greater the discrepancy between the current and the maximal state of the
resource, the greater must be the proportion of the population that uses the resource
in a sustainable way if the resource is to recover.
The estimated proportion of resource-sustaining users in the population is based
upon the behaviour of contact persons perceived by the person. According to the
ratio of contact persons using the resource sustainably and non-sustainably, the
person makes an estimate for the entire population. This factor is at present not
yet weighted by socially, or environmentally, oriented values. For example, people
who are not environmentally oriented overestimate the number of like-minded
people.
If no particular interventions take place, there is a danger that dynamics such
as those shown in Figure 4.7a will develop. With an optimistic starting value with
regard to average use within a population (over 50 = environmentally friendly
use), the state of the resource briefly improves. Due to its improved state, its value
declines (only goods in short supply are valuable), and the individuals in our
simulated population resume stronger exploitation of the resource. From the fifth
step onwards, this tendency results in clear over-utilization (average use under 50).
This increases the discrepancy between actual and sustainable use patterns, and
expectations of sustainable use correspondingly decrease. The individual is less
motivated to make a personal contribution to sustainable patterns of utilization
(‘… personal restraint on my own part would not make any difference; no one else
is showing restraint, so it is better for me to help myself to the resource so long as
it is still up for grabs …’). And so the state of the resource deteriorates. Its value
rises, which does not, however, lead to a marked reduction in utilization. There is
no stopping the course of these ‘downhill’ dynamics, and the resource is destroyed.
To counteract these negative dynamics, we ran a campaign in the fifth step that
aimed to (a) lead persons, in their own use behaviour, to orient themselves less to
other persons’ patterns of use and (b) lend heavier weighting to the importance of
the resource. Through this, the common property becomes more highly esteemed.
We found that the utilization behaviour of the population becomes more
environmentally sustaining (see Figure 4.7b). If individuals use a resource in an
environmentally sustainable fashion on the average, the state of the resource
improves. This line of development continues until the resource has regenerated.
At this time, the resource is available in over-abundance so to speak, whereby its
value again declines, and intensity of use increases. Downhill dynamics develop
until that crucial point where they are again brought under control. A dynamic-
stable balance has emerged, in which utilization continually adapts to the state of
the resource by means of inner personal factors. The resource will never be
completely destroyed.
Through the interaction of the social system and the resource, the resulting
system behaviour shows large fluctuations. Following Forrester (1972: 48ff.),
fluctuating system behaviour results when in a system of interlocking feedback
loops two or more temporal delays occur. In the present simulation model, there
are delays both within the social system and between the social system and the
resource system. The average expectation that one can make an effective
contribution to sustainable use of a resource reacts with a delay to the average
resource use behaviour of the social surround. The rise and fall of the average
value of the common good reacts with a delay to the development of the resource.
Our simulation approach is not directly comparable to Forrester’s approach.
Forrester observes only one macrosystem, while the simulation builds upon
numerous microsystems (individual persons) which join to form the macrosystem
(the social system).
Environmentally
`´ ýP
responsible/
resource maximal
100
80
60
40
20
0
Not 0 5 10 15 20 25 30
environmentally
responsible/ resource in % Runs
resource minimal average use
average value of common resource
average expectation of sustainable use
Environmentally
responsible/
resource maximal
100
80
60
40
20
0
Not 0 5 10 15 20 25 30
environmentally Runs
Resource in %
responsible/ Intervention
resource minimal Average use
Average value of common resource
Average expectation of sustainable use
Note: Also illustrates the course of significant inner variables. 4.7a (top): control situation; 4.7b (bottom):
with intervention.
Figure 4.7 Use of an environmental resource

Processing communicative influence
What is the effect of convincing attempts (persuasion) in populations, according to
individuals’ concern about the environment, knowledge of the environment, and
biases? Here we apply the Elaboration Likelihood Model (Petty and Cacioppo
1986a, b). People process information in varying degrees of thoroughness: the
‘intensity’ of processing is a function of processing motivation and processing ability;
areas of great personal relevance (concerned consternation about the environment)
and self-responsibility (here responsibility for the state of the environment) increase
motivation to process (Stahlberg and Frey 1993). Where deep, complete processing
of information occurs, the results of campaigns to convince are primarily dependent
upon the quality of the arguments presented; in the case of superficial processing,
superficial cues, such as status, credibility and attractiveness of the communicator,
gain more weight. Furthermore, individuals holding extreme attitudes or values
tend to process information in a biased way: when a person’s own position is too
removed from the attitude advocated by the influencing campaign, there can be a
‘boomerang’ effect, whereby the person changes his attitude in a direction opposing
the persuasive arguments. If the advocated position is close to one’s own, there
will be convergence between them. The principle mechanisms involved may be
illustrated by a simplified, information-campaign experiment.
The four external input variables in Figure 4.8 represent external influences on
the processing individual. The three variables in the upper section of the figure
represent influences stemming from the source of persuasion, a communicator.
The communicator communicates a particular attitude position with a certain
degree of persuasiveness. This persuasiveness variable corresponds to ‘argument
quality’ in Petty and Cacioppo (1986b). The fourth external input variable that is
processed is the (perceived) current state of the environment. From the left, two
free, internal input variables – self-responsibility and values – flow into the sub-
model. These are not affected by the processes within the model, just as the status
of the individual also remains constant.
Concern arises from the processing of information on the state of the
environment and from a person’s own values as related to the environment. Concern
corresponds to the discrepancy between these personal values (the desired value
for the state of the environment) and the actual value for the state of the
environment (the effective, perceived state of the environment). This means that
the greater the deviation between the individual’s values in relation to the environ-
ment and the perceived state of the environment, the greater the feeling of
concerned consternation. This is true only as long as those values are at a higher
level than the state of the environment, or in other words, only as long as the
environment is in worse shape than the individual would like to see.
The knowledge that the individual has, changes depending upon the degrees
of persuasiveness the individual has. Environmental knowledge, in its function as
an ability variable, determines the degree of processing depth, or intensity.
Processing motivation results from the sum (addition) of the feeling of self-
responsibility – how strongly the person experiences a sense of responsibility for
INDIVIDUAL
PROCESSING OF SOCIAL INFLUENCING UNDER CONDITIONS OF
KNOWLEDGE AND CONCERN
Communicator's persuasiveness
Communicator's attitude
Communicator's status
Self-responsibility
Processing
Status
motivation
Perception of state Concern
of environment
Processing intensity Attitude
Values
Persuasiveness
Processing ability
Knowledge
Figure 4.8 Representation of ‘processing of communicative influencing under conditions

of knowledge and environmental concern’
the state of the environment – and environmental concern. Processing intensity

(elaboration likelihood) arises in multiplicative fashion from processing motivation
and processing ability. In addition, it is influenced by the status of the communicator
(but only where neither processing motivation nor processing ability equals zero).
A communicator’s positive status raises processing intensity, while negative status
lowers processing intensity.
The difference between the attitude position of the individual and the position
advocated by the communicator determines attitude change – in dependency,
however, upon the influences from central and peripheral processing. If processing
intensity is low, peripheral processing will predominate, and the status of the
communicator is the main determinant of attitude change. High status of the
communicator leads to a change of attitude to the position advocated. Low status
of the communicator leads to change of attitude opposing the advocated position.
If processing intensity is high, however, central processing takes over, and the
persuasiveness of the communicator is the main determinant of attitude change.
High persuasiveness is expressed by high quality arguments, and the result is a
change of attitude in the desired direction. Low persuasiveness, accompanied by
weak arguments, leads to a change of attitude in the opposing direction.
Biased processing, moreover, modifies central processing. The individual’s
environmental values are compared with the attitude position advocated by the
communicator. If they agree (when both are environmentally friendly or
environmentally unfriendly), the communicator will successfully influence the
person in the desired direction. If there is a discrepancy, however, the person’s
attitude will change in the opposing direction.
To the individual’s initial persuasiveness, a central and peripheral proportion is
added in proportion to processing intensity. Where processing intensity is high,
the individual’s persuasiveness increases if the persuasiveness of the communicator
is higher than the individual’s persuasiveness. Here the individual integrates the
communicator’s good arguments into his own argumentation scheme. Where
processing intensity is low, the individual’s persuasiveness increases if the persuasive-
ness of the communicator is higher than the individual’s persuasiveness. In contrast,
if the persuasiveness of the communicator is low, the individual’s persuasiveness
decreases. This assumption is our own and is not based upon the Elaboration
Likelihood Model. The idea is that unreflected acceptance of weak arguments
tends to weaken a person’s persuasiveness.
We assume to this purpose (as an exception) that the individuals do not influence
each other mutually, but that in five steps they are presented with information
campaigns (the individuals do not discuss the subject of the environment among
themselves, but they all stand under the influence of the campaigns). The campaigns
promote strong environmentally friendly attitudes by means of good arguments
(leading to high persuasiveness) and/or by means of strong peripheral cues
(communicators having high status, as for example esteemed public figures). In
addition, we assume that in the campaigns, an appeal can cause persons’ self-
responsibility to increase. Average attitudes and values are low on environmental
friendliness, the extent of the sense of self-responsibility is low, and the state of the
environment is poor. Persons having environmentally friendly values experience
high concern under the condition of poor state of the environment. The following
interventions were ‘tested’ (see Mosler et al. 1998):
• an information campaign presenting good arguments (Figure 4.9, persuasive-

ness+);
• an information campaign which works with well-known and esteemed
personalities (Figure 4.9, status+);
• an information campaign based on good arguments and well-known, esteemed
personalities (Figure 4.9, persuasiveness+/status+);
• an information campaign with well-known, esteemed personalities that
moreover appeals to people’s sense of environmentally-oriented self-
responsibility (Figure 4.9, self-responsibility+/status+).
Let us first examine results with individual persons selected from the population.
Person A – who is biased due to an environmentally non-friendly attitude – places
a negative value even on good arguments (curve Persuasiveness+). In contrast,
good arguments effect even more environmental consciousness in Person B, who
has environmentally responsible attitudes from the start, as information processing
is hardly biased here.
A campaign using esteemed public figures is effective with both Persons A and
B, if they tend to process information relatively superficially (curve Status+). But
if Person A’s depth of processing is increased by means of self-responsibility (curve
Self-responsibility+/Status+), they then show biased processing, and their attitude
Environmentally
responsible
70 70
Person A Person B C C
60 60 B B
D D
A A
50 50
40 BC
40
DA
30 B B
C 30
C
20 20
BC
10 A A 10
DA
0 D D 0
Not 0 5 Intervention 10 15 0 5 Intervention 10 15
environmentally Runs
responsible Runs
Environmentally
responsible Campaigns with:
90
Population A Campaign: persuasiveness+
80 B Campaign status+
70 C Campaign: persuasiveness+/status+
C C
B B D
60 Campaign: self-responsibility+/status+
A A
50 D D Control situation
40
30
Not
environmentally 5 Intervention 10 15
responsible Runs
Note: Above, individuals in the population. Below, population values. The three graphs present
courses of attitude change in four different experiments and the control situation.
Figure 4.9 Information campaign experiments without environmentally relevant

communication
changes in the direction away from that advocated by the campaign. In the total
population, the average attitude becomes most changed by means of good
arguments and presentation by esteemed personalities. But as shown in Figure
4.9, we must always take into consideration that these average values within the
population hide very contrasting effects upon differing individuals.
For various experiments with multiplicators (persons who advocate their environ-
mental attitudes to others), a communication net was constructed. Each simulated
person ‘talks’ about environmental issues with five friends and one stranger during
each step, whereby the friendships are assumed to be lasting and reciprocal. The
relationships to strangers are different in each step. Initial values in the population
are the same as those described above for the information campaign experiment.
Within the framework of an environmental action campaign, 500 persons
(multiplicators) having very environmentally friendly attitudes receive training in
argumentation (increasing persuasiveness, see Gonzales et al. 1988). In a variant
of this experiment, the status of these persons is raised (as for example through
public commendation awards). The following interventions were carried out:
• Multiplicators with good arguments (Figure 4.10, Persuasiveness+).
• Multiplicators with raised status (Figure 4.10, Status+).
• Multiplicators with good arguments and raised status (Figure 4.10,
Persuasiveness+/Status+).
• Control situation.
In Figure 4.10 we again see two selected individuals A and B, who both stand
under the influence of the action campaign. A glance shows that these graphs are
not as smooth as the information campaign graphs. The curves would be smoother
if, after a short levelling-off phase, there were only the influence of a constant
group of friends. But the ever-changing influence of strangers, upon Persons A
and B and also upon their friends, causes fluctuations in attitudes. We can still
recognize certain tendencies in the development of attitudes, however. Person A’s
attitude becomes less environmentally friendly relatively quickly. Later changes to
a more responsible attitude are only brief and are repeatedly destroyed by the
environmentally unfriendly social surround. Person B shows initial swings in
attitude, but then develops a tendency towards a less environmentally friendly
attitude. Both Persons A and B react to the introduction of multiplicators having
high status (curve Status+) with an increasingly environmentally friendly attitude.
This effect is even stronger if multiplicators in addition show high persuasiveness
(curve Persuasiveness+/Status+). But if multiplicators demonstrate only
persuasiveness (curve Persuasiveness+), they have no effect upon Person A. There
is an initial effect in this case upon Person B, but the influence of the rather
environmentally unfriendly social surround cannot be cancelled out. The same
effects find expression at the level of the population: multiplicators with raised
status are just as convincing as those who also show increased persuasiveness.
Multiplicators who have only high persuasiveness at their disposal have a counter-
productive effect.
Summary and conclusions

The concept of the model, which is based on an elaborated rational choice theory,
has proved to be – in connection with the simulation method – a valuable tool for
the drafting, designing, and testing of environmental psychological interventions.
The most important results of the experiments with the various sub-models can
be summarized as follows:
People who as a minority wish to effect more environmentally responsible
attitudes in their group absolutely must have a low susceptibility to influence. If
this condition is fulfilled, no measures for networking the minority persons need
be undertaken.
Role models can be implemented more successfully if these models have high
status and their behaviour is highly visible.
To achieve appropriate use of a resource, it is important to lead persons to
orient themselves less to other people’s patterns of use and to pay more attention
to the state of the resource.
Environmentally
responsible
Person A Person B
100 100 C C C
C C
B B B
80 B
80
C CB B C
B A B
60 C B C 60 D
C B DA
B D
40 40 D D A
A A A D
20 20
C
D
A A A D
B D A
0 D D A A D 0
Not
environmentally 10 20 30 10 20 30
responsible Runs Runs
Intervention Intervention
Environmentally
responsible
80
Population
70
A Multiplicators: persuasiveness+
60
B Multiplicators: status+
50 C C C C C
B B C Multiplicators: persuasiveness+/status+
B B B
A D D D
40 D D D
A Control situation
A A A
30
20
Not
environmentally 10 20 30
responsible Runs
Intervention
Note: Individual graphs show the course of attitude change in three different experiments and the
control situation.
Figure 4.10 Experiments with multiplicators
Both in general campaigns and in specific action campaigns using multiplicators,

it is advantageous to work with people who have high status. If attitudes within a
population tend to be environmentally irresponsible, there is little to be gained
through the use of persuasiveness or appealing to people’s sense of environmentally-
oriented self-responsibility.
A common denominator of the results can be formulated as follows, in analogy
to Latané’s (1981) theory of social impact, according to which social influence is a
function of the strength, immediacy, and number of sources.
• Strength: the concepts of susceptibility to influence, suggestibility, status, and

persuasiveness belong here.
• Immediacy: this involves the way that persons are networked, the visibility of
behaviour, and the degree to which a person may be influenced by others.
• Number: while it is always advantageous to work with a large number of
pioneers, role models, and multiplicators, the financial means to do so may be
limited. In this case, the same effect can be achieved with a smaller number of
persons if measures are implemented to improve strength or immediacy.
In summary, we find a certain number of people equipped with a certain degree
of strength and immediacy who will stimulate others to behave in more
environmentally responsible ways. Success in getting 3, 5, or even 10 pe rcent of a
community to become active does, however, seem an impossible task. According
to the Swiss Environmental Survey (Diekmann and Franzen 1995), 16 per cent of
people in Switzerland are members of environmental groups. This existing
potential, dispersely distributed, needs to be seen as a resource that can be activated
and brought together within the framework of concerted environmental efforts. If
efforts are insufficiently co-ordinated, they fall flat.
The conception of our experiments, with various forms of intervention as well
as the application of terms suited to translation into action, might create the impres-
sion that we underestimate the problems of implementing in reality the forms of
intervention studied. We are conscious of the fact that there are great gaps
(theoretical as well as empirical) and incalculabilities between ‘population
simulations’ based upon simulated individual behaviour and possible processes in
real populations. Many sub-components are missing in our model or are as yet not
adequately designed or validated. In spite of this, from our perspective there is no
other alternative: we should indeed take on the challenge of examining the
complexity of such dynamic processes by means of appropriate instruments.
Simulation is an instrument well suited to enlarging our understanding of basic,
underlying social processes and developing further those explanatory approaches
that are almost exclusively static and based on one-person models. Against the
possible view that our approach is far from reality, it must be objected that the
forms of intervention we propose are founded upon empirical knowledge gained
in small-scale field experiments (see Dwyer et al. 1993; Mosler and Gutscher 1998).
In the next phase of research, we plan stage-by-stage testing of the simulation
model as a logical step in its proof in practice. Here, in the sense of a practice-
oriented ‘acid test’, we will strive towards process-oriented validation. To gain
well-founded understanding of the dynamics of existing social systems, data must
be collected in real social systems that are relevant to indications resulting from
the simulation and that can then be entered into the simulation itself. The focus of
the investigation would be the topic of dissemination of new behaviour patterns
throughout existing social networks. The optimal method would consist in large-
scale field experiments.
It is not our intention to propagate as ‘tested’ or ‘problem-free’ the application
of those forms of intervention judged effective on the basis of our simulations.
Rather, our purpose is to show environmentally conscious people, responsible
persons, and politicians how worthwhile it can be to expand the old triad of
‘traditional information campaigns’, ‘legal measures’, and ‘economic measures’
by means of additional, novel strategies and to test these potentially successful
forms of intervention in practice.
References
Ajzen, I. (1991) ‘The theory of planned behaviour. Some unresolved issues’, Organizational
Behaviour and Human Decision Processes, 50: 179–211.
Ajzen, I. and Fishbein, M. (1977) ‘Attitude–behaviour relations: a theoretical analysis and
review of empirical research’, Psychological Bulletin, 84: 888–918.
Ajzen, I. and Madden, T.J. (1986) ‘Prediction of goal directed behaviour: attitudes,
intentions, and perceived behavioural control’, Journal of Experimental Social Psychology,
22: 453–74.
Bandura, A. (1979) Sozial-kognitive Lerntheorie, Stuttgart: Klett-Cotta.
Bandura, A. (1986) Social Foundations of Thought and Action, New York: Prentice-Hall.
Becker, G.S. (1976) The Economic Approach to Human Behavior, Chicago: University of Chicago
Press.
Diekmann, A. (1996) ‘Homo ökonomicus. Anwendungen und Probleme der Theorie
rationalen Handelns im Umweltbereich’, in A. Diekmann and C. Jaeger (eds)
Umweltsoziologie, Sonderband 36 der Kölner Zeitschrift für Soziologie und Sozialpsychologie: 89–
118.
Diekmann, A. and Franzen, A. (1995) Der Schweizer Umweltsurvey 1994, Codebuch, Bern:
Institut für Soziologie.
Dwyer, W.O., Leeming, F.C., Cobern, M.K., Porter, B.E. and Jackson, J.M. (1993) ‘Critical
review of behavioural interventions to preserve the environment. Research since 1980’,
Environment and Behaviour, 25(3): 275–321.
Festinger, L. (1954) ‘A theory of social comparison processes’, Human Relations, 7: 117–40.
Forrester, J.W. (1972) Grundzüge einer Systemtheorie, Wiesbaden: Gabler.
Frey, D., Dauenheimer, D., Parge, O. and Haisch, J. (1993a) ‘Die Theorie sozialer Ver-
gleichsprozesse’, in D. Frey and M. Irle (eds) Kognitive Theorien 2, Bern: Hans Huber:
81–122.
Frey, D., Stahlberg, D. and Gollwitzer, P.M. (1993b) ‘Einstellung und Verhalten: Die Theorie
des überlegten Handelns und die Theorie des geplanten Handelns’, in D. Frey and M.
Irle (eds) Kognitive Theorien 2, Bern: Hans Huber: 361–98.
Gonzales, M.H., Aronson, E. and Costanzo, M.A. (1988) ‘Using social cognition and
persuasion to promote energy conservation: a quasi-experiment’, Journal of Applied Social
Psychology, 18: 1049–66.
Harsanyi, J.C. (1977) Rational Behaviour and Bargaining Equilibrium in Games and Social Situations,
Cambridge: Cambridge University Press.
Jorgenson, D.O. and Papciak, A.S. (1981) ‘The effects of communication, resource feedback,
and identifiability on behaviour in a simulated commons’, Journal of Experimental Social
Psychology, 17: 373–85.
Klandermans, B. (1984) ‘Mobilization and participation: social-psychological expansions
of resource mobilization theory’, American Sociological Review, 49: 583–600.
Latané, B. (1981) ‘The psychology of social impact’, American Psychologist, 86: 343–56.
Latané, B. and Wolf, S. (1981) ‘The social impact of majorities and minorities’, Psychological
Review, 88: 438–53.
Liebrand, W.B.G. (1986) ‘The ubiquity of social values in social dilemmas’, in H.A.M.
Wilke, D.M. Messick and C.G. Rutte (eds) Experimental Social Dilemmas, Frankfurt am
Main: Verlag Peter Lang: 113–33.
Liebrand, W.B.G., Wilke, H.A.M. and Wolters, F.J.M. (1986) ‘Value orientation and
conformity. A study using three types of social dilemma games’, Journal of Conflict
Resolution, 30: 77–97.
Mosler, H.-J. (1990) Selbstorganization von umweltgerechtem Handeln: der Einfluss von Vertrauensbildung
auf die Ressourcennutzung in einem Umweltspiel, Zürich: Zentralstelle der Studentenschaft.
Mosler, H.-J. (1993) ‘Self-dissemination of environmentally-responsible behaviour: the
influence of trust in a commons dilemma game’, Journal of Environmental Psychology, 13:
111–23.
Mosler, H.-J. (1995) ‘Umweltprobleme: eine sozialwissenschaftliche Perspektive mit
naturwissenschaftlichem Bezug’, in U. Fuhrer (ed.) Oekologisches Handeln als sozialer Prozess,
Basel: Birkhauser: 77–86.
Mosler, H.-J. (2000) Computersimulation sozialpsychologischer Theorien. Studien zur Veränderung von
Umwelteinstellung und Umweltverhalten, Weinheim: Psychologie Verlags Union.
Mosler, H.-J. and Gutscher, H. (1996) ‘Kooperation durch Selbstverpflichtung im Allmende-
Dilemma’, Kölner Zeitschrift für Soziologie und Sozialpsychologie, Sonderheft Umweltsoziologie,
36: 308–23.
Mosler, H.-J. and Gutscher, H. (1998) ‘Umweltpsychologische Interventionsformen für
die Praxis’, Umweltpsychologie, 2: 64–79.
Mosler, H.-J. and Tobias, R. (2000) ‘Die Organisation kollektiver Aktionen durch
Beeinflussung der individuellen Teilnahmeentscheidung. Eine Simulationsstudie’, Kölner
Zeitschrift für Soziologie und Sozialpsychologie, 52(2): 264–90.
Mosler, H.-J., Ammann, F. and Gutscher, H. (1998) ‘Simulation des Elaboration Likelihood
Model (ELM) als Mittel zur Entwicklung und Analyse von Umweltinterventionen’,
Zeitschrift für Sozialpsychologie, 29: 20–37.
Nowak, A. and Latané, B. (1994) ‘Simulating the emergence of social order from individual
behaviour’, in N. Gilbert and J. Doran (eds) Simulating Societies, London: UCL Press:
63–84.
Nowak, A., Szamrej, J. and Latané, B. (1990) ‘From private attitude to public opinion: a
dynamic theory of social impact’, Psychological Review, 97(3): 362–76.
Petty, R.E. and Cacioppo, J.T. (1986a) ‘The elaboration likelihood model of persuasion’,
in L. Berkowitz (ed.) Advances in Experimental Social Psychology, New York: Academic Press:
123–205.
Petty, R.E. and Cacioppo, J.T. (1986b) Communication and Persuasion: Central and Peripheral
Routes to Attitude Change, New York: Academic Press.
Spada, H. and Ernst, A.M. (1990) ‘Wissen, Ziele und Verhalten in einem ökologisch-
sozialen Dilemma’, Forschungsbericht, Freiburg: Psychologisches Institut der Universität
Freiburg.
Stahlberg, D. and Frey, D. (1993) ‘Das Elaboration–Likelihood–Modell von Petty und
Cacioppo’, in D. Frey and M. Irle (eds) Kognitive Theorien 2, Bern: Verlag Hans Huber:
327–60.
Stalder, J. (1985) ‘Die soziale Lerntheorie von Bandura’, in D. Frey and M. Irle (eds) Gruppen-
und Lerntheorien, Bern: Verlag Hans Huber: 241–72.
Tajfel, H. (1978) Differentiation Between Social Groups, London: Academic Press.
Tajfel, H. (1979) ‘Individual and groups in social psychology’, British Journal of Social and
Clinical Psychology, 18: 183–90.
Tanford, S. and Penrod, S. (1984) ‘Social influence model: a formal integration of research
on majority and minority influence process’, Psychological Bulletin, 95(2): 189–225.
Thibaut, J.W. and Kelley, H.H. (1959) The Social Psychology of Groups, New York: John Wiley.
Assessing the quality of different MCDA methods 97
Part II
Taking multiple criteria

into account
98 Andrea De Montis et al.
5 Assessing the quality of

different MCDA methods
Andrea de Montis, Pasquale De Toro,
Bert Droste-Franke, Ines Omann and Sigrid Stagl
Introduction
The awareness that environmental problems have a large impact on the economy
and society has risen continuously during the last three decades. Lifting sustainable
development to the political agenda nurtured the understanding that policy analysis
is characterized by conflicting economic, environmental, societal, technical, and
aesthetic objectives. Consequently numerous research projects on the inter-
dependence of environment, society and economy were conducted in this period.
With increasing knowledge about the economic and environmental processes and
their interlinkages, it became clear that decisions about environmental problems
are especially characterized by uncertainties, high stakes, urgency, and dispute
(Funtowicz and Ravetz 1990). These characteristics influence the choice of the
method.
In the field of environmental decision-making, multiple criteria decision aid
(MCDA) has become a widespread tool, able to tackle problems involving more
than one objective. A number of theoretical (e.g. Munda 1995; Castells and Munda
1999) and empirical studies (e.g. Janssen 1992; Andreoli and Tellarini 2000;
Mendoza and Prabhu 2000) showed that MCDA provides a useful tool for decision
aid in this field, as it allows for multiple objectives, for the use of different types of
data and the involvement of different stakeholders. The choice of a particular
method depends on the specific problem as well as on users’ demands. As the
number of existing methods is already quite large and is still increasing, the choice
of the ‘right’ method is not an easy task.
The aim of this chapter is the comparison of frequently used multiple criteria
methods. A number of authors have compared multiple criteria algorithms with
regard to their consistency, relative accuracy, speed, etc. (e.g. Olson 2001; Salminen
et al. 1997). Since the underlying philosophies and assumptions, and hence the
applicability, vary significantly, we chose, in contrast, a less mechanistic approach
and describe the methods with reference to a comprehensive list of quality criteria.
The criteria were chosen with particular emphasis on their relevance for questions
of sustainable development. For the evaluation we apply the list of quality criteria
to seven different MCDA methods (MAUT, AHP, Evamix, ELECTRE III, Regime,
NAIADE, and programming methods) and compare them. This list of criteria
provides a basis to assess the usefulness of each method for a specific application.
The chapter is structured as follows. First, the list of quality criteria are presented,
then the seven methods are characterized and evaluated by use of the list of criteria.
This evaluation leads to a comparison of the methods, which is followed by
conclusions.
Criteria for the quality assessment of methods

To ease the selection of the appropriate method for a specific decision-making
situation, a list of quality criteria was developed in De Montis et al. (2000) which
can be used to reveal strengths and weaknesses of MCDA methods with respect to
their application for environmental problems. These MCDA quality criteria can
be grouped according to three main aspects:
• operational components of MCDA methods,

• applicability of MCDA methods in the user context, and
• applicability of MCDA methods considering the problem structure.
Table 5.1 lists the criteria together with a summarizing description.

The first group of criteria includes operational components and characteristics
of the method resulting from its theoretical underpinnings. The first important
point for the method characterization is whether interdependence of criteria can
be addressed in its application (e.g. equity or efficiency). Different technical
procedures used to derive the decision may lead to different results in the same
decision-making situation, because they use different mechanisms to compare
criteria and alternatives. Some methods, for instance, do allow for incomparability,
while others do not. Additionally, with regard to the characteristics of the criteria
and the weighing process, the methods are distinct from each other.
Beside the theoretical foundations and the technical components of the method,
specific properties concerning the applicability of the method with respect to the
user’s/users’ situation are important factors for the method characterization.
Quality criteria concerning the project constraints (costs and time), and on the
other hand criteria addressing the envisaged type of the structure of the problem-
solving process (e.g. the possibility of stakeholder participation) are relevant factors.
Another type of criteria which are relevant in the context of the methods’
applicability, are criteria which give information about the possibility for
consideration of important issues within the specific problem-solving situation.
One group of these criteria aims to examine the possibilities for the inclusion of
indicators with specific characteristics necessary for the problem solution, e.g. the
evaluation of indicators over different geographical scales. Another group of criteria
characterizes the data situation, i.e. the type of data which can be used and the
situation of data availability for the indicators’ evaluation in the context of the
specific problem tackled.
According to the milestone textbooks by Roy (1985; 1996), multiple criteria
methods can be divided into three operative approaches (Roy 1996: 241):
Table 5.1 List of quality criteria for MCDA methods (De Montis et al. 2000)
Quality criterion Description
Operational components
Criteria
Interdependence Allowance for the interdependence of different criteria
Completeness Need for the completeness of the criteria
Non-linear preferences Possibility to express non-linear valuation patterns
Weights
Transparency of process, Type of the procedure of deriving values for the weights
type of weights
Meaning Interpretation and role of weights in the evaluation process
Solution finding Type of procedure used for the comparison of alternatives
procedure
Issues addressed by Interpretation of the results generated by the use of method
results
Applicability of MCDA methods – user context
Project constraints
Costs Implementation costs in the specific user situation
Time Implementation time in the specific user situation
Structure of problem
solving process
Stakeholder participation Possibility to include more than one person as decision
maker
Problems structuring Existence of mechanisms supporting the structuring of the
problem
Tool for learning Support of learning processes
Transparency Promotion of transparency in the decision making process
Actor communication Support of the communication between opposing parties
Applicability of MCDA methods – problem structure
Indicator characteristics
Geographical scale Applicability of different geographical scales for one case
Micro–macro link Applicability of different institutional scales for one case
Societal/technical issues Possibility for the consideration of both societal and
technical issues
Methods combination Possibility of methods combination
Data situation
Type of data Type of data supported as values for the indicators
Risk/uncertainties Possibilities for the consideration of evaluation risk and/or
uncertainties
Data processing amount Processing amount needed to compile the data required for
the method
Non-substitutability Possibility to consider sustainability standards and non-
substitutability
(a) methods based on the use of a single synthesizing criterion without
incomparability, (b) methods based on the synthesis by outranking with
incomparability, and (c) methods based on interactive local judgements with trial-
and-error iteration. While the first two groups embody a clear mathematical
structure, the third does not use any formalized or automatic procedure. Another
possible division refers to two groups of multiple criteria methods, namely discrete
and continuous, depending on whether the set of alternatives is finite or infinite
(Voogd 1983). In the following we will first present two types of discrete methods
(single synthesizing criterion and outranking methods) and then the continuous
methods (programming methods).
Single synthesizing criteria

MCDA methods using a single-criterion approach are trying to convert the impacts
concerning the different criteria into one criterion or attribute, which build the
base for the comparison of the alternatives. A well-known example for a single-
criterion approach is cost–benefit analysis (CBA), which is using the Net Present
Value as the criterion. This requires the conversion of impacts into monetary
values. Hence, CBA cannot be seen as an MCDA method, because by use of
MCDA methods each impact is evaluated on its most suitable scale; this scale can
be both cardinal (monetary and non-monetary) and ordinal (qualitative). In the
case of MCDA methods using a single synthesizing criterion, the aggregation of
all the evaluations concerning each criterion results in a single index associated
with each alternative and it expresses the performance of the alternative with
respect to all the evaluation criteria simultaneously. Usually the higher this index
the better ranks the alternative. Three such methods are described in the following:
multiple attribute value theory (MAUT), analytic hierarchy process (AHP) and
evaluation matrix (Evamix).
Multiple attribute value theory (MAUT)

The foundations of MAUT were laid by Churchman, Ackoff and Arnoff (1957)
who first treated a multiple criteria decision problem using a simple additive
weighting method. It was developed further using a quasi-additive and multi-linear
utility function. Its early development is attributed to Fishburn (1968, 1970, 1978).
Keeney and Raiffa devoted much of their work (e.g. 1976) to MAUT. Since then
the method has been developed further in the way that methodologies to elicit the
utility functions are elaborated, and the single-criterion utility functions are aggre-
gated to a multi-attribute utility function, and it now provides a formal mechanism
for handling multiple objectives, intangible factors, risk, qualitative data and time-
sequence effects in ex-ante appraisals based on the decision-maker’s preferences
(Dillon and Perry 1977). A frequently used software package, which is based upon
the estimation of an additive utility function and on interaction with the decision-
maker, is the so-called PREFCALC software. Other software packages are ‘Decide
Right’, which is based on SMART, the Simple Multi-Attribute Rating Technique
(for explanation, see later in this chapter) and ‘HIPRE 3+’, which integrates two
methods of decision analysis and problem solving, namely AHP – Analytic
Hierarchy Process (see also p. 106 and Chapter 6 in this volume) and SMART.
As the name MAUT indicates, it is based on utility theory (von Neumann and
Morgenstern 1947). It relies upon the basic von Neumann–Morgenstern axioms of
preference and thus upon a utility function, which allows the comparison of risky
outcomes through the computation of expected utility. Risk can arise in two ways, as
risk about the outcomes or about the attribute values. MAUT can be used for the
first form of risk1 (Hwang and Yoon 1981). MAUT uses directly assessed preferences
with general aggregation; this involves direct questioning of the decision-makers
and the choice on the basis of an aggregate measure for each alternative (Dillon and
Perry 1977). It can be seen as an extension of ordinary subjective expected utility
procedures to the case of choice between multi-attributed alternatives.
To prepare a multiple attribute decision by use of MAUT requires the following
steps (Dillon and Perry 1977):
1 specify the project alternatives (including combinations) as discrete entities,

2 elicit the decision-maker’s set of probability distributions for outcomes
associated with each project alternative in each attribute if there is risk,
3 elicit the decision-maker’s utility function ui(xi ) for the range of outcomes on
each attribute,
4 use the appropriate global multi-attribute utility function U(x) to find the
expected utility of each project alternative, and
5 choose the project or project combination with the highest expected utility;
thus the function U should be maximized.
Step 1 comprises the definition of the alternatives, the objectives, the attributes
and the outcomes of each alternative in each attribute. This step is not specific to
MAUT; this is how all discrete MCDA methods start.
Step 2 concerns risk viewed as subjective probability. If there is risk concerning
the impact of the actions on the attributes, probabilities pij can be assigned to the
outcomes x1, y1, … of the alternatives on the attributes. It is assumed that the value
of the action a for attribute i is given by the expected value of ui (Vincke 1985;
Dillon and Perry 1977): ui(xi ) = Sj pijui(xij ), where pij is the probability of the j-th
outcome in the i-th attribute or in continuous terms, and ui(xi ) = Úui(xi ) ¶i (xi )dxi,
where ¶i (xi ) is the probability distribution of outcomes in the i-th attribute.
There are different methods to elicit subjective probability distributions, which
are not without difficulties. Which method to choose depends on the problem and
often on the experience of the researchers. Direct involvement methods are simple
and allow the decision-makers to see and understand the built distribution. For
these reasons, such direct involvement methods seem preferable to others, such as
lottery-type questions (Dillon and Perry 1977). The simplest method of all is asking
for a few point estimates of probability; this is crude but practicable.
Step 3 requires the elicitation of the decision-maker’s utility function ui(xi ) for
the range of outcomes on each attribute. A utility function ui(xi) is developed for
each attribute. This is done by asking the decision-maker a series of questions
based on the axiom of continuity (e.g. Dillon and Perry 1977: 10). Additionally
weights wi are assigned to the attributes, presenting trade-offs, by asking for the
decision-maker’s order of importance of the attributes. The assessment of the
appropriate utility function is complex and intricate. Thus it is a crucial step in
MAUT. This procedure leads us to three difficult questions (Vincke 1985): (a) what
must the properties of the decision-maker’s preferences be so that U is a certain
function of ui ?; (b) how to test the properties?; and (c) how to construct the function
U? The method can only be applied if these questions can be answered in a
meaningful way.
After the establishment of the utility functions ui (xi ) and the elicitation of the
attribute probability distributions, the expected utility is aggregated across the
attribute distributions for each alternative in Step 4. For each attribute a function
ui is built and the functions ui are aggregated in a global criterion U, such that
U(u1(x1),…, un(xn )) > U(u1(y1),…,un (yn )), if the action represented by (x1,x2,…,xn ) is
better than the action represented by (y1,y2,…,yn ), when considering all the attributes
simultaneously (Vincke 1985; Roy and Bouyssou 1985). In that way the multiple
criteria problem is reduced to a single-criterion decision problem. In order to build
the aggregate function, different aggregation procedures are possible; the additive
or the multiplicative aggregations are most widely applied. The additive form is
the simplest form of a utility function, requiring two strong assumptions, namely
utility independence and preferential independence of any subset of criteria (Fandel
and Spronk 1985).2 SMART (simple multiple attribute rating technique) presents
one method to build the additive form (Martel, in Climaco 1997). If the additive
model holds, the expected utility of each attribute for each alternative is added.
Each attribute must get a weighting factor wi to get a common scale for the utility
functions ui(xi ). The utility for a certain alternative with uncertain consequences is
then given by: U(x) = Swi [ui (xi )], where in discrete terms, ui (xi ) = Sj pij ui (xij ), pij being
the probability of the j-th outcome in the i-th attribute; or in continuous terms,
ui(xi ) = Úui (xi ) ¶i (xi )dxi , where ¶i (xi ) is the probability distribution of outcomes in the
i-th attribute. If independence does not hold, the multiplicative form is usually
applied. Another form is the quasi-additive model, which can be used if neither
marginality nor preferential independence prevails; but this procedure becomes
very complicated for cases with more than three attributes (Martel, in Climaco
1997). Step 5 consists of summarizing the results and interpreting them.
A sub-methodology of MAUT is multiple attribute value theory (MAVT). The
two approaches deal with risk in different ways. While MAUT relies upon a utility
function, which allows the comparison of risky outcomes through the computation
of an expected utility, MAVT is unable to take risk into account. It uses a value
function to represent the outcome of the alternatives, not allowing for risky
outcomes. Value functions preserve deterministic ordering, whereas utility functions
preserve the expected utility hypothesis. MAUT is suited for ex-ante evaluations
for multiple objective problems with risky outcomes as long as one accepts the
assumptions on which the method is based.
MAVT is based on four key assumptions. First, a complete system of preferences
pre-exists, which is an adequate representation of the axioms in the decision-maker’s
mind. Second, the preference relation of the decision-maker is of a weak order,
which means that all states are comparable, transitivity of the preferences and
transitivity of the indifferences is given. The last one is particularly difficult as it
can lead to a lack of realism; see for example the famous cup-of-coffee example by
Luce (1956).3 Third, the main hypothesis underlying MAUT is that in any decision
problem there exists a real-valued function U defined on A (the set of actions),
which will be examined. This function aggregates the utility functions ui (Vincke
1985). Fourth, in the probabilistic case, when there is risk concerning the outcomes,
the axioms of von Neumann and Morgenstern (1947) must hold. In this way the
concept of rationality comes into MAUT, which can be seen as an extension of
ordinary subjective expected utility procedures to the case of choice between multi-
attribute alternatives. They imply an existing unique set of subjective probability
distributions for outcomes in each of the attribute dimensions associated with any
risk multi-objective alternative and a utility function whose expected value gives a
measure of the decision-maker’s preference for each of the risky multi-attribute
alternatives. If the axioms are accepted, there exists both ‘a unique set of subjective
probability distributions for outcomes in each of the attribute dimensions associated
with any risky alternative and a utility function U(xi) whose expected value (in
terms of the decision-maker’s subjective probabilities) gives a measure of the
decision-maker’s preference for each of the risky multi-attribute alternatives’ (Dillon
and Perry 1977: 9).
The development of MAUT was a milestone in the history of MCDA, as it
presented a formal and transparent method able to deal with multiple objectives,
risk concerning the consequences, the pattern of consequences over time, and a
start for using qualitative data. But MAUT’s main problem is the strong assumptions
that need to be fulfilled to derive the global utility function. Basically, MAUT
relies upon the same axioms as social welfare theory like CBA does and therefore
a number of the points of critique towards CBA also apply to MAUT. A review of
case studies shows that MAUT is used frequently for decision-making for economic,
financial, and actuarial problems as well as problems concerning environmental
issues, like water management, energy management, agriculture, but not really in
the field of sustainable development, which is much broader as it covers an
economic, a social, and an environmental dimension.
Concerning the first group of quality criteria (see Tables 5.1 and 5.2) for MCDA
methods, interdependence of attributes is not allowed; on the contrary the
independence assumption is crucial. This assumption makes it hard for use in
complex decision problems, as is the case for problems in the field of sustainable
development. The meaning and assignment of weights is clear (weights stand for
the importance of criteria) and transparent. The solution is a complete ranking of
actions, according to their expected global utility. This again is clear and transparent,
but the derivation of the utility function is the most delicate part of the whole
procedure. Different issues concerning the type of decision problem are allowed,
as long as they can be addressed by attributes; a utility function is designed for
each attribute. Thus the more issues are addressed the more time it takes and the
higher are the costs.
For most decisions relating to sustainable development, stakeholder participation
is necessary and desirable. Stakeholders are normally involved twice during the
decision-making process. First, when developing the utility function for each attribute,
they have to elicit their preferences. Second, when calculating the expected utility,
they have to assign probabilities to certain outcomes. Preferences are directly assessed
and generally aggregated and used to develop the utility functions. Under MAUT,
transparency of the process is given as well as a clear structuring. Concerning the
indicators and data, MAUT can handle different scales as well as different issues.
The required amount of data is usually quite high. The data itself can be of qualitative
or quantitative character. Risks of the outcome are accounted for by the expected
utility theory, while uncertainties are checked with a sensitivity analysis.
In sum, MAUT ensures in risky project choices the correspondence of the choice
with the decision-maker’s preferences and it is a mechanism for doing this in cases
too complex to be handled satisfactorily by intuition, as long as the underlying
assumptions are known and always in the minds of the researchers. Information
about the preferences of the decision-maker are necessary to build the singular
utility functions for each attribute and to build the aggregate utility function based
on all attributes. These individual utility functions build the base for the decision.
No social function is constructed, thus no interlinkages or interdependencies
between the individuals are taken into account. The building of a group or social
function is a problem in every MCDA method, as we learned from Arrow’s
impossibility theorem that it is not possible to aggregate individual preference
functions to group opinions. The individual preferences and estimations can be
seen as an acceptable proxy as long as the decision projects stay on a rather micro-
related level, like the decision about an irrigation system. But if we have to deal
with macro-related problems, additional influences gain importance, such as the
impact on the different dimensions of sustainable development or on different
groups in society. This decreases the appropriateness of individual utility functions
as an acceptable proxy for social preferences. Thus MAUT, like CBA, is hardly
suited for macro-related problems.
Analytic hierarchy process (AHP)

AHP was developed by Thomas Lorie Saaty. The method is implemented by the
software package ‘Expert Choice’. From a procedural point of view this approach
consists of three steps: (1) construct suitable hierarchies; (2) establish priorities
between elements of the hierarchies by means of pairwise comparisons; (3) check
logical consistency of pairwise comparisons (Saaty 1980, 1988; Saaty and Alexander
1989; Saaty and Forman 1993).
Step 1 concerns the construction of suitable hierarchies. This step is based on
findings indicating that when elaborating information, the human mind recognizes
objects and concepts, and identifies relations existing between them. Because the
human mind is not able to perceive simultaneously all factors affected by an action
and their connections, it helps to break down complex systems into simple structures:
this simplification is possible by means of a logical process which aims at the
construction of suitable hierarchies. Hence, hierarchies can be useful in helping
the human mind to make decisions by constructing a framework of different and
numerous elements, which are separated and connected at the same time. ‘A
hierarchy is a particular type of system, which is based on the assumption that the
entities, which we have identified, can be grouped into disjoint sets, with the entities
of one group influencing the entities of only one other group, and being influenced
by the entities of only one group’ (Saaty 1980: 7).
The simplest model of hierarchy consists of three steps. The first step coincides
with the main objective (called the ‘goal’) of the decision-making problem; the
second and third steps include criteria and alternatives. It is, however, possible to
develop more complex hierarchies (i.e. with more levels), which include a certain
number of sub-criteria. This means that factors affecting the decision are organized
in gradual steps from the general, in the upper level of the hierarchy, to the
particular, in the lower levels. With reference to the same decision-making problem
it is also possible to construct more than one hierarchy, for example, a first hierarchy
for benefits, a second one for costs, and a third one for risks. In order to obtain
satisfactory results, hierarchies should be large enough to capture all the major
factors relevant for the decision-making problem but small enough to remain
sensitive to changing crucial factors.
Step 2 establishes priorities between elements of the hierarchies by means of
pairwise comparisons (i.e. comparing elements in pairs with respect to a given
criterion). In the AHP approach, pairwise comparisons are used for establishing
priorities or weights among elements of the same hierarchical level. They are
compared in pairs with respect to the corresponding elements in the next higher
level, obtaining a matrix of pairwise comparisons. For representing the relative
importance of one element over another, a suitable evaluation scale is introduced
(Saaty 1988, 1992), called ‘Saaty’s scale’. It defines and explains the values 1 to 9
assigned to judgements in comparing pairs of elements in each level with respect
to a criterion in the next higher level. In particular, the following values which
express the intensity of importance between the elements are used:
1 if two elements have ‘equal’ importance;

3 if ‘moderate’ importance of one element over another is recognized;
5 if ‘strong’ importance of one element over another is recognized;
7 if ‘very strong’ importance of one element over another is recognized;
9 if ‘extreme’ importance of one element over another is recognized.
The numbers 2, 4, 6, 8 can be used to express intermediate values between two

adjacent judgements (i.e. between ‘equal’ and ‘moderate’, ‘moderate’ and ‘strong’,
etc.). When element i compared with j is assigned one of the above numbers, then
element j compared with i is assigned its reciprocal.
In particular, for each criterion C, an n-by-n matrix A of pairwise comparisons
is constructed. The components aij (i, j = 1, 2, …, n) of the matrix A are numerical
entries, which express (by means of Saaty’s scale) the relative importance of the
element i over the element j with respect to the corresponding element in the next
higher level. Thus the matrix A has the form:
a11 a12 a1n 

 
a21 a22 a2n 
A ≡  
 
 
an1 an 2 ann 
where:
1
aii = 1 a ji = aij ≠ 0
aij
In order to calculate relative priorities among the n elements considered, the

‘principal eigenvector’ (v, with usually Svi ¹ 1) of the matrix A is computed. Then
this eigenvector is normalized obtaining the ‘priority vector’ (x, with Sxi = 1), which
expresses the priorities among the elements belonging to the same node of the
hierarchy. Each component of the vector x represents the ‘local priority’ of an
element (i.e. a node of the hierarchy) of the pairwise comparisons; the ‘global
priority’ of that element is the product of its local priority with the global priority
of the upper node. Of course, the local priority of nodes at the first and second
levels is equal to their global priority. Thus each element at each level has a weight
(i.e. its global priority) assigned to it. The composite weight for each of the elements
at the final level (i.e. the alternatives) is obtained by multiplying the weights along
each path of the hierarchy from the apex to the final element and adding the
resultant weights from all paths to the element. The result is a vector of final
weights for the alternatives under consideration: the higher its weight the better
the alternative is. Moreover, it can be examined how priorities among alternatives
change according to the variation of weights assigned to the upper elements
(‘sensitivity analysis’).
Step 3 checks for the logical consistency of pairwise comparisons. In comparing
elements, inconsistency of a certain degree can arise: in the AHP approach the
‘maximum or principal eigenvalue’ (called lmax) of each matrix of pairwise
comparisons is computed for checking the degree of inconsistency.4 In particular
the consistency of a positive reciprocal n-by-n matrix is equivalent to the requirement
that its maximum eigenvalue lmax should be equal to n; if the matrix is inconsistent
lmax > n it is possible to estimate the departure from consistency by the difference
(lmax –1) divided by (n –1) (Saaty and Vargas 1991; Saaty 1994). The closer lmax is
to n the more consistent is the matrix A and its deviation from consistency may be
represented by the so-called ‘consistency index’ (C.I.):
l max - n
C .I . =
n -1
Saaty proposes a ‘consistency ratio’ in order to appreciate the consistency of the
matrix which is calculated by dividing the C.I. by R.I., where R.I. represents an
experimental ‘random index’ which increases as the order of the matrix increases.5
A consistency ratio of 0.10 or less is considered acceptable; if this ratio is more
than 0.10, it is necessary to reformulate the judgements by means of new pairwise
comparisons.
The AHP method can take into account the interdependence existing between
the different evaluation criteria by means of the construction of suitable hierarchies.
The elements of a hierarchy (i.e. criteria, sub-criteria, alternatives) are not linked
to all the others but they are grouped into disjoint sets; if an interdependence
between some elements is recognized, they are placed in the same hierarchical
level and connected to the same element of the next higher level. If the choice of
the number of hierarchical levels is unlimited (although, as a rule of thumb, no
more than seven elements for each node is best) it is possible to insert in the hierarchy
all the criteria required by the decision-making problem. In this sense a complete
(i.e. exhaustive) list of evaluation criteria is allowed. A cardinal weight is assigned
to each element of the hierarchy by means of opportune pairwise comparisons.
The weights express the importance of one element over the others. The final
result offered by the method is a complete ranking of the evaluated alternatives.
AHP facilitates stakeholder participation. Different social actors are invited to
assign weights to the evaluation criteria; in this perspective weights reflect different
social views. In comparing elements in pairs the different judgements given by
each actor can also be integrated – by means of an arithmetic average – obtaining
only one weight for each criterion, which expresses synthetically the points of
view of all the involved stakeholders. It is also possible to promote the actors’
communication if they are called to work together in order to identify general
objectives, evaluation criteria, sub-criteria, and deduce their importance.
The method has been applied to a range of decision-making problems (e.g.
corporate policy and strategy, public policy, political strategy, environmental
planning) at different geographical scales. Sometimes also using other evaluation
methods it is possible to make a pairwise comparison between criteria as carried
out in the AHP if normalized weights are required. In this sense a combination
with other methods is possible. AHP allows use of qualitative and quantitative
data; in particular, cardinal numbers can be directly normalized without the need
of pairwise comparisons.
Evaluation matrix (Evamix)

Evamix was developed by Henk Voogd. The evaluation matrix in this method
may include mixed (i.e. qualitative and quantitative) data. Both the method and
the software are called Evamix. From a procedural point of view this approach
consists of five steps: (1) make a distinction between ordinal and cardinal criteria;
(2) calculate dominance scores for all ordinal and cardinal criteria; (3) calculate
standardized dominance scores for all ordinal and cardinal criteria; (4) calculate
overall dominance scores; and (5) calculate appraisal scores (Voogd 1981, 1983).
Step 1 makes a distinction between ordinal and cardinal criteria. The first step
is the construction of an evaluation matrix E, which is an m-by-n matrix
characterized by m evaluation criteria and n alternatives. Its components are
qualitative or quantitative entries, which express by rows the performance of each
alternative with respect to a certain criterion. Given a set of evaluation criteria
j (j = 1, 2, …, m) and a finite set of alternatives i (i = 1, 2, …, n), the evaluation
matrix E will be characterized by its qualitative and quantitative6 components ej i :
Ê e11 e12 e1n ˆ

Áe e22 e2n ˜
21
E∫Á ˜
Á ˜
Áe emn ˜¯
Ë m1 em2
The set of criteria j is divided into two subsets, denoted O and C, where O is the
set of the ordinal (qualitative) criteria and C the set of cardinal (quantitative) criteria,
obtaining two distinct evaluation matrices: EO (ordinal criteria/alternatives) and
EC (cardinal criteria/alternatives). This way the differences among alternatives
can be expressed by means of two dominance measures: the first one based on
ordinal criteria and the second one based on cardinal criteria. In particular, in
order to construct the cardinal dominance score, the components ej i Œ EC are
standardized to a common unit; in this way all the quantitative criteria are expressed
on a scale from 0 to 1. All the standardized scores should have the same direction;
this means that a higher score implies a better score. In the case of criteria for
which lower scores mean better scores, the standardized scores are transformed by
subtracting them from 1.
Additionally, a qualitative weight can be assigned to each criterion constructing
a vector wj (j = 1, 2, …, m); the associated cardinal weights wj can be obtained – in
an approximated way – by looking at the ‘extreme weights sets’ which delimit the
values that metric weights may have.7 Because weights represent subjective value
judgements, it is preferable to use different priority sets so that the consequences
of different political viewpoints can be illustrated.
Step 2 requires calculation of dominance scores for all ordinal and cardinal
criteria. Differences between alternatives are expressed by means of two dominance
scores, for ordinal (aii¢) and cardinal (aii¢) criteria:
(
a ii ¢ = f e ji , e ji ¢ , w j ) ("j ŒO )
(
aii ¢ = g e ji , e ji ¢ , w j ) (" j ŒC )
where eji ( e ) represents the score of criterion j and alternative i ( i ¢ ), w j represents
ji ¢
the weight attached to criterion j, and f and g are two different functions. If each
extreme (quantitative) weight set w is substituted into the above formulae instead
of the qualitative weights w j , it is possible to define – for each extreme weight set
– a new dominance measure for ordinal ( a ii ¢ ) and cardinal ( aii ¢ ) criteria:
(
a ii ¢ = f e ji , e ji ¢ , w j ) ("j ŒO )
(
aii ¢ = g e ji , e ji ¢ , w j ) ( "j Œ C )
The scores expressed by means of these formulae reflect the degree in which
alternative i dominates alternative i¢ for ordinal and cardinal criteria respectively.
Step 3 calculates standardized dominance scores for all ordinal and cardinal
criteria. The dominance scores a ii ¢ and aii ¢ are standardized into the same
measurement unit in order to make them comparable. If h is a standardization
function, the ‘standardized dominance measures’ for all ordinal (d ii ¢ ) and cardinal
(dii ¢ ) criteria have the following expression:
δ ii ' = h(α ii ' )
d ii ' = h(a ii ' )
The standardization of the scores α ii ' and a ii ' can be done in three different
ways: the ‘subtractive summation technique’, the ‘subtracted shifted interval
technique’ and the ‘additive interval technique’.8 Also in this case the standardized
dominance scores reflect the degree in which alternative i dominates alternative
i ¢ for ordinal and cardinal criteria. Different scores are obtained using the three
different techniques.
Step 4 calculates overall dominance scores. The ‘overall dominance measure’
mii ' for each pair of alternatives (i, i ¢ ) – giving the degree in which solution i
dominates solution i ¢ – is calculated by means of the following formula:
mii ' = wO δ ii ' + wC d ii '
where wO represents the weight of the qualitative criterion set O and wC
represents the weight of the quantitative criterion set C:
wO = ∑ w j wC = ∑ w j
j∈O j∈C
Step 5 calculates appraisal scores. The overall dominance measure mii ¢ may also
be considered as a function k of the ‘appraisal scores’ si and si ¢ of the alternatives
i and i ¢ :
mii ' = k (s i , s i ' )
At the same time this means that the appraisal score of an alternative depends
on the overall dominance score such as calculated in Step 4. In particular, three
different appraisal scores of alternative i with respect to the n other alternatives
are obtained if the subtractive summation technique, subtracted shifted interval
technique or additive interval technique has been used in Step 3. Additionally, by
using the appraisal score of all three techniques (t = 1, 2, 3) a standardized ‘average
appraisal score’ ai for alternative i can be calculated:
3
Ê s - s- ˆ
ail = Â Á ti+ ti- ˜
t =1 Ë sti - sti ¯
where sti is the appraisal score of alternative i for each technique; sti- is the lowest
sti -score for each technique; sti+ is the highest sti -score for each technique. The
result is a complete ranking of alternatives and the higher ai is the better is
alternative i.
Since the starting point of Evamix is the construction of an evaluation matrix,
it is not possible with this method to take the interdependence between the different
evaluation criteria into account. But if we consider that there is no limit on the
choice of the number of criteria to insert in the evaluation matrix, it is possible to
consider all the criteria required by the decision-making problem. In this sense a
complete (i.e. exhaustive) list of evaluation criteria is allowed. An ordinal weight,
expressing the importance between all the relevant criteria, is assigned to each
criterion, and subsequently ordinal weights are transformed into cardinal ones.
Evamix supports stakeholder participation. Different social actors are invited
to assign weights to the evaluation criteria, and in this case the different weights
given by each actor are not integrated but they are used to show the different
points of view (i.e. economic, social, environmental) expressed by the stakeholders
involved. It is also possible to promote actor communication if they are called to
work together in order to identify general objectives, points of view, evaluation
criteria, sub-criteria, and deduce their importance. The method is used to cope
with different decision-making problems at different geographical scales. Most
frequently, Evamix been used in urban and regional planning. Evamix allows using
qualitative and quantitative data. In particular, quantitative data are expressed in
an ordinal scale and the cardinal numbers are standardized in a scale from 0 to 1.
Outranking methods
The development of the outranking methods was started in France in the late
1960s by Bernard Roy and his team (Roy 1985). Outranking methods were
developed in an attempt to manage with less strong assumptions (about existence
of utility function, additivity, etc.) and to require less information from decision-
makers (especially preference intensities, rates of substitution) than the methods
described above.
The outranking methods also differ fundamentally from other methods in the
approach insofar as they start from the assumption that the decision is a process
during which decision-makers may change their preferences in response to the
information provided and/or after thorough reflection about the problem. For
this reason the developers of outranking methods considered it important to allow
at the beginning for incomparability of options. This avoids a complete ranking
being identified too early; the aim is therefore to stimulate thorough consideration
and to avoid premature conclusions. Hence, more than other methods, these
methods encourage interaction between the model and the decision-maker/s. The
outranking algorithms also account for the political reality that options which
perform very poorly in one dimension are likely to face severe opposition from at
least some of the stakeholders and are therefore deemed unacceptable.
ELECTRE III
Roy (1985) developed one of the most widely used software packages, known as
ELECTRE, which is available in four different main releases (ELECTRE I, II,
III, IV) and two other versions (ELECTRE IS and TRI). A comprehensive
comparison of these approaches can be found in Vincke (1992). The underlying
principle of the ELECTRE approach stems from processing a system of pairwise
comparisons of the alternatives. This procedure involves the use of a system of
preference relations (SPR) that is based on the exploitation of the outranking
relation. According to this relation, an action will outrank another if the first action
is considered to be at least as good as the second. Roy (1996) defines a system like
this as a basic system of outranking relations. Recently it was pointed out that
release III is remarkable among the ELECTRE family, since it takes explicitly into
account the indifference and preference thresholds but, furthermore, it is based on
an outranking relation, which is less sensible to the variation of the input data
(Scarelli 1997). The system solves ranking problems, dividing the set A of actions
into equivalence classes (Scarelli 1997). In release III of ELECTRE, the process
can be sketched in the following steps.
A cardinal impact table is prepared for a finite number of alternatives ‘measured’
according to a set of pseudo-criteria. The latter are criteria whose SPR includes
weak preference, indifference and incomparability, and a coherent system of
thresholds has to be set. The program allows to set thresholds for indifference and
for preference, according to a linear relationship: T = a • g(a) + b, where T is the
total threshold, a is a proportional threshold, b is a fixed threshold and g is the
criterion function. This threshold is defined directly when it refers to the maximum
value of the criterion, and indirectly when it refers to the minimum value. This
allows constructing fuzzy outranking relations, given that it is possible to state how
much an action outranks another. In fact, a series of pairwise comparisons is set
through the use of concordance and discordance indices. A general framework for
the calculation of these indices is given by Voogd (1983). The degree of dominance
of alternative i over alternative i¢ is measured by the following index:
1/ a
È Â j Œy w aj ˘
cii ¢ = Í ˙
Í Â jwj ˙
a
Î ˚
where y is the set of criteria that indicate a preference for the alternative a with
respect to the alternative a ¢ , w the criterion weight and a a scaling parameter
greater than one. ‘By means of this scaling parameter the analyst is able to vary
the importance of the small weights and small divergences between the scores’
(Voogd 1983: 123). In many applications, the value of the scaling parameter is set
equal to one, which allows managing a simpler functional algorithm.
The degree of discordance between the same alternatives is measured by the
following index:
a 1/ a
dii ¢ =
È
Í (
Â j Œj w j | e ji - e ji ¢ | ˘˙ )
Í w | e - e ji ¢ | ˙˙
a
(
ÎÍ Â j j ji )
˚
where j is the set of criteria that do not agree about the preference of alternative
i with respect to alternative i ¢ .
In the generalized concordance analysis the values of the concordance and
discordance indices individuate the subsets of alternatives that are accepted and
rejected. The first group consists of the alternatives that fulfil the following
requirements: Tc e ≥ h and Td ≥ m.
In ELECTRE III, the concordance and discordance indices are calculated with
reference to the values that the criterion functions take for the consequences of
each alternative. In the same pattern, the threshold functions are modified as linear
functions of the pseudo-criterion g. For each criterion gj a preference threshold pj
and an indifference threshold qj are set.
Given the pseudo-criterion g and two variable alternatives i and i ¢ , two
concordance indices cj (i, i ¢ ) are calculated for the comparison of alternative i with
respect to alternative i ¢ and vice versa. For the comparison between alternative i
and i ¢ , the concordance index takes values between 0 and 1.
The discordance index dj (i, i ¢ ) is calculated with reference also to the veto
threshold vj . This threshold is associated with a criterion that is considered to be
very important. If the difference between two functional values in the case of a
number of alternatives is greater than the threshold value, this condition acts as a
veto to the affirmation: the first alternative outranks the second.
The construction of the outranking relation is based upon the credibility index
s(i, i ¢ ), which ranges between 0 and 1 and is a function of the concordance and of
the discordance index. The credibility index takes the following values:
s (i, i ¢ ) = c (i, i ¢ ), if dj (i, i ¢ ) < cj (i,);

otherwise
1 - d j (i , i ¢ )
s (i , i ¢ ) = c j (i , i ¢ ) ’
1 - c j (i , i ¢ )
The credibility index provides the information for the construction of two
rankings of the alternatives. First, from the worst to the best (ascending distillation)
and the second from the best to the worst (descending distillation). The intersection
of these distillations yields the final distillation graph. This graph indicates the
system of relationships among the alternatives, according to a system of oriented
arcs. This framework reveals incomparability relationships that the other distillations
do not consider.
Roy (1996) considers that in every decision-making situation, four problem areas
have to be taken into account: Pa (choice), Pb (sort), Pg(rank), and Pd (description).
With reference to this outline set forth by Roy (1996), the procedure does not
deliver a complete ranking of the evaluated alternatives, leaving partially unresolved
what Roy (1985) calls the problematic P. g (ranking of the alternatives). On the
other hand, the procedure of ELECTRE III is helpful for what Roy (1985) names
the problematic P.a (choosing the best action).
ELECTRE III is based on the assumption that the system of individual
preference relationships can be described by means of a sophisticated analysis of
the relationships of preference, indifference, and incomparability. As a matter of
fact, incomparability enters as a puzzling factor into the final ranking. On the one
hand, this complex analysis involving the assessment of several thresholds and
preference relationships may be judged as a remarkable effort. On the other hand,
it may lead to cumbersome work. Often it is very difficult to express credible
threshold values and the resulting ranking can end up being hardly understandable.
As a matter of fact, in the situation where results were described by a Kernel
graph, the complex meaning stemming from this kind of representation would be
hardly transmissible to a group of laypersons, such as a local community. Local
communities seem to be more keen to comprehend and discuss definitive positions
and comparable outcomes than to give personal interpretation of relationships.
The nature of the ranking, which is presented by a graph, allows also for incompar-
ability among the alternatives: such an incompletely resolved outcome challenges
the analyst who should then, if equipped with abilities of good advocacy, foster
communication among stakeholders to discuss the incomparability. Politicians often
are unwilling to accept the complexity of this structure and show a tendency to
sceptical behaviour against an incomprehensible technique for decision-making.
Often it is the expectation that the multiple criteria analysis will solve the problem
at hand completely. The obvious fact that MCDA supports decision-making, but
must not substitute decision-makers, becomes even more apparent with outranking
methods than with the previously discussed ones. These are the main reasons why
in Table 5.2 ELECTRE is rated highly with respect to criteria completeness and
interdependence, while it is rated poorly with respect to transparency and to
applicability of criteria. It should be noted that in terms of ‘meaning of weights’
ELECTRE received a zero score, since it does imply weighting procedures con-
nected to the combinatory use of the so-called ‘weights’ and to the thresholds.
Regime
The Regime method was developed by Hinloopen et al. (1983), assessed and refined
by Hinloopen (1985) and Hinloopen and Nijkamp (1990). This method can be
considered as a multiple criteria evaluation framework belonging to the family of
dominance analysis approaches, even though it has been pointed out that it is
quite distinct from other methods in this family, mainly because Regime is a
qualitative multiple criteria method. Qualitative multiple criteria methods aim to
provide a tool for analysing complex situations, which are impossible to model by
means of quantitative information. Hinloopen and Nijkamp (1990) list several
such methods: (a) the extreme expected value method (Kmietowicz and Pearman
1981; Rietveld 1980), (b) the permutation method (Mastenbroek and Paelink 1977),
(c) the frequency method (Van Delft and Nijkamp 1977), (d) the multidimensional
scaling method (Nijkamp and Voodg 1981), and (e) the mixed data method (Voogd
1983).
Since human activities and settlements are characterized by phenomena that
are often impossible to measure, studies have focused on the construction of
techniques able to deal with ‘soft’ and ‘fuzzy’ information. ‘These qualitative multi-
dimensional judgement methods are essentially more in agreement with “satisfying
behaviour” (based on Simon’s bounded rationality principle) than with conventional
“optimising decision making”, as imprecise information precludes the attainment
of an unambiguous maximum solution’ (Hinloopen and Nijkamp 1990: 38).
As in Evamix, an evaluation table is given and composed by eij scores of a
number n of alternative scenarios with respect to m criteria. In the case of ordinal
information the weight can be represented by means of rank orders wj in a weight
vector w: w = (w1, w2, …, wj )T. The higher the value of the weight, the better the
correspondent criterion.
The following main features characterize the multiple choice regime method.
First, with this method the evaluation table can contain cardinal as well as ordinal
data; this is accomplished by treating cardinal information as ordinal, with reference
to the ranking position of each alternative. Second, the basis of the method is the
regime vector. For two alternative choice options i and i ¢ , the difference of the
criterion scores is assessed by considering the difference between the scores sii¢j = eij
– eei¢j.
In case of ordinal information, the order of magnitude of sii¢j is not relevant,
but only its sign qii¢j . Therefore, if qii¢j = +, then alternative i is better than alternative
i¢ for criterion j; i¢ qii¢j = –, then alternative i is better than alternative i¢ for criterion
j. The extension of the pairwise comparison to all the alternatives leads to the
following J ¥ 1 regime vector: rii¢ = (sii¢1,sii¢2, sii¢3, … sii¢n ). The regime vector,
composed of + or – signs, or eventually zeros, ‘reflects a certain degree of (pairwise)
dominance of choice option i with respect to i¢ for the unweighted effects for all J
judgement criteria’ (Hinloopen and Nijkamp 1990: 41). For all the combinations
of comparisons, a regime matrix is drawn with the following pattern:
R = ÍÎ r12r13 … rI 1r21 ……… rIa ……… rI ( I -1) ˙˚
Usually, the regime vector is not completely composed of a series of ‘+’ or a

series of ‘–’. Instead it is often composed of both signs. Additional information
has to be processed by means of ordinal values, on the assumption that the ordinal
nature of the weights reflects the inability to measure human preferences, which
in principle are cardinal. This weight vector can be considered as a ‘rank order
representation of an (unknown) underlying cardinal stochastic weight vector’
(Hinloopen and Nijkamp 1990: 41): w* = (w1* , w2* ,…w*j )T with 0 < w*j < 1. The
consistency hypothesis implies that the weight vector is supposed to be consistent
with the quantitative information incorporated in an unknown cardinal vector.
The next assumption is that the dominance relation can be represented by
means of the following stochastic expression referred to the following weighted
linear summation of cardinal entities:
n ii ¢ = Â s ii ¢j w*j
j
If nii¢ is positive, then a dominance relation of choice option i over i¢ is revealed.

