IJASample

International Journal of
ADVERTISING
The Quarterly Review of Marketing Communications
Volume 32 Number 1 2013
Contents
Editorial: Hot topics in advertising research 7
Charles R. Taylor
Putting brands into play: how game difficulty and player experiences 17
influence the effectiveness of in-game advertising
Laura Herrewijn and Karolien Poels
Constitutive marketing: towards understanding brand community formation 45

Ana Kilambi, Michel Laroche and Marie-Odile Richard
The differential effects of position, ad and reader characteristics on readers’ 65

processing of newspaper ads
Edith G. Smit, Peter C. Neijens and Robert Heath
Rational and irrational influences on lucky draw participation 85

Gerard Prendergast and Edmund R. Thompson
Communication using celebrities in the non-profit sector: determinants of 101

its effectiveness
Maria del Mar Garcia de los Salmones, Rafael Dominguez and Angel Herrero
How banner ads can be effective: investigating the influences of exposure 121
duration and banner ad complexity
Kai-Yu Wang, Eric Shih and Laura A. Peracchio
Advertising and promotion budgeting during volatile economic conditions: 143

factors influencing the level of decentralisation in budgeting and its relations
to budget size and allocation
Yunjae Cheong, Kihan Kim and Hyuksoo Kim
The Advertising Association is not responsible for the opinions and

data contained in the editorial or articles in this journal
© Advertising Association 2013
Book Reviews 163
Hamish Pringle and Jim Marshall – Spending Advertising Money in

the Digital Age – How to Navigate the Media Flow
James Harris
Timothy De Waal Malefyt and Robert J. Morais – Advertising and

Anthropology: Ethnographic Practice and Cultural Perspectives
Stephanie O'Donohoe
Ex Libris – Winston Fletcher

Editorial Advisory Board
Tim Ambler Yung Kyun Choi
London Business School Dongguk University
Saravudh Anantachart George Christodoulides
Chulalongkorn University University of Birmingham
Bradley Barnes Shu-Chuan (Kelly) Chu
University of Kent DePaul University
Mickey Belch Jaihak Chung
San Diego State University Sogang University
Steve Bellman Bettina Cornwell
Murdoch University University of Michigan
Richard Beltramini Samuel Craig
Wayne State University New York University
Edoardo Brioschi Stephan Dahl
Università Cattolica University of Hull
Steven Burgess Micael Dahlén
University of Cape Town, South Africa Stockholm School of Economics
John Cadogan Nathalie Dens
Loughborough University University of Antwerp
Michael Capella Sandra Diehl
Villanova University Alpen-Adria University of Klagenfurt
Les Carlson Steve Edwards
University of Nebraska-Lincoln Southern Methodist University
Albert Caruana Richard Elliott
University of Malta University of Bath
Paulo Catalfomo Mike Ewing
University of Turin Monash University
Chingching Chang Ron Faber
National Chengchi University University of Minnesota
Dae Ryun Chang Kim-Shyan Fam
Yonsei University Victoria University of Wellington
Sejung Marina Choi Paul Farris
University of Texas at Austin University of Virginia
3
John B. Ford Kihan Kim
Old Dominion University Seoul National University
Rossella Gambetti Philip Kitchen
Università Cattolica del Sacro Cuore ESC Rennes School of Business
Maggie Geuens Eunju Ko
Ghent University Yonsei University
Stacy Grau Rick Kolbe
Texas Christian University Kent State University
Louisa Ha Scott Koslow
Bowling Green State University University of Waikato
Chris Hackley Dean Krugman
Royal Holloway University of London University of Georgia
Geraldine Henderson Hyokjin Kwak
University of Texas–Austin Drexel University
Janet Hoek Russ Laczniak
University of Otago Iowa State University
Jisu Huh Tobias Langner
University of Minnesota Bergische University Wuppertal
Kineta Hung Judie Lannon
University of Hong Kong Market Leader
Wim Janssens Doo Hee Lee
Hasselt University Korea University
John Philip Jones Wei-Na Lee
Syracuse University University of Texas
Michael Kamins Hairong Li
University of Southern California Michigan State University
Jikyeong Kang May Lwin
The University of Manchester Nanyang Technological University
Kiran Karande Edward Malthouse
Old Dominion University Northwestern University
Kevin Keenan Marlys Mason
American University in Cairo Oklahoma State University
Jeremy Kees Michael Maynard
Villanova University Temple University
Rachel Kennedy Rod McColl
Ehrenberg-Bass Institute, Uni SA ESC Rennes School of Business
4
Colin McDonald Herbert Rotfeld
McDonald Research Auburn University
Robert Meeds Cristel A. Russell
Qatar University American University
Paul Michell Julie Ruth
Leeds University Business School Rutgers School of Business–Camden
Mariko Morimoto Sela Sar
University of Georgia Iowa State University
Agnes Nairn Marko Sarstedt
E.M. Lyon Business School Otto-von-Guericke University Magdeburg
Peter Neijens Sheila Sasser
University of Amsterdam Eastern Michigan University
Shintaro Okazaki Don Schultz
Autonomous University of Madrid Northwestern University
Hye-Jin Paek Manfred Schwaiger
Michigan State University Ludwig Maximillians University
Stan Paliwoda John Scriven
University of Strathclyde South Bank University
George Panigyrakis Trina Sego
University of Athens Boise State University
Patrick De Pelsmacker Kim Sheehan
University of Antwerp University of Oregon
Joseph Phelps Akira Shimizu
University of Alabama Keio University
Barbara Phillips Janas Sinclair
University of Saskatchewan University of North Carolina
Sascha Raithel Edith Smit
Ludwig Maximilians University University of Amsterdam
Leonard Reid Dongyoung Sohn
University of Georgia Hanyang University
José I. Rojas-Méndez David Stewart
Carleton University University of California, Riverside
Sara Rosengren Yongjun Sung
Stockholm School of Economics University of Texas–Austin
John Rossiter Max Sutherland
University of Wollongong Bond University
5
Ralf Terlutter Gary Wilcox
University of Klagenfurt University of Texas at Austin
Brian Till R. Dale Wilson

Loyola University Michigan State University
Joyce Wolburg
Sunny Tsai
Marquette University
University of Miami
Chan Yoo
Eva A. van Reijmersdal University of Kentucky
Amsterdam School of Communication
Research ASCoR Seounmi Youn
Emerson College
Martin Waiguny
Brian Young
Auckland University of Technology
University of Exeter
Mike Waterson Yorgos Zotos
World Economics Cyprus University of Technology
Douglas West Peng Zou
King's College London Harbin Institute of Technology
6
Editorial
Hot topics in advertising research
Charles R. Taylor
Villanova University
Introduction
Among the questions I am most frequently asked when at academic or practitioner
conferences is ‘What are the current hot topics in advertising research?’ While answer-
ing this question involves some level of subjectivity, there are, nevertheless, some topic
areas for which many would agree there is a pressing need for more research. In this
editorial, I will outline some areas on which there is a clear need for more research and,
indeed, would like to see more research on over the coming years. One of the big chal-
lenges in determining what to include on a list of ‘hot’ topics is to make a judgement on
whether a topic that has been of importance for some time should remain on the list.
For example, measurement of return on investment has been a big focus of discussion of
key issues facing advertisers for the better part of the past decade (see e.g. Bruno et al.
2005; Ambler & Roberts 2008; Taylor 2010). However, while improved methods of
measuring ROI remain an important subject it is probably no longer accurate to refer
to is as being a hot topic.
In 2013 we are not seeing the immediate implementation of new technologies driving
things as much as we are seeing recent new technologies that have been applied to adver-
tising and promotion more fully playing themselves out. As the internet really did not
begin to become a well-established advertising medium until about 2005 (Taylor 2009), it
is not so surprising that what is really interesting eight years later is more how advertising
on the internet and in social media is playing out. In addition, examining how companies
integrate traditional advertising with digital advertising is of interest. The impact that
digital media has had on traditional media can also be viewed as a key topic and, again, is
based on some of the longer term impact of new technologies.
It is also the case that some topics that have been around for some time have gotten
to the point where research on them can be characterized as being in the mature stage of
the life cycle overall, but niche areas within the overall topic are badly in need of current
research. Below, I will list out eight topics that can be characterized as ‘hot’ – the first five
having to do with issues that emanate from how digital media has evolved (and impacted
traditional media) and the remaining four involving specific topics that remain hot within
a research area that has been around for some time.
International Journal of Advertising, 32(1), pp. 7–12

© 2013 Advertising Association
Published by Warc, www.warc.com 7
DOI: 10.2501/IJA-32-1-007-012
International Journal of Advertising, 2013, 32(1)
The Hot Topics
How to get people to be attentive to digital ads

Whether in traditional internet advertising, in social media, or via SMS, it has long been
viewed as a challenge to get people to pay attention to an ad in the digital environment.
Watching YouTube’s attempts at getting people to view ads, which include allowing view-
ers to watch a limited portion of an ad prior to viewing content, or watching a full ad
for access to particularly sought after content, is fascinating to any student of advertising.
It appears to remain a challenge to get people to view ads in the ‘private’ online context
and more research on how that can be done is needed. Bellman et al.’s (2012) article on
commercial loading in online video programs, and Southgate et al.’s (2010) examination
of determinants of viral video viewing, and Yu et al.’s (2012) study of remaining duration
prompts are excellent examples of the type of research that needs to be done, but more
is needed.
Enhancing attentiveness to television commercials in the digital video

recorder era
While research has documented that digital video recorders have not had the poten-
tially disastrous effects that some had predicted (see Brasel & Gips 2008; Bronnenberg
et al. 2010) it is nonetheless timely to monitor the degree to which viewers are attentive
to television ads. This involves both research such as that by Brasel and Gips (2008)
on whether and how an ad watched in fast forward mode can still be effective as well
as research on how programming context effects affect ad viewing. It is notable that in
the US many ‘on demand’ systems offering free programming now limit the speed at
which content can be fast forwarded. Clearly, advertisers, television networks, and cable/
satellite television operators are highly interested in this issue. In terms of programme
context effects there have been some good recent studies (e.g. Cauberghe et al. 2011;
Moorman et al. 2012).
Advergames
Perhaps the least surprising entry on this list of hot topics is advergames. Ha and McCann
(2008) provided an excellent model of clutter in both online and traditional media and
emphasized the challenge in getting an ad to stand out in the modern advertising envi-
ronment. Studies by Fuordigital (2009) and others have now documented that under the
right circumstances, game players actually like to see advertising in games and sometimes
even believe that it enhances the experience associated with the game. This advantage of
advergames is of key importance, and research on what makes an ad ‘fit’ a game and how
to maximize its effectiveness are clearly important topics. While much good research has
already been conducted on this topic (e.g. Cauberghe & DePelsmacker 2010; Vedrashko
2011), more is needed.
8
Editorial: Hot topics in advertising research
New approaches in social media

As social media outlets such as Facebook, Twitter and YouTube have achieved explosive
growth, there has been much enthusiasm about their financial viability. However, recently
much focus has turned to the real ability of such sites to bring in advertising revenue
and how such sites can be used. It is, indeed, interesting to read debates on issues such as
whether and to what degree advertising and other promotional techniques on Facebook
are effective. Paek et al.’s article on YouTube (2011), Muntinga et al.’s (2011) study of
consumers’ online brand-related activities (COBRAs), and Steyn et al. (2011) on source
effects in consumer-generated advertising are good examples of the type of research
needed. More specific studies of the effectiveness of various approaches (e.g. Facebook’s
use of consumer ‘likes’) are needed.
Customized communications and related privacy issues

It has been clear for some time that widespread use of the internet by the public leads to
a much greater ability to target consumers with customized messages based on various
preferences they have or exhibit (e.g. where they have travelled, what their favourite sports
team is, what they have purchased online, what websites they visit, what political party
they belong to, what charities they contribute to, etc.). Whether on the internet or in trad-
itional media, there has been a clear trend toward more ‘fine-grained’ targeting based on
obtaining personal information from consumers. While there is a clear and obvious upside
to this type of targeting, there are also privacy issues that raise concerns among many
consumers. Baek and Morimoto (2012) studied personalized advertising and found that
a message perceived to be more personalized is associated with less avoidance. This is an
interesting finding, the type of which should be built upon by subsequent issues. Within
the next few issues, International Journal of Advertising will publish a special section on
this issue so please stay tuned.
Measuring the impact of sponsorships and product placement

Both sponsorships and product placement have been growing worldwide, most likely
in response to the more cluttered advertising environment. While the need for better
measures of the impact of sponsorships on advertisers has been much discussed, there is
a pressing need for research on this topic as well as on product placement. While some
studies have arisen and look at a broader array of effects of sponsorship (e.g. Schwaiger
et al. 2010; Zdravkovic & Till 2012; Ruth & Strizhakova 2012) it is simply the case that
more are needed.
Advertising, promotion, nutrition and obesity

As always, public policy-related issues pertaining to advertising and promotion are of
interest to this journal. While there are several issues of interest, including privacy (see
9
above), advertising to children (subject of a forthcoming special issue of IJA in 2014)

and many others, the general issue of promotion of food and nutrition continues to be
in need of more research. The weight of the research evidence pertaining to advertising
appears to be against the notion that food advertising had a direct impact on consump-
tion choices at the product category level, let alone individuals’ dietary choices (Ambler
2006; Rotfeld & Taylor 2009). Nevertheless, research on the impact of nutritional
labelling programmes (see Hieke & Taylor 2012 for a review) and health promotion
campaigns is badly needed, as are deeper insights into how food advertising impacts
consumers.
Advertising trends in emerging markets

As I wrote in an Editorial in the second issue of 2012 on the need for more research on
the BRICs and other emerging markets (Taylor 2012) I won’t rehash all of the ration-
ale behind this assertion here other than to state what is relatively obvious. The world
is changing and the continued growth of global commerce and the advertising industry
in emerging markets will increasingly change the ways that global marketers do busi-
ness. Going forward it is very likely that emerging markets will be more and more a
part of the centre of strategic planning as opposed to something considered after mak-
ing plans for long developed markets. The advertising landscape in the BRICs, as well
as growing markets in other parts of Asia, Eastern Europe, Latin America and Africa
is likely to matter more and more to those who study advertising across the world in
the future.
Acknowledgement
In ending this Editorial I would like to acknowledge the contribution of Matthew
Coombs who has served as Publisher of the International Journal of Advertising during
my entire tenure as Editor and, indeed, well before. I recently heard that Matthew will be
leaving Warc. Matthew has been instrumental in making decisions that I believe have
moved the journal forward and has been a real pleasure to work with. On behalf of the
Associate Editors, Editorial Review Board Members and IJA staff, I would like to wish
him the best of luck in future endeavours and thank him for all of his contributions to
the journal.
References
Ambler, T. (2006) Does the UK promotion of food and drink to children contribute to their
obesity? International Journal of Advertising, 25 (2), 137–156.
Ambler, T. & Roberts, J.H. (2008) Assessing marketing performance: don’t settle for a silver
metric. Journal of Marketing Management, 24 (7,8), 733–750.
Baek, T. & Morimoto, M. (2012) Stay away from me: examining the determinants of
personalized advertising. Journal of Advertising, 41(1), 59–76.
10
Editorial: Hot topics in advertising research
Bellman, S., Treleaven-Hassard, S., Robinaon, J.A., Rask, A. & Vran, D. (2012) Getting the
balance right. Commercial loading in online video programs. Journal of Advertising, 41(2),
5–24.
Brasel, S.A. & Gips, J. (2008) Breaking through fast-forwarding: brand information and visual
attention. Journal of Marketing, 72(4), 31–48.
Bronnenberg, B.J., Dubé, J. & Mela, C.F. (2010) Do digital video recorders influence sales?
Journal of Marketing Research, 47(6), 998–1010.
Bruno, H.A., Parthasarathi, U. & Singh, N. (2005) Does marketing measure up? Performance
metrics: practices and impacts. Marketing Science Institute, Conference Summary Report
No. 05-301.
Cauberghe, V. & De Pelsmacker, P. (2010) Advergames: the impact of brand prominence and
game recognition on brand responses. Journal of Advertising, 39(1), 5–18.
Cauberghe, V., Geuens, M. & De Pelsmacker, P. (2011) Context effects of TV programme-
induced interactivity and telepresence on advertising responses. International Journal of
Advertising, 30(4), 641–663.
Fuordigital (2009) Video game advertising. A brave new world. Available from: http://www.
slideshare.net/fuordigital/the-video-game-advertising-landscape (accessed 8 January 2013).
Ha, L. & McCann, K. (2008) An integrated model of advertising clutter in offline and online
media. International Journal of Advertising, 27(4), 569–592.
Hieke, S. & Taylor, C.R. (2012) A critical review of the literature on nutritional labeling:
conceptual and methodological directions for future research. Journal of Consumer Affairs,
46(1), 120–156.
Moorman, M., Willemsen, L.M, Neijens, P.C. & Smit, E.G. (2012) Program-involvement
effects of commercial attention and recall of successive and embedded advertising. Journal of
Advertising, 41(2), 25–37.
Muntinga, D.G., Moorman, M. & Smit, E.G. (2011) Introducing COBRAs: exploring
motivations for brand-related social media use. International Journal of Advertising, 30(1),
13–46.
Paek, H., Hove, T., Jeong, H. & Kim, M. (2011) Peer or expert? The persuasive impact of
YouTube video producers and their moderating mechanism. International Journal of
Advertising, 30(1), 161–188.
Rotfeld, H.J. & Taylor, C.R. (2009) The advertising regulation and self-regulation issues ripped
from the headlines with (sometimes missed) opportunities for disciplined multi-disciplinary
research. Journal of Advertising, 38(4), 5–14.
Ruth, J.A. & Strizhakova, Y. (2012) And now, goodbye: consumer response to sponsor exit.
International Journal of Advertising, 31(1), 39–62.
Schwaiger, M., Sarstedt, M. & Taylor, C.R. (2010) Art for the sake of the corporation: Audi,
BMW Group, DaimlerChrysler, Montblanc, Siemens, and Volkswagen help explore the
effect of sponsorship on corporate reputations. Journal of Advertising Research, 50(1), 77–91.
Southgate, D., Westoby, N. & Page, G. (2010) Creative determinants of viral video viewing.
Steyn, P., Ewing, M.T., van Heerden, G., Pitt, L.F. & Windisch, L. (2011) From whence it
came: understanding source effects in consumer-generated advertising. International Journal
of Advertising, 30(1), 133–160.
Taylor, C.R. (2009) The six principles of digital advertising. International Journal of Advertising,
28(3), 411–418.
Taylor, C.R. (2010) Measuring return on investment from advertising: ‘holy grail’ or necessary
tool? International Journal of Advertising, 29(3), 345–348.
Taylor, C.R. (2012) Editorial: On advertising in the BRICs and other emerging markets.
11
Vedrashko, I. (2011) Game-based marketing: inspire customer loyalty through rewards,

challenges, and contests. International Journal of Advertising, 30(1), 189–190.
Yu, R., Chan, A.H.S., Zhao, P. & Gao, G. (2012) Enhancing memorability: do remaining
duration prompts affect advertising recall? International Journal of Advertising, 31(4),
861–876.
Zdravkovic, S. & Till, B.D. (2012) Enhancing brand image via sponsorship: strength of
association effects. International Journal of Advertising, 31(1), 113–132.
12
Managerial Summaries
Putting brands into play: how game difficulty and player 17
experiences influence the effectiveness of in-game advertising
• The study examines the impact of an individual’s psychological responses
in reaction to the content of a digital game (i.e. player experiences) on the
effectiveness of in-game advertising (IGA).
• A between-subjects experimental design was used, in which player experiences are
varied between three conditions by manipulating the difficulty (easy – medium –
hard) of the experimental game that contained several in-game advertisements.
• Results indicate that manipulating game difficulty affects the processing and
evaluation of the brands featured in in-game ad placements both directly and
indirectly (through the variation of player experiences).
• The findings of the study support the notion that the emotions and experiences
that players encounter while playing games are relevant context characteristics that
have to be taken into account while studying and integrating advertising inside
digital games.
Constitutive marketing: towards understanding brand 45

community formation
Ana Kilambi, Michel Laroche and Marie-Odile Richard
• This study helps understand the concept of interactive consumer, and the
transformation of the actual consumer to the imagined consumer of the brand
community through the operation of advertising.
• Our model of brand community formation outlines the notion that brand
communities are established through advertising persuasion and the rhetorical
action of branding, and the interplay among the brand, its message and its
consumers.
• Our model frames advertising efforts in a new light, demonstrating how brand
communities are constituted in and through advertising texts.
• Understanding the antecedents of brand community, marketers can direct, grow
and maintain their brand communities through subtle relationship marketing
initiatives, and in doing so, strengthen their brand equity.
• While marketers must always tread lightly when choreographing their efforts in
advertising, balance brand objectives with the consumer climate, this study offers
insights as to how brands can work to build and shape the foundations of a brand
community.
• It is thus imperative to include the brand community in the discussion of brand
equity and the brand valuation process.
13
managerial summaries
The differential effects of position, ad and reader characteristics 65

on readers’ processing of newspaper ads
Edith G. Smit, Peter C. Neijens and Robert Heath
• It is shown that with increasing levels of ad processing, influencing variables
moved from position in the newspaper, via mechanical ad characteristics, to
message content and reader interest characteristics.
• Ad position is most important for exposure to newspaper ads, while mechanical ad
characteristics determine ad noticing and brand awareness.
• Relevance and execution of ad content are most important for stimulating higher
processing levels.
• Support for earlier studies on a differential influence between mechanical and
message variables and between stimulus-specific characteristics and person-specific
factors.
Rational and irrational influences on lucky draw participation 85

Gerard Prendergast and Edmund R. Thompson
• The relative roles of consumer rationality and irrationality in chance-related sales
inducements is unknown.
• Previous research indicates irrational belief in luck influences opting for a lucky
draw instead of other promotional offers such as discounts.
• The extent to which lucky draw participation might also be influenced by rational
consideration of winning odds has been uninvestigated.
• Findings suggest both an irrational belief in luck and a rational appraisal of the
probability of winning significantly influence the choice of participating in a lucky
draw over other promotions.
Communication using celebrities in the non-profit sector: 101

determinants of its effectiveness
Maria del Mar Garcia de los Salmones, Rafael Dominguez and Angel Herrero
• Celebrity endorsement is a popular marketing technique in the non-profit sector,
but there is still a lack of studies about the effectiveness of this communication
strategy in this area.
• This study presents a causal model which analyses the sequence of relationships
from the antecedents of celebrity credibility to the determinants of the attitude
towards ads and the intention to collaborate with a social entity.
• Results indicate that celebrity credibility is directly explained by the fit perceived
and the attribution of altruistic motivation, and indirectly by the celebrity image
and the general attitude towards celebrity activism.
• The main antecedent of behavioural intentions is the social entity credibility,
but celebrities can have some influence as well through their impact on attitudes
towards the ad.
14
managerial summaries
How banner ads can be effective: investigating the influences 121

of exposure duration and banner ad complexity
Kai-Yu Wang, Eric Shih and Laura A. Peracchio
• This research examines the impacts of exposure duration and banner ad
complexity on advertising persuasion in a web advertising environment.
• Consistent with previous research on banner advertising, the present study shows
that banner advertising generates desirable effects for advertising and brand
judgements.
• Although increasing the duration of exposure enhances processing fluency under
some conditions, our findings suggest that exposure duration and ease of banner
ad processing jointly influenced attitudes towards the ad and brand judgements
among the participants in this study.
• The findings support the prediction that the priming banner ads would facilitate
the perceptual or conceptual processing of subsequently presented target ads, and
thereby lead to more favourable attitudes among the viewers.
• This research suggests that, in order to generate favourable effects (e.g. positive ad
and brand attitudes among consumers), advertisers should consider the complexity
of an ad when designing banner ad content.
Advertising and promotion budgeting during volatile economic 143

conditions: factors influencing the level of decentralisation in
budgeting and its relations to budget size and allocation
Yunjae Cheong, Kihan Kim and Hyuksoo Kim
• This paper provides a perspective on traditional budgeting approaches and
managerial processes in the midst of the severe economic downturn in the US.
• Specifically, this paper examines the extent to which various company, brand and
organizational factors are related to the level of decentralization in the budgeting
– the degree to which the power to influence budget-setting is shared with lower-
level marketing department personnel.
• The evidence presented in this paper indicates that companies engaged in
more decentralized budgeting processes are more likely to perform better in the
market, as indicated by the size of the company, and its profitability and brand
equity.
• In addition, the level of budgeting decentralization was positively related to the
total marketing budget size and more allocation of the marketing budget to
advertising in relation to promotion.
15
Your essential source of
ideas and inspiration.
Warc’s lively monthly magazine written by
expert practitioners around the globe.
Admap is an invaluable
source of insight and
May 2011
Ide as and
evi de nc e
wisdom. It tells me what

for ma rke
tin g pe op le
ww w.w arc
.co m/ ad ma
p May 20
11
is important and what

I need to know to keep
Shopper
Marketing
up and move forward.

It is practical and
inspirational, a must Influenc
the shop e
per
read. journey
Vanella Jackson, Worldwide CEO,

• Best Practic
e: Shopper
• Path To Pur Marketing
chase • Sports Spo
• Activation nsorship
Response • Organic
Measurem Planning and • 10 Global
ent • McCain’s Buying Consumer
eCRM Stra • Outdoor Trends
Hall & Partners

tegy Ad Effectiv
• Best Practic eness
ADM May
01 COVE e: Co-creatio
R.indd 1 n
4/19/2011
16:40:52
Subscribe at www.warc.com/admap
Follow us on twitter
@colin_admap
Admap_170x240mm_Advert_HiResPrintReady_GREY.indd 1 01.08.11 13:47

Putting brands into play
How game difficulty and player experiences
influence the effectiveness of in-game advertising
University of Antwerp
The study looks at the effectiveness of in-game advertising, focusing specifically on the impact
of medium context. Using an experimental design, the study investigates the relationship that
exists between the effectiveness of in-game ad placements and a player’s psychological experi-
ences in reaction to the content of the game. It is argued that the numerous emotions and
experiences that occur during game play (e.g. enjoyment, competence, tension) might have
an impact on how players process the game environment, including in-game ads. In order to
test this, a between-subjects online experiment was conducted in which participants played a
casual PC game that contained several in-game advertisements. We manipulated the difficulty
of the game, leading to differences in several player experiences between experimental condi-
tions. Results indicate that increasing game difficulty affects the processing and evaluation of
the brands featured in in-game ad placements both directly and, more importantly, indirectly
through the variation of player experiences.
Introduction
Digital gaming has become one of the largest entertainment sectors worldwide. Millions
of people from all socio-demographic groups are spending more and more of their leisure
time playing digital games. As a result, the advertising industry has taken an interest in
this medium, believing it may be a suitable vehicle for integrating advertisements and
reaching their target markets (Schneider & Cornwell 2005; Mau et al. 2008; Mackay et al.
2009; Nicovich 2010; Brightman 2011). This belief is further amplified by the fact that
integrating advertisements into digital games can provide brands with an opportunity to
become incorporated in the digital game experience, and engage with players in a fun,
vivid, interactive and immersive entertainment environment (Nelson 2005; Schneider &
Cornwell 2005; Mackay et al. 2009; Brightman 2011). Moreover, a lot of digital games are
now played online, enabling advertisers to dynamically place and vary ads in games based
on multiple criteria such as players’ demographic, regional and gamer profile, time of the
day, etc. (Schneider & Cornwell 2005; Bardzell et al. 2008).

DOI: 10.2501/IJA-32-1-017-044
Consequently, the use of digital games as a medium for the delivery of advertisements,
or in-game advertising (IGA), is predicted to grow steadily over the next several years,
eventually reaching $7.2 billion globally in 2016 (Brightman 2011). However, taking into
account the amount of time that is spent playing digital games, IGA revenues are still
relatively low compared to advertising revenues spent on other media (such as television
and the internet) (Brightman 2011). This implies that, as an advertising vehicle, digital
games are still under-utilised and not fully explored (Brightman 2011).
The aim of the current study is therefore to further analyse the practice of in-game
advertising, and shed more light on its effectiveness. Although academic research has
increasingly turned its attention to the use of digital games as an advertising medium in
recent years (e.g. Nelson 2005; Schneider & Cornwell 2005; Mau et al. 2008; Mackay
et al. 2009), little is known about the impact of the context in which the in-game ads
are encountered. This is surprising, since there is considerable evidence that advertising
effectiveness in traditional media (e.g. print, radio and television) is influenced by the
advertising context (Moorman 2003).
In literature, a distinction is made between context characteristics related to the receiver
of the advertisement and those related to the medium carrying the ad (Moorman 2003).
The receiver context generally describes the situational circumstances a person finds him-
self in while being exposed to an advertisement. This includes factors such as the person’s
physical (e.g. at home or elsewhere) and social (e.g. alone or with others) environment
(Moorman 2003; Bronner et al. 2007; van Reijmersdal et al. 2010). The medium context,
on the other hand, concerns the environment of the ad provided by the medium (e.g. a
television programme, an internet site). This can include factors such as genre, the amount
of congruence that exists between context and advertising, and individual psychological
responses in reaction to medium content (e.g. suspense, enjoyment) (Moorman 2003; van
Reijmersdal et al. 2010). These individual psychological responses in reaction to the con-
tent of a medium (in this case, digital games) will be the focus of the current study. More
specifically, the goal is to get an insight into the way player experiences, or the emotional
experiences while playing digital games, influence player responses to in-game advertising.
Digital games have the potential to evoke a wide array of general (e.g. pleasure, arousal)
and more specific (e.g. frustration, competence) emotional experiences (Ermi & Mäyrä
2005; Poels et al. 2009). Prior advertising research studying traditional media has already
showed that both the general (e.g. Pavelchak et al. 1988) and specific (e.g. Griskevicius et al.
2009) emotions people experience while watching an advertisement affect the way they
respond to the ad. Consequently, these emotions might also exert an influence on the way
people process elements in a digital game environment, including in-game advertisements.
In order to test this hypothesis, the current study employs a between-subjects experi-
mental design. Participants were asked to play an online PC game that contained several
background advertisements (billboards). Because previous studies indicated that player
experiences can change whenever the difficulty of the game is altered (Sykes & Brown
2003; van den Hoogen et al. 2008), we manipulated the difficulty of the game, resulting
in three experimental conditions: people had to play either an easy, medium or hard level.
After playing the online game, we assessed participants’ processing (i.e. recognition) and
18
evaluation (i.e. likeability) of the in-game ad placements. By examining the impact of

player experiences on the effectiveness of IGA, the study attempts to fill the gap in the
literature on player experiences as context characteristics in in-game advertising.
Player experiences as context characteristics

It is already well documented in advertising research studying traditional media that con-
text characteristics, like the emotional experiences people encounter while being exposed
to an advertisement, have an impact on the way they process the ad (Moorman 2003).
Research on the effects of context characteristics on advertising in digital games, however,
is limited. This is surprising, as gaming is an active experience where a person is drawn in
to a virtual world, and potentially confronted with numerous emotions and experiences.
Players have to concentrate to be able to successfully conclude certain tasks, and face feel-
ings of control, frustration, relief, euphoria, flow, etc., in the process (Sherry 2004; Ermi &
Mäyrä 2005; Klimmt et al. 2007; Poels et al. 2009).
Traditional advertising studies already showed that the arousal level and valence of the
emotions people feel while encountering an advertisement affect the way they process the
ad. For example, it has been documented that people have poorer recall and recognition of
commercial content in highly involving and arousing contexts (Norris & Colman 1992;
Coulter & Sewall 1995; Coulter 1998). Moreover, when compared to people who are in a
negative state of mind, people who are in a positive state of mind are proven to have a more
positive brand attitude, a greater intention to try advertised products (Owolabi 2009), and a
better recall of message stimuli (Gardner 1985) and brand names (Lee & Sternthal 1999).
Bearing this in mind, it is not unlikely that the state of mind people are in, and the emo-
tions and experiences they encounter while playing a digital game, in a similar way, affect
the way they observe and process the advertisements that are integrated into the game
environment. We therefore argue that player experiences are important context character-
istics that have to be taken into account when studying digital games and their effectiveness
as an advertising medium. First, we will give an overview of the general and more specific
emotional experiences people are potentially confronted with while playing digital games.
Subsequently, we will describe in greater detail our manipulation of game difficulty and the
effects we expect this manipulation to have on the experiences of the players.
Specific player experiences

Playing digital games has the potential to evoke a wide array of specific emotional experi-
ences, like challenge, tension, competence, boredom, flow and immersion (Poels et al. 2009;
Jennett et al. 2008; Gilleade & Dix 2004). In the following sections, we will describe these
experiences in greater detail, and elaborate on the consequences they might have on a
cognitive (e.g. memory, attention) and affective (e.g. feelings, emotions) level.
First of all, digital games generally provide players with an activity that is goal-directed.
Moreover, every game puts players to a certain challenge, which in combination with the
skills of the player will determine whether the player is able to reach this goal (Poels et al.
19
2009). When players fulfil a goal or master a specific skill, they will perceive a certain
amount of competence: they will feel successful and skilful, which often evokes positive
emotions like pride and euphoria (Klimmt et al. 2007; Poels et al. 2009; Trepte & Reinecke
2011). On the other hand, however, failing to reach these goals often leads to feelings of
tension, like frustration, irritation, and even anger (Gilleade & Dix 2004; Klimmt et al.
2007; Poels et al. 2009).
Another prominent player experience that is highly related to the challenge a game
imposes on a player, and the abilities he or she possesses to overcome this challenge, is
flow. Csikszentmihalyi (1990) originally used this concept to describe extremely positive
‘peak’ experiences in daily life. He defined flow as an ‘autotelic’, or self-motivating and
optimal experience, ‘a state in which individuals are so involved in an activity that nothing
else seems to matter’ (Csikszentmihalyi 1990; Sherry 2004). The state is realised when
a perfect balance between skills and challenges can be reached. Activities that are most
likely to create the flow state are those that have concrete goals with manageable rules,
make it possible to adjust opportunities for action according to skill level or capabilities,
provide immediate and clear feedback on score and progress made, and screen out dis-
traction and make concentration possible (Csikszentmihalyi 1990; Sherry 2004; Sweetser
& Wyeth 2005). Playing digital games has many of these aspects, making flow a highly
relevant concept when deliberating about player experiences (Sherry 2004).
An important remark to make in this respect is that such an optimal state of flow is rare
in everyday life – the exception rather than the rule. Mismatches between skills and chal-
lenges often prevent the flow state from being experienced (Sherry 2004). As noted above,
if the challenges imposed by a digital game outpace the skills the gamer possesses, feelings
of tension (e.g. frustration) will most likely be experienced (Gilleade & Dix 2004; Sherry
2004; Klimmt et al. 2007; Poels et al. 2009). On the other hand, if a gamer’s skills are too
developed in comparison with the challenges imposed by a digital game, the experience
of boredom will presumably occur (Sherry 2004; Poels et al. 2009). Although experiences
like tension and boredom can be defined as negative when deliberating about player
experiences, they also contribute to the challenge of a game, and are presumably essential
in order for the overall game experience to work (Gilleade & Dix 2004; Poels et al. 2009).
Finally, an experience that does not necessarily need to be optimal or fulfilling is the
player’s sensation of inhabiting the space represented on-screen. This phenomenon is
generally referred to as presence (Minsky 1980) or immersion (Murray 1998), although
the latter is the more commonly used term in popular and academic discussions of game
experience (Calleja 2011). A prerequisite of such a sense of immersion in digital games
is player involvement (Calleja 2011). According to Calleja, immersion can be seen as
a blending of a variety of experiential phenomena afforded by involving gameplay (for
example, involvement with the narrative and quest of the game, the aesthetic beauty of the
game environment, etc.). This sense of being transported to an alternative reality is what
makes games very interesting tools for advertisers. The more a person feels like he or she
is in the virtual environment represented on-screen, the greater the impact of the events
that are experienced in that environment, including any persuasive communication (such
as IGA) (Nicovich 2010).
20
General emotions as player experiences

When people play digital games, they will also experience several general or ‘lower-order’
emotions (Poels & Dewitte 2006). Lower-order emotions are largely uncontrollable and
spontaneous emotional reactions that are continuously present to some degree, like pleas-
ure, arousal and dominance, traditionally referred to as PAD-dimensions (Mehrabian &
Russell 1974; Poels & Dewitte 2006). The pleasure dimension refers to the pleasantness
or enjoyment of a certain experience, and ranges from unpleasant to pleasant (Ravaja
et al. 2005). The arousal dimension gives an indication of the level of physical and mental
activation associated with the experience, and ranges from very excited or energised at
one extreme to very calm or sleepy at the other (Ravaja et al. 2005). Finally, the dominance
dimension concerns the feeling of control and influence over others and surroundings,
and ranges from very dominant to very dominated (Gilroy et al. 2008; Klimmt et al. 2007).
In a gaming context, studies of van Reekum et al. (2004) and Ravaja et al. (2005) dem-
onstrated that different in-game events can lead to changes in these PAD-dimensions,
indicating that they are highly related to players’ gameplay and experience. For example,
compared to positive game events, negative game events (e.g. failing to complete a certain
objective) often lead to lower levels of pleasure and higher levels of arousal (van Reekum
et al. 2004; Ravaja et al. 2005). Studies also showed that the PAD-dimensions have an
influence on people’s motivations to play games, their play duration, game preferences, etc.
(e.g. Mehrabian & Wixen 1987; Sherry et al. 2006; Poels et al. 2012). Moreover, former
advertising research found that these general emotional dimensions have an impact on the
way people process advertisements, making them highly relevant to this study (e.g. Newell
et al. 2001; Morris et al. 2002; Poels & Dewitte 2008).
Relationships between specific and general player experiences

When people are playing digital games and are confronted with several specific player
experiences (e.g. challenge, tension), the consequences thereof will also be felt on the gen-
eral emotional level. For example, in a gaming context, the pleasure dimension refers to
the pleasantness or enjoyment of the gameplay. This game enjoyment is based on several
player experiences, like cyclic feelings of suspense and relief, the experience of control, and
the fascination from being immersed in an alternative reality (Klimmt 2003). Moreover,
each element of the flow experience (e.g. concentration, the balancing of challenge and
skills) also contributes to the pleasantness of a gaming episode (Sweetser & Wyeth 2005).
Arousal, then, is related to the physical and mental activation a certain gaming expe-
rience evokes. Former research focusing on player experiences in a gaming context has
already demonstrated that experiences of boredom and competence can be interpreted
as low arousal experiences, while experiences of tension, challenge and, to a lesser extent,
immersion and flow correlate with high arousal states (Grigorovici & Constantin 2004;
van den Hoogen et al. 2008).
Finally, the dimension of dominance gives an indication of the amount of control a person
perceives to have over others and his surroundings (Klimmt et al. 2007; Gilroy et al. 2008).
21
As such, dominance is highly related to challenge (which can also be defined as the struggle
for control), and associated player experiences like competence, tension and boredom.
The general emotions themselves are also interrelated. To some extent, the dimen-
sions of dominance and arousal can be seen as each other’s counterparts. As people feel
that they are losing their control over the game world, they will feel more challenged and
will consequently experience more arousal in the process (van den Hoogen et al. 2008).
The relationship between pleasure and the two other general emotions is more complex.
Depending on, among other factors, the skills of the player and the challenges of the
game, both arousal and dominance can either lead to tension (e.g. negative arousal in the
form of frustration when challenges are too difficult) or enjoyment (e.g. positive arousal
or excitement when challenges and skills are in balance) (Poels et al. 2009).
Manipulation of game difficulty

Because our interest lies in studying the impact player experiences might have on the
effectiveness of in-game advertising, it is important to know which factors can lead to
variations in these player experiences. Studies by Sykes and Brown (2003) and van den
Hoogen et al. (2008) showed that the intensity of different player experiences can change
whenever the difficulty of the game is increased from easy to medium to hard. Levels with
a higher difficulty were found to mostly lead to higher levels of arousal, and high arousal
experiences like frustration, challenge and, to a lesser extent, flow and immersion. Easier
levels, on the other hand, were found mostly to lead to lower levels of arousal, and low-
arousal experiences like pleasure, dominance, boredom and competence (Sykes & Brown
2003; van den Hoogen et al. 2008).
Based on these studies, we propose that the manipulation of game difficulty will have
an effect on a multitude of player experiences. In our experiment, we therefore manipu-
lated the difficulty of a digital game. However, we do not expect all player experiences
to vary between experimental groups in the current study. In our experimental design, a
2D casual PC game with a minimal story, and simple graphics, sounds and music, was
used (see the ‘Method’ section). These audio-visual and narrative game characteristics did
not vary between the experimental levels. Also, it was the first time that participants had
played the experimental game level, and they played it for only a short amount of time
(seven minutes). In such a gaming context, we do not expect the rather complex player
experiences of immersion and flow to be experienced intensely, nor do we expect them
to vary between conditions. For high levels of immersion to occur in such a short period
of time, a game generally needs impressive graphics and/or an engaging, profound story
with appealing game characters (Ermi & Mäyra 2005). Moreover, for people to be able
to encounter the optimal experience of flow, the game must be able to provide a highly
involving and intrinsically motivating context, and a state in which skills and challenges
are perfectly balanced needs to be reached. We do not expect this to be possible in the
relatively short time span of seven minutes.
Furthermore, we do not expect to observe significant changes in boredom between
experimental levels. Our experimental game was designed to be a fun little game that was
22
able to grab and hold the attention of the participants during the experiment. Moreover,
since it was the first time that participants had played the game, we do not expect them
to find the experience boring. Also, playing a digital game is in itself a rather enjoyable
experience that is generally less boring than completing a repetitive, non-challenging task.
The player experiences we do expect to vary between difficulty levels in the current study
are pleasure, arousal, dominance, challenge, competence and tension. These experiences are
less dependent on the audio-visual and narrative design of the game, and are more directly
related to the manipulation of game difficulty compared to, for example, immersion. More
specifically, based on the studies of Sykes and Brown (2003) and van den Hoogen et al.
(2008), mentioned earlier, we expect that, compared to the more difficult game levels, the
easier game levels will be less challenging and arousing, but also easier to master in a short
amount of time, thus leading to more pleasure, dominance and competence, and less tension.
Because higher levels of pleasure, dominance and competence will go together with
positive emotional experiences (e.g. enjoyment, pride, euphoria), we expect people to
experience more positive emotions in the easier levels. Furthermore, greater amounts
of tension, challenge and arousal are expected to lead to negative emotional experiences
in the case of the particular game used, because the levels with a higher difficulty were
designed to be significantly harder to master in a short amount of time, potentially leading
to negative experiences like frustration and irritation. Therefore, compared to easy levels,
we expect people to experience fewer positive emotions, more negative emotions and
more arousal in levels with a higher difficulty. In other game contexts, however, it is quite
possible that arousal and challenge are associated with positive emotional experiences like
enjoyment (e.g. when a satisfying balance between the challenges of the game and the
skills of the player can be reached) (Grigorovici & Constantin 2004).
Impact of game difficulty on the effectiveness of in-game

advertising
We expect that our manipulation of game difficulty will not only have an impact on sev-
eral player experiences, but that this will also have an effect on the processing and evalua-
tion of in-game advertising, both indirectly (by varying the valence and intensity of player
experiences) and directly. In the following sections we formulate hypotheses concerning
the indirect and direct effects of manipulating game difficulty on brand processing (i.e.
recognition) and brand evaluation (i.e. likeability).
Indirect effect: impact of player experiences
Effect on brand processing

Emotional valence: hedonic contingency hypothesis
People who are playing digital games can be confronted with both positive and negative
experiences. As already mentioned, traditional advertising studies showed that, compared
to people who are in a negative emotional state, those who are in a positive emotional state
23
have a better recall and recognition of message stimuli and brand names (Gardner 1985; Lee
& Sternthal 1999). The explanation for this effect is based on the hedonic contingency theory
(Wegener et al. 1995; Lee & Sternthal 1999). This theory states that people who are in a
positive state of mind are interested in sustaining their positive emotional state. As a conse-
quence, they will scrutinise the hedonic consequences of a particular action. This means that,
compared to individuals who are in a negative state of mind, individuals who are in a positive
state of mind will analyse persuasive messages more closely for their hedonic consequences,
resulting in a more elaborate processing of information, and thus better recognition scores
(Wegener et al. 1995). Because people may encounter varying levels of positive and negative
experiences (inducing either a positive or negative emotional state) while playing digital
games, we expect this theory to also apply in the case of in-game advertising. More specifi-
cally, because easier game levels are suggested to lead to more positive and fewer negative
emotional experiences in the current study, we propose the following hypothesis:
H1: Compared to the more difficult game levels, the easier game levels will gener-
ate more elaborate brand processing and thus lead to a better recognition of
in-game ad placements.
Arousal: limited capacity model of mediated message processing hypothesis

Game environments often present players with highly demanding, challenging and arous-
ing experiences (Grigorovici & Constantin 2004). Based on the limited capacity model of
mediated message processing (Lang 2000), we propose that this arousing character of digital
games might have consequences for the effectiveness of in-game advertising. The limited
capacity model states that a person’s ability to process information is limited. Moreover, it
has been shown that highly challenging and thus arousing environments consume more
of people’s cognitive resources, resulting in people focusing their attention primarily on
the most essential information and tasks at hand, while being distracted from secondary
information (e.g. commercial content) (Lang 2000; Klimmt & Vorderer 2003; Grigorovici
& Constantin 2004). For example, in television and magazine advertising settings, it has
already been demonstrated that viewers have poorer recall and recognition of commercial
messages in highly arousing contexts (Norris & Colman 1992; Coulter & Sewall 1995;
Coulter 1998). Because people often experience varying degrees of challenge and arousal
while playing digital games, we expect the limited capacity theory to also be valid in the
case of in-game advertising. In our experimental game, easier game levels are suggested
to lead to less arousal. Consequently, in the specific context of our game, we stick to the
expectations formulated in our first hypothesis (H1), stating that easier game levels will
lead to a more elaborate brand processing, and thus a better recognition of IGA.
Effect on brand evaluation

Emotional valence: spill-over hypothesis
In the case of brand evaluation, traditional advertising studies showed that people who
are in a positive emotional state evaluate brands more positively compared to those who
24
are in a negative emotional state (Owolabi 2009). An explanation for this effect is offered
by the mood congruency model (Bower 1981). Bower found that a person’s state of mind
influences their judgement by increasing the accessibility of congruent thoughts. Positive
feelings enhance persuasion by priming positive thoughts, while negative feelings prime
negative thoughts. In an advertising context, various studies showed that a person’s psy-
chological response to the context spills over to his or her attitude towards the advertise-
ment, influencing it in a similar direction (e.g. Goldberg & Gorn 1987; Murry et al. 1992;
Aylesworth & MacKenzie 1998). This so-called spill-over hypothesis (e.g. Pavelchak et al.
1988; De Pelsmacker et al. 2002; Moorman et al. 2006) therefore proposes that, when the
context in which an advertisement is shown conveys positive feelings, the advertisement
will also be evaluated positively (and vice versa for negative feelings). In a similar way,
we expect this theory to be relevant in the case of in-game advertising. More specifically,
because we expect easier game levels to lead to more positive emotions and fewer negative
emotions, we propose the following hypothesis:
H2: Compared to the more difficult game levels, the easier game levels will lead to
a more favourable evaluation of the in-game ad placements and thus a higher
brand likeability.
Direct effect of game difficulty

Apart from the indirect effect of game difficulty through the variation of player experi-
ences, we also expect that game difficulty will have a direct effect on the processing and
evaluation of in-game advertisements.
Effect on brand processing: limited capacity model of mediated

message processing
As for the direct effect of increasing game difficulty on brand processing, we refer once
again to the limited capacity model of mediated message processing (Lang 2000), stating that
people have a limited capacity for the cognitive processing of information. When people
decide to pay attention to a message or task (e.g. playing a digital game), they allocate
resources to information processing. When a message or task requires more concentration
and attention (e.g. when the difficulty of the game increases), fewer cognitive resources
can be spent on other tasks, resulting in people focusing their attention primarily on the
main task at hand (in this case: playing the game), while being distracted from second-
ary information (such as IGA) (Lang 2000; Klimmt & Vorderer 2003; Grigorovici &
Constantin 2004).
Consequently, because we manipulated the difficulty of a digital game in our current
study, we expect that people playing easier game levels will have to pay less attention to their
primary task (scoring points while avoiding obstacles), giving them more opportunities to
focus on the game environment, and the advertisements integrated into its background.
These expectations are in line with the predictions formulated in our first hypothesis (H1).
25
Effect on brand evaluation: mere exposure theory

Regarding the direct effect of game difficulty on brand evaluation, the possibility of
‘mere exposure’ effects (Zajonc 1980) has to be considered. Advertisements often occur
under incidental conditions, when a person’s attention is focused on something else – for
example, when reading a magazine, surfing the internet or, in this case, while playing a
digital game. Under such circumstances, stimuli such as advertisements that are inte-
grated into the background of the game are usually not the players’ central focus, and
as a result, they are not actively processed. However, since most emotional reactions and
learning transpire beyond awareness, these ads can still cause changes in people’s sub-
conscious minds (Zajonc 1980; Janiszewski 1993). More specifically, repeated exposures
to the ads might lead to the development of preferences and enhanced liking merely
because people are (subconsciously) familiar with them (Zajonc 1980; Janiszewski 1993).
Our manipulation of difficulty might lead to such mere exposure effects. Participants
in our experiment will have to play either an easy, medium or hard game level for seven
minutes. It is quite possible that people in the more difficult conditions will need more
time to successfully manoeuvre through the game level, because there are more obstacles
scattered around the game environment. This means that people in the easier conditions
will be able to move more quickly through the level, giving them more opportunities to
come into contact with the ads. Since frequency seems to breed familiarity, and thus like-
ability (Zajonc 1980; Janiszewski 1993), we propose that the expectations following the
mere exposure theory will be in line with our second hypothesis (H2).
Player experiences as mediators

If the results of our analyses indicate that increasing the difficulty of the game indeed
leads to variations in several player experiences and the effectiveness of IGA, we expect to
find that the player experiences that are significantly affected by our manipulation weaken
and therefore partially mediate the relationship between game difficulty and IGA effec-
tiveness. For this reason, we formulate the following hypotheses:
H3: The player experiences that are significantly affected by our manipulation of game
difficulty will partially mediate the effect of game difficulty on IGA recognition.
H4: The player experiences that are significantly affected by our manipulation of game
difficulty will partially mediate the effect of game difficulty on brand likeability.
Study
Materials
An online PC game, called Flight of the Strihuhn (Herrewijn & Ratarf Games 2010), was
developed for the purpose of this study. Flight of the Strihuhn is a 2D casual game in which
26
the principal character is a little bird. In order to score points, the player character (PC)
had to compete in a sport called ballhuhn. The goal of this sport was to collect balloons
and power-ups (e.g. red berries) while avoiding obstacles (e.g. spikes that are scattered all
around the level). Collecting a power-up produced a reward of 5 points, while collecting a
balloon produced a reward of 25 points. Every time the PC hit an obstacle, he lost some
of his health. After being hit by an obstacle five times, the PC would re-spawn at a certain
point in the level and lose 10 points.
The decision to develop a 2D casual game was made in order to be able to exert a large
amount of control on the composition of the digital game experience and placements of
the in-game ads, and the integration of the questionnaires in the game. This way it was
also possible to keep variations in certain variables (e.g. display resolution, time frame of
the levels) to a minimum. An online casual game is also easier for participants to down-
load and to learn in a short amount of time.
Procedure
A total of 99 participants were recruited by placing invitations for the online experiment
on several online game forums (e.g. 9lives, FOK!, ZGeek) and the social networking sites
Facebook and Twitter. The request to participate was made without revealing the purpose
of the study. Practical information about the experiment was provided on the website where
the PC game could be downloaded. Participants played the online game at home, without
the presence of a researcher, which is advantageous for the external validity of the study.
However, this also means we were not able to keep everything under control, like the size
of the PC monitor the participants used to play the game and the amount of attention the
participants devoted to the experiment. To keep variations in these variables to a minimum,
the online game was always played in full-screen mode and in the same display resolution
(1024 × 768). The participants were also explicitly asked to concentrate on the experiment as
much as possible, and to play the game and fill out the questionnaires without interruptions.
After downloading and starting the game, participants had to answer some questions
concerning their socio-demographic (age, gender) and gaming characteristics (game fre-
quency). Subsequently, the first phase of the experiment could start. This phase consisted
of playing a neutral game level without in-game advertising. This first level was similar
for all participants, and its main goal was to explain and train the basics of the game and
to get acquainted with the procedure of the experiment.
Upon completion of the first level, participants were automatically and randomly
assigned to one of the three experimental conditions. Because our manipulation con-
cerned the difficulty of the game, participants had to either play an easy, medium or hard
level. Compared to the more difficult game levels, the easier levels contained fewer obsta-
cles, making it easier for the player to navigate through the game environment and collect
power-ups and balloons (see Figure 1). The time frame to complete this second level
remained the same for each player (seven minutes), regardless of game difficulty. In the
upper-left corner of the game interface, there was a timer counting down from seven min-
utes. Players were aware of the fact that when the time ran out, the game would be over.
27
Figure 1: Screenshots of the easy, medium and hard levels
Apart from the difference in difficulty, these levels looked exactly the same. Each time,
five billboards of fictitious brands were integrated into the background of the game. The
choice to integrate fictitious instead of real brands was made in order to make sure that
participants would not be influenced by their prior experiences with specific brands. The
billboards included brands of shoes (B&L), a radio channel (Radio Cirius), a mobile phone
(iStar), a soft drink (Cuhna) and a fast-food chain (MacBurger) (see Figure 2).
Figure 2: Screenshot of a billboard (iStar) in the online PC game Flight of the Strihuhn
28
These billboards were all about the same size, and clearly visible and readable. During
game play, players passed each billboard several times. After finishing the experimental
level, participants were asked to rate the emotions and experiences they felt while play-
ing. These questions were followed by a questionnaire concerning their recognition of
and likeability towards the five brands shown in the game (see the ‘Measures’ section for
further details). On average, the experiment lasted about 15 to 20 minutes.
Participants
A total of 99 players (79 male 20 female) participated in the experiment. Every experi-
mental group contained 33 participants. The age of the participants varied between 16
and 50 years, with an average age category of 21 to 30 years old (16 to 20 years old:
18.2%. 21 to 30 years old: 58.6%. 31 to 40 years old: 17.2%. 41 to 50 years old: 6.1%);
77.8% of the participants played digital games for less than 15 hours per week, 22.2% of
the participants played for 15 hours or more. The random assignment of the participants
to the experimental groups ensured that the composition of the easy, medium and hard
group did not differ significantly concerning their socio-demographic and gaming char-
acteristics (gender (c2 (2, N = 99) = 0.13, NS), age (c2 (6, N = 99) = 4.02, NS) and game
frequency (c2 (2, N = 99) = 0.67, NS)).
Measures
General emotions
The general emotions the participants felt while playing the game were measured
retrospectively by using the Self-Assessment Manikin (SAM-scale) (Lang 1980). The
SAM-scale is a visual scale based on the PAD-dimensions of Mehrabian and Russell
(1974), and uses three 9-point visual scales on which participants have to indicate how
much pleasure, arousal and dominance they felt while playing the game. Scale values ranged
from 0 to 8, with ascending scores corresponding to higher pleasure, arousal and domi-
nance ratings. This method is being used and has been validated in both advertising (e.g.
Morris 1995; Morris et al. 2002) and gaming research (e.g. Chanel et al. 2008; van den
Hoogen et al. 2008, 2009).
Specif ic player experiences

Specific player experiences were also measured retrospectively by asking participants to fill
out the in-Game Experience Questionnaire (iGEQ) (IJsselsteijn et al. 2008) after playing
the experimental level. The iGEQ explicitly asks the participants to indicate how they
felt while playing the game. For each specific player experience, the iGEQ includes two
statements to which agreement is measured on a 5-point intensity scale ranging from 0
(‘not at all’) to 4 (‘extremely’). We included the following specific player experiences: com-
petence (‘I felt successful’, ‘I felt skilful’; r = 0.78, p < 0.001), challenge (‘I felt challenged’,
‘I felt stimulated’; r = 0.71, p < 0.001), tension (‘I felt frustrated’, ‘I felt irritable’; r = 0.78,
p < 0.001), immersion (‘I was interested in the game’s story’, ‘I found it impressive’; r = 0.77,
29
p < 0.001), boredom (in the iGEQ called negative affect: ‘I felt bored’, ‘I found it tiresome’;
r = 0.68, p < 0.001) and flow (‘I forgot everything around me’, ‘I felt completely absorbed’;
r = 0.72, p < 0.001).
The Game Experience Questionnaire has been used in several gaming studies and is
of sufficient quality to accurately report gameplay experience (e.g. Gajadhar et al. 2009;
Nacke et al. 2010).
IGA recognition
IGA recognition was measured on three levels. First, the recognition questionnaire pre-
sented participants with lists of product categories, next with lists of brand names, and
eventually with lists of pictures of billboards. Every time, participants had to indicate
which categories, brand names and billboards they remembered encountering in-game.
For each IGA recognition measure, the correct option was included, as well as four filler
items and an ‘I don’t know’ option. For every list that was presented, participants could
pick only one out of the six options.
The data that originated from this questionnaire were combined into three recognition
variables: recproduct, recbrand and recbillboard. These variables indicate how many product categories
(recproduct), brand names (recbrand) and billboards (recbillboard) each participant recognised correctly.
Brand likeability
To assess brand likeability, participants were successively shown the five billboards inte-
grated into the background of the game, and asked to rate the statement ‘I feel positive
towards this brand’ by means of 5-point Likert scales ranging from 0 (‘strongly disagree’)
to 4 (‘strongly agree’). The data that originated from this questionnaire were combined
into one variable: brandlike. This variable is an indication of each participant’s mean
likeability of the five brands (Cronbach’s alpha = 0.71). The term ‘brand likeability’ is
used instead of ‘brand attitude’, because the experiment incorporates fictitious instead
of real brands. Therefore, we cannot be sure whether we measure the attitude towards
the integrated brand, the attitude towards the generic product (e.g. fast-food chain) or
both. Brand likeability is a better option in this regard, as it is a general variable that says
something about the global reaction of a person concerning an advertisement (Walker &
Dubitsky 1994), while brand attitude is a more complex variable.
Results
Manipulation check
Before being able to analyse the effect of game difficulty on both the player experiences
and IGA effectiveness, we had to check whether our manipulation of game difficulty
had been effective. Therefore, we analysed the results of a table of means and one-way
ANOVA with the game score of each player as the dependent variable (DV) and game dif-
ficulty as the independent variable (IV) (see Table 1). The results showed that the score of
the players differed significantly between all difficulty groups (F(2, 96) = 1.52, p < 0.001),
30
providing initial confirmation that our

manipulation of game difficulty had Table 1: Table of means and one-way
been effective. ANOVA with the score of the players as
Moreover, it was expected that, in the DV and game difficulty as the IV
the case of the particular game used Score
in the experiment, the manipulation Difficulty M SD
of game difficulty would have a signifi- Easy (1) 309.09 79.63
cant impact on the player experiences Post hoc tests (Tukey) **2, ***3
of pleasure, arousal, dominance, compe- Medium (2) 245.91 82.67
tence, challenge and tension, but not on Post hoc tests (Tukey) **1, *3
the experiences of immersion, flow and Hard (3) 191.21 62.85
boredom. To check if this were the case, Post hoc tests (Tukey) ***1, *2
tables of means and one-way ANOVAs
F(2, 96) 20.12***
were conducted with the player experi-
*p < 0.05, **p < 0.01, ***p < 0.001
ences as the dependent variables and
game difficulty as the independent vari-
able (see Tables 2 and 3). These analyses showed that the variations in pleasure, arousal,
dominance, competence, challenge and tension between the levels were in line with our expec-
tations. The results demonstrate that, compared to the medium and hard level, the easy
level indeed led to higher levels of pleasure (F(2, 96) = 12.87, p < 0.001), dominance (F(2,
96) = 12.94, p < 0.001) and competence (F(2, 96) = 12.17, p < 0.001), and to lower levels
of arousal (F(2, 96) = 12.21, p < 0.001), tension (F(2, 96) = 12.65, p < 0.001) and challenge
(F(2, 96) = 10.28, p < 0.001). Further post hoc tests (PHT) revealed that the amounts of
pleasure, dominance, competence and tension people felt differed significantly for all experi-
mental groups (easy, medium and hard). The amounts of challenge differed significantly
for all conditions but the medium and hard group, and the amounts of arousal differed
significantly for all conditions but the easy and medium group.
Table 2: Table of means and one-way ANOVAs with the general player
experiences as the DVs and game difficulty as the IV
Pleasure Arousal Dominance

Difficulty M SD M SD M SD
Easy (1) 5.52 1.15 3.06 3.06 5.27 5.27
PHT (Tukey) *2, ***3 ***3 *2, ***3
Medium (2) 4.55 1.72 3.94 3.94 4.03 4.03
PHT (Tukey) *1, *3 *3 *1, *3
Hard (3) 3.61 3.61 5.15 5.15 2.88 2.88
PHT (Tukey) ***1, *2 ***1, *2 ***1, *2
F(2, 96) 12.87*** 12.21*** 12.94***
Notes: *p < 0.05, **p < 0.01, ***p < 0.001. The general player experiences were measured by using 9-point visual scales
ranging from 0 (very low) to 8 (very high)
31
Table 3: Table of means and one-way ANOVAs with the specific player experiences
as the DVs and game difficulty as the IV
Competence Tension Challenge Immersion Flow Boredom

Difficulty M SD M SD M SD M SD M SD M SD
Easy (1) 1.95 1.01 0.71 0.80 1.03 0.72 0.97 0.84 1.05 0.63 1.30 0.74
PHT (Tukey) *2, ***3 *2, ***3 *2, ***3 ***3
Medium (2) 1.41 0.88 1.29 0.79 1.52 0.81 1.17 0.90 1.45 0.78 1.20 1.08
PHT (Tukey) *1, *3 *1, *3 *1 *3
Hard (3) 0.85 0.83 1.91 1.24 1.95 0.94 1.32 1.04 2.02 0.91 1.08 0.73
PHT (Tukey) ***1, *2 ***1, *2 ***1 ***1, *2
F(2, 96) 12.17*** 12.65*** 10.28*** 1.17 12.70*** 0.57
Notes: *p < 0.05, **p < 0.01, ***p < 0.001. The specific player experiences were measured by using 5-point intensity scales
ranging from 0 (not at all) to 4 (extremely)
Furthermore, as expected, the amounts of immersion (F(2, 96) = 1.17, p = 0.32) and
boredom (F(2, 96) = 0.57, p = 0.57) that participants experienced were low and did not
differ significantly between the experimental conditions (see Table 3).
However, against expectations, the flow dimension did differ according to game dif-
ficulty (see Table 3). The results show that, compared to the more difficult game levels, the
easier levels led to lower degrees of flow (F(2, 96) = 12.70, p < 0.001), with further post hoc
tests revealing significant differences between all experimental groups. This is in line with
the study from van den Hoogen et al. (2008), which states that game levels with a higher
difficulty go together with higher levels of arousal, and high arousal experiences like flow.
Nevertheless, we argue that the measure of flow (‘I forgot everything around me’, ‘I felt
completely absorbed’), in this particular case, mainly gives an indication of the amount of
absorption and concentration that was needed to reach the goals of the game, rather than
give an indication of whether or not a perfect balance between skills and challenges was
attained. As the variations in the other player experiences (pleasure, arousal, dominance, compe-
tence, tension and challenge) demonstrate, players experienced higher levels of negative arousal
(e.g. frustration) and lower levels of pleasure in the more difficult game levels, which means
the game was probably not able to guide participants to an optimal state of flow.
Effects of game difficulty on IGA recognition and brand likeability

By altering the difficulty of the game, seven player experiences (pleasure, arousal, dominance,
competence, tension, challenge and flow) were thus manipulated. To check whether these
player experiences have an effect on IGA processing (recognition) and brand evaluation
(likeability), first of all, the influence of game difficulty on the effectiveness variables had
to be analysed. Again, tables of means and one-way ANOVAs were conducted with IGA
recognition (recproduct, recbrand and recbillboard) and brand likeability (brandlike) as the dependent
variables, and the difficulty of the game as the independent variable (see Table 4).
32
Table 4: Table of means and one-way ANOVAs with IGA recognition and brand
likeability as the DVs and game difficulty as the IV
Recproduct Recbrand Recbillboard Brandlike

Difficulty M SD M SD M SD M SD
Easy (1) 1.39 0.97 1.15 1.00 2.55 1.15 2.15 0.68
PHT *2, ***3 ***3 **2, ***3 *3
Medium (2) 0.85 1.12 0.73 0.88 1.73 1.28 2.04 0.35
PHT *1 **1, *3
Hard (3) 0.36 0.60 0.33 0.65 1.03 0.88 1.81 0.60
PHT ***1 ***1 ***1, *2 *1
F(2, 96) 10.30*** 7.56*** 15.22*** 3.01°
Notes: °p < 0.1, *p < 0.05, **p < 0.01, ***p < 0.001. The recognition variables indicate how many products/brands/billboards
people correctly recognised (0 to 5); brand likeability gives an indication of people’s mean likeability of the five brands, and
ranges from 0 (very negative) to 4 (very positive)
These analyses showed that the experimental groups differed significantly con-
cerning recproduct (F(2, 96) = 10.30, p <.001), recbrand (F(2, 96) = 7.56, p < 0.001) and
recbillboard (F(2, 96) = 15.22, p < 0.001). Further post hoc tests revealed that for recbillboard
all three conditions differed significantly. For recproduct all conditions but the medium and
hard group differed significantly, and for recbrand the tests showed significant differences
between the easy and hard group. The results of these analyses showed that, compared to
the medium and hard level, the easy level led to a better IGA recognition, confirming H1.
Finally, the differences between the groups concerning brandlike were marginally sig-
nificant (F(2, 96) = 3.01, p = 0.054). The post hoc results of these analyses showed that,
although in general the differences between groups were only marginally significant, the
easy level did lead to a significantly more favourable brand likeability (p < 0.05) compared
to the hard level, partially confirming H2.
Next, we checked for correlations between all player experiences, and between the
player experiences and the IGA effectiveness variables (see Table 5). This correlation
matrix showed that several player experiences are interrelated, that some of the player
experiences (especially pleasure, dominance and competence) are related to IGA recognition,
and that brand likeability is related only to pleasure.
Effects of player experiences on IGA recognition and brand likeability

Finally, to determine whether or not the different player experiences mediate the rela-
tionship between game difficulty and IGA effectiveness, we conducted a series of multiple
mediations using Preacher and Hayes’ (2008) bootstrapping methodology for indirect
effects (see Table 6). We included only those player experiences that were proven to cor-
relate with the corresponding IGA effectiveness variables in our analyses (see Table 5),
because the player experiences that do not correlate with these brand variables cannot, by
definition, serve as mediators.
33
34
Table 5: Correlation matrix of all player experiences and IGA effectiveness variables
Pleasure Dominance Arousal Competence Tension Challenge Flow Recproduct Recbrand Recbillboard Brandlike
Pleasure 1.00 0.54*** 0.08 0.44*** –0.47*** 0.04 –0.15 0.29** 0.25* 0.33** 0.26*
Dominance 0.54*** 1.00 –0.07 0.51*** –0.49*** –0.14 –0.21* 0.27** 0.21* 0.35*** 0.02
Arousal 0.08 –0.07 1.00 0.08 0.21* 0.57*** 0.41*** –0.06 –0.06 –0.21* 0.03
Competence 0.44*** 0.51*** 0.08 1.00 –0.34*** 0.05 –0.12 0.35*** 0.33** 0.38*** 0.08
Tension –0.47*** –0.49*** 0.21* –0.34*** 1.00 0.21*** 0.31*** –0.15 –0.14 –0.21* 0.01
Challenge 0.04 –0.14 0.57*** 0.05 0.21* 1.00 0.46*** –0.07 –0.06 –0.11 0.11
Flow –0.15 –0.21* 0.41*** –0.12 0.31** 0.46*** 1.00 –0.10 –0.25* –0.15 –0.03
Recproduct 0.29** 0.27** –0.06 0.35*** –0.15 –0.07 –0.10 1.00 0.70*** 0.71*** 0.20*
Recbrand 0.25* 0.21* –0.06 0.33** –0.14 –0.06 –0.25* 0.70*** 1.00 0.64*** –0.01
Recbillboard 0.33** 0.35*** –0.21* 0.38*** –0.21* –0.11 –0.15 0.71*** 0.64*** 1.00 0.09
Brandlike 0.26* 0.02 0.03 0.08 0.01 0.11 –0.03 0.20* –0.01 0.09 1.00
*p < 0.05, **p < 0.01, ***p < 0.001

Results of these multiple mediation tests showed that only competence was able to sig-
nificantly weaken the relationship between game difficulty and recproduct, recbrand and recbillboard.
Pleasure was also able to greatly diminish the effect of game difficulty on brandlike, although
the effect of pleasure on brandlike was not significant.
Following these results, we can partially confirm H3, as the player experience of com-
petence partially mediates the impact of our manipulation of the difficulty of the game
on IGA recognition.
H4, however, must be rejected, since none of our player experiences was able to signifi-
cantly weaken the effect of game difficulty on brand likeability.
Table 6: Multiple mediation analyses with IGA recognition and brand likeability
as the DVs, game difficulty as the IV and the relevant player experiences as
the mediators
BCa 95% CI
DV IV R² B t(98) Lower Upper
Recproduct Diff_M 0.21*** –0.41 –1.71
Diff_H –0.75 –2.76
Pleasure 0.04 0.56 –0.21 0.07
Dominance –0.001 –0.01 –0.17 0.15
Competence 0.19 1.68 –0.35 –0.001
Total –0.35 0.01
Recbrand Diff_M 0.17** –0.42 –2.02
Diff_H –0.60 –2.37
Pleasure 0.04 0.64 –0.19 0.08
Dominance –0.02 –0.48 –0.10 0.19
Competence 0.19 1.87 –0.34 –0.003
Flow –0.14 –1.22 –0.20 0.01
Total –0.41 0.01
Recbillboard Diff_M 0.30*** –0.55 –1.78
Diff_H –0.95 –2.35
Pleasure 0.09 1.03 –0.36 0.05
Dominance 0.05 0.78 –0.30 0.07
Arousal –0.08 –1.09 –0.27 0.03
Competence 0.24 1.69 –0.40 –0.001
Tension 0.15 1.14 –0.03 0.30
Total –0.63 0.04
Brandlike Diff_M 0.09* –0.05 –0.34
Diff_H –0.21 –1.38
Pleasure 0.06 1.68 –0.19 0.002
Notes: *p < 0.05, **p < 0.01, ***p < 0.001; n = 99, 5,000 bootstrap samples. Game difficulty was divided into three
dichotomous variables: Easy (Diff_E; reference variable), Medium (Diff_M) and Hard (Diff_H). BCa 95% CI = Bias Corrected and
Accelerated Bootstrapping Confidence Intervals. Confidence intervals containing zero are interpreted as not significant.
35
Discussion and conclusion