Since the information about w*j is unknown, the probability of the dominance of
i with respect to i¢ is introduced in the following pattern: pii¢ = prob (nii¢ > 0). Thus
an aggregate probability measure (or success score) can be defined as follows:
1
pi = Â pii ¢
I - 1 i ¢π i
Hence, pi is the average probability that alternative i is given a higher value

than another alternative. The rank order of choice options is then determined by
the rank orders (or the order of magnitude) of the pi. Since the assessment of the
pii¢ is the crucial point, a uniform density function is usually assumed for the
probability distribution of the w*j . This argument is referred to as the ‘principle of
insufficient reason’, and as the Laplace criterion, in case of decision-making under
uncertainties (Taha 1976). According to this statement, without any prior
information, ‘there is no reason to assume that a certain numerical value of w* has
a higher probability than any other value’ (Hinloopen and Nijkamp 1990: 42).
Regime differs in some important points from other outranking methods. While
ELECTRE requires cardinal rankings, the Regime method allows using mixed
data so that cardinal and ordinal criterion can be included. The Regime method
adopts ordinal weights; this feature seems to depict real systems of preferences
better than other multiple criteria weighting methods. This approach is based on
more prudent assumptions as it refrains from the difficult task of attaching cardinal
values to measure the intensity of criteria importance.
In terms of shortcomings and benefits, this method is in a sense the counterpart
of ELECTRE III. It has a simpler structure of the preference model than
ELECTRE III, since qualitative and quantitative data can be used for criteria. In
other words, this method is able to process mixed data, because it can adopt an
ordinal scale for the impact scores. The result of the evaluation process with Regime
is a complete ranking of the alternatives. The outcome of the analysis is quite easy
to communicate and discuss. However, the Regime method does not admit
incomparabilities among alternatives. Also it is not clear how the system operates
in the stochastic domain. This is mainly due to the random generation system
adopted for the weight vector: the analyst, involved in the task of explaining the
exact mechanism of the algorithm, may meet significant difficulties in
communicating the meaning and in involving the decisional community. Hence,
despite the apparently simple structure of the Regime method, the resulting
decision-making process may not be so transparent.
Novel approach to imprecise assessment and decision

environments (NAIADE)
This was a MCDA method developed by Giuseppe Munda (1995), whose impact
or evaluation matrix may include quantitative and qualitative data; in particular
the values assigned to the criteria for each alternative may be expressed in the
form of either crisp, stochastic, fuzzy numbers or linguistic expressions. Hence it
allows the use of information affected by different types of uncertainty. It is a
discrete method, and no weighting of criteria is used explicitly. The method is
implemented by a software application called NAIADE.
From a procedural point of view the method consists of three steps: (1) the
pairwise comparison of alternatives; (2) the aggregation of all criteria; (3) the
evaluation of alternatives (Munda et al. 1994; Munda 1995; Menegolo and
Guimarães Pereira 1996).
Step 1 requires pairwise comparison of alternatives. The first step is the
construction of an evaluation matrix E, which is an m-by-n matrix characterized
by m evaluation criteria and n alternatives. Its components are qualitative or
quantitative entries, which express by rows the performance of each alternative
with respect to a certain criterion. Given a set of evaluation criteria j ( j = 1, 2, …,
m) and a finite set of alternatives i (i = 1, 2, …, n), the evaluation matrix E will be
characterized by its qualitative and quantitative9 components eji:
Ê e11 e12 e1n ˆ

Áe e22 e2n ˜
21
E∫Á ˜
Á ˜
Áe emn ˜¯
Ë m1 em2
and the pairwise comparison of alternatives (with reference to each criterion) is

carried out by means of the concept of ‘semantic distance’ between two fuzzy
sets.10 More precisely, if S1 and S2 are two fuzzy sets and m1 and m2 are respectively
their membership functions, it is possible to define two new functions:
f ( x ) = k1m1 ( x ) and g ( y ) = k2m 2 ( y )
obtained by rescaling the ordinates of m1 and m2 through k1 and k2 such as:
+• +•
Ú
-•
f ( x )dx = Ú g ( y )dy = 1
-•
The semantic distance between the two fuzzy sets (that is the distance between
all points of their membership functions) is computed. The same distance concept
can be extended to stochastic measures considering f(x) and g (y) probability density
functions of the measures.
The pairwise comparison of alternatives is based on six ‘preference relations’:
1 much greater than (>>)

2 greater than (>)
3 approximately equal to (@)
4 very equal to (=)
5 less than (<)
6 much less than (<<)
expressed for each criterion starting from the distance between alternatives. The
above preference relations are analytically defined by means of six functions that
express (for each criterion on the base of the distance between alternatives) an
index of credibility of the statements that an alternative is ‘much greater’, ‘greater’,
approximately equal’, ‘very equal’, ‘less’ or ‘much less’ than another.11 This
credibility index goes, increasing monotonically, from 0 (definitely non-credible)
to 1 (definitely credible). In particular, given a criterion j and a pair of alternatives
i and i¢, it is possible to define six membership functions (different for crisp criterion
scores and for stochastic or the fuzzy criterion scores), which are respectively
named12:
m >> (i , i ¢ ) j (for much greater than)

m > (i , i ¢) j (for greater than)
m @ (i , i ¢ ) j (for approximately equal to)
m = (i , i ¢ ) j (for very equal to)
m < (i , i ¢) j (for less than)
m << (i , i ¢ ) j (for much less than)
Membership functions related to stochastic or fuzzy numbers consider in their

mathematical expressions the semantic distance defined above.
Step 2 requires aggregation of all criteria. In order to take into account all
criteria simultaneously, it is necessary to aggregate the evaluations related to the
pairwise performance of alternatives according to each single criterion. A
‘preference intensity index’ of one alternative with respect to another is calculated:13
m * (i , i ¢ ) = f ( m * (i , i ¢ ) j )
and then an aggregate ‘fuzzy preference relation’ can be obtained:

Ïm >> (i , i ¢ ) H ( >> )
Ôm ( i , i ¢ ) H (>)
Ô >
Ôm @ (i , i ¢ ) H (@)
Ì
Ôm = ( i , i ¢ ) H (=)
Ôm < ( i , i ¢ ) H (<)
Ô
Óm << (i , i ¢ ) H ( << )
where m* (i , i ¢ ) is the overall evaluation of a given fuzzy relation for each pair of
actions and H(*) is the associated entropy level.14
Step 3 is the evaluation of alternatives. The evaluation of the alternatives derives
from the information provided by the above aggregate fuzzy preference relation,
defining for each action two functions f+ and f –, from which separate rankings of
alternatives are obtained. The function f+(i) is based on the ‘better’ and ‘much
better’ preference relations and with a value from 0 to 1 indicates how the alternative
i is ‘better’ then all other alternatives. The second function f–(i) is based on the
‘worse’ and ‘much worse’ preference relations, its value going from 0 to 1 which
indicates how the alternative i is ‘worse’ than all other alternatives (Menegolo and
Guimarães Pereira 1996). The final ranking of alternatives comes from the
intersection of two separate rankings obtained by means of the functions f+(i) and
f-(i), taking into account that it can also be an incomplete ranking because non-
dominated alternatives are calculated.
Additionally, NAIADE allows for another type of evaluation. It analyses conflicts
between different interest groups and the possible formation of coalitions according
to the proposed alternative options. Besides the ‘impact matrix’, also an ‘equity
matrix’ is constructed, which contains linguistic evaluations of different social groups
for each alternative. In particular, ‘equity analysis is performed by the completion
of an equity matrix from which a similarity matrix is calculated. Through a
mathematical reduction algorithm, it is possible to build a dendrogram of coalitions
which shows possible coalition formation, and a level of conflict among the interest
groups’ (Menegolo and Guimarães Pereira 1996: 1).
Because the starting point of NAIADE is the construction of an evaluation
matrix, it is not possible with this method to take into account the interdependence
existing between some different evaluation criteria. Since there is no limit on the
choice of the number of criteria that can be inserted in the evaluation matrix, it is
possible to consider all the criteria required by the decision-making problem. In
this sense a complete (i.e. exhaustive) list of evaluation criteria is allowed. No weights
are assigned to the criteria. The final result offered by the method can be an
incomplete ranking of the evaluated alternatives. In fact, the final ranking comes
from the intersection of two separate rankings which allows to calculate also the
non-dominated alternatives.
Stakeholder participation is explicitly supported in NAIADE. The method allows
the construction of an equity matrix which reflects how the different social groups
involved evaluate the alternatives. It is also possible to promote the actors
communication if stakeholders are called to discuss their evaluation of the
alternatives. The method is used to cope with different decision-making problems
(it has especially been used for problems of unsustainability) at different
geographical scales. NAIADE allows using qualitative and quantitative data. In
particular, qualitative data are expressed by means of linguistic evaluations and
quantitative data may be expressed either in crisp, stochastic or fuzzy numbers.
Programming methods
Multi-Objective-Programming (MOP), Goal Programming (GP) and their variants
are continuous multiple criteria methods. In contrast to the methods presented so
far, the programming methods do not rank or sort a finite number of alternatives,
but the alternatives are generated during the solution process on the basis of a
mathematical model formulation.
The roots of GP lie in the work by Charnes and Cooper in the early 1960s.
After the mid-1970s it became widely applied due to seminal work by Lee (1972)
and Ignizio (1976). A number of specific software tools were developed, but often
general programming software packages (like GAMS) are applied. From a
procedural point of view MOP consists of two steps: (1) identification of Pareto
efficient (also called non-dominated) solutions; and (2) identification of the most
preferred alternative together with the decision-maker(s).
Step 1 consists of finding the non-dominated solutions. The problem is formul-
ated as a task of simultaneous maximization/minimization of several objectives
subject to a set of constraints. In a so-called criterion space it can be defined as
follows: ‘max’ q, s.t. q Œ Q, where Q is a so-called feasible region in the criterion
space. The set Q is not specified directly, but by means of decision variables as
usually done in single optimization problems: ‘max’ q = f(x) = (f1(x),...,fk(x)), s.t. x Œ
X, where X is a feasible set. The functions fi, i = 1, 2, ..., k are objective functions
(Korhonen 2001).
Since the simultaneous optimization of all objectives is impossible – given a
certain level of conflict between them in most real problems – MOP tries to find
the set of Pareto efficient solutions. According to Ballestero and Romero (1998) a
necessary condition to guarantee the rationality of any solution to an MCDA
problem is the following: ‘A set of solutions in a MCDA problem is Pareto efficient
(also called non-dominated), if their elements are feasible solutions such that no
other feasible solution can achieve the same or better performance for all the criteria
being strictly better for at least one criterion’ (p.7). The goal is therefore to divide
the feasible set into a subset of Pareto efficient solutions and a subset of Pareto
inefficient solutions.
Step 2 requires choosing the most preferred solution. Any choice from among
the set of efficient solutions is an acceptable and ‘reasonable’ solution, as long as
we have no additional information about the decision-maker’s preference structure.
The final solution q Œ Q is called ‘the most preferred solution’. It is a solution
preferred by the decision-maker to all other solutions (Korhonen 2001). Since we
are always facing a number of conflicting goals in multiple criteria problems, a
solution can never be optimal (or ideal), but rather a compromise. With regards to
how much we need to know about the decision-maker’s preferences, different
approaches were developed: there are (a) those that need knowledge of the decision-
maker’s preference; (b) those in which the decision-maker’s preferences are pre-
emptively determined; and (c) those that progressively reveal the preferences of
the decision-maker through man–machine interactions (interactive methods) (Kalu
1999). The latter typically operate with decision-maker’s aspiration levels regarding
the objectives on the feasible region. The aspiration levels are projected via
minimizing so-called achievement scalarizing functions (Wierzbicki 1980). No
specific behavioural assumptions, e.g. transitivity, are necessary (Korhonen 2000).
At the beginning of MOP research the expectation was that computing all
efficient extreme points and then asking the decision-makers to examine their
criterion vectors to select the best one could solve the problem. However, after
vector-maximum codes became available, it was found that the number of efficient
extreme points generated by MOP was far larger than had been expected. As a
result, interactive procedures for exploring efficient sets for the best solution became
popular. Many different non-reference point and reference point procedures are
available now (Steuer and Gardiner 1990).
In GP the impossibility of building a reliable mathematical representation of
the decision-maker’s preferences due to conflicts of interest and incompleteness
of available information is addressed in a different way. Two major subsets of GP
models can be distinguished. In the first type it is assumed that within this kind of
decision environment, decision-makers attempt to achieve a set of relevant goals
as close as possible to the set of targets established (Ballestero and Romero 1998).
The algebraic representation is given as the sum of unwanted deviations from
target values if a number of objectives are minimized (weighted GP):
k
min zÂ (ui ni + vi pi ), s.t. f i (x ) + ni - pi = bi , i = 1…Q , x ŒC5 ,
i =1
where fi(x) is a linear function (objective) of x, and bi is the target value for that
objective. ni and pi represent the negative and positive deviations from this target
value. ui and vi are the respective positive weights attached to these deviations in
the achievement function z (Tamiz et al. 1998: 570). The methodology rests on the
following basic scheme (Vincke 1992): (1) set the values one wishes to attain on
each criterion (the objectives); (2) assign priorities (weights) to these objectives; (3)
define (positive or negative) deviations with respect to these objectives; (4) minimize
the weighted sum of these deviations; and (5) perform a sensitivity analysis. In
practice, the weights are sometimes derived from the pairwise comparison of AHP.
Alternatively, they can be derived by use of an interactive MCDA method, like the
one by Zionts and Wallenius (1976) (as shown in Lara and Romero 1992). Variants
of this type (Chebyshev GP, MinMax GP, and Fuzzy GP) seek to minimize the
maximum deviation from amongst the set of unmet objectives instead. In the case
of another variant (Preference-Function GP) the relationship between the deviation
from the target and the penalty imposed is not the standard linear one, but it may
be piecewise linear, discontinuous, or even non-linear (Tamiz and Jones 1997).
In the other major subset of GP the deviational variables are assigned to a
number of priority levels and minimized in a lexicographic sense (lexicographic
GP). The assumption of non-continuity of preferences implies the impossibility
of ordering the decision-maker’s preferences by a monotonic utility function.
Lexicographic minimization means a sequential minimization of each priority
while maintaining the minimal values achieved by all higher priority level
minimizations. In mathematical terms this means:
Lex min a = (g1(n,p), g2(n,p), ... , gL(n,p)), s.t. fi(x) + ni – pi = bi, i = 1, ..., Q
The given model has L priority levels, and Q objectives. a is an ordered vector of
these L priority levels. ni and pi are deviational variables which represent the under-
and over-achievement of the i-th goal, respectively. x is the set of decision variables
to be determined. The g (within a priority level) function is given by
gl ( n, p) = ulq n1 + … + ulq nq + vl1 p1 + … + vlQ PQ ,
where u and v represent inter-priority level weights (Tamiz et al. 1998: 570). The
possibility to account for the non-continuity of preferences seems particularly
important for the analysis of environment problems (Spash 2000).
Instead of optimization, GP is based on Simon’s satisficing philosophy (Simon
1955, 1979). In comparison, GP is a more pragmatic approach. It was used in the
1960s and 1970s by practitioners in corporations and agencies and only slowly
received more attention by theorists. While the techniques applied are the same
programming techniques as for the other MOP methods, the underlying philosophy
differs. The idea of optimizing is abandoned; only satisficing is attempted.
MOP and particularly GP are the most frequently applied multiple criteria
methods. They have been applied to a wide array of problems, namely production
planning, oil refinery scheduling, health care, portfolio selection, distribution system
design, energy planning, water reservoir management, timber harvest scheduling,
problems of wildlife management, etc. Their popularity can be explained by their
flexibility; they can account for diverse variable types (continuous, integer, Boolean,
etc.) as well as constraints and objective functions (linearity, convexity,
differentiability, etc.) (Vincke 1992).
More recent additions to the field of MOP have been: (a) to allow for fuzziness in
the data which is a way to address the problem of uncertainty to some degree; and
(b) evolutionary algorithms which were developed as a response to the fact that for
models which are large in size and/or significantly non-linear, traditional solution
methods are often unable to identify the global optimum. In such cases, typically
linear approximations are applied instead in order to find a solution. Genetic
algorithms were shown to be applicable to examples of large non-linear models (e.g.
Mardle et al. 2000). They follow the concept of solution evolution, by stochastically
developing generations of solutions populations using a given fitness statistic.
MOP and GP allow addressing a completely different type of multiple criteria
problems. Instead of choosing from a limited number of alternatives, a continuous
set of alternatives is generated during the solution process. They are the only
methodologies in the field which provide the model-based calculation of non-
dominated alternatives. Another advantage is that restrictions which are for instance
needed for the analysis of decision-making processes based on criteria of strong
sustainability (non-substitutability) can be accounted for. These special features
turn out to be important for the decision-making situation in many case studies.
GP has often been criticized for being analytically inferior to MOP. While GP
is indeed a somewhat simpler method, it is often preferred because of its capacity
to handle large-scale operational real-life problems. MOP works efficiently for
moderate-size problems defined within a well-structured environment. When
programming methods are applied, interdependence between criteria is not
allowed. The completeness description of the phenomenon is desired, but not
explicitly mentioned. While MOP is very appealing theoretically it has the
disadvantage that for large problems and particularly if non-linear functions are
included, often no optimal solution can be found.
The development of interactive methods, which allow to reveal progressively
the preferences of the decision-maker through man–machine interactions, made
MOP much less reliant on doubtful assumptions about the decision-maker’s
preferences. A number of useful software tools were developed which allow the
decision-makers an important role in the decision process and enable ‘psychological
convergence’ (i.e. termination of the interactive process as and when the decision-
maker chooses). Hence, interactive MOP leads not only to a good structuring of
the problem but can also provide a useful tool for learning for the decision-maker
involved. A problem remaining with interactive methods is that mostly they do not
provide an objective basis upon which the decision-maker could base the decision
to terminate the interactive process (Schniederjans 1995).
The next step in perceiving the role of the decision-makers in a more democratic
way, namely accounting for stakeholder participation (group decision-making), is
however explicitly accounted for in few tools and is therefore exhausted in only
some case studies. Another drawback in the user context is that the process cannot
be made very transparent. MOP allows for the use of different geographical scales,
as well as different general scales (micro–macro levels). Concerning data, MOP is
able to deal with quantitative as well as qualitative and fuzzy data and allows for
risk via fuzziness. An interesting aspect with regard to the fuzzy version of MOP
was reported in several case studies: the large amount of time which is normally
needed to select all the data required can on some occasions be bypassed by using
the fuzzy relations (e.g. Chang et al. 1997). GP is not able to deal with a lot of
different data, such as qualitative and fuzzy sets.
Comparison of the methods

The evaluations of the individual methods using the quality criteria pointed to the
differences in the characteristics of the presented methods. The summary in Table
5.2 gives a detailed characterization of each method. Without reference to the
specific characteristics of the case study for which the methods should be applied,
the selection of the best applicable method is not possible. Consequently the aim
here is a characterization of the methods based on the quality criteria with regard
to sustainability issues in general. From the list of quality criteria we find the
following four groups most crucial:
• possibility to deal with complex situations (criteria, consideration of different

scales and aspects, i.e. geographical scales, micro–macro link, societal/technical
issues, type of data, uncertainties),
• possibility to consider non-substitutability (strong sustainability) issues,
• possibility to involve more than one decision-maker (stakeholder participation,
actors’ communication, and transparency), and
• information of stakeholders in order to increase their knowledge and change
their opinion and behaviour (problem structuring, tool for learning, trans-
parency, type of weights).
Some criteria could not be considered here because they are especially designed to
help the user in the selection of a method for a concrete decision-making situation,
and not designed for a theoretical method comparison with regard to general
sustainability issues without any specific application in mind, as done in this
comparison. Reference to the whole list of criteria was made in De Montis et al.
(2000), which included selected case studies of sustainable water use.
In handling complex situations, all methods show similar performance with
respect to the aspects and scales which can be considered; but they all show
weaknesses in the characteristics of the evaluation criteria which should be used
as operational components. Only AHP allows the interdependence of evaluation
criteria, while only MAUT and NAIADE allow non-linear preferences. Evamix,
ELECTRE III, and Regime do not permit either of these characteristics. In each
case quantitative and qualitative data can be addressed. In ELECTRE III
qualitative data have to be transformed first to a quantitative scale. Additionally,
some of the methods like GP/MOP and NAIADE include features to deal with
fuzzy data and stochastic numbers, which is the way that these methods deal with
risk and uncertainties. MAUT allows to deal with risk (not uncertainty) concerning
the outcomes of the alternatives by assigning probabilities to the utility functions
(von Neumann–Morgenstern utility functions). All of the methods provide the
possibility to carry out a sensitivity analysis or to apply qualitative data to consider
uncertainties and risks. Non-substitutability aspects can only be properly analysed
by using GP/MOP and ELECTRE III which allow to set constraints and thresholds
explicitly. Within the remaining methods non-substitutability can only be considered
by assigning flags to the respective criteria.
The performance to inform stakeholders in order to increase their knowledge
and change their opinion and behaviour is satisfactory for all methods. In MOP/
GP, ELECTRE III and Regime the transparency is only of medium quality, which
hinders stakeholder participation. The other methods are satisfactory or even good
Table 5.2 Summary of method comparison
MAUT AHP Evamix ELECTRE III Regime NAIADE MOP/GP
Operational components of methods
Criteria Inter- Not allowed Allowed Not allowed Not allowed Not allowed Not allowed Not allowed
dependence
Completeness Allowed Allowed Allowed Allowed Allowed Allowed Required
Non-linear Allowed Not allowed Not allowed Not allowed Not allowed Allowed Allowed for
preferences some variants
Weights Transparency Cardinal Cardinal Ordinal Cardinal Ordinal No weights Weights
of process, weights weights weights weights weights assigned assigned for
type of weights assigned assigned assigned assigned assigned some variants

Meaning Trade-offs Importance Importance Importance Importance No weights Importance
Solution-finding Complete Complete Complete Non- Complete Non- Non-
procedure ranking ranking ranking dominated ranking dominated dominated
option/s option/s option/s
calculated calculated calculated
Applicability in user context
Project Costs Depending on Depending on Depending on Depending on Depending on Depending on Depending on
constraints the numbers the numbers the numbers the numbers the numbers the numbers the numbers of
of attributes, of attributes, of attributes, of attributes, of attributes, of attributes, attributes,
stakeholders stakeholders stakeholders stakeholders stakeholders stakeholders stakeholders
involved, etc. involved, etc. involved, etc. involved, etc. involved, etc. involved, etc. involved, etc.
Time Depending on Depending on Depending on Depending on Depending on Depending on Depending on
the numbers the numbers the numbers the numbers the numbers the numbers the numbers
of attributed, of attributed, of attributed, of attributed, of attributed, of attributed, of attributes,
stakeholders stakeholders stakeholders stakeholders stakeholders stakeholders stakeholders
involved, etc. involved, etc. involved, etc. involved, etc. involved, etc. involved, etc. involved, etc.
Structure Stakeholder Necessary Necessary Supported Supported Supported Necessary Supported by a

of problem- participation few variants
solving process
Problems Via the Via the Via the Sometimes Via the Via the Model based
structuring construction construction construction cumbersome construction construction
of utility of suitable of the of the Regime of the
functions hierarchies evaluation vector evaluation
matrix matrix
Tool for Appropriate Appropriate Not Appropriate Interactive
learning appropriate MOP
Transparency High High High Medium Medium High Low
Actors’ Via the Via the Via the Via the Via the Via the Via the
communi- integration of integration of integration of integration of integration of integration of integration of
cation stakeholders stakeholders stakeholders stakeholders stakeholders stakeholders stakeholders
Applicability for problem structure
Indicator Geographical Different can Different can Different can Different can Different can Different can Different can
characteris- scale be treated be treated be treated be treated be treated be treated be treated
tics
Micro–macro Possible Possible Possible Possible Possible Possible Possible
link
Societal/ Different issues Different issues Different issues Different issues Different issues Different issues Different issues
technical are possible are possible are possible are possible are possible are possible are possible
issues
Methods’ Possible Possible Possible Possible Possible Not possible Possible
combinations
Data Type of data Qualitative and Qualitative and Qualitative and Mainly Qualitative and Qualitative and Mainly
situation quantitative quantitative quantitative quantitative quantitative quantitative quantitative
possible possible possible possible possible
Risk/ Risky Via sensitivity Via ordinal Via thresholds Via stochastic Via sensitivity
uncertainties outcomes: analysis criteria only and fuzzy analysis and
probabilities, numbers fuzzy numbers
sensitivity
analysis
Data processing High Medium Low High Medium Medium Medium
amount
Non- Not possible Not possible Not possible Not possible Not possible Not possible Possible via
substitutability constraints or
lexicographic
ordering
concerning these aspects. NAIADE does not allow to assign weights explicitly but
is the only methodology which provides an explicit structure for stakeholder
participation. For MOP/GP there are tools for interactive decision-making as well
as for group decision-making. While the majority of applications did not make use
of them, these tools have been available for more than ten years now.
The methods also differ in the possibilities for involving more than one decision-
maker in the process. For MOP/GP there are tools for group decision-making,
but again, they have not been applied much. Some methods like AHP support
consideration of preferences of several decision-makers. In practice, each decision-
maker is asked individually. Concerning MAUT, if n persons are involved in the
decision, each person is asked for their single utility function for any attribute;
these single functions are then aggregated to n multi-attribute utility functions. If
they are in conflict with each other, group solutions are possible, but are not really
treated by the literature about MAUT. Concerning stakeholder information,
MAUT, AHP, and Evamix provide good transparency and allow to give weights
explicitly. In a very interesting way, NAIADE supports the involvement of more
than one person in the decision-making process, namely with reference to their
different interests, which allows for an explicit conflict analysis. In general, an
important aspect in group decision-making is how transparent the decision-making
process is, as this enhances possibilities of participation and goal-oriented discussion
within the group.
Evamix, ELECTRE III, and Regime provide only satisfactory performance
for complex situations, so they are not the first choice in this context. Non-
substitutability aspects could best be considered by using ELECTRE III or GP/
MOP. MAUT and Evamix have acceptable performance in general, however they
lack important aspects crucial for sustainability issues, namely the ability to account
for uncertainty and non-substitutability. NAIADE is a specialist to involve more
than one decision-maker, while GP/MOP and ELECTRE III allow dealing with
non-substitutability. Thus, a clear ranking of the methods concerning the different
quality criteria is impossible, as each has its own strengths and weaknesses in
different areas. The characterization shows, however, that depending on the specific
problem the choice of one method over the other is advisable and that some
methods are largely useless for the analysis of sustainability problems.
Conclusions
The goal of more sustainable development has been studied extensively during
the last fifteen years; but a key criticism has been vagueness as to recommendations
and lack of operational applications. The acknowledgement of the limits of growth
has led to the revision of the paradigms of economics regarding human and natural
resources. Besides well-known goals like efficiency and income distribution, many
other conflicting factors need to be included in the analysis. MCDA constitutes a
powerful tool that is able to take into account several concerns and to foster partici-
pation and learning processes among politicians and citizens. Thus, MCDA is
useful for decision-making processes in the context of sustainability (e.g. Munda
1995; Castells and Munda 1999; Janssen 1992; Andreoli and Tellarini 2000;
Mendoza and Prabhu 2000).
In contrast to earlier studies which compare MCDA methods with respect to
more technical issues (e.g. Olson 2001, Salminen et al. 1998) we used a list of
quality criteria for the assessment (De Montis et al. 2000) and concentrated on the
most relevant criteria for sustainability issues. The comparison shows that the
indicator list is applicable and useful for the evaluation of MCDA methods in the
context of sustainability. As a result, a detailed characterization of the different
MCDA methods was derived. Some rough guidelines which can be given from the
comparison of the methods are:
1 If the respective decision problem is such that relying upon social welfare
theory and its assumptions is possible and if the data to build utility functions
is available (risk and qualitative data are possible) then MAUT is a good choice.
2 If working with different conflicting interest groups is important for the case,
NAIADE and AHP provide the best performance.
3 If the decision-makers involved should primarily learn from the application
of the MCDA tool, it is advisable to use MAUT or AHP.
4 If thresholds and constraints are central for the problem under investigation,
which means that there is non-substitutability of some criteria, ELECTRE
III or GP/MOP should be chosen.
5 If the problem is a continuous one, i.e. there is no discrete number of alterna-
tives which comes out of the specific situation, GP or MOP should be chosen.
6 If a complete ranking of the given alternatives as a result of the analysis is
indispensable, MAUT, AHP, Evamix, or Regime should be applied.
Further research should focus on the relationship between sustainability and

evaluation with respect to the process of consensus building. Recent research studies
(Hewitt 1995) state that general principles of environmental action consist of envir-
onment management and of democratic decision-making. In this sense, the techni-
cal features are not that important; rather, each multiple criteria technique should
be evaluated according to its capacity to foster learning processes and to allow
users to become aware of society’s strategic role in environmental decision-making.
Notes
1 Risk and uncertainty are sometimes used interchangeably in the MCDA literature. As the
difference is very important in the context of sustainability, we would like to distinguish between
them. We use ‘risk’ if a probability distribution can be assigned for the unknown effects; otherwise
the term ‘uncertainy’ is used (Knight 1921).
2 Utility independence: ui(xi ) is unaffected by levels of the other attributes. Preferential indepen-
dence: the preferences for some attribute levels do not depend on the levels fixed for the other
attributes.
3 No individual will sense a difference between a cup of coffee with i milligrams of sugar and one
with i + 1 milligrams of sugar, but will express a preference between a cup with a lot of sugar
and one with none; this contradicts transitivity of indifferences (Vincke 1992).
4 ‘In general what we mean by being consistent is that when we have a basic amount of raw data,
all other data can be logically deduced from it. In doing pairwise comparison to relate n activities
so that each one is represented in the data at least once, we need n - 1 pairwise comparison
judgements. From them all other judgements can be deduced simply by using the following kind
of relation: if activity A1 is 3 times more dominant than activity A2 and activity A1 is 6 times
more dominant than activity A3 then A1 = 3A2 and A1 = 6A3. It should follow that 3A2 = 6A3 or A2
= 2A3 or A3 = ½A2. If the numerical value of the judgement in the (2, 3) position is different
from 2 then the matrix would be inconsistent. This happens frequently and it is not a disaster’
(Saaty 1980: 18).
5 For more information on the ‘random index’ (R.I.) see Saaty (1980: § 1.4). In the following table
values of the random index for matrix of orders from 1 to 11 are reported:
Order of the 1 2 3 4 5 6 7 8 9 10 11
Random index 0.00 0.00 0.58 0.90 1.12 1.24 1.32 1.41 1.45 1.49 1.51
6 Elements of qualitative criteria are expressed by means of symbols like +++, ++, +, 0, –, – –,
– – –, where . … + + + ++ + 0 - -- - - - ….
7 An example of the way to obtain cardinal weights from ordinal preferences is given by Voogd
(1983: 175): ‘if w1 £ w2 £ w3 and it is assumed that weights meet the condition
m
Â wj =1
j =1
then the following extreme sets w , = 1, 2, 3 can be distinguished:
Ê1 1 ˆ Ê 1 1 1ˆ
w1 = (1, 0, 0), w2 = ÁË , , 0˜¯ , w3 = ÁË , , ˜¯ ' .
2 2 3 3 3
8 For more information about the mathematical aspects of Evamix, see Voogd (1983: ch. 10).
9 Using stochastic numbers, it is necessary to choose the probability density functions; using fuzzy
numbers, it is necessary to define their membership functions; using qualitative evaluations,
some pre-defined linguistic variables such as ‘good’, ‘moderate’, ‘bad’ and so on are considerate.
It is not possible to assign different types of measurements (i.e. crisp, stochastic, fuzzy, linguistic)
to the same criterion for different alternatives.
10 In the case of two crisp numbers, the ‘distance’ is defined as their difference; in the case of
stochastic and fuzzy numbers, the ‘semantic distance’ measures the distance between two functions
(i.e. probability density functions or fuzzy membership functions); the linguistic variables are
treated as fuzzy sets.
11 For linguistic variables,the equivalent preference relations may be used : much better than (>>);
better than (>), approximately equal to (@); very equal to (=); worse than (<); much worse than
(<<). In this case the six associated analytic functions express an index of credibility of the
statements that an alternative is ‘much better’, ‘better’, ‘approximately equal’, ‘very equal’,
‘worse’ and ‘much worse’ than another.
12 For more information about the mathematical aspects of NAIADE, see Munda (1995: ch. 7).
13 The symbol * stands for >>, >, @, < or <<.
14 ‘Entropy measure of fuzziness can be easily extended to the notion of fuzzy relations … to have
information on the diversity among the assessments of the single fuzzy relations, according to
each criterion’ (Munda 1995: 138). More precisely ‘entropy is calculated as an index varying
from 0 to 1 that gives an indication of the variance of the credibility indexes that are above the
threshold, and around the crossover value 0.5 (maximum fuzziness). An entropy value of 0
means that all criteria give an exact indication (either definitely credible or definitely non-credible),
whereas an entropy value of 1 means that all criteria give an indication biased by the maximum
fuzziness (0.5)’ (Menegolo and Pereira 1996: 5). For the mathematic formulae of entropy, see
Munda (1995: § 7.4.)
References
Andreoli, M. and Tellarini, V. (2000) ‘Farm sustainability evaluation: methodology and
practice’, Agriculture, Ecosystems and Environment, 77 (1–2): 43–52.
Ballestero, E. and Romero, C. (1998) Multiple Criteria Decision Making and its Applications to
Economic Problems, Boston/Dordrecht/London: Kluwer Academic.
Castells, N. and Munda, G. (1999) ‘International environmental issues: towards a new
integrated assessment approach’, in M. O’Connor and C.L. Spash (eds) Valuation and
the Environment – Theory, Method and Practice, Cheltenham, UK/Northampton, MA:
Chang, N.-B., Wen, C.G. and Chen, Y.L. (1997) ‘A fuzzy multi-objective programming
approach for optimal management of the reservoir watershed’, European Journal of
Operational Research, 99: 289–302.
Churchman, C.W., Ackoff, R.L. and Arnoff, E.L. (1957) Introduction to Operations Research,
New York: Wiley.
Climaco, J. (1997) Multiple Criteria Analysis. Proceedings of the XIth International Conference
on MCDM, Coimbra, Portugal: Springer, 1–6 August 1994.
De Montis, A., De Toro, P., Droste-Franke, B., Omann, I. and Stagl, S. (2000) ‘Criteria for
quality assessment of MCDA methods’, paper presented at the 3rd Biennial Conference
of the European Society for Ecological Economics, Vienna, 3–6 May 2000.
Dillon, J.L. and Perry, C. (1977) ‘Multiattribute utility theory, multiple objectives and
uncertainty in ex ante project evaluation’, Review of Marketing and Agricultural Economics,
45 (1, 2): 3–27.
Fandel, G. and Spronk, J. (eds) (1985) Multiple Criteria Decision Methods and Applications. Selected
readings of the First International Summer School Acireale, Sicily, September 1983,
Berlin: Springer.
Fishburn, P.C. (1968) ‘Utility theory’, Management Science, 14: 335–78.
Fishburn, P.C. (1970) Utility Theory for Decision Making, New York: John Wiley and Sons.
Fishburn, P.C. (1978) ‘A survey of multiattribute/multiple criteria evaluation theories’, in
S. Zionts (ed.) Multiple Criteria Problem Solving, Berlin: Springer: 181–224.
Funtowicz, S.O. and Ravetz, J. (1990) Uncertainty and Quality in Science for Policy, Dordrecht:
Kluwer.
Hewitt, N. (1995) European Local Agenda 21 Planning Guide – How to Engage in Long Term Environ-
mental Action Planning Towards Sustainability? Brussels: European Sustainable Cities and
Towns Campaign.
Hinloopen, E. (1985) De Regime Methode, MA thesis, Interfaculty Actuariat and Econometrics,
Free University Amsterdam.
Hinloopen, E. and Nijkamp, P. (1990) ‘Qualitative multiple criteria choice analysis’, Quality
and Quantity, 24: 37–56.
Hinloopen, E., Nijkamp, P. and Rietveld, P. (1983) ‘Qualitative discrete multiple criteria
choice models in regional planning’, Regional Science and Urban Economics, 13: 73–102.
Hwang, C. and Yoon, K. (1981) Multiple Attribute Decision Making, Berlin: Springer.
Ignizio, J.P. (1976) Goal Programming and Extensions, Lexington: Lexington Books.
Janssen, R. (1992) Multiobjective Decision Support for Environmental Management, Dortrecht:
Kluwer Academic.
Kalu, T.C.U. (1999) ‘An algorithm for systems welfare interactive goal programming
modelling’, European Journal of Operational Research, 116: 508–29.
Keeney, R.L. and Raiffa, H. (1976) Decisions with Multiple Objectives, New York: John Wiley.
Kmietowicz, Z.W. and Pearman, A.D. (1981) Decision Theory and Incomplete Knowledge,
Aldershot: Gower.
Knight, F. (1921) Risk, Uncertainty, and Profit, Boston: Houghton Mifflin.
Korhonen, P. (2001) ‘Multiple objective programming support’, in C.A. Floudas and P.M.
Pardalos (eds) Encyclopedia of Optimization, Dortrecht: Kluwer: Vol. 3, 566–74.
Lara, P. and Romero, C. (1992) ‘An interactive multigoal programming for determining
livestock rations: an application to dairy cows in Andalusia (Spain)’, Journal of the
Operational Research Society, 43: 945–53.
Lee, S.M. (1972) Goal Programming for Decision Analysis, Philadelphia, PA: Auerbach.
Luce, R.D. (1956) ‘Semiorders and a theory of utility discrimination’, Econometrica, 24:
178–91.
Mardle, S.J., Pascoe, S. and Tamiz, M. (2000) ‘An investigation of genetic algorithms for
the optimization of multi-objective fisheries bioeconomic models’, International
Transactions in Operational Research, 7(1): 33–49.
Mastenbroek, P. and Paelinck, J.H.P. (1977) ‘Qualitative multiple criteria analysis –
applications to airport location’, Environment and Planning A, 9 (8): 883–95.
Mendoza, G.A. and Prabhu, R. (2000) ‘Multiple criteria decision making approaches to
assessing forest sustainability using criteria and indicators: a case study’, Forest Ecology
and Management, 131(1–3): 107–26.
Menegolo, L. and Guimarães Pereira, A. (1996) NAIADE Manual, Joint Research Centre –
EC, Institute for System, Informatics and safety, Ispra (VA), Italy.
Munda, G. (1995) Multiple Criteria Evaluation in a Fuzzy Environment – Theory and Applications
in Ecological Economics, Heidelberg: Physika Verlag.
Munda, G., Nijkamp, P. and Rietveld, P. (1994) ‘Fuzzy multigroup conflict resolution for
environmental management”, in J. Weiss (ed.) The Economics of Project Appraisal and the
Environment, Aldershot: Edward Elgar.
Nijkamp, P. and Voogd, H. (1981) ‘New multiple criteria methods for physical planning by
means of multidimensional scaling techniques’, in Y. Haimes and J. Kindler (eds) Water
and Related Land Resource System, Oxford: Pergamon Press: 19–30.
Olson, D.L. (2001) ‘Comparison of three multicriteria methods to predict known outcomes’,
European Journal of Operational Research, 130(3): 576–87.
Rietveld, P. (1980) Multi Objective Decision Making and Regional Planning, Amsterdam: North
Holland.
Roy, B. (1985) Méthodologie Multicritère d’Aide à la Décision, Parigi: Economica.
Roy, B. (1996) Multiple Criteria Methodology for Decision Aiding, Dordrecht: Kluwer Academic.
Roy, B. and Bouyssou, D. (1985) ‘An example of comparison of two decision-aid models’,
in G. Fandel and J. Spronk (eds) Multiple Criteria-Decision Methods and Applications. Selected
readings of the First International Summer School Acireale, Sicily, September 1983,
Berlin: Springer: 361–81.
Saaty, T.L. (1980) The Analytic Hierarchy Process for Decision in a Complex World, Pittsburgh, PA:
RWS Publications.
Saaty, T.L (1988) Decision Making for Leaders, Pittsburgh, PA: RWS Publications.
Saaty, T.L. (1992) Multicriteria Decision Making – The Analytic Hierarchy Process, Pittsburgh, PA:
RWS Publications.
Saaty, T.L. (1994) Fundamentals of Decision Making and Priority Theory with the Analytical Hierarchy
Process, Pittsburgh, PA: RWS Publications.
Saaty, T.L. and Alexander, J.M. (1989) Conflict Resolution – The Analytic Hierarchy Process,
New York: Praeger.
Saaty, T.L. and Forman, E. (1993) The Hierarchon, Pittsburgh, PA: RWS Publications.
Saaty, T.L. and Vargas, L.G. (1991) Prediction, Projection and Forecasting, Dordrecht: Kluwer
Academic Publishers.
Salminen, P., Hokkanen, J. and Lahdelma, R. (1998) ‘Comparing multicriteria methods
in the context of environmental problems’, European Journal of Operational Research, 104:
485–96.
Scarelli, A. (1997) Modelli matematici nell’analisi multicriterio, Viterbo: Edizioni Sette Città.
Schniederjans, M.J. (1995) Goal Programming Methodology and Applications, Boston: Kluwer
Publishers.
Simon, H. (1955) ‘A behavioral model of rational choice’, Quarterly Journal of Economics, 69:
99–118.
Simon, H. (1979) ‘Rational decision making in business organizations’, American Economic
Review, 69: 493–513.
Spash, C.L. (2000) ‘Multiple value expression in contingent valuation: economics and
ethics’, Environmental Science & Technology, 34(8): 1433–8.
Steuer, R.E. and Gardiner, L.R. (1990) ‘Interactive multiple objective programming:
concepts, current status, and future directions’, in C.A. Bana e Costa (ed.) Readings in
Multiple Criteria Decision Aid, Berlin/Heidelberg/New York: Springer: 413–44.
Taha, H.A. (1976) Operations Research, New York: Macmillan.
Tamiz, M. and Jones, D. (1997) ‘An example of good modelling practice in goal pro-
gramming: means to overcoming incommensurability’, in R. Caballero, F. Ruiz and
R.E. Steuer (eds) Advances in Multiple Objective and Goal Programming, Berlin/Heidelberg/
New York: Springer: 29–37.
Tamiz, M., Jones, D. and Romero, C. (1998) ‘Goal programming for decision making: an
overview of the current state-of-the-art’, European Journal of Operational Research, 111:
569–81.
Van Delft, A. and Nijkamp, P. (1977) Multiple Criteria Analysis and Regional Decision Making,
The Hague/Boston: Martinus Nijhoff.
Vincke, P. (1985) ‘Multiattribute utility theory as a basic approach’, in G. Fandel and J.
Spronk (eds) Multiple Criteria-Decision Methods and Applications. Selected readings of the
First International Summer School Acireale, Sicily, September 1983, Berlin: Springer:
27–40.
Vincke, P. (1992) Multiple Criteria Decision-Aid, New York: John Wiley & Sons.
von Neumann, J. and Morgenstern, O. (1947) Theory of Games and Economic Behaviour,
Princeton, NJ: Princeton University Press.
Voogd, H. (1981) ‘Multicriteria analysis with mixed qualitative–quantitative data’, Delft
University of Technology, Department of Urban and Regional Planning, Planologisch
Memorandom, 81: 86.
Voogd, H. (1983) Multiple Criteria Evaluation for Urban and Regional Planning, London: Pion.
Wierzbicki, A. (1980) ‘The use of reference objectives in multiobjective optimization’, in
G. Fandel and T. Gal (eds) Multiple Objective Decision Making, Theory and Applications, New
York: Springer.
Zionts, S. and Wallenius, J. (1976) ‘An interactive programming method for solving the
multiple objective programming problem’, Management Science, 22(6): 652–63.
134 Wendy Proctor
6 MCDA and stakeholder

participation
Valuing forest resources
Wendy Proctor
Introduction
Making decisions about the environment often involves balancing the conflicting,
incommensurate and incompatible values of many users and uses of a resource.
One of the most fundamental and difficult tasks involved, therefore, is the effective
integration or synthesis of all values related to the resource issue in question.
Consideration and effective integration of all resource values, whether they are
environmental, economic or social, is a necessary first step to achieving and main-
taining ecologically sustainable development.
In 1992, after decades of environmentalist versus forest industry conflict, the
Australian Government embarked on the largest and most expensive environmental
planning exercise ever undertaken in Australia. This was the programme of Com-
prehensive Regional Assessment (CRA) of Australia’s forests that would result in
the signing of Regional Forest Agreements (RFAs) and ensure the sustainable
management of forests in Australia. One of the most difficult tasks for planners
and policy-makers was that of integrating all of the different forest values upon
which a final decision about reserved and unreserved areas could be made. Without
an adequate process of integration, important information could have been lost in
the volumes of complex assessment documents that were created. As well, the
process would fail to adequately take account of the preferences of all stakeholders.
This chapter outlines a practical application of the decision support tool multiple
criteria decision analysis, also known as multiple criteria analysis (MCA). The
case study concerns the integration of various forest values using a stakeholder
forum from one of the assessment regions. In recent years, MCA, as a structured
procedure for aiding complex environmental, social and economic decisions, has
gained popularity. However, MCA has had limited application in environmental
policy-making in Australia. Although not part of the official process, the case study
utilizes the results of the assessments themselves and the decision-making criteria
and priorities placed on different forest values provided by a representative stake-
holder group that was part of the CRA of the Southern Forest Region of New
South Wales, Australia.1
MCDA and stakeholder participation 135
Forest assessments
Of the 157 million hectares of forest in Australia, around 20 per cent comprise
rainforest and open eucalypt forests, both of which are major sources of timber
production. Around 30 per cent of this rainforest and open eucalypt forest area is
situated on state government timber reserves and 18 per cent in nature conservation
reserves. The day-to-day management of these areas rests largely under the control
of the relevant state government with the Commonwealth (i.e. Federal) Government
having jurisdiction over the issuing of woodchip export licences and obligations to
ensure sustainable forest management under various international protocols, such
as the Montreal Process (Commonwealth of Australia 1997a). However, the
management of Australia’s public forests has been characterized by intense public
debate for several decades. As a result, Australian governments are now carrying
out a series of CRAs of these forests that are designed to determine the major part
of Australian forest policy for the next twenty years.
In 1992, the Commonwealth and State Governments of Australia reached
agreement on a National Forest Policy that would provide a long-term management
strategy for Australia’s forests (Commonwealth of Australia 1992). As part of this
strategy, the governments agreed to: undertake CRAs of environment, heritage,
social and economic values of forests; set up Comprehensive, Adequate and
Representative (CAR) conservation reserves on the basis of these assessments; sign
RFAs about the long-term management and use of Australia’s forests
(Commonwealth of Australia 1997b). The CAR Reserve system is intended to
safeguard biodiversity, old growth, wilderness and other natural and cultural values
of forests. Forests outside of these reserves will be available for timber or other
types of suitable ecologically sustainable production. The conservation criteria set
by a panel of experts and agreed to nationally by governments (known as the
JANIS criteria) aim to reserve where practicable: 15 per cent of the estimated
extent of each forest ecosystem prior to European arrival; at least 60 per cent of
old growth forest; and 90 per cent or more of high quality wilderness. Other
criteria have been similarly set for flora and fauna species. The assessments have
resulted in volumes of documents for each region that need to be considered
carefully before the drawing up of a Regional Forest Agreement between the
Commonwealth government and the relevant state government. The RFAs will
be maintained for a period of twenty years. The long-term nature of these agree-
ments has been designed to provide a secure basis on which the timber industry
can make long-term investment decisions. These agreements are, however, to be
reviewed every five years, to ensure that the objectives are being satisfactorily met.
In 1994, the first of a series of extensive assessments of 25 million hectares of
forests began. Relevant government departments provide assessments on
environment and heritage, National Estate, wilderness, resource, economic and
social values of the forests under review. An important part of the decision-making
phase leading up to these agreements is the ‘integration’ process. That is, the process
that ideally allows a rigorous comparison of economic, environmental and other
forest values in order to allow decisions to be made about the extent, in these
forests, of reserved, logged and other use areas. The following is an overview of
136 Wendy Proctor
how these assessments are undertaken and how the official ‘integration’ process
takes place (for more information, see for example Commonwealth of Australia
1998).
The ‘Environment and Heritage Assessment’ covers biodiversity, endangered
species, old growth and world heritage areas. The biodiversity assessment includes
a compilation of flora and fauna data and an assessment of species and forest
ecosystems. The species assessment includes information on distribution, habitat,
life history and risk of extinction factors for each terrestrial and aquatic forest
flora and fauna species including those that are rare and threatened. The forest
ecosystems assessment attempts to estimate pre-1750 distributions of forest
ecosystems and compare these with present-day distributions. The old growth
assessment lists and maps areas classified under certain scientific principles. The
framework for the world heritage assessment is provided by a panel of experts
who identify themes of outstanding universal value relevant to Australia and related
to Australian forests. The analysis then seeks to identify if any of these themes can
be applied to the forest region under review.
National Estate Assessment uses a broad range of technical expertise and public
input. The assessment seeks to identify areas of both natural and cultural value. It
draws on a wide range of existing flora, fauna, historical, ecological and pre-logging
surveys as well as some new data collected especially for the purpose. Also included
is an ongoing assessment of indigenous cultural and heritage values.
The Wilderness Assessment has been largely based on either existing surveys or
on expert opinion or a combination of the two. Data on variation in wilderness
quality across the natural areas of the region are mapped. Measures of wilderness
quality are essentially ‘remoteness’ from access, structures and settlement and
‘naturalness’, both apparent and biophysical. Other guidelines agreed upon by
technical experts (such as a minimum of 25,000 hectares for an area to be classified
as wilderness) are also used.
Resource and Economic Assessment details the existing resources and uses of
the forest area and provides an economic evaluation of the current situation and
potential future scenarios. The usage categories may include: native forests
(production of sawlogs and residual logs), plantations (native and exotic species),
other forest produce (e.g. honey), recreation and tourism, water and mineral
resources. The net value of the timber industry is assessed using survey data
collected from existing sawmillers. Future returns to the industry are estimated
using a GAMS (Generalized Algebraic Modelling System) based constrained linear
optimization model of the local timber industry called FORUM (Forest Resource
Use Model), and estimates of sustainable yields provided by independent forecasting
models of forest growth and production for the particular region (Dann and Clarke
1996). Data on the values of other forest uses is provided by existing surveys. The
potential for mineral resources is assessed using standard geological survey tech-
niques that place varying probabilities on the existence of valuable minerals
according to the basic geological structure of the area.
Social Assessment uses survey, interview and workshop techniques. This analysis
provides a socio-demographic profile of the CRA area as well as details of the
current community infrastructure and an outline of community attitudes and
perceptions with regard to the use of forest resources. Various techniques based
on social assessment theory are employed in the analyses of the data.
‘Integration’ describes the official government process by which the various
assessments are brought together and used as the basis of coming to a decision
about reserved and other use areas in a region. The integration process differs
from state to state depending on the preferences of particular governments and on
the availability of data. In general, though, after the assessments have been
completed, a series of meetings are held and attended by representatives from the
Commonwealth, the state government and various stakeholder groups. The
objective of these meetings is to develop feasible scenarios, and subsequently, a
negotiated resource use option or set of options for the region, that is released for
public comment before a final decision is made.
A Geographic Information System (GIS) based reserve selection model called
‘C Plan’ is used extensively as the basis for developing resource use options. The
C Plan model maps areas of differing ‘irreplaceability’ and finds the smallest
possible set of areas that will achieve a nominated conservation goal (Pressy et al.
1995). Irreplaceability measures the extent to which a particular conservation
goal can be achieved. Two definitions of irreplaceability are used: one is the
likelihood that any of the areas in a region will be needed to achieve an explicit
conservation goal, the other is the extent to which the options for achieving an
explicit conservation goal are narrowed if any of the areas in a region are
destroyed or made unavailable for conservation. The conservation goals can
include the preservation of such items as particular forest ecosystems, species of
flora and fauna, etc. Also, the model maps areas of old growth forest, wilderness,
potential mineral deposits, high cultural and heritage values, timber production,
private and leasehold land as well as areas which have been targeted for indigenous
land rights claims.
A process of visual interpretation of the irreplaceability indices is followed and
each area of high irreplaceability investigated and selected for inclusion as a possible
reserve area. A visual interpretation of the areas of wilderness, old growth forests
and areas where threatened species are reported to exist is also carried out and
these areas included in the possible reserve system. Reserve design is also taken
into account. Despite the technical and structured nature of the C Plan method,
the reserve design process also includes a degree of ad hoc selection. For example,
this is displayed in the choice of criteria being used to select areas for reserve and
sometimes on the heavy reliance on local knowledge of conservation groups and
government officers. After several selections have been made, the extent to which
conservation targets have been met for the various flora and fauna species and
ecosystems are assessed. Next, a timber volumes model, the Forest and Resource
Management System (FRAMES) is run. This model estimates the amount of
commercial timber that could be harvested from the remaining area of forest that
had been excluded from reserve and is not on private or leasehold land (Turner
1998). The results are then fed into the FORUM economic model to determine
the likely consequences on the value of timber production in the area as a result of
138 Wendy Proctor
the change in reserve selection. FORUM provides an analysis of the commercial
value of the forest harvest, net returns to the regional industry and employment.
Further details on ‘integration’ can be found in Proctor (1999), however, it should
be noted that at no stage in the official integration process for the Southern New
South Wales (NSW) Region were the data obtained from the various assessments
comprehensively and methodically incorporated into the decision-making. It should
also be noted that, unlike previous assessments, for this region, the stakeholder
groups were largely excluded from the integration process. The Southern Region
integration process resulted in the description of five options, along with various
data associated with each option, being released for public comment. These options
represented five different reserve designs (with implications for economic indicators
as well as various levels of conservation and National Estate targets being met)
associated with levels of high quality sawlog production of 32,000 cubic metres
per annum, 35,000 cubic metres per annum, 45,000 cubic metres per annum,
55,000 cubic metres per annum and 65,000 cubic metres per annum. Details of
each option are provided later.
The following sections outline an MCA to utilize the forest assessment data in
a much more rigorous, objective and structured way than has been undertaken in
the RFA integration process. This method also allows for a much more participatory
approach to the decision-making process than has so far occurred for any of the
forest agreements. The guidelines for the CRA process (see Commonwealth of
Australia 1992) outline the need for public participation covering all interest groups
and accounting for all stakeholder preferences in a comprehensive and transparent
manner. In addition, stakeholder participation is a fundamental principle of
ecologically sustainable development. It will be shown that MCA is an effective
means of achieving stakeholder participation in the decision-making process.
Method
MCA is a means of simplifying complex decision-making tasks which may involve
many decision-makers, a diversity of possible outcomes and many and sometimes
intangible criteria by which to assess the outcomes. It is one means by which
structure and transparency can be imposed upon the decision-making process. Its
origins lie in the fields of public policy, mathematics and operations research and
it has had a great deal of practical usage by public planners in such areas as the
location of health facilities, motorways and nuclear reactors (Massam 1988).
An MCA seeks to make explicit the logical thought process that is implicitly
carried out by an individual when coming to a decision. In complex decision-
making tasks, which sometimes involve many objectives and many decision-
makers, this structured process of logic may be lost in the complexity of the
issues. In general, an MCA seeks to identify the alternatives or options that are
to be investigated and decided upon, a set of criteria by which to rank these
alternatives and the method by which the alternatives are to be ranked and
preferences aggregated. Finally, a sensitivity analysis is carried out on the results.
The final outcome is a preferred option or set of options that is based upon a
rigorous definition of priorities and preferences decided upon by the decision-
maker.2
In general, the following steps are carried out by the decision-maker in an MCA.
The first is to identify the feasible options or preferred outcomes involved in the
decision problem. Next the decision-maker seeks to identify the overall objective
that is to be achieved in the process and then identifies the criteria by which to
judge the selected options. An important part of the process is then to apply
appropriate weights on each of the criteria that reflect the particular preferences
of the decision-maker in how important each criterion is in relation to the overall
objective of the decision process.
The analyst then chooses an appropriate form of MCA with which to analyse
the data and aggregate the preferences. An assessment is then made of each of the
options based on the chosen MCA technique, the data on how each option performs
under the different criteria and the weightings supplied by the decision-maker. An
investigation of the sensitivity of the ranking of options with respect to, for example,
the chosen weights and method of aggregation can then be carried out. An iterative
process can then be pursued where the analyst reports back to the decision-maker
the findings of the MCA and the sensitivities. ‘Fine tuning’ can then be performed
by the decision-maker to aspects related to their input. Also this part allows for the
identification of trade-offs, if any, that may be made in the overall decision process.
A common criticism in public policy-making and in particular environmental
policy-making in Australia is the high incidence of ad hoc decision-making that
takes place under such processes (see, for example, Dovers 2000; Doyle and Kellow
1995; Ferguson 1996). The need for a more structured and objective approach has
been addressed in the public policy literature, with some authors providing general
guidelines on how this might be achieved (for example Hogwood and Gunn 1984).
Similarly, an important aspect of environmental policy-making in recent years has
been the increasing practice of incorporating public and stakeholder participation
into some part of the policy-making process. The need for a participatory frame-
work to environmental policy formulation in Australia has grown out of increasing
interest by individuals in environmental issues (Lothian 1994), a grass-roots desire
by the public to become involved in the policy-making process over regional issues
that affect them, and a recognition by governments that involving the general
community early in the process can bring benefits by avoiding disagreements and
conflicts in later stages and increasing the knowledge base that is available to policy-
makers (Howlett and Ramesh 1995).
Whereas much of the literature on both prescriptive methods for policy-making
and the incorporation of public participation into such policy-making has addressed
general requirements and concerns, the steps outlined under the MCA process
are more explicit and provide a more detailed and structured guide for practical
approaches to participatory and other public policy-making. In the context of this
research, MCA is primarily regarded as an aid in the process of decision-making
and not necessarily as a means of coming to a singular optimal solution. As such,
the MCA process is valued for the enlightenment and unravelling of issues that it
can provide in the decision-making problem. The process adds to the knowledge
140 Wendy Proctor
of the decision-maker and is greatly aided by including the decision-maker in
each step of the analysis. This is one reason why some forms of MCA are regarded
as superior to other decision-aiding techniques (for example, mathematical
programming methods) which are usually carried out in isolation from the decision-
maker. Such techniques are designed to provide an ‘optimal’ result at the end of
the analysis, and so they do not necessarily increase the understanding of the
important elements of the process, especially to those who are not familiar with
such models. Other techniques are not designed to unravel the decision-making
problem or provide a process that will enlighten the decision-maker and add to the
decision-maker’s knowledge of the problem. The non-transparency and singular
solution nature of such techniques may only result in increased mistrust in the
process by those who are not familiar with how a decision was finally derived
(Prato 1999). The approach followed in this current research is therefore similar to
the ‘heuristic’ approach emphasized by Stirling and Mayer (1999) in their ‘multi-
criteria mapping’ technique (see also Chapter 7 in this volume).
The analytic hierarchy process