Prior advertising research studying classical media has shown that the context in which an
advertisement is presented can have an influence on how the message is processed by the
audience (Moorman 2003). A person’s state of mind, and the emotions and experiences
they encounter while being exposed to an advertisement, are very important aspects of this
context (Bronner et al. 2007; van Reijmersdal et al. 2010). Yet, although advertising studies
acknowledge the importance of context effects on the effectiveness of advertising in tradi-
tional media, research concerning the effects of context characteristics on the effectiveness
of advertising in digital games is limited.
Therefore, the aim of this article was to focus on the relationship that exists between
the effectiveness of in-game advertising and one specific context characteristic: player
experiences, or the emotional experiences people encounter while playing digital games.
Playing digital games has the potential to evoke numerous general and more specific
player experiences, which might exert an influence on the way players respond to in-game
advertising. To be able to analyse the impact of player experiences on the effectiveness of
IGA, a between-subjects experiment was employed in which participants had to play an
online PC game that contained several advertisements (billboards). The difficulty of the
game was manipulated, which was expected to affect IGA processing and evaluation both
directly and, more importantly, indirectly by varying several player experiences.
First of all, the influence of game difficulty on the player experiences was analysed. In the
current study, seven player experiences varied between the conditions. Easier levels led to
higher amounts of pleasure, dominance and competence, and to lower amounts of arousal,
tension, challenge and flow. This means participants in the easier levels experienced more
positive emotional experiences, fewer negative emotional experiences, and less arousal and
flow. These results were in line with our expectations, except for the variations in flow. We
argue that, in this particular case, flow mainly gives an indication of the amount of con-
centration players needed to successfully conclude the goals of the game. As variations in
other player experiences demonstrate, players experienced more tension, less pleasure, less
competence, etc., in levels with a higher difficulty, suggesting that an optimal state of flow
(which is characterised by a perfect balance between challenges and skills) was generally
not reached by participants.
Subsequently, the impact of game difficulty on the IGA recognition and brand likeabil-
ity of participants was examined. The results of the analyses showed that, compared to
the more difficult levels, easier levels led to a more positive IGA recognition and brand
likeability. These findings seem to be in line with our hypotheses, which were based on
the hedonic contingency and limited capacity model theories regarding brand processing
(H1), and the spill-over and mere-exposure theories regarding brand evaluation (H2).
The hedonic contingency theory (Wegener et al. 1995; Lee & Sternthal 1999) suggests
that people who are in a positive emotional state are interested in sustaining this state, and
will therefore analyse persuasive messages more closely for their hedonic consequences,
resulting in better brand recognition scores. Our findings seem to provide initial support
for this theory, as easier levels were able to evoke more positive emotions than did the
36
more difficult levels, apparently leading to higher amounts of IGA recognition in the
easier conditions.
The limited capacity model of mediated message processing theory (Lang 2000) states
that higher levels of concentration, engagement and arousal while paying attention to a
message or task lead to lower amounts of cognitive resources being available to process
secondary information, such as (in-game) advertisements. Our findings seem to be in line
with this theory, as the more difficult levels led to higher amounts of arousal, challenge
and concentration (as shown by the flow experience), resulting in a negative effect on IGA
recognition.
The spill-over theory (e.g. Pavelchak et al. 1988; De Pelsmacker et al. 2002; Moorman
et al. 2006), then, proposes that when the context in which an advertisement is shown
conveys positive emotions, the ad will also be evaluated positively (and vice versa for nega-
tive emotions). This was also the case in the current study, as easier levels conveyed more
positive emotions compared to the more difficult levels, apparently resulting in a more
favourable brand likeability.
Finally, the mere exposure theory states that repeated exposures to stimuli (e.g. advertise-
ments) lead to the development of preferences and enhanced liking merely because people
are familiar with them (Zajonc 1980; Janiszewski 1993). We hypothesised that people in
the easier conditions would be able to move more quickly through the level, giving them
more opportunities to come into contact with the ads, breeding familiarity and thus like-
ability. Again, our results seem to provide initial support for this hypothesis, since easier
levels resulted in a more favourable brand likeability.
However, the specific hypotheses concerning the impact of the player experiences can be
conclusively confirmed only if the player experiences mediate the relationship that exists
between game difficulty and IGA effectiveness. The mediation analyses showed that, in
particular, competence was able to significantly weaken the relationship between game
difficulty and IGA recognition, partially confirming H3. These results seem to provide
support for the hedonic contingency hypothesis, with competence evoking positive emotions
like pride and euphoria, leading to a positive effect on brand processing (Wegener et al.
1995; Lee & Sternthal 1999). An additional explanation may be that, if people are better
at playing a game, more of their attention can be devoted to the surroundings (e.g. bill-
boards). This may also be related to the limited capacity model of mediated message processing
(Lang 2000), which predicts that more experienced (and thus competent) players utilise
fewer cognitive resources when playing digital games because the mental and motoric
actions they need to perform in order to complete the goals of the game eventually
become automatic (Lang 2000; Schneider & Cornwell 2005).
An explanation for the strong mediating effect of competence may be that compe-
tence was the dimension most closely related to our manipulation of game difficulty.
By increasing the difficulty of the game, our goal was to make it harder for people to
achieve their goals and complete their tasks. By influencing player performance, mastery
and control, we thus primarily manipulated players’ feelings of competence (Klimmt et al.
2007; Trepte & Reinecke 2011), but also related experiences such as tension, dominance
and pleasure.
37
Finally, the results of the mediation tests also demonstrated that the player experience
of pleasure was able to greatly diminish the effects of game difficulty on brand likeability.
However, the effect of pleasure on brand likeability could not be proven significant. This
means that we were not able to provide conclusive support for H4, and thus our spill-over
hypothesis. Nevertheless, the impact of pleasure on the effectiveness of IGA, and more
specifically on brand likeability, is an area worth exploring in future studies.
The current study also has a few limitations. One of these is that the attention of the
participants could not be measured while playing the game. This way we were not able
to analyse the amount of focus each player was placing on the activity and elements of
the game environment. We hypothesised that easier levels would lead to lower levels of
(needed) concentration, involvement and arousal, thus leaving more cognitive space to
pay attention to the game environment in general, and the integrated billboards more
specifically. Our results showed that easier levels indeed led to a better brand recognition,
but the attentional processes underlying this outcome could not be analysed in detail yet.
This means we were not able to provide conclusive support for our limited capacity model
of mediated processing theory yet, and could not fully untangle the scope and reach of the
direct and indirect effects of our manipulation of game difficulty.
Furthermore, because we did not measure attention in this study, we were unable to test
the precise impact of mere exposure effects. That is, the variance in brand likeability might
be (partially) due to mere exposure effects, which causes people to develop a preference
for things merely because they are familiar with them (Zajonc 1980). Since we were not
able to measure people’s attention, however, we cannot fully support or reject the mere
exposure theory. We thus argue that it is important for future research on the effectiveness
of IGA to determine the frequency with which participants encounter the advertisements
in-game, and the amount of focus each player is placing on the activity and elements of
the game environment. Using an eye-tracker to evaluate participants’ attention while play-
ing a game might be a possibility to measure the effectiveness of IGA in greater detail.
Another limitation of the study concerns the low number of significant mediation
effects found. A possible reason for this low number of significant mediation effects may
lie in the fact that by altering the difficulty of the game, a wide array of (to some extent
interrelated) player experiences were manipulated. This way we were able to get an overall
picture of the impact of player experiences, but the specific effects of each experience could
not be fully analysed yet. The challenge for future research therefore lies in finding ways
to manipulate the different player experiences separately, in order to be able to examine
their specific effects.
Another reason might be that the direct impact of our manipulation of game difficulty
overshadowed the effects of the varying player experiences. As already mentioned above,
it is of utmost importance to be able to measure people’s attention in future research
studying the impact of player experiences on in-game advertising effectiveness, in order to
untangle the different emotional and attentional processes arising from people’s gameplay.
Several limitations can also be formulated related to the design of the current study.
The choice to develop a 2D casual PC game was made in order to be able to exert a large
amount of control regarding the structuring of the gameplay, the player experiences, the
38
integration of the billboards and questionnaires in the game, etc. To be able to test the
impact of different player experiences on the effectiveness of IGA for the first time, such
a strict and controllable design was a necessary first step. However, a lot of different game
platforms (e.g. PC, Sony PlayStation 3, Microsoft Xbox 360) and genres (e.g. shooter
games, role-playing games, puzzle games) exist, and each platform/genre possibly affects
the emotions and experiences of the player differently. Research has already shown that
people often play different kinds of games for different reasons, and that these games
elicit different kinds of emotions and experiences in their players (Lazzaro 2004). Our 2D
casual game, for example, was low-immersive, and rather limited in scope, narrative and
audio-visual richness, provided the player with a rather easy task (‘collect power-ups while
avoiding obstacles’), and did not offer the possibility to play with other people. Other
games, on the other hand, might actively focus on the game’s challenge, strategic think-
ing and problem solving; elicit high levels of immersion and escapism because of their
possibilities to explore and discover things, narrative and/or three-dimensional richness
of their game world; offer social experiences in form of competitive or collaborative game
modes; and so on (Lazzaro 2004). These different aspects possibly evoke different kinds
of experiences and emotions (e.g. frustration, immersion, flow, enjoyment). Therefore,
although the current study provides relevant insights in the way player experiences influ-
ence the effectiveness of IGA, we cannot claim that our results apply to all games. It
would be interesting for future studies to examine the impact of player experiences on the
effectiveness of IGA in other games.
Furthermore, an important remark to make regarding the player experiences we took
into account in the current study, is that we focused only on the general and specific
emotional experiences people encountered while playing the game. Meta-emotions were
not taken into account. Meta-emotions are emotions that people have about their own
emotions ( Jäger & Bartsch 2006). An example of such a meta-emotion in the context of
playing digital games, is that sometimes a hard and frustrating struggle to gain control
can ultimately be interpreted as enjoyable by a player (i.e. the ‘hard fun’ of challenge and
mastery, in which the feelings of frustration will eventually pass into feelings of personal
triumph and accomplishment (Lazzaro 2004)). We did not take along these meta-
emotions in the current study because we did not expect a great discrepancy between
the emotions felt while playing the game, and the emotions people felt about their own
emotions afterwards. The game was a casual 2D game that lasted for only seven minutes.
This means that people did not have a lot of time to practise or optimise their strategies,
and they did not get a second chance. Hence, we did not expect players’ meta-emotions
to differ greatly from the emotions they felt while playing the game. However, it would
be interesting to analyse the influence of meta-emotions on the effectiveness of in-game
advertising in more complex and engaging 3D games. When analysing one’s own emo-
tions after playing a digital game, a person might think differently of a game in retrospect,
possibly influencing his or her attitude towards the game and the ads that are integrated
into its environment.
Finally, as already stated, the emotional reactions of the players were measured ret-
rospectively and subjectively by using self-report questionnaires. This way we were not
39
able to objectively assess the players’ emotions and experiences while they were actu-
ally playing the game. These limitations need to be tackled in future research (e.g. by
combining self-report measures with psychophysiological measures like facial electro-
myography and electrodermal activity (Poels & Dewitte 2006)) in order to get a more
detailed picture of the impact of player experiences on the effectiveness of in-game
advertising.
Summarised, the study illustrates that increasing the difficulty of a digital game suc-
cessfully manipulates several player experiences, and that this also has an effect on the
way people process in-game advertising, supporting the notion that the emotions and
experiences players encounter while playing games are relevant context characteristics
that have to be taken into account while studying the effectiveness of in-game advertising.
The findings of this study might offer new insights and create possibilities for both the
gaming and in-game advertising industry. For example, advertisements could be incorpo-
rated in digital games while taking into account the specific context of the placement and
anticipating the emotions and experiences of gamers. In the case of the particular game
used in our experiment, easier game levels led to more positive emotional experiences
(e.g. pleasure, competence) and less negative emotional experiences (e.g. tension), flow
and arousal, resulting in a better IGA recognition and more favourable brand likeability.
Because the player experience of competence proved to have the strongest effect in the
current study, in order to maximise IGA recognition, it would be advisable for advertisers
to integrate advertisements on moments in the game scenario when the player is feel-
ing very competent and skilful (e.g. right after he has been able to successfully achieve a
certain goal or kill a boss character, evoking positive emotions like pride), while avoiding
moments when the player is feeling all but successful (e.g. when the player character is
being overwhelmed by superior foes or losing a race). An important challenge for future
research lies in finding ways to manipulate the different player experiences separately, in
order to be able to examine their specific effects and formulate specific recommendations
for the practice of in-game advertising.
References
Aylesworth, A.B. & MacKenzie, S.B. (1998) Context is key: the effect of program induced
mood on thoughts about the ad. Journal of Advertising, 27(2), 17–31.
Bardzell, J., Bardzell, S. & Pace, T. (2008) Player engagement and in-game advertising.
Available online at: http://yavuzkaynar.com/wordpress/_all/wp-content/uploads/2009/03/
ingame_adv_report.pdf (accessed 9 December 2012).
Bower, G.H. (1981) Mood and memory. American Psychologist, 36(2), 129–148.
Brightman, J. (2011) Video games ‘underutilized as advertising medium’, ads in games to hit
$7.2 billion by 2016 – DFC’. Available online at: http://www.industrygamers.com/news/
video-games-underutilized-as-advertising-medium-ads-in-games-to-hit-72-billion-by-
2016---dfc/ (accessed 9 December 2012).
Bronner, F.E., Bronner, J.R. & Faasse, J. (2007) In the mood for advertising. International
Journal of Advertising, 26(3), 333–356.
Calleja, G. (2011) In-Game: From Immersion to Incorporation. Cambridge, MA/London: MIT
Press.
40
Chanel, G., Rebetez, C., Bétrancourt, M. & Pun, T. (2008) Boredom, engagement and anxiety
as indicators for adaptation to difficulty in games. In: Proceedings of the 12th International
Conference on Entertainment and Media in the Ubiquitous Era. Tampere, Finland, 13–17.
Coulter, K.S. (1998) The effects of affective responses to media context on advertising
evaluations. Journal of Advertising, 27(4), 41–51.
Coulter, K.S. & Sewall, M.A. (1995) The effects of editorial context and cognitive and affective
moderators on responses to embedded ads. In: Kardes, F.R. & Sujan, M. (eds) Advances in
Consumer Research. Provo, UT: Association for Consumer Research, 177–183.
Csikszentmihalyi, M. (1990) Flow: The Psychology of Optimal Experience. New York: Harper & Row.
De Pelsmacker, P., Geuens, M. & Anckaert, P. (2002) Media context and advertising
effectiveness: the role of context style, context quality and context-ad similarity. Journal of
Ermi, L. & Mäyrä, F. (2005) Fundamental components of the gameplay experience: analysing
immersion. In: de Castell, S. & Jenson, J. (eds) Changing Views: Worlds in Play. Selected
Papers of the 2005 Digital Games Research Association’s Second International Conference, 15–27.
Gajadhar, B.J., de Kort, Y.A.W. & IJsselsteijn, W.A. (2009) Rules of engagement: influence of
co-player presence on player involvement in digital games. International Journal of Gaming
and Computer-Mediated Simulations, 1(3), 14–27.
Gardner, M.P. (1985) Mood states and consumer behavior: a critical review. Journal of
Consumer Research, 12(December), 281–300.
Gilleade, K.M. & Dix, A. (2004) Using frustration in the design of adaptive videogames. ACE
2004, 3–5 June, Singapore.
Gilroy, S.W., Cavazza, M., Chaignon, R., Mäkelä, S., Niranen, M., André, E., Vogt, T., Urbain,
J., Seichter, H., Billinghurst, M. & Benayoun, M. (2008) An affective model of user
experience for interactive art. Proceedings of the 2008 International Conference on Advances in
Computer Entertainment Technology.
Goldberg, M.E. & Gorn, G.J. (1987) Happy and sad TV programs: how they affect reactions
to commercials. Journal of Consumer Research, 14(3), 387–403.
Grigorovici, D.M. & Constantin, C.D. (2004) Experiencing interactive advertising beyond rich
media: impacts of ad type and presence on brand effectiveness in 3D gaming immersive
virtual environments. Journal of Interactive Advertising, 5(1). Available online at: http://jiad.
org/article53 (accessed 9 December 2012).
Griskevicius, V., Goldstein, N.J., Mortensen, C.R., Sundie, J.M., Cialdini, R.B. & Kenrick,
D.T. (2009) Fear and loving in Las Vegas: evolution, emotion, and persuasion. Journal of
Marketing Research, 46(3), 384–395.
Herrewijn, L. & Ratarf Games (2010) Flight of the Strihuhn.
IJsselsteijn, W.A., de Kort, Y.A.W. & Poels, K. (2008) The Game Experience Questionnaire:
development of a self-report measure to assess player experiences of digital games. FUGA
Deliverable D3.3. Eindhoven University of Technology.
Jäger, C. & Bartsch, A. (2006) Meta-emotions. Grazer Philosophische Studien, 73(2006), 179–204.
Janiszewski, C. (1993) Preattentive mere-exposure effects. Journal of Consumer Research,
20(December), 376–392.
Jennett, C., Cox, A.L., Cairns, P., Dhoparee, S., Epps, A., Tijs, T. & Walton, A. (2008)
Measuring and defining the experience of immersion in games. International Journal of
Human–Computer Studies, 66(2008), 641–661.
Klimmt, C. (2003) Dimensions and determinants of the enjoyment of playing digital games:
a three-level model. In: Copier, M. & Raessens, J. (eds) Level Up: Digital Games Research
Conference. Utrecht University, Faculty of Arts.
Klimmt, C. & Vorderer, P. (2003) Media psychology ‘is not yet there’: introducing theories on
media entertainment to the presence debate. Presence, 12(4), 346–359.
41
Klimmt, C., Hartmann, T. & Frey, A. (2007) Effectance and control as determinants of video
game enjoyment. Cyberpsychology & Behavior, 10(6), 845–874.
Lang, A. (2000) The limited capacity model of mediated message processing. Journal of
Communication, 50(1), 46–70.
Lang, P.J. (1980) Behavioral treatment and bio-behavioral assessment: computer applications.
In: Sidowski, J.B., Johnson, J.H. & Williams, T.A. (eds) Technology in Mental Health Care
Delivery Systems. New Jersey: Ablex Publishing, 119–137.
Lazzaro, N. (2004) Why we play games: four keys to more emotion without story. Available
online at: http://www.xeodesign.com/xeodesign_whyweplaygames.pdf (accessed
9 December 2012).
Lee, A.Y. & Sternthal, B. (1999) The effects of positive mood on memory. Journal of Consumer
Research, 26(2), 115–127.
Mackay, T., Ewing, M., Newton, F. & Windisch, L. (2009) The effect of product placement
in computer games on brand attitude and recall. International Journal of Advertising, 28(3),
423–438.
Mau, G., Silberer, G. & Constien, C. (2008) Communicating brands playfully: effects of in-
game advertising for familiar and unfamiliar brands. International Journal of Advertising,
27(5), 827–851.
Mehrabian, A. & Russell, J.A. (1974) An Approach to Environmental Psychology. Cambridge,
MA: MIT Press.
Mehrabian, A. & Wixen, W. (1986) Preferences for individual video games as a function of
their emotional effects on players. Journal of Applied Social Psychology, 16(1), 3–15.
Minsky, M. (1980) Telepresence. OMNI magazine, 2, 45–52.
Moorman, M. (2003) Context considered. The relationship between media environments
and advertising effects. Unpublished doctoral dissertation. Amsterdam: University of
Amsterdam.
Moorman, M., Neijens, P.C. & Smit, E.G. (2006) The effect of program responses on the
processing of commercials placed at various positions in the program and the block. Journal
of Advertising Research, 48(1), 49–59.
Morris, J.D. (1995) Observations: SAM: the Self-Assessment Manikin. Journal of Advertising
Research, 35(6), 63–69.
Morris, J.D., Woo, C.M., Geason, J. & Kim, J. (2002) Predicting intentions: the power of
affect. Journal of Advertising Research, 42(2), 7–17.
Murray, H.J. (1998) Immersion. In: Murray, H.J. (ed.) Hamlet on the Holodeck. Cambridge, MA:
MIT Press, 97–125.
Murry, J.P., Lastovicka, J.L. & Singh, S.N. (1992) Feelings and liking responses to television
programs: an examination of two explanations of media-context effects. Journal of Consumer
Research, 18(3), 441–451.
Nacke, L.E., Grimshaw, M.N. & Lindley, C.A. (2010) More than a feeling: measurement of
sonic user experience and psychophysiology in a first-person shooter game. Interacting with
Computers, 22(5), 336–343.
Nelson, M.R. (2005) Exploring consumer response to ‘advergaming’. In: Haugtvedt, C.,
Machleit, K. & Yalch, R. (eds) Online Consumer Psychology. Understanding and Influencing
Consumer Behavior in the Virtual World. Mahwah, NJ: Lawrence Erlbaum Associates, Inc.,
156–183.
Newell, S.J., Henderson, K.V. & Wu, B.T. (2001) The effects of pleasure and arousal on recall
of advertisements during the Super Bowl. Psychology and Marketing, 18, 1135–1153.
Nicovich, S.G. (2010) The effect of involvement on ad judgment in a computer-mediated
environment: the mediating role of presence. International Journal of Advertising, 29(4),
597–620.
42
Norris, C.E. & Colman, A.M. (1992) Context effects on recall and recognition of magazine
advertisements. Journal of Advertising, 21(3), 37–46.
Owolabi, A.B. (2009) Effect of consumers mood on advertising effectiveness. Europe’s Journal of
Psychology, 2009(4), 118–127.
Pavelchak, M.A., Antil, J.H. & Munch, J.M. (1988) The Super Bowl: an investigation into
the relationship among program context, emotional experience, and ad recall. Journal of
Consumer Research, 15(3), 360–367.
Poels, K. & Dewitte, S. (2006) How to capture the heart? Reviewing 20 years of emotion
measurement in advertising. Journal of Advertising Research, 46(1), 18–37.
Poels, K. & Dewitte, S. (2008) Getting a line on print ads. Pleasure and arousal reactions reveal
an implicit advertising mechanism. Journal of Advertising, 37(4), 63–74.
Poels, K., de Kort, Y.A.W. & IJsselsteijn, W.A. (2009) Identification and categorization of
digital game experiences. A qualitative study integrating theoretical insights and player
perspectives. Working paper. Eindhoven University of Technology.
Poels, K., van den Hoogen, W.M., IJsselsteijn, W.A. & de Kort, Y.A.W. (2012) Pleasure to
play, arousal to stay. The effect of player emotions on digital game preferences and playing
time. Cyberpsychology, Behavior, and Social Networking, 15(1), 1–6.
Preacher, K.J. & Hayes, A.F. (2008) Asymptotic and resampling strategies for assessing and
comparing indirect effects in multiple mediator models. Behavior Research Methods, 40,
879–891.
Ravaja, N., Saari, T., Laarni, J., Kallinen, K. & Salminen, M. (2005) The psychophysiology of
video gaming: phasic emotional responses to game events. Available online at: http://www.
digra.org/dl/db/06278.36196.pdf (accessed 9 December 2012).
Schneider, L. & Cornwell, B.T. (2005) Cashing in on crashes via brand placement in computer
games. The effects of experience and flow on memory. International Journal of Advertising,
24(3), 321–343.
Sherry, J.L. (2004) Flow and media enjoyment. Communication Theory, 14(4), 328–347.
Sherry, J.L., Lucas, K., Greenberg, B.S. & Lachlan, K. (2006) Video game uses and
gratifications as predictors of use and game preference. In: Vorderer, P. & Bryant, J. (eds)
Playing Computer Games: Motives, Responses and Consequences. Mahwah, NJ: Lawrence
Erlbaum, 213–224.
Sweetser, P. & Wyeth, P. (2005) GameFlow: a model for evaluating player enjoyment in games.
ACM Computers in Entertainment, 3(3), 1–24.
Sykes, J. & Brown, S. (2003) Affective gaming: measuring emotion through the gamepad. CHI
Extended Abstracts 2003, 732–733.
Trepte, S. & Reinecke, L. (2011) The pleasures of success: game-related efficacy experiences as
a mediator between player performance and game enjoyment. CyberPsychology, Behavior &
Social Networking, 14(9), 555–557.
van den Hoogen, W.M., IJsselsteijn, W.A. & de Kort, Y.A.W. (2008) Exploring behavioral
expressions of player experience in digital games. In: Nijholt, A. & Poppe, R. (eds)
Proceedings of the Workshop for Facial and Bodily Expressions for Control and Adaptation of
Games. Amsterdam: ECAG 2008, 11–19.
van den Hoogen, W.M., IJsselsteijn, W.A. & de Kort, Y.A. (2009) Effects of sensory
immersion on behavioural indicators of player experience: movement synchrony and
controller pressure. In: Atkins, B., Kennedy, H. & Krzywinska, T. (eds) Breaking New
Ground: Innovation in Games, Play, Practice and Theory: Proceedings of the 2009 Digital Games
Research Association Conference. London: Brunel University.
van Reekum, C.M., Johnstone T., Banse R., Etter, A., Wehrle, T. & Scherer, K.R. (2004)
Physiological responses to appraisal dimensions in a computer game. Cognition and
Emotion, 18(5), 663–88.
43
van Reijmersdal, E., Smit, E. & Neijens, P. (2010) How media factors affect audience responses
to brand placement. International Journal of Advertising, 29(2), 279–302.
Walker, D. & Dubitsky, T.M. (1994) Why liking matters. Journal of Advertising Research, 34(3),
9–18.
Wegener, D.T., Petty, R.E. & Smith, S.M. (1995) Positive mood can increase or decrease
message scrutiny: the hedonic contingency view of mood and message processing. Journal of
Personality and Social Psychology, 69, 5–15.
Zajonc, R. (1980) Feeling and thinking: preferences need no inferences. American Psychologist,
35(2), 151–175.
About the authors

Laura Herrewijn is a PhD student at the Department of Communication Studies at the
University of Antwerp, Belgium. Her doctoral research addresses the effectiveness of
in-game advertising, focusing primarily on the impact of player processes such as player
involvement and immersion.
Karolien Poels is an assistant professor at the Department of Communication Studies,
University of Antwerp. She received her PhD from Ghent University. Her current
research focuses on the experience and consumption of digital media (e.g. digital games,
virtual worlds, social media) and strategic communication in digital worlds (e.g. in-
game advertising, virtual shops). Her work has been published in Journal of Advertising
Research, Journal of Advertising, Cyberpsychology, Social Networking and Behavior and Media
Psychology.
Address correspondence to: Laura Herrewijn, PhD student, Department of
Communication Studies, University of Antwerp, Sint Jacobstraat 2, 2000 Antwerp,
Belgium.
Email: laura.herrewijn@ua.ac.be
44
Constitutive marketing
Towards understanding brand
community formation
Ana Kilambi
SAY Media
Michel Laroche and Marie-Odile Richard
Concordia University
This paper theoretically and conceptually examines the formation of brand communities
through the mechanisms of constitutive rhetoric. Although research has identified the con-
ceptual tenets of brand communities, the mechanisms through which brand communities
are formed and established remain largely understudied. This paper addresses this limitation
through using communications theory to develop a model of brand community formation
and the examination through historical case studies of the most notable and prominent brand
communities: Apple, Nike and Harley-Davidson. The model outlines the notion that brand
communities are established through persuasion and the rhetorical action of branding, and
the interplay among the brand, its message and consumers. Understanding the antecedents
of brand community, marketers can set forth to direct, grow and maintain their brand com-
munities through subtle relationship marketing initiatives, and, in doing so, strengthen their
brand equity.
Introduction
Simply put, a brand represents the attributes, identity and positioning of a product of
a company, and how that product is to be perceived by consumers. Where brands rep-
resent goods, and goods carry meaning beyond their function and commercial value
(McCracken 1986, p. 71), brands are relevant to culture in that they bear communicative
power to convey cultural meaning.
The interpersonal relationship between a consumer and a brand is a common area of
research in relationship marketing. The traditional model of consumer–brand relations is
formed via the exchange of brand values, which are communicated in two dimensions –
emotional values, such as self-portrayal and image transfer (Aaker 1997) and rational val-
ues, such as tangible benefits (Aggarwal 2004). With the advent of digital media, the focus
of research has shifted to the consumer–brand–consumer model. For Schembri (2008), in
testing the theory of consuming as play, through consumption practices consumed objects

DOI: 10.2501/IJA-32-1-045-064
became the vehicles of interaction (2008, p. 9). Establishing common ground through
their rational attachment to the brand, and sharing emotional experiences, consumers
are bound together as nodes in a brand community network. Similar to the interpersonal
relationships between a consumer and a brand, the brand community sits at the cross-
roads of emotional and rational values, but builds by integrating multiple consumer–brand
relationships in its schematic.
Models of brand community
Basic model
A brand community is a ‘specialized, non-geographically bound community, based on
a structured set of social relationships among admirers of a brand’ (Muniz & O’Guinn
2001, p. 412). Akin to McLuhan’s ‘global village’ (1962), the brand community plays a vital
role in a brand’s value and legacy, and is marked by the same markers as traditional com-
munities: a shared consciousness, rituals and customs, and a sense of moral responsibility.
For Muniz and O’Guinn, brand communities are of ‘a particular strip, and of their time’
(2001, p. 415), indicating that they are neither everlasting nor equal across brand bar-
riers. They concluded that brands available for public consumption with a strong image,
a history and a fierce competition are most likely to develop communities around them.
Extending the concept from an interpersonal to a mass/social level, they showed that
‘[b]rand communities reveal the socially situated nature of brands as something more than
a summation of attitudes or impoverished critical stereotypes’ (2001, p. 428).
As illustrated in the left side of Figure 1, the brand community involves two individual
consumer relationships with a brand, and the interaction of these consumers with one
another. Expanded further, this model includes multiple consumers, and their relation-
ships to the brand and to one another. Although the traditional model is forthright in its
articulation of the relationships between the brand, the consumer and the co-consumer,
it is elementary in that it does not elaborate on the messages, perceptions, and contexts of the
community and its members. In short, the traditional model of brand community does not
elaborate on the process of community formation.
Expanded models of brand community

Building upon this model, recent studies enriched the concept of brand community.
Where Muniz and O’Guinn (2001) linked the consumer and brand to the co-consumer,
and concentrated on characteristics key to the brand community, others developed a more
thorough examination of the concept in light of potential mediators, moderators and
contextual shifts. Introducing a consumer-centric model, McAlexander et al. (2002) pre-
sented a means by which marketers can strengthen their brand communities by facilitating
shared consumer experiences and flattering the dynamic characteristics of brand commu-
nity. Also, in studying European car clubs, Algesheimer et al. (2005) showed the social
influences of brand communities in leading to either positive or negative consequences on
46
Figure 1: Traditional model compared to the proposed model of the links between
the brand, the advertisement and the audience in the formative process of the
brand community
First
Persona
Brand Myth
Second
Persona
Brand
Advertisement
Consumer Consumer
Brand
Audience
community
an individual level, as well as the moderating effects of brand knowledge and community
size on these outcomes. The study exemplified the ‘importance of purposely selecting, ini-
tiating, managing and controlling interactions among consumers when facilitating brand
communities’ (2005, p. 30). Carlson et al. (2008) extended these findings in studying the
psychological sense of brand community, and found that brand-related constructs, such as
identification with the brand, can moderate brand commitment by consumers.
In sum, brand community is an evolving and dynamic concept. The literature on
brand community established a rich and thorough examination of the concept. One
major limitation, however, is that brand community was examined primarily from a per-
formance perspective, creating a limited knowledge of the evolution and inception of brand
communities. The purpose of this research is to establish a conceptual framework for the
constitutive process of brand community formation. Practically, this study helps explain how
brand communities can be garnered and maintained through advertising rhetoric. This is
crucial, in that it offers marketers a perspective of brand community from its onset, and
a brand-instigated perspective. By demonstrating the theoretical steps towards directing
brand community formation and nurturing, this approach will be of interest to brands that
wish to engage in brand community formation.
Role of advertising in brand experience

Advertising influences brand experience, and in doing so informs brand community. Thus,
to understand the formative process of brand community, advertising must be viewed as
a mode of communication with transformative capabilities. The revised communication
model for advertising (Stern 1994) describes advertising as a multidimensional, interac-
tive process. Understanding advertising as ‘a network of different messages depending on
47
different codes and working at different levels of signification’ (Eco 1987, p. 425), Stern’s
model conceptualised ad messages as texts crafted by a rhetor, and envisioned consumers
as active agents engaged in the production of meaning. Within this model, the source is
split into a triad composed of sponsor, author and persona (Stern 1994). Stern understood
the message communicated by the source as a discourse, activated by the audience in their
allegiance to its codes and ability to decipher them. The message can deploy the persona
in one of three fashions: autobiographical, narrative or dramatic enactment. The mode
of communication used to disseminate a message, as well as the medium, can contribute
to both the adoption and viability of the myth among the target audience. Stern (1994)
viewed the consumer as three entities (implied, sponsorial and actual), each accentuating
agency and ability to create meaning through participation in, and interpretation of, the
discourse. Once a consumer engages in attitudes and behaviours supposed for the implied
consumer, he becomes an interactive consumer, aligning the three personas and respond-
ing to messages in real time.
Qualitative research methodology

To assess and comprehend how a community is formed, qualitative research in the form
of theoretical and historical case analyses is undertaken. Including literature from com-
munications theory (primarily rhetoric and rhetorical action) and the social sciences, and
applying them to the literature on brand communities, through theoretical analysis, rela-
tionships between concepts can be explored to generate a theoretical basis for the under-
standing of brand community formation. Once a formal understanding is hypothesised
through literature review, an intensive investigation of specific brands seeks to substantiate
the theoretical concepts. Case selection is based on and employs a practical method. The
current analysis is exploratory, and thus is best served by a case wherein strong brand
communities are in existence, and whose advertising materials are accessible and con-
temporary. The Apple, Nike and Harley-Davidson brand communities are exemplars, as
all were born of humble beginnings but rose beyond their starting point with the help of
strong brand personalities and brilliant advertising campaigns.
Marketing communication process

Previous research examined consumer responses to differences in ads (Toncar & Munch
2001; De Pelsmacker et al. 2002; Cline et al. 2003; Braun-LaTour & LaTour 2004;
Garretson & Neidrich 2004). However, research has not examined the internal dynamics
of advertising in identity formation and the constitution of collective bodies. Advertising is not
communication for its own sake. Although an ad and product offering are tailored to an
identified target market, the message can also ‘create’ or ‘call into being’ the target market
for whom the message was intended. This is when a message carries a function extending
beyond its informative role. Advertising is conceptualised in a framework that draws upon
its constitutive capabilities, its capacity to communicate cultural meaning and its ability to
resonate ideological tenets within consumers.
48
Advertising as rhetoric
Rhetorics are concerned with how humans use language to reach an agreement that
permits coordination of efforts. Advertising is a rhetorical event as it is a commercial
message, sent from a sponsor to a consumer through a medium, with the specific intent
of influencing purchase decisions or instigating thought on matters of consumption. In
advertising rhetoric, the rhetor forwards an argument, and audiences serve to legitimate
the rhetor and his claims, ideology and desires. To forward his argument, the rhetor must
produce a ‘first persona’ (Black 1970) or an implied author. The first persona bears an
ethos and dianoia, standing for everything that a brand is, and wishes to be recognised for.
Describing the relationship between the rhetor and his audience, Perelman (1969) noted
that, ‘for argumentation to exist, an effective community of minds must be realized at a
given moment’ (p. 14). Thus, the rhetor constructs his audience so that they will be per-
ceptive to his claims, and positively appraise his leadership and convictions – what Black
(1970) calls the ‘second persona’. The second persona is the implied auditor of a rhetor’s
address, an understanding of whom he wants his audience to be. Effective rhetoric and
argumentation demand that the rhetor positions himself according to the laws of supply
and demand, anticipating the unforeseen and compelling needs, expectations and prior
knowledge of the audience so that he may provide for them accordingly.
Audiences cannot be constructed out of a rhetor’s pure imagination. Fundamentally, all
members of a community know and understand themselves as a collective. Nonetheless,
being self-proclaimed is not enough to substantiate a community as a veritable group; it
must be hailed by an outside source to be legitimate. The message, the text or the ad is
where the rhetor’s creative liberty is exercised, in the explicit use of myth to create of indi-
viduals a community. Myth is exercised in the cultural significance of ads ( Johar, Holbrook
& Stern 2001). Images draw upon socio-cultural meanings to create identity, and reframe
these meanings as inherent to the product (Goldman 1992). Myth provides the parameters
of community membership, and represents those who adhere to a moment as generations
in and of themselves. Their identity is contained in the definition of myth – the ideology,
the axioms, the common threads of interest and being – and transgresses the usual quali-
fiers of age, geography, gender or income. The enactment of the myth is the establishment
of the rhetor’s audience. This community is reflexive, existing only by ways of the myth and
only in so far as the myth lives in their belief system. Persuasion depends on commonality
of thought, and a shared encoding and decoding process between the rhetor and the audi-
ence. The community is not a product of the myth, but a response to it, filtering it through
experience before turning it into conviction. The myth stresses a rethinking of the situation,
permits a judgement, and allows for members to band together around a cause.
Constitutive rhetoric
Constitutive rhetoric is the process of composing a collective subject; it is a theory that
understands the rhetor, audience and message as mechanisms in this process. This col-
lective is established through the positioning of the subject as transhistorical, and through
49
offering the audience the illusion of freedom, while constraining them to the narrative
that calls them into being (Charland 1987). For constitutive rhetoric to operate effec-
tively, members must first be successfully interpellated, and second adhere to the action
necessitated by the tautological logic of the message. Audiences are always here (Althusser
1994), interpellated through a process of identification with a textual position. To enter a
rhetorical narrative is to identify with Black’s (1970) second persona; this occurs through
a series of ideological associations arising from the narrative structure of the message
(Charland 1987). Once constituted as a collective, the narrative activates the members
to act and motivates them towards certain ends. For McGee (1975), ideology is material,
existing both in the realm of ideas and practices, animating members to act upon it and
structure their material world around its very principles and images.
Constitutive rhetorics are ideological because they provide members with narrative to
inhabit as subjects and motives to experience, but also because they insert ‘narrativised’
members-as-agents into the world – in particular, the constitution in action of a motivated
member that orientates those addressed towards particular future acts (Charland 1987,
p. 143). Interpellating members and calling them to action, the discourse of advertising
constitutes viewers as ‘deficient’ in some quality, attribute or value, which is happily rem-
edied through the consumption of material objects. The constitutive action of an ad occurs
in the transition of a consumer from the implied to the actual, from the real-world body
of individuals receiving a message, to the implied audience of that message. ‘[The implied
reader] differs from an actual reader in that he is created by the work and functions as
the work’s ideal interpreter. Only by agreeing to play the role of this created audience for
the duration of the reading can an actual reader correctly understand and fully appreci-
ate the work’ (Sulieman 1980, p. 8 as cited by Stern 1984, p. 11). Consumer goods carry
and communicate cultural significance, and advertising is a mechanism for and instru-
ment through which cultural meaning is transferred from consumer goods to individual
consumers (McCracken 1986). Constitution occurs with identification. In actualising the
roles prescribed in the implied audience, the actual consumer enters in, and forges, a com-
munity centred on the consumption of goods and the cultural meaning embedded therein.
A formative model for brand communities

The first step in calling a brand community into being, and offering it direction, is to
establish a strong character for the brand itself. This conceptualisation is then forwarded
to the consumer. Brands create personas, for themselves and their implied audiences, who
embody the messages and myths that define the brand. Thus, we propose:
P1: In order to direct the formation of a brand community, brands must first seek to
outline a persona for themselves (First Persona) and for their implied audiences
(Second Persona).
Once personas are established for the brand and its audience, these dispositions and roles
must be communicated to the audience for recognition, identification and empowerment
50
to take place. The process of identification for a collective is similar to that of interpersonal
brand relationships. Not unlike that between a brand and a consumer, the consumer–brand–
consumer relationships of brand communities are established through persuasion and the
action of branding. Advertising is a ‘conduit through which meaning constantly pours from
the culturally constituted world of consumer goods’ (McCracken 1986, p. 75). Where it is
understood as a reference body for cultural meaning, practices and norms, as well as con-
sumer expectations, persuasion is a result of this communication effort:
P2: Advertising messages are imbued with myths, articulating ideological frame-
work, brand personality, character traits and cultural meaning.
When a community appropriates the myths and ideas in a given message, they are
called to action to assume this role through consumption and beyond. This process confers
an identity to consumers, who are vested with a mission and then assembled as a collective
with co-consumers of the same mindset:
P3: In identifying with the message, members assume the role and expectations of
Second Persona, and are empowered as a brand community with those who
engage in the same act.
Called into being and persuaded by the cultural meaning adopted through the rhetoric
of marketing communications, individuals emerge as a collective in a consumer culture
flourishing only when the proper tools are at their disposal:
P4: The process of constitutive marketing is iterative and cyclical, insofar as without
constant nurturing, brand community development and growth will be stagnated.
Based on these propositions, the right side of Figure 1 depicts the proposed links between
the brand, the myth and the audience in the developmental process of brand community.
We then apply this model to three case studies: Apple, Nike and Harley-Davidson.
Case study 1: Apple

Apple has the highest brand loyalty of any computer manufacturer. Its product line is
more than just ‘Conversational Capital’ (Cesvet et al. 2009); it is a brand brought to life,
and a veritable religion to many of its consumers (Muniz & Shau 2005; Eaton 2011).
The Apple community is marked by all the ‘key markers’ of brand communities – image,
history and competitors – however, the emergence of this community is best understood
as a calculated exercise in constitutive marketing. Observers of the Apple community are
familiar with its founding myth:
This hallmark of independent thinking and reinvention was started by two young
men with no money, but plenty of good ideas … we hope that we are as creative,
inventive and free as Steve Jobs and Steve Wozniak. (Cesvet et al. 2009, p. 92)
51
The development of the Apple community began with the ‘1984’ television ad, which
launched the highly successful Macintosh computer.
‘1984’
The ‘Cult of Mac’ originates from the Macintosh introduction in January 1984, when
Apple urged the world to ‘Think Different’. The ad was a historical event. Broadcast only
once during Super Bowl XVIII, it launched a personal computer, and constituted a com-
munity of non-expert computer users who abandoned the technological imperative driv-
ing capitalistic socio-economic progress. The ad opens with a woman running through
a dark hallway to a monotonous beat. A sketch of a Macintosh computer is emblazoned
on her white shirt, and her red shorts set her apart from the uniformly dressed, herd of
grey-clad workers marching in unison, hypnotised by Big Brother, who addresses them
from a cinema-sized screen. Echoing statements like ‘one whim, one resolve, one course’,
the ad portrays a pre-Mac world with little choice, and conformity. Breaking through
the monotony and homogeny, the heroine tosses her hammer with the strength of an
Olympic athlete to shatter the screen, and by extension the old regime. As the screen
explodes, the room is lit and the crowd awakened as an off-screen narrator proclaims the
revolution: On January 24th Apple Computer will introduce Macintosh. And you will see why
1984 won’t be like ‘1984’. The success of this ad reached well beyond the product offer-
ing. Of course, the Macintosh was the first commercially successful personal computer to
feature a mouse and a graphical user interface, but the methods used to communicate the
added value and beyond-product offering were unparalleled.
‘Think Different’
In 1997, market share was 3% (Levy 2009, p. 92). While the foundation for the brand
community was laid, the loss of continuity from the initial articulation of the myth and the
loss of the myth’s founders did little to build Mac market share or strengthen the brand
community. With the return of Steve Jobs in 1997, Apple began to once more ‘Think
Different’. The campaign was launched with a television ad that was iconically Mac.
The ‘Crazy Ones’ featured images of influential figures throughout the century – Albert
Einstein, John Lennon, Amelia Earhart and Pablo Picasso, to name a few – and reinforced
how, not unlike these individuals who changed the world, Mac users ‘think differently’.
The ad marked a revitalisation of the Apple community, with a new generation of mem-
bers ready to take a stand against the capitalist way of thinking and doing.
‘Get a Mac’
Building upon the persona of the Mac user, the campaign reaffirms his role, status
and imperative. Each ad follows the same template, and reinforces the mantra of non-
conformity and elitism. It begins with Justin Long, portraying the hip, urban professional
Mac, and John Hodgman, standing in as the nebbish and more formal PC, introducing
52
themselves, then delves into an ‘us-versus-them’ strategy that witnesses the PC as ineffective
and Mac triumphing as the cool and humble victor, and closing with the white Mac com-
puter. Expressing both product-level and brand-level attributes, this ad serves as a nurturing
tool for the Apple community, reinforcing the original myth that Mac users are creative and
well educated, and can obtain the perfect balance between work and life (Fried 2002).
Case study 2: Nike

Started in 1965 as a small distribution company for Japanese-made Tiger shoes, the Nike
brand and swoosh were developed in 1971 – but only with the 1988 ad campaign did the
legend begin.
‘Just Do It’
‘It’s not just a shoe, it’s a community’ (Hesseldahl 2009). First broadcast in 1988, ‘Just Do
It’ fashioned a community of consumers, and resonated with an athletic audience as more
than a simple ‘power line’ (Cone 2008). Quick, bold and concise, ‘Just Do It’ serves as
the connective device and the summation of Nike’s personality and persona (Hunsberger
2008). ‘If you have a body, you are an athlete’ (Nike, Inc. 2009). The eight-letter phrase
‘sp[oke] to the restraint and inhibition in everyday life that keeps people from the experi-
ence of transcendence’ (Goldman & Papson 1998, p. 19), and served as an appeal for
consumers ready to take on the challenge of achieving their personal best.
‘Bo Knows’
Aired during the 1989 Major League Baseball (MLB) All-Star Game, the ‘Bo Knows’
campaign forwarded the ‘Just Do It’ mantra in a hip hail to athletes of all abilities and
disciplines. With Bo Jackson, the ad promoted a Nike cross-training shoe, and was tar-
geted at athletes who weren’t afraid to take on, and excel at, new and multiple things. It
proclaimed Jackson as an ‘athletic authority’ against athletes considered the best in their
domains, each proclaiming: ‘Bo knows’. It was more than a collection of celebrities shep-
herding their fans towards Nike – it set the tone for Nike. With its celebrity power and
tongue-in-cheek attitude, it curbed the distance between the world of top athletes and
that of the weekend warrior. The myth behind the campaign demonstrated grit, passion
and determination, serving as an appellation not nearly as overt as a simple ‘hey you’, yet
clear and uncomplicated. Through ‘Bo Knows’, the maxim became ‘cemented in consum-
ers’ minds as a rallying cry to “get off the couch, and play sports”, calling them … to just
act, to Just Do It’ (Goldman & Papson 1998, p. 19).
Michael Jordan
In 1997, Nike brought Jordan back in a play-off commercial demonstrating that even
Michael Jordan was human: I’ve missed more than 9,000 shots in my career. I’ve lost almost
53
300 games. I’ve failed over and over and over again in my life. And that is why I succeed.
Reminding its community that the greats are only doing their best, Nike reaffirmed its
mission and character in this humbling ad. Building upon the persona of the diehard,
never-quit athlete ready to take on a challenge and surpass it, these ads created a con-
tinuum in Nike’s efforts to build a solid brand community of determined and diverse
athletes.
‘Pretty’
‘Pretty’ continues to demonstrate the multifaceted character of its athletes, all the while
calling upon cultural code and significance to illustrate its message. It opens with Maria
Sharapova getting ready in front of a mirror, as a radio presenter talks about her chances
of winning and marvellous serve. As she exits her hotel room, the instrumental to the song
‘I Feel Pretty’ begins and is punctuated as she delivers a powerful unanswered return on
her opponent’s serve. The crowd silenced by awe, she positions herself for the next rally as
the ad ends with the swoosh and ‘Just Do It’. The epitome of a Nike athlete, Sharapova
is brawn and beauty.
Case study 3: Harley-Davidson (H-D)