There are many different types of MCA, as discussed in the previous chapter.
Saaty (1980) developed a method of analysing decisions based upon a hierarchy
of components of the decision, known as the Analytic Hierarchy Process (AHP).
His method is essentially an interactive one where a decision-maker or group of
decision-makers relay their preferences to the analyst and can debate or discuss
opinions and outcomes. His method largely stems from theories of human
behaviour including thought process, logic, intuition, experience and learning
theories (Saaty 1982). The hierarchies are made up of: a top level, which comprises
the overall objective of the decision process; intermediate levels, which comprise
the criteria and sub-criteria to analyse the decision; and the lowest level, which
lists the alternative plans that are to be analysed and decided between.
The AHP is based upon the construction of a series of ‘pairwise comparison’
matrices which compare criteria with one another. This is done to estimate a ranking
or weighting of each of the criteria that describes the importance of each of these
criteria in contributing to the overall objective. If the criteria are broken down
into a number of sub-criteria, the pairwise comparisons are repeated for each
level of the hierarchy. A pairwise comparison of J criteria (G1,…, GJ ) to reflect the
importance or weighting of each criteria in influencing the overall objective, involves
constructing a J by J matrix (G) which shows the dominance of the criteria in the
left-hand column with respect to each criteria in the top row.
A1 A1
a j1 a j1
Aj A1 a j1 a jI
a jI a jI
AI a j1 a jI
Each cell entry of G reflects a ratio scale of the underlying priority weights assigned
to each of the criteria, i.e. gjj ¢ = wj/wj ¢. The weightings (wj ) have to be derived from
the cell entries gjj ¢. A nine-point Intensity of Importance scale was developed by
Saaty to determine these. It is claimed that the scale is based on psychological
experiments and is designed to allow for an accurate reflection of priorities in
comparisons between two items whilst minimizing the difficulties involved in doing
so (Table 6.1).
In G, each cell entry is positive and the diagonal elements (gjj ) are a series of 1’s.
If it is assumed that transitivity of preferences prevails (i.e. that if G1 is preferred
by a scale of 5 to G2, then G2 is preferred by a scale of 1/5 to G1) then the reciprocal
property gjj’ = 1/gjj’ is satisfied and estimates need only be provided for those cells
which lie above the diagonal. Saaty proves that, if G displays ‘cardinal’ consistency,
in that gjj’ gj’j” = gjj”, then by normalizing the positive reciprocal matrix G so that the
columns sum to unity, a solution to w, the vector of overall priority weights, can be
obtained by reading any column of the matrix, as each column in this normalized
matrix will be identical (Saaty and Vargas 1982). Imposing cardinal consistency
on the matrix also means that only one row of the matrix needs to be entered and
all other values can be derived. If cardinal consistency were not imposed, then
each column vector may be different and it would then be necessary to average
across the rows to determine the overall priority weights. In this instance, Saaty
provides a ‘check for consistency’ measure and acceptable bounds within which
this measure should fall.
To gauge how each plan performs with respect to each of the criteria, another
series of pairwise comparisons are carried out. For j = 1, …, J criteria and i = 1,
…, I different plans, a pairwise comparison matrix comparing all plans under the
criteria Gj would be:
A1 A1
a j1 a j1
Aj A1 a j1 a jI
a jI a jI
AI a j1 a jI
Table 6.1 The AHP intensity of importance scale

Intensity of importance Definition
1 Equal importance of both elements
3 Weak importance of one element over another
5 Essential or strong importance of one element over another
7 Demonstrated importance of one element over another
9 Absolute importance of one element over another
Source: Saaty (1982)
142 Wendy Proctor
Again, the nine-point scale is used to provide estimates of the ratio weights, the aji,
and the positive reciprocal properties are assumed to hold. The overall priority
(OPi) of each of the plans with respect to all of the criteria can be estimated as
follows:
OP1 = a11(w1) + a21(w2) +…+ aJ1(wJ)

OP2 = a12(w1) + a22(w2) +…+ aJ2(wJ)
.
.
.
OPI = a1I(w1) + a2I(w2) +…+ aJI(wJ)
The following example clearly describes the above process. Assume that three
broad criteria groups (which could incorporate a number of sub-criteria) are
thought to be important in a decision concerning various natural resource use
options. These represent environmental, social and economic considerations. A
pairwise comparison for each of the criteria could then be as shown in Table 6.2.
The comparison of a criterion with itself always denotes equal importance and
so the central diagonal is a series of 1’s. In this example, environmental criteria are
considered to be weakly dominant over social criteria and so a 3 is recorded in the
appropriate cell. Its reciprocal is placed in the corresponding cell that compares
social with environment. This matrix is consistent in the sense that a value of 5/3
is recorded in the space when social criteria are compared with economic criteria.
This value was derived by imposing the cardinal consistency property on the matrix
although cases could exist where a different value is recorded here thus denoting
an inconsistency of preferences of the decision-maker. The matrix is then
normalized (by dividing each cell entry by its column total) so that meaningful
comparisons can be made amongst the components, and the relative priorities for
each criteria with respect to the overall objective can be read from any column.
This is shown in Table 6.3.
The vector of overall priorities shows the relative importance of each of the
three criteria in affecting the final decision. In this example, environmental criteria
are considered most important by a large margin and then follow social and finally
economic considerations. Using the criteria and priority weights estimated above,
three different plans are compared by first carrying out a pairwise comparison of
the performance of each plan against the others under the three different criteria
identified. An example is shown for the environmental criteria (Table 6.4). The
same procedure would then be carried out for the social and the economic criteria.
In this example, Plan a performs quite strongly under the social and economic
criteria and less so under the environmental criteria. Plan b performs poorly under
the environmental criteria and slightly better for the other two, whilst Plan c is
strongly preferred under environmental criteria but is least preferred when only
social or economic criteria are considered. To obtain an overall ranking of the
plans under all of the criteria combined, a weighted sum of each plan’s performance
under each of the criteria is calculated as in Table 6.5. From this analysis, Plan c is
Table 6.2 Example pairwise comparison

Environment Social Economic
Environment 1 3 5
Social 1/3 1 5/3
Economic 1/5 3/5 1
Table 6.3 Normalized matrix example

Environment Social Economic Priority weighting
Environment 0.65 0.65 0.65 0.65
Social 0.22 0.22 0.22 0.22
Economic 0.13 0.13 0.13 0.13
Table 6.4 Pairwise comparison of environmental criteria

Environmental Plan a Plan b Plan c Ranking (priority after matrix is normalized)
Plan a 1 3 1/2 0.30
Plan b 1/3 1 1/6 0.10
Plan c 2 6 1 0.60
Table 6.5 Overall ranking example

Environment Social Economic Overall priority
0.65 0.22 0.13
Plan a 0.30 (0.65) +0.50 (0.22) +0.48 (0.13) = 0.37
Plan b 0.10 (0.65) +0.25 (0.22) +0.32 (0.13) = 0.16
Plan c 0.60 (0.65) +0.25 (0.22) +0.19 (0.13) = 0.47
chosen as the preferred plan because of its strong performance with respect to the
environmental criteria and the associated weighting placed on the environmental
issues in achieving the overall objective.
Case study
The NSW Southern Region Regional Forest Forum was formed as part of the
RFA process. The Forum provided an opportunity for regional stakeholders to
share information with governments and offer recommendations. The Forum was
not, however, given an official decision-making role in the process. Members of
the Forum represented local state and federal government natural resource
management organizations, forest industry, conservation, farmer and beekeeping
lobby groups as well as academics and indigenous tribal elders and representatives.
The Forum met regularly for almost three years – around once every six weeks
between 1997 and 2000. Although not part of the ‘official’ process, the Forum
members agreed to take part in the MCA as part of their regular meetings, with
144 Wendy Proctor
the time allocated to collect information for the analysis forming only a small part.
Much of the information was gathered outside the meetings by mail survey, phone
or face-to-face interview.
The first stage of the analysis was to identify a complete set of criteria by which
to assess each alternative plan. The criteria identified relate directly to the stated
objectives of the RFAs as outlined in the National Forest Policy Statement. Figure
6.1 shows a hierarchy of objectives and criteria, developed in association with the
Forum members. Forum members were then asked to carry out the AHP pairwise
comparison exercise outlined previously. A survey asking for details about their
pairwise preferences of the identified criteria and sub-criteria was distributed and
information on how to fill out the survey was provided at one of the meetings.
Follow-up phone calls also provided assistance. A pairwise comparison of the three
broad criteria (Environment, Economic and Social) was requested as well as three
sub-criteria group comparisons. Information was only required on the matrix cell
entries which were above the diagonal. Reciprocals were later entered below the
matrix diagonals. Cardinal consistency was not imposed on the comparisons. Time
and other constraints did not allow for further interaction with the Forum for
‘fine-tuning’ of their preference data.
Results
Southern Forest Forum preferences

The purpose of this part of the analysis is to identify issues of relative disagreement
and agreement on priorities between the 22 members of the Forum. The potential
of the analysis is that those issues of substantial disagreement can be isolated and
be subject to more in-depth analysis and discussion so that possible compromises
or trade-offs may be found.
Criteria consistently given low priority can be excluded from the analysis as
they are regarded as being of relatively less importance in meeting the overall
objectives. The raw data from each respondent’s matrices were normalized (by
dividing the cell entry by the column total) to allow comparisons between criteria
and to show the range of relative priorities. The Cardinal Consistency Index was
within the acceptable limits for all responses. Figure 6.2 shows the differences in
priorities amongst Forum members with respect to the three major criteria.
The largest differences lie in the Environment and Economic categories with a
distinct polarization of priorities evident amongst many Forum members. In
general, those members classed as having forest industry affiliation and conservation
group affiliation, show very divergent priorities as expected. There is some
consistency with regards to the Social criteria, with most respondents considering
this to be of least (or equal) importance out of the three. As an indication of the
overall group response, following the method recommended by Saaty (1982), the
geometric mean over all respondents of each matrix cell entry was calculated and
the resulting geometric mean matrix normalized. For the three broad criteria
categories, the Environmental category was given greatest priority in influencing
Protect conservation values
Ensure the long-term ecologically sustainable management of forests Objectives
Develop an internationally competitive forest products industry
Effectively use other regional economic and social resources
Conservation of Maintenance of Maintenance of social and Criteria

environmental values long-term economic benefits cultural values
Apiary
Timber
Minerals
Old growth
Indigenous
Wilderness
Biodiversity
National estate Indicators
Other product values

Cultural and heritage
Recreation and tourism
Water and soil resources

Adequate hazard reduction
Employment and community needs
Forest contribution to global carbon cycles

Productive capacity, health and vitality of forest
MCDA and stakeholder participation
Figure 6.1 Hierarchy of objectives, criteria and indicators for the Southern Region RFA
145
146 Wendy Proctor
0.90
0.80
Normalized priorities
0.70
0.60
0.50
0.40
0.30
0.20
0.10
0.00
Environment Economic Social
Figure 6.2 Preferences of criteria
the overall RFA objective with a figure of 0.41. Next followed the Economic criteria
(0.34) and finally the Social criteria (0.25).
The normalized priorities of the sub-criteria were then calculated. Within the
Environmental sub-criterion, the most important criteria were identified as Water
and Soil Resources, the Productive Capacity of the Forest and Biodiversity (Figure
6.3). The normalized priorities of the Economic criteria show less dispersion than
those of the Environmental criteria. Disparities in preferences exist in the Productive
Capacity, Timber Values and Minerals Values categories although the range of
priorities for Productive Capacity is not spread evenly with one substantial outlier
influencing the results (Figure 6.4). Disregarding this outlier, the highest priorities
were given to Timber Values and Employment and Community Needs.
In the category of Social Criteria, the highest priorities were given to
Employment and Community Needs and Indigenous Values. The largest range of
priorities for the social criteria (ignoring outliers) appeared to be in the Employment
and Community Needs category (Figure 6.5). More detailed analysis of the
preference structures of the Forum can be found in Proctor (1999) and Proctor
(2001).
Performance of options against the criteria

The next part of the analysis was to assess each of the suggested options devised
for the Southern Region against the decision criteria identified by the Forum
members. Table 6.6 shows an Impact Table (measuring the performance and
implications of each option) of the only available data released on each option at
the time of writing. The available data fell well short of those that were identified
by the Forum as necessary to the decision-making process. An indication of the
relationship between the available data and the decision criteria is provided in
Table 6.6. It is important to note that impact data on some of the criteria consistently
regarded as important by the Forum members (for example, Water and Soil
0.60
0.50
0.40
0.30
0.20
0.10
0.00
Carbon
Hazard
Prod cap
Biodiversity
Wilderness
Water and
Old growth
soil
Figure 6.3 Preferences of environmental criteria
0.60
0.50
0.40
0.30
0.20
0.10
0.00
Productive Timber Minerals Apiary Other Employment Recreation and
capacity values values values and community tourism
Figure 6.4 Preferences of economic criteria
Resources and Indigenous Values) were not provided for the different proposed
options.
The individual Forum members’ preference matrices of the sub-criteria were
re-normalized, after excluding those sub-criteria for which data were not available.
Table 6.7 shows the estimated pairwise comparisons of options under the
148 Wendy Proctor
0.70
0.60
0.50
0.40
0.30
0.20
0.10
0.00
Employment and Recreation and National Cultural and Indigenous
community tourism estate heritage
Figure 6.5 Preferences of social criteria
environmental criteria based on the performance of each option outlined in the

Impact Table.
Using the priorities placed on the environmental criteria by the Forum members
(i.e. Biodiversity 0.47, Old Growth 0.29 and Wilderness 0.23), Option 1 (with an
overall priority of 0.43 = (0.442)0.47 + (0.490)0.29 + (0.333)0.23) ranks highest in
terms of the environmental criteria. Then follow Option 2 (0.20), Option 3 (0.14),
Option 4 (0.12) and Option 5 (0.11) (Table 6.8). Table 6.9 shows the pairwise
comparison of options under the economic criteria.
Using the priorities placed on the economic criteria as shown in Table 6.10 (i.e.
Timber Values, 0.39; Employment and Community Needs, 0.36; and Other Values,
0.24), the ranking of the options according to their performance under the economic
criteria was: Option 5 (0.40), Option 4 (0.26), Option 3 (0.16), Option 2 (0.09),
Option 1 (0.08). Using the pairwise comparisons of options under the social and
cultural criteria (Table 6.11) and the social and cultural criteria priorities (i.e.
Employment and Community Needs, 0.56; National Estate Values, 0.44), the
ranking of the options under the social criteria was: Option 5 (0.26), Option 1
(0.22), Option 4 (0.19), Option 3 (0.17), Option 2 (0.16) (Table 6.12).
The overall ranking of the options under all criteria was: Option 1 (0.26), Option
5 (0.24), Option 4 (0.19), Option 3 (0.16), Option 2 (0.15). These rankings are
depicted in Figure 6.6 with the priorities of the broad criteria and the individual
Forum member’s overall priorities are presented in Figure 6.7 and Figure 6.8. An
option representing an ‘optimal group decision’ has been provided here for
illustrative purposes in Figure 6.7, following the method of averaging over the
group’s preferences suggested in the AHP. However, a more valuable and
informative part of the MCA is provided by a closer scrutiny of the results than is
revealed in a simple geometric averaging procedure. For example, the range of
individual overall priorities of options (Figure 6.8) suggests that the greatest disparity
of rankings lies in the two extreme options with much greater agreement by Forum
members in placing the centre options on a lower priority. The following sensitivity
Table 6.6 Impact table for forest use options
Indicator Option 1 Option 2 Option 3 Option 4 Option 5
Volume of sawlogs/year (cubic 32,000 35,000 45,000 55,000 65,000
metres) (for 20 years and (for 20 years and (for 20 years and (for 20 years and (for 20 years and
at least 22,000 per at least 35,000 per at least 43,000 per at least 45,000 per at least 45,000 per
year thereafter) year thereafter) year thereafter) year thereafter) year thereafter)
Achievable targets* met in dedicated
reserves (%): 1
Forest ecosystems 80 67 62 61 60
Old Growth Forest 73 55 52 52 52
Fauna 79 73 72 68 66
Flora 84 74 71 67 63
Wilderness reserved (%) 1 90 88.5 88.1 87.7 87.2
National Estate Areas Reserved All All All Some Some
(areas with National Estate sites
identified within them and placed
into reserve) 2
National Estate Values (additions to High High High Medium Low
reserves because of high potential
National Estate values identified)2
Total Direct Mill Employment (no.)2,3 140–145 144–151 172–180 195–200 213
Total Harv. and Haul. Employment 36 38 43 50 55
(no.) 2,3
Gross Value Output ($m)3 15.5–16.5 16.4–17.6 19.3–22.3 22.2–25.4 25.0–27.6
Change in other employment (no.) 2,3 –48 –33 +16 +69 +115
Total Gross Value Output ($m)3 3,327.6 3,329.1 3,333.8 3,338.6 3,343.2
Source: Forest Taskforce and Resource and Conservation Division 2000, ‘A Proposal for a Regional Forest Agreement for Southern NSW’.
* Achievable targets refers to the extent to which the JANIS criteria are met on formal or dedicated reserves such as national parks, nature reserves and flora reserves.
1 These data address the broad criteria under Conservation of Environmental Values – Biodiversity (achievable targets met for Flora, Fauna and Forest Ecosystems); Old
Growth Forests (achievable targets met); Wilderness (percentage reserved).
2 These data address the broad criteria under Maintenance of Social and Cultural Values – Employment and Community Needs (direct mill employment, harvesting and
MCDA and stakeholder participation
haulage employment, change in other employment); National Estate (National Estate areas reserved, National Estate values).
3 These data address the broad criteria under Maintenance of Long-Term Economic Benefits – Timber (gross value of timber output); Other Products (total gross value of
output) Employment and Community Needs (direct mill employment, harvesting and haulage employment, change in other employment).
149
150 Wendy Proctor
Table 6.7 Pairwise comparison of options under environmental criteria
Biodiversity* Option 1 Option 2 Option 3 Option 4 Option 5
Option 1 1 3 3 4 4
Option 2 0.33 1 2 2 3
Option 3 0.33 0.5 1 2 2
Option 4 0.25 0.5 0.5 1 2
Option 5 0.25 0.33 0.5 0.5 1
Wilderness Option 1 Option 2 Option 3 Option 4 Option 5

Option 1 1 2 2 2 2
Option 2 0.5 1 1 1 1
Option 3 0.5 1 1 1 1
Option 4 0.5 1 1 1 1
Option 5 0.5 1 1 1 1
Old growth Option 1 Option 2 Option 3 Option 4 Option 5

Option 1 1 4 4 4 4
Option 2 0.25 1 2 2 2
Option 3 0.25 0.5 1 1 1
Option 4 0.25 0.5 1 1 1
Option 5 0.25 0.5 1 1 1
* Biodiversity figures were estimated using an average over achievable targets met for forest
ecosystems, flora and fauna from Table 6.7.
Table 6.8 Priority of options in terms of environmental criteria

Ranking of options for each sub-criteria**
Biodiversity Old growth Wilderness Priority
0.47* 0.29* 0.23*
Option 1 0.442 0.490 0.333 0.426
Option 2 0.217 0.189 0.167 0.195
Option 3 0.153 0.107 0.167 0.141
Option 4 0.111 0.107 0.167 0.121
Option 5 0.077 0.107 0.167 0.106
* Contribution of sub-criteria to overall priority.
** After normalizing the pairwise comparison matrices.
analyses are also important parts of the MCA to help to unravel the complexities
of the decision-making problem.
Sensitivity analysis
A sensitivity analysis was carried out to test the outcome of the overall ranking of
options when the priorities placed on the broad Environmental and Economic
criteria were both placed at their maximum values in turn.3 The outcomes are
depicted in Figure 6.9 and Figure 6.10, with an environmental preference leading
to an even higher ranking of Option 1 and an economic preference leading to
Option 5 being ranked first.
Table 6.9 Pairwise comparison of options under economic criteria
Timber Option 1 Option 2 Option 3 Option 4 Option 5
employment*
Option 1 1 0.5 0.33 0.25 0.2
Option 2 2 1 0.33 0.25 0.2
Option 3 3 3 1 0.5 0.33
Option 4 4 4 2 1 0.5
Option 5 5 5 3 2 1
*Includes direct mill and harvesting and haulage employment.
Timber values* Option 1 Option 2 Option 3 Option 4 Option 5
Option 1 1 1 0.25 0.25 0.17
Option 2 1 1 0.33 0.25 0.2
Option 3 4 3 1 0.5 0.33
Option 4 4 4 2 1 0.5
Option 5 6 5 3 2 1
*Gross value of timber output.
Other values* Option 1 Option 2 Option 3 Option 4 Option 5
Option 1 1 1 1 0.5 0.5
Option 2 1 1 1 0.5 0.5
Option 3 1 1 1 0.5 0.5
Option 4 2 2 2 1 0.5
Option 5 2 2 2 2 1
*Total gross value of output.
Table 6.10 Priority of options in terms of economic criteria

Timber Value emp. & comm. Other values Priority
0.39* 0.36* 0.24*
Option 1 0.062 0.061 0.141 0.080
Option 2 0.067 0.082 0.141 0.090
Option 3 0.178 0.169 0.141 0.164
Option 4 0.263 0.267 0.249 0.259
Option 5 0.429 0.421 0.327 0.397
Sensitivity analysis was also carried out to assess the order of rankings under
different assumptions about the sub-criteria. As the Employment and Community
Needs criterion displayed the greatest difference in the priorities placed on it by
Forum members, two scenarios were tested. The first used the maximum value
obtained for this criterion from the individual pairwise comparisons of both the
Economic and the Social sub-criteria (with other relative priorities maintained).
The effect was to change the ranking to Option 5 (0.29), Option 1 (0.22), Option
4 (0.21), Option 3 (0.16) and Option 2 (0.13). The second scenario used the
minimum value obtained from the individual pairwise comparisons for the
Employment and Community Needs criterion. Minimizing this criterion’s
152 Wendy Proctor
Table 6.11 Pairwise comparison of options under social criteria
National Estate* Option 1 Option 2 Option 3 Option 4 Option 5
Option 1 1 2 3 4 5
Option 2 0.5 1 2 3 4
Option 3 0.33 0.5 1 2 3
Option 4 0.25 0.33 0.5 1 2
Option 5 0.2 0.25 0.33 0.5 1
*The National Estate criterion has been given an arbitrary weighting because only
qualitative data were available.
Emp. & comm. Op ion 1 Option 2 Option 3 Option 4 Option 5
Option 1 1 1 0.25 0.2 0.17
Option 2 1 1 0.25 0.2 0.17
Option 3 4 4 1 0.25 0.25
Option 4 5 5 4 1 0.5
Option 5 6 6 4 2 1
Table 6.12 Priority of options in terms of social criteria

Emp. & comm. National estate Priority
0.56* 0.44*
Option 1 0.061 0.416 0.217
Option 2 0.082 0.262 0.161
Option 3 0.169 0.161 0.165
Option 4 0.267 0.098 0.193
Option 5 0.421 0.062 0.263
preference weighting only strengthened the lead of Option 1 further and only
slightly changed the lower rankings with an ordering of Option 1 (0.30), Option 5
(0.21), Option 2 (0.17), Option 4 (0.17), Option 3 (0.15).
A final sensitivity analysis was carried out on the relatively arbitrary values of
the National Estate impacts for the different options. Changing these values from
very small differences in impacts to very large differences in impacts had the effect
on the final rankings of swapping the placements of Options 1 and 5 between first
and second in each case. For small differences in National Estate impacts, the final
ranking was Option 5 (0.25), Option 1 (0.24), Option 4 (0.19), Option 3 (0.17),
Option 2 (0.15). For large differences in National Estate impacts, the final ranking
was Option 1 (0.27), Option 5 (0.24), Option 4 (0.18), Option 3 (0.15), Option 2
(0.15).
Conclusion
The outcome of the MCA based on the data released for the Southern Region
reveals that the preferred option is Option 1. Contrary to what may be thought
0.45
0.40
0.35
0.30
Priority weight
Environmental
0.25
Economic
0.20
Social
0.15
0.10
0.05
0.00
1 2 3 4 5
Option
Figure 6.6 Overall ranking of options
0.30
0.25
Priority weight
0.20
0.15
0.10
0.05
0.00
1 2 3 4 5
Option
Figure 6.7 Priorities of broad criteria
from initial observations of the data, however, is that the next preferred option is
Option 5. On closer inspection of the data, this outcome is seen to be the result of
the high performance of Option 1 under the environmental criteria when compared
with all the other options and, similarly, the much better performance of Option 5
under the employment criteria when compared with all the other options. This
result accurately depicts the polarized nature of the priorities of the Forum members
and shows the nature of the trade-off in the forest management debate as being
one between conservation and employment. The outcome of the analysis shows
154 Wendy Proctor
0.5
Max
Priority weight
0.25 Min
Av
0
1 2 3 4 5
Option
Figure 6.8 Range of individual priorities
0.40
0.30
Priority
0.20
0.10
0.00
1 2 3 4 5
Option
Figure 6.9 Environmental preference
that a ‘middle ground’ choice, such as Option 2, 3 or 4 may not be accepted

favourably by any stakeholders because none of these options perform strongly
enough under either the conservation or the employment criteria. This result is
also supported by the outcomes of the individual Forum member’s overall priorities.
The sensitivity analysis also supports this conclusion and indicates that the issue
of reserving or not reserving areas of land from timber production in the southern
New South Wales region is really reduced to a decision between the two extreme
forest use options as they currently stand. Maximizing the weighting of the
0.40
0.30
Priority
0.20
0.10
0.00
1 2 3 4 5
Option
Figure 6.10 Economic preference

Employment and Community needs sub-criterion did change the top ranked option
to Option 5. However, Option 1 was still ranked second under this variation.
Minimizing the weighting for Employment and Community Needs far increased
the priority of Option 1 even though Option 5 was still ranked second. The results
were sensitive to the arbitrarily assigned weightings used to portray the qualitative
differences in the National Estate data but this also only affected to any large
extent the rankings of Options 1 and 5. Further analysis of National Estate values
in the official process could have resulted in more precise measurement of the
impacts of this sub-criterion.
The effects of changing the two broad criteria weightings to an economic
preference or an environmental preference is to be expected due to the nature of
the forest debate. Changes to the social group weightings had little impact on the
final results. However, none of the sensitivity analyses performed had any impact
on raising the middle three options to the status of the chosen scenario. On the
basis of the impact data provided and the preferences of the Forum, the decision
appears to be reduced to one between the two extreme forest use scenarios and a
choice between conservation and jobs. In carrying out this MCA it seems that
there could have been great scope for revising the options (noting that the middle
options did not appear to have any impact in the analysis) to accommodate this
apparent trade-off between conservation and jobs. For example, one avenue which
could have been explored under this analysis is that of following the forest use
scenario of Option 1 but with increased emphasis on the encouragement of jobs
growth in industries that would not impact on forest ecosystems.
At the time of writing, no RFA has been signed by the New South Wales and
Commonwealth governments for the New South Wales Southern Region forests.
However, the New South Wales government has declared its own ‘Forest Agreement’
choosing, as a result of its own integration process, Option 3 as the preferred
156 Wendy Proctor
framework for its forest policy. It is hardly surprising that since announcing this
policy, intense public disagreement has resulted, with both conservation groups
and timber industry organizations criticizing the decision. Using an approach such
as MCA could have led policy-makers to foresee the response to such a proposal
and encouraged them to identify possible trade-offs in order to revise and fine-
tune the options. The results therefore show that overcoming the complex problems
involved in achieving ecologically sustainable development, such as those of
comparing multiple values and incorporating stakeholder participation into the
decision process, can be aided by the use of MCA.
Some specific merits and flaws deserve mention as a result of carrying out the
MCA. MCA imposed a rigorous structure on the decision-making problem that
allowed for a methodical approach to the breakdown of the problem which was
lacking in the approach to the official decision-making process. This structure and
methodical disentangling of the important issues also led to the identification of
several key areas of information that should have been collected, but were not, in
the official assessments. These included, for example, the issues of water and soil
quality and indigenous values under different options. Both of these criteria were
given high weightings of importance by most members of the Forum. As this
research shows, under the MCA process, data requirements would be identified
very early in the process, thus reducing the costs of collecting irrelevant or unneces-
sary information. Another advantage made evident from this analysis, is that the
objectives, criteria and indicators that are made explicit in the MCA technique,
could aid the process of future monitoring of environmental policies.
Some problems were also identified in using the MCA technique. For example,
there seems to be a lack of guidance in the MCA literature in carrying out inter-
active group analyses. Although the technique does facilitate a rigorous and
objective method of incorporating public participation into a policy-making process,
some parts of this process require skills that are not addressed in the MCA literature.
For example, in carrying out the first part of this analysis to elicit criteria, indicators
and preferences, a large amount of time was spent in interacting with the Forum
as a group and as part of their regular meetings. Conducting such a facilitating
role in issues that are highly contentious in nature requires certain skills, which are
not addressed in most textbooks on MCA. The Analytic Hierarchy Process appears
to be the only method which addresses (albeit briefly) the problem of conducting
interactive group decision-making. To be effective in this part of the process, it is
necessary for the analyst to borrow from the literature on negotiation, facilitation
and meeting techniques. There exists great scope for incorporating some other
techniques, for example, the deliberative approach of the Citizens’ Jury, into the
MCA. This approach would also aid the iterative phase that would take place at
the end of the MCA.
Notes
1 Under the RFA process, the Southern Region was divided into three smaller zones: the northern,
western and coastal sub-regions. Data in this research refer only to the coastal sub-region.
2 In this research, the term ‘decision-maker’ can also apply to multiple decision-makers that may,
for example, be comprised of stakeholders, bureaucrats, politicians or a combination of these.
3 The maximum weighting for the environmental and economic criteria were both around 0.75
and so, for the first test for example (Environmental Preference), the environmental weighting
was set at 0.75 and the relativities of the other two weightings were maintained as in the original
analysis at 0.15 for economic criteria and 0.10 for social criteria. The same was done for the
Economic Preference.
References
Commonwealth of Australia (1992) National Forest Policy Statement: A New Focus for Australia’s
Forests, Online. http://www.rfa.gov.au/nfps/contents.html (accessed April 2000).
Commonwealth of Australia (1997a) Australia’s First Approximation Report for the Montreal Process,
Canberra: Department of Primary Industries and Energy, Montreal Process Imple-
mentation Group.
Commonwealth of Australia (1997b) East Gippsland Resource and Economics Report, Joint
Commonwealth and Victorian Regional Forest Agreement (RFA) Steering Committee,
Canberra and Melbourne.
Commonwealth of Australia (1998) Central Highlands: Comprehensive Regional Assessment Report,
Canberra.
Dann, T. and Clarke, J. (1996) ‘An economic model for comprehensive regional forest
assessments, a case study – some issues and considerations’, paper presented at 40th
Annual Conference of the Australian Agricultural and Resource Economics Society,
Melbourne, 13–15 February 1996.
Dovers, S. (ed.) (2000) Environmental History and Policy: Still Settling Australia, Melbourne:
Oxford University Press.
Doyle, T. and Kellow, A. (1995) Environmental Politics and Policy Making in Australia, South
Melbourne: Macmillan.
Ferguson, I. (1996) Sustainable Forest Management, Melbourne: Oxford University Press.
Hogwood, B. and Gunn, L. (1984) Policy Making for the Real World, Oxford: Oxford University
Press.
Howlett, M. and Ramesh, M. (1995) Studying Public Policy: policy cycles and policy subsystems,
Oxford: Oxford University Press.
Lothian, J. (1994) ‘Attitudes of Australians towards the environment: 1975–1994’, Australian
Journal of Environmental Management, 1: 78–99.
Massam, B. (1988) Multi-Criteria Decision Making (MCDM) Techniques in Planning, Oxford:
Pergamon Press.
Prato, T. (1999) ‘Multiple attribute decision analysis for ecosystem management’, Ecological
Economics, 30: 207–22.
Pressy, R., Ferrier, S., Hutchinson, C., Sivertsen, D. and Manion, G. (1995) ‘Planning for
negotiation: using an interactive geographic information system to explore alternative
protected area networks’, in D. Saunders and J. Craig (eds) Nature Conservation 4: The
Role of Network, Sydney: Surry Beaty and Sons.
Proctor, W. (1999) ‘A practical application of MCA to forest planning in Australia’, paper
presented at IUFRO International Symposium: From Theory to Practice – Gaps and
158 Wendy Proctor
Solutions in Managerial Economics and Accounting in Forestry, Czech University of
Agriculture, Prague, 13–15 May 1999.
Proctor, W. (2000) ‘Towards sustainable forest management – an application of MCA to
Australian forest policy’, paper presented at the 3rd International Conference of the
European Society for Ecological Economics, Vienna, Austria, 3–6 May 2000.
Proctor, W. (2001) ‘Multi-criteria analysis and environmental decision-making: a case study
of Australia’s forests’, unpublished PhD thesis, Australian National University, Canberra.
Saaty, T. (1980) The Analytic Hierarchy Process, New York: McGraw-Hill.
Saaty, T. (1982) Decision-making for Leaders, California: Wadsworth.
Saaty, T. and Vargas, L. (1982) The Logic of Priorities, Boston: Kluwer-Nijhoff.
Stirling, A. and Mayer, S. (1999) ‘Rethinking risk: a pilot multi-criteria mapping of a
genetically modified crop in agricultural systems in the UK’, report for the UK
Roundtable on Genetic Modification, SPRU, University of Sussex, August 1999.
Turner, B. (1998) ‘Review of FRAMES data for the upper north east and lower north east
RFA regions of NSW’, unpublished consultant’s report for the State Forests of NSW
and the Bureau of Resource Sciences, Commonwealth DPIE.
Confronting risk with precaution 159
7 Confronting risk with

precaution
A multi-criteria mapping of
genetically modified crops
Andy Stirling and Sue Mayer
Introduction
The ‘precautionary principle’ is becoming an increasingly prominent theme in
the debate over technological risk. Many questions are raised over the implications
for policy making. In particular, concerns have been expressed over the relationship
between ‘precautionary’ and more traditional ‘science-based’ approaches to
decision making such as cost–benefit and risk analysis. Fears are sometimes raised
that – unlike risk assessment – a ‘precautionary approach’ is too ambiguous and
impractical to serve as a basis for real decision making, that it is somehow antag-
onistic to science and even that it threatens to stifle technological innovation and
economic growth.
The first part of this chapter takes a close look at some of the key practical and
theoretical issues bearing on the relationship between ‘science’ and ‘precaution’ in
the management of technological risk. It is found that – far from being in tension
– these two concepts might actually be seen as entirely consistent and even mutually
reinforcing. The real distinction is found to lie between a narrow ‘risk-based’ concept
of regulatory appraisal and broader precautionary approaches. A series of eight
criteria are developed against which appraisal techniques might be evaluated in
terms of both their scientific and precautionary validity. The second part of the
chapter then reports on a recent pilot exercise that seeks to apply these criteria to
the specific case of the regulatory appraisal of a genetically modified crop. The
main features of a novel ‘multi-criteria mapping’ methodology are outlined. Key
findings are reported, with particular attention paid to the extent to which the
method conforms to the scientific and precautionary quality criteria. Conclusions
are drawn concerning the degree to which the quality criteria can be made
operational in the practical business of regulatory appraisal.
Risk, regulation and precaution

Risk is a complex concept. Even under the most narrowly defined of quantitative
approaches, it is recognized that risk is a function of at least two variables – the
likelihood of an impact and its magnitude. However, it is only very rarely the case
that a series of technology, policy or investment options are seen to present only
160 Andy Stirling and Sue Mayer
one form of hazard. Normally, the characterization of risks associated with any
individual option requires the consideration of a wide variety of disparate risks.
In the energy sector, for example, risks can take forms including greenhouse gas
emissions, radioactive wastes, heavy metals, persistent organic pollutants, soil
erosion, thermal discharges, ambient noise, ecological disturbance or aesthetic
intrusion in the landscape. Each of these risks is manifest in a different way, with
different physical, biological, social, cultural and economic connotations.
The conventional analytical response to this breadth and diversity of issues in
regulatory appraisal is to identify a single major yardstick of performance and
seek to measure all the various aspects of risk using this as a metric. The chosen
unit of measurement in conventional risk assessment is usually human mortality
rates, although more complex appraisals sometimes also employ a variety of
measures of human morbidity effects. In some areas, the techniques of cost-benefit
analysis (CBA) are employed in seeking to subject a wider range of impacts to
measurement under a common monetary metric and so allow comparison with
the associated benefits. In this way, it is hoped that the multiplicity of magnitudes
typically confronted in the regulation of environmental and health risk may usefully
be reduced to a single key factor, thus apparently simplifying the process of
appraisal. This process of reduction is an essential element in what is sometimes
described as a ‘science-based’ approach to the regulatory appraisal of risk.
Of course, one crucial consequence of this artificial narrowing and conflation
of the full diversity of technological risks is effectively to exclude from consideration
many classes of effect. For instance, it is clear that only a minority of the types of
energy risks mentioned above is meaningfully addressed by a mortality, morbidity
or even a monetary metric. Moreover, even with respect to the single issue of
human health, risk is an inherently multi-dimensional concept. For instance, are
exposures voluntary or controllable? Are they manifest as disease, injuries or deaths?
How familiar are the risks? How immediately are they realized and how reversible
once identified? To what extent are they concentrated in large events or dispersed
in small routine incidents? How are they distributed across space, time and society?
Mortality, and even morbidity, indices fail to capture these important contextual
features.
Beyond this, further scope for divergent approaches to regulatory appraisal lies
in the characteriztics of the assessment process itself. Should appraisal take account
of social, economic, cultural and ethical issues, as well as environmental and health
factors? With respect to the more narrowly defined physical factors, to what extent
should appraisal seek to address the potential additive, cumulative, synergistic and
indirect effects associated with particular environmental and health risks? With
how wide an array of potential alternatives should each individual technological
or policy option be compared in appraisal? Should attention be confined simply
to the operation of the options concerned, or should it extend to the manufacture,
decommissioning and disposal of plant, as well as to the various inputs (such as
energy and materials) and associated risks at each stage? To what extent should
the relative benefits of different options be taken into account in appraisal so that
they can be offset against the associated risks?
In an ideal world, the appropriate response to factors such as these is easy to
determine. All else being equal, the regulatory appraisal of risk should be as
complete and as comprehensive as possible. However, aspirations to completeness
concerning the different classes and dimensions of risk and benefit and compre-
hensiveness concerning the different types of option provide only a rather loose
operational guidance in the practical regulation of risk. Moreover, even were
appraisal to be fully complete and comprehensive, in some hypothetical sense,
then there would still remain the problem of how the different aspects of risk
should be framed and prioritized in analysis. For instance, what assumptions should
be made about adherence to best practice in the various activities under appraisal?
What relative priority should be attached to different effects such as toxicity,
carcinogenicity, allergenicity, occupational safety, biodiversity or ecological integrity?
What weight should properly be placed on impacts to different groups, such as
workers, children, pregnant and breastfeeding mothers, future generations,
disadvantaged communities, foreigners, those who do not benefit from the tech-
nology in question or even to animals and plants as beings in their own right?
Even if they were practically feasible, objectives such as completeness or comprehen-
siveness do not assist in addressing issues of framing and prioritization of this
kind. No one set of assumptions or priorities may be claimed to be uniquely rational,
complete or comprehensive.
It is here that we come to a classic and well-explored dilemma in the field of
social choice theory, but one that is frequently forgotten in risk assessment and
regulatory appraisal. The disciplines of risk assessment, economics and decision
analysis have developed no single definitive way of addressing the problems of
comparing ‘apples and oranges’. Even the most optimistic of proponents of rational
choice acknowledge that there is no effective way to compare the intensities of
preferences displayed by different individuals or social groups (Bezembinder 1989).
Indeed, even where social choices are addressed simply in relative terms, the
economist Kenneth Arrow went a long way towards earning his Nobel Prize by
demonstrating formally that it is impossible definitively to combine relative
preference orderings in a plural society (Arrow 1963).
Put simply, the point is that ‘it takes all sorts to make a world’. Different cultural
communities, political constituencies or economic interests typically attach different
degrees of importance to the different aspects of environmental risk and look at
them differently. Within the bounds defined by the domain of plural social discourse,
no one set of values or framings can definitively be ruled more ‘rational’ or ‘well
informed’ than can any other. Even were there to be complete certainty in the
quantification of all the various classes and dimensions of risk, it is entirely reason-
able that fundamentally different conclusions over environmental risk might be
drawn under different – but equally legitimate – perspectives. It is a matter of the
science of risk assessment itself, then, that there can be no analytical fix for the
scope, complexity and intrinsic subjectivity of environmental and health risks.
The notion that there can be a single unambiguous ‘science-based’ prescription in
the regulatory appraisal of risk is not only naïve and misleading; it is a fundamental
contradiction in terms.
The depths of incertitude
This problem may seem serious enough. Unfortunately, the difficulties encountered
in the regulatory appraisal of risk are even more intractable than this. Thus far we
have considered only the issues associated with the characterization of the
‘magnitude’ aspects of risk. What of the likelihoods? Here we come upon some
profound limitations to the applicability and robustness of probabilistic approaches
that are as seriously neglected in regulatory appraisal as are the difficulties discussed
above concerning the comparison of magnitudes.
In economics and decision analysis, the well-established formal definition of
risk is as a condition under which both a comprehensive set of all possible outcomes
can be defined and a discrete set of probabilities (or a density function) can be
defined across this array of outcomes. This is illustrated in the top left-hand corner
of the diagram in Figure 7.1. This is the domain under which the various probab-
ilistic techniques of risk assessment are applicable, permitting (in theory) the full
characterization and ordering of the different options under appraisal. There is a
host of details relating to this picture (such as those hinging on the distinction
between ‘frequentist’ and ‘Bayesian’ understandings of probability), but none of
these alter the formal scientific definition of the concept of risk.
The strict sense of the term uncertainty, by contrast, applies to a condition
under which there is confidence in the completeness of the defined set of outcomes,
but where there is acknowledged to exist no valid theoretical or empirical basis
confidently to assign probabilities to these outcomes. This is found in the lower
left-hand corner of Figure 7.1. Here, the analytical armoury is less well developed,
with the various sorts of scenario analysis being the best that can usually be managed
(Funtowicz and Ravetz 1990). Whilst the different options under appraisal may
still be broadly characterized, they cannot be ranked even in relative terms without
some knowledge of the relative likelihoods of the different outcomes.
Both risk and uncertainty, in the strict senses of these terms, require that the
different possible outcomes can be clearly characterized and subject to measure-
ment. The discussion in the previous section has already made it clear that this is
often not the case – the complexity and scope of the different forms of environ-
mental risk and the different ways of framing and prioritizing these, can all-too-
easily render ambiguous the definitive characterization of outcomes. This may be
so, even where there is relatively high confidence in understandings of the likelihood
that at least some form of impact will take place (top right corner of Figure 7.1).
An illustrative example here might be the prospects for regional climatic, ecological
and socio-economic impacts arising from the human-enhanced greenhouse effect.
Where these problems are combined with the difficulties in applying the concept
of probability, we face a condition which is formally defined as ignorance (bottom
right corner of Figure 7.1) (Loasby 1976; Smithson 1989; Wynne 1992). This
applies in circumstances where there not only exists no basis for the assigning of
probabilities (as under uncertainty), but where the definition of a complete set of
outcomes is also problematic. In short, recognition of the condition of ignorance
is an acknowledgement of the possibility of surprises. Under such circumstances,
KNOWLEDGE KNOWLEDGE ABOUT OUTCOMES
ABOUT
continuum of set of discrete outcomes
LIKELIHOODS outcomes outcomes poorly defined
INCERTITUDE
RISK
Apply:
firm basis frequentist discrete AMBIGUITY

for distribution frequentist
probabilities functions probabilities
Apply:
shaky basis Bayesian discrete fuzzy logic

for distribution Bayesian sensitivity analysis
probabilities functions probabilities precaution
UNCERTAINTY IGNORANCE
no basis
for Apply: Apply:
probabilities scenario analysis diversity
sensitivity analysis flexibility
precaution precaution
Figure 7.1 Risk, uncertainty, ambiguity and ignorance defined
not only is it impossible definitively to rank the different options, but even their full
characterization is difficult. Under a state of ignorance (in this strict sense), it is
always possible that there are effects (outcomes) which have been entirely excluded
from consideration.
Figure 7.1 provides a schematic summary of the relationships between these
formal definitions for the concepts of risk, uncertainty, ambiguity and ignorance.
It is quite normal, even in specialist discussion, for the full breadth and depth of
these issues to be rolled into the simple concept of ‘risk’ (and sometimes ‘uncert-
ainty’), thus seriously understating the difficulties involved. In order to avoid
confusion between the strict definitions of the terms ‘risk’ and ‘uncertainty’ as
used here, and the looser colloquial usages, the term ‘incertitude’ can be used in a
broad overarching sense to subsume all four subordinate conditions. Either way, it
is not difficult to see that it is the formal concepts of ignorance, ambiguity and
uncertainty – rather than mere risk – which best describe the salient features of
regulatory decision making in areas such as energy technologies, toxic chemicals
and genetically modified organisms. Indeed, many of the most high profile
technologically-induced ‘risks’ of recent years – such as stratospheric ozone
depletion, endocrine disrupting chemicals and BSE, for instance – are all cases
where the problem lay not so much in the determination of likelihoods, but in the
anticipation of the very possibilities. They were surprises!
The crucial point is that intractable uncertainties, ambiguities and ignorance
are routinely treated in the regulatory appraisal of technology simply by using the
probabilistic techniques of risk assessment. This treatment of uncertainty and
ignorance as if they were mere risk effectively amounts to what the economist
Hayek dubbed (in his Nobel acceptance speech) ‘pretence at knowledge’ (Hayek
1978). Far from displaying a respect for science in regulatory appraisal, the effect
of such scientistic oversimplification is actually to ignore and undermine the
scientific principles on which risk assessment itself purports to be based. Given the
manifest inapplicability – in their own terms – of probabilistic techniques under
uncertainty and ignorance, this is a serious and remarkable error. The self-
contradictions in aspirations to a ‘science-based’ approach reliant solely on
quantitative risk assessment, already noted in the last section, are thus further
underscored and reinforced.
Why is it that pursuit of (and claims to) the definitive authority of ‘science-
based’ approaches continues to be so prominent in regulatory appraisal? It seems
that the elegance and facility of probabilistic calculus has had a seductive effect on
many risk analysts and their sponsors. This may be understandable, yet it is also
curious. Despite the intractability of the condition of ignorance, there is no shortage
of operational tactical and strategic ‘precautionary’ responses. Some specific
features of these approaches will be reviewed in some detail later. For the moment,
in the specific context of ignorance, the point is simply that there do exist practical
alternatives to the use of probabilistic methods. For instance, there exists a variety
of institutional procedures for including in the regulatory appraisal process a range
of different scientific disciplines and people with pertinent professional and local
knowledge and relevant socio-economic perspectives. By providing for the
identification of a wider range of possibilities, this effectively helps to convert
some part of the domain of ignorance into the more tractable condition of
uncertainty. Here, techniques such as scenario and sensitivity analysis can also
help systematically to characterize these neglected possibilities and explore their
implications under different perspectives.
Beyond this, there are a series of broader strategies that may be employed and
which will be returned to later. In particular, rather than focusing entirely on efforts
to characterize the ‘problem’ (ignorance) attention can also be devoted directly at
aspects of the ‘solution’. Although the manifestations of ignorance are, by definition,
not characterizable in advance, certain dynamic properties of the different options
themselves can offer valuable ways of hedging against ignorance. Here properties
such as flexibility, reversibility, resilience, robustness and adaptability are all
potentially valuable (Stirling 1999; Salo 1999). Perhaps even more important, are
the possible merits of deliberate diversification across a range of options. After all,
it is a well-established matter of common sense that, when we don’t know what we
don’t know, we don’t put all our eggs in one basket! It may be that a persistent
preoccupation with probabilistic methods has left these kinds of strategies unduly
neglected in the regulatory appraisal of technological risk.
The practical consequences for risk assessment

The problems discussed so far – the multi-dimensionality of environmental and
health risks and the conditions of uncertainty, ambiguity and ignorance – may all
seem a little abstract and theoretical. It is perhaps also partly for this reason that
they remain relatively neglected in the business of regulatory appraisal.
Unfortunately, however, they have some important practical consequences that,
though often concealed, hold profound implications for the interpretation of
orthodox risk assessment results in all fields, extending from the regulation of energy
options, through chemicals and industrial hazards to genetic modification
technologies.
In all these areas, the typical response to these difficulties in regulatory appraisal
is to reduce and simplify – focusing on those aspects that are either the most tractable
or the most ‘reasonable’ under certain dominant perspectives. In this way, individual
studies can construct a picture of environmental risks, which appears to be quite
unambiguous and precise. The scale of the discrepancies only becomes evident
on occasions when attention is extended to a series of different appraisal studies,
each applying subtly different – but equally ‘reasonable’ and ‘legitimate’ – framing
assumptions concerning the different dimensions of appraisal discussed here. When
this takes place, it becomes clear that the apparent relative riskiness of different
options can vary quite radically, depending on the framings and priorities attached
to the ‘hidden variables’ during the process of appraisal.
Figure 7.2 illustrates this by showing the results obtained in 32 large-scale risk
assessments of eight different energy technologies conducted in industrialized
countries over the past two decades. Here, environmental and health effects are
characterized using the techniques of CBA as monetary ‘external costs’ expressed
in standardized form per unit of electricity production (Stirling 1997). This case is
taken as an example because both the techniques employed, and this particular
field of application, might arguably be seen as being among the most mature and
intensively explored areas of application of comparative risk assessment. The
picture is not specific to these techniques or this field. A similar pattern may be
found in a variety of other regulatory fields, including transport, toxic chemicals
and food safety. The same pattern is also evident in the underlying physical and
mortality indices on which these monetary results are based. A number of salient
features can be seen.
First, individual studies present their results with great precision – often as a
single value rather than a range and sometimes expressed with as many as four
significant figures (one part in ten thousand). Yet, the variability in the results
obtained in the literature as a whole for any one option is radically larger. For
instance, the uppermost values of the highest range assessing the risks associated
with coal power amount to the equivalent of some twenty dollars per kilowatt-
Logarithmic scale
0.001 0.01 0.1 1 0 1 100 1000
ONSHORE WIND
PHOTOVOLTAICS
BIOMASS
HYDROELECTRICITY
NUCLEAR FISSION
GAS
OIL
COAL
0.001 0.01 0.1 1 0 1 100 1000
Figure 7.2 Ambiguity of ordering in risk assessments
hour of electricity production. The lowest values of the bottom range in Figure
7.2 are less than four hundredths of a cent per kilowatt-hour. The difference is
more than four orders of magnitude – a factor of more than fifty thousand! Detailed
analysis of the reasons for these discrepancies show that they do not arise as a
result of any single factor. It is not a simple matter of some studies being more
‘accurate’ or ‘reasonable’ than others in any definitive sense. Instead, the variability
is the cumulative consequence of the adoption of divergent assumptions and
priorities concerning the whole range of the different ‘dimensions of appraisal’
identified in the preceding sections (Stirling 1997).
The second crucial feature that is illustrated in Figure 7.2 concerns the ambig-
uities in the ordering of the different options under appraisal. The lowest values
obtained for the worst ranking option (coal) are lower than the highest values
obtained for the apparently best ranking options (wind). Since the effect of the
particular assumptions adopted in individual studies is to produce results at the
high end of the overall range for some options but lower in the distributions for
others, the overall picture yielded by the literature as a whole would accommodate
virtually any conceivable ranking order for these eight options! By the judicious
choice of framing assumptions, then, radically different conclusions can be justified
for regulation.
This evident disjuncture between precision and accuracy in supposedly ‘science-
based’ risk assessment paints a rather negative picture. One of the first and most
basic tasks in the management of risk is to construct some robust overall notion of
the relative merits of the different options under consideration from the point of
view of society as a whole. This then serves as a basis for regulatory intervention,
market-based measures or investment initiatives. Where this cannot be achieved
in any absolute (or even relatively robust) sense, then the value of appraisal lies in
systematic exploration of the relationships between different assumptions in analysis
and the associated pictures of the relative importance of different options. Where
aspirations to the ‘science-based’ appraisal of risk lead to the assertion of the
intrinsic authority of narrow risk assessment procedures, then these crucial
exogenous factors typically remain unacknowledged and unexplored. In this event,
the problem is not simply one of a lack of rigour concerning the theoretical
contradictions noted in the previous sections. The difficulties are also very concrete
and pragmatic. For, without a robust appreciation of the assumptions under which
appraisal yields differing pictures of performance, serious questions must be raised
over whether the associated results – no matter how confidently and precisely
expressed – are of any practical policy use at all.
‘Science’ and ‘precaution’ in the regulatory appraisal of

environmental risk
It is with increasing realization of these practical and theoretical limitations to the
value of orthodox risk assessment in regulatory appraisal, that interest is growing
in complementary and alternative approaches. In particular, the ‘precautionary
principle’ is becoming an ever more prominent feature of the regulatory debate
on environmental risks and of national and international legislation (O’Riordan
and Cameron 1994; Fisher and Harding 1999; Raffensberger and Tickner 1999;
O’Riordan and Jordan 2001). Although subject to a variety of different definitions,
in the broadest of terms, a ‘precautionary’ approach acknowledges the difficulties
in risk assessment by granting greater benefit of the doubt to the environment and
to public health than to the activities which may be held to threaten these things.
A host of different practical instruments and measures are variously proposed in
different contexts as embodiments of a ‘precautionary approach’ or as means to
implement a ‘precautionary principle’. For present purposes, attention will
concentrate on the way in which a precautionary approach offers a direct response
to the practical and theoretical problems in regulatory appraisal which have been
discussed so far.
One key theme in the current lively debate on these matters surrounds the
frequent assertion (and sometimes assumption) that – whatever form it takes – a
‘precautionary’ approach to the management of environmental risk is somehow
in tension with (or even antithetical to) the generally uncontroversial aspiration
that regulatory decision making should be based on ‘sound science’. Of course,
this does not address the extent to which orthodox ‘scientific’ approaches such as
comparative risk assessment may themselves be claimed to yield ‘sound’ results.
The thrust of the discussion thus far has been to raise serious doubts over this.
Nevertheless, the important question remains as to what exactly is the relationship
between so-called ‘science-based’ and ‘precautionary’ approaches to the regulation
of environmental risk?
A necessary starting point for this analysis is a clear characterization of exactly
what is meant by ‘science’ and ‘precaution’ in the context of decision making on
environmental risk. Drawing on a wide literature, Figure 7.3 displays some idealized
attributes of scientific approaches to regulatory appraisal (Stirling 1999). In short,
a scientific approach to the management of risk should, ideally and at minimum,
be transparent in its argumentation and substantiation, systematic in its analytical
methods, sceptical in its treatment of knowledge claims, subject to peer review,
independent from special interests, professionally accountable and continually open
to learning in the face of new knowledge. These aspirations may not always be
realized, but they represent fundamental, and relatively uncontroversial, principles
guiding any ‘science-based’ approach to regulatory appraisal.
Likewise, drawing on an equally extensive parallel literature, it is possible broadly
to characterize the essential features of a ‘precautionary’ approach to the manage-
ment of risk. In short, a precautionary approach involves the application of
principles that ‘prevention is better than cure’, that ‘the polluter should pay’, that
options offering simultaneously better economic and environmental performance
should always be preferred (‘no regrets’), that options should be appraised at the
level of production systems taken as a whole and that attention should be extended
to the intrinsic value of non-human life in its own right (a ‘biocentric ethic’). In
effect, this is variously taken to mean a certain humility about scientific knowledge
and an acknowledgement of the complexity and variability of the real world. It
implies recognition of the vulnerability of the natural environment and living
organisms and the prioritizing of the rights of those who stand to be adversely
SCIENTIFIC Precautionary
Risk-based approach
DISCIPLINE
approach
systematic
sceptical
peer-reviewed
independent
accountable
learning
Cornucopian Apocalyptic
BREADTH OF SCOPE
consider all effects acknowledge possibility of surprise

compare alternatives include many perspectives
account for pros as well as cons take broad social view
reverse onus of persuasion extend research and monitoring
Figure 7.3 A model of the relationships between risk, science and precaution
affected. It requires scrutiny of claims to benefits and justifications as well as risks
and costs, with full account given to the available alternatives. Finally, a
precautionary approach involves the adoption of long-term, holistic and inclusive
perspectives in regulatory appraisal (Stirling 1999).
In many ways, these attributes of a ‘precautionary’ approach can be seen to
concern different aspects of the breadth of the regulatory appraisal process. A
‘broad’ regime is one that takes account of a wide range of different types of
impact, including qualitative as well as quantitative issues and including indirect
as well as direct effects. Likewise, a ‘broad’ framework accommodates a diverse
array of different points of view (including, importantly, those of potential ‘victims’)
and anticipates a wide range of possibilities in the face of uncertainty and ignorance.
It extends consideration to the benefits and justifications associated with the
introduction of the technology in question and examines a variety of alternative
ways in which the benefits of a regulated technology might be realized at lower
levels of risk. Taken together, these features constitute a more ‘precautionary’
approach because they increase the number and intensity of the constraints that
any technological option must satisfy in order to be approved by the regulatory
process, thus making it more difficult for certain innovations to pass through the
regulatory ‘filter’. At the same time, however, such measures might equally serve
to encourage other technological innovations that might otherwise remain
neglected.
What is interesting about this characterization of ‘precaution’ in terms of the
‘breadth’ of the associated regulatory regime, is that it reveals an inherently con-
sistent – and in many respects complementary – relationship between ‘precaution’
and ‘science’ in the management of technological risk. Accordingly, Figure 7.3
distinguishes between different approaches to risk management based on the degree
to which each embodies the respective characteristics of ‘scientific discipline’ and
‘breadth of framing’ identified here. Of course, both the ‘broad’/‘narrow’ and
the ‘scientific’/‘unscientific’ dichotomies drawn here are highly stylized and
simplified. However, the general picture revealed in Figure 7.3 is at least richer
and more realistic than the prevailing one-dimensional dichotomy between ‘science’
and ‘precaution’. Taken together, the combination of these two dichotomies
generates a fourfold array of idealized permutations. The adoption of a ‘narrow’
regime without reference to scientific understandings or disciplines in appraisal
might be described as a permissive position. Taken to an extreme, this would amount
to an entirely uncritical ‘anything goes’ approach to the regulation of technology
of the kind associated with caricature ‘cornucopian’ visions of technological
progress. Likewise, a broad-based regime might be similarly unscientific. The
resulting restrictive position might be associated with a caricature ‘apocalyptic’
vision of technology. In the extreme, it would lead to a situation of paralysis under
which no new technological innovation that offends in the slightest respect would
ever be approved for deployment. The crucial point is that neither the ‘permissive’
(cornucopian) nor the ‘restrictive’ (apocalyptic) positions as defined here would be
subject to challenge or reversal by the disciplines of scientific discourse associated
with the vertical axis.
Clearly, neither the established procedures of risk regulation (based on relatively
narrowly framed risk assessment methods) nor the emerging precautionary
approach (based on broader perspectives and considerations) actually resemble
these stylized ‘permissive’ or ‘restrictive’ caricatures. Existing risk-assessment-based
regulation includes a host of effective checks and balances. It certainly does not
necessarily provide for the uncritical approval of any new technology that may be
developed. Likewise, even the most progressive formulations of a ‘precautionary
principle’ are circumscribed in their scope, admit an incremental series of instru-
ments and allow for regulatory approval under a host of favourable conditions.
Both approaches are compatible – at least in principle – with the requirements of
systematic methodology, scepticism, transparency, accountability, quality control
by peer-review, professional independence and an emphasis on learning which are
held for the purposes of this discussion to be among the key aspirations of a ‘science-
based’ approach.
Now let us return to the earlier discussion of the profound importance of the
conditions of uncertainty, ignorance and multidimensionality in risk assessment.
Questions over the scope of appraisal, the plurality of different value positions
and framing assumptions, the diversity of different anticipated possibilities and
the degree of confidence placed in the available knowledge are all matters that are
central to the ‘scientific’ status of the appraisal process. That probabilistic
approaches are inapplicable under strict uncertainty and ignorance flows directly
from the theoretical foundations of risk assessment, and CBA (and, indeed, all
‘rational choice’ approaches to decision-making on risk). Following equally directly
from these fundamental theoretical principles, different priorities, framing
assumptions and value systems cannot be definitively aggregated across different
groups. Thus, for both these reasons, there can be no analytical fix for the definitive
ranking of different technology or policy options in the social appraisal of risk. All
that can be done to maximize scientific rigour in appraisal is to ensure that the
process is as broadly based as possible in terms of the value systems and framing
assumptions that are included and the options and possibilities that are addressed.
Seen in this way, then, key elements of the ‘breadth’ of the regulatory regime
themselves become issues of ‘sound science’ in the management of environmental
risk, as well as institutional features of the wider regulatory regime. Precaution, in
this sense, is not just entirely consistent with science – it is a necessary pre-requisite
for a truly scientific approach to the regulatory appraisal of risk.
Implementing a precautionary approach to appraisal