Perhaps the founding of a brand community is not explicit in the development of H-D;
however, the act of consumption bled into the complexion of community though careful
nurturing of the H-D image, history and consumer base: ‘Harley-Davidson is a global
leader in fulfilling dreams and providing extraordinary customer experiences through
mutually beneficial relationships with our stakeholders’ (Harley-Davidson 2009). The
Harley Owners Group (HOG) encourages riders to ‘express [themselves] in the com-
pany of others’, and events on the local and international levels solidify bonds among
members, driving them towards ‘making the Harley-Davidson dream a way of life’
(Harley-Davidson 2009).
Selling individualism through camaraderie and brotherhood, the myth urges participa-
tion, and calls Harley riders to action with the promise of power and the joy of the ride.
The H-D archives span more than 100 years. While there have been ads created for radio
and television, H-D continues to be in print media. In the true spirit of brand community,
this targeted reach is in an effort to ‘advertise directly to its potential buyers’ (Wright 1983,
p. 219). Harley-Davidson’s earliest known ad appeared in the 2 January 1905 issue of Cycle
and Automobile Trade Journal (Supple 2002, p. 4). Lifestyle, persona and character were not
introduced until later, where fun finally trumped facts.
‘Get a Kick out of Life’

The ads that began to build and foster the ‘sense of Harley’ appeared in 1925. Campaigns
such as ‘Get a Kick out of Life’ resonated with Harley’s market, advocating freedom,
espousing individuality and relishing the American dream. The copy is ‘Get a kick out
54
of life with a Harley-Davidson’, and depicts a carefree couple riding on a motorcycle and
sidecar, enjoying the countryside. These components construct the Harley persona, and
demonstrate how consumers can develop and actualise this carefree, down-to-earth life-
style and vantage point. This trend in key claim and communication priority continues
throughout the first half of the 20th century, which laid the early tenets of development
for the brand community.
Love Affair
Throughout the 1960s and 1970s, fun became the standard for H-D ads, revitalising the
business with the same medicine that got it thriving. Catchy phrases like ‘Stand out in
’60 with a new Harley-Davidson’ (1960), ‘For a funderful you in ’62. For work, for school,
for play’ (1962), and ‘Ride the road to real adventure’ (1964) demonstrated pride in the
emotional over the rational. While the Harley ads hinged upon good times and easy rid-
ing, H-D merged with American Machine and Foundry (AMF) in 1969. Stripping H-D
of its independence, it did not change strategy – autonomy and fun. In the 1970s, the call
to action embedded in H-D print ads ushered consumers to ‘Pull the trigger’ (1971) and
hold on tight to the power of Harley, while underscoring the uniqueness of the brand,
in declaring that ‘Only one man could have done this’ (1977). With copy that weaves
product-level attributes and elements of brand persona, the ad is a cocktail of interpella-
tion that seeks to entice current consumers and a new generation of Harley riders.
‘The Eagle Soars Alone’

The AMF ownership was a difficult time. H-D owners and enthusiasts felt a loss of free
will and an inability to guide their own destiny. In 1981, the brand was reborn as senior
management participated in a leveraged buyback. This change saw a revision in quality
management, but stood for the return of H-D to its roots. The new ad is simple: at the
top, ‘THE EAGLE SOARS ALONE’ is proclaimed, in black text. At the foot, the time
stamp spells out what the layperson may miss: Commemorating the return to private owner-
ship of Harley-Davidson Motor Company. Ironically, the ad proclaims freedom, while also
underlining the importance of brotherhood in its connotative definition. An icon of the
American dream, the eagle stood for Harley’s appropriation of that dream, and for the
freedom, equality and brotherhood that embodied the brand. The eagle, like the H-D
rider, had a distinct edge and communicated the power and quest for adventure that lay
at the core of the company value system and persona.
Contributions of the model to understanding brand

community formation
The Apple, Nike and H-D archives show similar evolutionary patterns. At the core of each
campaign lies a myth that embodies a strong history, a distinct persona and a powerful call
to action that provides freedom and rebellion, while also ushering inclusivity in the brand
55
community. Each embodiment of first and second persona, while bearing traits relative to
the industry, ultimately speaks to freedom, and liberation from a greater, darker power –
Apple, from the overbearing PC; Nike, from the negative mantra of ‘It can’t be done’; and
H-D from the drudgery and conventions of life. The humble myths speak of underdogs
with visions, individuals who unchain themselves to transform dream into reality; and the
ads articulate these sentiments in overt and explicit means. Where Apple’s ad repertoire
evolved based on cultural context, Nike grew with its athletes, and H-D with the climate
and tone of the era. All three brands demonstrate an articulation of the model in action,
however differ in the dissemination of their messages; Apple and Nike took mass-media
approaches, and H-D used a more tailored reach. The following sections outline each
brand’s case in light of the model and propositions.
Apple
For Levy (2009), Apple has not lost its edge, being a ‘Mac person’ remains a strong state-
ment: a powerful ideology, launched with the ‘1984’ campaign, and careful and meticulous
nurturing through subsequent ads were imperative to the livelihood of the Apple commu-
nity. ‘If a product is successfully tied to an idea, branding persuades people … to consume
the idea by consuming the product’ (Walker 2008, p. 8). In their one-to-one interactions,
members’ shared consciousness is in sharing and calling upon the same myth.
Figure 2 illustrates the Apple case in light of the model. Indicative of P1, composing
and viewing themselves as ‘silicon artists’ and ‘a new kind of digital hipster … determined
to take down the faceless giants’ (Levy 2009, p. 92), the minds behind Apple constructed
their first persona and then forwarded the principle of this implied author. Imbuing the
persona with an ethos – a true moral character, and against repression – and a strong intel-
lectual virtue in its dianoia, this image was funnelled in to the implied audience, and used to
construct a second persona and myth. Weaving a myth shrouded in the ideology of change
and rebellion, and ‘identifying a threat to its very existence as a narrative’ (Charland 1987,
p. 146), ‘1984’ called upon a history of oppression and a David-vs-Goliath confrontation
to usher in a future of freedom. ‘Audiences are constituted as subjects through a process
of identification with a textual position. This identification occurs through a series of
ideological effects arising from the narrative structure of constitutive rhetoric’ (Charland
1987, p. 147). Substantiating P2, the cultural meaning and character codes articulated an
ideology and position with which members could identify. Hinging upon a sense of dread
for the future, ‘1984’ introduced the idea of the Apple user as freedom fighter and beacon
of hope. The Apple brand community was alive within the public, but it was energized
and given a cohesive breath of life by the ad and its message.
Establishing grounds for confirmation of P3, following this ad many users identified
themselves with the Mac myth and saw themselves as not unlike the ‘silicon artists’ craft-
ing the new Macintosh. The message, delivered in the dramatic enactment mode, guided
a collective of Macintosh users, branding them with the same identity as its founders, and
articulating, through its ad campaign, this identity and ideological framework of resist-
ance. It is the third ideological effect, the constitution in action of a motivated subject that
56
Figure 2: The Apple brand community, as understood via the proposed model
First persona
Myth
Geniuses/silicon artists
Brand Going against the
Apple grain/rebellion
against conformity
Second persona
Hip/cool
Advertisement
‘1984’/
Think different/
‘Get a Mac’
Brand Audience
community
orientates those addressed towards future acts. Since narrative offers totalising interpreta-
tions that ascribe transcendent meaning to individual acts, the maintenance of narrative
consistency demands that a certain set of acts be chosen (Charland 1987, p. 143). The
tautological logic of the message – traditional computers are repressive, Macs are liberat-
ing, be liberated and buy a Mac – provided the public with the proper motivation. While
many identified with the myth, and were called to action by its proponents, market figures
indicated that most consumers were not of this group. For these reasons, Apple continued
to build its brand community and reinforce its original efforts in subsequent campaigns.
While efforts fell short between 1985 and 1997, momentum was re-established when the
Apple focus was realigned with its creator and overall vision. Thinking of P4, the ‘Think
Different’ and ‘Get A Mac’ campaigns not only revitalised the original tenets of brand
community set in place in 1984, but also served as unique acts of constitutive rhetoric
themselves, interpellating new generations of consumers into the Cult of Mac. Apple
proved to be a top competitor in terms of its ability to command attention and enlist
engagement from its consumers in online realms (Klaassen 2009). This strength is paral-
leled in its offline efforts as well, where Apple continues to dominate in sales and earnings.
Nike
‘Just Do It’ was named by Advertising Age as one of the top five slogans of the 20th cen-
tury; the campaign is in the Smithsonian National Museum as a ‘true part of American
57
history’ (Nike, Inc. 2009). ‘Nike’s swoosh is a commercial symbol that has come to stand for
athletic excellence, a spirit of determination, hip authenticity and playful self-awareness’
(Goldman & Papson 1998, p. 1). To illustrate history, and situate its message, the Nike
ad repertoire is rife with cultural allusions and a referential system of meaning that calls
upon this core persona of an athlete on a mission. Appropriating culture, the Nike myth
addresses its audience with a ‘knowing wink’; furthermore, ‘Just Do It’ is empowering, and
speaks to the restraint that inhibits transcendence in everyday life (Goldman & Papson
1998, p. 19). The slogan represents the firm’s spirit, and its very essence allows for a mul-
tidimensional community of sports enthusiasts that come from all realms of athleticism
and all levels of discipline.
Figure 3 illustrates the Nike case in light of the model. Nike positioned itself as a brand
committed to, and governed by, victory, as suggested in P1. A band of athletes who saw
no obstacle as too big to keep them from achieving their goal, and attaining greatness,
the minds behind Nike created a First Persona akin to the goddess from whom their
moniker was borrowed. Interlacing this First Persona with an implicit call to action, and
a moral spirit that embraced a challenge, the persona was symbolised in the swoosh, and
summarised in the slogan. In this image, the Second Persona, the Nike athlete, was born.
Demonstrating the relationship between the brand and persona, confirming P2, through
its ads, Nike began the process of ‘[inscribing] real social actions within its contextual
structure of motives and [inserting] them into the world of practice’ (Charland 1987,
p. 142). ‘Bo Knows’ offered a voice to a community of ‘always ready’ subjects. Laden with
the notion of personal best, triumph and excelling at something new, the ad brought to
life the taunt of ‘anything you can do, I can do better’, and breathed life into the multi-
dimensional, cross-training Nike athlete – the athlete who was victorious, driven by the
challenge, yet humble; a person turned machine in the heat of the game. Verifying the
logic of P3, the meaningful and powerful message, and brand story rooted in a history of
rising to the occasion, was adapted into the personal narratives of the audience, initiating
their assumption of the role of the Second Persona. Charland (1987) articulates that ‘to
be an embodied subject is to experience and act in a textualized world’ (p. 144), this world
is one where identification and awakening occur at first interaction with the myth, and
initiation occurs at consumption. While the ambiguous call to action of ‘Just Do It’ could
be lent to various sports, goals and fitness objectives, one thing was undeniable: to Just Do
It, you had to Just Do It with Nike. Through rough patches in the 1990s, Nike continued
to thrive thanks to its brand community. In efforts to strengthen its base, and interpel-
late new members, Nike reverberated its myth and rallying call in ads like the Michael
Jordan series and ‘Pretty’, confirming P4 and the cyclical nature of the brand community
process. Demonstrating the necessity of continuity, through its various campaigns, Nike
established itself as an industry leader.
Harley-Davidson
H-D is much more than a motorcycle. The most powerful brands are those ‘that have
built their own mythology, or rather that have helped their loyal customers to build
58
Figure 3: The Nike brand community, as understood via the proposed model
First Persona
Top athletes/
Brand multifaceted Myth
Nike Athletic virtue/
victory
Second Persona
Driven/diverse
Advertisement
‘Bo knows’/
Michael Jordan/
‘Pretty’
Brand Audience
community
their mythology’ (Kevin Roberts, cited in Haig 2003, p. 77). Since 1903, H-D has pro-
duced the ‘most enduring and recognizable motorcycles in the world’ (Harley-Davidson
2009), and the very same can be said for the ad campaigns that accompanied these
products. Articulating the H-D myth of freedom and the open road, the ads piloted a
premier brand community of loyal enthusiasts willing to share their passion and show
their pride.
Figure 4 illustrates the H-D case in light of the model. The H-D First Persona,
and then Second Persona, are formed around the homegrown hero. Master of his own
domain, the H-D First Persona is a brother free from the chains of ‘Big Brother’ when
on his Harley. Demonstrating P1, it is with this image in mind that the myth of H-D
was spun. Laced with themes of brotherhood, freedom and power, the H-D myth is
difficult to articulate in that it is open for interpretation and adoption to a wide body
of consumers: ‘It’s like a fire, you can pull and use a color from the fire, but if you just
try to grab the whole thing, you’re going to get burned’ (Clyde Fessler, cited in Wright
1983, p. 223). Establishing a mystique about its product, H-D ensured that the only way
to experience the Harley myth is by riding a Harley. The focus of the H-D ad has not
always been myth centred: ‘We tried for a long time to convince people that motorcycles
had some utility value. Motorcycles have never been anything but pleasure’ (William H.
Davidson, cited in Wright 1983, p. 219). While originally the H-D ads spoke of practical
value, this focus shifted to integrate the H-D persona as the voice of media efforts. This
shift underscores the importance of P2, the fusing of the ad with myth to elicit brand
59
Figure 4: The Harley-Davidson brand community, as understood via the proposed model
First Persona
Homegrown hero/
Brand epitome of masculinity Myth
Harley-Davidson Freedom/power/
brotherhood
Second Persona
Master of his
own domain
Advertisement
‘Get a kick out
of life’/Love affair/
‘The eagle soars alone’
Brand Audience
community
experience and audience adherence. Once this change in direction was undertaken, the
ads began to espouse and expose the history and myth behind H-D. Articulating P3, the
H-D community flourished under the firm guidance and committal to the H-D myth.
The community persevered and persisted to build under AMF ownership, as ads contin-
ued to advance and feature the spirit of H-D, demonstrating the iterative nature of the
model. The formation of the HOG (1983) marks the height of the H-D community. A
de jure manifestation of what was de facto, the HOG served as a concrete indication of
the H-D community. Subsequently, brand-fests such as the pilgrimage to Milwaukee (‘the
ride home’) and H-D museum and factory tours demonstrate the potency of the brand
community and the value of a strong base of brand evangelists and enthusiasts. A 1981
study found that a H-D owner was a confirmation of the H-D Second Persona: ‘one
who sees himself as an individualist, but is a conformist who needs a sense of brother-
hood and camaraderie’ (Wright 1983, pp. 223–224). These statements are true for today’s
H-D rider, although the demographics have changed over the years. In the H-D case, the
principles of P4, while applied, are not full demonstrations of renewal in action. While
programmes like ‘Riders Education’ invite new members into the Harley community, the
pointed reach of the ads allows for little encouragement to new audiences. From 1987 to
2006, the median age of the H-D community, and specifically HOG, leapt from 35 to
47; while H-D riders are loyal, recruitment is low, and the brand has become less of an
attraction to new riders and more of a ‘middle-aged, nostalgia brand’ (Weber 2006). The
overall trend is a decline – both revenues and net income. Juxtaposing the H-D metrics,
60
versus those of Nike and Apple, it can be suggested that Harley casts a wider net and
exposes the application of the model to a larger audience.
Limitations and future research

The discourse of meaning making in advertising provides an interesting opportunity
for research in marketing communications. Our model of brand community formation
describes the brand community as a brand-initiated and nurtured organism. Drawing
upon the concept of constitutive rhetorics and the formation of a collective through
mythic messages, it describes interactive consumers (Stern 1994) acting in a collective, and
the transformation of the actual consumer to the imagined consumer through the opera-
tion of text. The study demonstrates the viability of the model through a historical case
analysis of the Apple, Nike and H-D communities, using major ad campaigns; however
this requires a more in-depth examination before it can be fully applied.
Future research should first seek to empirically test the model elements in order to
determine their concrete applicability and whether they exist as causal relationships.
Given the direction of the current study, neither ethnographic, grounded theory nor lon-
gitudinal studies were within the scope of exploration, but could prove to be interesting
means by which the model could be tested more exhaustively, moving it from theoretical
to practical.
The expansion of the model is also an interesting avenue for research, as is examining
in more detail mediators and moderators of the relationships. Specifically, the media used
for communication, as well as the various modes of message framing, could be articulated
within the model, as they could result in different audience responses. The model should
additionally be tested for differences that might emerge given the product or brand
type and its position in the product life cycle. For example, the model could prove to be
effective for new-to-market products and services, or those heralded as revolutionary to
their class, but less potent for line extensions. Moreover, low-involvement purchase items
may not merit or be able to sustain brand communities as well as high-involvement or
big-ticket products. It would be beneficial to test the model based on consumer-level
differences, such as the level of consumer engagement within the brand community, the
position of the consumer in the brand community life cycle, and the congruence between
brand community behaviour and the consumer’s everyday lifestyle.
Further to consumer-level variants, as our study explores the formative process of brand
community from a brand-engineered perspective, our model should be expanded by the
specific and detailed analysis of the role of consumer agency. As articulated, audience
agency plays a large role in brand community formation, but also continues throughout
all stages of the brand community life cycle. Muniz and Schau (2005) underscored the
grassroots potential and intense religiosity that brand communities can foster in their
study of the abandoned Apple Newton brand. Concluding that the brand community
continued to operate, bearing themselves the role of the brand in their maintenance and
development activities, they demonstrated the power that consumers play in the brand
community process.
61
Conclusion
Brand communities are more than fan clubs, more than social networks. The brand
community represents the dovetailing of both self-portrayal and relationship aspects of
brand value, and contributes to the brand value through its social positing and power. It
is therefore imperative to include the brand community in the discussion of brand equity
and the valuation process. Our study expands on the concept of brand community from a
brand-engineered and directed perspective. In doing so, the theoretical tenets proposed by
this model offer important insights for future research, as well as initial steps for market-
ers seeking to manage and direct their brand community efforts. While marketers must
always tread lightly when choreographing their advertising efforts, and balance brand
objectives with the consumer climate, our study offers insight as to how brands can build
and shape the foundations of a brand community.
Acknowledgements
The authors wish to thank the editor, Dr Charles R. Taylor, the associate editor and the
manuscript’s three anonymous reviewers for their insightful and constructive comments
and suggestions.
References
Aaker, J.L. (1997) Dimensions of brand personality. Journal of Marketing Research, 34(3),
347–356.
Aggarwal, P. (2004) The effects of brand relationship norms on consumer attitudes and
behavior. Journal of Consumer Research, 31(1), 87–101.
Algesheimer, R., Dholakia, U.M. & Herrmann, A. (2005) The social influence of brand
community: evidence from European car clubs. Journal of Marketing, 69(3), 19–34.
Althusser, L. (1994) Sur la philosophie. Paris: Gallimard.
Black, E. (1970) The Second Persona. Quarterly Journal of Speech, 56, 109–119.
Braun-LaTour, K.A. & LaTour, M.S. (2004) Assessing the long-term impact of a consistent
advertising campaign on consumer memory. Journal of Advertising, 33(2), 49–61.
Carlson, B.D., Suter, T.A. & Brown, T.J. (2008) Social versus psychological brand community:
the role of psychological sense of brand community. Journal of Business Research, 61(4),
284–291.
Cesvet, B., Babinski, T. & Alper, E. (2009) Conversational Capital: How to Create Stuff People
Love to Talk About. Upper Saddle River, NJ: Pearson.
Charland, M. (1987) Constitutive rhetoric: the case of the peuple québécois. Quarterly Journal of
Speech, 73, 133–150.
Cline, T.W., Altsech, M.B. & Kellaris, J.J. (2003) When does humor enhance or inhibit ad
responses? Journal of Advertising, 32(3), 31–45.
Cone, S. (2008) Powerlines: Words That Sell Brands, Grip Fans and Sometimes Change History.
New York: Bloomberg.
De Pelsmacker, P., Geuens, M. & Anckaert, P. (2002) Media context and advertising
effectiveness: the role of context appreciation and context/ad similarity. Journal of
62
Eaton, K. (2011) And on the eighth day, God created Apple. Fast Company, 19 May.
Eco, U. (1987) Introduction (0.1, 0.2). In: Lambropoulos, V. & Miller, D.N. (eds) Twentieth
Century Theory: An Introductory Anthology. Albany: State University of New York Press,
423–436.
Fried, I. (2002) Are Mac users smarter? CNET news. Available online at: http://news.cnet.
com/2100-1040943519.html (accessed 12 July 2002).
Garretson, J.A. & Neidrich, R.W. (2004) Spokes-characters. Journal of Advertising, 33(2), 25–36.
Goldman, R. (1992) Reading Ads Socially. London: Routledge.
Goldman, R. & Papson, S. (1998) Nike Culture. Thousand Oaks, CA: Sage.
Haig, M. (2003) Brand Failures. London: Kogan Page.
Harley-Davidson (H-D) (2009) Available online at: http://www.harley-davidson.com (accessed
4 September 2009).
Hesseldahl, A. (2009) Apple’s impressive quarterly numbers. BusinessWeek, 22 January.
Hunsberger, B. (2008) Nike’s ‘Just Do It’ slogan celebrates 20 years. The Oregonian, 18 July.
Johar, G.V., Holbrook, M.B. & Stern, B.B. (2001) The role of myth in creative advertising
design: theory, process and outcome. Journal of Advertising 30(2), Summer, 1–25.
Klaassen, A. (2009) Most social brands of 2008. Ad Age, 29 January.
Levy, S. (2009) All grown up. Wired, January, 90–93.
McAlexander, J.H., Schouten, J.W. & Koening, H.F. (2002) Building brand community. Journal
of Marketing, 66(1), 38–54.
McCracken, G. (1986) Culture and consumption: a theoretical account of the structure and
movement of the cultural meaning of consumer goods. Journal of Consumer Research, 13(1),
71–84.
McGee, M.C. (1975) In search of ‘the people’: a rhetorical alternative. Quarterly Journal of
Speech, 6l, 235–249.
McLuhan, Marshall (1962) The Gutenberg Galaxy: the Making of Typographic Man. Toronto,
Canada: University of Toronto Press.
Muniz Jr, A.M. & O’Guinn, T.C. (2001) Brand community. Journal of Consumer Research,
27(March), 412–433.
Muniz Jr, A.M. & Schau, H.J. (2005) Religiosity in the abandoned Apple Newton brand
community. Journal of Consumer Research, 31(4), 737–747.
Nike, Inc. (2009) Nikebiz: NIKE, Inc., official site, the world’s largest, leading athletic brand.
Available online at: http://www.nikebiz.com (accessed 4 September).
Perelman, C. (1969) The New Rhetoric: A Treatise on Argumentation. Chicago, IL: University of
Notre Dame Press.
Schembri, S. (2008) Reframing brand experience: the experiential meaning of Harley-
Davidson. Journal of Business Research, 62(12), 1299–1310.
Stern, B.B. (1994) A revised communication model for advertising: multiple dimensions of the
source, the message and the recipient. Journal of Advertising, 23(2), 5–15.
Sulieman, S.R. (1980) Introduction: varieties of audience-oriented criticism. In: Suleiman, S.R.
& Crosman, I. (eds) The Reader in the Text: Essays on Audience and Interpretation. Princeton:
Princeton University Press, 3–45.
Supple, J. (2002) 100 Years of Harley-Davidson Advertising. Boston: Bulfinch.
Toncar, M. & Munch, J. (2001) Consumer responses to tropes in print advertising. Journal of
Walker, R. (2008) Buying In: The Secret Dialogue Between What We Buy and Who We Are. New
York: Random House.
Weber, J. (2006) Harley just keeps on cruisin’. Businessweek, 6 November.
Wright, D. (1983) The Harley-Davidson Motor Company: An Official Eighty-Year History.
Osceola: Motorbooks International.
63
About the authors

Ana Kilambi is Senior Production Manager at SAY Media (Montreal, Canada). She
received her MSc in Marketing from Concordia University, Montreal (Canada). Her
areas of interest and expertise are brand engagement, communication theory and digital
marketing strategy.
Michel Laroche is the Royal Bank Distinguished Professor of Marketing, John
Molson School of Business, Concordia University, Montreal (Canada). He received his
PhD from Columbia University (New York). His main research interests are in con-
sumer behaviour, marketing communications and Internet marketing, services marketing,
and retailing. Within consumer behaviour, he is mostly interested in the role of culture
and brand decision processes. He has published more than 270 articles in journals and
proceedings, including Journal of Consumer Research, Journal of Advertising Research,
International Journal of Advertising, Journal of Business Research, Journal of Retailing, Journal
of the Academy of Marketing Science, International Journal of Research in Marketing and
Journal of Service Research.
Marie-Odile Richard is a Banting Post-doctoral Fellow in neuromarketing, John
Molson School of Business, Concordia University, Montreal (Canada). She received her
PhD in Marketing from HEC-University of Montreal (Canada). Her research inter-
ests are in marketing communications (including Internet and social media marketing),
neuromarketing, services marketing, and cultural effects on individual responses. She
has published more than 32 articles in journals and proceedings, including Journal of the
Academy of Marketing Science, Journal of Advertising Research, Journal of Business Research
and Journal of Social Psychology.
Address correspondence to: Michel Laroche, Royal Bank Distinguished Professor of
Marketing, John Molson School of Business, Concordia University, 1455 de Maisonneuve
West, Montreal, Quebec H3G 1M8 Canada.
Email: laroche@jmsb.concordia.ca
64
The differential effects of position, ad
and reader characteristics on readers’
processing of newspaper ads
Edith G. Smit and Peter C. Neijens

University of Amsterdam
Robert Heath
Bath School of Management
Building on previous research on the processing of newspaper ads, this comprehensive field
study, with 26,556 newspaper readers and 290 unique advertisements, investigated the com-
bined effects of position in the newspaper, ad characteristics and reader characteristics. The
results show a differential effect of stimulus-specific and person-related factors: with increasing
levels of ad processing, from passive to active, the influence moved from position in the news-
paper, via mechanical ad characteristics to message content and reader interest. The findings
affirm that advertisers who want to reach their audience need to first consider placement; but
to encourage actual processing of the ad they need to make the ad outstanding and relevant.
Introduction
What are the effects of an ad’s position in the newspaper, ad characteristics and readers’
characteristics on how readers process newspaper ads? An answer to this question is of
crucial importance for both advertising theory and practice. Surprisingly, the idea that
different factors have different influences on the processing of advertisements has not
really been subjected to any rigorous testing. Valuable research has been conducted into
factors influencing ad processing (e.g. Rossiter 1981; Stewart & Furse 1984, 1985; Finn
1988; Stewart & Koslow 1989; Nordfalt 2005), however these factors have never been
combined into a single framework, which prevents a real estimation of the influence of
different factors on different levels of ad processing. Methodological issues in previous
studies also make it difficult to provide an unambiguous answer. Many ad processing
studies are based on copy test research, in which respondents are forcefully exposed to
ads stripped from their natural media context. Both factors limit the external validity
of the studies. Another problem is that the limited number of advertisements and titles

DOI: 10.2501/IJA-32-1-065-084
included in these studies might have resulted in context-dependent results (Pieters &
Wedel 2004).
This paper studies a full range of variables related to ad processing among a representa-
tive sample of a great number of ads and newspaper titles, in a real-life context without
forced exposure and with ads included in the original publications, just as they were
published. Our theoretical starting point lies in Greenwald and Leavitt’s four-level model
of audience involvement in advertising processing, which we extend with theories that
hypothesise a differentiating role for bottom-up factors versus top-down ones (Greenwald
& Leavitt 1984; Pieters & Wedel 2004; Leigh et al. 2006).
The paper proposes and tests an integrated model that specifies a differential effect
of bottom-up and top-down factors on different levels of advertising processing. By
conducting a comprehensive study that simultaneously includes position, ad form,
ad content and reader characteristics, we are able to assess these effects corrected for
interaction. Furthermore, by including a large number of newspaper titles, advertise-
ments and respondents, we are able to generalise research findings. In addition, a real-
life study allowed us to overcome limitations due to forced exposure and evaluations
of advertisements outside the natural editorial context, thereby increasing the study’s
external validity. The paper also provides a much needed update of newspaper research
findings, since much of the limited academic research on newspaper advertising dates
back to the 1980s.
We present hypotheses based on ad processing theories and a thorough analysis of
previous studies of newspaper advertising, synthesising all previously tested relation-
ships between readers’ processing of newspaper ads and influencing factors. Field data
(N = 26,556 readers and 290 unique advertisements) were used to test these hypotheses
and to show how readers process newspaper ads and what characteristics most effectively
catch the reader’s eye.
Ad processing
Greenwald and Leavitt (1984) propose four levels of ad processing, basing their frame-
work on several influential papers, such as Krugman’s work on involvement (1965), and
that of Craik and Lockhart (1972) on levels of processing. The four distinct levels are pre-
attention, focal attention, comprehension and elaboration (Greenwald & Leavitt 1984,
p. 84). At the pre-attention level, a reader continuously scans the newspaper environment
in a subconscious and automatic way, filtering all incoming information. Only a small
selection of stimuli enters the second level (focal attention), where enough attention is
paid for the reader to determine what the ad is about. This orientation stage takes no more
than a second, allowing the reader to decide if the ad is worth paying more attention to. At
the third level – comprehension – the ad is analysed to assign meaning. Although the ad
will now leave some traces in memory, Greenwald and Leavitt hypothesise that there is as
yet no connection to existing knowledge structures, a process that takes place at the fourth
level, elaboration. At this level, the person’s full capacity is used to respond cognitively to
the ad, generating personal connections and imagery. It is assumed that these four levels
66
effects on readers’ processing of newspaper ads
require increasing processing capacity and have a hierarchical structure. Thus, if processing
at one level fails to evoke the next level, processing of the ad is terminated.
Others have criticised this model. Finn (1988) proposes a divergent processing model
in which it is possible to move directly from focal attention (the second level) to elabora-
tion (the fourth), because additional processing is dependent upon the individual’s goals.
Other researchers have argued that Greenwald and Leavitt do not allow for the effects of
ad processing in peripheral vision, which has been shown to influence consumer decision
making (Shapiro et al. 1997), and that they do not consider the possibility that emotional
and heuristic elements may be processed subconsciously (Heath 2009).
The ad processing concepts introduced by Starch (1931, 1966) play an important role
in empirical tests of the models developed by Greenwald and Leavitt (1984) and Finn
(1988), and other ad processing studies (Hendon 1973; Holbrook & Lehmann 1980;
Rossiter 1981; Zinkhan & Gelb 1986). Starch’s concepts include several measures: noted
scores (recognition of the ad), associated scores (brand awareness), and ad read scores,
conceptualising different levels of ad processing, from passive to active. In addition to
these Starch measures, we included in our study ad exposure, defined as ‘the number of
target audience individuals exposed to the advertising’ (Rossiter & Percy 1998, p. 447),
sometimes described as the number of people with open eyes/ears in front of advertising
space, which is comparable to ‘opportunity to see’ (OTS) measures used in media planning
(Slater 2004).
Factors influencing processing of newspaper ads

Which factors influence ad processing in newspapers? In our search for possible effects of
position, ad characteristics and reader characteristics, we followed two strategies: (1) we
analysed previous studies on advertising processing in newspapers; and (2) we added
insights from advertising processing theories.
Previous studies on processing of newspaper advertising

A keyword and reference search was conducted in several databases (Academic Search
Premier, Business Source Premier, Communication & Mass Media Complete, PsycINFO,
Sociological Abstracts, Online Contents, WARC, Web of Science). The following key-
words were used in combination with print or newspaper: ad(vertisement)(s), advertising,
attention, attend(ing), process(ing), readership. No limits were applied for year of publica-
tion. Studies into other types of advertising (such as Finn’s 1988 magazine study), or stud-
ies with separate test ads (such as that of Leigh et al. 2006) were not included. The search
showed only 12 published empirical studies in peer-reviewed outlets on the factors that
influence processing of newspaper advertising. Remarkably, only two newspaper studies
have been published this century, while most appeared in the 1980s.
The different studies were analysed by means of a vote-counting analysis in which all
empirically studied relationships were coded as a relationship between an independent
and a dependent variable. In the vote-count analysis the number of studies is counted
67
that found a positive, negative or no statistically significant relationship between the

variables: ‘A vote-counting analysis is considered the most suitable method of research
synthesis when a literature is small and heterogeneous as in the case of the advertising-
effects literature’ (Buijzen & Valkenburg 2003, p. 438; see also Cooper 1990; Bushman
1994; Eisend 2009).
Table 1 summarises the findings of the studies. The independent variables were put
into three categories: position in the newspaper, ad characteristics and reader character-
istics. The relationships found in the literature were coded as being positive, negative or
without direction, significant or not. A relationship was coded as being moderated when
the effect was not direct, but an effect of another variable on ad processing, with this
specific variable being involved.
Table 1 shows that these studies investigated 45 relationships and tested 21 different
independent variables: 12 independent variables were tested once, and nine were tested in
two or more studies. The most commonly investigated independent variable was ad size
(nine studies).
Before we discuss the findings in detail, three observations should be made about
the studies. First, the dependent variables in previous newspaper studies were generally
lower-level processing variables such as recall, recognition, ad noted and ad seen. Variables
that correspond with higher levels of ad processing (such as associated scores and ad read
scores) were not tested, nor were variables that measure ad exposure (opportunity to see).
Second, the majority of the independent variables in these studies were tested only once,
which is an unstable basis for drawing firm conclusions. Third, most studies considered
only a few independent variables, which makes it difficult to determine the direct (causal)
effects of these variables, corrected for the influence of other influencing variables.
What did these previous studies find? We discuss the findings summarised in Table 1
for the three categories of independent variables. First, with respect to position in the
newspaper: significant effects were found for news sections (versus other sections), ads on
right-hand pages (versus left), ads on the front page (versus other pages), single ads on a
page (versus cluttered ads), ads published on Thursdays and Fridays (versus Mondays and
Tuesdays), and ads in tabloids (versus broadsheets). However, though expected, no effects
were found for the variables page number, ads on editorial pages and position on the page.
Second, for ad characteristics, it was found that bigger ads, coloured ads, particular prod-
uct types (ads for women’s clothing and ads for electric appliances) were noted more. No
significant effects were found for ad layout (for example, copy–picture ratio) or number of
items in the ad. Third, for reader characteristics, brand use, medium use, attitudes towards
the ad (ad liking, interestingness, felt involvement), sex (women more) and age were found
to influence ad processing.
Although it is possible – as indicated above – that some of the relationships between
these independent variables and ad processing variables are spurious, and that they might
become insignificant if other independent variables were included, we formulated hypoth-
eses for each of the significant findings (see Tables 2a, 2b and 2c). Newspaper position
hypotheses are indicated as Hn1, Hn2, etc.; ad characteristics hypotheses as Ha1, Ha2, etc.;
and reader characteristics as Hr1, Hr2, etc.
68
Table 1: Summary of tested relationships (sorted per category)
Independent variable Effect1 Dependent variable References2

Position characteristics:
Section (news vs other) 3 (+ + *) Claimed ad noticed; aided recall Sol82a; Sol82b
Right vs left page 3 (+ + ns) Claimed ad noticed; recognition Fer35; Hul03; Wal98
Front page 1 (+) Claimed ad noticed Wal98
Single ad versus ad clutter 1 (+) Claimed ad noticed Wal98
Day of the week 1 (+) Ad recognition Fer35
(Thu and Fri vs Mon and Tues)
Newspaper format 1 (mo) Claimed ad noticed Wal98
(tabloid vs broadsheet) (by right vs left page)
Page number 1 (ns) Recognition Fer35
Next to editorial copy 1 (ns) Claim ad read Lew89
Position on the page 1 (ns) Recognition Fer35
Ad characteristics:
Ad size 9 (all +) Claimed ad noticed/read/seen; Bog88; Fer35; Hul03;
eye tracking; aided recall; Hyu06; Wal98; Sol82a;
recognition Sol82b; Sta98; Tro65
Ad colours 2 (all +) Claimed ad noticed Hul03; Wal98
Product type 3 (all *) Aided ad recall Sol82a
Ad layout (attractiveness; 3 (all ns) Claimed ad seen/read Tro65; Lew89
type; copy–picture ratio)
Number of items in the ad 1 (ns) Claimed ad seen Tro65
Reader characteristics:
Brand use (choice, preference, 4 (all +) Recall; recognition; eye tracking Bog70; Sta98
purchase intentions)
Medium use (article involvement, 2 (all +) Ad recall; claimed ad noticed Hyu06; Wal98
editorial readership)
Ad liking 1 (all +) Recall; recognition Sta98
Ad interestingness 1 (+) Recall; recognition Sta98
Felt involvement 1 (+) Ad viewing time Cel88
Sex (women) 4 (2+ 2mo) Recognition; claimed ad noticed Fer35; Sol82a; Sol82b
(by section and product type)
Age 1 (+) Claimed ad noticed; claimed Hul03
attention paid
Notes:
1
Number of studied relationships is listed. Coding between brackets is + or – for each positive or negative significant
relationship, mo means a moderated effect, ns means relationship is not significant, and * means a significant relationship, but
no direction given
2
References are abbreviated by the first three letters of the first author and the last two figures of the year, see also reference
list. Reading example (first row): the effect of section (news section versus other sections) was investigated three times in two
studies; a positive effect was found twice (ads in the news section were more frequently noticed than ads in other sections), a
significant relationship was noted once, but the direction of the relationship (positive or negative) was not noted. Dependent
variables in these studies were ‘claimed ad noticed’ and ‘aided recall’.
69
Table 2a: Hypotheses with respect to position in the newspaper
Confirmed in our study?