In the discussion so far, environmental and health risks have been treated at a high
level of generalization. Of course, it is obvious that at a greater level of detail,
different technological and policy options will vary radically in the magnitude and
character of the risks that they present. Likewise, the range of different possible
regulatory interventions do not fall neatly into ‘permissive’ or ‘restrictive’ categories,
but lie along a series of continua, and range between being more and less precau-
tionary in their effect. With respect both to precautionary and other approaches
to the regulation of risk, then, different measures will be appropriate in different
contexts. This general picture is the subject of a vast literature (e.g. O’Riordan
and Cameron 1994; Fisher and Harding 1999; Raffensberger and Tickner 1999;
Stirling 1999; Renn and Klinke 1999; O’Riordan and Jordan 2001). Here, however,
the purpose is to focus specifically on some key implications for regulatory appraisal.
As a result of the foregoing analysis, we construct a series of eight evaluative
criteria against which the regulatory appraisal of risk can be assessed both in terms
of its scientific rigour and its precautionary qualities (after Stirling 1999). In short,
these are as follows:
1 Humility: maintain a culture of humility in the face of the many sources of

uncertainty, ignorance and subjectivity in appraisal. Avoid claims to complete
or otherwise definitive knowledge.
2 Completeness: broaden the scope of the regulatory appraisal of technological
risk to address cumulative, additive, complex, synergistic and indirect effects
as well as more direct causal processes.
3 Benefits and Justifications: include systematic consideration of the claimed
benefits and justifications as well as adverse effects, in order to allow
determination of ‘net’ benefits under different contexts.
4 Comparison: conduct appraisal on a comparative rather than a case-by-case
basis, including account of a variety of technological and policy options and
the cumulative effects across different cases.
5 Participation: ensure full engagement by all interested and affected parties,
both to elicit all relevant knowledge and to include consideration of all pertinent
priorities and framing assumptions.
6 Mapping: express appraisal results not as discrete numerical values, but using
sensitivity analysis systematically to ‘map’ the consequences of different value
judgements and framing assumptions.
7 Transparency: use the most straightforward of methods. Minimize the number
of hidden variables. Provide for detailed auditing of how particular results
derive from particular inputs.
8 Diversity: extend appraisal to address the way that diverse mixes of different
options may help to hedge against uncertainty and ignorance and help
accommodate divergent social perspectives.
These are the considerations that have informed the development of the ‘multi-
criteria mapping’ technique employed in the present case study (Stirling 1997;
Stirling and Mayer 1999). In short, the motivation behind this approach is to seek
to combine the openness and qualitative flexibility of participatory deliberation
with the clarity and focus of quantitative assessment. The specific case study with
which this method has been piloted concerns the hotly contested debate over the
use of genetically modified crops in UK agriculture.
Applying the multi-criteria mapping technique
The UK debate over GM crops

The agricultural use of GM technologies is held in some quarters to promise great
benefits. On the other hand, there is general agreement that there exists at least
the potential for serious, irreversible harm. In the UK, formal regulatory appraisal
of GM crops has centred on the question of whether or not they are ‘safe’ for the
environment and for human consumption. However, there is considerable scientific
incertitude over the form and magnitudes of the possible effects and, as yet (by
contrast with chemical or nuclear risks), little accumulated practical experience to
draw upon. This has led to the evolution of a set of controls which are intended to
be precautionary in nature – where it is accepted that action to avoid harm may
be taken in the absence of scientific proof – with the conduct of risk assessment
being required before experimental or commercial use of a particular genetically
modified organism is allowed.
Despite this somewhat precautionary approach to risk regulation enshrined in
the European Commission’s Deliberate Release Directive (90/220/EC) and the
Novel Foods Regulation, the regulatory appraisal process has failed to gain
confidence, either of Non-Governmental Organizations (NGOs), private industry
(Mayer et al. 1996) or the general public (EPCAG 1997; Grove-White et al. 1997).
This lack of confidence arises among other reasons: because the scope of the
regulatory appraisal is still held in many quarters to be too narrow; because there
is a general lack of trust in official reassurances of safety (particularly in the wake
of BSE); and because justifications and benefits are not explicitly included in the
evaluation process. Industry and regulators have expressed frustration, believing
that the precautionary approach is being invoked in too burdensome a fashion,
with unrealistic demands being made concerning absolute proof of safety.
It has also been almost impossible to gain agreement between European Union
Member States over whether particular commercial releases of GM crops are
environmentally ‘safe’, despite a supposedly common approach to their risk assess-
ment (Von Schomberg 1998; Wynne and Mayer 1999). Disputes routinely emerge
over the appropriate scope of regulatory appraisal. Even where there is agreement
over the possibility that effects will occur, notions of what constitute adverse effects
remain strongly contested. These sorts of problems with the current regulatory
appraisal of GM crops are typical of those that beset the use of conventional risk
assessment and cost-benefit analysis in other areas. Taken together, these charact-
eristics of the debate over the use of GM technologies in agriculture make it a
challenging candidate for a case study concerning the application of the general
principles of scientific and precautionary appraisal enunciated above.
The pilot study

The details of the multi-criteria mapping methodology and this particular
application in the GM field in the UK are described at length elsewhere (Stirling
1997; Stirling and Mayer 1999, 2000). Only those features bearing on the
examination of the quality criteria given above will be outlined here. The pilot
study took place between April 1998 and September 1999. It was funded by the
transnational food firm Unilever, but was entirely independent in its conception,
design, implementation and reporting. The two authors come from an academic
and NGO background and the project was overseen by a steering group comprising
a wide range of environmental, consumer, farming and food industry represen-
tatives. The specific topic chosen for the study was the production of oilseed rape
in the UK. Six ‘basic options’ were identified in advance for the purposes of
comparison (Table 7.1).
In consultation with the steering group, twelve individual high profile
protagonists in the GM food debate were selected as participants. These individuals
came from a variety of backgrounds, including academics and government advisers,
environmental, consumer and religious organizations and representatives from
the farming, food and biotechnology industries. Of course, it is impossible to claim
any ‘statistically representative’ status for such a small sample. However, care was
taken that, between them, the group of participants covered a full ‘envelope’ of
specialist and socio-political perspectives – ranging, for instance, from strongly
opposed to strongly in favour of the use of GM crops. Given the character of the
UK GM debate at the time of the study (1998–9), it was necessary to give
undertakings of anonymity in order to secure participants’ involvement. However,
the viewpoints of participants are reproduced individually, being identified by a
code letter and an affiliation with one of four general groupings of perspectives:
‘academic’, ‘NGO’, ‘industry’ or ‘government’ (Table 7.2).
Table 7.1 The definitions of the ‘basic options’ appraised by all participants
Option Definition
Organic agriculture All farming and food production conducted under present-
day organic standards.
Integrated pest All farming and food production conducted using systems
management designed to limit but not exclude chemical inputs and with
greater emphasis on biological control systems than
conventional systems.
Conventional All farming and food production conducted under present-
agriculture day intensive systems.
GM oilseed rape with Labelling based on the presence of foreign DNA or protein in
segregation and present the final product.
systems of labelling
GM oilseed rape with Monitoring for effects (mainly environmental) conducted on
post-release monitoring an on-going basis after commercialization.
GM oilseed rape with Areas of growing of GM oilseed rape restricted on a
voluntary controls on voluntary basis to avoid unwanted effects such as gene-flow
areas of cultivation and cross-fertilization of non-GM crops.
Up to six additional Any option of participant’s choice including combinations of
options to be specified the above if desired.
by participant
Table 7.2 The participants
Area Code
Agriculture and food industry B, L, H, K
Academic scientists C, J
Government safety advisors E, F
Religious and public interest groups A, D, G, I
During a 2–3 hour individual interview, each participant undertook a four-

stage process (Figure 7.4): (i) the identification of additional options; (ii) the defining
of appraisal criteria under which the options should be assessed; (iii) the scoring
of the performance of each option under each criterion; and (iv) the weighting of
each criterion in terms of its relative importance. A straightforward ‘linear additive’
multi-criteria procedure was employed using a simple spreadsheet model mounted
on a portable computer in order to display to the participant the results of the
scoring and weighting process as it developed. The process then iterated cyclically
until the participant was satisfied that the results accurately reflected their own
personal perspective and professional judgement on the issues at hand.
One crucial feature of the scoring process was the deliberate eliciting of a range
of performance scores for each option under each criterion. Rather than providing
a single ‘best guess’, this allowed consideration of a range of optimistic and pessim-
istic assumptions under each perspective. The discussion of the associated issues
was then carefully documented and provided an interesting reflection of the
importance of technical uncertainties in the final picture of performance derived
under each perspective.
In the period following the interviews, participants were each contacted and
asked to consider one final aspect of appraisal. This concerned the role of deliberate
diversification among options as a means to implement two crucial features of the
precautionary approach discussed earlier in this chapter: the desire to hedge against
intractable uncertainties and ignorance and to accommodate a variety of socio-
political viewpoints. For this purpose, a simple index of diversity was borrowed
from mathematical ecology and information theory in order to model different
trade-offs between the value attached to the performance of individual options
and the value attached to diversity in the mix of options. The details of this method
are described in more detail elsewhere (Stirling 1994, 1997; Stirling and Mayer
1999). In short, participants can choose to assign a ‘zero’, ‘low’, ‘medium’ or ‘high’
weighting to diversity. Where this takes a zero value, then only the best-performing
option is included in the favoured mix. Where higher weightings are placed on
diversity, progressively greater contributions are included from lower-performing
options – drawn on in proportion to their relative performance under the perspective
in question. In this way, the study was able to elicit a crude notion of the importance
of ignorance and pluralism under the different perspectives.
As analysis proceeded, participants were asked to review the results of a thorough
sensitivity analysis conducted on their own set of weightings. Along with an
1
CHOOSE
OPTIONS
OPTIONS
CRITERIA WEIGHTS A B C D E
environment
2 health 5
DEFINE
CRITERIA economics
SCORES CALCULATE
RANKS
other pessimistic / optimistic
RANKS
3
ASSESS
4
SCORES
ASSIGN
WEIGHTS
EXAMPLE
GM 0.14
pessimistic assumptions 0.14 1 optimistic assumptions
GM0.572 0.5717 1
GM0.566 0.5658 1
con0.252 0.2521 0
organic IPM0.142 0.1423 1
org 0.21 0.21 1
integrated pest management
conventional
GM - current
GM - monitoring
GM - voluntary controls
low performance high performance
Figure 7.4 The multi-criteria mapping process

invitation that participants retain the spreadsheet model and freely experiment
with it at their leisure, this provided a further means to verify that the final results
represented a robust reflection of the different viewpoints. Finally, all participants
were invited to a concluding workshop at which the results were again confirmed
and used as a basis for wider-ranging discussion.
Findings
The procedure summarized above generated a rich body of empirical material,
reflecting a very wide range of issues bearing on the regulatory appraisal of GM
technologies in agriculture. This is reported on in more detail elsewhere (Stirling
and Mayer 1999, 2000) but the key points can be summarized here. In addition to
the six ‘basic options’ a further 18 alternative agricultural strategies were identified
and evaluated by participants. These included many different labelling and control
regimes and the application of a series of different regulatory assessment procedures
and criteria. They also involved a variety of different agricultural strategies,
including a focus on hypothetical strategies using GM technologies under an organic
farming regime. Interestingly these tended to perform relatively well under the
perspectives of participants from both sides of the GM debate (Table 7.3). A total
of 117 appraisal criteria were defined by different participants. Although some of
these are apparently closely related, all embody important differences of emphasis
or framing. It is interesting that criteria typically reflect issues not only of consumer
choice but also of citizenship and wider questions of participation and agency.
Collectively they cover a very wide range of considerations, including environ-
mental, agricultural, health, economic, social and ethical issues. Most of these
issues are very remote from the factors included in orthodox risk assessment. Indeed,
for no participant was it true that the full range of their concerns is fully addressed
by existing regulatory appraisal procedures in the UK (Table 7.4).
Table 7.3 Additional options defined by participants

Labelling GM crops with segregation, full labelling and post- A
and/or other release monitoring and legally binding growing contracts
controls GM crops with segregation, current labelling and post- F
release monitoring
GM crops with segregation, full labelling and post- H
release monitoring
GM crops with segregation and labelling according to G
means of production and source of gene, plus post-
release monitoring
GM crops with segregation, comprehensive labelling I
based on process and generic restrictions on some
classes, e.g. in centre of origin
GM crops within controlled sectors (compulsory control) A
GM crops with legally binding threshold for gene transfer A
to non-GM stream
Agricultural GM crops, IPM system G
system GM crops, IPM system J
No GM crops – conventional and organic as now K
GM crops in conventional and organic systems K
GM crops, organic agricultural system, plus J
segregation, labelling and other regulations as required
Assessment GM crops with quality B
criteria GM crops with assessment of indirect agricultural I
impact and assessment of need
Other GM crops only in USA A
No GM commodity crops A
Complete public control over choice C
Table 7.4 Broad groupings of criteria defined by participants
Environment Biodiversity e.g. field boundary ecology, other environmental
risks
Chemical use e.g. reduction in use of existing herbicide sprays,
benefits of contact herbicides versus soil acting
residuals, longer term pollution of air and water
Genetic pollution e.g. gene flow to other crops and native flora
Wildlife effects e.g. impact of enhanced weed control efficiency on
wildlife, other practices affecting wildlife value of
agricultural systems
Unexpected effects e.g. potential for effects not foreseen under this
scheme
Visual e.g. amenity impacts
Aesthetics e.g. feelings about environment
Health Allergenicity e.g. from food consumption
Toxicity e.g. human or animal health
Nutrition e.g. to consumers
Unexpected effects e.g. unexpected interactions between ingredients,
stability of genetic insert
Ability to manage e.g. traceability and ease of recall
Agriculture Weed control e.g. invasive volunteers and weedy relatives
Food supply e.g. sustainability, tendency to monocultures, global
stability food security
Agricultural e.g. farmers’ rights, choice and quality of life, land
practice requirements
Economy Consumer benefit e.g. retail price
Producer benefit e.g. shorter term costs, yield or longer term value
added
Benefit to processor e.g. profitability
Socio-economic e.g. welfare of small farmers, substitutions for
impact developing countries
Society Individual impacts e.g. consumer choice, transparency, accessibility,
participation, pluralism
Institutional e.g. concentration of power, institutional trust,
impacts regulatory complexity
Social needs e.g. new opportunities, opportunity costs, misuse of
science, employment, quality of life
Ethics Fundamental e.g. animal welfare, taking care of nature
principles
Knowledge base e.g. hubris about scientific knowledge
The pattern displayed by the performance scores assigned by participants under

the various groups of criteria shows that the differences between perspectives are
not simply due to variations in the weighting placed on different criteria. Instead,
the picture yielded in this exercise is very similar to that illustrated above with
respect to risk assessment of energy technologies (Figure 7.3). Participants adopt a
variety of different ‘framing assumptions’, resulting in significant differences in
the scores assigned by different participants under the same criteria. This has
implications for conventional multi-criteria analysis, in which scoring is often
conducted by a separate body of experts, with an assumption that different value
judgements can be captured simply in the weightings. Despite these differences,
however, the patterns evident in the scoring do allow certain general conclusions.
For instance, under all perspectives the organic option tends to perform well under
environmental criteria. Also, the evident divergence between the patterns of
performance under health and environmental criteria challenge conventional
assumptions that these aspects of performance will necessarily be well correlated.
Notwithstanding the complications raised with respect to scoring above, the
patterns evident in the weightings do illuminate some interesting, if rather unsur-
prising, features. Perspectives drawn from the biotechnology industry and food
supply chain are conspicuous in their relative under-emphasis of the social and/
or environmental and safety considerations which are prominent under all other
perspectives. The perspectives adopted by government advisers have the distinctive
characteristic of being at the same time relatively narrow in scope whilst emphas-
ising environmental and safety considerations. The perspectives expressed by the
non-industry participants share a markedly lower emphasis on economic or
agricultural considerations.
Perhaps the most important outputs of the multi-criteria mapping procedure
are the patterns in the final rankings obtained under the various participants’
perspectives. Indeed, it is this picture which constitutes the ‘map’ referred to in the
name of the technique, reflecting the way that final results vary with starting
assumptions. Figure 7.5 displays the results obtained in the present exercise for the
ten participants from whom it was possible to elicit all the necessary quantitative
inputs. For reasons discussed more fully in the full report (Stirling and Mayer 1999)
the remaining two – G and H – expressed difficulties either in the scoring or
weighting process. Each chart represents the rankings elicited from a single
participant. The sequence of basic options is the same in each case, running top to
bottom from organic farming, through integrated pest management and conven-
tional intensive agriculture to GM crops with (respectively) segregation and
labelling, monitoring and voluntary controls. The extension of the horizontal bars
to the left reflects the overall performance of the option under that perspective –
the further to the left, the better the performance. The scale is in arbitrary subjective
linear units of performance. The length of the bars shows the difference between
the performance under optimistic and pessimistic assumptions under each
perspective, providing a useful indication of the relative importance of uncertainties.
The differences of ordering in the left- and right-hand ends of the bars show the
practical implications for option performance of adopting pessimistic or optimistic
assumptions under each perspective.
A number of features can be distinguished in this picture. First, it is clear that
there exist profound differences between the practical implications of the
perspectives adopted by different participants. In an orthodox regulatory appraisal
procedure (such as risk assessment) such variability would typically be concealed
by the emphasis on a single aggregated position. In the present analysis, on the
other hand, the idiosyncrasies of each individual position are clearly displayed.
Second, it is clear that the differences between rankings obtained under optimistic
and pessimistic scores are generally rather small compared with the differences
between perspectives. This suggests that it is not generally the technical dimensions
Academic scientists
C J
org org
ipm ipm
cnv cnv
gm1 gm1
gm2 gm2
gm3 gm3
Government safety advisers

E F
org org
ipm ipm
cnv cnv
gm1 gm1
gm2 gm2
gm3 gm3
Public interest
A D
org org
ipm ipm
cnv cnv
gm1 gm1
gm2 gm2
gm3 gm3
Industry
I B
org org
ipm ipm
cnv cnv
gm1 gm1
gm2 gm2
gm3 gm3
K L
org org
ipm ipm
cnv cnv
gm1 gm1
gm2 gm2
gm3 gm3
Each chart shows one perspective on the relative performance of six agricultural
options for producing oilseed rape. Individual perspectives are grouped and
colour coded according to their affiliation with four different constituencies. In
descending sequence on the vertical axis, the options are organic farming,
integrated pest management, conventional intensive agriculture and three
different strategies for the implementation of genetically modified crops. The
horizontal axis shows option performance and uncertainties on a subjective
interval scale (low to left, high to right).
Figure 7.5 Final individual rankings for the basic options

of uncertainty which are the key issue, but rather more intangible qualitative aspects
concerning the divergent interests, values and framing assumptions adopted by
different participants. Third, despite the variability in the evaluations of different
participants, there emerges a series of consistencies. With only a few exceptions,
the organic option tends to perform relatively well. Conventional intensive agri-
culture tends to perform uniformly badly. The voluntary controls regime for GM
crops is regarded rather poorly by industry, academic and NGO participants alike,
being viewed favourably only by the government advisers.
The final issue concerns the merits of diversification as part of a precautionary
response to uncertainty, ignorance and divergent values. Figure 7.6 shows the diverse
mixes of agricultural strategies that were favoured by each of the seven participants
who were able to respond to this part of the analysis. The partitioning of the pie
charts represents some arbitrary measure of importance, such as ‘share of output’
or ‘share of land in production’. For present purposes, the details do not affect the
broad-brush conclusions. What is interesting is that six of these participants –
drawn from all sides of the debate – favour the deliberate pursuit of some degree
of diversity. The remaining participant – a senior UK government safety adviser –
displays a uniquely high degree of confidence in the validity and robustness of
their own perspective. Together with the idiosyncrasies evident in the rankings
displayed in Figure 7.5, this is, in itself, an informative result. Taken as a whole,
this part of the multi-criteria mapping analysis might raise questions over the
extent to which R&D and regulatory policy making should be geared towards
active encouragement of a variety of techniques rather than assuming or
emphasising a single particular trajectory. Given the broad support expressed from
all sides of the debate for the principle of diversity, questions might also be raised
over how to treat options that display characteristics that militate against diversity.
In summary, then, this pilot multi-criteria mapping exercise permits a series of
quite subtle and specific observations to be made concerning the character of the
debate and the positions of the different protagonists. However, it also provides a
basis for some general normative conclusions concerning the options under
appraisal. These findings are all the more robust for being set firmly in the context
of systematic exploration and acknowledgement of the underlying variability and
dissent. In particular, in the present case, the picture is quite unequivocal with
respect to the relatively negative performance of conventional intensive farming
and voluntary controls on the use of GM crops. If equal attention is paid to each
perspective, then it is clear that the organic farming and integrated pest
management options tend to perform significantly better overall.
Conclusions
This chapter has sought to cover quite a bit of ground. It has been shown that
there are significant theoretical, methodological and practical grounds for concern
over the utility of orthodox ‘narrow risk assessment’ techniques in the regulatory
appraisal of risk. The frequently cited dichotomy between ‘science-based’ and
‘precautionary’ approaches to regulatory appraisal has been shown to be unfounded
Notes:
Participants were asked their opinion concerning diversity among options.
The results for each participant are shown as a pie chart displaying the
proportional importance of each option in their favoured mix of options.
The six basic options 1–6 are colour coded. The basic options are: 1
organic, 2 IPM, 3 conventional, 4 GM with segregation and current
labelling, 5 GM with monitoring, 6 GM with voluntary controls. The
additional options are: I7 GM with segregation, labelling, monitoring
and other restrictions; I8 GM with full assessment of impacts and need;
B7 GM with quality; K7 conventional and organic without GM; K8 mix
of GM, conventional and organic; J7 GM in regulated organic system,
J8 GM in IPM system; F7 GM with segregation, labelling and monitoring.
Figure 7.6 Diverse options mixes favoured by individual participants

and misleading. Indeed, it is argued that there exist strong scientific reasons for
holding the broader scope of precautionary approaches to be more consistent
with the scientific foundations of rational choice and probability theory than are
conventional ‘narrow risk assessment’ techniques. The imperatives both of science
and precaution can therefore be seen to pull in the same direction. The regulatory
appraisal of risk should become more systematic and broader in scope. In particular,
a set of criteria are developed concerning the need for greater humility, complete-
ness, transparency and participation in regulatory appraisal, with specific attention
to the comparison of different options (including mixtures of options), the consider-
ation of benefits and justifications and the systematic ‘mapping’ of the way in
which different framing assumptions lead to different pictures of performance.
In order to explore how these criteria might be practically operational, the
second part of the chapter examined as a case study a pilot exercise applying a
novel ‘multi-criteria mapping’ method to the regulatory appraisal of a genetically
modified (GM) crop. The results of this exercise are more complete than orthodox
risk assessment, in that they embody consideration of an unlimited array of issues.
They also include consideration of a wide range of different strategic alternatives
to the use of GM technologies. Equal attention is given to the justifications and
benefits, as to the impacts and costs associated with the various options. The
appraisal is inherently participatory, in that the results fully reflect the perspectives
of a wide range of different socio-economic and cultural interests. The relative
simplicity and auditability of the weighting and scoring approach retains a high
degree of transparency. A central feature is the focus on the systematic mapping
of the way in which specific results relate to specific assumptions in appraisal.
This, together with the explicit attention given to more diverse mixes of options,
embodies a greater degree of humility in the face of ignorance and pluralism than
is normally displayed in risk assessment.
In this way, it may be concluded that there seems no reason in principle why
conventional regulatory appraisal might not be adapted to better address the
imperatives both of science and precaution. The present pilot exercise was not
prohibitive in scale, being conducted by two researchers working part-time over a
period of 18 months. Although necessarily provisional, the consistencies displayed
in the results show that it is possible – despite the breadth of the process – to draw
some quite robust conclusions over the relative performance of options such as
organic farming, conventional intensive agriculture and voluntary control regimes
for the use of GM crops. Although the techniques of risk assessment continue to
offer essential tools in addressing specific aspects, there seem to be no compelling
theoretical, methodological or practical reasons for persisting in basing the entire
business of regulatory appraisal on these inherently constrained and limited
methods.
Acknowledgement
This chapter is based on an article published in the International Journal of Occupational
and Environmental Health (Vol. 6, 2000: 296–311), the material being reproduced by
permission.
References
Arrow, K. (1963) Social Choice and Individual Values, New Haven: Yale University Press.
Bezembinder, T. (1989) ‘Social choice theory and practice’, in C. Vlek and G. Cvetkovitch
(eds) Social Decision Methodology for Technological Projects, Dordrecht: Kluwer.
EPCAG/Biotechnology and the European Public Concerted Action Group (1997) ‘Europe
ambivalent on biotechnology’, Nature, 387: 845–7.
Fisher, E. and Harding, R. (eds) (1999) Perspectives on the Precautionary Principle, Sydney:
Federation Press.
Funtowicz, S. and Ravetz, J. (1990) Uncertainty and Quality in Science for Policy, Amsterdam:
Kluwer.
Grove-White, R., Macnaghton, P., Mayer, S. and Wynne, B. (1997) ‘Uncertain world.
Genetically modified organisms, food and public attitudes in Britain’, Lancaster:
Lancaster University, Centre for the Study of Environmental Change.
Hayek, F. (1978) New Studies in Philosophy, Politics, Economics and the History of Ideas, Chicago:
Chicago University Press.
Loasby, B. (1976) Choice, Complexity and Ignorance: An Inquiry into Economic Theory and the Practice
of Decision Making, Cambridge: Cambridge University Press.
Mayer, S., Hill, J., Grove-White, R. and Wynne, B. (1996) ‘Uncertainty, precaution and
decision making: the release of genetically modified organisms into the environment’,
briefings no 8, Sussex: University of Sussex, Global Environmental Change Programme.
O’Riordan, T. and Cameron, J. (eds) (1994) Interpreting the Precautionary Principle, London:
Earthscan.
O’Riordan, T. and Jordan, A. (2001) Interpreting the Precautionary Principle, 2nd edn, London:
Cameron May.
Raffensberger, C. and Tickner, J. (eds) (1999) Protecting Public Health and the Environment:
Implementing the Precautionary Principle, Washington: Island Press.
Renn, O. and Klinke, A. (1999) ‘Risk evaluation and risk management for institutional
and regulatory policy’, field study for ESTO project on ‘Technological risk and the
management of uncertainty’, Sevilla: Institute for Prospective Technological Studies.
Salo, A. (1999) ‘Technological risk and the management of uncertainty: the role of decision
analytic modeling’, study for ESTO project on ‘Technological risk and the management
of uncertainty’, Sevilla: Institute for Prospective Technological Studies.
Smithson, M. (1989) Ignorance and Uncertainty: Emerging Paradigms, New York: Springer.
Stirling, A. (1994) ‘Diversity and ignorance in electricity supply investment: addressing the
solution rather than the problem’, Energy Policy, 22(3): 195–215.
Stirling, A. (1997) ‘Multi-criteria mapping: mitigating the problems of environmental
valuation’, in J. Foster (ed.) Valuing Nature, London: Routledge.
Stirling, A. (1999) ‘On science and precaution in the management of technological risk –
final report’, Sevilla: Institute for Prospective Technological Studies.
Stirling, A. and Mayer, S. (1999) ‘Rethinking risk: a pilot multi-criteria mapping of a
genetically modified crop in agricultural systems in the UK’, report no 21, Sussex:
University of Sussex, SPRU.
Stirling, A. and Mayer, S. (2000) ‘Precautionary approaches to the appraisal of risk: a case
study of a GM crop’, International Journal of Occupational and Environmental Health, 6(4):
296–311.
Von Schomberg, R. (1998) ‘An appraisal of the working in practice of directive 90/220/
EEC on the deliberate release of genetically modified organisms’, Luxembourg:
European Parliament, Scientific and Technological Options Assessment Unit (STOA).
Wynne, B. (1992) ‘Uncertainty and environmental learning: reconceiving science and policy
in the preventive paradigm’, Global Environmental Change, 2(2): 111–27.
Wynne, B. and Mayer, S. (1999) ‘The release of genetically modified organisms to the
environment’, in EU 98, Copenhagen: European Environment Agency.
Consumer valuation and citizen deliberation 185
Part III
Deliberation,
participation and value
expression
186 Jonathan Aldred
8 Consumer valuation and

citizen deliberation
Towards a comparison
Jonathan Aldred
Introduction
There is now a relatively well-developed critique of the contingent valuation method
(CVM) for adding decisions about environmental problems. Theories of
deliberative democracy have been invoked which question the individualistic,
preference-based calculus of contingent valuation. A particular deliberative institu-
tion which has recently received much attention is the citizens’ jury (CJ). This is
taken here to involve around 16 ordinary members of the public, selected to
represent a cross-section of the local community. The jury comes together over a
period of usually four days to consider an important public policy question. The
members of the jury are briefed by and question expert ‘witnesses’, and discuss
the issues with each other in small and large groups, chaired by an independent
moderator. However, while the original impetus for the adoption of deliberative
institutions in environmental policy-making was undoubtedly dissatisfaction with
the CVM and the search for an alternative, a perhaps unexpected development is
that advocates of deliberative institutions no longer see them as direct alternatives
to the CVM. For instance, as the CJ methodology has been elaborated in more
detail, researchers have increasingly realized that the CJ ‘valuation exercise’ serves
different purposes, answers different questions, and makes different assumptions
about the individual participants than the CVM.
Although much of interest must be set aside in order to pursue any sort of
comparison between the CVM and CJ, it is still possible to explore the relative
merits of public discussion about the environment, vis-à-vis private reporting of
individual judgements. What does the presence of discussion per se add to the
environmental decision-making process? A number of putative advantages of
discussion are explored in detail, and in particular the critique of these deliberative
virtues, coming from both neoclassical environmental economists and the ‘rational
choice’ school of political theorists, is assessed. An aim has been to address not
just existing criticisms of deliberative institutions from this perspective, but to
anticipate and respond to criticisms which are likely to be made in the same spirit.
Rather than repudiate this perspective altogether (although I do not share it), I
show that the criticisms do not even succeed on their own terms. That is, I hope to
show that the arguments against discussion fail, in terms which the critics of
deliberative democracy must acknowledge.
188 Jonathan Aldred
The chapter begins by exploring the extent to which the CVM and CJ can be
meaningfully compared. It is argued that this is possible to a limited extent, by
treating both as explicit decision-making processes, with only CJs involving
discussions before decision, and restricting the comparison to an assessment of
epistemic advantages and disadvantages of that discussion. The assessment in this
chapter is made from the perspective of rational choice theory, because that
approach has been influential amongst critics of deliberative democracy. A final
section examines the desirability of using CJs to generative explicit decisions,
including a brief discussion of methodologies which attempt to combine the CVM
and CJ.
Comparing the CVM and CJ

Theories of deliberative democracy1 are often proposed as an alternative to
aggregative models, where given preferences2 are aggregated in order to determine
a collective decision. At the level of institutional design these two approaches will
be said to generate deliberative and aggregative institutions respectively (‘voting’
will be used as a synonym for aggregation in what follows). In the context of
environmental decision-making, the CVM might be regarded as an aggregative
institution, while CJs could be seen as one type of deliberative institution.3 But can
the CVM and CJ be meaningfully compared?
Certainly they should be regarded as competing alternatives only with caution.
The CVM and CJ answer to different institutional needs, cultural roles and social
contexts (O’Connor 2000). Moreover they are incompatible in an important sense
– they assume competing, arguably mutually exclusive, accounts of rationality.
The validity of one approach, then, would seem to cast doubt on the theoretical
underpinnings of the other. But this tension does not rule out a comparison. Public
bodies charged with environmental policy are increasingly required to justify their
policy decisions through public consultation, and see the CVM and CJ as alternative
methods (among others) for doing so (Jacobs 1997a). So the motivation for pursuing
a comparison is clear. Public bodies do compare the CVM and CJ, and it would be
inadequate to answer a decision-maker seeking advice on the appropriate
consultation process with the reply ‘it depends on the kind of consultation you
seek’. A less question-begging response would be based on an assessment of which
consultation process is superior. A central claim of this chapter is that the
comparison involved in such an assessment is coherent, and that much of what is
essential to the comparison can be captured by a relatively straightforward question:
What are the relative epistemic merits of preceding decision-making by discus-

sion, compared with mere private reporting of individual judgements? (Q )
It is not immediately obvious that the comparison specified by this question maps
neatly and helpfully on to that between the CVM and CJ. A number of features of
the comparison must be clarified.
Decision-making processes?
(Q) implies the end point of the process is a decision. There are two separate issues
at stake here. Roughly speaking, the first may be regarded as the question of whether
the input to the process is open or closed; the second issue is whether the output is
open or closed. In both cases, there is no dichotomy but a continuum of varying
degrees of openness. The process input is said to be closed if the participants’
have no control over the inputs to the process. These inputs are principally the
task the participants are set, and the information they are given at the beginning
of the process, or are able to obtain during it. The more open the process input,
the more control participants have over their role, and in particular the subject
matter of their thinking, during the process. The process output is said to be closed
if the outcome of the process is narrow and determinate. The more open the
output, the more it admits diverse interpretations, and the less structure it will
have. While these notions of openness and closedness of outputs and inputs seem
intuitive, they are hard to define clearly in isolation. A few examples should help to
clarify them.
The following process outputs are (arguably) progressively more open: decision,
valuation, recommendation, competing recommendations, process report
suggesting consensus but making no recommendations, process report suggesting
dissent. Clearly the openness of the output will be influenced by the openness of
the input – for instance, if the participants are set a relatively closed task in being
asked to reach a decision. This example might suggest that a relatively closed
input entails a closed output. However this is mistaken for two reasons. First, in
the example, the participants may be asked to reach a decision, but the process
input may be very open in other respects, implying a relatively open input overall.
Second, the participants may simply ignore the exhortation to reach a decision.
The openness of inputs and outputs are in general independent of one another,
and only the former concerns the degree of control the participants possess. A
very strict requirement to reach a decision is likely to be followed by a closed
output; but it is the closed input rather than the closed output which is responsible
for the participants’ lack of control; in contrast, a decision may emerge from a
process in which the participants were given an extremely open, sketchy task, but
chose to narrow this down into a particular decision problem. Another example
of confusion between open inputs and outputs is captured in the dubious claim
that if a process requires only a report of deliberation rather than a decision, then
the participants have more control. Certainly a decision is a less open output, but
this does not entail a less open input. In other words, there is no a priori reason why
the requirement to produce a report on a specific task gives participants any more
control than the requirement to produce a decision. It does, however, demand less
agreement among them.
While the process input is principally classified according to control, the focus
in classifying the process output is agreement – a closed output will demonstrate
more agreement among the participants.4 In a later section, this feature will be
used to distinguish different types of CJ, but the more immediate concern is the
190 Jonathan Aldred
comparison between the CVM and CJ. The CVM methodology presumes and
necessitates a closed input in which the set of policy alternatives, among which a
single choice is to be made, is exogenously determined. CJ methodology makes no
such assumption. With an open input, the set of alternatives is not well defined,
perhaps because there is no fact of the matter, or established understanding,
regarding the nature of the problem. The public authority may present the CJ
with an open input because the authority’s understanding of the problem does not
permit closure. Part of the resolution of the problem is in framing the decision to
be made – ‘asking the right questions’. While the open input approach may be the
richer and more realistic characterization, it remains incompatible with the CVM
methodology. If a comparison of the CVM and CJ is to be pursued, a closed input
must be assumed.
Another difficulty for the comparison between the CVM and CJ is that they
typically have different outcomes. A CJ provides a report, while the CVM gives a
valuation. Since comparing a report and a valuation directly is impossible, it is
helpful to construe the comparison between the CVM and CJ as one between
different decision-making procedures. So, in the absence of unanimity, deliberation
in the CJ is terminated with a voting or aggregation procedure; it may then be
compared with ‘CV-in-CBA’, which also yields a decision, rather than the CVM
alone.5 Once both the CVM and CJ are defined to stretch as far as a final decision,
a direct comparison is possible. However adopting this approach is unnecessary in
order to compare the CVM and CJ indirectly. Referring again to (Q), it is still
possible to examine whether, and in what respect, discussion affects participants’
views. But the comparison is clearly much sharper if the effect of discussion can
be traced through to a decision.
To sum up, the CVM and CJ are regarded here as decision-making processes
in two respects: participants in both are set the task of deciding between exogenously
specified options (closed input), and both processes reach a decision (closed output).
These features are clearly exhibited by some, but not all, CVs and CJs in practice.
For the purpose of tractable comparison, they will be assumed in what follows,
although I shall return to the possibility of open input and output processes at the
end of the chapter. This approach also has the virtue of consistency with the more
abstract comparison between deliberative and aggregative institutions. Insofar as
they have been compared, both are assumed to terminate with a decision (Elster
1998a). The comparison between deliberative and aggregative institutions is then
a rather strange one – between aggregation preceded by discussion, and aggregation
alone. Most advocates of deliberative democracy would of course reject the claim
that the only difference between deliberative and aggregative institutions is that in
the latter, voting is not preceded by discussion. But this approach reflects the ‘thin’
view of deliberation, as discussion, adopted here.
Deliberation and discussion

While CJs are usually said to involve deliberation rather than discussion, (Q) adopts
the latter term. This is because deliberation is here understood simply as discussion
(Fearon 1998).6 There are a number of reasons for this approach. First, the meaning
of discussion as a concrete activity is intuitively clear. Second, CJs involve a relatively
small number of people meeting to make a decision. Discussion is not obviously
applicable to society at large, whereas deliberation is often urged on this scale.
Third, one may deliberate with oneself, while the concern here is with public
discussion. Finally, and most importantly, deliberation carries much theoretical
baggage. For example, many deliberative democracy advocates distinguish the
‘diagnostic’ approach of preference discovery in the CVM from the ‘constructive’
approach of judgement creation in CJ (Sagoff 1998: 221). But this distinction is
not essential to the comparison between deliberative and aggregative institutions.
Aggregation procedures as defined by social choice theory are formal structures
alone, with no requirement that their inputs be preferences rather than judgements.
The picture of the agent constructing her judgements afresh in response to the
information given to her is compatible with an aggregative institution. Conversely,
the possibility of a utility maximizer entering into discussion with other utility
maximizers, is compatible with participation in a deliberative institution.
Epistemic merits
The approach taken here might appear to bias the debate heavily in favour of
deliberative methods; the comparison between aggregation preceded by discussion,
and aggregation alone, is, to repeat, a strange one. But the debate is far from over,
since advocates of deliberative democracy must claim:
decisions reached after discussion will be better decisions than those reached
without it (B)
As it stands, the term ‘better decision’ is so vague as to be almost meaningless. Yet

the extent to which many deliberative democracy advocates regard (B) as obviously
true suggests that there is a strong intuitive notion of ‘better decision’ in circulation.
Despite the vagueness, claims such as (B) are common. The list in Table 8.1 is
adapted from Elster (1998b: 11).
Table 8.1 Possible advantages of discussion

Positive features
Epistemic Encourages an appeal to principles
Pools private information
Lessens or overcomes the impact of bounded rationality
Processual Legitimizes the ultimate choice
Is desirable for its own sake
Makes for a larger consensus
Improves the moral or intellectual qualities of the participants
Distributional Makes for Pareto-superior decisions
Makes for better decisions in terms of distributive justice
192 Jonathan Aldred
This chapter defends (B) in regard to the content of the decision rather than
the process by which it is reached; it appeals to the epistemic virtues of discussion,
rather than any processual or distributional merits of the process. There are three
reasons for this focus. First, it is difficult to see how the legitimating powers of
discussion, its educative or moral value to the discussants, or its desirability per se,
could be defended without some prior commitment to the virtues of deliberative
democracy. Thus, in the spirit of avoiding such commitments in order to engage
with the critics of deliberative democracy, the focus must be on the epistemic
merits of deliberation, seen as practical consequences of the concrete activity of
discussion.
Second, another sense in which assessment of these epistemic virtues engages
with the critics of deliberative democracy is that it shifts the debate on to the
critics’ favoured methodological ground. The list above interprets the advantages
of discussion in a way that facilitates analyses drawing on rational choice theory
and game theory.7 While this list represents one perspective on the advantages of
discussion, there are at least three others, which will not be discussed further here:
1 Recent work by deliberative democrats has focused on the feasibility rather

than the legitimacy of the deliberative ideal, examining various mechanisms
which might generate true ‘deliberative uptake’ (Bohman 1996, 1998;
Gutmann and Thompson 1996; Macedo 1999).
2 There is empirical research in psychology, particularly social psychology, which
supports many of the theoretical claims about discussion made here. While
rational choice theory appeals to homo economicus to justify agents behaving in
ways which generate beneficial features of discussion, social psychology
explains these same benefits using an altogether different model of the agent.
There is evidence of non-self-interested behaviour increasing as a result of
discussion (Orbell et al. 1988, 1990), and violations of homo economicus more
generally (Lane 1991; Liebrand et al. 1992). For a synopsis, see Sally (1995).
3 There has been tentative discussion of the framework that social choice theory
might provide for exploring how discussion might help consensus be reached
(Buchanan 1954a, 1954b; Hylland 1986; Miller 1992; Sen 1995b, 1997; Aldred
2000).
An obvious drawback of viewing deliberation through the lens of rational choice

theory is that much of the richness of the deliberative account of decision-making
is inevitably lost. Most obviously, the processual advantages appear to be overlooked.
However, these benefits tend to be by-products of the epistemic virtues (Elster
1986b, 1995; Fearon 1998; Gambetta 1998). Pursued directly, they can be self-
defeating. For instance a main aim of CJs for many advocates is to aid in rebuilding
the trust of ordinary citizens in government and political processes (Coote and
Lenaghan 1997). Telling the participants in a CJ that it is a trust-building exercise
will not achieve that effect. So a third reason for the focus on epistemic merits is
that a policy-maker attempting to secure the processual advantages directly is
probably doomed to failure.
Appealing to principles8
At the heart of deliberative democracy are two claims: an empirical one, that
public discussion will encourage a particular mode of justifying arguments (Goodin
1986), and a theoretical claim, that this justificatory requirement will have a
beneficial effect on the quality of the final decisions of a deliberative institution.
The justificatory requirement is that arguments invoked during discussion cannot
be narrowly concerned with self-interest, but must be acceptable to all. In short,
arguments must appeal to widely shared principles. There are two broad reasons
why we might expect this justificatory requirement to be satisfied in practice (Elster
1995; Fearon 1998).
Strategic reasons
A speaker (S) might advance justifications which appeal to the public interest for
what have been termed ‘strategic’ reasons (Elster 1995). The basic idea is
straightforward: if S invites others to choose proposal P merely because P serves S’s
private interest, few will be persuaded. Johnson (1998) is characteristic of the
deliberative democracy critics in seeing the presence of self-interested motives behind
apparently impartial arguments as crucially undermining the normative weight, which
deliberative democracy advocates attach to the argumentative content of deliberation.
The epistemic justification of deliberative democracy, which appeals to ‘the force of
the better argument’, is subverted by the self-interested motives of participants to
the deliberative process. The objection runs that listeners are swayed by arguments
which have no objective merit but have been proposed simply because they serve the
speaker’s interests. Johnson concedes that ‘psychological pressures’ generated by the
process of discussion might over time make speakers sincerely come to embrace a
principle which they earlier invoked for self-interested reasons alone – one example
of Elster’s ‘civilizing force of hypocrisy’ (Elster 1995). But this possibility does not
nullify the objection – Johnson (1998: 172) argues that such a mechanism which
‘first induce[s] actors to engage in self-censorship and eventually cause[s] them to
embrace the resultant views … does not so much generate a “reasoned agreement”
as induce a conformity that is at once rather shallow and normatively suspect’. The
reasons for suspicion are not stated, although in a note Johnson (1998: 172 n. 47)
worries about the autonomy of actors subject to such mechanisms and quotes Elster
(1986a: 113) approvingly: ‘Dissonance reduction does not tend to generate
autonomous preferences’. He summarizes his critique by implying that deliberative
democracy advocates must demonstrate both ‘(a) that impartial claims are, even
tacitly, quite widespread in public debate and (b) such claims are animated by
mechanisms other than … strategic calculation and attendant psychological pressures’
(Johnson 1998: 172). It is argued here that the case for deliberative institutions such
as CJs need not satisfy (b) and that there are reasons to expect (a) will be satisfied, at
least in the ideal case.
Regarding (b), a claim can be impartial, in the sense of appealing to widely
shared principles or interests, but still be ‘animated by strategic calculation’.
194 Jonathan Aldred
Arguments that appeal to the public interest, to widely shared principles, have
normative force not because they purportedly reveal the noble character of the
speaker, but because they are much more likely – in the ideal case, the only way –
to secure the consent of the listener. The motives of the speaker are not in general
relevant to winning this consent. For example, in a debating contest S may be
acting as a ‘devil’s advocate’, but this does not in itself undermine S’s arguments,
even if the listeners know of S’s rejection of S’s arguments. Of course motives
matter: if the listener (L) is aware that S’s self-interest is best served by the action
urged by S, then deliberative democracy recommends (and expects) that L will
examine S’s claim that the public interest is served by the action all the more
carefully. L will also be sceptical of emotional appeals to the public interest, or
uncorroborated factual claims, in this context. But L should in general be able to
establish, independently of S’s motives, whether the action urged by S serves the
public interest – and in particular, whether it does so for the reason(s) advanced in
S’s argument.
Johnson also misrepresents Elster’s ‘civilizing force of hypocrisy’. While
advocates of deliberative democracy are not always clear, the essential point is
that agents, through discussion, become aware of new ideas and the positions of
others. In particular, as S advances arguments in which S has no conviction in
order to persuade others, S may come to realize their merits over opposing argu-
ments S privately held at first – a situation I shall term devil’s advocacy conversion.
It occurs directly through the recapitulation of arguments to others in order to
persuade them: finding new perspectives which enjoy common support, realizing
how the position can be strengthened further, or simply coming to understand it
more fully.
Returning to Johnson’s claim (a): as already suggested, whenever S seeks to win
consent, whether for self-interested reasons or otherwise, S will seek to advance
impartial arguments – arguments showing how the public interest is served – because
by showing how diverse private interests are simultaneously served, an argument
will command widespread support. This motivation for appeals to the public interest
would not seem to apply to majority groups: a subset of deliberators, if they form
a majority of the whole group, will be free to make blatant appeal to their collective
self-interest, that is the collective interest shared by the members of the subset but
not shared by all. Yet there is anecdotal evidence that even majorities appeal to a
broader public interest than their own constituents. This suggests there must be
other reasons besides strategic ones for justifications in terms of the public interest.
Avoiding appearing selfish

A second reason why S might advance justifications which appeal to the public
interest is that S will not wish to appear selfish to other members of the deliberative
institution, whatever S’s true motives are. A related reason why a speaker or group,
particularly a majority group, might not wish to appear selfish is reciprocity.
Although the majority group may be able to command majority support in the
current argument by appealing to (its) self-interest, in a later argument (which
refers to the same final decision) the members of the majority group may be on
opposing sides. Thus the members of the majority group may wish to adhere to
strictly public-interest justifications now in the hope of ensuring only public-interest
justifications from other majority groups in the future.
The desire not to appear selfish is arguably less significant than the strategic
motive for public-interested justifications in discussion, for two reasons. First, the
desire not to appear selfish is parasitic upon the existence of some widely accepted
account of ‘unselfish justifications’ – it depends on the possibility, or at least the
belief in, genuinely impartial or public-interested arguments after all. Second, it
may have less relevance in problematic cases of environmental decision-making.
For the embarrassment or shame a person feels from appearing ‘selfish’ is shorthand
for shame from appearing to ignore a moral norm. However in the environmental
case there will often be clashing norms, value conflicts and moral dilemmas where
no matter which outcome is chosen, it will be wrong in the sense of a norm being
broken (Nagel 1979; Williams 1979; Gowans 1987). More generally, in such cases,
the judgements ‘public-interested’ and ‘self-interested’ are at the least ambiguous,
and often wholly inappropriate.
Pooling private information

The information might take the form of, or refer to, facts, preferences, or beliefs. It
is often noted that discussion permits participants to express intensities of preference
as well as mere orderings. In response, it may be argued that the CVM is also able
to capture intensities of preference, subject as always to the caveat that preferences
(and judgements) can be adequately represented through the private monetary
format. However, while the CVM can capture preference intensities, it is surely
unable to capture the subtle nuances of natural language, or even all the nuances
of ‘natural language expressions of preference’. This may appear a misleading
comparison, because, except when the participants are unanimous, a deliberative
institution will ultimately be limited to what can be expressed through a vote. But
the point is that nuanced information – both ‘preference’ and ‘factual’ – can be
shared with others before a vote is conducted. This has obvious benefits, since I
may have good reasons to change my preference once aware of your information,
such as altruism, reciprocity, or trust in your judgement. It is not possible with the
CVM.
For critics of deliberative democracy, the chief difficulty with these arguments
is that they overlook the strategic incentives an agent may have to lie or misrepresent
information in an attempt to sway the preferences of others (the so-called ‘cheap
talk’ models, e.g. Farrell and Gibbons 1989; Austen-Smith 1990a, 1990b; Krehbiel
1991). Of course most voting procedures, including consumer valuation, are
susceptible to strategic manipulation by voters exaggerating their preferences. But
this is irrelevant here, since a deliberative institution will typically be terminated
by voting. The key question is: ‘Given preference diversity and thus incentives to
misrepresent, is a group better off simply voting or discussing things before voting?’
(Fearon 1998: 47).
196 Jonathan Aldred
Much of the ‘cheap talk’ literature offers a very nihilistic answer, claiming that
because ‘talk is cheap’ – reported private information is unverifiable and/or lies
cannot be punished – no credible communication can occur (Riker 1982; Austen-
Smith and Riker 1987; Austen-Smith 1990a, 1990b). The idea is that, because L
knows that ‘talk is cheap’, L will ignore S’s communications, regardless of whether
S in fact has an incentive to lie. (More precisely, L will not engage in the costly
business of establishing S’s incentives or verifying communications.) There are a
large number of problems with the game-theoretic cheap talk models used to
justify the nihilistic conclusion (Mackie 1998 offers a good critique), but three
main ones are that the models are too static, the results are dramatically weakened
once repeated interaction is assumed, and equally weakened once multiple speakers
and listeners are modelled. The models are too static because they ignore the
agents’ responses to the credibility problems. If S knows that L will ignore S’s
private information then S will use different information, information that is public
but overlooked or private but verifiable by an independent third party, because the
original information has no persuasive power. Moreover L has an incentive to
help S in S’s attempt to reveal information credibly. Repeated interaction improves
the credibility of communication for obvious reasons: ‘When Aristotle was once
asked, what a man could gain by uttering falsehoods; he replied, “not to be credited
when he shall tell the truth”’ (Johnson 1963, quoted in Mackie 1998: 85). With
multiple speakers and listeners, the formal results are complex, but confirm
intuitions such as multiple speakers with conflicting interests having an incentive
to correct each other.
Due to the focus in the cheap-talk literature on the difficulty of credibly commun-
icating any private information, and the idea that L will ignore all communications,
the extent to which L might filter the transmission of misleading information is
overlooked. The distinction between communication which is not heeded, and
misleading communication which is acted upon, is an important one. Only in the
latter case can there be a putative argument against discussion. In the case of
cheap talk and other attempted strategic manipulation, how might deliberative
institutions prevent misleading communication? Put simply, they should be
constructed to allow participants to heed Aristotle’s advice, by paying particular
attention to the credibility of a speaker when evaluating the speaker’s claims:
That credibility has both epistemological dimensions – the speaker must have
good sense and be reliable in the formation of judgements – and ethical
dimensions – the speaker must have the moral character that allows us to trust
their utterance and there must be grounds for believing that they are not
inclined to impart falsehoods to their audience.
(O’Neill 1997)
The Aristotelian account encourages listeners to recognize their inability to evaluate

certain claims in the absence of specialist training, and concentrate on discerning
on whose authority they can rely for such an evaluation. The emphasis is on assessing
the character of the speaker, rather than the speaker’s claims (O’Neill 1993: Ch. 8).
Little more can be said in general, but in terms of institutional design, the
following features of a CJ may be helpful: a neutral facilitator selects witnesses,
factual information is conveyed through witnesses, witnesses are able to criticize
each other’s statements, more time is usually allocated for jurors to question
witnesses than for their statements, and jurors are able to recall witnesses at a later
point in the proceedings if necessary. In a CJ that I facilitated (Aldred and Jacobs
2000), jurors were able to pick up on inconsistencies in witness presentations. The
jurors were choosing between various land management options for wetland areas
in East Anglia, UK. In response to a question about its cost, one advocate of a
particular option claimed that the option would be ‘largely self-managed’ (reducing
labour costs). But later the same witness maintained that the option would ‘create
significant numbers of jobs’. The advocate had probably not actively tried to mislead
the jury, but the jury were able to force the advocate to be more specific about the
advantages proposed. In sum, although the evidence so far is limited and there are
obvious problems, it seems that CJs do at least have the potential to allow listeners
to assess the credibility of speakers.
As well as overlooking many of the cognitive difficulties associated with elaborate
schemes for strategic manipulation, critics of deliberative democracy, like the
rational choice theories on which they draw, also tend to assume that more inform-
ation is always better. But without some constraint on the revelation of private
information, perhaps through discussing rules of relevance or time limits,
‘information overload’ will be a danger for many decision problems (Fearon 1998:
65 n. 9). Although an aggregative institution might face overload problems because
the voting instrument provides the voters with too much information about the
alternatives, a deliberative institution, with (at least in principle) open-ended
opportunities for private information revelation, appears more susceptible to these
difficulties. For a supporter of deliberative democracy, the obvious response to this
challenge, besides acknowledging the necessity of rules of relevance and time limits,
is to demonstrate how discussion may help participants to overcome the broad
class of ‘bounded rationality’ problems, including information overload.
Bounded rationality and new perspectives