Hypotheses based on vote count
Hn1: Ads included in the news section get noticed more No
Hn2: Ads on right-hand pages get noticed more No
Hn3: Ads on the front pages get noticed more No
Hn4: Single (compared to cluttered) ads on a page get noticed more No
Hn5: Ads published on Thursdays and Fridays get noticed more Yes (weekend less)
Hn6: Ads in tabloid newspapers get noticed more No
Added hypotheses
Hn7: Ads included in the news section get more exposure Yes
Hn8: Ads on the front page get more exposure Yes
Hn9: Ads on the back page get more exposure No
Hn10: Ads on the front page of a section get more exposure No
Hn11: Ads on the back page of a section get more exposure No
Hn12: Ads at the beginning of the newspaper get more exposure Yes
Hn13: Ads in thinner newspapers get more exposure Yes
Hn14: Position in the newspaper is important for ad exposure Yes
Table 2b: Hypotheses with respect to ad characteristics

Ha1: Bigger ads (compared to smaller ads) get noticed more Yes
Ha2: Coloured ads (compared to black and white ads) get noticed more Yes
Ha3: Advertised product type influences on ad noticing No
Added hypotheses
Ha4: The relative influence of ad characteristics compared to reader interest Yes
characteristics is stronger for ad noticing than for ad exposure, brand
awareness and ad read
Ha5: The relative influence of mechanical variables compared to message variables Yes
is stronger for ad noticing than for ad exposure, brand awareness and ad read
Added: effects on ad exposure

Previous studies did not investigate ad exposure. What would we expect? By definition, ad
exposure, operationalised as opening the page with the ad on it, cannot be influenced by
the form or content of the ad. We assume that ad exposure is a function of position in the
newspaper and reader characteristics. We expect that a prominent position, such as the news
section, front page, back page, front page of a section, back page of a section, page number
(lower numbers, more exposure), size of the newspaper (thicker newspapers, less exposure),
have an effect on ad exposure. We added these hypotheses in Table 2a (Hn7 to Hn14).
70
Table 2c: Hypotheses with respect to reader characteristics

Hr1: Users of the brand process the ads of the brand more strongly Yes
Hr2: Newspaper readers who use the medium more intensively process ads
Yes
more strongly
Hr3: Liked ads are processed more strongly Yes
Hr4: Women process ads more strongly Yes
Hr5: Older readers process ads more strongly Yes
Added hypotheses
Hr6: The relative influence of reader interest characteristics compared to ad
Yes
characteristics is higher for brand awareness and ad read than for ad noticing
Added: differential influence of stimulus-specific versus person-related

characteristics
As we saw in Table 1, no study of newspaper advertising has investigated the effects
of higher-level processing variables, such as brand awareness and ad read. Which
effects would we expect? We hypothesise that stimulus-specific characteristics will
be more important for catching the reader’s eye (ad noticing), whereas person-
related characteristics will be more important for more intensive ad processing
(Rossiter 1981; Greenwald & Leavitt 1984; Wedel & Pieters 2000; Leigh et al. 2006).
Wedel and Pieters (2000) distinguished between ‘attractors of attention’ and ‘retainers
of attention’. Attractors include features of the ad; retainers include person-related fac-
tors like product involvement, and experience with the advertised product and brand.
Ad processing driven by these different determinants is believed to occur in distinct
but connected areas of the brain (Pieters & Wedel 2004). Features of the ad, such as
ad size and shape, determine perceptual salience and capture attention to ad elements
rapidly and almost automatically, even when someone is not actively searching for them.
Person-related factors encourage people to pay more or less attention to advertisements
and their elements.
This view is also reflected in the distinction between bottom-up and top-down pro-
cesses (Rosbergen et al. 1997; Pieters & Wedel 2004, 2007). Bottom-up processing is a
rapid, automatic form of attention that depends on the intrinsic properties of the stimulus
to which someone is exposed (Eysenck & Keane 2005). Top-down processing is a voli-
tional, focal, mechanism that depends on personal motivation; it enhances processing of
the selected item (Eysenck & Keane 2005). These considerations lead to the following
hypotheses: the relative influence of ad characteristics compared to reader interest char-
acteristics is stronger for ad noticing than for ad exposure, brand awareness and ad read
(see Ha4 in Table 2b); the relative influence of reader interest characteristics compared to
ad characteristics is higher for brand awareness and ad read than for ad noticing (see Hr6
in Table 2c).
71
Added: differential influence of mechanical versus message characteristics

The number of variables that may be used to characterise ads is endless (see, e.g., Rossiter
1981; Finn 1988; Hoeken 1995). A well-known distinction within this category is between
‘mechanical’ versus ‘message’ variables (Ferguson 1935; Valiente 1973). The former include
variables such as ad size and use of visuals, and the latter ones such as communication
strategy, use of humour, or fear appeal. It has been hypothesised that mechanical variables
are relatively more important for explaining ad noticing, whereas message variables that
relate to readers’ interest are more important for higher-level processing variables: brand
awareness and ad read (Hendon 1973; Holbrook & Lehmann 1980; Rossiter 1981). For
example, in a magazine study, Rossiter (1981) showed that about 40–45% of the variance
in initial attention (noted scores) and associated scores is due to visual variables (pic-
ture size) and verbal ones (such as nouns in headline, fewer verbs in headline, personal
reference in headline). Holbrook and Lehmann (1980) showed that message variables such
as creativity, believability and status increased ad read. This leads to the following hypothesis:
the relative influence of mechanical variables compared to message variables is stronger for
ad noticing than for ad exposure, brand awareness and ad read (see Ha5 in Table 2b).
Method
Procedure and sample

The advertisements in our study were taken from the database of the Advertising
Response Research (ARR). The ARR is a standardised research instrument commis-
sioned by the Dutch marketing association for newspaper publisher Cebuco. The ARR is
based on the recognition method and uses computer-assisted self-interviewing: electronic
questionnaires are sent to respondents, completed offline and the answers returned via the
internet. The respondent is asked to pick up his/her copy of the previous day’s paper in
order to measure advertising reactions in a natural context. The ads that are included in
the ARR are selected by Cebuco in collaboration with advertisers.
Each day a number of respondents are asked to fill out the questionnaire, assuring an
equal spread over the week. Respondents are selected from the research company’s access
panel, which consists of a nationally representative sample of the Dutch population. To
participate, respondents must have read or seen the particular issue (i.e. the previous day’s
newspaper). The total number of respondents was 26,556; 63% of these respondents were
male, and the average age was 45 (SD = 13). The group of respondents is representative
for the Dutch population of newspaper readers.
The analysed data included 290 ads. Between 70 and 405 respondents were inter-
viewed for each ad, with an average of 151 per ad (SD = 43.10). Most respondents (82.9%)
evaluated one ad; the others evaluated more than one ad. The average number of ads
evaluated by a respondent was 1.65 (SD = 1.44).
The 290 ads were content-analysed by trained coders from Cebuco, using a coding
system that was developed and tested in earlier research on newspaper ads, magazine ads
72
and television commercials (Van der Lee & Smit 2001; Neijens & Smit 2003; Steevels
& Van den Putte 2004). To assess coder reliability, a random sample of 20% of the adver-
tisements was selected and coded by a second coder, from the University of Amsterdam.
Measures
The dependent ad processing variables were operationalised in line with Starch (1931,
1966), Zinkhan and Gelb (1986), Finn (1988), and Rossiter and Percy (1998). Ad exposure
was measured by the question (referring to the page containing the particular ad), ‘Did
you open page X? We do not mean that you read or saw this page, only that you opened
this page’ (answer: yes/no). Ad noticing was measured by the question, ‘Did you see this ad
on page X, no matter how briefly?’ (answer: yes/no). Brand awareness was measured by the
question, ‘Were you aware that this ad concerns brand X?’ (answer: yes/no). Ad read was
measured with the question, ‘Did you read some or all of this ad?’ (answer: yes/no) (see,
e.g., Wells 1964; Krugman 1985; Zinkhan and Gelb 1986; Leigh et al. 2006).
The independent variables in the hypotheses were measured as follows: newspaper
position characteristics were measured by means of the content analysis, considering news-
paper size (number of pages, M = 41), page number (ranging from 1 to 134), position in
the newspaper (front page, back page, front page of section, back page of section, left or
right page), context (news, business, sports, culture, real estate, science), page type (mostly
editorial, mostly ads, ads with classifieds, combination), and day of the week. The news-
paper’s format was also coded: tabloid (5%) or broadsheet (95%).
In addition, mechanical ad characteristics were coded in the content analysis: ad size
(surface size in millimetres, recoded into ‘smaller than ¼ of the page’ to ‘one full page or
more’), colour (black and white, one colour, more colours), product type (based on Nielsen:
food, 3.1%; consumer products, 5.2%; transportation, 15.5%; ICT/telecom, 9.3%; indus-
trial products, 3.4%; retail, 21.4%; leisure, 6.2%; financial services, 19.7%; other, 16.2%),
logo surface in millimetres (logo could include any sign that characterizes the brand, such
as an image or the typical typeface of the brand), placing of the logo (upper left, lower
left, middle, upper right, lower right). Message strategy was included in the study by dis-
tinguishing between an action strategy (which aims at immediate action and has a higher
sense of urgency) versus a theme strategy (which communicates brand values) (Van den
Putte & Dhondt 2005).
The quality of the coding of the variables was assessed by recoding 20% of the
advertisements, as was mentioned above. The inter-coder reliability index was based on
Perreault and Leigh (1989), which corrects for chance agreement caused by the number of
categories (e.g. Kolbe & Burnett 1991). The index exceeded the recommended minimum
of 0.80 for all coded variables. The newspaper position variables were perfectly coded
(0.99). The ad content variables included in the study had an index of 0.90, which is good.
The content analysis variables in this study are straightforward (mainly newspaper pos-
ition and mechanical ad characteristics), which contributes to the high reliability scores;
subjective variables (such as ad liking) were not part of the content analysis and were
included in the study as respondent variables, as they depend on readers’ evaluations.
73
The following reader characteristics were measured for each respondent: demographic
data, including age, sex, income and education; medium use, measured by the question
how intensively the paper was read (almost completely, largely, about half, some pieces,
hardly anything, nothing); ad liking, measured on a 10-point scale (from 1 = ‘not at all’
to 10 = ‘very much’); and respondents’ experience with the brand, measured by asking the
respondents whether they ever had bought the advertised brand before (yes/no). The lat-
ter variable (‘brand use’) was included in the data collection only at a later stage, so the
number of respondents that answered this question was much smaller than for the other
variables (N = 348).
Analysis
We created two databases: first, a respondent-ad database with the responses of each
respondent (N = 26,556) to each ad that (s)he evaluated (the average number of ads
evaluated by a respondent was (rounded) 1.65; this resulted in a total of 43,755 person-
ad observations in the respondent-ad database); second, an ad database (N = 290) with
scores for each ad based on the content analysis, the ad processing scores (proportion of
respondents that were exposed to the ad, noticed the ad, etc.), the respondents’ evalua-
tions of the ad (ad liking), measured with mean scores across respondents who evaluated
the ad, and aggregated values for respondent characteristics. To give an example: in the
respondent-ad database, the four ad processing measures are dichotomous (yes/no); in the
ad database, these measures are aggregated scores with values between 0 and 1, based on
the mean scores per ad. Each table includes a reference to the database used.
The analysis comprised several steps. First, bivariate analyses were conducted between
all independent variables and the four dependent variables. Based on this first step, vari-
ables were selected that showed a significant bivariate relationship with one or more of the
dependent variables. Second, multivariate regression analyses were conducted to explain
the variation in the four dependent variables by the independent variables selected in
the first step. The first step was necessary to limit the number of variables, permitting a
multivariate analysis.
The multivariate analyses at the ad level employed hierarchical linear regression: pos-
ition in the newspaper and ad characteristics were entered in the analysis after controlling
for possible effects of respondent characteristics. This was done to rule out possible effects
of different respondents evaluating different ads. At the respondent level, the multivariate
analyses employed logistic regression because the dependent variables consist of ‘yes’ (1)
and ‘no’ (0) answers.
Results
Ad processing
On average, 86% of all respondents opened the page with the investigated ads; 74% of
these respondents remembered the ad on that page (thus 64% of all respondents); 73%
74
of the respondents who noticed the ad were aware which brand was mentioned in the ad
(47% of all respondents); and 69% of the respondents who noticed the ad indicated that
they had read the complete ad or part of it (44% of all respondents).
What explained ad processing at the ad level?

Hierarchical regression analysis was used to explain ad processing by position in the
newspaper (second step) and ad characteristics (third step), after controlling for possible
effects of respondent characteristics (first step). The results of the regression analyses (ad
database; unit of analysis is the ad) are presented in Table 3.
The explained variance per block of predictors indicates that position characteristics
best explain the chance that a respondent is exposed to an ad (50% explained variance as
compared to 1% for ad characteristics). Position characteristics were also important for ad
noticing (13% explained variance) and brand awareness (15%), but not important for ad
read. Ad characteristics were not important for ad exposure (1%), but explained all other
ad processing variables: ad noticing (11%), brand awareness (12%) and ad read (10%).
Table 3: Effects of newspaper position and ad characteristics (N = 290 ads)1
Ad exposure Ad noticing Brand awareness Ad read

Step 2 (position characteristics):
News section 0.21 *** –0.02 –0.10 –0.06
Front page 0.15 ** 0.00 0.01 0.05
Page number –0.50 *** 0.09 –0.29 *** –0.06
NP size (pages) –0.17 ** –0.13 * –0.04 0.01
Editorial page 0.01 –0.12 0.14 0.04
Ad page –0.02 –0.02 –0.01 –0.03
Weekend 0.04 –0.15 * –0.07 –0.07
∆ R2 (Step 2) 0.50 *** 0.13 *** 0.15 *** 0.03
Step 3 (ad characteristics):
Size ad 2 –0.05 0.23 * 0.21 * 0.09
Colours in ad 0.04 0.18 ** 0.25 *** 0.19 **
Message strategy –0.02 0.22 *** 0.18 ** 0.25 ***
∆ R2 (Step 3) 0.01 0.11 *** 0.12 *** 0.10 ***
Total R2Adjusted 0.50 0.21 0.23 0.11
F(13, 273) = 23.13 *** 6.90 *** 7.45 *** 3.60 ***
Notes:
Ad database; N = 290. Table includes Beta values based on hierarchical linear regression (method Enter).
1
Control variables (Step 1) include the aggregated values for respondent characteristics. This step did not significantly add to
the model and is not reported in the table.
2
Ad size is corrected for size of the newspaper page; *p < 0.05, **p < 0.01, ***p < 0.001. Weekend: 0 = week; 1 = weekend.
Message strategy: 0 = theme ad; 1 = action ad. Reading example: first row shows a significant effect of news section (ads
placed in the news section versus ads in other sections) on ad exposure, after controlling for respondent characteristics, with
a standardised regression coefficient of 0.21. Ad noticing is controlled for ad exposure. Brand awareness and ad read are
controlled for ad noticing.
75
The explained variance in ad exposure is considerable; the explained variance in ad notic-

ing, brand awareness and ad read is moderate.
These findings show that position characteristics are most important for ad exposure
(confirming Hn14), but that these variables also affect the higher-level ad processing vari-
ables, namely ad noticing and brand awareness. Mechanical ad variables are most import-
ant for ad noticing and brand awareness, while ad message strategy is more important for
ad read (confirming Ha4 and Ha5).
When considering how position characteristics affect ad processing, Table 3 shows that
ads in the news section (Hn7), on the front page (Hn8), the beginning of the newspaper
(Hn12) and ads in thinner newspapers (Hn13) were more likely to be seen (ad exposure).
We also found a weekend effect (Hn5): ads at the weekend – controlled for number of
pages, since newspapers at the weekend have many more pages – were noticed less. On
the other hand, our hypotheses about the news section (Hn1), right-hand pages (Hn2),
front page (Hn3), single ads (variables ‘editorial page’ and ‘ad page’) (Hn4), ads in tabloids
(Hn6) on ad noticing were not confirmed. Our hypotheses with respect to ads on the back
page (Hn9), on the front page of sections (Hn10) and back page of sections (Hn11) on ad
exposure were not confirmed either.
Three ad characteristics contributed significantly to more ad processing in terms of ad
noticing, brand awareness and ad reading. First, bigger ads were more often noticed than
smaller ones, and respondents were more aware of the advertised brand (confirming Ha1).
Second, ad noticing and brand awareness were much higher for full-coloured ads than
black and white, and more respondents claimed to have read the ad (confirming Ha2).
Action ads also got more attention than theme ads (confirming Ha5). Our hypothesis
with respect to product type (Ha3) was not confirmed.
What explained attention at the respondent level?

Logistic regression was used to determine which reader characteristics explain ad process-
ing (see Table 4; respondent-ad database). Odd ratios lower than 1 indicate a negative
relationship, and ratios higher than 1 indicate a positive relationship. The high number
of respondents means that even small effects are significant. We therefore base our inter-
pretation on effect sizes. As there is much discussion about whether or not to apply stand-
ardised coefficients in logistic regression, and if so, which coefficient to use (Williams
2011), we express the relative strength of the independent variables in standardised linear
regression coefficients (betas). The effects of brand use are presented separately, because
the number of respondents for whom this variable was measured was much smaller (see
Table 5).
Demographics appeared relatively unimportant in explaining the variation in ad pro-
cessing. Female readers notice slightly more ads (confirming Hr4) and older readers score
higher on ad processing variables (confirming Hr5). However, brand use, reading intensity
and ad liking have a stronger effect on ad processing levels (confirming Hr1, Hr2 and Hr3).
These findings confirm that reader characteristics are more important for higher-level
processing variables (Hr6).
76
Table 4: Effects of reader characteristics

Nagelkerke R2 0.18 0.10 0.14 0.12
Age 1.02 *** 1.00 *** 1.00 * 1.00
[1.01, 1.02] [1.00, 1.01] [1.00, 1.01] [0.99, 1.01]
(0.08 ***) (0.02 ***) (0.01 *) (–0.01)
Sex 0.89 *** 0.96 ** 1.19 *** 0.95
[0.83, 0.94] [0.87, 0.97] [1.12, 1.26] [0.90, 1.01]
(–0.02 ***) (–0.02 **) (0.03 ***) (–0.01)
Reading intensity 2.35 *** 1.56 *** 1.18 *** 1.41 ***
[2.28, 2.42] [1.51, 1.60] [1.14, 1.22] [1.37, 1.46]
(0.31 ***) (0.17 ***) (0.06 ***) (0.13 ***)
Ad liking 1.10 *** 1.45 *** 1.85 *** 1.66 ***
[1.07, 1.13] [1.42, 1.48] [1.80, 1.90] [1.61, 1.70]
(0.03 ***) (0.20 ***) (0.31 ***) (0.26 ***)
Notes:
Respondent-ad database; N = 39,772 (Ad Exposure); 34,251 (Ad Noticing); 25,332 (Brand Awareness), and 25,335 (Ad Read).
Table shows odd ratios based on logistic regression (eB); *p < 0.05, **p < 0.01, ***p < 0.001 (based on Wald); 95% confidence
interval [lower, upper]; between brackets (): standardised beta values based on linear regression analysis. Sex: female = 0, male
= 1. Ad noticing is controlled for ad exposure. Brand awareness and ad read are controlled for ad noticing.
Table 5: Effects of brand use

Brand use (%) 79.7 73.7 80.4 91.5
No brand use (%) 75.8 55.3 53.8 8.5
Total (%) 79.0 70.5 76.8 67.0
Chi2 (df = 1) 0.47 6.33 * 8.88 ** 8.29 **
Notes:
Respondent-ad database. Table shows % of respondents scoring ‘yes’ on the included processing variables. Based on Chi2;
*p < 0.05, **p < 0.01, ***p < 0.001. N = 348 (Ad Exposure); 275 (Ad Noticing); 194 (Brand Awareness), and 194 (Ad Read).
Conclusion and discussion

Some ads in newspapers are processed better than others, and some readers are better
ad processors than others (Nördfalt 2005). What explains differences in ad processing
between ads and between readers? To answer this question, we studied how position in the
newspaper, ad characteristics and reader characteristics influence ad processing at different
levels: exposure, noticing, brand awareness and ad read.
Our study provides the following answers. First, exposure to newspaper ads was largely
determined by the ad’s position in the newspaper: ads in the first part of the newspaper,
ads on the front page and ads in the news section had a higher chance of being seen.
77
Second, mechanical ad characteristics determined ad noticing and brand awareness, in

particular. Bigger ads and ads with colours were noticed better, and readers were more
aware of the advertised brand. Third, ad read – the highest ad processing level considered
in our study – was especially influenced by how attractive the ad was for the reader: the
relevance of the message content (brand users scored higher, action ads scored higher
than theme ads) and the ad’s execution (liked ads scored higher). Fourth, reading intensity
affected all ad processing variables: readers who read the newspaper more intensively were
more exposed to ads, noticed more ads, were more aware of the advertised brand and read
the ads more than less intensive newspaper readers.
Our study shows that, with increasing levels of ad processing, influencing vari-
ables moved from position in the newspaper (especially important for ad exposure),
via mechanical ad characteristics (important for ad noticing and brand awareness)
to message content and reader interest characteristics such as ad liking and brand
use (which affect ad noticing, brand awareness and ad read). These findings support
earlier studies, which also found a differential influence between mechanical and message
variables (Holbrook & Lehmann 1980; Rossiter 1981), and between stimulus-specific
characteristics and person-specific factors (Wedel & Pieters 2000; Pieters & Wedel
2004). Stimulus-specific characteristics, in particular mechanical variables, induce a
bottom-up process and are most important as attractors of attention; person-specific
factors such as involvement with the brand or interest in the product category induce a
top-down process, work as retainers of attention and are more important for further ad
processing.
The four levels of advertising processes that we distinguished were based on theory
proposed by Greenwald and Leavitt (1984) and also used in the empirical (magazine)
study of Finn (1988). The model proved to be a useful starting point, to which we added
the differential effects of media position, ad and reader characteristics on the various ad
processing variables based on theories of bottom-up and top-down processes developed
by Pieters and Wedel (2004). Finn also added effects to the Greenwald and Leavitt model,
but in his article, ‘Extended model of the audience readership of print ads’, he tested only
effects of ad characteristics on ad noticing. Greenwald and Leavitt’s model is a strict
hierarchical model: if processing at one level fails to evoke the next level, processing of the
ad is terminated. Finn additionally found that it is possible to move directly from focal
attention (the second level) to elaboration (the fourth) (see also Edell & Staelin 1983).
Which route is selected depends upon the specific goals of the reader (Mitchell 1983).
Pieters and Wedel (2007) found that ad informativeness is goal contingent: the goal of
learning about a brand, for example, promotes readers to concentrate on processing certain
parts of the body text instead of attending to all text.
In addition, our study confirmed other hypotheses examined in previous studies. For
example, a prominent position in the newspaper is important for ad exposure: in the news
section, on the front page and at the beginning of the newspaper. Furthermore, ads in
thinner newspapers and ads on weekdays get more exposure. Ad size and use of colour
were also found to be important for ad processing. Reader characteristics had an import-
ant role as well, particularly reading intensity, liking of the ad and brand use.
78
Some hypotheses from previous studies were not confirmed. We did not find a (posi-
tive) effect of right-hand pages, single (versus cluttered) ads, tabloids, front and back of a
section and product type. Nor did demographics play a major role in explaining ad pro-
cessing in our study. Differences in study design might explain the different results of our
study compared to previous studies. We included ad exposure and ad read in our study,
something other newspaper studies have not, and we conducted multivariate analysis in
which the effects of the independent variables were controlled for one another, where
previous studies have generally studied only a few independent variables. The fact that we
did not find significant differences between ads in tabloids and those in broadsheets could
be explained by the very small percentage of tabloids in our study (5%).
Our study contributes to the discussion of accountability in media planning. On the
basis of our results, the question can be raised whether all attention measures are import-
ant for the value of a newspaper (or any other medium) in the advertising market. If gen-
erating an audience is the primary media objective, media should be judged on their ability
to expose ads to an audience and only ‘page open’ is relevant. However, when the objective
is to direct attention to the ad, one should also consider the medium’s performance on
‘ad noticing’ or ad awareness when determining its value. This study shows, however, that
these criteria are influenced not only by media position and reader characteristics, but by
ad characteristics as well. And these ad characteristics are beyond the control of the news-
paper publisher.

The strength of our study is its external validity, with real newspapers and real ads investi-
gated in a natural context, using a large number of ads (N = 290), multiple measures for ad
processing, and a huge and random sample (N = 26,556) of the population of newspaper
readers. These advantages are associated with some limitations, which lead to suggestions
for further research.
First, some measures in our study are based on single-item measures and self-reports,
such as intensity of reading, ad processing, ad liking and brand use. Although we believe
that these concepts are ‘concrete’ perceptual concepts for which single-item measures
suffice (Rossiter 2011), further research on the validity and reliability of these measures
might be useful. In particular, eye tracking would enrich the ad processing concepts, and
future research would benefit from including this metric.
The number of variables included in our content analysis was limited in comparison
with the endless list of possible options as suggested by Hoeken (1995). Given our goal
to study the relative influence of four types of variables (position, ad form, ad content and
reader), we selected variables from these four categories without intending to be necessar-
ily exhaustive (if that were possible at all). However, we admit that the number of variables
included in the analysis could be extended – in particular, variables that indicate message
content. A promising candidate would be, for example, brand differentiation (inclusion
of USP and brand image elements in the ad), which proved to be a successful predictor
in the television commercial studies of Stewart and Furse (1984, 1985), and Stewart and
79
Koslow (1989). It should be noted that these ‘higher order’ variables are harder to code
and require more coders – four in the studies of Stewart and Furse (1984, 1985), and
Stewart and Koslow (1989). Including more variables that measure the four types of
independent variable would probably result in higher R-squares for the explanations of
the processing variables.
Another restriction is that our data were collected in one country. Although we believe
that the Netherlands is a representative western country, and our results are in line with
ad processing studies in, for example, the United States, and confirm general advertising
theories, further research is necessary to investigate to what extent our findings are robust
across media landscapes, advertiser practices and populations. A study by Callow and
Schiffman (2002), for instance, showed that respondents from the Philippines expressed
a greater ability to derive meaning from implicit visual ads compared to the American
respondents.
Note that our study measured only explicit processing levels. Our dependent variables
were based on Starch scores, which provide useful information about the effectiveness
of an ad (Zinkhan & Gelb 1986; Smit & Neijens 2011). As exposure without conscious
awareness can have ad effects ( Janiszewski 1988, 1993; Kitchen & Spickett-Jones 2003;
Buijzen et al. 2010), we might have missed these influences. We also did not study inter-
esting persuasion effects such as attitudinal and behavioural change.
An interesting question is whether processing of newspaper advertising is different
from that of other print media such as magazines. Some characteristics of these media are
similar: both are print media in which ads are combined with editorial content. Others
are not: newspapers are general interest media, while magazines are special interest media;
media use, reading motivations and gratifications differ (McQuail 2005; Bronner &
Neijens 2006). It is an open question to what extent factors influencing ad processing in
these two media match.
Our final remark concerns causality. Although we use the term ‘effects’ in our title, we
did not measure effects in terms of experimentally controlled effects with a stimulus and
an effect of the different conditions. In our study, we estimated effects from variances in
the content on the one hand and reactions of readers on the other. These reactions could
have been caused by other non-experimental factors. Despite this lack of internal valid-
ity of effects, our study shows a high external validity of found relations between content
and reactions. These relations could add to the principles of evidence-based advertising
formulated by Armstrong (2011).
Building on previous research on processing of newspaper ads, the present study
investigated the combined effects of position in the newspaper, ad form and content, and
reader characteristics. The results of this real-life study confirm ad processing theories that
specify a differential effect of stimulus-specific and person-related factors: with increasing
levels of ad processing, from passive to active, the influence moved from position in the
newspaper, via mechanical ad characteristics, to message content and reader interest vari-
ables. Advertisers who want to reach their audience need to first consider placement; but
to encourage the audience to actually process the ad, they need to make the ad outstanding
and relevant.
80
References
Notes: references included in the vote count are indicated with *; abbreviations between
brackets [...] refer to the references listed in Table 1.
Armstrong, J.S. (2011) Evidence-based advertising: an application to persuasion. International

*Bogart, L. & Tolley, B.S. (1988) The search for information in newspaper advertising. Journal
of Advertising Research, 28(2), 9–19. [Bog88]
*Bogart, L., Tolley, B.S. & Orenstein, F. (1970) What one little ad can do. Journal of Advertising
Research, 10(4), 3–13. [Bog70]
Bronner, A.E. & Neijens, P.C. (2006) Audience experiences of media context and embedded
advertising: a comparison of eight media. International Journal of Market Research, 48(1), 81–100.
Buijzen, M. & Valkenburg, P.M. (2003) The impact of television advertising on materialism,
parent–child conflict, and unhappiness: a review of research. Journal of Applied
Developmental Psychology, 24(4), 437–456.
Buijzen, M., van Reijmersdal, E.A. & Owen, L.H. (2010) Introducing the PCMC model: an
investigative framework for young people’s processing of commercialized media content.
Communication Theory, 20(4), 427–450.
Bushman, B. (1994) Vote-counting procedures in meta-analysis. In: Cooper, H. & Hedges, L.V.
(eds) The Handbook of Research Synthesis. New York: Russel Sage Foundation, 193–213.
Callow, M. & Schiffman, L. (2002) Implicit meaning in visual print advertisements: a cross-
cultural examination of the contextual communication effect. International Journal of
Advertising, 21, 259–277.
*Celsi, R. & Olson, J.C. (1988) The role of involvement in attention and comprehension
processes. Journal of Consumer Research, 15(2), 210–224. [Cel88]
Cooper, H.M. (1990) Integrating Research: A Guide for Literature Reviews (2nd edn). Newbury
Park, CA: Sage.
Craik, F.I.M. & Lockhart, R.S. (1972) Levels of processing: a framework for memory research.
Journal of Verbal Learning and Verbal Behavior, 11(1), 671–684.
Edell, J.A. & Staelin, R. (1983) The information processing of pictures in print advertisements.
Journal of Consumer Research, 10(1), 45–61.
Eisend, M. (2009) A meta-analysis of humor in advertising. Journal of the Academy of Marketing
Science, 37(2), 191–203.
Eysenck, M.W. & Keane, M.T. (2005) Cognitive Psychology: A Student’s Handbook. New York:
Psychology Press.
*Ferguson, L.W. (1935) The importance of the mechanical features of an advertisement.
Journal of Applied Psychology, 19(5), 521–526. [Fer35]
Finn, A. (1988) Print ad recognition readership scores: an information processing perspective.
Greenwald, A.G. & Leavitt, C. (1984) Audience involvement in advertising: four levels. Journal
of Consumer Research, 11(1), 581–592.
Heath, R.G. (2009) Emotional engagement: how TV builds big brands at low attention.
Journal of Advertising Research, 47(1), 62–73.
Hendon, D.W. (1973) How mechanical factors affect ad perception. Journal of Advertising
Research, 13(4), 39–45.
Hoeken, H. (1995) The design of persuasive texts: effects of contents, structure, and style on
attitude formation. Doctoral dissertation, University of Tilburg, the Netherlands.
Holbrook, M.B. & Lehmann, D.R. (1980) Form versus content in predicting Starch scores.
81
*Hultén, S. & Wilberg, E. (2003) Unveiling the true value of ads. ESOMAR Print Audience
Measurement LA, June. [Hul03]
*Hyun, Y.J., Gentry, J.W., Park, C. & Jun, S. (2006) An investigation of newspaper ad memory
as affect context involvement and ad size – a Korean case. Journal of Current Issues and
Research in Advertising, 28(1), 45–56. [Hyy06]
Janiszewski, C. (1988) Preconscious processing effects: the independence of attitude formation
and conscious thought. Journal of Consumer Research, 15(2), 199–209.
Janiszewski, C. (1993) Preattentive mere exposure effects. Journal of Consumer Research, 20(3),
376–392.
Kitchen, P.J. & Spickett-Jones, G. (2003) Information processing: a critical literature review
and future research directions. International Journal of Market Research, 45(1), 73–98.
Kolbe, R.H. & Burnett, M.S. (1991) Content-analysis research: an examination of applications
with directives for improving research reliability and objectivity. Journal of Consumer
Research, 18(2), 243–250.
Krugman, H.E. (1965) The impact of television advertising: learning without involvement.
Public Opinion Quarterly, 29(3), 349–356.
Krugman, H.E. (1985) Point of view: measuring memory – an industry dilemma. Journal of
Advertising Research, 25(4), 49–51.
Leigh, J.H., Zinkhan, G.M. & Swaminathan, V. (2006) Dimensional relationships of recall and
recognition measures with selected cognitive and affective aspects of print ads. Journal of
Advertising, 35(1), 105–122.
*Lewis, W. (1989) Readership of buried ads versus ads placed beside reading matter. Newspaper
Research Journal, 10(2), 67–74. [Lew89]
McQuail, D. (2005) McQuail’s Mass Communication Theory (5th edn). London: Sage.
Mitchell, A.A. (1983) The effects of visual and emotional advertising: an information-
processing approach. In: Percy, L. & Woodside, A.G. (eds) Advertising and Consumer
Psychology. Lexington, MA: Lexington Books, 197–217.
Neijens, P.C. & Smit, E.G. (2003) De Kracht van Adverteren in Dagbladen en Dagbladmagazines
[The Power of Advertising in Newspapers and Newspaper Magazines]. Amsterdam:
SWOCC.
Nördfalt, J. (2005) Track to the future? A study of individual selection mechanisms preceding
ad recognition and their consequences. Journal of Current Issues and Research in Advertising,
27(1), 19–30.
Perreault, D.W. & Leigh, L.E. (1989) Reliability of nominal data based on qualitative
judgments. Journal of Marketing Research, 26(2), 135–148.
Pieters, R.F.G.M. & Wedel, M. (2004) Attention capture and transfer by elements of
advertisements. Journal of Marketing, 68(2), 36–50.
Pieters, R.F.G.M. & Wedel, M. (2007) Goal control of attention to advertising: the Yarbus
implication. Journal of Consumer Research, 34(2), 224–233.
Rosbergen, E., Pieters, R.F.G.M. & Wedel, M. (1997) Visual attention to advertising: a
segment-level analysis. Journal of Consumer Research, 24(3), 305–314.
Rossiter, J.R. (1981) Predicting Starch scores: visual and psycholinguistic mechanical variables
can predict advertisement and brand recognition as accurately as ‘message’ variables. Journal
of Advertising Research, 21(5), 63–68.
Rossiter, J.R. (2011) Measurement for the Social Sciences. The C-OAR-SE Method and Why it
Must Replace Psychometrics. New York: Springer.
Rossiter, J.R. & Percy, L. (1998) Advertising and Promotion Management (2nd edn). Boston:
McGraw-Hill.
Shapiro, S., MacInnis, D.J. & Heckler, S.E. (1997) The effects of incidental ad exposure on the
formation of consideration sets. Journal of Consumer Research, 24(1), 94–104.
82
Slater, M.D. (2004) Operationalizing and analyzing exposure: the foundation of media effects
research. Journalism & Mass Communication Quarterly, 81(1), 168–183.
Smit, E.G. & Neijens, P.C. (2011) The march to reliable metrics: a half-century of coming
closer to the truth. Journal of Advertising Research, 50(1), 112–123.
*Soley, L.C. (1982) Predicting display ad readership. Newspaper Research Journal, 3(4), 6–14.
[Sol82a]
*Soley, L.C. & James, W.L. (1982) Estimating the readership of retail newspaper advertising.
Journal of Retailing, 58(3), 59–75. [Sol82b]
*Stapel, J. (1998) Recall and recognition: a very close relationship. Journal of Advertising
Research, 38(4), 41–45. [Sta98]
Starch, D. (1931) Buying Power of the American Market: National and Local Markets Measured by
Family Incomes in all States – a Comprehensive Measure of America’s Buying Power. New York:
Daniel Starch.
Starch, D. (1966) Measuring Advertising Readership and Results. New York: McGraw-Hill.
Steevels, R. & Van den Putte, B. (2004) The ABCs of magazine advertising: practical guide for
choosing advertising strategies. In: Neijens, P.C, Hess, C., Van den Putte, B. & Smit, E.G.
(eds) Content and Media Factors in Advertising. Amsterdam: Spinhuis Publishers, 75–88.
Stewart, D.W. & Furse, D.H. (1984) Analysis of the impact of executional factors on
advertising performance. Journal of Advertising Research, 24(6), 23–26.
Stewart, D.W. & Furse, D.H. (1985) The effects of television advertising execution on recall,
comprehension and persuasion. Psychology & Marketing, 2(3), 135–160.
Stewart, D.W. & Koslow, S. (1989) Executional factors and advertising effectiveness: a
replication. Journal of Advertising, 18(3), 21–32.
*Troldahl, V.C. & Jones, R.L. (1965) Predictors of newspaper advertisement readership. Journal
of Advertising Research, 5(1), 2–27. [Tro65]
Valiente, R. (1973) Mechanical correlates of ad recognition. Journal of Advertising Research,
13(3), 13–18.
Van den Putte, B. & Dhondt, G. (2005) Developing successful communication strategies: a
test of and integrated framework for effective communication. Journal of Applied Social
Psychology, 35(11), 2399–2420.
Van der Lee, A.M. & Smit, E.G. (2001) Advertising framed: an exploratory study on
advertising frameworks in financial service advertising. Document Design, 2(2), 156–171.
*Walker, J. & Cardillo, D. (1998) Does size matter (or position, or colour, or context …)?
Admap, January, 26–28. [Wal98]
Wedel, M. & Pieters, R.F.G.M (2000) Eye fixations on advertisements and memory for
brands: a model and findings. Marketing Science, 19(4), 297–312.
Wells, W.D. (1964) EQ, son of EQ and the reaction profile. Journal of Marketing, 28(4), 45–52.
Williams, R. (2011) Standardized Coefficients and Scaling Issues in Logistic Regression.
Downloaded course text. Available online at: http://www.nd.edu/~rwilliam/stats3/L06.pdf
(accessed 9 September 2011).
Zinkhan, G.M. & Gelb, B.D. (1986) What Starch scores predict. Journal of Advertising
Research, 26(4), 45–50.
About the authors

Edith G. Smit (PhD, University of Amsterdam) is a full professor of Media & Advertising
in the Amsterdam School of Communication Research, University of Amsterdam.
Current research interests include targeted advertising, privacy concerns, measuring
advertising effectiveness, and health-related persuasion campaigns. Her papers have been
83
published in the International Journal of Advertising, Journal of Advertising Research, Journal

of Advertising, Cyberpsychology, Behavior and Social Networking, Internet Research, Medical
Internet Research and elsewhere.
Peter C. Neijens (PhD, VU University Amsterdam) holds the Chair of Persuasive
Communication in the Amsterdam School of Communication Research, University of
Amsterdam. His research interests include media and advertising, public opinion and per-
suasion. His papers have been published in the European Journal of Marketing, International
Journal of Advertising, Journal of Advertising, Journal of Advertising Research, Psychology and
Marketing, Journal of Communication, Journal of Computer-Mediated Communications,
Journal of Health Communication, Public Opinion Quarterly, the International Journal of
Public Opinion Research and elsewhere.
Robert Heath (PhD, Bath School of Management) is a Senior Lecturer in Advertising
Theory at Bath School of Management. His research focuses on the emotional influence
of advertising at low levels of attention. His journal publications have received numerous
awards, and his latest book, Seducing the Subconscious: The Psychology of Emotional Influence
in Advertising (Wiley-Blackwell), was published in April 2012.
Address correspondence to: Prof. Dr Edith G. Smit, Amsterdam School of
Communication Research, University of Amsterdam, Kloveniersburgwal 48, 1012 CX
Amsterdam, the Netherlands.
Email: e.g.smit@uva.nl
84
Rational and irrational influences
on lucky draw participation
Gerard Prendergast
Hong Kong Baptist University
Edmund R. Thompson
University of Bath
Belief in luck is known to influence opting for a lucky draw instead of other promotional offers
such as discounts. This suggests that irrational factors influence consumers’ participation in
lucky draws, but expected utility theory suggests such participation should also be influenced
by rational considerations about the probability of winning. A survey (n = 665) conducted in
Hong Kong confirmed this proposition and found that both an irrational belief in luck and a
rational appraisal of the probability of winning significantly influence the choice of participat-
ing in a lucky draw over other promotions. Theoretical and practical implications are discussed.
Introduction
Sales promotion is an aspect of the marketing mix that has been demonstrated to have a
strong impact on short-term buying behaviour (Laroche et al. 2003). Constituting ‘a direct
inducement that offers an extra value or incentive for the product to the sales force, dis-
tributors, or the final consumer with the primary objective of creating an immediate sale’
(Haugh 1983), sales promotion tools include coupons, price discounts, sampling, value
packs, refunds, contests, premiums, and many other techniques including lucky draws.
The benefits that such promotions offer consumers may be hedonic, functional, or both.
Understanding what makes one sales promotion tool more effective than another is
important, especially given the increased use of sales promotions (Simpson 2006), and
the growing academic and practitioner focus on marketing productivity (Rust et al. 2004).
Research by Shi et al. (2005) investigated the extent to which the different sales promo-
tion techniques induce different behavioural responses from consumers. They found that
price discounts, buy-one-get-one-free offers and coupons were most effective for encour-
aging stockpiling, purchase acceleration and additional spending. Lucky draws were found
to be relatively less effective at inducing the five behaviours investigated. Their descriptive
study, however, did not delve into any explanation as to why one form of sales promotion
might induce a stronger behavioural response than another. Addressing this issue, some
studies have looked at how sales promotion preferences can be predicted using various