A central feature of ‘bounded rationality’ is that when making decisions our
reasoning and calculating capacities are both limited and prone to error (Simon
1978). Even if a decision was in the hands of a single individual with dictatorial
powers and all relevant information, the decision-maker might begin by discussing
the alternatives with another, in order to clarify their nature and perhaps suggest
new solutions altogether. This suggests two benefits of discussion, namely the
sharing of arguments, ideas, and understandings which were previously familiar
to only a subset of participants, and the generation of new ideas, etc. which were
previously familiar to none of the participants (Fearon 1998: 49–50). So for complex
decisions, will many heads be better than one?
The subject matter of the decision is crucial. For decisions where the relevant
arguments can be readily evaluated in a way that would command the assent of all,
198 Jonathan Aldred
then in general it seems likely that many heads will be better. Examples include a
player – or team – aiming to win at chess (Fearon 1998: 50), or an individual talking
to a friend when trying to decide upon the purchase of some complex financial
product such as a pension or insurance. A chess champion or financial expert would
be best, but much can be gained through discussion per se. In these cases, certain
arguments can be shown to be objectively superior to others, and discussants will
recognize and accept their arguments being overthrown by stronger ones. Excepting
pathological cases, there is little possibility of social pressure forcing discussants to
abandon stronger arguments in favour of weaker ones. However, this case clearly
does not apply to most environmental decisions. Rather than cleanly resolving the
complexities of the decision in the manner of the chess example, discussion may
become blocked by intractable disagreements arising from value conflicts.
As described so far, the sharing of arguments and ideas appears to be simply
another form of private information pooling. But this interpretation is mistaken,
due to the treatment of uncertainty in the Bayesian framework, which underpins
rational choice theory. For the Bayesian, there is no such thing as ‘pure’ uncertainty;
rather, all uncertainty is ‘risk’ where agents always know the probabilities (subjective
or otherwise) of all possible alternatives. Note that as a prerequisite on this account,
agents must know the complete list of possible ‘answers’ to some question of uncert-
ainty. This may be adequate with respect to the weather (tomorrow it could be
fine, wet, windy or snowy, etc.) and at least coherent with respect to some forms of
scientific uncertainty. But it is incoherent with respect to the arguments, interpret-
ations and perspectives of another. Arguments are either known or unimagined,
rather than on a list of possibly existing arguments. Thus if L hears a new
(unimagined) argument from S it may have a profound and creative effect, whereas
if S reveals private information, then according to rational choice theory, it will
already have been considered as a possibility by L.
So far in this section S and L have been assumed to have common interests.
How can I trust the arguments of another with divergent interests? The consider-
ations mirror those of the private information case, although particularly here, a
speaker’s communication may not have the intended effect. S may intend to mislead
with a bogus argument, but simply because new arguments are generated by creative
leaps of the imagination, S’s communication may trigger as a by-product a new
idea or argument which serves L’s case. Or S may not intend to mislead but never-
theless present a partial argument, because bounded rationality makes S simply
unaware of the argument’s limitations. This latter case is common with environ-
mental problems where the participants bring a number of differing perspectives
and interpretations to the deliberative institution. Indeed the idea that CJ
participants are exposed to a much wider range of points of view than they would
be through responding privately to a CVM survey has been a key claim of CJ
advocates (Jacobs 1997b). In terms of the framework here, this case is one of
divergent interests without intent to deceive. The arguments and interpretations
offered by any speaker will inevitably be partial and incomplete because they reflect
that speaker’s particular perspective and experiences. However the partiality is
neither strategic, nor necessarily conscious, and notwithstanding the presence of
divergent interests, a speaker will be willing to modify and amend her perspective
if she seeks to understand, communicate with, or persuade others.
Thus the distinction between the sharing of arguments and the pooling of
private information applies a fortiori when there are divergent interests. When L
hears S claim to report S’s private information, L can always use S’s speech to
update L’s probability estimates regarding S’s true private information, because L
can conceive of what it might be. In contrast, it does not matter whether S has
actually entertained, let alone expressed, an argument or idea once L has conceived
of it, and if L has not conceived of it, then L cannot attach a probability to its
existence. Moreover an argument by S can often have indirect creative benefit for
L regardless of (whether L knows) S’s interests. In sum, the sharing of arguments
is more than simply the pooling of private information when new perspectives are
introduced.
Bounded rationality considerations are important for discussion in another way:
they suggest that the standard for a better argument should not be overstated. An
argument is better just in case it can be recognized as superior to existing arguments
by all the deliberative institution participants, that is impartially superior. It need
not be flawless or objectively compelling; rather, in the presence of bounded ration-
ality the introduction of a superior but flawed argument by one participant should
be seen as beneficial not because it is objectively compelling but because it demands
a revised response from those who disagree. This point hints at one possible explan-
ation of why critics of deliberative democracy are unwilling to accept that the
quality of collective decision-making can be improved as a result of arguments
advanced for purely strategic reasons. For such critics, it may be that there can be
no impartially superior arguments, or at any rate it is not possible for listeners to
identify them when the arguments emanate from speakers who may be partial. A
related but distinct explanation, at least for those critics influenced by rational
choice theory, appears to be that the possibility of new ‘altruistic’ arguments is
widely denied. Either an argument made by S to L benefits only S, or if it benefits
both S and L, or L alone, L would already be aware of it. The critics of deliberative
democracy tend to overlook the possibility of bounded rationality preventing L
from being aware of information or arguments which serve L’s case.
In summary, the last three sections have examined some key claims regarding
the epistemic virtues of discussion in the light of criticisms from a rational choice
theory perspective. At the level of this abstract comparison, it has been argued
that the claimed epistemic merits withstand this critique. One possible virtue of
the approach here is that it seeks to persuade neoclassical environmental economists
of the merits of deliberative democracy in terms which they must acknowledge.
But a drawback has been the necessity of assuming the CJ as well as the CVM
yields a closed output. While the assumption of a closed output – roughly, the
requirement that the process yield a ‘decision’ – has been useful for comparing the
CVM and CJ, it is important to assess the merits of imposing this requirement on
the process. This is the subject of the next section, and it turns out to be indirectly
relevant to the CVM/CJ comparison after all, because it permits evaluation of
some methods which seek to combine both the CVM and CJ.
200 Jonathan Aldred
Agreement and deliberative valuation
For many advocates of CJs and other deliberative institutions, the acid test of
the ‘success’ of some deliberative institution is whether the preferences of
participants changed during the deliberations – crudely, whether the deliberation
‘made a difference’ (see Chapter 12). Probably because the notions of ex ante and
ex post preference in this context are controversial and little understood, this test
appears to have received no explicit discussion in the deliberative democracy
literature, and there is no consensus on how it should be operationalized. However,
without ex ante unanimity, CJ participants will not reach agreement on the decision
facing them unless at least some participants’ preferences have shifted. Thus the
fact of agreement usually implies a change in preferences. While agreement
may be a sufficient condition for preference change, it is clearly not necessary.
Too much of the abstract deliberative democracy literature has assumed that it
is – that genuine deliberation, involving preference change, guarantees agreement
on the decision faced (Bohman 1998). No assumption that discussion per se will
generate agreement has been made here; the epistemic merits of discussion
discussed above all tend to encourage, but not ensure, agreement. Rather, the
assumption of a closed output has been made on the basis that discussion will
usually be followed by a vote. But should the output of a CJ be closed by the
imposition of some aggregation mechanism? In the rest of this chapter, I shall
briefly consider two related questions:
1 Should some procedure be imposed on the CJ which ensures a closed output,

or should the participants be left to reach as much or as little agreement as
they feel able to?
2 If a closed output is deemed desirable, what form should it take? In particular,
should it be a qualitative or quantitative output?
Beginning with the first question, the problem is not the emergence of a closed
output per se, but the extent to which this is forced. The requirement that a closed
output be reached, is a form of closed input which reduces the control the
participants have over the process. In contrast, theories of deliberative democracy
generally encourage as much openness as possible. Deliberative democracy
advocates argue that, by demanding a decision on a specific ‘charge’ or question,
the body which commissions a CJ may exclude important issues, information and
perspectives. At best, the CJ may be biased, while if the exclusion is intentional
then the CJ is manipulated. At the other extreme:
Experiments with complete juror control of the process have found that
participants, in the initial stages, do not have enough of an overview on a
subject to deal competently with setting the charge, agenda organisation or
witness selection.
(Smith and Wales 1998: 12)
Thus there is a trade-off between the openness of inputs and deliberative
competence. How this trade-off should be resolved depends on the role attributed
to the CJ. One view holds that the role of a CJ is simply to ‘read off ’ the given
opinions or tastes of the jurors. On this view, it may be plausible to separate entirely
the framing of the question from opinions about its answer. In contrast, supporters
of CJs usually interpret their role as centrally involving matters of ethical concern,
where the neoclassical economists’ mantra de gustibus non est disputandum is irrelevant.
Jurors must debate their judgements of ‘right’ and ‘wrong’ rather than simply
express their tastes of ‘like’ and ‘dislike’, and a central part of this debate is to
explore the issues at stake. It seem indefensible that this exploration be subject to
an exogenous constraint.
Part of the process of reaching the ‘right answer’ in environmental problems is
framing the question correctly, just as it is for, say, abortion or capital punishment
issues. For such problems, it seems that any defence of imposing a specific question
on a CJ must rely on roughly the claim that ‘the commissioning body knows best’.
Whether this claim is based on the view that the commissioning body is more
experienced at making complex problem-framing decisions, or better informed,
or more public-interested than the jury, it faces many difficulties. An obvious one
is that a commitment by the commissioning body to a closed input reflects a lack
of trust in participants. This may trigger a corresponding mistrust among
participants – in particular a suspicion that the commissioning body will ‘cherry-
pick’ recommendations from the report of the CJ as it suits them (Aldred 1998;
Aldred and Jacobs 2000). It is difficult to overemphasize the need for jurors to
maintain trust in the process if a CJ is to be successful (Armour 1996; Crosby
1996, 1999; Coote and Lenaghan 1997; Smith and Wales 1999; Ward 1999).
A less obvious difficulty is an apparent tension between the claim that ‘the
commissioning body knows best’ and rational choice theory. For on the broader –
and arguably more consistent – account of rational choice theory which embraces
public choice theory, it is not just ordinary market agents who act as if homo
economicus, but politicians and bureaucrats too. Politicians are neither better
informed, nor more public-interested, than anyone else. Within this tradition, it is
very difficult to defend any claim that ‘the commissioning body knows best’, yet it
is from within this tradition that environmental economists emerge who most
strongly argue in favour of requiring closed outputs to the CVM and CJ.
In summary, reconciling the demand for a closed output with a view of CJs as
calling forth ethical judgements of right and wrong, proves difficult. Suppose
however that such a reconciliation can be achieved, or that this view of CJs is
rejected, and that in answer to the first question, a closed output is deemed desirable.
Moving to the second question, should this output be quantitative or qualitative?
A common early view of the role of CJs was that they could offer a useful space
for discussion before proceeding to a standard CVM exercise: the CJ would finish
with the now well-informed participants each privately filling in a CVM
questionnaire, a close variant of the ‘focus group CV’ (Brown et al. 1995; Aldred
1996; Jacobs 1997b; Brouwer et al. 1999). Both Ward (1999) and James and Blamey
(2001) are more radical in seeking an average valuation from the jury not as a
202 Jonathan Aldred
statistical average of individual CVM responses, but as the direct response to asking
the jurors collectively for their valuation. They both hope to avoid some of the
problems of imposing forced agreement: they hope that an average valuation will
emerge naturally as a kind of quantitative consensus from group discussion, rather
than being enforced by means of a vote. Nevertheless, they both limit jurors’ control
in the sense discussed above; they require a closed output, to be achieved by some
kind of voting procedure, if consensus does not emerge. This new approach, termed
‘deliberative monetary valuation’ (DMV) (see the introduction to this volume),
seeks to combine the advantages of discussion with the merits of a valuation output.
Since DMV aims to be more readily comparable with the CVM than conventional
CJs, with their qualitative output, it will undoubtedly attract much attention.
In practice, probably the main advantage of the DMV approach is the appeal
of quantitative measures in contemporary policy discourse in market economies;
they have overtones of scientific rigour and accuracy. However, at a theoretical
level, the DMV approach appears to have two main virtues over a conventional
CJ. They mimic two of the arguments made for favouring the CVM over CJ. First,
DMV captures preference intensities. If, as seems likely, the deliberation in both
CJ and DMV processes does not generate unanimity, then some kind of aggregation
mechanism must be employed to generate a decision. If the aggregation procedure
used in DMV at this point measures preference intensities, then it is argued to be
superior because it captures more information about individual preferences. Second,
this additional information enables DMV to avoid standard Arrovian results in
social choice theory under which there exists no mechanism to aggregate
preferences which also satisfies certain weak and plausible properties (Sen 1970).
The threat to post-deliberation preference aggregation posed by Arrovian
impossibility is arguably a weak one – indeed, as noted above, the language and
concepts of social choice theory may help make clear the extent to which
deliberation can generate consensus (Miller 1992; Aldred 2000). But the argument
that individual valuations provide extra information seems more robust, and can
be generalized to the claim that DMV does more than simply decide amongst a
given set of policy options, because it provides a willingness to pay which is
applicable in broader resource allocation contexts. For instance, the willingness to
pay value could help a public authority make budgetary decisions beyond the
options facing the jury – say between health care and environmental preservation
expenditure.
The fundamental difficulty with this argument, and the DMV method, can be
easily stated: ‘is the collective valuation question meaningful?’ Ward (1999: 78)
writes: ‘The question posed should be “what is a particular level of provision of
environmental quality worth to the average citizen?”’ (1). Later he adds (p.79):
‘Jurors are not asked to express their personal evaluations but their judgements
about what environmental quality is worth to society as a whole’ (2). In contrast,
James and Blamey’s (2000, emphasis added) stylized valuation question takes the
form: ‘What is the maximum amount the citizens of City X would be prepared to
pay to ensure the proposed harbour development proceeds, if they knew what you
now know about it?’ (3). In the actual DMV exercise they conducted, the format was
different: ‘How high would a park management levy have to be, before the jury
would recommend Option 1 rather than Option 4 below? In other words, how
high would the levy have to be before the New South Wales public would be no
better off under Option 4 than Option 1?’ (5).
It seems possible that these five different question formats may elicit five
substantively different kinds of response. The variety of question formats may
reflect an underlying ambiguity in the variable they seek to measure. Although
both Ward and James and Blamey suggest they seek to capture ‘citizen’ rather
than ‘consumer’ responses (Sagoff 1988), it is not obvious that all the formats
would achieve this, particularly formats (1) and (3). Formats (1) and (3) might both
be interpreted by jurors as asking them to make some kind of prediction regarding
what others (non-jurors) would be willing to pay. Format (5) also appears to ask for
a prediction, but this time relative to welfare rather than willingness to pay. In
contrast, formats (2) and (4) seem to ask the jurors for their valuations, although
particularly in (2) it remains unclear how they are to interpret the question, and,
correspondingly, the meaning to be attached to an answer. One conclusion to
draw is simply that questions in DMV exercises need very careful formulation,
another is that DMV exercises reflect similar problems of meaning as do the CVM
exercises, if not more so. There is no space to go into this very large issue,9 but the
kinds of questions which both the CVM and DMV exercises raise are easily stated.
In the DMV context, do the jurors assume all members of the affected community
have to pay? And an equal amount? Or just those who express a hypothetical
willingness to do so? Or are the jurors being asked how much they would voluntarily
contribute to a joint effort to secure the specified outcome? It is difficult to see how
the DMV format, or any decision procedure which seeks a valuation, will ever be
as readily comprehensible to participants as the underlying ‘what is best for society?’
in a CJ.
Conclusions
This chapter has discussed how a comparison between the CVM and CJ might
proceed, focusing on the epistemic merits of discussion. It was argued that the
rational choice theory critique of these merits largely fails on its own terms. For
the CJ/CVM comparison, the focus in this chapter on the feasibility of competent
deliberation seems more informative than the early emphasis in the deliberative
democracy literature on an ideal of deliberation. However this focus is subject to
a number of caveats. The social facts which constrain and determine the deliberative
setting are not fixed but contingent. There is already some evidence of a dynamic
feedback effect from deliberative institutions to the social conditions, which
determine their feasibility. The evidence is generally positive – increasing the
presence of deliberative institutions may improve the confidence, capacity and
willingness of ordinary citizens to make decisions about public policy questions
(Smith and Wales 1998, 1999). Thus it is important to set the detailed discussion
of feasibility constraints in this chapter in context: they will be subject to – and
may be nullified by – the deliberative institutions, which they presently constrain.
204 Jonathan Aldred
One important feasibility constraint which has not been mentioned here (it receives
little attention from the rational choice theory perspective), but which deserves detailed
treatment, is inequalities of voice in any community – what might be termed the
problem of ‘willingness to say’ (O’Neill 1997). It is arguable that income and wealth
inequalities may generate the same unequal distribution of power in influencing the
decision, whether they work through constraining willingness to pay in CVM or
constraining willingness to say in CJs. More worrying still: the very requirement that
discussion be reasoned and serious to qualify as deliberation may be enough to
disenfranchise many members of society (Sanders 1997).
The question (Q) invokes a comparison both narrower and more general than
that between the CVM and CJ: it captures only one element of any overall com-
parative assessment (although it is not clear in what other respects the methodologies
can be helpfully compared), but it applies to the more general comparison of
deliberative versus aggregative institutions. It will by now be apparent that the
advantages of discussion canvassed here – and the objections raised – are highly
interdependent. Together they generate a complex web of forces (see in particular
Bohman 1996: 57–66), which even on the account of rationality advanced by
rational choice theory, create some significant benefits of discussion.
With these caveats firmly in mind, the discussion here has a number of policy
implications. Very broadly, where bounded rationality is likely to be severe, public-
interested principles are called forth, or important private information is present,
the epistemic merits of discussion are likely to be more significant. CJ appears
more attractive in these contexts. This conclusion aims to stimulate further research
rather than general judgements about the superiority of one method in all cases.
In a particular policy context, there is no substitute for a careful, empirical, case-
by-case comparison. While both the CVM and CJ have drawbacks, the final section
suggests that appeal to a mixed method such as DMV may not be the best way to
overcome them.
Acknowledgements
The author would like to thank Alan Holland, Michael Jacobs, Rosemary James,
Jürgen Meyerhoff, Simon Niemeyer, John O’Neill, Graham Smith, Andrew Stirling
and Hugh Ward for useful discussions on these issues.
Notes
1 See Bohman (1998) for a survey. The leading early statements of deliberative democracy are all
in Bohman and Rehg (1997), which also contains important papers. Another key collection is
Elster (1998a).
2 I shall ignore the distinction between judgements and preferences in what follows. Much of the
literature to be discussed uses ‘preference’ in contexts where ‘judgement’ seems more appropriate
– where the content of the ‘preference’ can be assessed for reasonableness. I shall follow the
usage of this literature while acknowledging its shortcomings.
3 There is substantial work discussing CJs in an environmental context (Armour 1996; Crosby
1996; Coote and Lenaghan 1997; Jacobs 1997b; Kuper 1997; Aldred 1998; Crosby 1999; Rippe
and Schaber 1999; Smith and Wales 1999; Ward 1999; Aldred and Jacobs 2000), in addition to
the contributions elsewhere in this volume.
4 Of course absence of agreement does not ensure an open output. There may be no agreement
among the CVM participants, yet an apparently closed output – at least if the CVM is regarded
as having only one output, an aggregate valuation, and data on the distribution of individual
valuations or protest bids is ignored.
5 The distinction (Jacobs 1997b) between an institution with decision-making power, and one
with power only to make a recommendation to a decision-making authority, is ignored in this
chapter. A recommendation is termed a decision regardless of whether those making the recom-
mendation have power to enforce it. (Of course the possession or absence of such power will
typically affect the recommendation.)
6 Although the term deliberative institution will be retained instead of discursive institution.
7 In this chapter, rational choice theory is understood in a strong sense as comprising both the
formal axioms which make up decision theory (Anand 1993 offers a detailed critical treatment),
and a substantive assumption of self-interest – in short, the agent is homo economicus. See Pettit
(1998).
8 The arguments in the next three sections are developed in more detail in Aldred (2001).
9 There is a large literature on what the CVM question means in theory (as opposed to
methodological difficulties in the practical application of the method), most of it written by
non-economists. For three different, and powerful, critiques by economists which cast doubt on
the coherence of the CVM question, see Sen (1995a), Sugden (1999) and Vatn and Bromley
(1994).
References
Aldred, J. (1996) ‘Value conflicts in environmental decision-making’, PhD thesis, Faculty
of Economics and Politics, Cambridge: University of Cambridge.
Aldred, J. (1998) ‘Land use in the fens: lessons from the Ely citizens’ jury’, Ecos, 19(2):
31–7.
Aldred, J. (2000) ‘Deliberation and social choice theory’, paper presented at ISUS
Conference, Winston-Salem/North Carolina, March 2000.
Aldred, J. (2001) ‘Citizens’ juries: discussion and rationality’, Risk, Decision and Policy, 6(2):
71–90.
Aldred, J. and Jacobs, M. (2000) ‘Citizens and wetlands: evaluating the Ely citizens’ jury’,
Anand, P. (1993) Foundations of Rational Choice under Risk, Oxford: Clarendon.
Armour, A. (1996) ‘The citizens’ jury model of public participation’, in O. Renn, T. Webler
and P. Wiedemann (eds) Fairness and Competence in Citizen Participation: Evaluating Models
for Environmental Discourse, Dordrecht: Kluwer.
Austen-Smith, D. (1990a) ‘Credible debate equilibria’, Social Choice and Welfare, 7: 75–93.
Austen-Smith, D. (1990b) ‘Information transmission in debate’, American Journal of Political
Science, 34(1): 124–52.
Austen-Smith, D. and Riker, W. (1987) ‘Asymmetic information and the coherence of
legislation’, American Political Science Review, 81 (3): 897–918.
Bohman, J. (1996) Public Deliberation, Cambridge, MA: MIT.
Bohman, J. (1998) ‘The coming of age of deliberative democracy’, Journal of Political
Philosophy, 6 (4): 400–25.
Bohman, J. and Rehg, W. (eds) (1997) Deliberative Democracy, Cambridge, MA: MIT.
Brouwer, R., Powe, N., Turner, R.K., Bateman, I.J. and Langford, I.H. (1999) ‘Public
attitudes to contingent valuation and public consultation’, Environmental Values, 8(2):
325–47.
206 Jonathan Aldred
Brown, T., Peterson, G. and Tonn, B.E. (1995) ‘The values jury to aid natural resource
decisions’, Land Economics, 71(2): 250–60.
Buchanan, J. (1954a) ‘Individual choice in voting and the market’, Journal of Political Economy,
62: 334–43.
Buchanan, J. (1954b) ‘Social choice, democracy and free markets’, Journal of Political Economy,
62: 114–23.
Coote, A. and Lenaghan, J. (1997) Citizens’ Juries, London: Institute of Public Policy
Research.
Crosby, N. (1996) ‘Citizens’ juries: one solution for difficult environmental questions’, in
O. Renn, T. Webler and P. Wiedemann (eds) Fairness and Competence in Citizen Participation:
Evaluating Models for Environmental Discourse, Dordrecht: Kluwer.
Crosby, N. (1999) ‘The citizens jury process and environmental decisions’, in K. Sexton
and T. Burkhardt (eds) Better Environmental Decisions: Strategies for Governments, Businesses
and Communities, Washington, DC: Island Press.
Elster, J. (1986a) ‘The market and the forum: three varieties of political theory’, in J. Elster
and A. Hylland (eds) Foundations of Social Choice Theory, Cambridge: Cambridge University
Press.
Elster, J. (1986b) The Multiple Self, Cambridge: Cambridge University Press.
Elster, J. (1995) ‘Strategic uses of argument’, in K. Arrow (ed.) Barriers to Conflict Resolution,
New York: Norton.
Elster, J. (ed.) (1998a) Deliberative Democracy, Cambridge: Cambridge University Press.
Elster, J. (1998b) ‘Introduction’, in J. Elster (ed.) Deliberative Democracy, Cambridge: Cambridge
University Press.
Farrell, J. and Gibbons, R. (1989) ‘Cheap talk with two audiences’, American Economic Review
79(3): 1214–23.
Fearon, J. (1998) ‘Deliberation as discussion’, in J. Elster (ed.) Deliberative Democracy,
Gambetta, D. (1998) ‘Claro! An essay on discursive machismo’, in J. Elster (ed.) Deliberative
Democracy, Cambridge: Cambridge University Press.
Goodin, R. (1986) ‘Laundering preferences’, in J. Elster and A. Hylland (eds) Foundations of
Social Choice Theory, Cambridge: Cambridge University Press.
Gowans, C. (1987) Moral Dilemmas, Oxford: Oxford University Press.
Gutmann, A. and Thompson, D. (1996) Democracy and Disagreement, Cambridge, MA:
Harvard University Press.
Hylland, A. (1986) ‘The purpose and significance of social choice theory’, in J. Elster and
A. Hylland (eds) Foundations of Social Choice Theory, Cambridge: Cambridge University
Press.
Jacobs, M. (1997a) ‘Alternatives to contingent valuation’, paper presented at CSEC/Green
Alliance Conference, London, 23 January 1997.
Jacobs, M. (1997b) ‘Environmental valuation, deliberative democracy and public decision-
making institutions’, in J. Foster (ed.) Valuing Nature?, London: Routledge.
James, R. and Blamey, R. (2000) ‘Deliberative non-market valuation’, paper presented at
ISEE 2000 Conference: People and Nature: Operationalising Ecological Economics,
5–8 July.
Johnson, J. (1998) ‘Arguing for deliberation: some skeptical considerations’, in J. Elster
(ed.) Deliberative Democracy, Cambridge: Cambridge University Press.
Johnson, S. (1963) ‘The adventurer, Saturday 28th April 1753’, in W. Bate (ed.) ‘The Idler’
and ‘The Adventurer’, New Haven: Yale University Press.
Krehbiel, K. (1991) Information and Legislative Organization, Michigan: Ann Arbor.
Kuper, R. (1997) ‘Deliberating waste: the Hertfordshire citizens’ jury’, Local Environment,
2(2): 139–53.
Lane, R. (1991) The Market Experience, Cambridge: Cambridge University Press.
Liebrand, W., Messick, D. and Wilke, H. (eds) (1992) Social Dilemmas: Theoretical Issues and
Research Findings, Oxford: Pergamon Press.
Macedo, S., (ed.) (1999) Deliberative Politics, Oxford: Oxford University Press.
Mackie, G. (1998) ‘All men are liars’, in J. Elster (ed.) Deliberative Democracy, Cambridge:
Cambridge University Press.
Miller, D. (1992) ‘Deliberative democracy and social choice’, Political Studies, 40: 54–67.
Nagel, T. (1979) ‘The fragmentation of value’, in T. Nagel (ed.) Mortal Questions, Cambridge:
O’Connor, M. (2000) ‘Pathways for environmental evaluation: a walk in the (Hanging)
Gardens of Babylon’, Ecological Economics, 34(2): 175–93.
O’Neill, J. (1993) Ecology, Policy and Politics, London: Routledge.
O’Neill, J. (1997) ‘Deliberation and its discontents’, working paper presented at the
conference Environmental Justice, Global Ethics for the 21st Century, October 1997,
Melbourne: University of Melbourne.
Orbell, J., van der Kragt, A. and Dawes, R. (1988) ‘Explaining discussion-induced co-
operation’, Journal of Personality and Social Psychology, 54(5): 811–19.
Orbell, J., Dawes, R. and van der Kragt, A. (1990) ‘The limits of multilateral promising’,
Ethics, 100: 616–27.
Pettit, P. (1998) ‘Institutional design and rational choice’, in R. Goodin (ed.) The Theory of
Institutional Design, Cambridge: Cambridge University Press.
Riker, W. (1982) Liberalism against Populism: A Confrontation between the Theory of Democracy and
the Theory of Social Choice, San Francisco: Freeman.
Rippe, K. and Schaber, P. (1999) ‘Democracy and environmental decision-making’,
Environmental Values, 8(1): 75–88.
Sagoff, M. (1988) The Economy of the Earth, Cambridge: Cambridge University Press.
look beyond contingent pricing’, Ecological Economics, 24(2–3): 213–30.
Sally, D. (1995) ‘Conversation and cooperation in social dilemmas’, Rationality and Society,
7: 58–92.
Sanders, L. (1997) ‘Against deliberation’, Political Theory, 25(3): 347–76.
Sen, A. (1970) Collective Choice and Social Welfare, Amsterdam: North Holland.
Sen, A. (1995a) ‘Environmental evaluation and social choice: contingent valuation and
the market analogy’, Japanese Economic Review, 46(1): 23–37.
Sen, A. (1995b) ‘Rationality and social choice’, American Economic Review, 85(1): 1–24.
Sen, A. (1997) ‘Individual preference as the basis of social choice’, in K. Arrow, A. Sen
and K. Suzumara (eds) Social Choice Re-examined, London: Macmillan.
Simon, H. (1978) ‘Rationality as process and as product of thought’, American Economic
Review, 68 : 1–16.
Smith, G. and Wales, C. (1998) ‘Citizens’ juries and deliberative democracy’, mimeo,
Glasgow: University of Strathclyde, Department of Government.
Smith, G. and Wales, C. (1999) ‘The theory and practice of citizens’ juries’, Policy and
Politics, 27(3): 295–308.
Sugden, R. (1999) ‘Alternatives to the neo-classical theory of choice’, in I. Bateman and
K. Willis (eds) Valuing Environmental Preferences, Oxford: Oxford University Press.
Vatn, A. and Bromley, D. (1994) ‘Choices without prices without apologies’, Journal of
Environmental Economics and Management, 26: 129–48.
208 Jonathan Aldred
Politics, 8(2): 75–96.
Williams, B. (1979) ‘Conflicts of values’, in A. Ryan (ed.) The Idea of Freedom, Oxford:
Three approaches to valuing nature 209
9 Three approaches to
valuing nature
Forest floodplain restoration
Wendy Kenyon and Nick Hanley
Introduction
Given the extensive literature documenting problems related to economic valuation
techniques such as the contingent valuation method (CVM), interest has developed
in approaches which provide an alternative source of information on public value
judgements for natural resource decision-making (Brown et al. 1995). In particular,
some authors have suggested that deliberative methods might have a role to play
as alternatives or complements to more traditional project appraisal techniques
(Jacobs 1997; Sagoff 1998). A number of researchers have assessed the role of the
Citizens’ Jury (CJ) as discussed in the preceding chapter (see also, Stewart et al.
1994; Aldred and Jacobs 1997; James and Blamey 1999). Here we take the CJ to
consist of a small group of people, selected to represent the general public rather
than any particular interest group or sector, who meet to deliberate upon a policy
question.
The CJ offers a number of advantages over economic appraisal methods, and
in particular over the CVM. First, respondents to CVM surveys must understand
exactly what it is they are to value if useful information about preferences is to be
elicited (Arrow et al. 1993). However, recent literature highlights the fact that many
respondents do not appear to be well informed about the issues or the good to be
valued (Brown et al. 1995; Jacobs 1997). As Munro and Hanley (1999) show,
changing people’s information sets can be expected to change their willingness to
pay (WTP). CJs tackle this ignorance problem by combining information, time,
scrutiny and deliberation in the preference elicitation process (Coote and Lenaghan
1997). They allow participants to question witnesses, discuss witnesses’ evidence
with other jurors, and thereby gradually learn about and reach a richer
understanding of the issue (Sagoff 1998). CJs therefore address the information
problem better than the CVM.
Second, economists and others have suggested that a CVM questionnaire asks
respondents the wrong question, assuming that consumers think about
environmental goods in the same way they do about private goods (Blamey 1996;
Jacobs 1997; Sagoff 1988). Blamey (1996) suggests that respondents should not be
treated as consumers of environmental goods, but rather as citizens who think of
the welfare of the community when responding to environmental issues. In other
210 Wendy Kenyon and Nick Hanley
words, individuals approach decision-making relating to environmental goods as
citizens rather than consumers. Whilst the validity of this as a universal description
fitting all cases of environmental management can be disputed, the use of CJs as
a method of preference revelation allows consumers to be asked what Sagoff and
Jacobs call ‘the right question’, as it allows deliberation on the environmental issue
in terms of what is best for society. Indeed, while the question for the jury can be
framed in the context of individual consumer values and preferences if necessary,
CJs were developed specifically to determine opinions that represent the general
public, rather than any individual interest (Coote and Lenaghan 1997; Brown et
al. 1995).
Third, CJs may also be useful in dealing with the equity and distributional
issues which the CVM has attracted criticism over. Economic value is effectively
determined by preferences, underpinned by ability to pay. Therefore, in the CVM,
any value that a consumer places upon a good is not registered unless she is able to
pay for it. CJs however allow participants’ opinions and preferences to be expressed
and registered regardless of their ability to pay (Crosby 1995).1
Fourth, the United Nations Conference on Environment and Development has
suggested that environmental decisions will not be sustainable unless local
communities participate fully in the decision-making process (United Nations 1993).
The CVM does not include the participation of the community in a central way,
and therefore may not encourage sustainable actions. The use of CJs is, in contrast,
a means by which public participation can be more fully incorporated into the
environmental decision-making process.
Finally, the notion of value construction suggests that respondents do not have
well-defined preferences for many complex environmental goods prior to the elicit-
ation process, but that these preferences are constructed during this elicitation
process itself (Gregory et al. 1997; Payne and Bettman 1999). The way in which
people construct their preferences is important as it is argued that decision-makers
should attach more weight to the preferences of someone who knows both sides of
the argument, than to someone whose knowledge of the problem is more limited
(Elster 1983; Payne and Bettman 1999). According to Elster (1983) however,
methods which explicitly try to determine how people construct their preferences
(for example by aiding value construction, such as Gregory et al.’s (1997) decision
pathways, or by asking respondents to ‘think aloud’ whilst determining their value
as Schkade and Payne (1994) did) are likely to be contradictory – akin to telling
someone to ‘be spontaneous’! CJs, however, provide information about the process
of preference construction as a by-product of the process rather than as a central
role. Therefore, using CJs may offer a means of circumventing the contradiction
inherent in helping respondents construct their values, and provide information to
decision-makers on the weight which should be given to those preferences expressed.
Elster (1983) argues that because the CVM is based on a ‘thin theory of
rationality’, which requires only consistency in the expression of preferences, poor
policy choices may be made on the basis of CVM results. CJs on the other hand
appear more consistent with a ‘broad theory of rationality’ which examines not
only the consistency of expressed preferences, but also the beliefs and desires behind
decisions. He argues that decisions should be made by the ‘public and rational
discussion about the common good, not the isolated act of voting to private
preferences’ (Elster 1983: 35). CJs facilitate this type of rational discussion more
than the use of economic appraisal methods.
Despite the advantages that CJs appear to have over the CVM, they do not
provide an economic estimate of the value of any particular project, nor whether
it constitutes an efficient use of resources. These are important weaknesses. A
third approach therefore seems potentially desirable, which allows deliberation,
but also the estimation of the economic value of a project. Such a method might
aim to combine the strengths of the CVM and CJ. This chapter reports on an
application of three approaches to one decision-making context, a proposed
floodplain restoration project in the south of Scotland. A CVM, a CJ and a
combined approach (the ‘valuation workshop’) were conducted to assess the value
of the project to local communities. The valuation workshop was an attempt to
combine the quantitative outputs of the CVM with the participatory, deliberative
and preference construction attributes of the CJ.
The rest of the chapter is organised as follows. We briefly describe the context
of the three approaches. Then we report on the design and the results of the
CVM. The next section describes the design and results of the CJ, and the following
section relates to the workshop approach. Then we assess the methods, and look at
the implications of the CJ and valuation workshop for dealing with problems related
to the CVM. The chapter concludes with a discussion of the complementarity of
the three approaches.
The Ettrick Valley Forest floodplain restoration project

The context for the three approaches is a forest floodplain restoration project in
the Borders Region of Scotland. Forest floodplains have almost disappeared from
temperate areas in Europe and North America. In Britain, most forest floodplains
were cleared in prehistoric and early historic times (Peterken and Hughes 1995).
Until the eighteenth century some tracts survived on large rivers, but interventions
by man in the early nineteenth century have meant that much of the remaining
forest floodplains were wiped out. Actions such as ‘river rectification’ meant that
flooding regimes were altered so that agriculture could take place on land previously
prone to flooding.
Ecologically the decline of forest floodplains is significant as they are one of the
‘richest components of the landscape’ (Peterken and Hughes 1995: 191). Dechamps
et al. (1987) prove that bird species richness depends on the variety of vegetation
types, and the fauna depends on the variety of food sources available. The richness
of flora found in forest floodplain areas therefore leads to a richness of wildlife
inhabiting such areas. The forest floodplain is therefore important in protecting
and restoring biodiversity. Restoration of forest floodplains may therefore go some
way to maintaining or restoring biodiversity in suitable areas. In addition, Petts
(1990, cited in Peterken and Hughes 1995) suggests that the restoration of forest
floodplains might have a number of other benefits including timber production,
reduction of agricultural surpluses, fishing, water quality and pollution control,
and landscape benefits.
In the last two decades interest has grown in conserving, restoring and expanding
those areas of forest floodplain that have survived. In 1981, forest floodplains
were the subject of a special study by the Council of Europe (Yon and Tendron
1981). In 1982, the Council of Ministers passed a resolution that they should be
protected (No. R(82)12). The World Wide Fund for Nature (WWF) launched a
project to conserve central European floodplains in 1985, and ten years later, in
1995, WWF Scotland responded to concerns about the loss of forest floodplain
habitats by commissioning a review of their status in Scotland. The study concluded
that such habitats were scarce and under threat in Scotland (McGhee Smith
Associates 1995).
One of the most ecologically interesting areas of floodplain identified in the
review was the Upper Ettrick site, which contains a variety of woodland, wetland
and grassland habitats of nature conservation interest. The different habitats are
distributed in a mosaic of small patches which results in high biodiversity for the
area, including species which are recognised as locally and nationally scarce or
endangered. The Upper Ettrick presents great potential for the expansion of this
valuable habitat, utilising areas which are at present of limited conservation interest
such as conifer plantation and improved grassland. Increasing the areas of valuable
habitat would both protect the species which are already present and encourage
others which would have been present in these habitats in the past. Specifically the
project aimed to create 25 hectares of native woodland; to restore 30 hectares of
floodplain habitat, including scrub, fen, haymeadow and wetland; to convert 30
hectares of conifer plantation to native broadleaf; to manage 15 hectares of willow
scrub and to create 3 km of footpaths, boardwalks and tracks. A group of non-
government and government organisations in partnership with the local community
and landowners co-ordinated the habitat restoration project. The project was led
by a local community organisation, the Borders Forest Trust, who were guided by
technical and local community steering groups. The Trust was keen to work in
partnership with the local community in all aspects of the project, including its
evaluation, and were also keen to illustrate the economic benefit the project provided
to current and potential future funders. The project therefore provided an
appropriate context for the application of the three methods.
The contingent valuation method survey

There is much debate in the literature about the appropriate design of CVM
questionnaires, with respect to a number of issues, and in particular the choice of
elicitation format and payment vehicle. Boyle and Bergstrom (1999: 197) state
that ‘despite early evidence that payment vehicles can influence responses to CVM
questions no published research has been conducted to address this concern’.
However, recent recommendations suggest that a mechanism where the respondent
has no choice but to pay (e.g. taxation), is most appropriate. Carson et al. (1999)
suggest that the use of charitable donations as a payment vehicle within a CVM
survey may lead to strategic behaviour by the respondent. However, others suggest
that a charitable bid vehicle is appropriate in certain circumstances, and indeed
recommend that the information and questionnaire be in a campaigning style to
imitate real scenarios of this nature (Macmillian et al. 1998). Macmillian et al.
(1998) found that a charitable donation achieved a high level of convergence with
real payments. One of the other important design issues in using the CVM
surrounds the elicitation format used. The National Oceanic and Atmospheric
Administration panel favoured the dichotomous choice (DC) format, but this
contrasts with other recommendations that the most conservative questionnaire
design is most appropriate (Arrow et al. 1993). The DC format consistently produces
higher estimates of mean WTP than do open-ended elicitation formats (Ready et
al. 1996; Boyle et al. 1996), and is therefore not the most conservative design choice.
Notwithstanding the preceding discussion, one of the most important
considerations in the design of a CVM questionnaire is to make the scenario
believable (Mitchell and Carson 1989; Boyle and Bergstrom 1999). In the Ettrick
study a charitable donation payment vehicle and an open-ended elicitation format
were used. There were two main reasons for this choice. First, a number of focus
groups were carried out during the design of the questionnaire, where different
payment vehicles and elicitation formats were considered. Participants of these
groups indicated that they were most comfortable with the open-ended format
and the charitable donation bid vehicle. Second, this format and vehicle is one
which respondents to the survey are familiar with, especially given the local nature
of the project. A wildwood recreation project in the Borders Region of Scotland
was campaigning for funds at a similar time that the questionnaire was designed,
tested and conducted. In this real situation an open bid in conjunction with a
payment card type elicitation method was used with a charitable donation as the
payment mechanism. This method is often seen in other charitable campaigns in
the UK. Respondents to the Ettrick survey could therefore be expected to be familiar
with this payment context, making the choice most suitable in the circumstances.
The questionnaire consisted of three sections. The first requested general
information about the respondent’s residential status, their participation in outdoor
activities, and attitudes towards the environment. The second provided information
about the forest floodplain project and asked a payment principle question, as well
as the WTP question.2 This section also reminded the respondent of their budget
constraint, that the money would go to fund the Ettrick project only, and that the
project would not go ahead if sufficient funds were not raised from public donation.
It also contained a question which allowed protest bids to be identified. The final
section of the questionnaire requested the usual socio-economic data from
respondents. The questionnaires were completed in the late summer of 1998, by
an independent market research company, using a face-to-face interview method.
Nine towns in the Borders Region of Scotland were selected as sites for the
questionnaires to be carried out, and responses were collected from a stratified
sample of the Borders population, and a small number of visitors to the Region.
The contingent valuation questionnaire provided information on respondents’
WTP for the project and on their views on the project as a whole. Of the 336
respondents, 29 per cent were protest bids. This number is relatively high and the
reason most often stated was that they thought the government, the lottery or
some other body should pay. Those respondents who did not provide a reason for
bidding zero were classified as protesters. Protest bids were removed from the data
set for further analysis, although this means that average WTP is higher than would
be the case if protesters were allocated a zero value and included. Table 9.1 shows
the resultant descriptive statistics for the project. The mean WTP was £13.18 per
person for the project as a one-off payment. Descriptive statistics are shown in
Table 9.1.3
Given the results shown in Table 9.1, the benefit of the project can be estimated
at £568,677 (based on a once-only mean WTP aggregated over the population of
the Borders Region) and can be compared with the costs of the project at £350,000.4
Costs of the project were considered in the initial years only, as the aim of the
project was that it would be self-sustaining, perhaps with some small continued
help from the local community. The composition of costs and benefits indicates
that the project might pass the cost-benefits test and may therefore be considered
value for money. However, this data does not provide any recommendations for
management, nor does the method involve the participation of the local community
in a central way. An alternative approach which does meet these two criteria was
therefore also conducted – a CJ.
The citizens’ jury

The CJ participants were selected from the CVM questionnaire respondents. The
final question on the CVM questionnaire asked respondents whether they would
be willing to attend a group meeting to discuss local issues in more depth. Those
who responded positively to this question formed the pool from which jurors were
selected.5 The CVM questionnaire provided extensive socio-economic information
about respondents, so that it was possible to select jurors to be representative of
the Borders population. Eleven jurors finally took part in the CJ.
The jury met over three days and one evening in December 1998. The aim was
to assess the project site and to provide qualitative information on its value and
importance to the local community. The jury was carried out in collaboration
with a local community environmental organisation, the Borders Forest Trust,
who were keen to encourage consultation and participation of the local community
in the project. The jurors were asked to deliberate on the following specific questions:
(i) What should individual land use and environmental projects in southern Scotland
such as the Ettrick Forest Floodplain Project aim to achieve? and (ii) How might
Table 9.1 Descriptive statistics for WTP for the Ettrick project
Mean Std dev Median Range 95% Confidence intervals N
(£ per person)
13.18 69.71 1 0–1,000 4.15–22.1 232
the success of such projects be determined? Clearly, ‘economic efficiency’ could
be one answer to either of these questions, but may not be a priority for participants.
To provide information with which the jury could assess the project, ten
witnesses attended sessions to provide ‘evidence’ and to answer questions. The
witnesses were selected in consultation with the Borders Forest Trust, and in
discussions with stakeholders from all sides of the debate to ensure balanced
representation. The witnesses came from a variety of backgrounds including
Scottish Natural Heritage, the local council, the Forestry Commission, the Scottish
Tourist Board, and environmental consultants. Witnesses made short
presentations to the jury of 10 to 15 minutes followed by a discussion session
with the jury of about 30 to 40 minutes. In addition to sessions involving witnesses,
the process included a number of jury-only sessions, where the jurors discussed
particular issues as a whole unit, or in smaller groups using participatory appraisal
techniques. Participatory tasks were designed to including listening, hearing and
watching. According to Pretty et al. (1995: 24) participants are likely to remember
‘10 per cent of what they read, 20 per cent of what they are told, 30 per cent of
what they see and 50 per cent of what they hear and see’. Although participants
did not need to remember long term what went on in the session, ensuring they
were engaged in the tasks was vital. The use of brainstorming techniques,
organising comments written on cards, and using large flip-chart paper served a
number of purposes: to provide a focus for attention while the discussion ensued;
to stimulate discussion between people with different levels of numeracy and
literacy; to re-enforce spoken issues; to provide a means of cross-checking what
was said, and to record the proceedings. The final recommendations were
achieved entirely by discussion and consensus, and approved by all of the jurors.
A report on the process and outcome of the jury was written and sent to the
jurors for approval, before being sent to the Borders Forest Trust and other
interested parties (Kenyon 1999).
The jury made recommendations about how the Ettrick project should be
managed and co-ordinated in order to achieve the environmental and social
objectives which the jurors felt were desirable in southern Scotland. The jury also
noted a number of aspects related to the project they felt were valuable (Table
9.2). Interestingly, not only environmental issues were identified, but also aspects
which might be classed as social values, such as community involvement, and the
role of the project for education.
In addition to the aspects of value identified in Table 9.2, the jury noted two
areas of concern, namely access and future management. Jurors felt that visitors
should be allowed access to the site, but were concerned about the damage they
might inflict. Specific recommendations were made regarding arrangements to
promote benign access, and the provision of information to visitors to encourage
careful use of the site. The second area of concern related to future management
of the site and the availability of funding to secure it. The jurors felt that it may be
possible to start a trust fund dedicated to the Ettrick project to make sure that
money was still available for future management. However, after speaking with a
member of the local community, the jurors became less concerned as they were
Table 9.2 Positive issues identified with the forest floodplain project
Environmental issues Social issues
Preservation of a natural ecosystem – Good demonstration scheme for copying
a world resource
Flood control Education of the young and encouraging
educational studies
Balance of different habitats and getting Decrease the number of sheep and fencing
back to nature of sensitive areas
Encouraging wildlife and preservation of Getting rid of ugly blanket plantation
indigenous life forms forestry
Monitoring of species Community involvement
assured that the local community was fully involved in the project and that they
would ensure that the site was well managed into the future.
The jury were able to look at the Ettrick project in the wider context of the
south of Scotland, and made a number of recommendations regarding the
management of individual environmental projects in the region. They felt that a
variety of projects were needed in the region, which met a number of different
needs, but that all of these projects should be co-ordinated in an integrated way.
For example, some environmental projects might aim to attract tourism to the
area, but such projects must be situated in less environmentally sensitive areas.
Others might aim to increase biodiversity, but not aim to attract tourists. Finally,
the jury considered how the success of environmental projects such as the Ettrick
Forest Floodplain Project, might be assessed. Table 9.3 shows the criteria offered
by the jury, who also suggested that these measures could be considered as
elements of future project design and planning. It is interesting to note that the
jury considered environmental and social elements important in judging the
success of the project, but that they did not seem to consider economic criteria
important.
The valuation workshop

The final approach used to evaluate the Ettrick Floodplain Restoration Project
was a hybrid of the CVM and CJ, named the valuation workshop. It aimed to
build on the strengths of the CVM and CJ exercises, and contained elements of
each. In particular, the intention was to combine the quantitative outputs of the
CVM with the participatory, deliberative and preference construction aspects of
the CJ. The valuation workshop was conducted in three parts. After an introduction,
the participants were each given a CVM questionnaire and asked to complete it
individually. Next, participants were divided into groups of between 4 and 7 people
for discussion. The first task each group was given was to discuss the good aspects
of the Ettrick project. Participatory methods (such as those discussed above) were
used to facilitate the discussion, and provide a focus for the task (Pretty et al. 1995).
Table 9.3 Jury suggestions measuring environmental project success
Environmental issues Social issues
Has the variety of wildlife improved? Has it got community and farmer approval?
Is wildlife being protected? Has community spirit improved?
Is the site attractive? Has the project created any unseen
problems?
Each group was asked to provide a ranked list of good points related to the Ettrick
project. The second task was to consider the problems of the project, and the
concerns they had with it. Each group were asked to discuss problems, rank their
relative importance and offer suggestions of how each problem might be mitigated
or solved. The final task in this section of the workshop asked the participants to
make recommendations on how the success of the project might be judged. This
involved creating a list of means by which the project could be assessed, with
discussion surrounding each suggestion. In the third part of the valuation workshop
participants were asked to complete some further questions individually in a survey.
These questions asked whether the participants would change the WTP they stated
at the beginning of the workshop, and explain why. Participants had now had a
chance to discuss the project, assimilate information more fully and consider the
project from different perspectives, given the interaction with other participants.
The valuation workshops were conducted in December 1999 and took around
three hours each. Two workshops were carried out in each of two towns in the
Borders, giving a total of four workshops and 44 participants. In order to select
participants, 500 letters were sent out to a random selection of addresses in or
near each of the towns. The letter invited the addressee to one of the workshops,
explained that the Scottish Agricultural College was carrying out research on the
local environment and were interested in their views, and asked them to return an
enclosed form if they wished to participate. A pool of possible participants was
then drawn up from those who replied, and participants were chosen from this
pool to be as representative of the local population as possible. Each participant
was paid £20 for their attendance.
The workshop provided a range of qualitative and quantitative data on the
Ettrick project. Of the 44 participants attending the workshops, 27 (61 per cent)
made genuine CVM bids initially, the rest being either protest bids (5 per cent) or
‘don’t know’ responses (34 per cent) to the valuation question. After the discussion,
in the second part of the questionnaire, the number of valid bids increased as 2 of
the ‘don’t know’ respondents offered a new valid bid. Of the 44 participants, only
6 (14 per cent) changed their bid after the discussion (Table 9.4). Participants were
clearly considering their budget constraints carefully as the most common reason
given for not changing their bid was that their financial circumstances had not
changed and they could not therefore afford a higher amount even if they now
wanted to. Interestingly none of the participants revised their bid downwards.
Although the preceding part of the workshop comprised discussion of both the
good and bad aspects of the project, the discussion of the good points clearly
carried more weight.
Table 9.4 Original and revised bids with reasons for change
Original bid Revised bid Reason
25 50 Wish it would happen
10 20 It would have a beneficial effect on the local community,
Borders community, and is a valuable recreational and
educational asset
25 50 Awareness of the problem
20 ?? Would be willing to sponsor the project on a monthly
basis
DK 10 More knowledge
DK 25 Because of points for the project brought up in the
discussion
Table 9.5 and Table 9.6 display the descriptive statistics for pre-discussion and
post-discussion bids. Although the mean has increased, a two-tailed paired T-test
showed no significant difference between the two workshop mean WTPs at 10 per
cent significance (t = –0.57, p = 0.57).6 Although the discussion did have an impact
on individual bids, it did not therefore have a statistically significant impact on
mean WTP. A Mann–Whitney U-test gives the same results, the hypothesis that the
pre-discussion median is equal to the post-discussion median cannot be rejected (W
= 747.0, p = 0.6735 adjusted for ties).
In addition to this quantitative output, the valuation workshop provided
information about what the participants felt was good and bad about the project,
and how the project’s success might be measured. Environmental aspects of the
project were in all but one of the 8 groups (2 per workshop) considered the most
important. The one group who dissented felt that the promotion of education and
tourism in the area was most important. Table 9.7 shows the results from one
discussion group which is typical of the responses from other groups.
One strength of the valuation workshop approach relative to the CVM is that it
provides additional information on participants’ views of the project – both positive
and negative. Participants were asked not only to discuss issues which they thought
were problematic, but also to offer solutions to these problems. Table 9.8 shows the
typical concerns and the solutions offered by one of the workshop groups. Every
group considered the cost of the project to be a problem and many groups
differentiated between start-up costs and running costs. One of the other most
discussed problems was that of access. Roads to the site were single track and not
suitable for a large number of cars, and there was virtually no public transport
available to the site. Despite having been provided with very little information,
participants were able to offer solutions to these problems in most cases, clearly
some more practical than others (Table 9.8).
Finally, the participants were asked to suggest how the success of the project
might be measured. A range of ideas was suggested including monitoring flora and
Table 9.5 WTP descriptive statistics for initial valuation
Mean Median Std dev Range 95% CI N
11.07 10 12.59 0–50 6.09–16.05 27
Table 9.6 WTP descriptive statistics for final valuation

Mean Median Std dev Range 95% CI N
13.59 10 12.59 0–50 7.64–19.54 29
Table 9.7 Good aspects of the Ettrick project

Good points Examples Rank
Protected area Place of scientific interest 1
Protected area
Wildlife Increase wildlife 2
Varied wildlife
Protect wildlife
Jobs Create jobs 3
Attract people to the area
Local/outdoor interest Somewhere to visit 4
More interesting countryside to walk in
Makes people more aware of the environment
locally
Reduce chemicals on land Reduce chemicals on land 5
Less cultivated Less cultivated, more natural 6
fauna, evaluating visitor levels, jobs, the level of community involvement, and interest
in the site by educational establishments. As with the CJ, no direct financial criteria
were suggested here, although many of the criteria have an economic element,
such as jobs and visitor levels.
Conclusions
Both the CVM and CJ provide policy-relevant information, the information
provided by each being useful in different ways. The results of the CJ identify the
project needs and how it should develop. Table 9.2 lists those issues that the jurors
felt to be positive, and that might be used as objectives by the managers of the
project. Similarly, by identifying concerns relating to the project, jurors provide
direction to managers and policymakers as to the development of the project. The
information provided by the jury therefore plays a practical role in directing the
management of the project.
The results of the CVM provided different, but equally policy-relevant
information. The CVM is able to measure both the intensity and direction of
Table 9.8 Project problems, solutions and the importance of bad aspects
Problems Rank Solutions
Costs – who pays 1 Start-up cost: central funding, lottery, EU
Running cost: 3 for admission fee, 3 against
Prevent damage by visitors
Rural depopulation – 2 Change/expansion of jobs related to the
reduction of number exercise, e.g. catering, maintenance, paths,
of viable farms bridges, etc.
Rangers
Tourism facilities
Farm shop/equipment hire shop
Access roads 3 Straighten roads
Bridges
Improved public transport, i.e. minibuses
Agreement with 4 No solutions offered
landowners
Damage by visitors 5 Designated picnic area and toilets
Designated play area
Car parking 6 No solutions offered
preferences, and provides data that is consistent with welfare economic analysis
which can feed into cost-benefit analysis. The benefit of the project estimated at
£568,677 could be compared with the costs of the project at £350,000. The project
passes the cost-benefit test, and thus offers a potential Pareto improvement.
The workshop approach is an attempt to amalgamate aspects of the two methods
and provide both practical guidelines to the project managers and cost-benefit
estimates, whilst maintaining the participation of the local community in the project
evaluation. The results of the workshop are comparable on the one hand with the
CVM results and on the other with CJ results. A T-test shows that neither the pre-
discussion workshop mean nor the post-discussion workshop mean WTP are
significantly different from the CVM mean bid at 10 per cent level (t = –0.41, p =
0.68 and t = 0.01, p = 0.99 respectively). However, a Mann–Whitney U-test does
shows a significant difference between the medians (W = 4228.5, p = 0.0350 and
W = 4265.5, p = 0.0268 respectively). The results of the discussion of the workshop
correspond quite closely to the recommendations provided by the jurors.
Workshop participants were also asked to evaluate the final aggregate CVM
result at the end of the session. The calculation was explained to the group, and
the final figure presented to them. The comments from all the groups can be split
into four categories. First, that it is not possible to put a value on such a project.
Second, comments surrounded whether or not this aggregate figure would be
obtained in reality, as due to the poor economic situation in the Borders there
were more important things to spend money on. Third, the clear feeling that money
for such projects should come from government, lottery or EU funds was expressed.
Finally, there was discussion that information on costs (which was not made
available) was required before any pronouncement was made on its value. This
additional feature provides interesting information on the CVM figures, and on
the authenticity of the economic estimates. Overall the valuation workshop
approach goes some way to developing a method which provides economic
estimates of environmental projects whilst adding value to traditional CVM
exercises, and offering qualitative recommendations. However, the workshop results
suffer in this study from a small sample size relative to the CVM survey.
The research also related to problems identified with the CVM in the literature,
and the role that more deliberative methods may have in mitigating these concerns.
One relates to the provision of information to CVM respondents. The respondents
to the CVM survey were given visual and verbal information about the project
site, and asked, ‘Do you prefer the site with or without the project?’ Thirteen per
cent of respondents did not know whether they preferred it with or without. This
may indicate that the information was not sufficient for them to be able to determine
their preference. This did not appear to be a problem in the CJ. Jurors were all
able to determine their own preferences regarding the Ettrick project, and further
were able to identify those aspects they preferred most (Table 9.2). The workshop
participants were given the same information as the CVM respondents, but were
allowed to discuss it in a social setting. The fact that two participants changed their
WTP bid from a ‘don’t know’ response to a positive response implies that the
discussion did allow the participants to better understand the project they were
asked to pay for.
A number of researchers suggest that CVM questionnaires ask the wrong
question and assume that respondents act as consumers and not citizens when
responding. Our results indicate that this assumption may be valid (Sagoff 1998;
Blamey 1996). Analysis of the CVM results showed that the most significant variable
in influencing whether a respondent would be willing to pay anything was whether
she was likely to visit the site if the project went ahead (Kenyon and Nevin 2001).
This may indicate that the respondents acted as consumers and not citizens when
responding to the questionnaire.7 If this is the case, the CJ can be seen as a comple-
ment to the CVM by evaluating the project from a citizen’s point of view. The
valuation workshop incorporates elements of both a citizen and consumer
approach. The initial CVM responses were from a consumer standpoint, but the
output from the discussion was clearly from a citizen standpoint, since community
issues were considered an important aspect of the Ettrick project. However, this
result also suggests some relationship between a ‘consumer’ standpoint and the
respondent accounting mainly for use and option values (and not other components
of total economic value). It is clear that further research is required on the relation-
ship between these two issues.
The issue of value construction is also a matter of debate within the environ-
mental valuation literature. The CVM results provide no evidence regarding the
detailed construction of final values. However, comments from jury participants
in the evaluation discussion and questionnaire at the end of the process suggested
that breaking down the decision-making process into tasks made the whole task
less daunting, and more manageable. The by-product of these tasks is to provide
information that may indicate how jurors constructed their final response. The
evidence of the valuation workshop is mixed on this point. Whilst the pre-discussion
WTP was constructed in the normal CVM way, the post-discussion response had
the benefit of discussion in a social setting, and the outcome of the discussion may
be an indicator of the reasons for a (limited) changed bid.
Finally, in a climate where both national and international agreements exist
which seek to enhance public participation in environmental decision-making, it
seems clear that the process should rely on more than just economic estimates of
value as provided by traditional CVM surveys. Whilst such estimates are still useful,
policy-makers are increasingly required to incorporate decision-making, planning
and management into one process, and include both expert and lay opinion. New
methods which are able to provide a more holistic approach to environmental
policy are needed so that these national and international targets can be met. The
CJ is able to offer such an integrated approach. Evidence from the CJ suggests
that the jurors were able to think holistically, offering comment on benefits and
pitfalls of the project and a wide range of management recommendations. One
of the recommendations was that the Ettrick project should not be considered in
isolation, but as part of a suite of projects, to ensure that the Borders environment
developed in an integrated way. The valuation workshop also provided a more
integrated approach, in which participants were able to complement economic
estimates with wider indicators of values and preferences.
In many instances where environmental project evaluation is required,
conducting both a CVM and CJ would be prohibitive in terms of time and money.
The challenge for the future may therefore lie in developing complementarities.
Building on the valuation workshop approach, and increasing sample sizes may
go further in offering a more appropriate combined approach, so that decision-
makers can benefit from both a wide range of policy-relevant information when
evaluating environmental projects.
Acknowledgements
This work was funded by the Economic and Social Research Council under the
Global Environmental Change Programme, and the Scottish Executive Rural
Affairs Division. The authors would also like to thank the Borders Forest Trust,
and the people of the Borders for their input and enthusiasm.
Notes
1 Although self-selection of the participants in a CJ could mean that those with higher education
and income, for example, may be more likely to participate and the preferences of the ‘excluded’
may not be registered.
2 How much money (on a once-only basis) would your household be willing to donate to the fund
in order to ensure this project went ahead? Please bear in mind that this money would only go
towards paying for the management of the natural environment in the Ettrick Valley. You might
also like to think about the spending on other items that you might have to give up if you did
make this payment.
3 More detailed statistical analysis of the CVM results can be found in Kenyon et al. (2000).
4 Figure obtained from the Borders Forest Trust.
5 The Borders Jury was selected in this way to minimise recruitment costs. Despite the final CJ
being a reasonable representation of the Borders population (when compared by socio-economic
characteristics), this is probably not the best means of jury selection. Selection processes utilising
a random draw from the electoral register are considered to provide a better sample of jurors.
6 The two respondents who changed from a ‘don’t know’ to a positive bid were taken out of the
sample for the purpose of the test.
7 It could be argued that the most important feature of the project was therefore its use or option
use value.
References
Aldred J., and Jacobs, M. (1997) ‘Citizens and wetlands. What priority, if any, should be
given to the creation of wetlands in the fens?’, report of the Ely citizens’ jury, Centre
for the Study of Environmental Change (CSEC), Lancaster: Lancaster University.
Arrow K., Solow R., Portney P.R., Leamer E.E., Radner R. and Schumann H. (1993)
‘Report of the NOAA panel on contingent valuation’, Federal Register, 58: 4601–14.
Blamey, R.K. (1996) ‘Citizens, consumers and contingent valuation: clarification and the
expression of citizen values and issue-options’, in W.L. Adamowicz, P.C. Boxhall, M.K.
Luckert, W.E. Phillips and W.A. White (eds) Forestry Economics and the Environment, Oxford:
CAB International.
Boyle, K.J. and Bergstrom, J.C. (1999) ‘Doubt, doubts and doubters’, in I. Bateman and
K.G. Willis (eds) Valuing Environmental Preferences. Theory and Practice of the Contingent Valuation
Method in the US, EU and Developing Countries, Oxford: Oxford University Press.
Boyle, K.J., Johnson, F.R., McCollum, D.W., Desvouges, W.H., Dunford, R.W. and Hudson,
S.P. (1996) ‘Valuing public goods: discrete versus continuous valuation responses’, Land
Brown, T.C., Peterson, G.L. and Tonn, B.E. (1995) ‘The values jury to aid natural resource
decisions’, Land Economics, 71: 250–60.
Carson, R.T., Groves, T. and Machina, M.J. (1999)‘Incentive and informational properties
of preference questions’, paper presented at the European Association of Resource
and Environmental Economists Conference, Oslo, June 1999.
Coote, A. and Lenaghan, J. (1997) Citizens’ Juries: Theory into Practice, London: Institute for
Public Policy Research.
Crosby, N. (1995) ‘Citizens juries: one solution for difficult environmental questions’, in O.
Renn, T. Webler and P. Wiedemann (eds) Fairness and Competence in Citizen Participation,
Dordrecht: Kluwer.
Dechamps, H., Joachim, J. and Lauga, J. (1987) ‘The importance for birds of the riparian
woodlands within the alluvial corridor of the River Garonne, SW France’, Regulatory
Rivers Restoration Management, 1: 301–16.
Elster, J. (1983) Sour Grapes. Studies in the Subversion of Rationality, Cambridge: Cambridge
University Press.
Gregory, R., Flynn, J., Johnson, S.M., Sattefield, T.A., Slovic, P. and Wagner, R. (1997)
‘Decision-pathways surveys: a tool for resource managers’, Land Economics, 73: 240–54.
Jacobs, M. (1997) ‘Environmental valuation, deliberative democracy and public decision
making institutions’, in J. Foster (ed.) Valuing Nature? Economics Ethics and Environment,
London: Routledge.
James, R.F. and Blamey, R.K. (1999) ‘Public participation in environmental decision-making
– rhetoric to reality?’, paper presented to the International Symposium on Society and
Resource Management, Brisbane, Australia, July 1999.
Kenyon, W. (1999) Report of the Galasheils Citizens’ Jury, Edinburgh: Scottish Agricultural
College.
Kenyon, W. and Nevin, C. (2001) ‘The use of economic and participatory approaches to
assess forest development: a case study in the Ettrick Valley’, Forest Policy and Economics,
3(1–2): 69–80.
Macmillian, D., Smart, T.S. and Thornburn, A.P. (1998) ‘The importance of realism to
experiments comparing cash and CV charitable donations: the case of the isle of Eigg
Trust’, paper presented at the Agricultural Economics Society Conference, Reading,
March 1998.
McGhee Smith Associates (1995) ‘Wild rivers. Aspects of the location and ecology of the
floodplain woodlands in Scotland’, report to World Wildlife Fund.
Mitchell, R.C. and Carson, R.T. (1989) Using Surveys to Value Public Goods: The Contingent
Valuation Method, Washington, DC: Resources for the Future.
Munro, A. and Hanley, N. (1999) ‘Information, uncertainty and contingent valuation’, in
I. Bateman and K.G. Willis (eds) Valuing Environmental Preferences. Theory and Practice of the
Contingent Valuation Method in the US, EU and Developing Countries, Oxford: Oxford University
Press.
Payne, J.W. and Bettman, J.R. (1999) ‘Measuring constructed preferences: towards a building
code’, Journal of Risk and Uncertainty, 19: 243–70.
Peterken, G.F. and Hughes, F.M.R. (1995) ‘Restoration of floodplain forests in Britain’,
Forestry, 68: 187–202.
Petts, G. (1990) ‘Forested river corridors: a lost resource’, in D. Cosgrove and G. Petts,
(eds) Water, Engineering and Landscape: Water Control and Landscape Transformation in the Modern
Period, London: Belhaven Press: 12–34.
Pretty, J., Guijit, I., Thompson, J. and Scoones, I. (1995) Participatory Learning and Action: A
Trainers Guide, London: International Institute for Environment and Development.
Ready, R., Buzby, J.C. and Hu, D. (1996) ‘Differences between continuous and discrete
contingent valuation estimates’, Land Economics, 72: 397–411.
Sagoff, M. (1988) The Economy of the Earth: Philosophy, Law, and the Environment, Cambridge:
look beyond contingent pricing’, Ecological Economics, 24: 213–30.
questions: a verbal protocol analysis of willingness to pay for an environmental
regulation’, Journal of Environmental Economics and Management, 26(1): 88–109.
Stewart, J., Kendall, E. and Coote, A. (1994) Citizens’ Juries, London: Institute of Public
Policy Research.
United Nations (1993) The Global Partnership for the Environment and Development. A Guide to
Agenda 21, post-Rio edition, New York: United Nations.
Yon, D. and Tendron, G. (1981) Alluvial Forests of Europe, Nature and Environment Series
200, Strasbourg: Council of Europe.
Deliberation and economic valuation 225
10 Deliberation and economic

valuation
National park management
Rosemary F. James and Russell K. Blamey
Introduction
Among the wide variety of methods available for incorporating community attitudes
and values into environmental decision-making are public hearings, calls for sub-
missions, referenda and opinion polls. Each of these approaches has its limitations.
Concerns also surround the use of stated preference techniques such as the
contingent valuation method (CVM) and choice modelling as inputs to environ-
mental decision-making, in part because responses are based on limited levels of
information and deliberation. Techniques based on the theory of deliberative
democracy have the potential to correct these adverse conditions. Citizens’ juries
(CJs), deliberative polls, consensus conferences and related techniques offer
participants time, information and opportunities for social interaction and
deliberation, which may result in the development and articulation of robust
preferences.1
Under the tenets of participatory democratic theory, public participation is
functionally, logically and ethically central to the operation of democracy (Webler
and Renn 1995). As Thomas-Slatyer (1995: 11) noted (in relation to community
participation in project development in developing countries): ‘Central to these
approaches is the belief that ordinary people are capable of critical reflection and
analysis and that their knowledge is relevant and necessary.’ Public participation is
required in order to ensure that the aspirations and needs of citizens are heard by
decision-makers and is considered to produce many benefits including definition
and communication of the public will (Rousseau 1968), popular sovereignty
(Rosenbaum 1978), political equality (Rosenbaum 1978, Van Valey and Petersen
1987), protection of citizens’ interests (Rosenbaum 1978), definition of the public
will (Dienel 1989) and enhancement of personal (Daneke et al. 1983) and social
(Rosenbaum 1979) development. An experimental application of the CJ method
to economic valuation is described in this chapter. The theoretical basis of the
method is first described, followed by a detailed description of this application of
the method. Some of the methodological and theoretical issues which arose from
this work are discussed.
226 Rosemary F. James and Russell K. Blamey
Public participation through cost–benefit analysis
There are numerous methods commonly used in developed countries to allow
public input into environmental decision-making, including methods in which
respondents are wholly self-selected (such as calls for submissions) and those in
which potential respondents are identified by some other agency (such as opinion
polls). Another form of public participation is provided by the market, where
people’s preferences are reflected in their market transactions. In the absence of a
market, surrogate markets and economic valuation surveys may provide information
regarding the preferences of citizens. An assessment of the value (benefits or costs)
accruing to the community from an actual or proposed course of action is an
important component of environmental cost–benefit analysis (CBA), of which the
objective is to inform decisions on the allocation of resources so as to maximize
the excess of benefits over costs to society as a whole.
The basis of the practice of environmental CBA lies in a form of preference
utilitarianism in which it is assumed that preferences reflect utility functions, and
willingness to pay (WTP) or willingness to accept (WTA) is taken as an indicator
of strength of preference. The democratic basis of environmental CBA and
associated estimates of community value can thus be summarized as ‘one dollar,
one vote’. Environmental CBA involves an assessment of the implications for
community environmental preferences, typically in WTP terms, of proposals under
investigation, which may have impacts on the natural environment. Stated
preference methods alone are potentially capable of estimating the full range of
environmental values, while revealed preference methods only provide information
on direct use values (see Chapter 1).
Stated preference surveys involve eliciting information pertaining to strength
of preference through the use of community surveys in which respondents are
requested to state what they would do under prescribed circumstances. Respondents
are questioned in such a manner as to elicit their WTP for a programme involving
some potential environmental improvement (or prevention of some potential
environmental deterioration) or WTA compensation for a programme involving
some environmental deterioration or lack of environmental improvement. The
two most commonly used stated preference methods are the CVM and choice
modelling. The CVM, and to a lesser extent choice modelling,2 have been criticized
on several grounds.
One argument is that ‘stated preference surveys represent a highly constrained
environment for preference construction’. Whilst some respondents to stated
preference questionnaires may reveal their pre-existing preferences for the attributes
and/or goods and services under examination, it is probably more common for
respondents to construct their preferences at the time of completion of the
questionnaire. Environmental goods and services are complex and often unfamiliar,
and the method chosen to elicit WTP should allow respondents the opportunity to
articulate their values for the goods and services under consideration and then to
construct their preferences. Contingent valuation in particular is not seen as a
method which permits this to occur (Gregory et al. 1993; Gregory and Slovic 1997;
Sagoff 1998). CVM respondents are required to make choices in the face of
necessarily limited time and information and in the absence of the capacity to
seek clarification of many of the issues of concern, which may severely constrain
the construction of informed and deliberated preferences.
Stated preference surveys, in common with other types of surveys, are subject
to coverage, sampling, non-response and measurement error. Much, if not most,
of the controversy surrounding techniques such as CVM and choice modelling
focuses on measurement errors, particularly those attributable to respondents and
their interactions with the questionnaire. Sources of such error include yea-saying
and nay-saying (when respondents answer positively or negatively to valuation
questions without regard to the amount involved in the scenario), the non-
identification of protest responses (objections to the use of a tax or levy to fund
environmental improvements), strategic responses and amenity misspecification
biases. Methodological criticisms of the CVM include the production of
unacceptably large estimates for total WTP or WTA and the insensitivity of WTP
or WTA estimates to price and scope variations (Perman et al. 1999).
There are a range of philosophical concerns associated with environmental
valuation surveys. These relate both to the use of the methodology itself and to
respondent motivations and decision processes (Blamey and Common 1999). The
‘one dollar, one vote’ assumption of environmental CBA raises the question as to
whether the wealthy should have more say than others. Ethical concerns are raised
regarding the valuation of the non-monetary components of environmental goods
in monetary terms (Blamey and Common 1992; O’Neill 1997; Gregory and Slovic
1997). The potentially reductionist and atomistic nature of stated preference methods
is seen as problematic. Alternative methods which are more pluralistic in terms of
community perspectives and values may offer a more appropriate and legitimate
input into environmental decision-making. Environmental issues may then best be
addressed by individuals acting as citizens with a societal perspective, rather than as
consumers acting on the basis of individual preferences (Sagoff 1998).
Deliberative and discursive democracy

In the last 30 years or so, several public participation methods have been developed
which may surmount some or all of these difficulties and thus provide for more
effective environmental value assessment. These methods, founded in deliberative
and discursive democracy,3 provide participants with the opportunity for the
collection and analysis of information and for discussion and deliberation of the
issue under consideration. Those participating in deliberative processes are
characterized by Dryzek as ‘amenable to changing their minds and their preferences
as a result of the reflection induced by deliberation’ (Dryzek 2000: 31). Deliberative
methods include consensus conferences, citizens’ juries, deliberative polls, planning
cells, citizen advisory committees, study groups and many more. Processes such as
these which have greater claims to being reflective, evolutionary, empowering and
conflict-resolving may offer significant advantages over stand-alone stated
preference methods in the determination of environmental values. Deliberative
processes may also be more acknowledging and accommodating of the complexity
and uncertainty inherent in environmental problems and decision-making. The
recent application of one of these methods is the CJ.
Peter Dienel in Germany and Ned Crosby in the United States independently
developed similar techniques for deliberative input by members of the public to
decision-making on matters of public policy significance. The Planungszelle, or
planning cell, method, was developed in Germany in 1969 and the CJ method
was developed in 1971 in the United States (Crosby 1996). The CJ method has
since been used in many parts of the world in numerous types of application.
Recent publications describing citizens’ juries include Jefferson Center (1995, 1996,
1997), Coote and Lenaghan (1997) and Aldred and Jacobs (2000). The principal
features of the CJ method have been well detailed in Crosby (1991, 1996). Essent-
ially, a small group representing a larger population is selected, most usually by
quota sampling, and asked to spend several days as part of the CJ. The CJ hears
and questions witnesses, discusses and deliberates on a pre-specified charge, and
reaches a conclusion. The result is, under ideal circumstances, consensual.
The CJ technique may be applied in a number of ways to enable public partici-
pation in decision-making on environmental and other matters. The charge may
be structured to provide for:
1 assessment of support: the level of support for a proposed action is determined,

e.g. Does the jury think the harbour area should be developed?;
2 information gathering: the views of participants are sought on some matter,
e.g. What does the jury think are the most important aspects of the harbour?
or What does the jury think are the most important impacts of the proposed
harbour development?;
3 option choice: a choice is made between several options, which may include
cost data, in regard to a proposed development, e.g. Does the jury think the
harbour should be developed and, if so, which is the preferred development
plan?;
4 deliberative valuation: the valuation of some proposed change in resource
allocation or access using stated preference or other means, e.g. for anticipated
declines in resource allocation or access: What is the maximum amount the
jury would be prepared to pay to prevent the proposed harbour development?
or What is the minimum amount the jury would be prepared to accept as
compensation for the proposed harbour development?4
The potential for use of the fourth approach, deliberative valuation, has been
discussed under various names by Brown et al. (1995), Jacobs (1997) and Sagoff
(1998), but no examples of the practice appear to be reported in the literature to
date (although see other chapters in this volume). The conduct of a stated preference
method in the context of a CJ has the potential to reduce some of the difficulties
associated with stated preference methods. The use of a discursive rather than
questionnaire-based stated preference valuation method in a CJ may provide
additional benefit.
Design and conduct of this application
The CJ reported in this chapter was conducted as a research application of the
technique, rather than commissioned by a government agency in response to an
existing environmental issue. The specific topic examined in the CJ was the
management of national parks5 in the most populous Australian state, New South
Wales. This CJ formed part of an examination of the potential for use of the
technique in environmental decision-making in Australia.
There are several ways in which the CJ technique can be applied to non-market
valuation, i.e. used for deliberative valuation. These are illustrated in Figure 10.1.
The suite of work undertaken in this project is shown in italics. The CJ reported
herein included the following elements: a conventional option choice charge which
did not involve deliberative valuation; two approaches to deliberative valuation,
involving a further charge in which the jurors were asked to determine societal
WTP for a specified programme involving environmental improvement; and a
stated preference (choice modelling) survey, completed by the individual jurors
before and after the CJ proceedings. The principal focus of this chapter is on the
first of these deliberative valuations, in which a process of group discussion and
decision-making was used to arrive at a valuation (WTP) for a posited programme
involving environmental improvement.
A scenario based on consideration of five major national park management
activities was developed and then refined using focus groups. The scenario specified
options differing in terms of the levels of effort allocated to the following key
national park management activities: fire management, weed control, feral animal
control, maintenance of visitor facilities and management of historic sites.6 The
relationships between park management activities and biophysical attributes such
as wildlife and native plants were detailed in an explanatory document provided
to the jurors. The same scenario format was used for both the conventional and
deliberative valuation charges and the choice modelling questionnaire. Further,
the same base allocation of effort (Option 1), representing the current situation,
was used in all three approaches. The charges are shown in Table 10.1 and
Table 10.2.
The jury was given a set of information and asked which of three options they
preferred before the introduction of a fourth option. They were told that the
National Parks and Wildlife Service manages about 5 million hectares of parks in
New South Wales (6.3 per cent of the state) in 479 parks and reserves. This includes
national parks, nature reserves, historic sites, state recreation areas and regional
parks. They were told that the term ‘national park’ included all these different
types of parks and reserves. They were told that budgets for national park
management are always tight and that there would never be enough money to do
all the things which might be required to meet the expectations of the public
about national park management. A possibility was described as to continue
spending the existing budget as shown in Figure 10.2 and in Option 1. Further
information about these management activities was provided in a pamphlet.
Another possibility described was to keep the total amount of money spent on
Citizens’ Jury
Without monetary valuation -

With monetary valuation conventional citizens' jury
(CJ constructions (i) - (iii))
Using conventional stated Using other valuation

preference methods methods
(CJ construction (iv))
Framing of charge
Jury to estimate WTP for a
specified option
Jury to select preferred
option from several given;
each option includes a cost
per potential payer
Framing condition Framing condition

Citizen Citizen
Ambiguous/not specified Ambiguous/not specified
Consumer Consumer
Response basis Response basis

Group response Group response
Individual response Individual response
Figure 10.1 CJs applied to non-market valuation
national park management the same, but change how it is spent. For example,
some money could be taken away from maintaining facilities and spent on feral
animal control. This is what is proposed in Options 2 and 3 of the charge.
The jurors were told that another possibility was to raise more money to spend
on national parks. State governments could introduce a park management levy to
help pay for increased levels of current activities. The levy would be paid by all
Table 10.1 Charge 1: CJ group preference among three options
Outcomes of national park management Option 1 Option 2 Option 3
current
situation
Number of national parks with good fire 100 40 160
management
Area of feral animal control each year 50,000 ha 100,000 ha 30,000 ha
(hectares)
Area of weed control each year (hectares) 3,000 ha 1,000 ha 10,000 ha
Proportion of facilities that are well 35% 45% 25%
maintained
Number of well-protected historic sites 7,000 6,000 7,500
Table 10.2 Charge 2: introducing Option 4

Outcomes of national park management Option 1 Option 4
current
situation
Number of national parks with good fire management 100 400
Area of feral animal control each year (hectares) 50,000 ha 160,000 ha
Area of weed control each year (hectares) 3,000 ha 9,600 ha
Proportion of facilities that are well maintained 35% 100%
Number of well-protected historic sites 7,000 7,900
Recurrent levy on income tax $0 $??
those who live in NSW and pay income tax, and would be a set amount per year,
paid as part of the income tax return system. They were then told: ‘Another way
of raising extra funds would be to introduce entrance fees for more national parks
or to increase those fees that already exist. Unfortunately, opportunities to
significantly increase the funds raised from these fees are very limited. The costs
of collecting entrance fees are often higher than the money collected. For this
reason, entrance fees are not considered in this charge. We have put together a
new option, Option 4, that involves a substantial increase in expenditure for all
five park management activities.’ The results of the straw preference poll were
used in establishing the priorities for this option. The jury was told that what
Option 4 would cost to implement is unknown. However, it would cost a lot more
money than the NSW National Parks and Wildlife Service currently has to spend
on these activities. For the purposes of this scenario, jurors were asked to assume
that the only way that Option 4 could be achieved is via the introduction of a park
management levy. They were then asked ‘How high would a park management
levy have to be before the jury would recommend Option 1 rather than Option 4?
Protection of historic sites
and items (13%)
479 parks covering
5 million hectares
Feral animal control (6%)
Maintenance of Fire management (15%)

facilities (60%)
Weed control (6%)
Figure 10.2 Spending on parks excluding salaries
In other words, how high would the levy have to be before the NSW public would
be no better off under Option 4 than Option 1?’
The population relevant to the scenario was identified as the 3.8m population
of the Australian state of New South Wales between 20 and 69 years of age. The
jurors were selected using a two-phase telephone survey, in which various
demographic and attitudinal data were collected. A sample was selected to be as
representative as possible of the population on the basis of the latest available
Australian Bureau of Statistics data on the following criteria: gender, age, place of
residence, ranking of the environment in relation to other social issues, occupation,
income, income source and education. One selected juror failed to attend.
Jurors were provided with written instructions regarding both their roles and
the roles of witnesses and with training in note-taking and strategic questioning.
All travel, meal and accommodation costs were met and the jurors were paid a
small sum (A$100 per day) to compensate them for out-of-pocket expenses incurred.
Five witnesses with expertise in particular aspects of the charges were selected (i.e.
fire, weeds, feral animals, facilities and historic sites), in addition to two general
witnesses on national park management. Witnesses were provided with written
instructions regarding the content of their presentations, their roles and the roles
of the jurors in the jury process.
The jury met over the period 5–7 October 1999. The first day of the process
was devoted to preparation of the jurors and the jury for their roles over the
following two days. The second day was largely occupied by presentations from
and questioning of the witnesses; on the third day, the jury consulted two further
witnesses and reached decisions on the two charges. The second charge was
presented to the jury only after a decision had been reached on the first charge.
Option 4 in the second charge was developed from the results of an initial
confidential identification by the jurors of their individually preferred option from
the first charge. The jurors were aware that Option 1 in each choice set represented
the current situation. They were also informed that the results of this experimental
application of the CJ method would be relayed to the NSW National Park Service
for use in budgeting and in the design of public participation programmes.
Little has been documented about appropriate decision-making methods for
use in citizens’ juries; there does not appear to be a method consistently applied as
part of the CJ process. It was necessary, therefore, to develop a consistent, robust
method for facilitating group consensus, which would prevent false or forced
consensus and would allow for alternative strategies if a true consensus could not
be reached. In this work, different decision-making methods and decision rules
were used for the two charges. The processes followed are outlined at Figure 10.3
and Figure 10.4. For the first charge, jurors were asked to choose their preferred
option privately and to provide written reasons for this choice. The results of this
work were then displayed without citation of the source of the individual comments.
This confidential approach was adopted in order to ensure the jurors first
determined their own preferences and then considered the reasons cited by others
for their choices without consideration of their authorship. This was done in order
to avoid contamination of the ideas by jurors’ perceptions of the status of and
personal relationships with the author. As anticipated, this was effective for a short
time, after which the individual jurors in the minority (those who did not select
Option 1) declared their choices and the discursive process commenced. Assessment
of support for each option was then discussed openly by the group and consensus
reached. When the second charge was presented, the approach adopted for the
first charge was modified in anticipation of the need for the jurors to express their
views regarding both the proposed introduction of some type of payment and the
payment type proposed. Following considerable discussion by the jurors, public
polling on the options was conducted in several stages.
Results
After their initial and independent consideration of the options, nine jurors chose
Option 1, one chose Option 2 and three chose Option 3 as their preferred option.
After consideration of the reasons for the choice and discussion amongst the jurors,
the jury reached consensus and selected Option 1 (the status quo) as the preferred
option. The outcome was stated by the jury as follows:
We reached a consensus on Option 1. It involved the four people who in the

straw poll had expressed a preference for either Option 2 or Option 3. We
listened to their points of view about the issues they had, and they agreed
that, for the purpose of consensus, they would be prepared to move across to
Option 1. All of us shared some of the concerns that were identified by those
four people. For that reason, as part of the consensual process, we sought to
add what we call a ‘rider’, or ‘tag’ on to the consensual decision, which comes
in two parts:
1 that the feral animal, weed and fire experts be told that we can see benefits,
based on their separate presentations to us, in their getting together. Some
of the programs have some common benefits, and we think they could all
be more efficient as a result of that liaison.
Framing statements presented

by facilitator and project team members
Silent ballot and listing of reasons for choice

by jurors
Results of ballot and lists of reasons put up on noticeboards

by facilitators
Consideration and discussion

by jury
Open discussion of individual juror's views

by jury
Consensus reached
by jury
Figure 10.3 Process for decision making under ‘charge 1’
2 that we are all very concerned about the adequacy of the programs in
Option 1, and that the level of funding which permits that type of com-
bination of programs is inadequate. We would like to send a strong message
to the politicians that, as we are a representative jury in the community,
there ought to be more funding, which could be put to better use in
developing these programs.
3 A sub-issue under that (relates) to the National Parks and Wildlife Service –
it is possibly a little unfair because we have not heard from anyone from
the National Parks and Wildlife Service – is that we can see from what the
experts told us that better management would possibly enhance the way in
which these programs work.
Framing statements
by the facilitator and project team members
Discussion re tax –
occurrence
methods of collection
by jury and facilitator
Open discussion of individual juror's views

by jury
Vote then conditional consensus reached

by the jury
Figurer 10.4 Process for decision making under ‘charge 2’
The second charge required the jury to undertake a deliberative valuation. The
jury entered into considerable debate, in stages as follows: (i) discussion of the
charge and its interpretation – some jurors had initial difficulty with the idea of
maximum willingness to pay; (ii) clarification regarding the timing of collection of
the proposed levy (weekly or annual) and the eventual use of the funds collected;
(iii) discussion about the way in which the levy should be determined; one juror
put a contribution model7 in the jurors’ minds by stating ‘If we don’t know how
much it is going to cost for each one of those (activities), how do we know the
amount of the levy to impose to cover the costs?’;8 (iv) discussion about the
acceptability of the levy to jurors, the type of levy (flat or scaled rate) and other
potential sources of money to fund Option 4; and (v) discussion amongst pro-levy
jurors about an appropriate type and amount of levy. At this point, in a public
poll, the jury voted nine to four in favour of Option 4 (the ‘with levy’ option).
Further discussion followed regarding the nature of the levy, during the course of
which one additional juror came to support a levy. A lengthy discussion then
followed about appropriate fixed levy amounts. The jurors who now supported a
levy eventually voted ten to zero in favour of a proportional rather than a fixed
price levy. The levy was to be calculated as a percentage of gross income. Following
a discussion of appropriate percentage figures, a vote was taken to determine
support for the two figures which had been proposed by the jurors, 0.1 per cent
and 0.25 per cent of gross income. A majority rule preference of a 0.1 per cent
levy on income tax resulted, by a vote of eight to two. Three jurors maintained
their objection to the use of a levy. Using the proposed levy of 0.1 per cent, an
additional A$109.7m (1997/8 income data), equivalent to €70.2m, would be
collected annually for use on national park management. Total annual expenditure
on the five programmes was A$15.2m in 1999–2000.
Conclusions
The conduct of this deliberative valuation exercise has raised a number of
significant methodological and theoretical issues which require resolution in order
that the method may be rendered more rigorous. These include methodological
issues covering consensus formation, decision rules for polling or voting, equality
of juror impact, and provision of information. Theoretical issues which arose
concern framing, representation, non-consensual outcomes, and the economic
interpretation of the results.
The development of a consensual decision is considered the ideal outcome of
a CJ, although alternative approaches can be accommodated. The selection of an
appropriate concluding process is necessarily strongly influenced by the wording
of the charge; as is illustrated by the case studies discussed by Coote and Lenaghan
(1997) it may involve voting, listing of information or some other method. The
importance accorded to consensus in the CJ method suggests a need for clarity
and exactitude in identifying consensus. When is a consensus a consensus?
Consensus requires that the eventual decision is agreed to by all members of the
group (Center for Conflict Resolution 1981). As this jury considered the first charge,
it was apparent that jurors differed markedly in their degree of participation in
the discussion of the options. This is hardly surprising; the same situation pertains
to most circumstances where a group is given a complex task to complete. However,
the task in a CJ is to reach a consensual decision; thus there should be clear evidence
that the group decision was willingly accepted by all and that apparently consensual
outcomes are not false. Factors which may lead to false consensus include: compliant
behaviour by jurors whose initial views are in the minority, driven by, for example,
a need to accord with a group norm or with the views of particular jurors
(conformity bias); and failure to participate in the process by individual jurors (due
to, for example, lack of communication or interpersonal skills, or to irritation or
boredom with the process). A protocol to reach consensus and robust methods of
testing the validity of apparently consensual outcomes are required in order to
prevent acceptance of false consensus. Such testing should necessarily be capable
of distinguishing between the ‘workable agreements’ referred to by Dryzek (2000:
48), arrived at by participants agreeing to a common outcome for different reasons
(i.e. content-related) and an apparent but false consensus developed in response to
some failure within the process of a particular CJ in response to factors such as
juror boredom (i.e. process-related). One possible approach would be to examine
the decision-making processes of individual jurors by means of a post-decision
questionnaire. Further, juries may fail to arrive at a consensus. At what point should
the possibility of a non-consensual decision be revealed to the jury? If this is done
too early in the deliberation phase, abandonment of the consensus-seeking process
may result. If done too late, a false consensus may have been reached. In addition,
at what point in the deliberation and under what conditions should the jury be
encouraged to implement other decision rules? What alternative decision rules
should be used?
As noted above, if a jury fails to reach a consensus, alternative decision rules
are necessary. If these involve voting or polling, care is required to establish an
appropriate procedure, as the polling process may affect the outcomes of a CJ
significantly. There are four major methodological aspects of polling which
require consideration: the form of the question posed, the polling method used,
the structure of the deliberative and polling processes and the decision rule
employed. The question to be considered may be framed in one of several forms,
dependent in part on the construction of the charge. Possible approaches include:
a choice between ‘yes’/‘no’/‘abstention’; forced choice ballots, in which all
answers but ‘yes’ are taken as ‘no’; and various forms of rating and ranking
(Gastil 1993: 56–60). The polling method used – confidential or public polling –
may also influence poll results. The use of confidential straw or preliminary
polls (as used in initial consideration of the first charge in this CJ) would seem to
offer advantages, by allowing jurors to make their initial choice uninfluenced by
the choices of others. The use of a subsequent anonymous tabling of reasons for
these preliminary choices permits analysis of those reasons unclouded by social
dynamics associated with the status of and personal feelings towards the source.
The structure of the deliberative and polling processes may also influence the
poll outcome. For example, multiple polls separated by deliberative phases are
more likely to lead to consensus on controversial issues than a single poll, by
allowing those holding differing views to consider and reflect on both their own
views and those of other jurors. The decision rule to be applied also requires
consideration. Should a simple majority vote, some proportional threshold or
some other measure be used? The decision on this aspect of polling will clearly
exert a strong influence on the poll results.
Bohman noted that deliberation may be considered to have been successful
when ‘participants in the joint activity recognize that they have contributed to and
influenced the outcome, even when they disagree with it’. He considered that
equality, at a minimum, required ‘equal standing and effective voice’ for every
citizen, both in terms of access to deliberative processes and participation within
specific processes (Bohman 1996: 33, 35). A lack of within-process equality was
evident in this CJ. Preliminary analysis shows that, of the thirteen jurors, one
generally spoke only when directly questioned and several others made minimal
contributions to the discussions. Whilst it might be that these jurors were not
distressed by their low profile in the jury and their minimal impact on the outcome
(and may have in fact agreed with the decisions reached), their lack of contribution
suggests a less than ideal deliberative process. The sources of this within-process
inequality lie in the wide range of social and cognitive skills and capacities, and of
prior knowledge, found in a jury derived from effective sampling of a diverse
population. To what extent it is desirable and possible to reduce this inequality is
a fundamental issue affecting the legitimacy of both deliberative and non-
deliberative public participation methods.
Two key aspects of the provision of information to jurors are: the appropriate
level of juror access to witnesses; and how best to deal with inaccurate or misleading
information. Witnesses will vary both in their personal appeal to the jury and the
level of interest engendered by their subject matter. Hence some witnesses may be
sought after by the jury to provide additional information. Equality of access
appears reasonable on procedural grounds. However, the purpose of the witness
presentations is to provide the jurors with the information they consider necessary
to address the charge. Hence, one could argue for unlimited jury access to each
witness. In adversarial citizens’ juries, however, interested parties who consider
themselves disadvantaged by relative lack of access to the jury could claim the
outcome was biased. On the other hand, refusing the jury access to additional
information from witnesses could be considered to subvert the jury process, by
forcing jurors to reach a decision based on potentially inadequate information.
Further, witnesses will vary in their capacity to provide information to the jury, in
terms of accuracy, relevance and pedagogic effectiveness. Means are required to
deal with these potential distortions in information delivery. The provision of a
neutral expert to assist the jury in assessing information provided by witnesses and
the use of some type of discovery process (as used in the legal system) have been
proposed as possible mechanisms for dealing with inaccurate, misleading or
irrelevant information (Crosby 1995). The effectiveness of these approaches has
yet to be assessed in practice.
A citizen rather than a consumer frame is considered appropriate for most
citizens’ juries, although it may produce results differing from the estimates of
Hicksian measures of welfare change sought for environmental CBA. A citizen
frame, for example, encourages participants to think about costs and benefits in
aggregate societal terms which may not fit well with the notion of individual WTP.
Further, to the extent that individuals do consider costs and benefits at the individual
level, they may adopt a contribution model rather than the purchase model typically
assumed in environmental economics. This is problematical for environmental
CBA, as are several other features of citizen responses to valuation questionnaires
(Blamey 1996, 1998). In the study reported in this chapter, the charges were framed
so as to have jurors provide WTP estimates according to a purchase model and
hence potentially of use to environmental CBA, whilst maintaining a citizen rather
than consumer perspective.
A CJ engaged in deliberative valuation may provide either a maximum WTP
or a minimum WTA figure for some course of action or programme involving
altered levels of environmental goods and services. How should that figure be
extended to reflect the societal valuation? The most obvious approach to this issue
is to extrapolate the figure across the relevant population, as was done for this CJ.
However, the question of the degree of representation of the sample arises in this
context. If the jurors are acting fully as citizens – behind a Rawlsian veil of ignorance
– and have perfect knowledge of the full range of relevant societal circumstances
and views, then representation is of limited importance in terms of determination
of the societal view on the issue under consideration.9 That is, if the construction
of the citizen frame of reference has been successful and the jurors are clearly
deliberating and deciding in a citizen rather than an individual consumer context,
the influence of juror characteristics may be low, if sufficient information about
the relevant circumstances of other citizens is available to them. Relevant knowledge
would include details of the costs of substitute and complementary goods and
services and the amount of discretionary expenditure possible for all those potent-
ially liable to pay for the posited environmental change. This is of particular
importance where the jury is not representative of the population in terms of
factors such as income which may exert a strong influence on WTP or WTA.10 If
this is not the case, however, and a jury is selected which is not representative of
the wider population in terms of income and other characteristics and has less
than perfect knowledge of societal circumstances, biased WTP or WTA estimates
may result. These estimates may not reflect the WTP/WTA which would apply to
the population under the same conditions. The importance of this effect will be
determined by several factors, including: the degree to which the jurors are acting
as citizens; their knowledge of the relevant circumstances of the population; and
the impact of any ignorance of relevant circumstances on the resultant collective
estimate of WTP or WTA.
Decisions reached via voting are unlikely to satisfy the tenets of neoclassical
economics because the strength of each juror’s preference is not taken into account
in reaching an outcome. An argument can perhaps be made for the use of some
kind of averaging process under these circumstances. In this CJ, three jurors were
not in favour of any payment and ten were in favour of a payment of 0.1 per cent
of taxable income. Should an average WTP be determined in those circum-
stances?11 Or should the majority view of 0.1 per cent be accepted and used in
estimation of the societal WTP?
The interpretation of WTP recommendations reached by consensus raises some
issues also associated with results from applications of stated preference methods.
These include: the transferability of results; the consensual WTP determined by a
jury represents the maximum levy any individual would agree to after having
participated in a CJ on the topic under the same conditions of information,
deliberation, decisiveness and responsibility; miscounting of benefits and costs;
altruistic motivations and inclusion of bequest values by jurors may result in biased
estimates of WTP (see Quiggin 1998 and Milgrom 1993). For reasons such as
these, the results of deliberative valuation exercises may be better interpreted in
terms of social welfare functions and social optimality than individual utility
functions and Pareto optimality. Using this approach, the results of a deliberative
valuation WTP exercise should be interpreted as the maximum amount of money
which could be charged to members of the public using a given payment vehicle
in order to obtain the posited environmental improvement programme without
leaving social welfare as a whole worse off than prior to the change. This is an
advantage of deliberative valuation over stand-alone applications of stated
preference methods such as the CVM.
The application of the CJ method to environmental decision-making and
specifically to deliberative valuation of environmental goods and services has been
considered in this chapter. The use of deliberative valuation clearly has potential
to overcome some of the deficiencies evident in valuation methods such as the
CVM. The provision of information, and of time and opportunity for deliberation,
allow for the development and expression of more robustly-held preferences than
is considered to be the case in both CVM and choice modelling. However, some
of the difficulties associated with the CVM and, to a lesser extent, choice modelling,
still remain with deliberative valuation. Further, the use of the CJ method in deliber-
ative valuation raises some additional methodological and theoretical issues. Aspects
including representation, management of the jury process to ensure equity of
participation, robust and effective decision rules and minimization of various forms
of bias, all require the development of appropriate protocols. It is likely that research
into the management and behaviour of legal juries will prove to be a rich source
of procedural guidance in regard to these matters, particularly in relation to the
conduct of adversarial citizens’ juries. There is also no doubt that potential future
litigation surrounding the results of adversarial citizens’ juries will provide much
impetus to the development and refinement of the method.
Acknowledgements
Mick Common of the University of Strathclyde originated this project whilst at
the Australian National University and has provided valuable advice during its
course. Michael Lockwood of Charles Stuart University and Rebecca Smith of
the Australian National University assisted in the conduct of the CJ. Mick Common,
David Stern and Jonathan Aldred provided useful comments on this material.
Notes
1 Writing with reference to aid projects, Slocum and Thomas-Slatyer (1995: 3) defined public or
popular participation to be ‘… active involvement of people in making decisions about the
implementation of processes, programmes and projects which affect them’. The definition applies
equally well to projects in developed countries.
2 Whilst choice modelling may provide a means to address some of the flaws cited with the
CVM, the extent to which this is so remains to be established. To date, choice modelling in
environmental applications has not been subject to the same degree of rigorous examination
and litigious critique as CVM.
3 Whilst these terms tend to be used interchangeably, Dryzek’s (1990) discursive theory has stronger
roots in the critical theory of Habermas (Dryzek pers. com.; Blaug 1999). Dryzek (2000) discusses
the distinction in some detail.
4 Constructions (i), (iii) and (iv) may result in adversarial proceedings, in which the witnesses and
perhaps the jurors perceive advantage from strategic behaviour and act accordingly. Hence
careful management of the jury process is necessary in order to minimize bias.
5 The term was used to describe several categories of land managed by the NSW National Parks
and Wildlife Service including national parks, nature reserves, state recreation areas and regional
parks.
6 Fire management in this context referred to the preparation for wildfires; this included training
of staff, controlled burning and the maintenance of fire trails and fire breaks. Weed and feral
(or pest) animal control are necessary in order to protect the natural biophysical resources of
the national parks from damage.
7 Repeated framing was necessary to ensure the jurors were operating on a purchase rather than
a contribution model. The distinction, as enunciated by Kahneman et al. (1993), is critical;
respondents must focus on the amount they are WTP or WTA when purchasing the posited
environmental change, in order to enable estimation of societal WTP or WTA. If respondents
instead adopt a contribution model, under which they nominate their willingness to contribute
to achieve environmental change, the resultant amounts cannot be aggregated to provide
estimates of the usual Hicksian measures of welfare change.
8 The jurors were unaware of the current levels of expenditure on the five management activities
during their consideration of the two charges.
9 It may, nonetheless, be of political importance in gaining widespread acceptance of the results
of the jury’s deliberations.
10 For example, this jury was under-representative in the lowest age (20–29 years) and income (less
than A$10,400 per annum) categories. Can the WTP figure from this jury be validly extrapolated
to the population? In this case, it is perhaps considered legitimate to do so, as the amount to be
paid by each individual was set at a percentage of taxable income. In situations where juries opt
for absolute payments rather than payments scaled in relation to income, however, the
development of societal WTP or WTA by extrapolation from a jury value is more questionable.
11 Using this approach, a per-taxpayer contribution of 0.077 per cent of taxable income resulted.
The resultant WTP for the improvements posited under Option 4 would be A$84.5m.
References
Aldred, J. and Jacobs, M. (2000) ‘Citizens and wetlands: evaluating the Ely CJ’, Ecological
Economics, 34 (2): 217–32.
Blamey, R.K. (1996) ‘Citizens, consumers and contingent valuation: clarification and
expression of citizen values and issue opinions’, in W. Adamowicz, P. Boxall, M.K.
Luckert, W.E. Phillips and W.A. White (eds) Forestry, Economics and the Environment,
Wallingford, UK: CAB International: 103–33.
Blamey, R.K. (1998) ‘Contingent valuation and the activation of environmental norms’,
Ecological Economics, 24: 47–72.
Blamey, R.K. and Common, M. (1992) ‘Sustainability and the limits to pseudo-market
valuation’, in M. Lockwood and T. De Lacy (eds) Valuing Natural Areas: Applications and
Problems of the Contingent Valuation Method, Albury, Australia: Johnstone Centre, Charles
Stuart University: 117–46.
Blamey, R.K. and Common, M. (1999) ‘Valuation and ethics in environmental economics’,
in J. van den Bergh (ed.) Handbook of Environmental and Resource Economics, Northhampton:
Blaug, R. (1999) Democracy, Real and Ideal: Discourse Ethics and Radical Politics, Albany, NY:
State University of New York.
Bohman, J. (1996) Public Deliberation, Cambridge, MA: MIT Press.
Brown, T.C., Peterson, G.L. and Tonn, B.E. (1995) ‘The values jury to aid natural resource
decisions’, Land Economics, 71(2): 250–60.
Center for Conflict Resolution (1981) ‘Building united judgement’, Madison, WI: Center
for Conflict Resolution.
Coote, A. and Lenaghan, J. (1997) ‘Citizens’ juries: theory into practice’, London: Institute
for Public Policy Research.
Crosby, N. (1991) ‘Citizens’ juries as a basic democratic reform’, Minneapolis, MN: Jefferson
Center.
Crosby, N. (1995) ‘Citizen’s juries: one solution for difficult environmental questions’, in
O. Renn, T. Webler and P. Wiedemann (eds) Fairness and Competence in Citizen Participation:
Evaluating Models for Environmental Discourse, Dordrecht: Kluwer: 157–74.
Crosby, N. (1996) ‘Creating an authentic voice of the people’, paper presented to the
Panel on Deliberation on Democratic Theory and Practice, Chicago, IL: Midwest
Political Science Association.
Daneke, G.A., Garcia, M.W. and Delli Priscoli, J. (eds) (1983) Public Involvement and Social
Impact Assessment, Boulder, CO: Westview Press.
Dienel, P.C. (1989) ‘Contributing to social decision methodology: citizen reports on
technological projects’, in C. Vlek and G. Cvetkovich (eds) Social Decision Methodology for
Technological Projects, Dordrecht: Kluwer: 133–51.
Dryzek, J.S. (1990) Discursive Democracy: Politics, Policy and Political Science, Cambridge:
Dryzek, J.S. (2000) Deliberative Democracy and Beyond: Liberals, Critics, Contestations, Oxford:
Gastil, J. (1993) Democracy in Small Groups: Participation, Decision Making and Communication,
Philadelphia, PA: New Society Publishers.
Gregory, R. and Slovic, P. (1997) ‘A constructive approach to environmental valuation’,
Gregory, R., Lichenstein, S. and Slovic, P. (1993) ‘Valuing environmental resources: a
constructive approach’, Journal of Risk and Uncertainty, 7: 177–97.
Jacobs, M. (1997) ‘Environmental valuation, deliberative democracy and public decision-
making institutions’, in J. Foster (ed.) Valuing Nature? Ethics, Economics and the Environment,
London, UK: Routledge: 211–31.
Jefferson Center (1995) Citizens’ Jury on Hog Farming: Final Report, Minneapolis, MN: Jefferson
Center.
Jefferson Center (1996) Citizens’ Jury on Comparing Environmental Risks, Minneapolis, MN:
Jefferson Center.
Jefferson Center (1997) Citizens’ Jury on Dakota County’s Comprehensive Land Use Plan,
Minneapolis, MN: Jefferson Center.
Kahneman, D., Ritov, I., Jacowitz, K.E. and Grant, P. (1993) ‘Stated willingness to pay for
public goods: a psychological perspective’, Psychological Science, 4–5: 310–5.
Milgrom, P. (1993) ‘Is sympathy an economic value? Philosophy, economics and the
contingent valuation method’, in J.A. Hausman (ed.) Contingent Valuation: A Critical
Assessment, Amsterdam: North-Holland: 417–40.
O’Neill, J. (1997) ‘Value pluralism, incommensurability and institutions’, in J. Foster (ed.)
Valuing Nature? Ethics, Economics and the Environment, London: Routledge: 75–88.
Perman, R., Ma, Y., McGilvray, J. and Common, M. (1999) Natural Resource and Environmental
Economics, New York: Pearson Education.
Quiggin, J. (1998) ‘Existence value and the contingent valuation method’, Australian Economic
Papers, 37(3): 312–29.
Rosenbaum, N. (1978) ‘Citizen participation and democratic theory’, in S. Langton (ed.)
Citizen Participation in America, Lexington, KY: Lexington Books: 43–54.
Rosenbaum, W. (1979) ‘Elitism and social participation’, in Lincoln Filene Center for
Citizenship and Public Affairs Citizen Participation Perspectives, proceedings of the National
Conference on Citizen Participation, Tufts University, Washington DC, 28 Sept.–1
Oct. 1978.
Rousseau, J. (1968) The Social Contract, Harmondsworth: Penguin.
look beyond contingent valuation’, Ecological Economics, 24: 213–30.
Slocum, R. and Thomas-Slatyer, B. (1995) ‘Participation, empowerment and sustainable
development’, in R. Slocum, L. Wichhart, D. Rocheleau and B. Thomas-Slatyer (eds)
Power, Process and Participation – Tools for Change, London: Intermediate Technology
Publications: 3–8.
Thomas-Slatyer, B. (1995) ‘A brief history of participatory methodologies’, in R. Slocum,
L. Wichhart, D. Rocheleau and B. Thomas-Slatyer (eds) Power, Process and Participation –
Tools for Change, London: Intermediate Technology Publications: 9–16.
Van Valey, T.L. and Petersen, J.C. (1987) ‘Public service science centers: the Michigan
experience’, in J. DeSario and S. Langton (eds) Citizen Participation in Public Decision Making,
Westport: Greenwood Press: 39–63.
Webler, T. and Renn, O. (1995) ‘A brief primer on participation: philosophy and practice’,
in O. Renn, T. Webler and P. Wiedemann (eds) Fairness and Competence in Citizen Participation:
Evaluating Models for Environmental Discourse, Dordrecht: Kluwer: 17–33.
244 Katja Arzt
11 The challenges of
stakeholder participation
Agri-environmental policy
Katja Arzt
Introduction
Some environmental economics approaches based on neo-classical theory attempt
to reveal people’s values related to the natural environment by aggregating single
statements, e.g. the contingent valuation method (CVM). Such approaches are
based on the assumption that a person has complete information about the natural
good at stake and that they already have formed preferences. However, people
tend to form values and attitudes over time and may change their value perception
of the natural environment as their knowledge and experiences increase (e.g. Jiggins
and Röling 1998; O’Hara 2000). The case study presented in this chapter is based
on the hypothesis that values for nature emerge from interactions between people
(see O’Connor 1998) and could therefore be linked to Sahlins’s (1976) remark that
valuation approaches should be recognized as a subjective, cultural and contextual
phenomenon. This study is based on the observations that ecosystems are dynamic
and that many of their changes are unpredictable (see Hanna et al. 1996; Berkes
and Folke 1998; Holling et al. 1998). Therefore an optimal solution for the manage-
ment of different ecosystems cannot be presumed beforehand (see Munda 2000).
Complexity and unpredictability in physical and social systems has thus to be
considered when valuing environments.
An ‘interactive valuation process’ is here defined as different stakeholders
concerned with a particular environmental issue coming together to exchange
their views and values in order to decide together about future programmes, plans
or targets for managing the environment. This kind of decision making pays
attention to ‘plural rationality’ (Söderbaum 1999) and thus could be termed a
participatory valuation process as different orientations and attitudes are to be
considered beside monetary assumptions. A discursive process will enable indivi-
duals, in groups, to construct values rather than express prior preferences and is
hence a deliberative approach which leaves it to the respondents to clarify among
themselves explicitly what they are valuing and why. The deliberative group can
define precisely which aspect(s) of the public good it values and by how much
(Sagoff 1998: 224). According to Sagoff (1998: 227): ‘If individuals … do not
come with predetermined preferences but must construct them, then the process
of construction may legitimately involve social learning, since this is precisely what
The challenges of stakeholder participation 245
occurs in other contexts in which people work out their values’. Hence, it is
hypothesized in this chapter that a mutual learning process is central for a successful
decision making process. If different stakeholders learn about each others’
perceptions and sets of values it is more likely that their actions can be co-ordinated
towards mutual agreements. As Ostrom (1998a: 71) has already explained for
common-pool resources: ‘When all appropriators of a common-pool resource share
a common set of values and interact within a complex set of arrangements, there
is a much greater probability that they will develop adequate rules and norms to
manage resources’.
This study supports that learning refers to the accumulation of insights into
system causes (see Meppem and Gill 1998). The intention is to promote the develop-
ment of a more unified and shared impression of the underlying processes of
cause and effect that describe the issues and problems under investigation. ‘Experts’
and citizens alike can be instrumental in the general learning process. This general
consultative process is proposed as an alternative to the more exclusive, funda-
mentally ‘expert’ driven, and certainly less learning-oriented, decision-making
framework of neo-classical economics (Meppem and Gill 1998; Jiggins and Röling
1998). Learning to communicate about values needs a process which goes beyond
the supply of information. DeBono (1993, quoted in Meppem and Gill 1998: 118)
states: ‘There are people who believe that if you get enough information then the
information will do your thinking for you, however, it is the concept through which
we perceive the information that gives it any value’.
The interactive valuation process brings different orientations together; some
participants may join the process because of individual interests while others may
do so for altruistic reasons. Hence, it is the aim of the ‘learning’ institution that the
limited number of participants in that deliberative approach act as citizens as
opposed to individuals who try to choose an outcome most benefiting their
individual interests. In other words, consumer preferences reflect conceptions of
the good life that individuals seek for themselves, while citizens’ preferences reflect
conceptions of the good society offered for the consideration and agreement of
others (Sagoff 1998: 215). A well-performed learning process is one where the
citizens’ preferences are emphasized; decision in the learning process may depend
more on the force of the better argument about public interest (Habermas 1982).
Discourse processes attempt to bring such different orientation together, and
constitute a conscious decision process, in which individuals may reconsider their
way of thinking (O’Hara 1999, 2000). A decentralized institution for valuation
may therefore have the following positive impacts compared with those valuation
methods which are based on neo-classical theory: it is a ‘learning institution’ in
which new experiences and knowledge as well as local knowledge can be integrated
continuously; decision results of such an institution are based on norms and rules
which correspond to the local, social and natural environment; results can therefore
be better accepted and transformed into action.
An interactive valuation process which aims for sustainable solutions and
sustainable actions needs an institutional design, capable of achieving a mutual
learning process. At present different institutional settings for interactive decision
246 Katja Arzt
making are being discussed. Different tools – such as citizens’ juries, consensus
conferences, roundtable groups, and focus groups – are used for different purposes;
some exist only for a short time to solve specific problems within communities
while others persist over time in order to be established as a new democratic
institution. The aim of this study is therefore to find the most significant circum-
stances aiding the creation of an institutional arrangement for decision making at
the local level. The first section draws some theoretical assumptions for a successful
institutional design. The second section presents a research project in northeast
Germany, which established a ‘roundtable group’ in order to decentralize decisions
about agri-environmental policies and schemes. In the conclusion the interactive
valuation process in this project is argued to have been only partially successful
due to an inappropriate institutional setting.
An analytical framework for interactive valuation

processes
A successful institutional design for interactive valuation based on mutual learning
is defined here as one which exists over a period of time, is able to create sustainable
decisions which are also accepted at higher levels of authorities and leads to actions
to improve the environment. Such an arrangement should be flexible in order to
react towards changing social, economic and environmental conditions. It is
successful when it meets the needs for a democratic institution in which different
stakeholders are involved deliberately. Thus success means airing and assessing
(and possibly combining) different views rather than finding an ‘optimal’ solution
of one underlying perception or view (for instance a result of a cost-benefit analysis,
or a naturalistic view). In this respect the process which leads to the decision is
important. This definition of success corresponds with Munda (2000: 12), who
says that processes for environmental valuation should be considered under a
procedural rationality, in which the process is evaluated: ‘Since an optimal solution
is non-existent, the important factor is the quality of the process leading to the
decision. Optimal solutions are hence replaced by satisfactory solutions’ (Simon
1983).
In order to create an analytical framework, a theory is needed which makes
assumptions about the way a group can reach mutual learning to attain satisfactory
decisions and actions. Here Ostrom’s (1998b) development of ‘second generation
models of rationality for social dilemma situations’ is used which is a behavioural
approach to the rational choice theory of collective action. Although the situation
discussed in this chapter differs from a social dilemma situation, the assumptions
made by this theory help to understand when co-operation in small groups can
take place, hence a mutual learning process is initiated which will lead to a successful
institutional design.
Ostrom analysed different laboratory research results of behaviour in social
dilemma situations and concluded that the core elements of co-operative behaviour
are trust, reputation and reciprocity. Only if those elements are developed positively
within the group is co-operation possible. Ostrom defined trust as the expectation
of one person about the actions of others that affects the first person’s choice,
when an action must be taken before the actions of others are known. In the
situation of decision making for environmental goods, individuals may agree upon
certain programmes or measures only when they are sure that other members will
also make compromises. For example, farmers might agree on special measures
which will be good for certain birds. However, they may only agree upon a
programme if they have the commitment from the government for monetary
compensation. In situations where trust and reciprocity among the participants is
difficult or impossible to create, then it is likely that people will not share information
with others in order to gain advantages. This phenomenon is quite common during
the implementation of agri-environmental policies. Farmers, for instance, are
knowledgeable about their land and its vegetation. Also, they often realize that
administrators know very little about it and use this fact to accept payments for
extensions, although they would never seriously consider using this land for intensive
cultivation (Hagedorn 2000). Trust affects whether an individual is willing to initiate
co-operation in the expectation that it will be reciprocated. According to Ostrom
(1998b), rational individuals enter situations with an initial probability of using
reciprocity based on their own prior training and experience. At the core of a
behavioural explanation are the links between the trust that individuals have in
others, the investment others make in trustworthy reputations, and the probability
that participants will use reciprocity norms. This mutually reinforcing core is
affected by structural variables as well as participants’ past experiences. This study
concentrates on the following structural variables which feed into that mutually
reinforcing core: continuous discourse, participating stakeholders, flow of
information, property right structure of the issues discussed and decision procedure.
One important structural variable is the possibility for the participants to have
face-to-face communication (Ostrom 1998b; Ostrom et al. 1994). With a repeated
chance to see and talk with others, a participant can assess whether they trust
others sufficiently to try to reach a simple contingent agreement regarding the
level of joint effort and its allocation (Ostrom 1998b). Continuous discourse there-
fore raises the possibility to co-operate, in order to exchange mutual commitments,
creating and reinforcing norms and developing group identity. Given the complexity
of the physical world that individuals frequently confront, they are rarely ever able
to get the rules ‘right’ on the first or second try (Ostrom 1990). In highly
unpredictable environments a long period of trial and error is needed before
individuals can find rules that generate substantial positive returns over a sufficiently
long time-horizon (Ostrom 1998b).
Another variable which feeds into the mutually reinforcing core is certainly the
mixture of stakeholder representatives participating in the process and what
knowledge they have from past experiences (see also Ostrom 1998b). Participants
for the interactive valuation process should come from different spheres of society
in order to be able to meet the interests of different stakeholders. On the one hand
it is possible to involve stakeholders who are already elected by parties, clubs or
hold certain administration positions; on the other hand stakeholders can be people
who may not be linked to any constituency but are closely working with nature
248 Katja Arzt
and are affected by ecological changes. Representatives should act for the group
they present; however, this could make them quite inflexible in changing their
point of view or in their decision competence, because they may come with only
limited decision abilities. Nevertheless, representatives have the respect of the people
they stand for, and it might be more likely that their decisions are accepted by
others. Yet, a learning process with many elected representatives tends to consolidate
the hegemony of the expert in decision making (Meppem and Gill 1998; Röling
and Wagemakers 1998). On the contrary, stakeholders who are not representatives
of an organization but are interested in discussing different issues might have a
greater personal interest in the problem at hand, and might be better informed or
interested in others’ perceptions. However, it is essential for an interactive valuation
process that decisions are made by those who will bear the main consequences
(O’Hara 2000; O’Conner 1998; Meppem and Gill 1998; Jiggins and Röling 1998).
Most studies of deliberative institutions so far claim that only a limited number
of stakeholders can be involved in an interactive decision-making process; a
group of over twenty people is difficult to moderate for the purpose of a learning
process. In this respect the selection of the ‘right stakeholders’ can be difficult,
and so far no institutionalized methods (e.g. election process) exist to choose
them. Many organization teams working with deliberative approaches try to
inform as many people as possible about the process and then leave it to ‘the
community’ who will participate. However, some thought should be given to the
ability of the stakeholders to be involved in a process. Ostrom (1998a) stated
that getting the institutions right is a difficult, time-consuming, conflict-invoking
process. Time availability is certainly a limiting criteria; people who have to
work long hours are unlikely to join the process over a long period, or only if the
issue in question is in their own interests or of strong concern. Gender aspects
are to be considered too, as for instance mothers with young children have
disadvantages in joining processes which often take place in the evenings and
during weekends when public child care is unavailable. People from big
organizations or from governmental authorities, on the contrary, are likely to
join the process, as it may correspond with their duties; they do not tend to use
their spare time for the process. Still, their presence seems to be important because
it assures that the process of decision making is already linked to existing
institutions and there is a greater chance that the outcome of the deliberative
process will be accepted at higher levels if authorities are involved from the
beginning. The process can also be driven by the participation of ‘innovative’
local persons who are active and well respected by different stakeholders (Kolloge
2000). Generally, stakeholders only participate when they gain benefits from
doing so. In the case of an interactive valuation process the benefits are non-
monetary in the short term; on the contrary, the participants have to invest time.
However, most people will join the process when they are dissatisfied with the
present situation, hope to gain a better image or would like to improve their
knowledge. Once the process shows positive results more sceptical people may
join the process. The kinds of benefits people expect when joining the process
will feed considerably on the core relationship of trust, reputation and reciprocity.
The flow of information is a key aspect of a learning organization and should
be organized in a way that it internalizes different value aspects and makes
information transparent for all stakeholders. Professional moderation and
visualization techniques may be one helpful method, another is the use of language.
In this respect, O’Hara (2000) states:
Fair and competent discourse must mean giving particular attention to the
differences in verbal skill among discourse participants. The distribution of
verbal skill can become just as much a source of power and distortion as the
distribution of wealth in price based valuation. The broader rationality of
discursive reason may not at all be supported by featuring culturally rewarded
skills like articulated speech, deductive logic, or abstract thinking to the
exclusion of less articulate, empathetic, or traditional thinking.
The more concrete the objects of decisions and the more obvious the consequences
of action, the better is a mutual learning process initiated. Environmental problems
which are worked out in a learning process should be obvious for all participants.
Only then will these instruments be better than existing institutional arrangements.
The source of information is also important for the learning and trust-building
process. Some sociological studies in the UK (see Clark and Murdoch 1997) showed
that in the past farmers lost trust in scientific results and ideas. However, those
findings cannot be generalized; in other areas it may be observed that farmers
depend on and use scientific research. In any case, while implementing a valuation
process, researchers should be aware of the information used, as Bernabo (1998b)
observed: ‘When researchers apply assessment models as mechanistic predictive
tools, they risk producing results that are not sufficiently relevant and are prone to
misuse in decision making (Bernabo 1998a). Policy-relevant assessment is a more
organic process involving researchers and stakeholders in an interactive exercise
that builds understanding rather than trying to define “truth”. The emphasis must
be on providing practical information in an easily usable form.’
We may find that participants decide easily on one issue, but fail on another.
Property rights theory certainly gives useful interpretations as to why people value
items the way they do, as the owner of rights of a special natural good will have
the pleasure of the corresponding utility, or, on the contrary, if the corresponding
right is a duty, the owner has to carry the negative effects and costs (Hagedorn
2000; O’Neill and Spash 2000). For example, property rights on natural goods
such as soil or water do not have homogeneous property right structures, but are
differentiated by their different ecological characteristics; correspondingly the cost
and benefits will vary, leaving room for different institutional designs of duties and
rights (see Hagedorn 2000; Bromley 1997, 1998; Ostrom 1998a). The learning
process should clarify through discourse which property right structures are
connected to which participants and how much room is left to change the
institutions. Lawrence (2000: 16) suggested to distinguish between ownership,
control and use of property resources in environmental policy definition and
applications. Soil erosion, for example, is different for farmers who actually are
250 Katja Arzt
the owners of land as opposed to being tenants. Owners may have an interest in
keeping erosion as small as possible to maintain the fertility of the land for their
children or for future sales, whereas tenants may tend to reduce the fertility in the
short run for the time of the contract. However, if arable land should be converted
into meadow for ecological reasons, the owner of the land may not agree to convert
the land, because the rent received will be less for meadows. In this case the property
right structures are difficult to overcome and a decision will not be easily obtained.
In order to build reciprocity within the group it seems necessary that the property
right structures are clarified. This also corresponds with the control mechanisms.
Only if participants are convinced that breaking the rules about the decisions
made by the group (e.g. to reduce nitrogen fertilizer) can be controlled, is there
then a chance that the group formulates rules and norms which consolidate trust,
reciprocity and reputation.
The urgency of a problem can also be an important factor for participants
reaching consensus. For example, if farmers and environmentalists of a certain
region have problems with the water regime, the farmers would like to reduce the
water level in order to keep on farming, whereas the environmentalists would like
to have a higher water level to increase biodiversity. For both, the situation is urgent
and needs a solution in order to plan future activities. In this case the participants
may invest more time and money to reach a consensus compared with a situation
where nobody’s activities are considerably harmed.
Apart from the learning process the decision procedure itself strongly influences
the outcome. In the case of a roundtable group, decisions are made by consensus,
and not by voting or other mechanisms. Consensus means that all participants have
to agree upon the results by accepting the arguments addressed during discussion.
In this case verbal skills and power structures may have an important impact. Also
analytical methods such as multi-criteria analysis which recognize the plurality of
values that inform environmental choices can be combined with more deliberative
methods that enable citizens to articulate values and which are sensitive to concerns
about fairness in procedures and outcomes (O’Neill and Spash 2000).
In Table 11.1 the above-mentioned variables and their characteristics are
summarized. The next section will analyse a project that was carried out in Germany
to see how far it meets the idealized model presented here. It will do so by evaluating
the variables described above. The data for the criteria were collected from
qualitative interviews before and during the process and from meeting observations.
Conducting Brandenburg’s roundtable