DOI: 10.2501/IJA-32-1-085-100
consumer personality traits. For instance, Laroche et al. (2003) found that personality
traits such as variety seeking, perceiving financial well-being, busyness, brand loyalty and
store loyalty all have an impact on consumer responses to sales promotions.
Individual differences
Looking specifically at lucky draws, these promotional tools are not without controversy.
It has been claimed that lottery organisers tend to exploit lower socio-economic groups
through using counterfactual thoughts to promote them (Landman & Petty 2000).
Landman and Petty are not the only authors to raise the ethical issue, with Wisman
(2006) arguing that lotteries sometimes exploit the poor. Taylor and Kopp (1991)
explained that there are often irregularities and ambiguities in relation to the regulation
of such ‘prize promotions’. As a result, sometimes vulnerable individuals might be hurt
because of having unrealistic expectations. Such research has highlighted the importance
of profiling the rationality or irrationality of the lucky draw-prone individual. At the level
of the individual, however, research explaining why some people participate in lucky draws
and some do not has been limited, as mentioned by Ariyabuddhiphongs (2011) in a recent
review of lottery gambling. The few studies that do focus on luck-based promotions have
used either an attribution theory perspective or an individual difference approach ( Jiang
et al. 2009). One such individual difference is irrational belief in luck, with Prendergast
and Thompson (2008) finding evidence that an irrational belief in luck predicts a prefer-
ence for a lucky draw promotion over a coupon or a price discount.
The demonstrated effect of consumers’ irrational beliefs about luck raises questions
about their rationality in assessing probabilities of success in sales promotions. All con-
sumer promotion tools require consumers to be able rationally to appraise ‘extra value’.
However, there is debate about the extent to which consumers are rational (Shugan 2006).
Some activities, such as signalling (Kirmani & Rao 2000), require consumers to use
extrinsic cues to make supposedly rational decisions. Other marketing activities are argued
to require irrational consumers (Taylor et al. 2004). Kramer and Block (2008), for example,
highlighted the role of irrationality in their study finding that Taiwanese consumers, who
consider the number eight to be lucky, were more likely to purchase a portable radio when
it was priced at the ostensibly lucky price point of NT$888 than when the same radio was
priced at NT$777. In this case their belief in luck was essentially irrational because the
resultant behaviour disregards the calculable benefits. In contrast to an irrational belief in
luck, sales promotion strategies normally offer some sort of ‘extra value’ (Haugh 1983).
Expected utility theory would then predict that rational appraisal of the odds of winning
a lucky draw, and therefore actually gaining the additional value, would at least partly
determine consumers’ willingness to participate in lucky draws.
Extending previous research

Given this confluence of rational and irrational influences, research is needed on consum-
ers’ irrational beliefs about luck and their rational appraisal of winning odds combined in
86
Rational and irrational influences on lucky draw participation
order to understand why they might be drawn to a promotion that offers them a proba-
bilistic outcome ( Jiang et al. 2009). The primary contribution of our research is that it
addresses this issue by examining the relationship between irrational belief in luck and
the rational appraisal of odds of winning a lucky draw in determining lucky draw partici-
pation. We adopt an individual difference approach by identifying stable differences in
irrational beliefs about luck and then examining how people with these differing irrational
beliefs respond to opportunities to participate in lucky draws that have different odds of
winning. Given global variations in beliefs in luck, our study is not only of theoretical
importance in terms of unravelling the rational versus irrational motivations of consum-
ers, but it is also of practical importance to marketers considering the use of lucky draw
promotions across countries and cultures that may vary in their luck beliefs.
Background and hypotheses
The hedonic value of games

There are several ways to classify sales promotion tools. One such classification might
aggregate competitions and lucky draws together as games. The difference, though, is that
success in competitions is primarily determined by skill, whereas success in a lucky draw
is wholly determined by chance. Games are different from other sales promotion tools
since they require skill or chance, and merely participating in them is in itself sometimes
considered by consumers to be enjoyable and therefore an intrinsic part of a game’s value.
Evidence of such an enjoyment benefit has been offered by Ward and Hill (1991), who
suggest that a game’s perceived intrinsic hedonic value in part determines its appeal. The
hedonic value may be undermined, however, if a consumer loses. Ward and Hill suggest
that, in order to reduce the possible negative effects of losing, consumers can be mollified
by suggesting that any loss would be due to luck and not to their ability, a factor arguing
for the popularity of lucky draws. Also looking at games, Darke and Freedman (1995)
found that, similar to Ward and Hill’s (1991) idea of intrinsic value, a benefit acquired
through chance is enjoyed as much as one acquired skilfully, suggesting that winning
is enjoyable no matter whether it depends on chance or skill. Also consistent with this
notion of hedonic benefits, Chandon et al. (2000) have shown that sales promotions may
be attractive to highly promotion-prone consumers for reasons beyond price savings,
which suggests that promotions such as games can be effective simply due to the enjoy-
ment consumers experience through taking part in the promotion.
Irrational versus rational influences on lucky draw participation

Irrational influences
Regardless of the level of enjoyment and/or price savings a promotion offers, a question
remains about the extent to which participation in a lucky draw is rational. Since a lucky
draw is regarded by the consumer as a game of chance as opposed to one of skill (Ward &
Hill 1991; Belch & Belch 2007), they offer a suitable vehicle for examining rational versus
87
irrational determinants of consumer behaviour. The chance of winning can be manipu-

lated to assess rational consumer calculations and, when paired with irrational belief in
luck, insights can be gained into the respective influences of rational and irrational pro-
pensities in lucky draw participation.
Darke and Freedman (1997a) developed a ‘belief in good luck’ scale that comprises two
subscales: General Belief in Luck, and Belief in Personal Good Luck. Prendergast and
Thompson (2008) have shown that Belief in Personal Good Luck is significantly associ-
ated with lucky draw participation, but General Belief in Luck is not. High scores on the
Belief in Personal Good Luck subscale reflect individuals’ beliefs that they themselves
are personally lucky, and this belief appears to encourage them to participate in activities
requiring luck, such as lucky draws, since they have a heightened, albeit irrational, confi-
dence in their ability to influence the outcome of chance events.
This, Prendergast and Thompson (2008) suggest, is consistent with research results
suggesting that operant conditioning and positive reinforcement have a role in estab-
lishing and maintaining gambling behaviour (Delfabbro & Winefield 2000). Positive
consequences have been shown to increase the probability of repeated gambling, with the
experience of winning generating increased confidence in future luckiness (Gilovich &
Douglas 1986; Ladouceur et al. 1988; Darke & Freedman 1997b; Dixon et al. 1998). Such
irrational beliefs have been found in other areas of sales promotion. Work by Kramer and
Block (2008) demonstrated that superstitious beliefs have a robust influence on decision
making in the face of unknown outcomes. Their experiment indicated that consumers in
the (negative) superstition-salient condition tended to be less willing to gamble than those
in the control condition.
Rational influences
Despite the findings of Prendergast and Thompson (2008), a predominant assumption
underlying many sales promotion designs is that buyers take part in promotions not
because they feel lucky but because they are rational utility maximisers. Discounts, free
gifts, buy-one-get-one-free and purchase coupon sales promotions, for example, each
assume that buyers will rationally calculate that they will be better off by participating. In
the case of a lucky draw, buyers may attempt rationally to calculate winning probabilities.
This would be in line with expected utility theory (EUT), as they would be choosing
rationally between alternatives by putting a value on each outcome and appraising the
probability of achieving that outcome (von Neumann & Morgenstern 1944). According
to this theory, consumers are utility maximisers. If the utility values of two outcomes are
the same, people will choose the outcome with the greatest probability of occurring. This
predicts a linear relationship between expected utility and choice. People rank their prefer-
ences and discard alternatives offering lower expected utility.
Of course, the rationality or irrationality of consumers is still the subject of active inves-
tigation. Some marketing activities, such as signalling, are thought to require highly rational
consumers to be effective (Kirmani & Rao 2000); others such as loyalty programmes are
designed to appeal to consumers who are irrational (e.g. Taylor et al. 2004). There have been
conflicting empirical findings about consumers’ rationality. Ariyabuddhiphongs (2011)
88
argues that it is possible that gamblers do not solely use rationality or expected value in
their gambling. He cites the study by Pelletier and Ladouceur (2007), which showed that
participants with knowledge of mathematics had more erroneous perceptions of gambling.
In another study, improvement in knowledge and skill in gambling odds was not associated
with any decreases in actual gambling behaviour (Williams & Connolly 2006). Gamblers
seemed to switch off their rational belief during gambling and switch it on again after
gambling (Sevigny & Ladoucer 2003). As a result of such studies, some researchers now
dismiss marketing models that assume extreme consumer rationality, while others, at the
opposite extreme, reject models that assume less than perfect rationality (Shugan 2006).
When it comes to lucky draw participation, we argue that consumers ought to be
influenced by both rational and irrational motives. In addition to an irrational belief in
luck (belief in good luck, belief in personal good luck), a rational consideration of the odds
of winning ought also to influence their decision to participate, with higher probabilities
of winning increasing the likelihood of participation. Rational considerations about the
probability of winning ought to therefore complement the individual’s less rational luck
motivations. Our interest here is not so much in how a belief in luck influences lucky draw
participation in and of itself, but rather how rational considerations such as the probability
of winning combine with such irrational beliefs. In the light of the evidence suggesting
that both rational and irrational considerations influence consumer behaviour, and drawing
on theory related to belief in luck as well as EUT, the following hypotheses are proposed:
H1: Rational appraisal of winning odds and irrational belief in good luck will each
independently influence lucky draw participation.
H2: Rational appraisal of winning odds and irrational belief in good luck will inter-
actively influence lucky draw participation.
H3: Rational appraisal of winning odds and irrational belief in personal good luck
will each independently influence lucky draw participation.
H4: Rational appraisal of winning odds and irrational belief in personal good luck
will interactively influence lucky draw participation.
Methods
Procedure
A survey was conducted in Hong Kong to test these hypotheses. A questionnaire was
developed in English, translated into Cantonese, and the translation was then checked
using a translation and back-translation procedure (Brislin et al. 1973; Tamanin et al.
2002). Three versions of the questionnaire were administered, each to a separate sub-
sample. Respondents were asked to imagine that they were buying a mobile phone, and
were offered several different sales promotions in connection with the purchase. They
could accept none of the promotional offers, choose to be entered in a lucky draw, receive
89
a supermarket discount coupon or get a price discount. The lucky draw was thus a non-
monetary offering involving chance, whereas the price discount and coupon represented
a monetary offering not involving chance. The three questionnaires were identical except
for the chances of winning the lucky draw (1 in every 20, 1 in every 100 or 1 in every 200
people winning). The options were presented to the respondents as follows.
You want to buy a mobile phone. You are in SHOP ONE that sells the make and
model you want for HK$3,000. In the same road there are three other mobile phone
shops selling the same phone, but with three different offers.
SHOP TWO sells the same phone for HK$3,000, but you get automatic entry into
a prize draw for a 40 inch LCD TV (1 in every 20, 1 in every 100, or 1 in every
200 people can win).
SHOP THREE sells the same phone for HK$2,980, i.e., with a HK$20 discount.
SHOP FOUR sells the same phone for HK$3,000, but you also get with it a
voucher worth HK$40 at either ParknShop supermarket or Wellcome supermarket.
What would you do – continue to buy the mobile phone in Shop 1 or go to one of
the other shops, and if this latter, which shop would you decide to go to?
Although expected utility theory predicts that the manner of presenting the alternatives
should not influence the choice, it was considered prudent to vary the order of presenta-
tion of the options to counter any potential order effects.
A mobile phone was chosen as the purchasing context because it represents a product that
the majority of Hong Kong adults have had the experience of buying, and because mobile
phone retailers in Hong Kong often use sales promotion techniques. This product also
helped to ensure the respondents’ interest and involvement in the topic (Helgeson et al. 2002;
Van Kenhove et al. 2002) and reduced the chance of uninformed responses (Graeff 2007).
The price discount and coupon (certain prospects) and the lucky draw (uncertain pros-
pect) were chosen as the three sales promotion options not only because they represent
certain and uncertain prospects, but also because they represent the three sales promotion
techniques most commonly used in Hong Kong in relation to mobile phones. To enhance
external validity, interviewees were informed that the coupons are redeemable at the two
largest supermarket chains in Hong Kong (Wellcome and ParknShop), which would be
well known to any Hong Kong resident.
Some justification is required for the choice of Hong Kong as the study location.
Hong Kong is an advanced Chinese city and a source of social influence to neighbour-
ing countries in south-east Asia (Prendergast & Huang 2003). Since its handover to
Mainland China in 1997, it is apparent that there has been increased cultural convergence
between Hong Kong and Mainland China (Barnes et al. 2009). That luck plays a special
role in Chinese culture has long been well understood. As an example, the number 8 is
90
associated with prosperity and good luck, so the 2008 Beijing Olympics commenced at
8 minutes past 8pm on 8/8/08. In line with this, Simmons and Schindler (2003) have
shown that the prices in Chinese advertisements indicate a distinct tendency to end with
an 8. Furthermore, Ang’s (1997) research has revealed that brand names containing a
lucky number of letters were perceived more favourably than those that did not. Lucky
colours, lucky numbers and lucky draws are part and parcel of life in Chinese culture (Shi
et al. 2005). Many festive occasions, such as weddings and Chinese New Year functions,
and most office functions will have a lucky draw as part of the entertainment. The reason
for such luck beliefs is unclear, but it might have its roots in Feng Shui beliefs that, while
originally grounded in the use of astronomy to orientate buildings, have since evolved into
a range of superstitious beliefs. For example, Feng Shui and astrology are concerned with
enhancing qi, an apparently beneficial energy. This qi is usually associated with bright
places, and good luck symbolism is invariably bright, with red being the best colour and
black considered taboo. Given its influence in Asia and the salience of luck beliefs, Hong
Kong was therefore considered a suitable location for a study of this nature.
Sample
A team of 15 field workers was briefed and sent out to shopping malls in three areas of
Hong Kong: Hong Kong Island, Kowloon and the New Territories. These three areas
are the three most populous regions of Hong Kong, and including all of them provides a
representative sample of Hong Kong people. A quota sampling procedure was followed in
an attempt to divide the sample equally between males and females and by age; 702 people
were personally interviewed, with 699 of these responses being deemed suitably valid to
make up the final sample. There were roughly 233 valid interviewees for each version of
the questionnaire. The fieldwork was conducted over a period of three weeks.
Winning probabilities
The winning probabilities that were suggested to the respondents were extensively tested
using 93 students from undergraduate classes. This testing revealed that, even with this
relatively highly educated population, winning probabilities are not easily communicated
or understood. However, the testing confirmed that the 1 in 20 probability was thought to
be higher than the 1 in 100 probability, which in turn was thought to be higher than the 1
in 200 probability. Given these findings, the probability of winning the LCD TV in shop
2 was set at 1 in 20, 1 in 100 or 1 in 200 in the three different versions of the questionnaire.
These three probabilities were selected to ensure that the respondents perceived them as
representing three different levels of chance.
Dependent variable
The dependent variable was the promotion chosen by the respondent: the lucky draw, the
discount or the coupon. Based on the choice, a dichotomous dependent variable was created
91
with 1 representing a respondent’s choosing the lucky draw and 0 choosing the discount or
coupon. Those cases (n = 34) where respondents chose to stay in Shop 1 and enjoy none
of the sales promotion options were removed from the dataset. This was done because our
interest was not to examine whether or not some kind of sales promotion would be chosen,
but rather the selection of a sales promotion with an element of chance against one without.
Belief in luck
Each respondent’s irrational belief in luck was measured using Darke and Freeman’s Belief
in Good Luck Scale (BIGLS; Darke & Freedman 1997a). This scale measures individual
perceptions of the causal properties of luck. A higher score indicates that the respond-
ent thinks of luck as a personal and stable influence in his or her daily life, and one that
they can control. People who have a low score on this scale believe that luck is nothing
more than random chance and that chance outcomes are stochastically determined. The
measure appears to be unrelated to self-esteem, general optimism, academic pessimism,
achievement motivation or desire for control (Darke & Freedman 1997a).
The BIGLS has been found by Darke and Freedman (1997a) to incorporate two sub-
scales that Prendergast and Thompson (2008) have labelled Belief in Personal Good Luck
and General Belief in Luck. In Prendergast and Thompson’s research, the Cronbach’s
alpha representing the internal consistency of the full 12 items of the BIGLS was 0.71.
A 4-item Belief in Personal Good Luck subscale had a Cronbach’s alpha of 0.86, and a
4-item General Belief in Luck subscale had an alpha of 0.68. These values can be con-
sidered satisfactory for predictive research and a 4-item measure (Nunnally & Bernstein
1994). The correlation between the two subscales was low (r = 0.10, p ≤ 0.05), supporting
Darke and Freedman’s (1997b) suggestion that they are largely independent constructs.
Given that Prendergast and Thompson (2008) found that a general belief in luck did
not predict lucky draw participation, this was omitted from the hypotheses and the sub-
sequent analysis, and the focus instead was on the full belief in good luck scale and the
belief in personal good luck.
Controls
Since personal utility for the LCD TV may influence the choice of sales promotion
options, each subject’s utility for an LCD TV was measured and entered as a control
variable in the subsequent analyses. Utility was measured using a 5-item Likert-type
scale, which gauged the extent to which an LCD TV was considered desirable, valuable,
attractive, special and needed. The utility scale proved to have a unidimensional structure
with a Cronbach’s alpha of 0.74. Existing owners of LCD TVs would also be less likely
to choose the lucky draw option. Therefore LCD TV ownership was also identified and
entered as a control variable in the subsequent analyses. Note that, to enable a valid com-
parison between the three versions of the questionnaire, it was necessary to hold constant
the payoff of the lucky draw (altering the prize in each of the lucky draw scenarios while
holding the other sales promotion options constant would have created obvious problems
92
in terms of controlling for the effect of prize utility). Therefore the prize in the lucky
draw (an LCD TV) was kept constant across the three versions of the questionnaire, as
too was the value of the price discount and coupon. Only the probability of winning the
lucky draw was manipulated.
Germane demographic variables were also measured and entered as control variables
in the subsequent analysis. Lower-income people may feel compelled to seek monetary
incentives that have a high degree of certainty (such as a discount or coupon), so personal
income was controlled for. Despite the observations with the college students, it is pos-
sible that more educated people may have a clearer understanding of probabilities and
expected value, and thus might better appreciate that even the least attractive odds in the
survey (1 in 200) still represented a quite good chance of winning. Therefore education
was controlled for. Because studies of gambling propensity in Hong Kong suggest that
higher rates of gambling are associated with males (Wong & So 2003), sex was controlled
for. Andre (2006) has reported differences in luck perceptions with age, so age was also
controlled for using, in view of the extreme reluctance of some female respondents in
Hong Kong to give their exact age, five-year categories.
Results
Looking first at the sample’s demographics, 55% of the interviewees were aged less than
40. Thirty-five per cent of the interviewees claimed to have attained a high school edu-
cation, with 25% claiming less than that, and 40% claiming to have completed tertiary
education (technical college or university). Around one-third of the interviewees said they
earned less than HK$8,000 (US$1,000) per month, while around one-third claimed to
earn HKD$10,000 (US$1,300) or more per month.
Based on the three versions of questionnaire administered, two dichotomous dummy
variables were created to represent the odds of winning. The first was coded 1 to represent
a 1 in 20 chance of winning, and 0 for a 1 in 100 or 1 in 200 chance of winning; the second
dummy was coded 1 for a 1 in 200 chance of winning, and 0 to represent a 1 in 20 or 1 in
100 chance of winning. The 1 in 100 odds was used as the reference category.
Preliminary analysis of the data revealed that any significant correlations were low (i.e.
less than 0.30), suggesting that multicollinearity was not a problem (Tabachnick & Fidell
2001). To test the hypotheses, the relationships between belief in good luck, belief in per-
sonal good luck and the odds of winning the lucky draw, and the dichotomous dependent
variable, were analysed. Given this dichotomous dependent variable, hierarchical logistic
regression was employed. Tables 1 and 2 show hierarchical logistic regression coefficients.
Table 1 focuses on belief in good luck while Table 2 focuses on belief in personal good luck.
Looking first at Table 1, Models 1–7 examine the independent and interactive effects
of belief in good luck, the 1 in 20 odds dummy and the 1 in 200 odds dummy. Model 1
regressed the lucky draw dummy against the controls to form a baseline model against
which to examine the independent effects of the odds variables on selecting the lucky
draw option. Model 2 added only belief in good luck to the controls, which had a sig-
nificant positive beta of 0.72, and added to the pseudo R2 (Δ Nagelkerke’s R2 = 0.014).
93
94
Table 1: Effects of belief in good luck, and high and low odds related to lucky draw selection
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7

Constant –0.45 (0.71) –3.68 (1.88)** –0.56 (0.72) –0.36 (0.71) –0.57 (0.73) –3.80 (1.20)** –0.62 (0.72)
Control variables
Income dummy 1 –0.33 (0.21) –0.29 (0.21) –0.34 (0.21) –0.34 (0.21) –0.34 (0.21) –0.31 (0.21) –0.30 (0.21)
Income dummy 2 –0.27 (0.28) –0.28 (0.28) –0.23 (0.28) –0.24 (0.28) –0.23 (0.28) –0.24 (0.29) –0.24 (0.29)
Education dummy –0.27 (0.22) –0.24 (0.23) –0.29 (0.23) –0.26 (0.22) –0.29 (0.23) –0.26 (0.23) –0.27 (0.23)
Age 0.00 (0.04) 0.00 (0.04) 0.00 (0.04) 0.00 (0.04) 0.00 (0.04) 0.00 (0.04) 0.00 (0.04)
Sex dummy –0.71 (0.19)*** –0.74 (0.19)*** –0.70 (0.19)*** –0.71 (0.19)*** –0.70 (0.19)*** –0.73 (0.19)*** –0.73 (0.19)**
LCD TV dummy 0.00 (0.25) –0.01 (0.25) –0.03 (0.25) –0.01 (0.25) –0.03 (0.25) –0.04 (0.25) –0.06 (0.26)
Prize utility 0.57 (0.22)* 0.55 (0.22)* 0.61 (0.22)** 0.56 (0.22)** 0.61 (0.22)** 0.58 (0.22)* 0.58 (0.22)*
Independent variables
Belief in good luck – 0.72 (0.29)* – – – 0.72 (0.29)* 0.64 (0.33)*
Odds dummy 1 – – 0.45 (0.19)* – 0.45 (0.22)* 0.42 (0.19)* 0.42 (0.19)**
Odds dummy 2 – – – –0.23 (0.20) 0.01 (0.24) – –
Belief in good luck × odds dummy 1 – – – – – – 0.14 (0.35)
Log–likelihood 725 718 719 723 719 713 719
Nagelkerke’s R2 0.068 0.082 0.080 0.071 0.080 0.092 0.092
Δ Nagelkerke’s R2 from Model 1 0.014 0.012 0.003 0.012 0.024 0.024
The dependent variable was ‘opted for lucky draw participation rather than a discount or coupon’, lucky draw coded 1. Gender was coded female = 1 and others = 0; education was coded 1 = tertiary
or above and 0 = secondary or below; LCD ownership was coded 0 = no current LCD TV and 1 = currently have an LCD TV. Odds dummy 1 was coded 1 = 1 in 20, and 0 = 1 in 100 or 1 in 200; Odds
dummy 2 was coded 1 = 1 in 200, and 0 = 1 in 20 or 1 in 100. For personal income, the contrast was HK$10,000–19,999. Standard errors are shown in parentheses.
N = 665. *Signifies significance at the p ≤ 0.05 (**p ≤ 0.01, ***p ≤ 0.001) level.
Model 3 added only the 1 in 20 odds dummy, and it also had a significant positive beta,
of 0.45, adding 0.012 to the pseudo R2. The evidence from Model 3 supports part of H1,
that rational appraisal of winning odds will influence lucky draw participation. Support
for this part of H1 is further reinforced by Model 4, which indicates that the 1 in 200
odds dummy was not significantly related to lucky draw participation. In Model 5, both
odds dummies were added, with only the 1 in 20 odds dummy having a significant posi-
tive beta, again supporting part of H1.
Model 6 adds both belief in good luck and the 1 in 20 odds dummy, and both have
significant positive betas and contribute to the higher pseudo R2, supporting H1’s sugges-
tion that a rational appraisal of winning odds and irrational belief in good luck will each
independently influence lucky draw participation. Model 7 adds the interaction between
belief in good luck and the 1 in 20 odds dummy. Belief in good luck and the 1 in 20 odds
dummy both maintained their significant positive betas, but their interaction was not
significant. H2, that a rational appraisal of winning odds and irrational belief in good luck
will interactively influence lucky draw participation, was not supported.
Turning to Table 2, these results concern specifically belief in personal good luck.
Model 1 is again a baseline model including only controls. Models 8–10 examine the
Table 2: Effects of belief in personal good luck, and low odds related to lucky
draw selection
Model 1 Model 8 Model 9 Model 10

Constant –0.45 (0.71) –3.14 (1.05)** –3.29 (1.06)** –0.62 (0.72)
Control variables
Income dummy 1 –0.33 (0.21) –0.29 (0.21) –0.30 (0.21) –0.30 (0.21)
Income dummy 2 –0.27 (0.28) –0.28 (0.28) –0.24 (0.28) –0.24 (0.28)
Education dummy –0.27 (0.22) –0.27 (0.23) –0.29 (0.23) –0.29 (0.23)
Age 0.00 (0.04) 0.00 (0.04) 0.00 (0.04) 0.00 (0.04)
Sex dummy –0.71 (0.19)*** –0.78 (0.19)*** –0.76 (0.19)*** –0.76 (0.19)***
LCD TV dummy 0.00 (0.25) –0.06 (0.25) –0.10 (0.26) –0.09 (0.23)
Prize utility 0.57 (0.22)* 0.56 (0.22)* 0.59 (0.22)** 0.59 (0.22)**
Belief in personal good luck – 0.51 (0.19)** 0.50 (0.19)** 0.52 (0.23)**
Odds dummy 1 – – 0.43 (0.19)* 0.43 (0.19)*
Belief in personal good luck ×
– – – –0.07 (0.39)
Odds dummy 1
Log–likelihood 725 717 711 712
Nagelkerke’s R2 0.068 0.084 0.095 0.095
Δ Nagelkerke’s R2 from Model 1 0.016 0.027 0.027
The dependent variable was ‘opted for lucky draw participation rather than a discount or coupon’, lucky draw coded 1. Gender
was coded female = 1 and others = 0; education was coded 1 = tertiary or above and 0 = secondary or below; LCD ownership
was coded 0 = no current LCD TV and 1 = currently have an LCD TV. Odds dummy 1 was coded 1 = 1 in 20, and 0 = 1 in 100 or
1 in 200. For personal income, the contrast was HK$10,000–19,999. Standard errors are shown in parentheses.
N = 665. *Signifies significance at the p ≤ 0.05 (**p ≤ 0.01, ***p ≤ 0.001) level.
95
independent and interactive effects of belief in personal good luck and the 1 in 20 odds
dummy. Model 8 indicates that belief in personal good luck is positively related to lucky
draw participation, with a significant positive beta and a change in the pseudo R2 of 0.016.
Model 9 adds the 1 in 20 odds dummy. In this model, belief in personal good luck main-
tained its significance, while the 1 in 20 odds dummy was also significant, supporting H3
(rational appraisal of winning odds and irrational belief in personal good luck will each
independently influence lucky draw participation). Model 10 includes belief in personal
good luck, the 1 in 20 odds dummy and their interaction. Belief in personal good luck and
the 1 in 20 odds dummy maintained their significant positive betas, but their interaction
was not significant. H4, that a rational appraisal of winning odds and irrational belief in
personal good luck will interactively influence lucky draw participation, was not supported.
Discussion and contribution

The findings indicate that an irrational belief in luck and a rational appraisal of winning
odds independently influence consumer participation in a lucky draw promotion, suggest-
ing that buyers are simultaneously rational and irrational. Belief in good luck and a belief
in personal good luck are both positively related to lucky draw participation. In addition
to belief in good luck and a belief in personal good luck, and consistent with expected
utility theory, a rational consideration of the odds of winning also influences lucky draw
participation. That there is no interaction between the rational and irrational dimensions
suggests that irrationality and rationality are discrete psychological traits that operate
independent of each other. Those who believe in good luck or personal good luck are not
necessarily more attracted to good odds than those who do not believe in good luck or
personal good luck.
From a theoretical perspective, this study has extended knowledge of the range of
psychological variables that influence sales promotion preferences, and has taken a small
step towards resolving the debate raised by Shugan (2006) about whether consumers are
rational or irrational. Consumers seem, in fact, to be both rational and irrational at the
same time. Knowledge of these rational and irrational influences should facilitate predic-
tion of those circumstances in which consumers are likely to participate in a lucky draw.
In order for the findings to have meaning to a broader audience, they also need to
have real-world implications (Taylor 2011). Marketers planning lucky draw promotions
need to design them to appeal to both consumers’ rational and irrational decision making.
Initially, this means that marketers considering a lucky draw promotion ought to consider
doing preliminary research to determine the extent to which their market and market
segments believe in luck. For market segments with a strong belief in personal good luck,
marketers might try to ‘personalise’ the lucky draw process so that consumers feel they can
exert what they regard as their personal good luck to influence the outcome. Participants
might, for example, be offered their own choice of ‘lucky’ numbers or colours when it
comes to their choosing their entry form or number (Prendergast & Thompson 2008).
Catering to luck beliefs, however, is not the only way to influence lucky draw participation.
Even when consumers believe in luck, they will still be rationally attracted to a draw that
96
offers a good chance of winning. Marketers may therefore consider offering more prizes
of less value rather than fewer prizes of greater value, to increase the odds of winning
something. And it might serve them well to publicise the odds of winning if they are good
or are perceived to be good, and omit mentioning the odds if they are not, where this is
legally permissible. Improving the odds of winning, and therefore increasing the number
of winners, may also help strengthen consumers’ belief in personal good luck, which will in
turn encourage future participation.

This research was limited in its geographical context (Hong Kong), and the mall-
intercept sample was not perfectly random. As mentioned previously, luck plays a special
role in Chinese culture, be it Mainland China or Hong Kong. Yet although luck beliefs
are pronounced in Hong Kong, this is not to say that luck beliefs do not exist in other
cultures. Luck beliefs are probably universal to a greater or lesser degree, and theoretically
there is no reason to expect that our findings would not be replicated in a culture that
does not place such a high priority on luck beliefs: our sample was not composed of luck
believers, but individuals who vary in their luck beliefs. A second limitation of the study
is that the research was based on a single buying scenario. The extent to which the find-
ings can be generalised to other product categories or to other prize, discount and coupon
combinations requires further research. Third, we did not control for hedonic utility, since
we assumed that the propensity to participate in lucky draws because of this utility is
evenly distributed throughout any population. Fourth, research suggests that situational
influences such as mood might influence perceptions of promotional incentives (Hsu &
Liu 1998), with consumers’ perceptions of transaction value being greater when they are
in a positive mood than when they are in a negative mood. Such situational influences
were not controlled for in this study. Fifth, while the relatively large sample facilitates
robust analyses, it might be averred that it produces significant but small effect sizes in
terms of variance explained. However, our aim has been to examine the respective effects
of the rational and irrational variables in lucky draw selection rather than to build a more
comprehensive model of variables affecting such selection, and in doing this, a larger
sample is to be preferred for producing reliable results. Finally, this research considered
one dimension of rationality (winning odds) and one dimension of irrationality (belief in
luck). Further research is required to determine whether or not other types of irrationality
and rationality operate in the same independent manner.
Conclusions
It has been shown in previous research that belief in luck influences lucky draw participa-
tion. Expected utility theory, however, suggests that rational considerations may also come
into play. Our study has shown that both irrational and rational influences independently
affect lucky draw participation. Belief in luck is irrational, but it is not divorced from
rationality. This study has extended previous research on sales promotion by considering
97
the dual roles of rational and irrational influences in predicting lucky draw participa-
tion. Further research is now required in order to generate evidence either verifying or
refuting our finding that lucky draw selection is a combination of rationality and
irrationality.
References
Andre, N. (2006) Good fortune, luck, opportunity and their lack: how do agents perceive
them? Personality and Individual Differences, 40(7), 1461–1472.
Ang, S.H. (1997) Chinese consumers’ perceptions of alpha-numeric brand names. Journal of
Consumer Marketing, 14(3), 220–233.
Ariyabuddhiphongs, V. (2011) Lottery gambling: a review. Journal of Gambling Studies, 27,
1533.
Barnes, B.R., Siu, N.Y.M., Yu, Q. & Chan, S.S.Y. (2009) Exploring cosmetics advertising in
southern China: a comparative investigation of Hong Kong and Shenzhen. International
Belch, G. & Belch, M. (2007) Advertising and Promotion – An Integrated Marketing
Communications Perceptive – International Edition (7th edn). New York: McGraw-Hill
Irwin.
Brislin, R.W., Lonner, W.J. & Thorndike, R.M. (1973) Cross-Cultural Research Methods.
New York: Wiley Publishing.
Chandon, P., Wansink, B. & Laurent, G. (2000) A benefit congruency framework of sales
promotion effectiveness. Journal of Marketing, 64(October), 65–81.
Darke, P.R. & Freedman, J.L. (1995) Nonfinancial motives and bargain hunting. Journal of
Applied Social Psychology, 25(l8), 1597–1610.
Darke, P.R. & Freedman, J.L. (1997a) The belief in good luck scale. Journal of Research in
Personality, 31, 486–511.
Darke, P.R. & Freedman, J.L. (1997b) Lucky events and beliefs in luck: paradoxical effects on
confidence and risk-taking. Personality and Social Psychology Bulletin, 23(4), 378–388.
Delfabbro, P.H. & Winefield, A.H. (2000) Predictors of irrational thinking in regular slot
machine gamblers. Journal of Psychology, 134, 117–128.
Dixon, M.R., Hayes, L.J., Rehfeldt, R.A. & Ebbs, R.E. (1998) An adjusting procedure for
studying outcomes of risk-taking. Psychological Reports, 82, 1047–1050.
Gilovich, T. & Douglas, C. (1986) Biased evaluations of randomly determined gambling
outcomes. Journal of Experimental Social Psychology, 22, 228–241.
Graeff, T.R. (2007) Reducing uninformed responses: the effects of product-class familiarity and
measuring brand knowledge on surveys. Psychology & Marketing, 24(8), 681–702.
Haugh, L.J. (1983) Defining and redefining. Advertising Age, 14 February, 44.
Helgeson, J.G., Voss, K.E. & Terpening, W.D. (2002) Determinants of mail-survey response:
survey design factors and respondent factors. Psychology & Marketing, 19(3), 303–328.
Hsu, C. & Liu, B. (1998) The role of mood in price promotions. Journal of Product and Brand
Management, 7(2), 150–160.
Jiang, Y., Cho, A. & Adaval, R. (2009) The unique consequences of feeling lucky: implications
for consumer behavior. Journal of Consumer Psychology, 19(2), 171–184.
Kirmani, A.Y. & Rao, A.R. (2000) No pain, no gain: a critical review of the literature on
signaling unobservable product quality. Journal of Marketing, 64 (April), 66–79.
Kramer, T. & Block, L. (2008) Conscious and nonconscious components of superstitious beliefs
in judgment and decision making. Journal of Consumer Research, 34(April), 783–793.
98
Ladouceur, R., Gaboury, A., Dumont, M. & Rochette, P. (1988) Gambling: relationship
between the frequency of wins and irrational thinking. Journal of Psychology, 122, 409–414.
Landman, J. & Petty, R. (2000) It could have been you: how states exploit counterfactual
thought to market lotteries. Psychology & Marketing, 17(4), 299–321.
Laroche, M., Pons, F., Zgolli, N., Cervellon, M. & Kim C. (2003) A model of consumer
response to two retail promotion techniques. Journal of Business Research, 56( July), 513–522.
Nunnally, J.C. & Bernstein, I.H. (1994) Psychometric Theory (3rd edn). New York: McGraw-Hill.
Pelletier, M.F. & Ladouceur, R. (2007) The effect of knowledge of mathematics on gambling
behaviours and erroneous perceptions. International Journal of Psychology, 42(2), 134–140.
Prendergast, G.P. & Huang, C.H. (2003) An Asian perspective of offensive advertising on the
web. International Journal of Advertising, 22(3), 393–411.
Prendergast, G.P. & Thompson, E.R. (2008) Sales promotion strategies and belief in luck.
Psychology & Marketing, 25(11), 1043–1062.
Rust, R.T., Ambler, T., Carpenter, G., Kumar, V. & Srivastava, R.K. (2004) Measuring
marketing productivity: current knowledge and future directions. Journal of Marketing,
October, 76–89.
Sevigny, S. & Ladoucer, R. (2003) Gamblers’ irrational thinking about chance events: the
‘double switching’ concept. International Gambling Studies, 3, 149–161.
Shi, Y.Z., Cheung, K.M. & Prendergast, G.P. (2005) The effectiveness of sales promotion tools
in Hong Kong. International Journal of Advertising, 24(4), 467–486.
Shugan, S.M. (2006) Are consumers rational? Experimental evidence? Marketing Science, 25(1),
1–7.
Simmons, L.C. & Schindler, R.M. (2003) Cultural superstitions and the price endings used in
Chinese advertising. Journal of International Marketing, 11(2), 101–111.
Simpson, L. (2006) Enhancing food promotion in the supermarket industry: a framework for
sales promotion success. International Journal of Advertising, 25(2), 223–246.
Tabachnick, B.G. & Fidell, L.S. (2001) Using Multivariate Statistics. (4th edn). New York:
Allyn & Bacon.
Tamanin, T., Ancona, C. & Rodrigues-Netto, N. (2002) Translation, validation and cross-
cultural adaptation into Portuguese language of the King’s Health Questionnaire.
Heidelberg: International Continence Society Annual Conference.
Taylor, C.R. (2011) Editorial: on advertising research that makes an impact. International
Taylor, C.R. & Kopp, S.W. (1991) Games, contests, and sweepstakes run afoul: a state of legal
disorder. Journal of Public Policy & Marketing, 10(Spring), 199–213.
Taylor, S.A., Celuch, K. & Goodwin, S. (2004) The importance of brand equity to customer
loyalty. Journal of Product and Brand Management, 13(4), 217–227.
Van Kenhove P., Wijnen, K. & De Wulf, K. (2002) The influence of topic involvement on
mail-survey response behavior. Psychology & Marketing, 19(3), 293–301.
von Neumann, J. & Morgenstern, O. (1944) Theory of Games and Economic Behavior. Princeton,
NJ: Princeton University Press.
Ward, J.C. & Hill, R.P. (1991) Designing effective promotional games: opportunities and
problems. Journal of Advertising, 20(3), 69–81.
Williams, R.J. & Connolly, D. (2006) Does learning about the mathematics of gambling
change gambling behavior? Psychology of Addictive Behaviors, 20, 62–68.
Wisman, J.D. (2006) State lotteries: using state power to fleece the poor. Journal of Economic
Issues, XL(4), 955–966.
Wong, L.K. & So, E.M. (2003) Prevalence estimates of problem and pathological gambling in
Hong Kong. American Journal of Psychiatry, 160(7), 1353–1354.
99
About the authors