The GRANO1 project established a roundtable group to carry out an interactive
valuation process in order to decentralize a decision about agri-environmental
schemes and measures. The project’s aim is to test institutional approaches for
interactive valuation in order to make statements about their ability to be transferred
to other regions. The idea for the project arose from different stakeholders in the
region of Brandenburg who were dissatisfied with the present situation of top-
down policy approaches, because the existing agri-environmental schemes were
Table 11.1 Variables feeding into trust, reciprocity and reputation

Continuous Stakeholder Information flow Issue Decision
discourse and information procedure
Frequency Representatives Adequate level of Property rights Considered
Time availability information considered aspects
generated
Duration Expected benefits Credibility of Control Participation
Gender aspects information mechanism Fairness
Information about Language Urgency of the
past Moderation and problem to the
presentation stakeholders
techniques
only partially in agreement with the practice of farming or were argued not to
include enough ecological benefits (Müller et al. 2000). A roundtable group operates
on the basis of seeking consensus between people with different viewpoints on the
issues discussed. In this approach each participant has an equal opportunity to
voice their option. Those involved should speak and listen to others with a view to
seeking areas of consensus. Agreements come out of the discussion process. This
approach works mainly over a series of meetings. In the case of the Agri-
Environmental Forum (AEF) the discussion for each meeting was prepared by a
project team and led by a professional moderator. For the project team it was
important to find an institutional arrangement that is quite simple to handle and is
time- and cost-efficient, because the idea of the interactive valuation institution
should be transferable to other regions and should be adopted by organizations
within the region.
On the basis of the common regulation of the European Union (EU), Germany
developed different agri-environmental schemes at the Länder (county) level.
Compared with other EU countries, where schemes are developed by central
governments (e.g. the Countryside Stewardship Scheme in the UK), Germany
already has a more decentralized approach. However, the weight given to different
environmental measures and the types of environmental consideration that are
important are clearly influenced by political priorities, policy objectives and cultural
attitudes to the farmed environment (see also Curry and Winter 2000). Local
farmers’ or local conservationists’ opinions about the different measures for the
schemes are not taken into consideration when establishing new programmes, yet
different experts are consulted by the local governments when setting up schemes.
In Brandenburg, two environmental schemes for farmers exist: one is called the
KULAP (cultural environmental landscape programme), the other is the Contract
Nature Conservation Programme. In KULAP farmers can choose among different
measures and apply for them. However, due to limited funds, the scheme operates
on a ‘first come, first served’ basis. KULAP is mainly financed by the EU and at a
national level (see EWG VO 2078/92 [Agri-environment Regulation (EEC) No.
2078/92] and VO 2052/88 [Council Regulation (EEC) No. 2052/88]). The
Contract Nature Conservation Programme is financed completely at the Länder
252 Katja Arzt
level. Farmers enter into individual contracts for their farms based on discussions
with environmental officers. Only farmers whose land lies within a protected area
can apply for the scheme (Hampicke et al. 1998). However, decisions about the
measures are not taken in consensus with different stakeholders, and the targets
for the scheme are fixed by experts.
For the implementation of the project, a region in the northeast of the
Uckermark (Brandenburg county, 180 km north of Berlin) was chosen. The
selection was made in arrangement with the local department for agriculture and
environment. The region contains 11,000 ha of agricultural land, is one of the
most thinly populated districts of Germany (53 inhabitants/km) and has a high
unemployment rate. An important factor for the selection of this particular region
was its strong agricultural character, the variety of landscapes and the low density
of protected areas.
The project team is interdisciplinary and consists of several natural scientists
and socio-economic scientists as well as a moderator. These professionals come
from different universities and institutions. The different backgrounds of the project
team made it difficult to find a general understanding of the definition of ‘value’
and ‘learning processes’ beforehand. The team could not create a common
understanding about the institutional arrangement as discussed in this chapter.
However, they already had agreed upon a phase- and process-oriented process.
The strategy was to choose relevant topics with the stakeholders and subsequently
each problem was to be discussed with regard to (a) reasons for the appearance of
the problem and the expected long-term consequences; (b) solutions that should
be achieved, including justifications; (c) measures to achieve these solutions,
including a monitoring-process; (d) financing of the measures; and (e) action
planning. However, this procedure could only be partially realized, because most
stakeholders considered it to be too abstract and too long. It became increasingly
evident by the third meeting that the stakeholders’ main interest was to talk chiefly
about measures and their financing. Therefore the project team had to plan each
meeting separately.
After the region was chosen the project team contacted different stakeholders
for initial interviews, in order to get to know the region and its problems, and also
to voice the intention of the project. The results of those interviews were presented
at a community meeting, to which over 100 different stakeholders had been invited
(selected using telephone books and contacts), but only 43 stakeholders attended.
During that meeting the project team asked for volunteers from different stake-
holder groups who would like to participate in the roundtable group. At the end
23 stakeholders wanted to attend the AEF. From October 1999 to April 2000 six
roundtable meetings took place (Figure 11.1), one each month, except in the summer
months when farmers are too busy on the farm to engage in other activities. The
process continued until December 2001, during which time 13 meetings took place.
The project team found another regional organization who organizes the meetings
until the present time but with a different focus. Todays meetings are more seen as
‘information meetings’, to inform farmers and environmentalists about agro-
political subjects.
35
30 Agricultural institutions
Farmers
25
Environmentalists
20
Conservation association
15 Tourism
10 Politics
5 Administration
Trade
0
Others
1
6
Scientists
g
g
t in
t in
t in
t in
t in
t in
ee
ee
ee
ee
ee
ee
M
Figure 11.1 Participants at the first six meetings of the AEF
The actual proceedings are outlined in Table 11.2. During the first meeting
much time was spent on gettting to know each other and clarifying expectations.
The different stakeholders revealed their interest in starting mutual discussions in
order to find solutions for regional ecological problems, but most of them were
unsure about the process itself and about the functioning of a roundtable discussion.
At the first meeting, the stakeholders identified the topics that should be discussed
over the following months. In order to have a better understanding of the current
environmental problems in the region, the scientists were asked to prepare a general
lecture about the environmental situation and a more detailed lecture on landscape
characteristics for the second meeting. However, it became clear in the following
meeting that the different participants including the project team had different
intentions in mind or understood the topics differently.
Accordingly, during the second meeting – which started with a scientific lecture
about the reasons for structuring the landscape (e.g. biodiversity, biotope network,
and other ecological functions) – several farmers complained that the financial
aspects were of major concern, but were not included in the lecture. They said for
instance: ‘Well as a human I find all those aims important, but as a farmer, I
cannot accept all those targets, because I must work on my fields in order to survive’,
and ‘I would do anything if I got enough cash’. The participants wanted to
concretize the topic (‘We should not talk so scientifically, otherwise this evening
does not mean anything to me.’) and decided to talk about Sölle (water bodies
developed during the ice age, which are a typical element of the regional landscape)
during the next meeting. As a result of this ‘unsuccessful’ meeting in terms of a
mutual learning process about the topic, the project team had to decide how the
topic of Sölle could be introduced and how stakeholders could come to a consensus
about a regional programme. Prior to the next meeting the organization team sent
information materials about Sölle to the stakeholders.
Table 11.2 Proceedings at the agri-environmental forum
What have been the What methods were used during the What came out of the discussion process?
expectations of the meetings?
organizers?
1. AEF To become acquainted with Questioning of expectations and Participants want a priority catalogue for the most
254 Katja Arzt
20.10.99 the participants backgrounds, visualization of answers important environmental problems

Find relevant subjects Point evaluation Structuring of the landscape as the main subject
2. AEF Presentation of expert Scientific lecture No discussion and questions about the lectures
3.11.99 knowledge Structured discussion No structured discussion about problems and targets
Problem analysis of Visualization Concretize the subject: the group wants to talk chiefly
landscape structures about ‘Sölle’
To find targets
3. AEF Ecological information about Questioning about attitudes about ‘Sölle’ Good information flow about ecological targets
2.12.99 ‘Sölle’ Expert presents knowledge within a Discussion was about proceedings at the AEF
Agreement on an agricultural dialogue Farming examples are wanted (experiences)
programme for ‘Sölle’ Open discussion
Visualization
4. AEF Agreement on an agricultural A farmer presents his farm Agreement on different measures for ‘Sölle’
12.1.00 programme for Sölle Presentation of different measures and Scientists were asked to present a regional programme for
their costs by the organization team ‘Sölle’
Discussion about transformation of those
results to other farms
Discussion about a regional programme
Visualization
5. AEF Presentation and reflection of Scientific lecture No comments about the lecture
24.2.00 expert knowledge about soil A farmer presents his experiences with Confusion about proceeding
erosion different measures
Discussion about experiences Open discussion about measures
Visualization
6. AEF Agreement on an agricultural Lecture about a method to find out the Method was not accepted
11.4.00 programme erosion potential Discussion about the usefulness of an erosion programme
Presentation of practical experiences No agreement about a regional soil erosion programme
Structured discussion
At the start of the third meeting all participants were asked to write down what
they knew about Sölle and why they thought it was important to protect them.
Then, one expert on Sölle answered questions as required and explained their
ecological functions. Contrary to the expectations of the organization team, the
stakeholders did not have detailed knowledge about the Sölle in their region. Instead
the stakeholders expected the natural scientist to know which treatment is the best.
The following discussion about a regional scheme for Sölle was unsatisfactory
because different scientists dominated the discussion while debating a general
proceeding (what it is important to know in order to be able to decide on a Sölle
programme). This confused those participants who were only interested in the
measures and actions for Sölle.
The fourth meeting was very structured, dealing with different measures for
Sölle as prepared by the scientists. Each measure was subsequently discussed by
the experts and stakeholders. The successful connection between practice and theory
corresponded especially to practitioners. At the end they had a list of different
measures for Sölle and how much payment should be given to a farmer when
applying each of them. The scientist for agricultural policy put those measures
into a programme for Sölle which they presented at the ministry of agriculture.
The group decided to close the discussion about Sölle.
The next meetings dealt with soil erosion, because the natural scientists in the
team found that to be the most relevant ecological problem for the region. The
participants agreed to discuss this problem, although they had never mentioned its
urgency or importance before. During the fifth meeting a new scientific ‘expert’
was introduced who explained why soil erosion is a common problem and how
important it could be to reach consensus about that topic among the stakeholders.
However, instead of creating an atmosphere of dialogue, the expert just presented
results and measures and left little room for the stakeholders to talk about their
problems. Instead of a mutual learning process by scientists and stakeholders it
was a lecture by one expert to the ‘others’. There was also a farmer who presented
his experiences with some soil erosion measures, but his findings were commented
on only by the expert.
The sixth meeting was attended by eleven stakeholders, with only three farmers
amongst them. It dealt with measures and if and how payments for soil erosion
should be directed. In this case, the scientists tried to direct the discussion towards
the regional risk of soil erosion in different parts of the landscape. They showed
with different techniques that soil erosion is not a problem on all farms and suggested
therefore that payments should be directed to high-risk farms only. The discussion
which followed the presentation of different techniques was highly controversial
and dominated by the people from the agricultural administration, who discussed
the pros and cons of such directed payments. The farmers themselves were quiet
during the discussion and avoided direct statements. However, they complained
that not enough farmers had attended this meeting, and they feared that their
decisions would not be accepted. Yet, this discussion left room to express different
perceptions of the problem. Although a structured discussion was planned in order
to reach consensus about a regional programme for soil conservation, the stake-
256 Katja Arzt
holders talked about the general usefulness of soil erosion programmes and no
consensus was reached.
Analysis of the agri-environmental forum

In this section I discuss how the AEF performed with respect to stakeholders
and representation, issues raised, information flow, and decision procedure. The
roundtable group consisted of different stakeholder groups: representatives from
clubs (e.g. the farmers’ union, nature conservation club), officers from different
institutions (agricultural and environmental ministries), private farmers and
people interested in the subject (e.g. tourist industry). There was a great fluctuation
in numbers and in backgrounds of participants during the process. A minority
of the representatives did not join the discussion in a creative way but just observed
the meeting(s) or made some sarcastic comments; it seemed that they were sent
by their constituency to observe the process but had no abilities or consent to
make decisions. Other participants, however, enjoyed the interesting mix of
different stakeholders: ‘It is good that many different people come together’, ‘it
is important that agriculture and conservationists come together, there is not
such a mixture in other institutions’. From the local agricultural ministry two
different representatives attended alternately but had different opinions about
the topics discussed.
The stakeholder group of farmers complained that not enough farmers joined
the process and that their number was decreasing during the process. Also, not all
types of farmers joined the process. At the AEF only three farmers from the former
Agricultural Production Co-operative (LPG) took part, as well as newcomers to
the region, who have smaller farm units. Typical for the region are farmers who
worked formerly in the LPG and started their own business after the transformation
process. Their farms are usually smaller in size and run by family labour, whereas
the former LPG are usually very big (up to 2000 ha). However, it was very difficult
for the organization team to invite farmers from those ‘smaller’ farms to the AEF.
They complained about time limits. One farmer said that he was sceptical about
sitting together with farmers from the big co-operatives, because he knew that
those big co-operatives have more power and were always dominant in consensus-
building processes in the past; therefore he was unwilling to come to the AEF.
Another imbalance to representation was the small number of women contributing
to the discussion process (3 women out of 23 participants). Women suffered most
from the agricultural institutional change, leaving most of them unemployed.
However, women are seldom in leading positions (in administration, on farms, in
trade, etc.) and were therefore not able to speak for a constituency. Even those
women who participated in the process hardly joined in the discussion. The
organization team (scientists) were the dominating stakeholder group not only in
number, but also in activity. As they prepared the meetings they had a broad
influence on the process. However, it became clear during the process that the
scientists themselves did not share a common vision about the process, because
they all had different interests in the outcome.
The topics discussed were not urgent problems for most stakeholders even though
they chose the topics themselves, as some participants admitted during the
interviews. They did not identify with the problems of Sölle. This point was difficult
to handle for the organization team, because the stakeholders did not contribute
their own ideas, interests and knowledge, but expected the scientists to do so. The
topic of soil erosion was introduced by natural scientists but was not seen as an
urgent problem by the participants. Some said that it is an issue for an information
service and need not be discussed in a big group. While soil erosion control
mechanisms were mentioned during the discussion, no consensus could be reached.
Topics which seemed to be of great interest for most stakeholders, such as hedge
planting, were not chosen in the first round of meetings because the participants
mentioned the difficult property right structures. If someone wants to plant a hedge
the owner and the farmer both have to be consulted; both of them might reject the
planting proposal because the farmer will lose land for crop growing and the owner’s
rent income will go down. Although hedges can help reduce soil erosion and improve
biodiversity – important for both owner and farmer – a solution is difficult to find,
unless property right structures are changed.
The stakeholders criticized the lack of a clear structure; they were confused
when the scientists started to discuss the general proceeding and project aims. The
participants liked the fourth meeting the best, because the organization team
presented a clear structure. They called it ‘effective’ working; the results were clear.
The mutual learning process only began to develop during the third meeting. The
stakeholders could first express their own knowledge and then the experts
formulated their knowledge on demand; it was not a lecture about the ‘right’
knowledge but a communication about the problem at hand. During the second
and fifth meetings, however, scientific lectures were given, leaving little room for
the stakeholders to formulate their problems, but giving insight into the concept
of one scientist (or scientific community). However, some participants mentioned
that for them the scientific input was very important: ‘The scientist should tell us
what is good for the environment’, ‘scientists should show us how to solve problems’.
In particular, natural scientists seem to have a very high position to formulate
environmental targets in the perception of most participants. Nevertheless, when
talking about economic data, the knowledge of economic scientists was criticized.
Some of the farmers complained that some of the economic data presented by the
scientists for the different measures were unclear: ‘You never know what kind of
data they use, … each farm is so different.’ During the meeting it could be observed
that frequently scientific experts (‘professors’) dominated the discussion and if they
could not attend a meeting, then people from the administration dominated the
process. The participants did not complain about the language used during the
process. However, people who were ‘used to talking’ at meetings/in large groups
were found to be dominant.
A decision procedure at the AEF was performed during the fourth meeting,
when participants decided on measures for Sölle. Decisions were taken on a
consensus basis, yet during the discussion it did not become sufficiently clear as to
which criteria the stakeholders divided the measures into those for direct payments
258 Katja Arzt
for farmers, and others classified as ‘codes of good agricultural practice’ without
direct payments. Usually they considered costs during the discussion. One
administrator from the environmental board, for example, said: ‘the long
transportation of the cut grass is very costly and a farmer could hardly pay for
that’ and ‘I do not think that a farmer is interested in cutting the grass in that
place, therefore I agree to a payment’. The decisions exclude consideration of
long-term consequences for the region’s ecology and different stakeholder groups.
Conclusions
Certainly, six meetings are insufficient to acknowledge if the AEF will be
institutionalized as an interactive valuation process in which learning processes
are initiated. Within four meetings the participants agreed upon a list of different
measures for Sölle. The list was transformed into an agri-environmental programme
for Sölle and presented by the organization team to the local government of agri-
culture, but it has not yet been accepted as part of a new agri-environmental
scheme in the future. For the second subject, soil erosion, no consensus could be
reached after three meetings within the group. Overall, it was the first time that
local stakeholders together with scientists decided and reached consensus about
agricultural measures, and on this basis the process can be called a success. They
exchanged their value perceptions on different issues. However, the decline in
numbers of participants over the series of meetings signalled that not all participants
were satisfied with the process itself.
This study hypothesized that a good learning process is a result of well-
implemented variables (e.g. selection of stakeholders, issues discussed, flow of
information, etc.). In the case of Sölle it can be concluded that the participants
learned about them during the meetings and were helped to create a value
perception about the issue. It also became clear that they needed different
information sources. The third meeting, for instance, was successful in stimulating
learning about the ecology of the subject under consideration, but participants
failed to decide about a local programme because more information was needed.
After the presentation of practical experiences from farmers a decision procedure
was possible. In the case of soil erosion, however, a decision process could not be
stimulated. The issue may not be of any relevance to the participants or the infor-
mation given to them did not correspond to their needs. However, it was shown
that the presentation of scientific lectures alone neither stimulated a dialogue among
the participants nor helped to create an atmosphere of trust.
What could be improved in future settings in order to stimulate a mutual learning
process? With regard to stakeholder selection it should be ensured that the
representatives of different organizations back-check with and can speak on behalf
of the constituency group, because it happened that two representatives from the
same organization had opposite opinions. This fact did not help to build reciprocity
among the participants. The group tended to exclude some stakeholder groups,
like farmers from family businesses, women, or people who do not belong to any
organization. More thoughts should be given how those groups could be integrated
or how the selection of stakeholders should be organized in order to have an
adequate representation of the county community. The organization team should
seek to be more independent of the other stakeholder groups. Scientists therefore
should be restricted to an advisory function, and should only make a contribution
when requested.
The chosen topics did not correspond to the interests, knowledge or urgent
concerns of the participants. Therefore a better method for choosing topics seems
to be necessary. However, the method still has to be quite simple so that it can be
transferred to other regions. This requirement precluded a careful analysis of needs
of the participants through social scientific methods (e.g. interviews) prior to the
process. Conversely, to start with conflicting problems may impede a constructive
discussion and may not help to build trust among the participants in the beginning.
Therefore it might be best to start with an issue which is relatively uncontroversial.
Property right structures (ownership, control and use) have not been exclusively
mentioned during the discussion but this should be changed so that action can
easily follow the decision process and reciprocity among the participants can emerge.
The information flow during the process was strongly influenced by scientists.
There is unmistakably a big gap between a scientist’s way of thinking and that of
practitioners. Scientific training emphasizes the hierarchical organization of an
issue: analysis of the situation, targets, instruments, implementation and testing. A
practitioner, on the other hand, tends to be more interested in concrete decisions
about actions. The organization of information flow is central for the learning
process. It helps people if a discussion can be followed easily in order to contribute
fruitfully to the debate. Scientific lectures should only be held on urgent request.
Further it was observed that participants with good verbal skills dominated the
discussion process; a more balanced level of participation by all stakeholders could
be achieved by using better moderation techniques or a different procedure.2
The decision procedure about the Sölle programme was inefficient. It lacked the
participation of all stakeholder groups and did not consider enough different aspects.
A well-performed MCA might improve the procedure and also stimulate a learning
process. However, institutional difficulties within the decision procedure may be
overcome by better institutional design or by learning about the process itself. Certainly
more information and perceptions are exchanged in such deliberative institutions
compared with monetary evaluations, but at the same time it is costly and time-
consuming. Yet, getting the institution right is a time-consuming process. The
stakeholders are dissatisfied with the present situation of agri-environmental policy
decisions but at the same time sceptical about the success of such new valuation
institution(s). So far none of the stakeholder groups or the organization team know
the direction that a change of the existing institutional setting in agricultural policy
may be taking. The decision procedure may improve over time and with more
experience, but the acceptance of those procedures from existing institutions can
still threaten the whole process. Therefore it is important to fit such new valuation
institutions in with existing institutions right from the beginning.
The framework used in this study certainly helped to analyse the process with
its different aspects. However, at this early stage in the deliberative process of the
260 Katja Arzt
case study it is difficult to draw any conclusions on the relevance of each element
to the process and whether there are other important elements which might lead
to a successful performance of interactive valuation processes. Only the comparison
of different case studies will clarify this question.3 Certainly more information is
needed about the reasons why people are involved in the process and their
expectations in order to understand their needs. This information might shed more
light on the adequate design of the interactive valuation process. Another question
is whether all environmental problems of agriculture can be solved with these
kinds of participatory methods. More research has to be done on which co-
ordination structures correspond to which kinds of transaction and property right
structures. In this respect a more hierarchical decision procedure might address
some problems better when transaction costs, and also fairness, are considered.
Finally, more knowledge is needed about the underlying assumptions of the core
elements of trust, reciprocity and reputation and how these stimulate a learning
process among different stakeholder groups. Further investigations are needed as
to how institutions enhance or restrict the building of mutual trust, reciprocity,
and reputation. In this respect future investigations should also be interdisciplinary
in their design.
Acknowledgements
This chapter arose from work commissioned by the German Federal Ministry of
Research and Education (Bundesministerium für Bildung und Forschung); it is
part of a scientific co-operative project called GRANO. The author is grateful to
the steering group for helpful discussions, and also to the Chair of Resource
Economics and Prof Dr K. Hagedorn, the scientists at the Countryside and
Community Research Unit in Cheltenham, UK, and especially to John Powell.
The author would like to thank Simon Andrews, Claudia Carter and the colleagues
who joined the discussions during the ESEE Conference 2000 in Vienna. The
author alone is responsible for any errors and omissions.
Notes
1 GRANO project stands for ‘Approaches for a sustainable agricultural production: Application
for North-eastern Germany’. GRANO is a scientific co-operative project supported by the Federal
Research Ministry. GRANO chose the districts of Uckermark, Barnim, and Elbe-Elster to search
for, develop, and test new ideas and solution strategies. The ‘round table’ project is one among
other projects carried out by GRANO.
2 The organization team discussed that problem, together with some participants, and decided
that bilateral interviews about the programme measures (in this case measures against soil erosion)
should precede the next meeting. Consequently the organization team prepared a questionnaire
and asked all participants separately during the summer months when no meeting took place,
and presented the results during the next meeting. This procedure was very time and resource
consuming.
3 At the moment a PhD student is working on that question.
References
Berkes, F. and Folke, C. (1998) Linking Social and Ecological Systems. Management Practices and
Social Mechanisms for Building Resilience, Cambridge: Cambridge University Press.
Bernabo, J.C. (1998a) ‘Improving integrated assessments for application to decision making’,
in S. Lee and T. Schneider (eds) Risk Assessment: Linking Science to Policy, Amsterdam:
Elsevier Science.
Bernabo, J.C. (1998b) ‘Perspectives and principles from other assessments’. Online.
Available at: http://www.agci.org/publications/EOC97/eoc97session2/Bernabo.html
(accessed 16 June 2004).
Bromley, D.W. (1991) Environment and Economy: Property Rights and Public Policy, Cambridge,
MA: Basil Blackwell.
Bromley, D.W. (1997) ‘Property rights in environmental economics’, in H. Folmer and T.
Tietenberg (eds) The International Yearbook of Environmental and Resource Economics 1997/
1998: A Survey of Current Issues, Cheltenham: Edward Elgar.
Clark, J. and Murdoch, J. (1997) ‘Local knowledge and the precarious extension of scientific
networks: a reflection on three case studies’, Sociologia Ruralis, 37(1): 37–60.
Curry, N. and Winter, M. (2000) ‘The transition to environmental agriculture in Europe:
learning processes and knowledge networks’, European Planning Studies, 8(1): 107–21.
DeBono, E. (1993) Parallel Thinking, Ringwood: Penguin Books.
Habermas, J. (1982) Theorie des kommunikativen Handelns, Frankfurt/Main: Suhrkamp.
Hagedorn, K. (2000) ‘Umweltgenossenschaften aus institutionenökonomischer Sicht’, in
M. Kirk, W. Jost and S. Kramer (eds) Genossenschaften und Kooperation in einer sich wandelnden
Welt. Festschrift zum 65. Geburtstag von Prof. Dr. Hans-H. Münkner, Münster: LIT Verlag:
267–91.
Hampicke, U., Müller, K., Meyer-Aurich, A. and Kachele, K.U. (1998) ‘Agrarumweltpolitik
nach dem Subsidiaritätsprinzip. Teil III Fallstudie Brandenburg/Uckermark’, working
paper of the Universität Greifswald und Zentrum für Agrarlandschafts- und
Landnutzungsforschung e.V.
Hanna, S., Folke, C. and Mäler, K.-G. (eds) (1996) Rights to Nature, Washington, DC: Island
Press.
Holling, C.S., Berkes, F. and Folke C. (1998) ‘Science, sustainability and resource
management’, in N.G. Röling and E. Wagemakers (eds) Facilitating Sustainable Agriculture,
Jiggins, J. and Röling, N. (1998) ‘Interactive valuation’, paper presented at the Conference
of the European Society for Ecological Economics, University of Geneva, Switzerland,
5 and 6 March 1998.
Kolloge, S. (2000) ‘Die Wirkung der Agenda 21. Eine institutionentheoretische Wirkungs-
analyse am Beispiel von sechs Kommunen im ländlichen Raum Großbritaniens und
Deutschlands’, unpublished PhD thesis, Landwirtschaftliche-Gärtnerische Fakultaät
der Humboldt-Universität zu Berlin, Germany.
Lawrence, R. (2000) ‘Property, rights and fairness’, in C.L. Spash and C. Carter (eds) EVE
Policy Research Brief Series, no. 6, Cambridge: Cambridge Research for the Environment.
Meppem, T. and Gill, R. (1998) ‘Planning for sustainability as a learning concept, survey’,
Müller, K., Bork, H.-R., Dosch, A., Hagedorn, K., Kern, J., Peters, J., Petersen, H.-G.,
Nagel, U.J., Schatz, T., Schmidt, R., Toussaint, V., Weith, T., Werner, A. and Wotke, A.
(eds) (2000) Nachhaltige Landnutzung im Konsens: Ansätze für eine dauerhaft-umweltgerechte Nutzung
der Agrarlandschaften in Nordostdeutschland, Giessen: Focus Verlag.
262 Katja Arzt
Munda, G. (2000) ‘Conceptualising and responding to complexity’, in C.L. Spash and C.
Carter (eds) EVE Policy Research Brief Series, no. 2, Cambridge: Cambridge Research for
the Environment.
O’Conner, M. (1998) ‘Walking in the Garden(s) of Babylon: an overview of the VALES
project’, Paris: C3ED Rapport de Recherche.
O´Hara, S. (2000) ‘The challenges of valuation: ecological economics between matter
and meaning’, in C. Cleveland, R. Costanza and D. Stern (eds) The Nature of Economics
and the Economics of Nature, London: Edward Elgar.
O’Hara, S., Shandas, V. and Vazquez, J. (1999) ‘Communicating sustainable development
options – who evaluates the trade-offs?’, in I. Ring, B. Klauer, F. Waetzold and B.
Mansson (eds) Regional Sustainability. Applied Ecological Economics Bridging the Gap between
Natural and Social Sciences, Heidelberg: Physica Verlag.
O’Neill, J. and Spash, C.L. (2000) ‘Conceptions of value in environmental decision-making’,
in C.L. Spash and C. Carter (eds) EVE Policy Research Brief Series, no. 4, Cambridge:
Cambridge Research for the Environment.
Ostrom, E. (1990) Governing the Commons: The Evolution of Institutions for Collective Action,
Ostrom, E. (1998a) ‘The institutional analysis and development approach’, in E.T. Loehman
and D.M. Kilgour (eds) Designing Institutions for Environmental and Resource Management,
Cheltenham: Edward Elgar: 68–90.
Ostrom, E. (1998b) ‘A behavioral approach to the rational choice theory of collective
action’, presidential address, American Political Science Association, American Political
Science Review, 92(1): 1–22.
Ostrom, E., Gardener, R. and Walker, J. (1994) Rules, Games, and Common-Pool Resources,
Ann Arbor: University of Michigan Press.
Röling, N. and Wagemakers, E. (eds) (1998) Facilitating Sustainable Agriculture, Cambridge:
look beyond contingent pricing’, Ecological Economics, 24(2–3): 213–30.
Sahlins, M.D. (1976) Culture and Practical Reason, Chicago, IL: University of Chicago Press.
Simon, H.A. (1983) Reason in Human Affairs, Stanford, CT: Stanford University Press.
Söderbaum, P. (1999) ‘Values, ideology and politics in ecological economics’, Ecological
Preference transformation through deliberation 263
12 Preference transformation
through deliberation
Protecting world heritage
Simon Niemeyer
Introduction
Preferences are prone to change. Yet much effort in environmental economics has
been applied to the measurement of static preferences for use in policy. This
‘endogenous preference’ orthodoxy has recently been challenged on at least two
fronts. One criticism stems from findings that support the idea of ‘constructed’
preferences in view of exogenous influences (for example Slovic 1995; van den
Bergh et al. 2000). A second criticism, with implications for environmental policy,
is the emerging tension between the normative goal of sustainability and
aggregation of sovereign preferences (for example Common and Perrings 1992).
These have led to arguments within ecological economics for policies that will
lead to so-called ‘sustainable’ preferences (Norton et al. 1998), although it is uncertain
what is needed to achieve such a normative goal.
The fact that preferences can change invokes notions of preference quality –
that some preferences might be ‘better’ than others. Most economists do not make
such distinctions: a stated preference is a legitimate preference where consumers
are sovereign and free to choose as they wish. Others, such as O’Neill (1997) and
Elster (1983) are less sanguine. For them, the process of producing ‘better’ (or
reasonable, in the sense that they have been through a process such that chosen
means match justified ends) preferences is more important than simply realising
given preferences – the reason for a preference is as important as the preference
itself.
Deliberation is one (if not the) process for reasonable preference formation. It
is the process of deliberation, be it interactive deliberation with others, or
deliberation within oneself (e.g. Goodin 2000) whereby one recursively constructs
those means that suit desirable ends. In the case of environmental policy, the most
vaunted end is sustainability although it is unclear and possibly indefinable. Many
authors are turning to deliberation in the search for mechanisms capable of
elevating environmental considerations in policy making (for example Sagoff 1998).
An important element of this ‘deliberative turn’ in environmental policy is the
tendency for deliberation to transform preferences in the direction of ‘nature-
loving’, or sustainability. The relationship between preference for nature and
sustainability is partially accounted for by the simple premise of ‘more nature
264 Simon Niemeyer
preserved is more nature sustained’; but sustainability is inherently more complex
and infused with a specific meaning depending on context.
One danger in attempting to produce ‘sustainable’ preferences is to pre-empt
outcomes that are nominally designated sustainable when that goal itself is not
well defined. Just as Elster (1983: 9) states that utility maximisation as a single
endeavour is futile because happiness eludes those who actively strive for it, the
same may also be true when actively seeking to achieve sustainability. Instead of
defining a sustainable end, a better approach is to focus on those processes that
tend to produce ‘sustainable’ outcomes; better still if those same processes help
clarify the debate concerning definitions thereof. Fortuitously, deliberation is also
identified as such a process through which sustainable outcomes are defined and
achieved as part of the same process (Rydin 1999). In other words, ‘deliberators’
formulate those ends deemed sustainable and form judgements about which policy
means are best suited.
This chapter seeks to explore the ‘deliberative alternative’ and its ability to
transform preferences to suit so-called sustainable ends. The case study concerns
a controversial road constructed through World Heritage Listed rainforest in Far
North Queensland, Australia, on which a Citizens’ Jury (CJ) was held in January
2000. CJs are seen here as an example of a formal deliberative process that has
some similarity to a legal jury where decisions are made collectively by a randomly
selected group of citizens based on evidence presented (see other chapters in this
volume and also Smith and Wales 2000). The focus of this chapter is on the margins
of ‘preference shift’, or those mechanisms through which preferences are
transformed by the deliberative process. The aim is to measure both changes to
(rank ordered) preferences and the underlying subjectivity for those preferences.
Subjectivity is measured using Q method, a form of ‘quantitative discourse analysis’,
which is also compared with actual jury discourse and interview data.
The following discussion begins by considering tentative mechanisms whereby
preferences are transformed through group deliberation in an information-rich
environment. This is followed by a description of the case study, its history and
politics, considered important because, as will be shown, issue context is a major
factor in preference dynamics. The impact of the deliberative processes on both
preference and subjectivity is then discussed with particular mechanisms
explaining the processes that contribute to preference shifts. Conclusions are
then drawn about the role of deliberative processes in improving environmental
policy outcomes.
Discourse and preference shift

The following discussion attempts to identify mechanisms whereby preferences
change through discourse, particularly during participation in a formal deliberative
process. The use of the term preference varies according to disciplinary perspective.
Here preference refers simply to the process of ranking choices, such as in the act
of voting. Preferences then reflect the tendency to favour one option over another.
Aldred (this volume, Chapter 8) argues that the use of the term preference is often
confused with judgement. According to Elster (1983: 16) judgement is defined as
‘the capacity to synthesize vast and diffuse information that more or less clearly
bears on the problem at hand, in such a way that no element or set of elements is
given undue importance’. A relationship between judgements and preferences exists
to the extent that judgements represent the reasoning behind a preference (Keat
1994: 335). It is posited here that preferences are shaped during deliberation by
three inter-related phenomena: (i) information (knowledge acquisition and
understanding of consequences); (ii) attitude (perspective adopted during
judgement, attitude transformation); and (iii) process (information absorbed and
attitudinal lens adopted by virtue of process, degree of preference construction).
Each of these impacts is briefly discussed in turn below.
Information is a widely accepted influence on preference. Elementary economic
theory is premised on ‘full information’; that is, the consumer is assumed to be aware
of all costs and benefits, including opportunity cost, when making choices about
consumption. In reality no individual has recourse to the entire suite of information
when making choices, uncertainty is implicit to the act of choice (values also play an
important role in information acquisition and impact on preference). Where provision
of a good requires social cooperation the problem of information deficit is particularly
strong.1 The problem is exacerbated for environmental goods, which are characterised
by complexity and uncertainty, making the ‘cost’ of acquiring information particularly
high. That political choices are poorly informed increases scope for preference
manipulation. To reduce the burden of information acquisition, individuals may
adopt simplifying strategies in order to make sense of the issue. Intuitively appealing
conclusions may be drawn from the little information available, which is in no small
part a product of the political environment. If the issue is polarised, with interest
groups adopting extreme positions, individuals may choose to follow these external
cues, resulting in issue polarisation being reflected in individual preference. The
mechanism relies on the tendency for interest groups to adopt extremes to elevate
their cause within the political sphere; and for individuals to replicate these extremes
in their own discourse and preferences rather than the messy middle ground (Elster
1983: 19). Environmental issues in particular are often dominated by polarised
interests which play an important role in defining the lines of discourse surrounding
issues within the public sphere (for example Harrison et al. 1999). This results in
preferences that reflect symbolism rather than substance, an outcome referred to as
symbolic preference (Sears 1993). Where symbolic preferences dominate an issue,
an information-rich deliberative process may improve the ability of the individual to
judge the merits of policy alternatives, or at least decouple them from the posturing
of external parties. The result is that improved information about the public good in
question then changes perception regarding the outcome of a particular policy. The
individual is armed with sufficient knowledge, giving them the fortitude to make an
independent assessment of the issue at hand, rather than defaulting to intuitively
appealing arguments proffered by entrenched political interests. Even where issues
are not dominated by symbolic politics, information will tend to transform individual
preferences by improving the understanding of consequences attached to certain
choices.
266 Simon Niemeyer
Preferences are more than a function of information. At least as important are
attitudes, or values. These help to frame both the desired end and the way that
information is received and processed. Attitude refers to the belief system, or
subjectivity, that comes into play during the process of judging alternative environ-
mental policies as far as they relate to subjective values, or value orientations. In
other words, attitude is the value attached to a particular outcome, rather than a
judgement regarding its likelihood. There are a number of attitudinal perspectives
that may coexist in relation to an issue, even within the same individual (Etzioni
1986). In relation to environmental attitudes, the most celebrated example of this
is that of Sagoff ’s (1988) citizen–consumer schism, which describes in very simple
terms how the same problem can be conceptualised from the position of consumer
or citizen, each yielding entirely different outcomes. Keat (1994) argues that, because
environmental goods are public goods requiring collective action, (sustainable)
environmental decisions require that individuals address them from the perspective
of the citizen. This implies other-regarding behaviour and explicit consideration
of ethical dimensions in contrast to the self-interested consumer (see Spash and
Hanley 1995). Some authors contend that deliberation tends to foster citizenry in
making choices (Ward 1999). Goodin (1996) posits that deliberation can extend
other-regardedness to that of Nature itself, an otherwise silent partner in democratic
processes. If individuals change their perspective during deliberation, adopting
different sets of attitudes through which they judge the process, it is anticipated
that preferences will change accordingly to reflect adjustments in the desired end.
Should group deliberation produce changes to attitudes such that they tend to
favour collectively desirable sustainable outcomes (over the satisfaction of individual
wants) then these should be reflected in the preferences of individuals.
Construction process concerns both the cognitive act of turning information
and attitudes/values into preferences and the context in which these are formed.
There has been much discussion in recent literature, particularly concerning
contingent valuation, regarding the construction of preference (Slovic 1995;
Schkade and Payne 1994). It is posited that an extensive process of information
acquisition, evaluation and judgement – such as for a formal deliberative process
– leads to better-constructed preferences than those based on superficial (or
symbolic) assessment of consequences.2 Information availability forms an important
part of this context because it facilitates the judgement process by filling gaps in
knowledge and providing linkages between cause and effect, improving translation
of attitudes regarding consequences into preferences. During the construction
process, information and attitude conjointly influence preferences as well as each
other: information transforms attitudes and attitudes impact on the acquisition of
information. With respect to ecology, Callicot (1982) argues that increased
knowledge ‘changes our values by changing our concepts of the world and of
ourselves in relation to the world’ (which resonates with the ‘biophilia hypothesis’,
see Kellert and Wilson 1993). Changes to attitudes, or subjectivity, will also bear
upon the acquisition of information for the judgement process (Elster 1983: 19).
If this contention is correct, a positive feedback occurs where increased valuation
of nature stimulates the desire for more knowledge about nature, and this in turn
increases environmental values. Such processes occur over a longer time-frame
than a deliberative forum, but may nonetheless be observable during deliberative
processes.
Case study: the Bloomfield Track

The case study for deliberations concerned policy options for the Bloomfield Track
in the Daintree Region in Far North Queensland, Australia. Figure 12.1 shows the
location of the Track, which traverses 30 km of mostly lowland coastal rainforest
Figure 12.1 The Bloomfield Track

268 Simon Niemeyer
from Cape Tribulation north to the Bloomfield River near the Wujal Wujal
township. It lies almost entirely within the Wet Tropics World Heritage Area, the
boundaries of which are shown as the continuous white line circumscribing the
region. The map also shows the Great Barrier Reef passing along the coast. Smaller
inshore reefs abut the coast along the region, creating a unique situation where
tropical rainforest adjoins coral reef.
The Bloomfield Track provides a link, albeit a rudimentary one, between the
remote Bloomfield region which is home to approximately 1000 residents, mostly
located in the Wujal Wujal Aboriginal Community, and major southern centres
such as Mossman and Cairns. Prior to its construction, local residents were restricted
to northern access via a rough track to the more remote township of Cooktown
approximately 60 km away – and from there a three-hour-plus inland drive south
to reach Cairns. The satellite image in the background for Figure 12.1 provides
some indication of the ruggedness of the area, most of which is covered by dense
tropical rainforests and steep terrain. The intended community-access Bloomfield
Track is currently used mostly for tourism. From the beginning of construction in
1982 the Bloomfield Track has been highly controversial. Despite stated objectives
to increase community access, many attribute its construction to the political
dynamics of the time. The period was one in Australian politics where
environmental issues were beginning to peak, in electoral terms, at the federal
level (Papadakis 1993: 142). In Queensland, however, state politics remained
dominated by development interests and strong state sentiment against federal
intervention. Tension between the federal government and the development-
oriented Queensland state government was reaching a particularly strong climax
(see Kellow and Niemeyer 1999: 210). Significantly, there were no formal processes
established for public input into the decision.3 Construction was undertaken by
the Douglas Shire Council with Queensland state government backing, amidst
pitched battles between environmentalists and the state police. The Track was
constructed in haste; a single bulldozer cleared a trail, sometimes swerving around
protesters buried along its path to prevent construction. Even during construction,
the green movement intensely lobbied the federal government to nominate the
area for World Heritage and intervene in the road issue, as it had recently done in
the state of Tasmania, without success. The thrust of public debate against the
track focused on highly emotive and symbolic issues such as the intrusion into a
significant wilderness (for example, Wilderness Action Group 1983). Pictures of
pristine rainforest ‘ruined’ by bulldozers dominated the campaign.
Environmentalists cited damage caused to the onshore reefs by sediment run-off
from the Track.
Seven years after the construction of the Bloomfield Track a state election in
Queensland replaced the incumbent government. This resulted in politically aligned
governments at both state and federal levels and provided the opportunity to declare
the Daintree region a World Heritage Area managed by a joint federal–state body,
the Wet Tropics Management Authority. The Bloomfield Track has remained a
thorny issue – a product of a different political era – one that has yet to be properly
resolved.4 There are strong, and opposing, interests, and interconnected pressures
within the region make the formulation of policy in the region extremely difficult.
Although at first indifferent, there is strong support for the Track from within the
vocal local community groups, who express resentment at what they view as ‘outside’
interference in their own affairs by those who call for its closure. The Track has
also become an important, if symbolic, link between the indigenous Wujal Wujal
community and their southern kinship groups, historically separated by forced
resettlement. There is consequently intense pressure from local residential groups
to keep the Track open and upgraded, though some residents also resent the
increasing intrusion by tourists. The combined forces of geographic isolation and
a sense of disenfranchisement have enhanced suspicion of outside interference in
local affairs. Environmental groups have consistently maintained their calls that
the Track be closed, and down-graded to a walking trail. Polarisation of interest
groups is, in part, an artefact of the political history of the area and the absence of
mechanisms for consultation and mediation. The continued policy stagnation of
the Bloomfield Track issue is a constant reminder of its bitter political history.
The Far North Queensland CJ was conducted as part of research into the use
of CJs for environmental policy in Australia. The jury, which was recruited on a
random-stratified basis from the region, was given the task of considering
management options for the Track over a four-day period. In broad terms the jury
process consisted of four stages: site familiarisation; presentations by expert witnesses
and questions; presentations by community witnesses and questions; and
deliberation and report writing. Site familiarisation consisted of a field trip to the
Track.5 This was followed by two days of presentations during which short talks
were given by witnesses. Expert witnesses gave evidence pertaining to technical
aspects under the jury remit. Themes which these witnesses were recruited to
address included tourism; environmental impact to both reef and rainforest;
engineering; and economics and regional planning. Community witnesses were
recruited to give voice to the needs and aspirations of the community and to provide
the deliberations with a social context. Community representatives included the
mayors of the Douglas and Cook Shire Councils and a local Bloomfield councillor
to provide a community perspective. A local indigenous perspective was provided
for as far as was possible.6 Report writing took place on the final day of the process,
during which jurors developed policy recommendations.
Information about the manner in which participants in the CJ changed attitudes
towards the Bloomfield Track issue and their policy preferences was obtained by
surveying participants at the outset of deliberations, at the midway point (in the
case of attitudes),7 and immediately after the proceedings. The empirical tool used
to probe subjectivity was Q method. The method provides insights into predominant
attitudes through the analysis of discourse (Barry and Proops 1999). It is a
qualitative, but statistical approach and differs from more conventional methods
of statistical analysis because it compares subjectivity across individuals rather
than comparing predefined variables (Brown 1993).8 The basic approach to Q-
methodology involves the identification of a concourse, or ‘running together of
ideas’, that emerge around an issue. Sample statements (which can also involve
pictures or some other medium relevant to the issue) are drawn from this concourse.
270 Simon Niemeyer
Respondents are asked to order statements from ‘agree’ to ‘disagree’, by ranking
them under a forced distribution (Brown 1993: 1–11).9 Individuals are then
correlated according to their responses and factor analysis is applied to draw out
dominant discourses among the group. These factors are then related back to the
original statements so that major characteristics, similarities and differences between
factors can be investigated. The Q-sort for the CJ was developed from a variety of
sources, both directly and indirectly related to the Bloomfield Track.10 The
statements were printed on individual cards, each with nine boxes labelled from
–4 to +4. Jurors were first asked to run through the cards and score each statement.
They were then asked to sort the cards from ‘most agree’ to ‘most disagree’ using
the original scores as a guide, although they were free to change the orderings
where they felt appropriate. The 42 cards were subsequently sorted into a quasi-
normal distribution (3, 4, 5, 6, 6, 6, 5, 4, 3 cards from disagree to agree). Analysis
of the Q data revealed a number of characteristic patterns. The resulting factors
each represent a particular type of subjectivity. The degree to which each juror
conformed to that factor is expressed in terms of a factor loading (a loading of 0
representing no conformance, 1 for complete conformance and –1 for complete
non-conformance). The factors were extracted by analysing all three sorts for all
individuals together, as though they were one group. Doing so facilitated comparison
between different sorts so that changes to each individual’s loading on a factor
could be examined. Measurement of preference shifts during the CJ was done
using a rank ordering of policy options. The options used for the preference rank
were obtained from information about the policy issue from which five distinct
alternatives were chosen, broadly representative of the major policy options. The
options included bituminisation of the Track (Bituminise); upgrading the Track so
it was passable by all vehicular types, but with a gravel surface (Upgrade);
stabilisation of the steeper slopes to reduce run-off (Stabilise); leave the Track as it
is under the current management regime (Status quo); and close the Track and re-
vegetate, with the possibility of using it for a walking track (Close).11
Results
In the following discussion the results from the CJ are briefly outlined. This is
followed by more substantial analysis and discussion to establish the implications
of deliberation for preferences and attitudes, and mechanisms for change thereof.
Attitudes
Measurable changes to attitudes did occur during the deliberative process. Yet
there were a number of constant aspects – in terms of their relative weighting.
Most significant of these was the tendency for jurors to place considerable emphasis
on long-term solutions to the issue, and its impact on future generations. For
example, jurors consistently agreed that a short-term perspective was inappropriate
for deciding the issue (see statement 30 in Table 12.1, x1 x3 = –2.08, s1 = 1.44, x3 =
–2.25, s3 = 1.22); and that inaction would damage the opportunities of future
generations (statement 25, x1 = 3.00, s1 = 0.74, x3 = 2.83, s3 = 0.94). There was
also a strong, though variable, commitment to conservation in the Daintree
irrespective of cost (statement 20, x1 = 2.67, s1 = 1.97, x3 = 2.42, s3 = 2.27). All
jurors tended to agree that the Bloomfield Track issue was an important one
(statement 12, x1 = 1.58, s1 = 1.00, x3 = 1.67, s3 = 0.78). One of the most intractable
issues among the jurors throughout the process was the level of permissible
development within the World Heritage area (statement 10, x1 = 0.92, s1 = 2.39,
x3 = 1.25, s3 = 3.39).12 There was also widespread disagreement regarding whether
the track should stay, given that it had already been constructed, although there
tended to be a net shift away from this position (statement 39, x1 = 0.75, s1 = 1.54,
x3 = –0.83, s3 = 2.66). As will be seen in the preference analysis below, this was
reflected in a division among jurors over the preferred policy option. Using
conventional Q analysis, described above, it is possible to establish an overview of
the attitudinal changes within the jury as a result of deliberation. By tracking
changes to juror’s loadings on each factor – representing the degree to which they
‘agreed’ or ‘disagreed’ with that subjective position – we can draw insights into
changes to attitudes during the deliberative process. The combined data for all
three Q sorts by the jurors produced four distinct factors: Preservation, Optimism,
Pragmatism and Symbolism. Statement scores for each factor are provided in Table
12.1. The following four paragraphs briefly describe the contours of each factor
and the degree to which their role in the attitudes of jurors changed during the
deliberative process. The description of each factor takes the combined form of
narrative regarding its defining features and includes a summary of factor scores
for relevant statements.
The first and most dominant factor – accounting for 40 per cent of variation
between sorts – is Preservation. Based on the statement scores for the factor given in
Table 12.1, Preservationists feel a personal commitment to protection of the Daintree
area. Statements with high scores for this factor tended to reflect an anti-development
sentiment towards the area (such as responses to statement numbers 10 and 32).
There is also a perception that the Bloomfield Track is a contributing factor to future
development (statement 2). (This is significant given, as previously mentioned, that
there was an increasing sense among the jury during deliberation that the Bloomfield
Track was part of an interconnecting web of regional impacts, which cannot be
properly considered in isolation.) Also evident is the belief that the use of vehicles on
the Bloomfield Track is detrimental to both rainforest (statement 21) and, to a lesser
extent, the reef (statement 7). Preservationists also feel they would be better off, or at
least no worse off, because of protection of the area (statement 11). Figure 12.2
shows how jurors changed their loadings on Factor 1 during deliberation. Each of
the bars represents the loading of each juror on the factor before, during and at the
conclusion of the CJ. (The names in the figure are the abbreviated codenames chosen
by jurors.13) The left-most bar represents the initial loading at the beginning of
deliberation. The middle (thinner) bar shows loadings midway and the darker right-
hand bar shows the loading of jurors at the end of the process. From Figure 12.2,
most jurors’ loadings on Preservation have changed substantially.14 Eight of the 12
jurors are significantly loaded at the beginning of deliberation (based on a level of
Table 12.1 Statement scores for Q attitudinal factors
Factor Score
No. Statement 1 2 3 4
1 Laying bitumen on the Bloomfield Track would be beneficial for the environment. It may even help reduce fuel usage –3 0 –3 1
and the greenhouse effect.
2 I don’t know if improving the Bloomfield Track would lead to a rapid acceleration of development in the area to the –2 3 –1 0
272 Simon Niemeyer
detriment of the environment.

3 In deciding on what to do with the Bloomfield Track I don’t know whether it’s more important to meet the needs of the –2 2 –2 0
community or the environment.
4 Whilst impacts on locals in the Bloomfield area are a concern, it is the broader community that should carry more weight 2 2 2 3
when deciding what to do with the Bloomfield Track.
5 I don’t know what the people of Bloomfield think about the Bloomfield Track. 1 2 2 –2
6 The road is just the ‘thin edge of the wedge’. Further improvement of the road will lead to more development in the area 3 –1 4 3
resulting in environmental damage. This may not happen for a long time, but it will happen.
7 Erosion from the Bloomfield Track is permanently damaging the coral reefs that fringe the beaches below. 1 –1 –4 3
8 When it comes to the Bloomfield track, people living in Cairns are in no position to judge what the interests of the local 0 1 0 0
residents of Bloomfield are.
9 If the Bloomfield Track is sealed (bitumenised) there will not be a rapid increase in environmentally damaging –3 –1 –2 –2
development in the Daintree area in the future. It may even benefit the environment there.
10 No development should be permitted in World Heritage areas such as the Daintree. 3 –3 1 0
11 I would be worse off if more of the Daintree rainforest is protected. –4 –2 –1 –1
12 The Bloomfield Track issue is important for Queensland. 2 1 1 0
13 The Bloomfield Track is bad for World Heritage values. 0 0 1 1
14 I’m not sure if the future of the Bloomfield Track should be determined by locals, outsiders or both. –1 1 –1 –1
15 The fate of the Bloomfield Track is of no concern to me. –4 –2 –3 –4
16 Economic development associated with the Bloomfield Track will provide more opportunities for future generations in –1 1 –1 1
North Queensland.
17 The future of the Bloomfield Track should be determined by everyone and not just by those who live in the Bloomfield 3 2 3 1
area.
18 I don’t know how, but I think that there must be some way in which everybody benefits from protecting the rainforest 2 4 1 2
near the Bloomfield Track.
19 The Bloomfield Track is important because it allows quick access to remote areas of the North. 0 2 0 2
20 Conservation in the Daintree area is worthwhile at whatever cost. 4 3 –2 4
21 Using cars on the Bloomfield Track is bad for the rainforest. 2 0 –1 0
22 Any decision about the Bloomfield Track will greatly affect people like me. 1 0 –1 2
23 I have no idea what the people in the Bloomfield area think about the Bloomfield Track. 1 1 1 0
24 Erosion from the Bloomfield Track does not cause siltation or damage to the fringing inshore reefs between Cape –1 –1 0 –3
Tribulation and Cooktown.
25 If we don’t take steps to protect the Daintree rainforest, future generations will miss out on the opportunity to experience 4 4 3 4
the area as we do now.
26 We don’t need to worry too much about environmental damage in the Daintree region because future generations will be –3 –4 –4 –4
better able to deal with these problems than we are.
27 There is no reason to believe that the Daintree Rainforest is under threat. –4 –3 0 –4
28 If future generations could have their say about the Bloomfield Track, they would be less concerned about the –2 –4 –2 –2
environmental impacts than many people make out.
29 The protection of plants and animals in the Daintree is OK so long as it doesn’t affect me. –2 –3 –4 –4
30 Let’s fix the problems in the Daintree just for now. The future will take care of itself. –3 –4 –3 –2
31 The more that it is possible for the average North Queensland resident access the Bloomfield Track the better. –2 –2 –2 –1
32 I don’t like how development is creeping further and further North into the Daintree and beyond because of its effect on 2 0 2 0
the environment.
33 The coral reefs along the foreshore below the Bloomfield Track are not badly affected by the road. 0 0 4 –3
34 Native animals in the Daintree need protection because they have a right to life which cannot be traded against economic 3 3 1 1
considerations.
35 The Wujal Wujal community is better off now that the Bloomfield Track has been built. 1 0 3 2
36 The most important use of the Bloomfield Track is for tourism. 1 –2 0 –2
37 I’m concerned that I will be made worse off by any decision about the Bloomfield Track. 0 –1 0 –2
38 I think that both short and long term perspectives are needed in deciding what should be done with the Bloomfield Track, 0 3 2 2
but I don’t know which one is more important.
39 The Bloomfield Track may not have been the best idea, but I guess there is probably little point in closing it now that it –1 –1 4 2
has been built.
40 I don’t really know who benefits most from the protection of rainforest in the Daintree. 0 1 0 –4
41 A long term perspective on the Bloomfield Track is essential. 4 4 3 4
42 When it comes to the Bloomfield Track, it’s not important to worry about what the future will hold. We need to worry –1 –3 2 –1
about now.
43 Everyone in Queensland is better off for having a road like the Bloomfield Track. –1 –2 –3 –2
274 Simon Niemeyer
100
90
80
Factor loading (x100)
70
60
50
40
30
20
10
0
ADV ASW BOA JAN JUL KEI KOD MAT PEA RAS SNO TAM
Juror
First sort Second sort Third sort
Figure 12.2 Juror loadings on factor 1 ‘preservation’
significance of 0.4).15 During deliberation there has been a net increase in factor
loadings ( x1 = 0.57, x3 = 0.61), with the exception of Adventure, Boat, Julie and
Keith, who have all decreased their loadings.
The Optimist archetype is characterised by someone who seeks an outcome where
everybody wins – environment and humans (exemplified by statement 18). Though
optimistic, they are very uncertain about a number of issues, including damage to
the area because of development (statement 2). They feel that the continued existence
of the Track will probably not lead to more development (statement 6), or destroy
the fringing reef (statement 7). Possibly in view of this, they are more predisposed to
development in the region than others (statement 10). Optimists are not inured to
future consequences. Indeed, they are less concerned about the present than others
(statement 42), but they are uncertain about whether now or the future is the more
important (statement 38). This could be a result of the level of uncertainty attached
to future outcomes, consequently greater significance is attached to the present – a
crude kind of discounting of future benefits.16 The Optimist is anthropocentric in
orientation.17 They feel that assessment of environmental problems should be based
on a perception of the efficacy of technological solutions.18 Greatest emphasis is on
how to achieve the best outcome. When questioned after deliberation, jurors most
strongly loaded on Optimism felt that closure of the Bloomfield Track should only
be undertaken once the costs are known to outweigh the benefits. These benefits
should be given proper account (statement 19). They feel that the Daintree could in
fact be under threat, but its conservation may not require the closure of the Bloomfield
Track (statements 20 and 27). Optimists tended to prefer technologically based ‘win–
win’ solutions. Optimism accounts for 9 per cent of variation between sorts. Net
loadings on the factor have steadily declined ( x1 = 0.26, x2 = 0.21, x3 = 0.17). As
90
80
70
60
50
40
30
20
10
0
-10
ADV AS W BO A J AN J UL KEI KO D MAT P EA RAS S NO TAM
-20
J uror
Figure 12.3 Juror loadings on faction 2 ‘optimism’
shown in Figure 12.3, only a few of the jurors subscribed to Optimism at any stage
of deliberation. Notably, Julie and Keith – the main dissenters at the end of
deliberation – have increased their loadings on the factor.
Pragmatism has many similarities to Optimism (with a correlation of 0.59).
Dominant features that distinguish Pragmatists from Optimists include scepticism
about anything but firm facts; confidence in their held beliefs; and less emphasis
on ‘win–win’ solutions. In contrast to the ecologically cautious approach of the
Preservationist – where the burden is to prove that activities do not damage the
area – Pragmatism focuses on the available facts. For example, because existing
research could not sustain the contention of reef damage, it is dismissed almost
entirely (statements 33, 7 and to a lesser extent 24). The pragmatic view of science
reflects this. Science is considered as the definitive input into decisions, rather than
a conservative and heuristic process that yields probabilistic outcomes. At its extreme
the logic is simply ‘no evidence, therefore no damage’ (and therefore, no road-
closure). As discussed below, there is a strong correlation between this position and
those jurors who prefer the Status quo option, provided that there is strict regulation
of Track use. The Pragmatist attaches more weight to present consequences than
future ones (statement 42), although they believe prudent decision making should
also benefit the future (statement 30). They are less ecocentric than Optimists
(statement 34); but they are not against preservation per se, believing that develop-
ment in World Heritage areas should be curtailed (statement 10). Because of the
conservatism of the Pragmatist they tend to place the burden of proof onto those
who wish to change the Status quo. Like Optimism, because damage to the Daintree
area does not appear significant (statement 27), conservation should be weighed
against benefits that development of the area may bring (statement 20). (However,
276 Simon Niemeyer
unlike the Optimist, Pragmatists are prepared to accept winners and losers
(statement 18).) Now that the Track is built and no definitive evidence says that it
should go, it may as well stay (statement 39); after all, it provides a benefit to the
community north of the Bloomfield River (statement 35). Pragmatists tend not to
be in favour of upgrading the Track (statements 6 and 9). As can be seen in Figure
12.4, pragmatism saw a marginal overall increase in factor loadings, although
these were far from uniform. Most of the increase can be attributed to three jurors
(Adventure 11 to 37, Janine –6 to 36 and Snoopy 7 to 33).
The last factor, Symbolic, resonates with the symbolic preferences discussed
earlier. Symbolists emphasise the impact of the road on fringing reefs, as well as its
benefit to the local community. There are only four statements that define the
Symbolic factor. Two of these concern strong belief in damage to the reef caused
by the Bloomfield Track (statements 7 and 33). There is a good deal of confidence
among Symbolists in ascribing benefits of protection (statement 40) as well as the
feelings of locals in the Bloomfield area (statement 5). Apart from these statements,
there is a high degree of affinity with factor 1, the Preservationist – with a high
level of correlation between the two (r = 0.63).19 Figure 12.5 shows changes to
loadings for Factor 4, Symbolism. Of all factors, it suffered the biggest decline
during deliberation. Beginning with five jurors loaded at the 0.15 level, by the end
of deliberation not one juror was significantly loaded – except for Keith and
Tamarra who were negatively loaded.
Preference shift
Attention is now turned to how preferences changed during deliberation. Table
12.2 shows the rankings given for each of the policy options before and after the
CJ process. To facilitate comparison of attitudes to preference, preference rakings
were subjected to factorial analysis similar to the attitude Q sort. This resulted in
four preference factors: Close Road, Stabilise, Bitumenise and Status quo.
Archetypal preference rankings for each preference factor position are provided in
Table 12.3. Figure 12.6 shows the loadings of jurors on each preference factor
before and after the deliberation. It can be clearly seen from both Table 12.3 and
Figure 12.6 that preferences changed substantially during deliberation. Two jurors
in particular – Snoopy and Aswad – experienced a near complete reversal of their
original orderings. During deliberation, preferences among jurors converged,
although they did not reach consensus.20 At the beginning there was a small
preference for minor road works to stabilise the Track and reduce run-off, or to
maintain the status quo. By the end of deliberation, road closure was clearly the
most preferred option.
Correlating preferences and subjectivity