Gerard Prendergast (PhD, Massey) is Professor and Head, Department of Marketing,
Hong Kong Baptist University. He previously held faculty positions in a variety of coun-
tries, having taught Bachelor, Master and Doctoral programmes in the UK, Singapore,
Africa and the Middle East. His research focuses on marketing communications and he
has published in journals such as The Journal of Advertising, Journal of Advertising Research,
Psychology & Marketing, International Journal of Advertising, Public Relations Review and
the European Journal of Marketing.
Edmund R. Thompson (PhD, London School of Economics) is Professor of
International Management at the School of Management, University of Bath, England.
Previously he was Professor of International Business at the Graduate School of
Management, Ritsumeikan University, Japan, and earlier a faculty member at the School
of Business, University of Hong Kong, and the National University of Singapore. He
has published in numerous international business, marketing, psychology and economics
journals, including Psychology & Marketing, Entrepreneurship Theory & Practice, Journal of
Cross-Cultural Psychology, Economic Geography and Group & Organization Management.
Address correspondence to: Gerard P. Prendergast, Professor and Head, Department of
Marketing, Hong Kong Baptist University, Kowloon Tong, Hong Kong.
Email: gerard@hkbu.edu.hk
100
Communication using celebrities
in the non-profit sector
Determinants of its effectiveness
Maria del Mar Garcia de los Salmones,
Rafael Dominguez and Angel Herrero
University of Cantabria
Nowadays celebrity endorsement has become a popular marketing technique in the non-profit
sector. However, there is still a degree of light and shade as regards the determinants of the
effectiveness of this communication strategy. Taking into account the significant lack of stud-
ies in this particular field, the current research presents a causal model with nine hypotheses,
which analyse the sequence of relationships that cover from the background of celebrity cred-
ibility, to the determinants of the attitude towards adverts and the intention to collaborate. We
surveyed 329 adults, who rated a fictitious ad with a celebrity without perceived experience in
the non-profit area. As a result, it is observed that the celebrity’s credibility depends on the fit
perceived, the attribution of altruistic motivation, the celebrity image and the general attitude
towards celebrity activism. Furthermore, it is found that a credible celebrity has a strong influ-
ence on the attitude towards the advert and, indirectly, on behavioural intentions.
Introduction
Many things have changed since Kotler (1979, p. 38) said ‘Third Sector administrators
must begin to think like marketers’, in the sense that the same tactics and strategies
employed to market goods and services can be used to sell social ideas (Hill & Moran
2011). Nowadays, the non-profit sector is becoming a very competitive and a fast-growing
advertising area (Wheeler 2009), and organisations are positioning themselves, communi-
cating their values and struggling to raise funds through communication (Becker-Olsen
& Hill 2006). In this sophisticated environment of fundraising competition, conceptual
and empirical guidance is needed when designing charity appeals (Chang & Lee 2010)
in order to maximise the impact of the communication strategy.
In this context, the use of celebrities in advertising is a very common practice worth
studying. Specifically, a celebrity endorser has been defined as any individual who enjoys
public recognition (actors, athletes, entertainers, models, politicians and singers, among
others) and who uses this recognition on behalf of a consumer good by appearing with
it in an advert (McCracken 1989). This definition reflects the main line of research so

DOI: 10.2501/IJA-32-1-101-119
far, i.e. the analysis of the effects and determinants of success of celebrity endorsements
within the context of manufacturing organisations and consumer brands. However, and
despite the fact that nowadays the relationship between celebrities and social causes has
no precedents, there is an important gap with regard to the impact and effectiveness of
these endorsers on the non-profit sector (Samman et al. 2009; Wheeler 2009; Van den
Bulck et al. 2010).
The idiosyncrasy of this sector makes this line of research interesting. Specifically,
non‑profit organisations (NPOs) rely on celebrities because they are very effective fund-
raisers; they have skills in mass communication, so they may reach a wide audience, attract
people’s attention, raise awareness of social issues and even access the upper echelons of
political power, among other issues (Samman et al. 2009; Wheeler 2009; Dominguez
2010). For these reasons, NPOs look for celebrities who are prepared to get involved in
their causes while, in turn, celebrities seek to join an NPO or create their own foundations.
They do not receive money for that, but other rewards. According to Dominguez (2010),
celebrity activism can be explained by the globalisation of communication, their growing
autonomy to promote their own personal brand, and the need to differentiate themselves,
gain publicity and promote their image. Apparently with this collaboration both celebri-
ties and organisations win; however they must be careful because the activism may be as
effective as questionable, and this alliance may cause resistance and scepticism on the part
of the consumer with regard to the real motives – selfish versus altruistic – of the celebrity
(Petty et al. 1983). This duality may have repercussions for consumer response.
Taking into account these issues and the lack of studies in this particular area, the object-
ive of the paper is to go more deeply into the effectiveness of celebrity endorsement in
advertising campaigns for humanitarian aid organisations. Specifically, the paper presents a
causal model with nine hypotheses, and analyses the determinants of the attitude towards
advertising with celebrities and the influence of this strategy over intentional behaviour. The
research considers the perceived fit between celebrity and social cause (Kamins & Gupta
1994; Lynch & Schuler 1994), the motivation attribution (Folkes 1988; Becker-Olsen et al.
2006), the attitude towards the celebrity (Choi & Rifon 2007; Samman et al. 2009), and the
general attitude towards the involvement of celebrities in social causes – general perceived fit
– as potential antecedents of endorser credibility. Moreover, it is expected that both celebrity
credibility and the credibility of the organisation influence the attitude towards advertis-
ing (Lafferty & Goldsmith 1999; Goldsmith et al. 2000) and the intention to collaborate
(MacKenzie & Lutz 1989; Ranganathan & Henley 2007). All these relationships will be
measured and analysed with a structural equations model.
Review of the literature and hypothesis for research
Celebrity and corporate credibility as direct antecedents of the

attitude towards an advertisement
Celebrities are traditionally considered to be sources with high persuasive power. NPOs
try to influence individuals through the emotions, so these endorsers can be very effective
102
Communication using celebrities in the non-profit sector
due to their power of attraction, and their ability to produce feelings and emotional
responses (Assael 1984). In the non-profit sector, and according to the results of Samman
et al. (2009), most people consider that celebrity involvement in charity can be beneficial.
However, this must be handled carefully because celebrity activism is also criticised, in the
sense that many people are sceptical about the real motives of the alliance and question
celebrities’ sincerity or legitimacy, and their opportunistic behaviour (Samman et al. 2009).
Taking into account these issues, celebrity credibility becomes, more than ever, a key vari-
able for the success of communication.
Credible endorsers enhance advertising effectiveness (Eisend & Langner 2010)
because, when their credibility is high, consumers are more likely to accept arguments pre-
sented in the ad, whereas they tend to reject arguments coming from a source with a low
degree of credibility (Grewal et al. 1994). Specifically, the credibility of the source is com-
posed of two dimensions: expertise and trustworthiness (Hovland et al. 1953). ‘Expertise’
is a cognition-based attribute (Eisend & Langner 2010), derived from knowledge of
the subject, and it refers to skill and competence in the communication topic, whereas
‘trustworthiness’ refers to the honesty and believability of the source when expressing an
objective opinion on the product (McGinnies & Ward 1980). This model of source cred-
ibility has been the most widely used in the research on this topic, although there are other
proposals that include attractiveness as a third dimension (Ohanian 1990).
The direct influence of endorser credibility on advertising effectiveness and attitudes
towards ad has been extensively proven in the profit sector (Goldberg & Hartwick 1990;
Lafferty & Goldsmith 1999; Stafford et al. 2002), but not in the NPO area. In fact,
research has shown that celebrity credibility, and therefore its effectiveness, varies by prod-
uct (Stafford et al. 2002). It is interesting, therefore, to analyse the product ‘social cause’,
and the relationship between endorser credibility and attitude towards the ad. Taking into
account the lack of literature on this issue, we consider the previous arguments and the
results obtained in the profit area in order to propose that:
H1: In the non-profit sector, celebrity credibility is directly and positively related to
the attitude towards the advertisement.
Together with the endorser, the company can be considered another source of the mes-
sage (Newell & Goldsmith 2001). When consumers are familiar with a company (due
to information and experience acquired over time) they develop perceptions about its
credibility, and these perceptions and thoughts will make up part of the information indi-
viduals use to evaluate new information received on it – for example, ads – (MacKenzie
& Lutz 1989; Goldsmith et al. 2000). Specifically, what consumers think about a brand
directly influences their attitudes towards the brand’s ad, in the sense that firms with
good credibility are in a better position to get people to respond positively and believe
their advertising claims (LaBarbera 1982; Goldberg & Hartwick 1990). Lafferty and
Goldsmith (1999), Goldsmith et al. (2000) and Lafferty et al. (2002) found precisely that
corporate credibility – composed of two dimensions: perceived expertise and trustwor-
thiness – positively influences attitudes towards an ad. More specifically, these authors
103
describe the causal relationship that exists between corporate and endorser credibility
and attitude towards an ad, resulting in the Dual Credibility Model. Despite the lack of
studies focused on the non-profit sector, we consider these arguments totally valid for
this sector, and take into account the Dual Credibility Model when proposing the second
research hypothesis:
H2: In the non-profit sector, the credibility of the organisation is directly and posi-
tively related to the attitude towards the advertisement.
The determinants of celebrity credibility as indirect antecedents of the

attitude towards the advertisement
Celebrity credibility is a key variable in the success of communication, so we devote this
section to analysing its determinants. Specifically, we study four potential antecedents,
both direct (celebrity–social cause fit and motivation attribution) and indirect (attitude
towards the celebrity and general perceived fit).
Direct antecedents of celebrity credibility: f it and motivation attribution

One of the main variables with influence over the success of advertising communication
with endorsers is match, fit or congruence (Kamins & Gupta 1994; Lynch & Schuler
1994; Amos et al. 2008), referring to the degree of similarity and compatibility that con-
sumers perceive between a social cause and an endorser (Lafferty 2007). In this sense, the
fit can be, on the one hand, functional, if there is similarity between the functions of the
product and the event or social cause supported (Gwinner 1997; Lafferty et al. 2004), or
between the advertised product and the celebrity – for example, an energy bar appropriate
for physical training and a renowned athlete (Till & Busler 2000) – or, in the non-profit
area, the case of Michael J. Fox, a Parkinson’s disease victim, supporting a Parkinson’s
disease association, an example used in the Wheeler (2009) research. On the other hand,
there is image congruence when the two parts of the alliance share the same image or the
same core values. This type of compatibility is more symbolic (Bigné et al. 2009) and it
represents the consumer’s perception that the pairing is acceptable for reasons different
from the link with the product’s characteristics, overriding any functional incompatibility
triggers (Trimble & Rifon 2006). Of the two types, the latter one is the more determinant
(Gwinner & Eaton 1999; Lafferty et al. 2004) and, in fact, it is advisable that the adver-
tising agency or company first determines the symbolic values of the product, and then
chooses a celebrity who represents those relevant symbolic properties (Byrne et al. 2003).
In the non-profit sector the image fit is, in our opinion, more prominent because, as we
mentioned previously, nowadays it is common for celebrities to support a cause, regardless
of their functional fit. We will therefore focus the study on this type of congruence.
In general, a good fit enhances celebrity credibility, and consequently the effectiveness
of communication – an affirmation supported by the Schema Theory (Rumelhart 1980).
According to this theory, individuals use categories – schemas – to organise and structure
the information in the memory. These schemas become an instrument that makes it
104
possible to process and store a large amount of data in the memory without a great cog-
nitive effort (Gwinner & Eaton 1999), guiding perception and consumer thinking. For
example, celebrities can make up schemas since they are well known to consumers due
to their work and frequent appearances in the media, so individuals have pre-established
knowledge and attitudes towards them, even before they appear in advertising (Choi &
Rifon 2007). When new information or a new stimulus (e.g. the social cause) is linked
to a pre-existing schema (e.g. the celebrity), this will produce different effects on people,
depending on the perceived fit between them (Lynch & Schuler 1994). Specifically, when
there is a good connection between consumers’ experience and cognitive consistency with
their existing schemas, they are more likely to perceive the endorser as having more skills
and experience when it comes to associating them with the cause (Bigné et al. 2009),
moreover producing feelings of sincerity that increase the credibility of the endorser
(Becker-Olsen et al. 2006). In contrast, when there is no fit, consumers will engage in
more elaboration on the ad in an attempt to resolve the perceived incoherencies (Lynn
& Schuler 1994; Gwinner & Eaton 1999), trying to assimilate the new information in
the pre-existing schema. Consumers value consistency and respond negatively to lack of
it; therefore their thoughts will include negative feelings towards any observed discon-
nectedness, resulting in unfavourable attitudes towards both partners (Becker-Olsen et al.
2006) and poorer evaluation of company credibility (Webb & Mohr 1998). Therefore, it
can be said that the perception of the credibility of the source is a function of the con-
gruence of the image of the spokesperson in relation to the product image (Kamins &
Gupta 1994), an affirmation especially valid in the non-profit sector. This argument leads
to Hypothesis 3:
H3: In the non-profit sector, a good image fit between celebrity and social cause is
directly and positively related to celebrity credibility.
Second, a very important determinant of credibility in the non-profit area may be

the motivation attribution (Kelley 1973), whereby consumers will try to understand why
companies (or celebrities) link themselves with social causes (Folkes 1988; Becker-Olsen
et al. 2006). Specifically, the recipients will perceive altruistic or intrinsic motives if they
consider that the brand/endorser is supporting a cause because he/she wants to benefit
society or raise awareness for that specific cause. On the other hand, consumers will per-
ceive selfish or extrinsic motives if they think that the company/endorser is seeking only
to make a profit, increase sales, boost their image (Moore et al. 1994; Drumwright 1996;
Becker-Olsen et al. 2006) or, in the specific area of celebrities and NGOs, to obtain their
own publicity and self-promotion (Samman et al. 2009). When people attribute selfish
motivations to brands – or celebrities – they react badly because they perceive that they
are being manipulated or deceived (Forehand & Grier 2003), and the endorser will be
less credible (Mowen & Brown 1981; Sparkman 1982; Moore et al. 1994). By contrast,
attributing an altruistic motive to the sponsor of the social cause is an indicator that
helps individuals to evaluate honesty, sincerity and the good faith of the endorser (Webb
& Mohr 1998), positively influencing credibility (Rifon et al. 2004; Ellen et al. 2006;
105
Bigné et al. 2009). These arguments are in line with traditional literature on persuasion
in social psychology (Hovland et al. 1953), according to which a source without apparent
self-interest in the promoted message will be more credible. All these arguments lead to
Hypothesis 4:
H4: In the non-profit sector, the attribution of altruistic motivations to the celebrity
is directly and positively related to his or her credibility.
Finally, a direct relationship is expected between the fit perceived and the motivation
attribution. In the profit sector, consumers are likely to understand that companies need
to make a profit to survive (Rifon et al. 2004), so it is expected that, to a certain extent,
individuals link the social initiatives of firms with a marketing strategy mainly motivated
by corporate self-interest (Speed & Thompson 2000). However, the sponsor is hoping
that consumers, when exposed to the message, will react with minimal perception of
these extrinsic motives, in order to achieve more effective communication. For this reason,
the congruence or fit between a sponsor and the cause will be one of the key elements.
This fact, which can be extrapolated to the relationships between celebrities and social
causes, is explained by the Schema Theory. Specifically, and just as has been mentioned
previously, a lack of fit would stimulate the elaboration of the message, and this greater
cognitive production would make consumers scrutinise in greater depth the reasons that
have led the brand – or in the case analysed, the celebrity – to associate him or herself
with social causes (Menon & Hahn 2003; Bigné et al. 2009). As a result, this is likely to
raise consumer doubt and scepticism as to their real motives (Petty et al. 1983), inferring
selfish attributions on the part of the brand (Hastie 1984; Speed & Thompson 2000) or
celebrity. In contrast, a high degree of congruence may not generate as many elabora-
tions, which could minimise the consumer’s judgement about endorser motives. In such
cases individuals rely on peripheral persuasion routes and express opinions on the basis
of heuristic clues, such as endorsers who are congruent with the company/product and so
valid in evaluating the firm (Ahluwalia & Gürhan-Canli 2000). This lower elaboration
of information favours individuals inferring altruistic motives on the part of the endorser
(Rifon et al. 2004). All these arguments lead to the following hypothesis:
H5: In the non-profit sector, a good image fit between the celebrity and the social
cause is directly and positively related to the attribution of altruistic motivation
of the celebrity.
Indirect antecedents of celebrity credibility: the attitude towards the celebrity and the
general f it
Taking into account the importance of the fit, it is interesting to analyse its determinants,
which become indirect antecedents of celebrity credibility. In the non-profit sector, a good
strategy may be to achieve the commitment of the celebrity to the cause off-camera – that
is, genuine support (Magnini et al. 2008) – as people usually respect celebrities they feel
are genuinely committed to the causes they support (Samman et al. 2009). In these cases,
106
and over the long term, people can get used to linking the celebrity with the social cause,
and the fit can develop in a natural way.
In the short term, it is accepted that, in many situations, consumers may not be able
to evaluate the fit of the celebrity objectively, and make inferences using known informa-
tion (Huber & McCann 1982), overall impressions (Boatwright et al. 2008) or even the
attractiveness of the source (Kahle & Homer 1985; Till & Busler 2000). Specifically, the
variable ‘attractiveness’ is usually measured with items such as ‘sexy’, ‘elegant’, ‘classy’ or
‘beautiful’ (Ohanian 1990), so its use may not be appropriate in the case of the non-profit
sector (Wheeler 2009). However, and going more deeply into these issues, attractiveness
also includes non-physical variables, such as similarity, familiarity and likeability (Kahle
& Homer 1985), charisma or grace (Langmeyer & Shank 1994), and it can be viewed
as positive attitude or affect towards the endorser (Silvera & Austad 2004). Nowadays,
celebrities are well known to consumers due to their frequent appearances in the media,
so consumers already have pre-established knowledge and attitudes towards them before
they appear in advertising (Choi & Rifon 2007). This perception or image of the celebrity
can influence the response to endorsements (Alperstein 1991), an argument supported by
Samman et al. (2009) in the non-profit area.
There are no specific studies that measure the direct influence of the celebrity image
on perceived fit in the non-profit sector, however there is enough evidence to consider
that attitudes and feelings towards celebrities, resulting from their behaviour both profes-
sionally and in their private lives, may influence the assessment of the endorsement and
condition other issues – for example, his or her image fit. This is particularly important in
the first stages of the strategy – that is, when people are not yet used to linking the social
cause with the celebrity, and when the functional fit is not evident. All the above leads us
to propose Hypothesis 6:
H6: In the non-profit sector, the attitude towards the celebrity is positively related
to the perceived fit between the celebrity and the social cause.
The last variable analysed is the general attitude towards this type of endorsement
in the non-profit sector. With regard to this, there are hardly any empirical studies that
measure public opinion on the phenomenon of celebrity activism, with the exception of
Samman et al.’s (2009) and Van den Bulck et al.’s (2010) studies. In spite of some suspi-
cions, Samman et al. (2009) conclude that, in most cases, people feel this collaboration is
good and beneficial for the causes celebrities support. In the same vein, Van den Bulck
et al. (2010) observes that a majority of respondents believe NPOs can benefit from
using celebrities, in the sense that organisations can get their messages across more suc-
cessfully. These results lead us to think that, if people are in favour of using celebrities in
the communicationof social causes, due to the associated benefits or other reasons, they
will feel in general more predisposed to consider any celebrity appropriate to represent a
non‑profit organisation. The general attitude towards this strategy or general perceived
fit has not been considered in previous studies, however in this field it is reasonable to
propose that:
107
H7: In the non-profit sector, the positive attitude towards the involvement of celeb-
rities in social causes is directly related to the perceived fit between the celebrity
and the social cause.
Determinants of behavioural intentions

A determinant of purchase intention is the attitude towards the ad (MacKenzie & Lutz
1989; Goldsmith et al. 2000; Lafferty et al. 2002). According to Shimp (1981), it is
important to produce positive attitudes towards advertisements in the minds of individu-
als in order to influence customer choice behaviour. This relationship is seen to a greater
extent when affective responses are evoked (Petty et al. 1983) – for example, in the com-
munication of social causes. In this sense, Ranganathan and Henley (2007) prove that a
favourable attitude towards an advertisement has a positive influence over the intention
to donate to charity. With regard to the use of celebrities, these ‘stars’ transmit emotional
feelings, and theoretically can encourage individuals to get personally involved in a cause,
e.g. by donating money (West & Orman 2003). Although people claim not to feel per-
sonally influenced by these (Samman et al. 2009; Van den Bulck et al. 2010), Van den
Bulck et al. (2010) observed that celebrity-based campaigns seemed more successful in
creating support. Therefore, it can be said that:
H8: In the non-profit sector, the attitude towards the advertisement is directly and
positively related to behavioural intentions.
Finally, it is extensively accepted that the attitude towards the brand has a direct and
positive influence on behavioural intentions (Goldsmith et al. 2000; Lafferty et al. 2002),
and advertiser credibility acts as a central processing clue in all the process (MacKenzie &
Lutz 1989). This relationship has also been proved in the non-profit sector (Ranganathan
& Henley 2007), so it can be said that:
H9: In the non-profit sector, the credibility of the organisation is directly and posi-
tively related to behavioural intentions.
The global causal model to be tested is represented in the Figure 1.
Methodology
The hypotheses are contrasted with a causal model, analysing the relationships between
the variables through a Structural Equations Model (SEM) with the software EQS 6.1.
Compared to other papers reporting on experiments with student samples and organisa-
tions, often fictitious (Lafferty & Goldsmith 1999; Lafferty et al. 2002; Silvera & Austad
2004; Wheeler 2009), we designed a causal model, a methodology that tests and analyses a
system of relationships, and the simultaneous and causal sequence established (Goldsmith
et al. 2000).
108
Figure 1: Causal model
Attitude H6
celebrity
Fit celebrity H3
– NPO
H1
General fit H7 H5 Celebrity
credibility
Motivation H4
attribution
H2
Entity credibility Attitude
towards ad
H8
Behavioural intentions
H9
Specifically, we measured adult individual perceptions of a realistic but fictitious and,

therefore, unfamiliar ad, with a real celebrity and a real NPO. In our research, the fit
is a variable of the model, both independent (because it explains other constructs) and
dependent (because it is explained by other variables). We thus followed the Currás et al.
(2009) methodology, presenting to those surveyed two ads with different congruence, in
order to average the endorser fit value in the set of final data and ensure the variability
needed to check the hypothesis in the causal model. The research was focused on the
effectiveness of the endorsement strategy in the short term – that is, before people had
learned to associate the celebrity with the social cause. We thus ensured that people had
no knowledge of the endorsement beforehand.
UNICEF (the United Nations International Children’s Emergency Fund) was chosen
as the NPO as it is a very popular global brand. The quantitative study was designed in
the form of a personal questionnaire for adults; consequently it was necessary to analyse
a familiar organisation to ascertain which individuals had information and held specific
beliefs. UNICEF is a well-known organisation, a characteristic that a priori rules out the
adverse effects of delegitimisation as the result of a bad social reputation (Dean 2003).
Subsequently, a pretest was carried out to identify the more appropriate NPO–
celebrity combinations. Forty people aged between 20 and 40 participated in the study. It
was important that the combination was realistic but unfamiliar, so we asked about both
the congruence perceived between the celebrity and the social cause, and about the knowl-
edge of the endorsement. As a result, we finally decided to carry out the research into the
combination UNICEF–Pau Gasol as the endorsement with the best fit and UNICEF–
Penélope Cruz as the one with the poorest fit (the t-test for related samples confirmed
the significant difference between the means, t = 4.19, p < 0.01). The ads showed a colour
picture of the celebrity in close-up, along with the UNICEF logo and the headline ‘We
109
need your aid’. There was a call to action, with the text ‘Join us, collaborate’, and the web
address was given at the bottom of the ad. Each advert was identical, with the exception
of the celebrity, and both were produced by a graphic designer.
The study population comprised Spanish people over 18 years old. The sample was
chosen by convenience, although age and gender were controlled. The sample size was
329, once those indicating knowledge of the collaboration between the celebrity and the
organisation had been ruled out (specifically, 162 people evaluated the UNICEF–Pau
Gasol combination, and 167 people the UNICEF–Penélope Cruz endorsement). The
questionnaire included 20 questions, related to previous collaboration with NPOs, aware-
ness of the entity and the celebrity, evaluation of the ad, the celebrity and the entity,
involvement with the social cause, perceived fit (specific and general), motivation attribu-
tion, behavioural intentions, personality traits of the respondent and demographic issues.
Scales
All items used in the study were measured on a 7-point Likert scale considering previ-
ous papers (see Appendix 1, which presents detail of the scales and the literature taken
into account). Before carrying out the SEM and testing the research hypotheses, we
checked the reliability and validity of all the concepts by means of a confirmatory analysis
(Anderson & Gerbing 1998). Previously, we had calculated second-order factors with
the constructs ‘celebrity credibility’ and ‘NPO credibility’. We proposed that the two
types of credibility measure expertise and trustworthiness, which was confirmed, since all
dimensions converged appropriately on those factors. The results indicate good reliability
and the convergent validity of all the scales studied globally, and goodness of fit indexes
exceed the 0.9 threshold, which confirms the suitability of the model presented (Table 1).
Discriminant validity was evaluated using the confidence intervals for pairs of latent con-
cepts. None of the confidence intervals calculated for the pair-wise comparisons of the
model variables included unity, so the proposed measurement model is correct.
Results
The proposed hypotheses were tested with a structural equation model. Goodness of
fit indexes confirm the validity of the global model (BBNFI = 0.93; BBNNFI = 0.96;
CFI = 0.96; IFI = 0.96; RMSEA = 0.05).
The conceptual model provides support for the entire hypothesis (Figure 2). First, it is
observed that the attitude towards the ad is directly conditioned for NPO credibility but,
above all, for celebrity credibility, so Hypotheses 1 and 2 cannot be rejected. The effect
of celebrity on the evaluation of the ad is stronger than the effect of perceived expertise
and trustworthiness of the organisation, a result which confirms that, both in the profit
and non-profit sectors, celebrities are powerful transmitters of feelings, and exert a strong
influence over consumers through peripheral persuasion routes. However, it is important
to take into account that the celebrity must be credible in order to be effective. In this sense,
credibility is directly conditioned by the perceived fit between the source and the NPO
110
Table 1: Confirmatory factor analysis of the final model
Latent variable Measured variable Standard lambda R2 Cronbach’s α

Celebrity credibility EXPCEL 0.89 0.79
0.91
TRUCEL 0.93 0.87
Entity credibility EXPCEL 0.90 0.82
0.91
TRUCEL 0.91 0.83
Attitude towards the ad AGOOD 0.91 0.83
APOSITIVE 0.97 0.93 0.96
AFAVOR 0.96 0.92
Attitude towards the celebrity CEGOOG 0.88 0.78
CEPOSITIVE 0.91 0.82
CEFAVOR 0.93 0.86
0.98
CENICE 0.97 0.94
CEPLEA 0.96 0.93
CELIKES 0.95 0.92
Perceived fit COHE 0.91 0.83
COMPATIB 0.94 0.90
0.96
SENSE 0.90 0.80
MATCH 0.93 0.88
General perceived fit GCOHE 0.91 0.83
GCOMPATIB 0.93 0.90
0.96
GSENSE 0.96 0.80
GMATCH 0.94 0.88
Motivation attribution INTEREST 0.79 0.62
ALTRUI 0.94 0.90 0.92
SOCIAL 0.95 0.91
Behavioural intentions COLABO 0.73 0.53
ENCOUR 0.88 0.77
0.92
SPEAK 0.92 0.85
ADVIS 0.95 0.90
Goodness of fit indices: BBNFI = 0.94, BBNNFI = 0.97, CFI = 0.98, IFI = 0.98, RMSEA = 0.04
(Hypothesis 3 accepted), and for the motivation attribution (Hypothesis 4 accepted). In

fact, the fit perceived is a key variable in the model because, on the one hand, the better
the fit, the more the expertise and trustworthiness of the celebrity is perceived. On the
other hand, the perceived fit enhances the attribution of social motivations (Hypothesis
5 supported). When individuals perceive high congruence between the celebrity and the
NPO, they attribute more social and altruistic motives to his/her behaviour as opposed to
self-interested motives, and this is important, as the motivation attribution also influences
celebrity credibility. In this context, the fit is mainly explained by the attitude towards
the celebrity (Hypothesis 6 supported). Celebrities are not used to being experts in social
causes because that is not their usual work, but they can be credible and very persuasive.
111
Figure 2: Results of the causal model
Attitude H6 (0.63, t=12.21)***

celebrity
Fit celebrity H3 (0.72, t=11.66)***
– NPO
H1 (0.62, t=10.04)***
General fit Celebrity
H7 (0.46, t=10.02)*** H5 (0.68, t=10.57)***
credibility
Motivation
H4 (0.29, t=4.68)***
attribution
H2 (0.12, t=2.65)***
Entity credibility Attitude

towards ad
H8 (0.11, t=2.53)***
***p < 0.001 H9 (0.66, t=10.24)***
Specifically, their powers of persuasion derive from their image, from the attitudes and
feelings people have about them. The more people like them, the better the fit perceived
with the social cause and so the more effective will be the advertising communication.
The general attitude towards celebrity activism also has a positive influence over the fit
perceived (Hypothesis 7 supported), although to a lesser extent. This result means that
people who approve the link between celebrities and social causes will tend to value the
endorsement strategy more favourably.
It is interesting to observe that the attitude towards the ad has a significant influence
over behavioural intentions (Hypothesis 8 supported). Nevertheless, this variable is mainly
explained by central processing clues, such as the organisation’s credibility (Hypothesis 9
supported). Therefore, the NPO must pay particular attention to its behaviour in order to
reflect expertise and trustworthiness, and so be credible.
Discussion
This paper tries to contribute to the development of the research line of communication
with celebrities in the non-profit area. Specifically, a causal model was carried out, and
nine hypotheses were proposed regarding the determinants of celebrity credibility and
the effectiveness of communication using these endorsers. As a result, we can prove that
behavioural intentions are mainly explained by the NPO credibility. Therefore individuals
take into account mainly central persuasion routes when they have to make a decision
regarding buying or, in this case, collaborating. Organisations must therefore be hon-
est and transparent, and work seriously in their field, communicating the work they do
through the media and, for example, public relations campaigns.
In this sense, an increasingly common strategy is to enlist the support of celebrities in
112
advertising campaigns. Taking into account the results of the model, it can be said that
individuals trust central routes when making certain evaluations, but they are also carried
away by emotions, so a valued celebrity may have a positive effect on a social organisa-
tion through his or her impact on attitudes towards the ad. Celebrities persuade through
peripheral routes. They are very popular, and possess special features, with outstanding
performances in their field, and enviable lifestyles that turn them into symbolic and
admired figures. However, not all of them are effective. The key question is celebrity cred-
ibility and, above all, his or her image.
On the one hand, the better the fit, the more credible will be the celebrity. If people
perceive some type of link between the celebrity and the social cause, the expertise and
trustworthiness of the celebrity will not be questioned. As a result, the endorsement will
be effective. This fit is mainly conditioned by the attitudes towards the celebrity. When
people perceive celebrities as supporting a social cause more for selfish motives rather than
social ones, their credibility will fall off and, therefore, they will have a negative impact on
individuals’ responses. This motivation attribution is directly determined by the fit, and
indirectly by attitudes towards the celebrity. Therefore, image is the key to ‘opening other
doors’, at least in the short term and in the case of celebrities without previous experience
in this field.
Celebrity activism has been criticised, but a lot of people are conscious of the bene-
fits of this link for NPOs. Due to the benefits perceived, these individuals can be more
favourably predisposed to these endorsements, and this strategy will be less questioned
and thus more effective. In order to arouse these feelings the link has to be taken seriously
for both parties, as people usually respect celebrities they feel are genuinely committed to
the causes they support (Samman et al. 2009).

This paper is not exempt from limitations, which therefore open the field to some future
lines of research. First, the findings are limited to the context analysed. We used well-
known celebrities and social organisations that usually carry out campaigns with popular
individuals who become ‘Goodwill Ambassadors’. Featuring other NPOs and different
combinations would extend the scope of the findings.
Second, we have analysed the data globally. It would be interesting to evaluate the
model with multisample analysis, taking into account greater or lesser involvement with
social causes, different features of personality for those involved, or demographic vari-
ables. For example, it may be expected that individuals very involved in social causes trust
peripheral routes in a minor extension and, therefore, the celebrities have less impact in
the behaviour process. This argument needs further investigation.
Third, celebrities can also benefit from the endorsement. In this same vein, it would be
interesting to analyse how supporting a social cause enhances celebrity image. Finally, the
data have been analysed globally to achieve variability in the causal model. Other studies
with the two celebrities separately could be carried out, in order to analyse the specific
contribution of each one to the NPO.
113
Appendix 1 (scales)
Celebrity credibility EXPCEL The celebrity (X) is right because he/she has Ohanian 1990; Wheeler 2009
experience in the social area
(X) is qualified to represent an entity of Ohanian 1990; Wheeler 2009
this type
(X) is knowledgeable about social problems Ohanian 1990; Wheeler 2009
TRUCEL (X) is reliable Ohanian 1990; Wheeler 2009
(X) is sincere Ohanian 1990; Wheeler 2009
(X) is honest in his/her declarations Ohanian 1990; Wheeler 2009
Entity credibility EXPENT The entity (Y) has a great amount of Newell & Goldsmith 2001
experience in the social area
(Y) is skilled in what they do Newell & Goldsmith 2001
(Y) has a great expertise in the social area Newell & Goldsmith 2001
TRUENT (Y) is trustworthy Newell & Goldsmith 2001
(Y) is honest Newell & Goldsmith 2001
(Y) makes truthful claims Newell & Goldsmith 2001
Attitude towards the ad AAD1 Good impression McKensie & Lutz 1989; Lafferty
et al. 2002; Klaus & Bailey 2008
AAD2 Positive impression Davis 1995
AAD3 Favourable impression Davis 1995; McKensie & Lutz
1989; Lafferty et al. 2002
Attitude towards the ACEL1 I have a good image of (X) McKensie & Lutz 1989; Lafferty
celebrity et al. 2002: Klaus & Bailey 2008
ACEL2 I value (X) positively Davis 1995
ACEL3 My attitude towards (X) is favourable Davis 1995; McKensie & Lutz
1989; Lafferty et al. 2002
ACEL4 (X) is likeable Tripp et al. 1994; Wheeler 2009
ACEL5 (X) is nice Tripp et al. 1994; Wheeler 2009
ACEL6 I like (X) Tripp et al. 1994; Wheeler 2009
Perceived fit FIT1 The link between (X) and (Y) goes together Till & Busler 2000
FIT2 The link is meaningful Bigné et al. 2009
FIT3 The celebrity and the social cause are Bigné et al. 2009
compatible
FIT4 The celebrity and the social cause fit Till & Busle 2000
together
General perceived fit GFIT1 The link between celebrities and social Till & Busler 2000
causes goes together
GFIT2 The link is meaningful Bigné et al. 2009
GFIT3 Celebrities and social causes are compatible Bigné et al. 2009
GFIT4 Celebrities and social causes fit together Till & Busler 2000
Motivation attribution MA1 (X) is motivated by interest in society (not Strahilevitz 2003; Becker Olsen
by self-interest) et al. 2006; Bigné et al. 2009
MA2 (X) has altruistic motives (not selfish) Strahilevitz 2003; Becker Olsen
et al. 2006; Bigné et al. 2009
MA3 (X) seeks social objectives (not individual Strahilevitz 2003; Becker Olsen
objectives) et al. 2006; Bigné et al. 2009
114
Behavioural intentions BEHINT1 It is likely I collaborate with (Y) in the future Lafferty & Goldsmith 1999;
Wheeler 2009
BEHINT2 If I wanted to collaborate with any entity, I Zeithaml et al. 2006
would turn to (Y)
BEHINT3 I could say positive things about (Y) Zeithaml et al. 2006
BEHINT4 I would recommend (Y) if somebody asked Zeithaml et al. 2006
my advice
Appendix 2: Questionnaire examples

7-point scales (1 = total disagreement; 7 = total agreement)
Celebrity credibility
Indicate your level of agreement or disagreement with the following statements about Pau Gasol/Penélope Cruz and
his/her link to UNICEF in this advertising campaign …
Pau Gasol/Penélope Cruz is qualified to represent an entity of this type 1 2 3 4 5 6 7
Entity credibility
Indicate your level of agreement or disagreement with the following statements about UNICEF:
UNICEF has a great amount of experience in the social area 1 2 3 4 5 6 7
Attitude towards the ad

What has been your general impression on seeing this ad, with the Pau Gasol/Penélope Cruz–UNICEF link? … Rate
your opinion:
… good impression 1 2 3 4 5 6 7
Attitude towards the celebrity

Indicate your level of agreement or disagreement with the following statements about Pau Gasol/Penélope Cruz:
I have a good image of Pau Gasol/Penélope Cruz 1 2 3 4 5 6 7
Perceived fit
Indicate your level of agreement or disagreement with the following statements about the link between Pau Gasol/
Penélope Cruz and UNICEF:
The link between Pau Gasol/Penélope Cruz and UNICEF goes together 1 2 3 4 5 6 7
General perceived fit