It was evident from jury discourse even before the empirical results from the CJ
were analysed that jurors had experienced dramatic changes in their attitudes
concerning the Bloomfield Track issue. Jurors related during interviews, and in
80
70
60
50
40
30
20
10
-10
ADV AS W BOA J AN J UL KEI KOD MAT P EA RAS S NO TAM
-20
J uror
Figure 12.4 Juror loadings on factor 3 ‘pragmatism’
70
60
50
40
30
20
10
0
-10
-20
-30
ADV ASW BOA JAN JUL KEI KOD MAT PEA RAS SNO TAM
-40
Juror
Figure 12.5 Juror loadings on factor 4 ‘symbolism’
questionnaires, how they felt they had changed their opinions and the emphasis of
issues they felt were most important. Although this supports the hypothesis that
preferences change with attitudes, alone it does not reveal a causal relationship.
One way of detecting possible relationships between preference and attitude is to
compare whether those individuals correlated to a preference factor also tended
to load to a corresponding attitude factor. These overall correlations before and
Table 12.2 Rank ordering of preferences for management options
Rank before CJ Rank after CJ
278 Simon Niemeyer
Codename Bitumenise Upgrade Stabilise Status quo Close Bitumenise Upgrade Stabilise Status quo Close
Adventure 5 4 3 2 1 5 4 3 1 2
Aswad 4 3 1 2 5 5 4 3 1 2
Boat 5 3 1 2 4 5 4 2 3 1
Janine 1 4 3 2 5 5 4 3 2 1
Julie 2 1 4 4 5 5 3 1 2 4
Keith 4 3 2 1 5 5 3 1 2 4
Koda 2 5 1 4 3 4 5 2 3 1
Matilda 5 4 2 1 3 5 4 3 2 1
Pearl 4 3 2 1 5 5 4 3 2 1
Rastus 4 3 1 2 5 5 4 3 2 1
Snoopy 1 2 3 4 5 5 4 3 1 2
Tamarra 2 5 3 4 1 5 4 3 2 1
Overall rank 3 4 1 2 5 5 4 3 2 1
Average rank 3.25 3.33 2.17 2.42 3.92 4.92 3.92 2.50 1.92 1.75
St dev 1.54 1.15 1.03 1.24 1.56 0.29 0.51 0.80 0.67 1.14
Table 12.3 Factor scores for preference ranking
Factor Description Rank preference for factor
Bitumen Upgrade Stabilise Maintain Close
1 Close road 5 4 3 2 1
2 Status quo 5 3 2 1 4
3 Bitumenise 1 4 3 2 5
4 Stabilise 2 5 1 4 3
Figure 12.4 Correlation between preference and Q factors loading

Preference factor Close Status quo Stabilise Bitumenise
Stage of deliberation Before After Before After Before After Before After
Preservation 0.77 0.84 –0.26 –0.80 –0.09 0.16 0.41 –0.35
Attitude Optimism –0.51 0.32 0.00 0.43 0.24 0.07 0.06 –0.20
factor Pragmatism –0.15 –0.67 0.65 0.64 –0.31 0.23 –0.44 0.16
Symbolism 0.05 0.04 0.04 –0.02 0.51 –0.05 –0.09 –0.14
after deliberation are shown in Table 12.4. The following discussion explores these
relationships and mechanisms that might explain such correlations. It does so by
considering each preference factor in turn and those attitude factors that are
significantly correlated to it.
From Table 12.4 it can be seen that loading on the Close preference factor is
strongly correlated with the Preservation attitude. It is also negatively correlated
with both Optimism (decreasingly so) and Pragmatism (increasingly so). The nature
of the changes in correlation between attitude and preference can be illustrated by
plotting each juror’s loading on both factors at the stages before and after
deliberation. Figure 12.7 shows loadings on the Preservation and Close Road factor
spectrum. The asterisks represent jurors’ initial loadings (before deliberation), the
solid dots represent the comparative loading of jurors on each factor at the
conclusion of the deliberative process, and the slender arrows indicate the
‘migration’ of preference and attitude during deliberation. The more significant
shifts (those that are subject to further scrutiny herein) are highlighted using bold
arrows. The two heavily dashed regression lines (the lighter line indicates the
relationship at the beginning of deliberation and the darker after deliberation)
indicate the relationship between preference and attitude before and after
deliberation respectively. From Figure 12.7, most jurors have migrated north-east,
increasing loadings on both factors indicated by the broad arrow. Others (Keith,
Julie and Boat) have moved north-west, increasing loading on Close, but decreasing
on Preservation. Adventure’s shift is particularly notable: beginning the process at
the top of the Preservation–Close spectrum, but retreating somewhat from this
position as deliberation has progressed.
That Preservation and Close should be so strongly correlated is intuitively
appealing: given the way they feel that the track impacts on the region in concert
with multiple pressures, a Preservationist would feel it to be the only solution to
280 Simon Niemeyer
Before deliberation
100
80
60
40
20
-20
-40
-60
-80
-100
ADV1 ASW1 BOA1 JAN1 JUL1 KEI1 KOD1 MAT1 PEA1 RAS1 SNO1 TAM1
Close Status quo Bitumenise Stabilise
After deliberation
100
80
60
40
20
-20
-40
ADV3 ASW3 BOA3 JAN3 JUL3 KEI3 KOD3 MAT3 PEA3 RAS3 SNO3 TAM3
Close Status quo Bitumenise Stabilise
Figure 12.6 Changes to preference factor loadings before and after deliberation
JAN3 TAM MAT PEAADV1RAS
100
BOA3 ASW SNO3 KOD3
First Sort Final Sort
ADV3
80
60
Factor loading (close road) × 100
40
KEI3 JUL3 TAM1
20 BOA1MAT1
0
KEI1 PEA1 KODA1
-20
RAS1ASW1
-40
JAN1
-60
-80
SNO1 JUL1
-100
0 10 20 30 40 50 60 70 80 90 100
Factor loading (preservation) ×100
Figure 12.7 Changes to loadings on ‘preservationist’ and ‘close road’ factors
protecting the area. However, while Preservation is the most significant attitude
factor throughout, track closure only reached prominence by the end of the process.
Why did Preservationists not vote more strongly for track closure from the outset?
This anomaly might be partially accounted for by the possibility that jurors also
initially loaded on other attitude factors which ‘polluted’ their preferences. The
role of these other factors is discussed below. Another related possibility is that the
preferences were not fully ‘constructed’ at the outset, the correspondent preference
position crystallising during deliberation. There was some evidence to support the
claim that jurors did not initially vote in a manner true to their convictions. One
juror in particular expressed that she would have liked to rank track closure as her
first preferences, but did not think that it was a realistic outcome – her opinion,
and conviction, on this matter changed dramatically during deliberation.
As revealed in Table 12.4, preference for the Status quo is positively correlated
with Pragmatism throughout deliberation. During the process it has also become
correlated with Optimism and strongly negatively correlated with Preservation. The
position of jurors on the Pragmatism–Status quo spectrum is shown in Figure 12.8.
From the figure it appears that only Adventure and Snoopy were influenced by
Pragmatism such that they increased preference for the Status quo – the rest having
more or less vertical or horizontal shifts. Keith and Julie did not increase their loadings
on Pragmatism, but remained the most significantly loaded of the jurors within this
spectrum. There is evidence here regarding how different attitudinal factors interact
in forming preferences. Those individuals who increased their loadings in this
spectrum were also significantly loaded on the Preservationist. However, while they
clearly sympathised with the Preservationist attitude, they hesitated in whole-heartedly
advocating road closure because they felt more proof was needed beyond that
presented to prove the case beyond doubt.21 In Adventure’s case in particular, this is
282 Simon Niemeyer
consistent with a decline on loading on both Preservation and Close (see Figure
12.7), as well as the jury discourse. Like Adventure, the impact of Pragmatism also
marginally increased Snoopy’s preference for the Status quo, but migrating from an
entirely different overall attitudinal position. From Figure 12.7 it can be seen that
Snoopy experienced the most dramatic shift of all jurors, having begun the process
at the low end of the Preservation–Close Road spectrum. Figure 12.9 shows the
relationships between Optimism and Status quo. Snoopy began deliberation as an
Optimist, decreasing loading significantly by the end.
Based on the changes observed in Figure 12.9, increases in loading on the Optimist
attitudinal factor appears to have affected the preferences of Keith and Julie at the
end of deliberation. Both felt very strongly that the track should remain open to
permit access to the area by anyone who wished to visit. Like other jurors, they also
agreed with the substance of the preservationist position that the area is important
and under threat, but sought to find outcomes whereby everyone could benefit. Keith
in particular was keen to explore other solutions outside the mainstream of discourse
among the jurors – advocating the use of alternative approaches such as hovercraft,
or even a monorail. One plausible explanation for their position is that Keith and
Julie were the elder members of the jury. They were concerned that Closure would
effectively exclude access to those who, like them, were not able to use the walking
track that would replace it. They were optimistic that a solution could be found to
the satisfaction of all users. Due to restriction of time and evidence, the merits of
these alternative options were not fully deliberated – with view to their cost and
feasibility. Optimism alone does not account for the attitude of those jurors who
loaded significantly on the Status quo. It appears instead to act in concert with
Pragmatism. As can be seen in Figure 12.8, Keith and Julie were also highly loaded
100 BOA1 J UL3

RAS1 AS W1 KE I3
PE A
80 KE I1
60
Factor loading (Status quo)
MAT
40
SNO3ADV3
AS W
20
BOA3
TAM3 ADV1 KODA1 J AN3 J UL1
RAS3 MAT3 PE A3
0
J AN1
SNO1 KOD3
–20
–40
–60
TAM1
–80
–20 –10 0 10 20 30 40 50 60 70 80
Factor loading (Pragmatist)
Figure 12.8 Changes to loadings on ‘pragmatist’ and ‘status quo’ factors

BOA1 KEI3 JUL3

ASW1 RAS1
90
PEA1
KEI1
70
50
MAT1
Factor loading (Status Quo)
ADV3
SNO
30 ASW3
BOA3
10 JAN3
JUL1 TAM3 KOD1
PEA3 ADV1 MAT RAS
JAN1
–10
SNO1
KOD3
–30
–50
TAM1
–70
–10 0 10 20 30 40 50 60 70 80
Factor loading (Optimist)
Figure 12.9 Changes to loadings on ‘optimist’ and ‘status quo’ factors
on the Pragmatism factor. Optimism and pragmatism shared many features but
were distinguished by emphasis on different facets of the issue.
Stabilise and Bituminise are considered together here because of both their
similarity in preference structure and interaction with attitudes. Most strongly
associated with either stabilising the track, or bituminising is Symbolism (Table
12.4). The correlation between Symbolic and Stabilise is significant at the outset,
but not by the end of deliberation. The explanation lies in the ability stabilisation
or bituminisation (which is second preference for the factor, see Table 12.2) to
address symbolic issues of reef damage and community access. For a Symbolist,
choosing other options would place the competing goals of preservation and
community access in tension with one another. As deliberation progressed and
symbolic concerns dissipated, Symbolism was supplanted by emphasis on indirect
regional and long-term impacts, epitomised by the Preservationist factor.
Subsequently the loadings on Symbolism declined, along with a statistically
detectible relationship with the Stabilise preference factor.
Conclusions
Overall there appears to be a significant shift in attitudes resulting from discourse
in the CJ. Jurors adopted a ‘big picture’ perspective of the issue and generally
increased their predisposition toward preservation of the Daintree area. When
questioned about these changes, many jurors cited that improved knowledge helped
them better understand the implications of different actions (information effect).
The decline of symbolic attitudes, with concomitant changes to preference ranking
was one of the most striking outcomes of the process. This appears to reflect the
ability of deliberation to decouple attitudes from extreme positions transmitted to
284 Simon Niemeyer
the public domain through the political discourse of interest groups and issue
polarisation.
The evidence presented to the jury was based on a very small amount of research
directly relevant to the road. Damage to the fringing reefs is one of the few impacts
of the Bloomfield Track that has been studied to date. However, there was enough
information to address many of the myths and symbols that had dominated the
issue to that date. Most notable was the effect of evidence on the impact of the
road on the fringing reefs. Findings from this research, as presented to the jury, do
not support this contention. Rather than focus on these issues, jurors considered
the broader and longer-term implications of the future of the Bloomfield Track.
The fact that jurors were encouraged by the process to look at the issue from
the perspective of the citizen, rather than what benefited them individually (as
consumer) was also cited by jurors as a significant feature in changes to attitudes
and preferences. Based on observations and interviews with jurors, this perspective
change was most important insofar as it served as a pretext for the manner in
which information was weighted and used in the judgement process. There is
evidence in the Q sort for an attitudinal change that impacted on preferences.
There is also evidence of ‘better construction’ of preferences, to the extent that
individuals are better informed about what policy means best suit their desired
ends and a concomitant increase in correlations between attitudes and preferences.
Another lesser factor is increased fortitude of jurors to vote for those options they
most preferred.
Preferences tended to converge during deliberation, although discernible differ-
ences remained. Contours of difference emerged around different interpretations
of evidence. These reflected individual predispositions, such as whether they
believed science could provide certain answers, or whether technology could solve
otherwise intractable problems. All jurors agreed on the importance of the area.
However for some, the fact that science could not prove beyond doubt that the
track caused irreparable damage led some jurors to adopt a slightly more conser-
vative and pragmatic approach, deferring closure until more evidence becomes
available. Two other jurors saw opportunities for ‘win–win’ in the evidence
presented: that technically sophisticated options might satisfy competing ends.
Differences also reflected a form ‘starting point bias’, where jurors who began
loaded on one extreme of an attitude–preference spectrum, maintained a residual
affinity with that position. The contours of difference, though still significant,
became more clearly identifiable and therefore arguably easier to address. The
question remains what role further deliberation directly addressing these difference
may have. It is possible that directly addressing these issues might have the same
impact as evidence by dissipating symbolic concerns, leading to acceptance by the
rest of the jury. Based on the evidence from the CJ on the harmonising capacity of
deliberation, it is less likely to lead to increasing polarisation, but this is dependent
on a number of factors, not least being willingness of jurors to tolerate scrutiny of
their formed position by others. In either case it seems that deliberation is never
complete. Despite emergent differences among the jury, the overwhelming trend
was toward increased concern for the longer-term and interconnected issues with
the Bloomfield Track, and a concomitant preference for track closure. This outcome
is very different to preferences and attitudes elicited at the beginning of the process,
which varied widely but tended to support maintaining the status quo.
This chapter is part of a broader project of ecological economics to organise
human institutions to elevate the cause of ecological sustainability. It is specifically
concerned with processes that shape preferences to be consistent with this goal.
Within this aim, formal deliberative processes such as CJs are an arguably better
foundation for public input into environmental policy than aggregation exercises
such as referenda or the contingent valuation method (see Niemeyer and Spash
2001, and also Aldred this volume, Chapter 8). The deliberative process helps to
improve judgement as far as preferences tend to match means to desired ends
better than when elicited without subjecting individuals to the deliberative process.
The matching of preferred means to ends suits the goal of sustainability if, because
of deliberation, each individual desires that collective outcome. Indeed, it appears
that the attitudes of participants changed in a manner conducive to the provision
of the common good.
Deliberative approaches may also help to avoid a potential trap for ecological
economics as the ‘science of sustainability’ where predefining outcomes subverts
both democratic process and the achievement of sustainability. Although the trend
favouring preservation was a strong feature of the CJ, the detail of jury discourse
has revealed more subtle differences on the margins of debate about what is the
most preferred outcome. The main driver of these differences was perceptions
among the jurors about the nature of the scientific evidence and the ability of
technical solutions to produce optimal outcomes. If unanimity is the benchmark
for deliberative success, that the CJ produced multiple options could be considered
a less than optimal result. An alternative view is that democratic process should
help to identify substantive differences (for example Benhabib 1996). Within this
view, the CJ marks an important step in defining substantive issue contours to be
addressed by policy. These results are arguably more ‘democratic’ because the
information and deliberative process provided individuals with the necessary tools
to make their own judgements, rather than defer to information they might normally
receive through conventional channels of political discourse. Where conventional
democratic processes might normally provide individuals with predefined policies
deemed by experts or elites to be ‘sustainable’, the deliberative process has produced
positive contributions by identifying a number of possible paths. The result places
the onus back on elites to establish better information from research, or better
educate the public about the use of science in policy. Even if this was the only
outcome of formal deliberative processes, they provide a surer footing for the
development of sustainable policies.
There are important caveats to the conclusions drawn here. First, the results
are only from one deliberative process. Others have conducted similar processes,
with mixed results (for example see Pelletier et al. 1999 and also the previous chapter
in this volume). If anything, this proves that the outcome of deliberative forums is
highly sensitive to context and design. Therefore, they are amenable to problems
of political manipulation and legitimacy. An important question also concerns
286 Simon Niemeyer
whether changes to preferences observed within the CJ could be replicated in the
broader public if the political environment was more ‘discursive’ (in the sense of
Dryzek 2000). If so, rather than rely exclusively on manufacturing formal
deliberative process, a better approach is to facilitate critical deliberation in the
public sphere.
It is possible that adopting formal deliberative approaches in the medium term
will facilitate this transformation, over the long term empowering citizens in the
democratic process through improved and relevant information. However, with
scope for manipulation, the adoption of deliberative approaches will depend on a
genuine desire on the part of organisers and elites to replicate discursive ideals.
An important conclusion is that the ability of deliberation to transform preferences
cannot be divorced from an appreciation of political context. This provides even
greater impetus for ecological economics to embrace the political dimension of
the science of sustainability.
Notes
1 For example, in the case of voting, Downs (1957) observed that voters had little incentive to
acquire information about alternatives because of the small marginal gain and individual impact
on election outcome.
2 There is a possibility that deliberation may not always lead to better preferences. Symbolic
preferences may not dissolve in the process of acquiring information during the deliberative
process, but become entrenched. Depending on an individual’s ability to assimilate information,
it is entirely possible that too much detail can lead to information overload and reduce capacity
for judgement, for that is what contributes to symbolic preference in the first place. Within the
deliberative process there is an ongoing tension between the acquisition and assimilation of
information, which if not properly balanced can lead to dramatically different results (e.g. Pelletier
et al. 1999).
3 There was a statutory obligation under state law that an environmental impact assessment be
prepared for major public works, often treated as no more than a symbolic requirement in an
administrative environment geared for development. To satisfy this requirement, the shire
engineer who oversaw the construction of the road prepared a short impact assessment, which
failed to find against road development.
4 A combination of factors has meant that the Bloomfield Track has effectively been in policy
stasis since construction. This was partially broken in 1998 when, due to prohibitive cost of
maintenance of a dirt track following steep contours in a heavy rainfall region, the Douglas
Shire Council applied to the Wet Tropics Management Authority to upgrade the steepest section
and lay gravel along major sections. They granted an approval for the works on the condition
that the local council prepared a comprehensive environmental impact assessment with
recommendations for providing for long-term management. To date the impact assessment has
yet to be commissioned.
5 Jurors were accompanied on this trip by an engineer from the Douglas Shire who was on hand
to answer questions about construction of the road, and a well-known local and tour operator
in the Daintree area who gave the issue a historical and local context as well as pointing out
major features.
6 Representation of indigenous interests was mainly achieved by the fortuitous recruitment of a
Kuku-Ylanji (traditional owners of the Daintree area) woman as part of the jury.
7 Preferences were not surveyed at this midway point because of concern that, because of the
simpler nature of the questionnaire, preferences surveyed too close together might be biased by
jurors feeling that they should be consistent with previous preference orderings.
8 Although for this research there is a mixing of methodologies by assessing changes within
individuals over time.
9 This approach, referred to as ‘forced distribution’ is commonly used in Q method to facilitate
analysis by keeping means equivalent, but does not affect the actual result so long as the rest of
the study is well designed (Brown 1980: 198). Some of the jurors found it difficult to conform to
this imposed regime and were given the freedom to depart from the template where they felt
doing so better represented their position.
10 Statements were grouped into whether they pertained to beliefs about the consequences or the
fundamental beliefs about the importance of those consequences. Statements were also grouped
according to whether they involved a level of knowledge about the issue, the way in which the
issue was framed or conceived – a citizen or a consumer, for example – and the time frame over
which the implications of decisions about the issue are considered. Statements within each
category were phrased either as a positive statement, a negative statement, or a statement
concerning the level of confidence in a particular belief. Expression of certainty is an important
element of the concourse and including it in the survey assists the researcher to detect changes
in confidence as a result of the deliberative process.
11 Importantly, when making recommendations at the end of the process the jury was not limited
to these options. The ranking of preferences was simply used to follow the way in which their
ordering of these particular options had changed.
12 The loading on statement 12 tended to reflect the juror’s predisposition to finding technical
solutions (of which one was the construction of a mono, or sky rail) or regulatory solutions to
the Bloomfield Track issue.
13 Because of the controversial nature of the issue, jurors were given the option of choosing
codenames to protect their identity in publications arising from this research.
14 Significantly, most revealed markedly higher shifts during the first half of the process, in some
cases ‘settling back’ toward their original position, though still substantially changed. Due to
space restrictions, the implications of this cannot be discussed here.
15 Based on a level of significance of 1/ n 0.15 all jurors are significantly loaded on the factor at
the outset, with only Keith defecting. However, although statistical significance is the commonly
accepted test for difference, Q method is based on a measure of ‘psychological significance’,
which is a phenomenologically rather than statistically determined level (Brown 1980: 198).
Using the statistically accepted level for significance suggests a higher level of agreement among
jurors than was actually observed during the CJ proceedings. For the purposes here, because of
the high level of correlation between factors and significant loading of jurors on more than one
factor, a level of 0.4 is adopted.
16 This type of response is consistent with the ‘pioneering’ ethic ascribed to remote areas of
Northern Australia. Many authors portray early pioneers in terms of rapacious developers with
little concern for the environment and future consequences (e.g. Lines 1992). Passmore (1974)
illuminates the issue from a different perspective whereby actions were driven precisely because
of concern for future generations, but unconditioned by awareness, or belief in, negative environ-
mental consequences. If there is certainty that exploitation of environmental resources has
negative consequences, then there is a case for changing course.
17 There are a number of similarities with the Individualist position identified in Steg and Sievers
(2000).
18 This is also consistent with the pioneering ethic, as was epitomised by the Premier of Queensland
at the time the Bloomfield Track was constructed (Passmore 1974: 77; Fitzgerald 1984).
19 Symbolic contrasts most strongly with the Pragmatic (r = 0.38), with disagreement on the question
of the reef (7, 24, 33) and impact of the road on the rainforest (27).
20 One participant expressed pleasure that consensus was not reached because: ‘the fact that we
did not come to a definite solution, I think, shows how successful the jury process is at tackling
any problem … this problem does not have a definite solution’.
21 Many of the technical witnesses themselves pointed out the poor quality of the evidence directly
relevant to the road gathered to that date.
288 Simon Niemeyer
References
Barry, J. and Proops, J. (1999) ‘Seeking sustainability discourses with Q methodology’,
Benhabib, S. (1996) ‘Toward a deliberative model of democratic legitimacy’, in S. Benhabib
(ed.) Democracy and Difference: Contesting the Boundaries of the Political, Princeton, NJ: Princeton
University Press.
Brown, S.R. (1980) Political Subjectivity: Applications of Q Methodology in Political Science, New
Haven: Yale University Press.
Brown, S.R. (1993) ‘A primer on Q methodology’, Operant Subjectivity, 16: 91–138.
Callicot, J.B. (1982) ‘Hume’s is–ought dichotomy and the relationship of ecology to
Leopold’s land ethic’, Environmental Ethics, 4: 163–73.
Common, M. and Perrings, C. (1992) ‘Toward an ecological economics of sustainability’,
Downs, A. (1957) An Economic Theory of Democracy, New York: Harper & Row.
Dryzek, J.S. (2000) Deliberative Democracy and Beyond: Liberals, Critics, Contestations, Oxford:
Elster, J. (1983) Sour Grapes: Studies in the Subversion of Rationality, Cambridge: Cambridge
University Press.
Etzioni, A. (1986) ‘The case for a multiple utility conception’, Economics and Philosophy, 2:
159–83.
Fitzgerald, R. (1984) From 1915 to the Early 1980s: A History of Queensland, St Lucia, Australia:
University of Queensland Press.
Goodin, R.E. (1996) ‘Enfranchising the earth and its alternatives’, Political Studies, 44:
835–49.
Goodin, R.E. (2000) ‘Democratic deliberation within’, Philosophy & Public Affairs, 29: 81–
109.
Harrison, C.M., Burgess, J. and Clark, J. (1999) ‘Capturing values for nature: ecological,
economic and cultural perspectives’, in J. Holder and D. McGillivray (eds) Locality and
Identity: Environmental Issues in Law and Society, vol. 3, Dartmouth: Ashgate: 85–110.
Keat, R. (1994) ‘Citizens, consumers and the environment: reflections on “The economy
of the earth” ’, Environmental Values, 2: 333–49.
Kellert, S.R. and Wilson, E.O. (1993) The Biophilia Hypothesis, Washington, DC: Island
Press.
Kellow, A. and Niemeyer, S. (1999) ‘The development of environmental administration in
Queensland and Western Australia: why are they different?’, Australian Journal of Political
Science, 34: 205–22.
Lines, W.J. (1992) Taming the Great South Land: A History of the Conquest of Nature in Australia,
North Sydney: Allen & Unwin.
Niemeyer, S.J. and Spash, C.L. (2001) ‘Environmental valuation analysis, public delibera-
tion, and their pragmatic syntheses: a critical appraisal’, Environment and Planning C, 19:
567–86.
Norton, B.G., Costanza, R. and Bishop, R.C. (1998) ‘The evolution of preferences: why
‘sovereign’ preferences may not lead to sustainable policies and what to do about it’,
O’Neill, J. (1997) ‘Managing without prices: the monetary valuation of biodiversity’, Ambio,
26: 546–50.
Papadakis, E. (1993) Politics and the Environment: The Australian Experience, St Leonards, NSW:
Allen & Unwin.
Passmore, J. (1974) Man’s Responsibility For Nature: Ecological Problems and Western Traditions,
London: Duckworth.
Pelletier, D., Kraak, V., McCullum, C., Uusitalo, U. and Rich, R. (1999) ‘The shaping of
collective values through deliberative democracy: an empirical study from New York’s
North Country’, Policy Sciences, 32: 103–31.
Rydin, Y. (1999) ‘Can we talk ourselves into sustainability? The role of discourse in the
environmental policy process’, Environmental Values, 8: 467–84.
Sagoff, M. (1988) The Economy of the Earth: Philosophy, Law and the Environment, Cambridge:
look beyond contingent pricing’, Ecological Economics, 24: 213–30.
question: a verbal protocol analysis of willingness to pay for an environmental
regulation’, Journal of Environmental Economics and Management, 26: 88–109.
Sears, D.O. (1993) ‘Symbolic politics: a socio-psychological theory’, in S. Iyengar and W.J.
McGuire (eds) Explorations in Political Psychology, vol. 5, Durham: Duke University Press.
Slovic, P. (1995) ‘The construction of preference’, American Psychologist, 50: 364–71.
Smith, G. and Wales, C. (2000) ‘Citizens’ juries and deliberative democracy’, Political Studies,
48: 51–65.
Spash, C.L. and Hanley, N. (1995) ‘Preferences, information and biological diversity’,
Steg, L. and Sievers, I. (2000) ‘Cultural theory and individual perceptions of environmental
risks’, Environment and Behaviour, 32: 250–69.
van den Bergh, J., Ferrer-Carbonell, A. and Munda, G. (2000) ‘Alternative models of
individual behaviour and implications for environmental policy’, Ecological Economics,
32: 43–61.
Politics, 8: 75–96.
Wilderness Action Group (1983) The Trials of Tribulation, Mossman: Wilderness Action
Group.
290 Index
Index
Aarhus Convention (1998) 4 Ballestero, E. and Romero, C. 122

agri-environmental forum (AEF) see Bamberg, S. 52; et al. 65
GRANO project Bandura, A. 78
agri-environmental schemes 14 Barry, J. and Proops, J. 269
Agricultural Production Co-operation Bateman, I.J. and Langford, I.H. 51,
(LPG) 256 54
Ajzen, I. 52, 75; et al. 9; and Fishbein, M. Bayesian theory 198
75; and Madden, T.J. 75 Becker, G.S. 30, 72
Aldred, J. 26, 192, 201, 204; and Jacobs, behaviour see human behaviour
M. 197, 201, 205, 209, 228 Benhabib, S. 285
analytic hierarchy process (AHP) 9, 103, Berkes, F. and Folke, C. 244
140–3, 148; application of 109; check Bernabo, J.C. 249
logical consistence of pairwise bid curve analysis 7
comparisons 108–9; consistency Bingham, G. et al. 31, 32
108–9, 130 (4n); construct suitable biodiversity 24, 26, 43, 45 (4n)
hierarchies 106–7; development of biophysical systems 4
106; establish priorities between Bishop, R.C. and Heberlein, T.A. 53; and
hierarchies 107–8; in practice 109; Welsh, M.P. 29
random index 109, 130 (5n) Blackburn, M. et al. 28
Anand, P. 205 Blamey, R.K. 31, 209, 238; and
Andreoli, M. and Tellarini, V. 99, 129 Common, M. 227; et al. 51
Andreoni, J. 46 Blaug, R. 240
Aristotle 196 Bloomfield Track (Australia), attitudes/
Armour, A. 201, 204 responses 270–6, 287 (14n–19n);
Arrow, K.J. 46, 161; et al. 23, 33, 45, 209 conclusions concerning 283–6;
attitudes, CJs/Bloomfield Track 270–6, construction of 268–9, 286 (3n, 4n);
283–4; group influences upon 76–8; correlating preferences/subjectivity
information 15; as predictors of WTP 276–83, 287 (21n); deliberation on
within CVM 52–4; New Environ- 284–5; geographical location/
mental Paradigm (NEP) 52–3; towards background 267–8; optimism
the environment 75; towards money as concerning 274–5, 279, 281–3;
measure of value 60; towards nature/ polarisation of interests in 269;
climate change 57–9 pragmatism concerning 275–6, 279,
Austen-Smith, D. 195, 196; and Riker, W. 282–3; preference shift 276, 287 (20n);
196 preservation 271, 274, 279, 281;
Australia 15; forest regions 9–10, 134–56; results of study 270–83; study
parkland study 229–36, 240–1 (5n, methodologies 269, 286–7 (7n–11n);
6n); wet tropics area 267–86 symbolism 276, 283; use of CJs
Austria 7, 24, 32–43 269–70, 286 (5n, 6n)
Index 291
Bohman, J. 192, 200, 204, 237; and Rehg, Common, M. et al. 27, 31; and Perrings,
W. 204 C. 31, 263
Borders Forest Trust (BFT) 212 commons dilemma 83–6
Borders Region (Scotland) 12, 211–12 communicative influence 87; and attitude
Boulding, K.E. 27, 28, 29, 45 position 88; and biased processing 88;
bounded rationality 197–9 external input variables 87;
Boyce, R. et al. 27, 31 interventions 89–91; personal/actual
Boyle, K.J. and Bergstrom, J.C. 212, 213; (state) values discrepancy 87; and
et al. 213 persuasiveness of communicator 89;
Brandenburg (Germany) 250–60 and self-responsibility 87–8
Breusch–Pagan test 63–4 Comprehensive, Adequate and
Bromley, D.W. 249 Representative (CAR) Reserve system
Brouwer, R. et al. 201; and Slangen, 135
L.H.G. 55 Comprehensive Regional Assessment
Brown, S.R. 269 (CRA) (Australia) 9, 134, 135, 136–7
Brown, T.C. et al. 5, 201, 209, 210, 228 Conlisk, J. 31
Buchanan, J. 192 consumer boycotts 16
Burgess, J. et al. 29, 32 consumer–citizen dichotomy 24, 30–2,
Burney, J. 2 43, 46 (12n, 15n)
context 16
Callicot, J.B. 266 contingent valuation method (CVM) 2–3,
Cameron, J.I. 31 5, 7–8, 17, 51, 65, 209, 212–14, 225,
Cardinal Consistency Index 144 244, 285; and CBA 43; criticisms of
Carinthia (Austria) 24; descriptive 43, 226–7; (dis)agreements concerning
overview of empirical results 33–5; 23; extended model 24, 26–32; and
design of survey 33; empirical results home-grown values 29; and
of survey 32–43, 44, 46–7 (18n–23n); hypothetical bias 23, 28; and
explaining WTP-bids 35–7, 43; individual willingness to pay 23;
importance of extended valuation institutional context 28–30;
framework 35–7, 43; social context 32 lexicographic preferences/willingness
Carson, R.T. 26; et al. 212 to exchange 24–7, 46 (8n); and
Carter, J.R. and Irons, M.D. 29 respondents’ beliefs/perceptions 24;
Castells, N. and Munda, G. 99, 129 social context 24, 30–2; standard
CHAID (Chi-squared Automated model 24, 45 (2n); in theory 205 (9n);
Interaction Detector) 60, 61–3 usefulness of 44; and valuation
Chang, N.-B. et al. 124 questions 29; and value construction
charitable giving 7, 213 29; and welfare economics 43, see also
choice modelling 226, 240 (2n) CVM/CJs comparison
Churchman, C.W. et al. 102 Contract Nature Conservation
citizens’ juries (CJs) 14, 15, 17, 204–5 Programme (Germany) 251–2
(3n), 225; advantages of 209; Coote, A. and Lenaghan, J. 192, 201, 204,
application of 228; Bloomfield Track 209, 210, 228, 236
269–83, 285; cf contingent valuation cost–benefit analysis (CBA) 1, 23, 43, 75,
method 11–12; design/conduct of 102, 160, 226–7; and attitudes 52–4;
229–33; development of 228; Ettrick and benevolence/malevolence 27;
Valley project 214–22; issues counter-critiques 17; criticisms of 1–2,
concerning 240; park management 3–4, 27; and heroic ethic 27;
229–40, 241 (8n–10n); and pioneering popularisation of 2; and preferences
ethic 287 (16n); results of deliberations 26; validity in applying 4
233–6, see also CVM/CJs comparison Crosby, N. 201, 204, 210, 228
civil rights 4 Cummings, R.G. et al. 28
Clark, J. and Murdoch, J. 249 Curry, N. and Winter, M. 251
Climaco, J. 104 CVM/CJs comparison 11–12;
climate change 57–9 advantages/weaknesses 211;
292 Index
agreement/deliberative valuation Dillon, J.L. and Perry, C. 102, 103, 104,
200–3; background 187–8; bounded 105
rationality/new perspectives 197–9; discursive democracy 227–8, 240 (3n)
decision-making processes 189–90, Dovers, S. 139
205 (4n, 5n); deliberation/discussion Downs, A. 286
190–1; economic appraisal methods Doyle, T. and Kellow, A. 139
209; epistemic merits 191–2; equity/ Dryzek, J.S. 227, 236, 240
distributional issues 210; feasibility Dubourg, W.R. et al. 45
constraint 203–4; policy implications Dunlap, R.E. and van Liere, K.D. 52
204; pooling private information Dwyer, W.O. et al. 93
195–7; questionnaires 209–10; and
rationality 210–11; restoration project Eagly, A.H. and Chaiken, S. 52, 58
211–22; and value construction 210 ecological systems 4
economic utility theory 26
Daneke, G.A. et al. 225 economic values, focus on CVM 6; and
de Dios, E.S. 26 psychology of behaviour 6–9; specific
De Montis, A. et al. 100, 125, 129 validity issues 6–7
DeBono, E. 245 economics 28–30
Dechamps, H. 211 Edwards, S.F. 26
decision making 31–2, 112, 139, 157 (2n), Ehrlich, P. and Ehrlich, A. 26
211; and AHP 106–9, 109; and Elaboration Likelihood Model 87–9
ELECTRE III 113–15; and Evamix ELECTRE III 9, 113–15, 117
109–12; by lexicographic ordering 27; Elster, J. 190, 192, 193, 210, 211, 263,
and MAUT 102–3, 102–6, 105; and 264, 265, 266
MOP/GP 121–4; and NAIADE Encarnación, J. 26
118–21; processes 189–90; quality of Environment and Heritage Assessment
4; and Regime 115–18; social/ (Australia) 136
institutional context 24, 45 (3n); environmental assets 55
software packages 102–3, see also environmental behaviour, assumptions 75;
multiple criteria decision analysis and attitudes 92; background 69–70;
(MCDA) commons dilemma 83–6; discussion of
deliberative democracy 187, 190–1, 205 findings 91–3; experiments/results
(6n), 225, 227–8, 240 (3n); and 76–91; framework model 71–6; group
agreement 200–3; as alternative to influences upon attitudes 76–8;
aggregative models 188, 204 (2n); internal/external input variables 72,
appealing to principles 193–5, 205 73; and low susceptibility to influence
(8n); avoiding appearing selfish 194–5; 91; pioneers 69–70, 78, 287 (16n);
and better dicisions 191–2; and processing communicative influence
‘cheap-talk’ literature 196; and 87–91; and role models 91; simulation
credible communication 196; critics of procedure 70–1; and social influence
195–7, 199; diagnostic/constructive 92–3; social learning of behaviour
distinction 191; and divergent interests 78–83; and state of the resource 91;
198–9; empirical/theoretical 193; and sub-models 74–5; turn-arounds in 70
rational choice/game theories 191; ethics 9, 27, 287 (16n)
strategic reasons 193–4; theories of Ettrick Valley Forest floodplain project 12,
188, 204 (1n); and trust 198, see also 211–12; access/future management
discursive democracy concerns 215–16; benefit of 214;
deliberative monetary valuation (DMV) 4, charitable donation payment/open-
5–6, 11–14, 15, 17, 201–2, 204 ended elicitation format 213; CJ
deliberative process 263–4, 266, 284–6 (2n) survey 214–16, 223 (5n); CVM survey
dichotomous choice (DC) format 213 212–14, 222 (1n), 223 (3n); decision-
Diekmann, A. 58, 69, 72; and Franzen, A. making process 221–2; deliberations
93 214–15; funding for 220–1; and
Dienel, P.C. 225, 228 policy-relevant information 219–20,
Index 293
222; positive/negative views on 218, Gastil, J. 237
219; qualititative/quantitative data on Generalized Algebraic Modelling System
217–18; questionnaires used 221, 223 (GAMS) 136
(7n); recommendations concerning genetically modified (GM) crops 10;
215; research conclusions 219–22, 223 application of MCM to 172–80;
(7n); valuation workshops 216–19; and criteria 177–8; findings 175–80;
willingness to pay 213–14 options 176; pilot study on 172–5;
Etzioni, A. 30, 266 ranking of options 177–80; UK debate
evaluation matrix (Evamix) 9; appraisal over 172; use of MCM 172–82
scores 111–12; development of 109; Geographic Information System (GIS)
dominance scores for ordinal/cardinal 137
criteria 110–11; ordinal/cardinal Georgiou, S. et al. 24, 52
distinction 109–10; overall dominance Germany 14, 250–60
scores 111; in practice 112; Getzner, M. 28, 32
standardized dominance scores for Gonzales, M.H. et al. 90
ordinal/cardinal criteria 111 Goodin, R.E. 266
existence value 3 Goodman, S. et al. 57
expanded rational choice model 72–6; governments 10, 16
attitude 72; behavioural intentions Gowans, C. 195
75–6; commons dilemma 83–6; Gowdy, J. 26; and Olson, P.R. 26
conclusions concerning 91–3; group GRANO project 260 (1n); analysis of
influences upon attitudes 76–8; agri-environmental forum 256–8;
knowledge 74; motives 74; criticisms by participants 256;
persuasiveness 74; processing development of 250–2; discussion of
communicative influence 87–91; state measures 255–6; financial aspects 253;
of resource 74; self-responsibility 74; implementation of 252; inefficent
situational factors/incentives 74; social decision-making 259–60; information
learning of behaviour 78–83; status flow 253, 259; interdisciplinary team
74; sub-models 74–5; use 72; values 252; knowledge concerning 255;
74; and volition 75–6 learning process 253, 258–9;
Exxon Valdez 2 meetings/proceedings 253–6;
relevance of topics 253, 259; results
Farrell, J. and Gibbons, R. 195 258; stakeholders 252
Fearon, J. 191, 192, 193, 195, 197, 198 Green, C. 65; et al. 64; and Tunstall, S. 65
Ferguson, I. 139 Green Party 16
Fischhoff, B. 31, 53; and Furby, L. 31 Greene, W.H. 63
Fishbein, M. and Ajzen, I. 45, 52 Greenhouse Effect 2
Fishburn, P.C. 102 Gregory, R. et al. 210, 226; and Slovic, P.
Fisher, E. and Harding, R. 167, 171 226, 227
Fitzgerald, R. 287 Grove-White, R. et al. 172
focus groups 5, 16, 201, 213 Gutmann, A. and Thompson, D. 192
Forest and Resource Management System
(FRAMES) 137 Habermas, J. 240, 245
Forest Resource Use Model (FORUM) Hagedorn, K. 247, 249
136, 137–8 Hahnemann, W.M. 46
forest resources see Southern Forest Hampicke, U. et al. 252
Region (New South Wales) Hanley, N. et al. 26, 29; and Spash, C.L. 1
Fransson, N. and Gärling, T. 65 Hanna, S. et al. 244
Frey, B.S. 31 Harrison, C.M. et al. 265
Frey, D. 76; et al. 75, 76, 77 Harrison, G.W. 28
Funtowicz, S.O. 4; and Ravetz, J. 99, 162 Harsanyi, J.C. 69
Hayden, F.G. 24
Gambetta, D. 192 Hayek, F.A. 164
Garrod, G.D. and Willis, K.G. 51 hedonic pricing 5
294 Index
Hinloopen, E. 115; et al. 115; and Jorgenson, D.O. and Papciak, A.S. 81
Nijkamp, P. 115, 116, 117 Journal of Economics Perspective 29
HIPRE 3+ 103
Hodgson, G.M. 24 Kahneman, D. et al. 52; and Knetsch, J.L.
Hoehn, J.P. and Randall, A. 46 46; and Tversky, A. 3
Hofer, D. 57 Kaiser, G.F. et al. 53
Hogwood, B. and Gunn, L. 139 Kapp, K.W. 1
Holland, A. 32 Keat, R. 46, 266
Holling, C.S. et al. 244 Keeney, R.L. and Raiffa, H. 102
Holm-Müller, K. et al. 29 Kellert, S.R. and Wilson, E.O. 266
homo oeconomicus 23, 29 Kenyon, W. 215; et al. 5
Howlett, M. and Ramesh, M. 139 Klandermans, B. 75
human behaviour 3–4; altruism 30; Kmietowicz, Z.W. and Pearman, A.D.
assumptions concerning 71; CBA 116
component 8; and cognition 69; Knetsch, J.L. 3; and Sinden, J.A. 4
economic/non-economic motives 7–8; Knight, F. 129
and environmental consciousness Korhonen, P. 121, 122
69–93; and ethics 9; expanded rational Kosz, M. 31
choice model for 72–6; free-rider 29; Kotchen, M.J. and Reiling, S.D. 52, 53,
importance of attitudes 7–8; personal/ 59, 64
social factors 69; and ‘pioneers’ 8, 287 Krehbiel, K. 195
(16n); predictability of 65; and KULAP (cultural environmental
pressure to conform 77–8; and landscape programme) 251
processing of information 8–9; Kuper, R. 204
psychology of 6–9; rational 30; and
self-concept 77; and self-confirmation Lane, R. 192
81; social comparison processes 76–7; Lara, P. and Romero, C. 122
social learning of 78–81; social/ Latané, B. 79, 80, 92; and Wolf, S. 79
personal factors 8; and status 81; Lazo, J.K. et al. 29
sustainability motives 8 learning process 245, 249–50
Hwang, C. and Yoon, K. 103 Lee, S.M. 121
Hylland, A. 192 lexicographic preferences 26–7, 43, 44, 45
hypothetical bias 28 (7n), 46 (8n–10n)
Liebrand, W.B.G. 83; et al. 83, 192
Ignizio, J.P. 121 Loasby, B. 162
incertitude 163–5 Lockwood, M. 26, 27
indifference curve 26 Lothian, J. 139
interactive valuation process, analytical Luce, R.D. 105
framework for 246–50; background Luhmann, N. 26, 46
244–6
McDaniels, T.L. and Roessler, C. 29
Jacobs, M. 5, 31, 188, 201, 204, 209, 210, Macedo, S. 192
228 Mackie, G. 196
Jakobsson, K.M. and Dragun, A.K. 28 Mann–Whitney test 56
James, R. and Blamey, R.K. 201, 202, Mardle, S.J. et al. 123
203, 209 Martinez-Alier, J. et al. 4
JANIS criteria 135 Marwell, G. and Ames, R.E. 29
Janssen, R. 99, 129 Massam, B. 138
Jeffrey, R.C. 26 Mayer, S. et al. 172
Jiggins, J. and Röling, N. 244, 245, 248 Mendoza, G.A. and Prabhu, R. 99, 129
Johansson, P.-O. 26 Menegolo, L. and Guimarães Pereira, A.
Johnson, J. 193–4 118, 120
Johnson, S. 196 Menger, C. 26
Jorgenson, B.S. et al. 55 Meppem, T. and Gill, R. 245, 248
Index 295
Milgrom, P. 239 National Oceanic and Atmospheric
Miller, D. 192, 202 Administration (NOAA) 3, 52, 213
Mitchell, R.T. and Carson, R.C. 33, 213 natural goods 23, 29
Moisseinen, E.M. 53, 66 nature 59
monetary valuation 60; concerns about 4; New South Wales 13; forest region
as essential 1; methods used 1 134–56; parkland study 229–36,
Montreal Process 135 240–1 (5n, 6n)
Mosler, H.-J. 71, 72, 81, 83; et al. 89; and Nijkamp, P. and Voodg, H. 116
Gutscher, H. 83, 93; and Tobias, R. 75 non-use values 54–5, 64–5
Müller, K. et al. 251 Norton, B.G. et al. 263
multi-criteria mapping (MCM) 10–11, 14, novel approach to imprecise assessment
171; application of 172–80 and decision environments (NAIADE)
Multi-Objective-Programming/Goal 9; aggregation of criteria 119–20;
Programming (MOP/GP) 9, 121; development of 118; evaluation of
additions to 123; application 123; alternatives 120; fuzzy numbers
choosing most preferred solution 118–20, 130 (10n, 14n); linguistic
121–2; criticisms of 124; development variables 119, 130 (11n); pairwise
of 124; expectations 122; non- comparisons 118–19; stakeholder
dominated solutions 121; and participation 121
problems of reliability 122–3 Nowak, A. et al. 79; and Latané, B. 79, 81
multiple attribute utility theory (MAUT)
102; based on utility theory 103; O’Connor, M. 46, 188, 244, 248; et al. 4
development of 105; and O’Hara, S. 32, 244, 245, 248, 249
interdependence of attributes 105–6; Olson, D.L. 99, 129
key assumptions 105; preparation for O’Neill, J. 1, 11, 13, 16, 196, 204, 263;
103–4; software 102–3; and and Spash, C.L. 249, 250
stakeholder participation 106; sub- Orbell, J. et al. 192
methodology of 104; summary of 106 O’Riordan, T. and Cameron, J. 167, 171;
multiple attribute value theory (MAVT) 9, and Jordan, A. 167, 171
104–5 Ostrom, E. 13, 245, 246, 247, 248, 249
multiple criteria analysis (MCA) 9–11, 45, outranking methods 112–13; ELECTRE
134, 250, 259; and AHP 140–3; as aid III 113–15; NAIADE 118–21; Regime
to decision-making process 139–40; 115–18
different types of 140; forest study
143–56; in practice 138–9; and WTP parkland study (New South Wales), choice
determinants 60–4 of jurors 232; and consensual decision-
multiple criteria decision analysis making 236–7; deliberative valuation
(MCDA), background 100; 235–6; design/conduct of CJs 229–33;
comparison of methods 124–8; discussion/decision-making processes
concluding remarks 128–9; outranking 232–3; issues raised 236; options
methods 112–21; programming chosen 233–5; options given 229–32;
methods 121–4; quality assessment of and provision of information 238;
methods 100–2; single synthesizing results 233–6; and sample
criteria 102–12; and stakeholder representation 238–9, 241 (10n);
participation 134–56, see also decision scenario for 229; structure of
making deliberative/polling processes 237; and
multiple regression analysis 63–4 within-process (in)equality 237–8
Munda, G. 99, 118, 128–9, 130, 246; et al. participatory approach 4, 5, 11–15, 16;
118 forest resources 134–56; GRANO
Munro, A. and Hanley, N. 209 project 250–60; in MAUT 106; park
management 225, 226–7, 240 (1n);
Nagel, T. 195 stakeholders 5, 16–17, 106, 112, 121,
National Estate Assessment (Australia) 134, 247–8, 250–60; through cost–
136 benefit analysis 226–7
296 Index
Passmore, J. 287 Regime 9, 115–18
Pate, J. and Loomis, J.B. 63 Regional Forest agreements (RFAs) 134,
Payne, J.W. and Bettman, J.R. 210 135, 143–4, 155
Pelletier, D. et al. 285, 286 Regional Forest Forum (Australia) 143–4
Perman, R. et al. 227 regulatory appraisal of risk 10, 159;
Peterken, G.F. and Hughes, F.M.R. 211 applying MCM technique 172–80;
Pettit, P. 205 conventional reponses 160; divergent
Petts, G. 211 approaches 160–1; implementing
Petty, R.E. and Cacioppo, J.T. 87 precautionary approach 170–1;
Peukert, H. 26 incertitude concerning 162–5;
planning cell method (Planungszelle) 228 practical consequences of assessment
policy 16, 44–5 165–7; science/precaution in 167–70,
Prato, T. 140 180
precautionary approach 180, 182; Reise, K. 51
implementation of 170–1; and Renn, O. and Klinke, A. 171
regulatory appraisal of environmental representation 14, 16
risk 167–70 Resource and Economic Assessment
precautionary principle 159, 167 (Australia) 136
PREFCALC 102 resource environment valuation 83–6
preferences 13, 15; and attitudes/values Rietveld, P. 116
266–7; background 263–4; case study rights-based beliefs 4, 27
267–83; and CBA 26; and Riker, W. 195
construction process 266; convergence Rippe, K. and Schaber, P. 205
of 284; and deliberation 263–4, 266, risk 103, 129 (1n), 198; evaluative criteria
284–6, 286 (2n); endogenous 263; and 171; and ignorance 162–3;
hypothetical bias 28; individual 106; implementation of precautionary
and information 265, 286 (1n); approach 170–1; and outcomes 162;
lexicographic 26–7, 43, 44, 45 (7n), 46 practical consequences of assessment
(8n-10n); quality of 263; shift in 165–7; precautionary approach
264–7, 276; and subjectivity 276–83; 167–70, 180, 182; and probability
sustainable 263–4; symbolic 265; use 162–3; science-based approach
of term 264–5 159–61, 164, 167–70, 180, 182; and
Pressy, R. et al. 137 uncertainty 162–5
Pretty, J. et al. 215 Rosenbaum, N. 225
principle of compatibility 54 Rottumeroog island 54
probability 198 Rousseau, J. 225
Proctor, W. 138 Roy, B. 100, 113, 115; and Bouyssou, D.
Productive Capacity of the Forest and 104
Biodiversity 144 Royal Commission on Environmental
programming methods 121–4 Pollution (1998) 16
prominence hypothesis 27 Rydin, Y. 264
prospect theory 3
psychology 3–4, 6–9, 71, 72, 76 Saaty, T.L. 106, 107, 130, 140, 144; and
public goods 29, 30 Alexander, J.M. 106; and Forman, E.
106; and Vargas, L. 141
Queensland 15 Sachs, L. 60
Quiggin, J. 239 Sagoff, M. 27, 30, 191, 203, 209, 210,
227, 228, 244, 245, 263, 266
Raffensberger, C. and Tickner, J. 167, 171 Sahlins, M.D. 244
Randall, A. 65 Saimaa seal 53
rational choice theory 161, 192, 198, Sally, D. 192
205 (7n) Salminen, P. et al. 99, 129
Ready, R. et al. 213 Salo, A. 164
reasoned action 52–3 Sanders, L. 204
Index 297
Scarelli, A. 113 Taha, H.A. 117
Schkade, D.A. and Payne, J.W. 31, 46, Tajfel, H. 76
210, 266 Tamiz, M. et al. 122, 123; and Jones, D. 123
Schulze, W. et al. 28 Tanford, S. and Penrod, S. 77
Schumann, S. 58 tax 1, 18n
Scotland 12, 211–12 Theory of Planned Behaviour 75
Sears, D.O. 265 Theory of Resource Mobilization 75
Sen, A. 192, 202, 205 Thibaut, J.W. and Kelley, H.H. 76
sensitivity analysis 150–2 Thomas-Slatyer, B. 225
Silberman, J. and Gerlowski, D.A. 55 Toman, M.A. 32
single synthesizing criteria 102; AHP trade-offs 4, 26, 27, 43, 201
106–9; Evamix 109–12; MAUT 102–6 travel cost method 4
Slocum, R. and Thomas-Slaytyer, B. 240 trust 13–14, 247, 249
Slovic, P. 263, 266 Turner, B. 137
SMART (Simple Multi-Attribute Rating Tversky, A. et al. 27
Technique) 102, 104, 129 (2n)
Smith, G. and Wales, C. 200, 203, 205 Uckermark (Germany) 14, 252
Smithson, M. 162 uncertainty 4, 162–5, 198
social choice theory 161 utilitarianism 30, 226
social dilemma situations 246–8
social/institutional context 7 valuation theory 1–6, 11–15
socio-economic systems 4 Van Delft, A. and Nijkamp, P. 116
Söderbaum, P. 244 van den Bergh, J. et al. 263
Solow, R.M. 26 Van Valey, T.L. and Petersen, J.C. 225
Southern Forest Region (New South Vatn, A. and Bromley, D.W. 27, 205
Wales) 10; background 134, 157 (1n); Vincke, P. 104, 113, 122, 123, 129
case study 143–4; forest assessments Von Neumann, J. and Morgenstern, O.
135–8; forum preferences 144–6; 105
merits/flaws of study 156; outcomes Von Schomberg, R. 172
152–6; performance of options against Voogd, H. 102, 109, 114
criteria 146–50; problems 156; results
144–50; sensitivity analysis 150–2; Wadden Sea 7, 54; background 51–2;
study method 138–43 discussion/conclusions of survey 64–5;
Spada, H. and Ernst, A.M. 83 influence of attitudes on WTP 57–60;
Spaninks, F.A. et al. 55 multiple analysis of WTP
Spash, C.L. 2, 4, 7, 8, 9, 12, 123; and determinants 60–4; and protest bids
Carter, C. 1; et al. 4, 7; and Hanley, N. 55–6; study design/main results 4–6,
7, 26, 27, 266; and Simpson, I.A. 4 66 (1n-4n)
SPSS-CHAID analysis 61–3 Ward, H. 5, 201, 202, 203, 205, 266
Stahlberg, D. and Frey, D. 87 Water and Soil Resorces 144
Stalder, J. 78 Webler, T. and Renn, O. 225
stated preferences surveys 226–7 welfare theory 13
Steuer, R.E. and Gardiner, L.R. 122 Wet Tropics World Heritage Area
Stevens, T.H. et al. 27, 31, 35, 46 (Australia) 15, 268
Stewart, J. et al. 209 wetlands see Bloomfield Track (Australia)
Stirling, A. 164, 165, 166, 169, 171, 172, Wierstra, E. 51, 57, 60, 64
174; and Mayer, S. 140, 171, 172, 174, Wilderness Assessment (Australia) 136
176 Williams, B. 195
Stork, P. and Viaene, J.-M. 26 willingness to accept (WTA) 3, 7
substitutability 24, 26 willingness to pay (WTP) 2, 6, 7–8, 13, 23,
Sugden, R. 205 24, 31, 65, 213–14, 222–3 (2n), 226,
sustainability 4, 8, 75, 83–6, 263–4 241 (10n, 11n); attitudes as predictors
Sutherland, R.J. and Walsh, R.G. 63 of 52–4; decisions concerning 239,
system theory 46 (13n) 241 (11n); deliberative evaluation
298 Index
process 239; empirical results of survey Wujal Wujal Aboriginal Community 15,
32–43; explaining WTP-bids 32, 35–7, 268
43; influence of attitudes on 57–60; Wynne, B. 162; and Mayer, S. 172
logit model estimated coefficients for
39–42; multiple analysis of Yon, D. and Tendron, G. 212
determinants 60–4; regression model
variables 38; and use values 55 Zionts, S. and Wallenius, J. 122

Alternatives For Environmental Evaluation

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Alternatives For Environmental Evaluation

Загружено:

Авторское право:

Доступные форматы

Alternatives for Environmental

Environmental cost–benefit analysis was developed by economists in the belief

Michael Getzner is Associate Professor of Economics at the University of

Sigrid Stagl is Lecturer and coordinator of the Ecological Economics MA

2 Oil Wealth and the Fate of Tropical Rainforests

3 The Economics of Climate Change

4 Alternatives for Environmental Valuation

ISBN 0-203-41287-7 Master e-book ISBN

ISBN 0-203-34079-5 (Adobe eReader Format)

List of figures vii

1 Exploring alternatives for environmental valuation 1

2 A framework for valuing nature: regional biodiversity 23

3 Non-use values and attitudes: wetlands threatened by climate

4 Modelling environmental behaviour: socio-psychological

5 Assessing the quality of different MCDA methods 99

7 Confronting risk with precaution: a multi-criteria mapping of

8 Consumer valuation and citizen deliberation: towards a

9 Three approaches to valuing nature: forest floodplain restoration 209

10 Deliberation and economic valuation: national park management 225

11 The challenges of stakeholder participation: agri-environmental

12 Preference transformation through deliberation: protecting world

2.1 The ‘standard’ and ‘extended’ model of valuation 25

2.1 Consumer vs. citizen and students’ perspectives 36

Jonathan Aldred, Department of Land Economy, Cambridge University, UK,

AEF agri-environmental forum

1 Exploring alternatives for

Contributions to the debate in this book

Economic values and the psychology of behaviour

Taking multiple values into account

Participation, deliberation and valuation

2 A framework for valuing

The extended context of valuation: some important issues

Lexicographic preferences and the willingness to exchange

The ‘standard ’ model of explaining WTP bids

Elements of the institutional Elements of the social

Figure 2.1 The ‘standard’ and ‘extended’ model of valuation

While there might be different approaches to the substitutability issue in natural

obtain knowledge about a system by changing its inputs and outputs of

The social context: ‘consumer’ vs. ‘citizen’

Explaining WTP-bids: incorporating the social context

Pr (WTPi > 0 ) = f (Si , Pi , I i , Ci ) and (3.1a)

Descriptive overview of the empirical results

Explaining WTP bids: the importance of the ‘extended’

amount). Explaining the respondent’s yes-vote is done stepwise. Four different

Table 2.3 WTP regression model variables

Variable name Dimension Description

Model 1 Model 2 Model 3 Model 4

Influence –1.3761 –2.0176** –1.6679 –2.3378** –2.3451 –2.8739***

Model 1 Model 2 Model 3 Model 4

Thoughts 0.7090 1.3168 0.2858 0.4951 0.6531 0.9298

Model 1 Model 2 Model 3 Model 4

Influence –0.6708 –1.2698 0.2236 0.5317 0.2430 0.5889

Model 1 Model 2 Model 3 Model 4

0.4076 0.3232 0.3428

Discussion and conclusions

3 Non-use values and

Attitudes as predictors of WTP within the CVM

Study design and main results

Table 3.1 Distribution of interviews and in-principle WTP

WTP Yes No Yes No

Attitude towards nature and climate change

• Household net income

Figure 3.1 A conceptual model of WTP determinants

Influence –1.3761 –2.0176 –1.6679 –2.3378 –2.3451 –2.8739***