Indicate your level of agreement or disagreement with the following statements about the link between celebrities
and social causes (in general terms):
The link between celebrities and social causes goes together 1 2 3 4 5 6 7
Motivation attribution
Indicate your level of agreement or disagreement with the following statements about the motivations of Pau Gasol/
Penélope Cruz to join an organisation like UNICEF:
Pau Gasol/Penélope Cruz is motivated by interest in society (not by self-interest) 1 2 3 4 5 6 7
Indicate your level of agreement or disagreement with the following statements:
It is likely I will collaborate with UNICEF in the future 1 2 3 4 5 6 7
115
References
Ahluwalia, R. & Gürhan-Canli, Z. (2000) The effect of extensions on the core brand
evaluation: an accessibility–diagnosticity perspective. Journal of Consumer Research, 27(2),
371–381.
Alperstein, N.M. (1991) Imaginary social relationships with celebrities appearing in television
commercials. Journal of Broadcasting and Electronic Media, 35(1), 43–58.
Amos, C., Holmes, G. & Strutton, D. (2008) Exploring the relationship between celebrity
endorser effects and advertising effectiveness. International Journal of Advertising, 27(2),
209–234.
Anderson, J.C. & Gerbing, D.W. (1998) Structural equation modelling in practice: a review
and recommended two-step approach. Psychological Bulletin, 103(3), 411–423.
Assael, H. (1984) Consumer Behavior and Market Action. Boston, MA: Kent Publishing
Company.
Becker-Olsen, K.L. & Hill, R.P. (2006) The impact of sponsor fit on brand equity. The case of
nonprofit service providers. Journal of Service Research, 9(1), 73–83.
Becker-Olsen, K.L., Cudmore, A.B. & Hill, R.P. (2006) The impact of perceived corporate
social responsibility on consumer behaviour. Journal of Business Research, 59, 46–53.
Bigné, E., Currás, R. & Sánchez, I. (2009) Brand credibility in cause-related marketing: the
moderating role of consumer values. Journal of Product & Brand Management, 18(6), 437–
447.
Boatwright, P., Kalra, A. & Zhang, W. (2008) Should consumers use the halo to form product
evaluations? Management Science, 54(1), 217–223.
Byrne, A., Whitehead, M. & Breen, S. (2003) The naked truth of celebrity endorsement.
British Food Journal, 105(4/5), 288–296.
Chang, C. & Lee, Y. (2010) Effects of message framing, vividness congruency and statistical
framing on responses to charity advertising. International Journal of Advertising, 29(2),
195–220.
Choi, S.M. & Rifon, N.J. (2007) Who is the celebrity in advertising? Understanding
dimensions of celebrity images. Journal of Popular Culture, 40(2), 304–325.
Currás, R., Bigné, E. & Alvarado, A. (2009) The role of self definitional principles in consumer
identification with a socially responsible company. Journal of Business Ethics, 89(4), 547–564.
Dean, D.H. (2003) Consumer perceptions of corporate donations: effects of company
reputation for social responsibility and type of donation. Journal of Advertising, 32(4),
91–102.
Dominguez, R. (2010) Celebridades y Cooperación al Desarrollo: Manejar con Cuidado.
Documento de trabajo del Real Instituto Elcano, ARI 142, 1–8.
Drumwright, M.E. (1996) Company advertising with a social dimension: the role of
noneconomic criteria. Journal of Marketing, 60(4), 71–87.
Eisend, M. & Langner, T. (2010) Immediate and delayed advertising effects of celebrity
endorsers’ attractiveness and expertise. International Journal of Advertising, 29(4), 527–546.
Ellen, P.S., Webb, D.J. & Mohr, L.A. (2006) Building corporate associations: consumer
attributions for corporate socially responsible programs. Journal of the Academy of Marketing
Science, 34, 147–157.
Folkes, V.S. (1988) Recent attribution research in consumer behaviour: a review and new
directions. Journal of Consumer Research, 14(March), 548–565.
Forehand, M.R. & Grier, S. (2003) When is honesty the best policy? The effect of stated intent
on consumer skepticism. Journal of Consumer Psychology, 13(3), 349–356.
Goldberg, M.E. & Hartwick, J. (1990) The effects of advertiser reputation and extremity of
advertising claim on advertising effectiveness. Journal of Consumer Research, 17(2), 172–179.
116
Goldsmith, R.E., Lafferty, B.A. & Newell, S.J. (2000) The impact of corporate credibility and
celebrity credibility on consumer reaction to ads and brands. Journal of Advertising, 29(3),
43–55.
Grewal, D., Gotlieb, J. & Marmorstein, H. (1994) The moderating effects of message framing
and source credibility on the price-perceived risk relationship. Journal of Consumer Research,
21(1), 145–153.
Gwinner, K.P. (1997) A model of image creation and image transfer in event sponsorship.
International Marketing Review, 14(3), 145–158.
Gwinner, K.P. & Eaton, J. (1999) Building brand image through event sponsorships: the role of
image transfer. Journal of Advertising, 28(4), 47–57.
Hastie, R. (1984) Causes and effects of causal attribution. Journal of Personality and Social
Psychology, 46( January), 44–56.
Hill, R. & Moran, N. (2011) Social marketing meets interactive media: lessons for the
advertising community. International Journal of Advertising, 30(5), 815–838.
Hovland, C., Janis, I. & Kelly, H. (1953) Communication and Persuasion. New Haven, CT: Yale
University Press.
Huber, J. & McCann, J. (1982) The impact of inferential beliefs on product evaluations. Journal
of Marketing Research, 19(August), 324–333.
Kahle, L.R. & Homer, P.M. (1985) Physical attractiveness of the celebrity endorser: a social
adaptation perspective. Journal of Consumer Research, 11(March), 954–961.
Kamins, M.A. & Gupta, K. (1994) Congruence between spokesperson and product type: a
matchup hypothesis perspective. Psychology & Marketing, 11(6), 569–587.
Kelley, H.H. (1973) The processes of casual attributions. American Psychologist, February, 107–128.
Klaus, N. & Bailey, A. (2008) Celebrity endorsements: an examination of gender and
consumers’ attitudes. American Journal of Business, 23(2), 53–61.
Kotler, P. (1979) Strategies for introducing marketing into nonprofit organizations. Journal of
Marketing, 43(1), 37–44.
LaBarbera, B.A. (1982) Overcoming a no reputation liability through documentation and
advertising regulation. Journal of Marketing Research, 19(2), 223–231.
Lafferty, B.A. (2007) The relevance of fit in a cause–brand alliance when consumers evaluate
corporate credibility. Journal of Business Research, 60(5), 447–453.
Lafferty, B.A. & Goldsmith, R.E. (1999) Corporate credibility’s role in consumers’ attitudes
and purchase intentions when a high versus a low credibility endorser is used in the ad.
Journal of Business Research, 44(2), 109–116.
Lafferty, B.A., Goldsmith, R.E. & Hult, T.M. (2004) The impact of the alliance on the
partners: a look at cause–brand alliances. Psychology & Marketing, 21(7), 509–531.
Lafferty, B.A., Goldsmith, R.E. & Newell, S.J. (2002) The dual credibility model: the influence
of corporate and endorser credibility on attitudes and purchase intentions. Journal of
Marketing Theory and Practice, 10(3), 1–12.
Langmeyer, L. & Shank, M. (1994) Managing beauty – products and people. Journal of Product
& Brand Management, 3(3), 27–38.
Lynch, J. & Schuler, D. (1994) The matchup effect of spokesperson and product congruency: a
schema theory interpretation. Psychology & Marketing, 11(5), 417–436.
MacKenzie, S.B. & Lutz, R.J. (1989) An empirical examination of the structural antecedents of
attitude toward the ad in an advertising pretesting context. Journal of Marketing, 53(April),
48–65.
Magnini, V.P., Honeycutt, E.D. & Cross, A.M. (2008) Understanding the use of celebrity
endorsers for hospitality firms. Journal of Vacation Marketing, 14(1), January, 57–70.
McCracken, G. (1989) Who is the celebrity endorser? Cultural foundations of the
endorsement process. Journal of Consumer Research, 16(3), 310–321.
117
McGinnies, E. & Ward, C.D. (1980) Better liked than right: trustworthiness and expertise as
factors in credibility. Personality and Social Psychology Bulletin, 6(3), 467–472.
Menon, S. & Kahn, B. (2003) Corporate sponsorships of philanthropic activities: when do they
impact perception of sponsor brand? Journal of Consumer Psychology, 13(3), 316–327.
Moore, D.J., Mowen, J.C. & Reardon, R. (1994) Multiple sources in advertising appeals: when
product endorsers are paid by the advertising sponsor. Journal of the Academy of Marketing
Science, 22(3), 234–243.
Mowen, J.C. & Brown, S.W. (1981) On explaining and predicting the effectiveness of celebrity
endorsers. In: Monroe, R.B. (ed.) Advances in Consumer Research, Association for
Consumer Research, Chicago, IL, 8, 437–441.
Newell, S.J. & Goldsmith, R.E. (2001) The development of a scale to measure perceived
corporate credibility. Journal of Business Research, 52(3), 235–247.
Ohanian, R. (1990) Construction and validation of a scale to measure celebrity endorsers’
perceived expertise, trustworthiness, and attractiveness. Journal of Marketing, 19(3), 39–52.
Petty, R.E., Cacioppo, J.T. & Schumann, D. (1983) Central and peripheral routes to advertising
effectiveness: the moderating role of involvement. Journal of Consumer Research, 10(2),
135–145.
Ranganathan, S. & Henley, W. (2007) Determinants of charitable donation intentions: a
structural equation model. International Journal of Nonprofit and Voluntary Sector Marketing,
13, 1–11.
Rifon, N., Choi, S.M., Trimble, C.S. & Li, H. (2004) Congruence effects in sponsorship: the
mediating role of sponsorship and consumer attributions of sponsor motive. Journal of
Rumelhart, D.E. (1980) Schemata: the building blocks of cognition. In: Spiro, R.J. et al. (eds)
Theoretical Issues in Reading Comprehension. Hillsdale, NJ: Lawrence Erlbaum.
Samman, E., McAuliffe, E. & MacLachlan, M. (2009) The role of celebrity in endorsing
poverty reduction through international aid. International Journal of Nonprofit and Voluntary
Sector Marketing, 14, 137–148.
Shimp, T. (1981) Attitude toward the ad as a mediator of brand choice. Journal of Advertising,
10(2), 9–15.
Silvera, D.H. & Austad, B. (2004) Factors predicting the effectiveness of celebrity endorsement
ads. European Journal of Marketing, 38(11/12), 1509–1526.
Sparkman, R.M. (1982) The discounting principle in the perception of advertising. Advances in
Consumer Research, 9, 277–280.
Speed, R. & Thompson, P. (2000) Determinants of sports sponsorship response. Journal of the
Academy of Marketing Science, 28(2), 226–238.
Stafford, M.R., Stafford, T.F. & Day, E. (2002) A contingency approach: the effects of
spokesperson type and service type on service advertising perceptions. Journal of Advertising,
31(2), 17–35.
Strahilevitz, M. (2003) The effects of prior impressions of a firm’s ethics on the success of
a cause-related marketing campaign: do the good look better while the bad look worse?
Journal of Nonprofit and Public Sector Marketing, 11(1), 77–92.
Till, B.D. & Busler, M. (2000) The match-up hypothesis: physical attractiveness, expertise, and
the role of fit on brand attitude, purchase intent and brand beliefs. Journal of Advertising,
29(3), 1–13.
Trimble, C.S. & Rifon, N.J. (2006) Effects of consumer perceptions of compatibility in
cause-related marketing messages. International Journal of Nonprofit and Voluntary Sector
Marketing, 11(1), 29–47.
Tripp, C., Jensen, T.D. & Carlson, L. (1994) The effects of multiple product endorsements by
celebrities on consumers’ attitudes and intentions. Journal of Consumer Research, 20(4), 535–547.
118
Van den Bulck, H., Panis, K., Van Aelst, P. & Hardy, A. (2010) Celebrity activists in social
profit campaigning: a survey with the Flemish public on views and effectiveness. Annual
Conference of the International Communication Association, Singapore, June.
Webb, J.D. & Mohr, L.A. (1998) A typology of customers responses to cause related
marketing: from skeptics to socially concerned. Journal of Public Policy and Marketing, 17(2),
226–239.
West, D. & Orman, J. (2003) Celebrity Politics. New Jersey: Prentice Hall.
Wheeler, R.T. (2009) Nonprofit advertising: impact of celebrity connection, involvement and
gender on source credibility and intention to volunteer time or donate money. Journal of
Nonprofit & Public Sector Marketing, 21(1), 80–107.
Zeithaml, V.A., Berry, L.L. & Parasuraman, A. (1996) The Behavioral Consequences of
Satisfaction decisions. Journal of Marketing Research, 17(4), 460–469.
About the authors

Maria del Mar Garcia de los Salmones is an Associate Professor at the Department of
Business Administration, University of Cantabria. Current research interests include
social marketing, branding and influence of corporate social responsibility on consumer
behaviour. Her papers have been published in, among others, Journal of Business Ethics,
European Journal of Marketing, International Journal of Bank Marketing and Journal of Risk
Research.
Rafael Dominguez is an Associate Professor at the Department of Economics,
University of Cantabria. He is Director of the International Cooperation Chair. His
research focuses on corporate social responsibility and international cooperation issues.
Angel Herrero is an Associate Professor at the Department of Business Administration,
University of Cantabria. His research interests are in regional branding and information
technology in marketing. He has published in The Service Industries Journal, Behaviour and
Information Technology, Electronic Commerce Research and Applications, Journal of Business
Ethics and elsewhere.
Address correspondence to: Maria del Mar Garcia de los Salmones, Associate
Professor, Department of Business Administration, University of Cantabria, Avenida de
los Castros, s/n, 39005 Santander, Spain.
Email: gsalmonm@unican.es
119
www.ijmr.com
The essential Advance your
source of new understanding of marketing
and social research
thinking on
Gain access to cutting edge
marketing and papers
social research Understand new tools and
how they are applied
Keep abreast of new
Premium developments, innovations
subscriptions and technical advances
give you online
access to issues Learn from the world’s
leading research thinkers
from 2000
onwards Online and print
subscriptions available
How banner ads can be effective
Investigating the influences of exposure duration
and banner ad complexity
Kai-Yu Wang
Brock University
Eric Shih
Sungkyunkwan University
Laura A. Peracchio
University of Wisconsin-Milwaukee
This research examines the impacts of exposure duration and banner ad complexity on adver-
tising persuasion in a web advertising environment. Processing fluency is used to explain the
underlying process that occurs among consumers during exposure to advertisements, and
refers to the ease of stimulus encoding and processing that is facilitated by prior exposure to
a banner ad. Based on previous studies (e.g. Reber et al. 1998), this research used a priming
phase and a testing phase, in which respondents viewed two banner ads for the same brand.
A banner ad presented in the priming phase facilitates viewer processing of a target banner ad
in the testing phase due to processing fluency. The findings show that, when a banner ad is
difficult to process in the priming phase, increasing the duration of exposure to the ad in the
priming phase causes a linear increase in respondent attitudes towards the target ad and brand
in the testing phase. When the priming banner ad is moderately difficult to process, increas-
ing the exposure duration in the priming phase first increases, and then decreases, respondent
attitudes towards the target ad and brand (an inverted-U pattern) in the testing phase. When
the priming banner ad is easy to process, increasing the exposure duration in the priming
phase first decreases, and then increases, respondent attitudes towards the target ad and brand
(a U pattern) in the testing phase.
Introduction
According to the Internet Advertising Bureau, online advertising revenue reached nearly
$11 billion in the first half of 2009 (IAB 2009a). Despite the popularity of other forms
of internet communication (Cheng et al. 2009; Lee & Youn 2009), internet banner ads –
display advertisements that are usually hyperlinked to the websites of advertised products
– continue to be one of the leading advertising formats on the web, accounting for 34% of
online ads (IAB 2009a). Ideally, banner ads should encourage consumers to click on an ad,

DOI: 10.2501/IJA-32-1-121-141
linking the consumer to the brand’s target communication and to additional information
about the brand. With an estimated 1% click-through rate, however, banner ads appear
to have limited success. In fact, empirical studies suggest that the use of click-through, or
direct-response measures, may undervalue banner ads as an advertising medium (Briggs
& Hollis 1997; Dreze & Hussherr 2003; Yoon & Lee 2007). An increasing number of
online advertisers are designing numerous versions of banner ads, and changing elements
such as colour, font, image and message in an attempt to come up with an optimal formula
to draw clicks (Clifford 2008).
Although banner ads are primarily used as a means to increase web traffic, viewing
banner ads has been linked to enhanced brand awareness, brand preference and purchase
intentions among consumers (Briggs & Hollis 1997). The IAB (IAB & MBI 1997) and
comScore Networks (2008), an internet marketing research company, have also found
that involuntary exposure to banner ads with no click-through generates an increase in
brand and advertising awareness in much the same way as ad exposure does in conven-
tional media. Therefore, identifying the critical elements that influence attention and
attitude towards banner ads is an important issue for scholars and practitioners. Previous
studies have examined the effects of web-page background complexity (Stevenson et al.
2000) and congruity between the product foci of the advertiser and website content, as
well as banner colour and colour–text contrast, on viewer attention and attitudes towards
websites, banner ads and advertised brands (Moore et al. 2005). However, little research
has considered the effect of the complexity of banner ads on viewer attitudes. Due to the
nature of internet browsing behaviour, users spend less than an average of 40 seconds
viewing a web page (Filloux 2009). Far less attention and time is paid to banner ads. In
the context of incidental ad exposure, banner advertising complexity is an important elem-
ent to investigate when evaluating the effectiveness of banner advertising. It would seem
to be of great interest and benefit to online advertisers to explore how to enhance viewer
attitudes towards ads and brands during brief periods of exposure to advertised products.
Cho et al. (2001) have demonstrated that, when banner ads are presented over multiple
exposures, the ads generate desirable effects for the brand (i.e. favourable attitudes towards
the banner ad and brand, and higher purchase intention among consumers). In explor-
ing the effects of ad exposure on viewer attitudes, most research investigates the effect of
repeated exposure to an ad, rather than the duration of exposure to an ad. An approach
that investigates the effects of exposure duration on viewer attitudes is important from
both a theoretical and a practical perspective, because the amount of time for which view-
ers are exposed to ads may influence their processing fluency and enhance their brand
judgement. Given that internet users may not encounter the same banner ad more than
once while surfing online, it is important from a practical standpoint to learn whether –
and how – the duration of exposure to banner ads influences the effects of advertising for
consumers. Do the same rules of repeated exposure apply to exposure duration? That is,
can advertisers expect analogous results from consumers seeing an ad repeatedly and the
amount of time that consumers spend viewing a banner ad? Even though determining the
effectiveness of internet advertising has been a primary theme for researchers investigat-
ing internet advertising (Kim & McMillan 2008), few studies, if any, have investigated
122
whether banner ads viewed a single time with a longer duration of exposure generate
favourable effects for the brand. Also, is it possible that banner ads viewed a single time
with shorter exposure duration might generate more favourable effects for the brand in
comparison to the effects from longer durations of exposure?
Processing fluency has recently received widespread attention in the area of cognitive
psychology. Some studies in marketing and consumer research have explored the effects of
processing fluency on consideration-set membership and brand evaluation (Shapiro 1999;
Lee & Labroo 2004; Labroo & Lee 2006). Reber et al. (1998) showed that processing flu-
ency increases brand preference, and that the experience of fluency is affectively positive.
Their research suggests that people prefer objects that are presented for longer durations
of time over objects that are presented for shorter durations, presumably because longer
exposure durations increase processing fluency. Based on existing processing fluency lit-
erature, however, this research proposes a potential boundary for this effect and argues
that the effect of processing fluency on affective judgement is not likely to be monotonic.
This research proposes that the effects of processing fluency on ad and brand judgements
depend on both exposure duration and banner ad complexity. Thus the objective of this
research is to examine how exposure duration and banner ad complexity influence affec-
tive judgements via processing fluency in online marketing environments.
Theoretical background and hypothesis development
Processing fluency and exposure duration

Research has suggested that prior exposure to an ad has a positive effect on subsequent ad
evaluation (Yoo et al. 2009). We propose that this effect is related to processing fluency.
The processing fluency model was proposed by researchers in order to explain mere expo-
sure effects (Bornstein & D’Agostino 1994). The model argues that repeated exposure to
a stimulus can lead to processing fluency, and that individuals attribute this processing
fluency to liking a stimulus based on contextual cues. Processing fluency is defined as
the ease of identifying and recognising a stimulus (Lee & Labroo 2004). It consists of
perceptual and conceptual fluency. Perceptual fluency is defined as the ease of identifying
the physical features of a stimulus ( Jacoby & Dallas 1981; Shapiro 1999; Lee & Labroo
2004; Reber et al. 2004), whereas conceptual fluency refers to the ease of understanding
the meaning of a stimulus (Lee & Labroo 2004). Both perceptual and conceptual fluen-
cies are known to be improved by prior exposure to a stimulus, and perceptual fluency
is sensitive to changes in surface features ( Jacoby & Dallas 1981). Empirical studies in
implicit memory research demonstrate that prior exposure to a stimulus often benefits
subsequent task performance, such as word completion and the solution of anagram
tasks that involve identification of the perceptual features of a stimulus (Schacter 1987;
Roediger & McDermott 1992). Moreover, the experience of perceptual fluency has been
found to be positively valenced (Reber et al. 1998; Winkielman & Cacioppo 2001). Thus
it seems possible that mere exposure to banner ads, a type of perceptual priming, might
enhance consumer preference for brands featured in banner ads. Research suggests that
123
perceptual priming appears to increase the ease of processing for primed ads and brands,
even when the primes are presented at subliminal levels (e.g. a duration of exposure less
than 50 milliseconds) (Zajonc 1980; Bornstein 1989). On the other hand, conceptual flu-
ency benefits from elaboration, and is therefore not affected by surface changes ( Jacoby
& Dallas 1981). When primes are presented at subliminal levels, they are less likely to
activate semantic processing. However, when the exposure duration of the primes lasts
long enough to register at conscious levels ( Jacoby & Whitehouse 1989), conceptual flu-
ency is likely to occur, and to influence consumer evaluations of ads and brands. In sum,
the expectation is that even very brief exposure to banner ads during online surfing will
promote processing fluency and thereby impact ad and brand evaluations in positive ways
among consumers. This research attempts to test this hypothesis by exploring the ways
in which the two types of processing fluency, perceptual and conceptual, jointly influence
consumer evaluation of ads and brands in the context of online banner advertising.
According to the processing fluency model ( Jacoby et al. 1989; Bornstein & D’Agostino
1994), previous exposure to an ad leads a consumer to experience the subjective feeling of
processing fluency. The experience of processing fluency subsequently impacts preference
by the consumer. More specifically, prior exposure to a stimulus, encoded in the memory,
will facilitate the encoding and processing of the stimulus when it is subsequently encoun-
tered and will enhance processing fluency. In an existing study by Reber et al. (1998), per-
ceptual fluency was manipulated by varying the duration of initial presentation (100, 200,
300 and 400 milliseconds) in a priming phase for viewers before subsequent exposure to
target objects. The results indicated that people preferred target objects that were primed
for longer initial durations over objects that were primed for shorter durations. The study
argued that the findings were the result of a longer duration of exposure to an initial prim-
ing phase enhancing the perceptual fluency when processing subsequent target objects.
While Reber et al. (1998) found that longer periods of exposure to priming ads were
more effective for establishing consumer preference, other studies have suggested that
shorter exposure durations are superior to longer ones. For example, a study by Bornstein
(1989) indicated that longer exposures to a stimulus may impair perceptual fluency. A
study by Bornstein and D’Agostino (1992) demonstrated that stimuli that were briefly
processed (5 milliseconds) were more favourably evaluated by consumers than stimuli that
were processed for a longer duration of time (500 milliseconds). A study by Labroo et al.
(2008) also showed that target products primed for a very brief duration of time (16 milli-
seconds) were selected more often, and were more favourably evaluated, than target prod-
ucts primed for longer periods of time (3 seconds). Existing literature on the processing
fluency model indicates that prior exposure to a stimulus can lead to processing fluency,
and that individuals make attributions for the positive experience of processing fluency
based on context and consumer memory. Accordingly, when individuals find another
salient potential attribution for the source of fluency, their preference for the stimulus
decreases. The study by Jacoby and Whitehouse (1989) showed that when individuals
are aware of the source of fluency (e.g. by previously seeing the stimulus), the awareness
reduces the likelihood of an individual to use a fluency heuristic as a basis for judgement.
Similarly, the current study expects that, if the duration of an initial prime is too long,
124
respondents may become aware of the priming process, and possibly experience reactivity
and a loss of interest in the ad. Thus under certain conditions, a downturn in affective
response by viewers towards target brands and ads is expected following longer exposure.
As discussed, previous studies present mixed findings on the effects of exposure dura-
tion on affective judgements. Instead of advocating either side of the previous findings,
this research attempts to uncover possible factors that might be contributing to the incon-
sistent findings. What determines whether the exposure duration of a banner ad has a
positive or negative effect on brand judgements? In addition to exposure duration, banner
ad complexity is likely to influence processing fluency. The following section addresses
how banner ad complexity may impact processing fluency, as well as brand and advertis-
ing judgements by consumers. Further, the paper explores how the complexity of initial
banner ads may interact with exposure duration to influence target brand judgements and
ad processing by consumers.
Processing fluency and banner ad complexity

A study by Berlyne (1970) argued that the amount of time that a consumer has available
for product message processing is not the only factor influencing their evaluations of the
brand and ad. The time required to process a message, which is determined by the ease of
processing the message, also has an effect on consumer evaluations. Similarly, ease of pro-
cessing may influence processing fluency and judgements of ads and brands, with ads that
are easier to process often being preferred (Reber et al. 2004). The study by Reber et al.
(1998) demonstrated that any prior exposure to an ad facilitates processing of the target
ad and enhances preference for the target ad, even after a single exposure. Multiple vari-
ables seem to influence ease of processing, including figure attractiveness, figure–ground
contrast, stimulus repetition, symmetry, prototypicality, visual or semantic priming, and
stimulus complexity (Reber et al. 2004). The current study focuses on stimulus complexity
(or banner ad complexity in the context of our study) as a variable that influences ease of
banner ad processing.
The complexity of a banner ad may be influenced by the presence of pictures versus
words. Picture–word effects have been extensively examined in cognitive psychology
(Shepard 1967; Nelson et al. 1976; Carr et al. 1982), in bricks-and-mortar (Sojka & Giese
2006) and in online (Kim & Lennon 2008) marketing communications. Numerous studies
have demonstrated that pictures are less complex, more easily recalled or recognised, and
are perceived as more familiar than words (Paivio 1969; Lutz & Lutz 1977; Hirschman
1986). The effect of pictures on memory has been found to be longer lasting than the
effect of words (Childers & Houston 1984; Garnder & Houston 1986). Further, previous
research about the processing of non-attended material indicates that ads consisting pri-
marily of pictures are easier to comprehend than advertising that comprises mainly words.
In addition, picture ads produce more favourable attitudes and improve memory among
consumers (McQuarrie & Mick 2003). Such findings suggest that processing pictorial
banner ads might be easier than processing verbally orientated banner ads. Additionally,
this study expects that banner ads consisting of both pictorial and verbal information
125
should be less complex and easier to process than banner ads with a purely verbal message,
but more complex and difficult to process than banner ads composed mainly of pictures.
Interaction effects of exposure duration and banner ad complexity on ad and brand

evaluations by consumers
Processing fluency consists of perceptual and conceptual fluencies. Perceptual fluency,
which focuses on the analysis of features, is likely to appear at subliminal levels (Bornstein
1989). Conceptual fluency, which relies on semantic analysis, is likely to occur at conscious
levels ( Jacoby & Whitehouse 1989). It seems that both stimulus complexity and time
influence these two types of processing fluency. Therefore, we argue that the interaction
effect of exposure duration and banner ad complexity on ad and brand evaluations is
determined by the relative influences of these two types of processing fluency. Consistent
with the findings of Reber et al. (1998), this study expects that a monotonic increase in
judgement towards a target ad will occur as the duration of exposure lengthens for highly
complex banner ads (words only). Short initial exposure to a highly complex banner ad
impairs both perceptual and conceptual fluencies, as well as evaluation of the target ad.
However, increasing the initial duration of exposure to a highly complex banner ad will
enhance perceptual and conceptual fluencies, and facilitate processing of the subsequent
target ad. Thus a linearly increasing exposure duration-evaluation relationship towards
target brands and ads is expected for highly complex ads.
Regarding low durations of exposure to moderately complex banner ads (involving
pictures and words), the presence of pictures in the moderately complex ad allows for per-
ceptual fluency to occur for viewers. Perceptual fluency relies on a match between phys-
ical features in the priming and test phases, and it is likely to occur at subliminal levels.
In contrast, it is unlikely for conceptual fluency to occur at low durations of exposure. As
exposure duration increases, conceptual fluency kicks in because viewers are able to bet-
ter grasp the meaning of words and pictures at longer durations of exposure. An additive
effect of perceptual and conceptual fluencies enhances evaluation of target brands and
ads. Past a certain point when exposure duration is high, both perceptual and concep-
tual fluencies fade away because longer exposure duration implies attribution of fluency
sources, and boredom with a stimulus ( Jacoby et al. 1989). This phenomenon reduces the
effectiveness of banner ads at longer durations of exposure for moderately complex banner
ads. Accordingly, an inverted-U shape pattern is expected.
For less complex banner ads (involving pictures only), perceptual fluency occurs at
low durations of exposure due to the utilisation of pictures in the ads (Bornstein 1989;
Bornstein & D’Agostino 1992; Reber et al. 1998; Labroo et al. 2008). Conceptual fluency,
however, does not occur for less complex banner ads at low exposure durations. When
exposure duration increases from low to moderate levels, viewers may become aware of
the fluency source, and experience tedium and a loss of interest in the ad. Thus the per-
ceptual fluency effect weakens, resulting in a decrease in affective response. Concurrently,
semantic processing among viewers might occur to a limited extent, but we do not expect
this to have a significant impact on viewer evaluation of target ads. As exposure duration
increases to higher levels, conceptual fluency kicks in as a result of the semantic processing
126
by viewers of the pictures in less complex banner ads (Smith & Dorfman 1975; Anand
& Sternthal 1990). Such semantic processing of pictures allows viewers to appreciate the
meaning of the pictures, and thereby enhances overall evaluation of ads. Existing scene-
perception literature also suggests that contextual scenes activate a schema for a theme,
resulting in product expectations and enhancing identification of advertised products
(Biederman et al. 1982). Consistent with a study by Whittlesea (1993), due to semantic
processing, conceptual fluency towards a less complex ad increases, and thus enhances the
overall evaluation of ads by viewers. A U pattern effect is predicted in such circumstances.
Previous research indicates that consumer attitudes about ads significantly influence
consumer attitudes about brands (MacKenzie et al. 1986; Brown & Stayman 1992).
That is, positive consumer attitudes towards an advertisement impact brand preference,
consciousness, and loyalty. In their web ad processing model, Yoon and Lee (2007) dem-
onstrated that consumer attitude towards an ad is positively correlated with brand judge-
ment. Further, Zinkhan et al. (1986) found that consumer attitude towards an ad and
brand judgements are strongly related to brand recognition. Therefore, this study expects
that a similar pattern of findings will emerge between consumer attitudes towards target
ads and consumer attitudes towards brands overall. Accordingly, the following hypotheses
are proposed.
H1: As the duration of exposure to a banner ad increases in the priming phase, the
attitudes of respondents towards the target ad and the advertised brand in the
testing phase will increase when levels of banner ad complexity are higher.
testing phase will increase and then decrease (an inverted-U pattern) when
levels of banner ad complexity are moderate.
testing phase will decrease and then increase (a U pattern) when levels of ban-
ner ad complexity are lower.
Methods
Experimental design
This study investigates whether exposure duration and banner ad complexity will impact
the attitudes of respondents towards target ads and brands. Two different products, pasta
sauce (in the hedonic product category) and a camcorder (in the utilitarian product cat-
egory), were utilised to assess the robustness of predicted effects. We employed a 3 (refer-
ring to varying exposure durations of 16, 150 or 500 milliseconds) × 3 (referring to varying
levels of banner ad complexity, including highly complex, moderately complex or less
127
complex) × 2 (referring to both the pasta sauce and camcorder products) mixed factorial
design. Exposure duration and banner ad complexity were the between-subject variables,
whereas the product category was the within-subject variable. To test the hypotheses,
multivariate analysis of variance (MANOVA) was used to examine the means.
Stimuli development
An online information-type website, consisting of several web pages, was created. In order
to create a natural advertising environment, web pages and banner ads were adapted from
an existing site (www.weather.com) and modified to fit the study objectives. The websites
contained a portal web page, and informational web pages about seasonal health, air qual-
ity, allergies and fitness.
Banner ads for pasta sauce (a hedonic product) and a camcorder (a utilitarian product)
were developed for use in the study. Fictitious brand names were developed to inhibit
any base-level brand awareness or prior attitudes among respondents that could have
produced confounding effects for analysis. Three versions of banner ads were developed
for each product (Pasto pasta sauce and a PCW camcorder). The more complex versions
contained only text messages – ‘Italy – now available in Pasto Classic take-home size’
and ‘Without a PCW Traveller, no one will ever believe you’ – along with the respective
brand logos. The second version of moderately complex ads for the products consisted
of product pictures, the text messages and the respective brand logos. The third and final
version of less complex ads for the products contained related pictures meant to evoke
context, along with pictures of the products and the brand logos. Specifically, the pasta
sauce ad contained a landscape picture of countryside with rolling hills. The camcorder ad
displayed a picture of a mountain climber hanging on to a cliff while holding a camcorder.
The brand logos were presented in the same position, the bottom right corner, in all three
versions of the ads.
A pretest was performed to assess the effectiveness of our manipulation of banner ad
complexity. Seventy-six respondents viewed the pasta sauce banner ad, and 74 respondents
viewed the camcorder banner ad. For each product category, respondents were randomly
assigned to one of the three conditions (highly complex, moderately complex or less com-
plex), and respondents subsequently viewed the corresponding banner ad for 150 milli-
seconds. Banner ad complexity was measured using a two-item, 7-point scale. Specifically,
the two items asked respondents to assess whether the banner ads were difficult to
process/easy to process and difficult to understand/easy to understand (rpasta sauce = 0.89,
rcamcorder = 0.85). Higher (lower) scores are associated with greater ease (difficulty) of
banner ad processing. As anticipated, the results showed that the text-only versions of
the ads were more difficult to process than the versions with pictures and text (pasta
sauce: M text = 1.50, Mpicture and text = 2.54; F = 11.62, p < 0.001; camcorder: Mtext = 1.54,
Mpicture and text = 2.65; F = 14.43, p < 0.001). Further, the picture with text versions of both
ads were more difficult to process than the picture-only versions of the ads (pasta sauce:
Mpicture and text = 2.54, Mpicture = 3.33; F = 6.24, p < 0.05; camcorder: Mpicture and text = 2.65,
Mpicture = 3.71; F = 12.38, p < 0.001).
128
Target ads were developed for the pasta sauce and camcorder, respectively. The target
ads featured both verbal and pictorial components, as well as other information (e.g. prod-
uct features) related to the advertised product. Portions of the target ad information (e.g.
text, picture or both) had been previously shown to the respondents in the priming phase.
Sample and experimental procedure

The sample included 219 respondents enrolled in marketing classes. Twenty-five respond-
ents were removed from the final sample for data analysis because they did not remember
the target brands in the study. The study was conducted in a campus computer lab. The
sample consisted of 40% females. The average age of respondents was 22.5 years, with a
range from 19 to 31 years. On arrival, respondents were randomly assigned to one of the
nine experimental conditions and were seated in front of a computer. They were then
asked to access the designated website. The first web page that respondents viewed was
an instructional page informing them of the same fabricated study objective (i.e. website
design evaluation) in all conditions. Respondents were not instructed to attend to either
banner ads or editorial content. After reading the instructions regarding the purpose of
the study, the respondents were asked to click the start button and begin the experiment.
Respondents were then exposed to the website portal and were told to browse the
website at their own pace. The web pages contained no active links, so respondents could
browse only the experimental website. Respondents were instructed to click through to
the next web page when they finished browsing the current web page. Previous research
has demonstrated that target pictures preceded by a facilitating prime are judged more
favourably than pictures not associated with such a prime (Whittlesea 1993; Reber et al.
1998). It was expected that a facilitating banner ad in the priming phase would influence
the evaluation of the target ad in the testing phase. Thus every respondent proceeded at
their own pace through web pages consisting of a priming phase (a brief exposure to the
banner ad) and a testing phase (viewing the target ad) for the Pasto pasta sauce and PCW
camcorder during the experiment.
The order of ad appearance on the web pages was: banner ad 1 (priming phase), fol-
lowed by target ad 1 (testing phase); banner ad 2 (priming phase), followed by target ad
2 (testing phase). The order in which the two products were introduced was counter
balanced across respondents. In the priming phase, a filler banner ad was shown for a
short period of time (2 seconds) prior to presentation of the manipulated banner ad.
The filler banner ad was subsequently replaced with one of the three priming banner ads
(highly complex, moderately complex or less complex) shown for 16, 150 or 500 mil-
liseconds on the same web page. Following exposure to the banner ad for the prescribed
time duration, another filler banner ad was viewed to minimise any short-term memory
effects. The filler banner ad was displayed until the respondent clicked through to the next
web page. On the following web page, the corresponding target ad (either the pasta sauce
or the camcorder) was presented, featuring both verbal and pictorial components, as well
as other information (e.g. product features) about the advertised product. As mentioned,
portions of the target ad information (e.g. text, picture or both) had been previously
129
shown to the respondents in the priming phase. Respondents viewed the target ad at their
own pace. The target ad remained on the web page until respondents moved to the next
web page.
The banner ads shown in the priming and testing phases were placed above the main
editorial content of the web pages. The order of content on the web pages was seasonal
health, air quality, allergies and fitness. The two filler banner ads in the first priming phase
featured bank mortgages and loans, respectively, whereas the two filler banner ads in the
second priming phase featured car rental companies. The content order on the web pages
and the order of the filler banner ads were the same in all conditions.
After exposure to the priming and testing phases of the web pages for both products,
respondents were directed to the questionnaire site. Respondents were asked to rate their
attitude towards the website on a five-item, 5-point scale (Chen & Wells 1999). These
questions were used to mask the actual purpose of the study. The respondents were asked
to respond to items assessing their attitudes towards the target ads. Subsequently, the
brand names of each of these two products were presented, and respondents were asked
to generate thoughts about each of the two target brands. Finally, several demographic
questions were administered, and the respondents were debriefed and thanked.
Banner ad complexity
The three levels of banner ad complexity utilised were highly complex, moderately com-
plex and less complex levels, and the condition was manipulated by varying the banner
ad content. The banner ads contained only text, a picture with text, or only a picture to
represent ads that are highly complex, moderately complex and less complex, respectively.
Exposure duration
Each banner ad, sized at 450 × 100 pixels, was presented for three exposure durations
(16, 150 or 500 milliseconds). Several studies have found that stimulus exposure times
of less than 50 milliseconds – e.g. 5 milliseconds, according to Bornstein & D’Agostino
(1992); 8 milliseconds, according to Seamon et al. (1995); and 16 milliseconds, accord-
ing to Bornstein (1989) and Labroo et al. (2008) – are precognitive, and that they rely
more strongly on perceptual processing than on cognition. The findings of Seamon et al.
(1995) suggested that subliminal processing (8 milliseconds) increases perceptual fluency.
Bornstein (1989) further showed that stimulus recognition associated with longer dura-
tions of exposure (200 vs 50 milliseconds in experiment 1, and 600 vs 16 milliseconds in
experiment 2) to a prime stimulus may inhibit perceptual fluency. These previous studies
manipulated the exposure duration to a prime and showed that brief exposures lead to
perceptual fluency. The present study utilised exposure durations of 16 (unaware) and
500 (aware) milliseconds, because these durations have been shown in previous studies
to result in priming effects (Zajonc 1980; Bornstein 1989; Jacoby & Whitehouse 1989;
Bornstein & D’Agostino 1992; Reber et al. 1998; Labroo et al. 2008). An exposure dura-
tion of 150 milliseconds was selected as a midpoint.
130
Dependent variables
Dependent variables included respondent attitudes towards the target ad and the adver-
tised brand. Target ad attitude was measured based on respondent ratings on a three-item,
7-point scale aimed to determine how pleasant, how attractive and how likeable the
respondents found the target ad to be. As these items loaded on to a single factor for the
two products, the items were averaged to form two separate ad attitude scales (α = 0.95
and α = 0.94 for the pasta sauce and the camcorder, respectively).
The attitudes of respondents towards the advertised brand were assessed by coding
all the responses about the brand. Two independent judges, who were blind to treatment
of the products, classified the valences of thoughts by respondents about the advertised
brand. The thoughts were classified as positive (e.g. ‘it looked like a nice brand’), negative
(e.g. ‘I did not like the brand very much’) and neutral (e.g. ‘I don’t have particular thoughts
about the brand’). The judges maintained an extensive list of thoughts by category to
ensure consistent coding. The intercoder reliability was 0.88 and all disagreements were
resolved through discussion. A net favourable brand index was obtained by subtracting
the number of negative thoughts from the number of positive thoughts, and then dividing
the difference by the total number of thoughts (Peracchio & Meyers-Levy 1994, 2005).
This index was subjected to arcsine transformation in order to render the variable suit-
able for multivariate analysis of variance (MANOVA) and univariate analysis of variance
(ANOVA).
Results
Preliminary analysis revealed no significant effects for either the order of the evaluation
and thought-listing tasks, or for the order of presentation of the pasta sauce and cam-
corder products (Fs < 1). Dependent variables were subjected to a 3 × 3 × 2 MANOVA,
and the data analyses revealed that the patterns of means for the two products were not
significantly different across the two dependent variables (Fs < 1). Therefore, the means
of the two products were collapsed (Smith & Cooper-Martin 1997; Barone et al. 2000;
Roberts & Pirog 2004), and a 3 × 3 MANOVA was performed. The results are summa-
rised in Table 1.
The MANOVA results showed that the main effect was significant for exposure dura-
tion (Wilks’ lambda = 0.91, p < 0.01) and banner ad complexity (Wilks’ lambda = 0.89,
p < 0.01). The univariate ANOVA results revealed that the main effect of exposure dura-
tion was significant on ad attitude (F (2, 185) = 7.51, p < 0.001), and marginally significant
on brand attitude (F (2, 185) = 2.56, p < 0.08). In addition, the main effect of banner ad
complexity was marginally significant on ad attitude (F (2, 185) = 2.54, p < 0.09), but not
significant on brand attitude (F < 1). The MANOVA results indicated that the interac-
tion of exposure duration and banner ad complexity was significant (Wilks’ lambda = 0.64,
p < 0.001). We further evaluate the univariate ANOVA results for the attitudes of
respondents about the ads and brands. The interaction of exposure duration with ban-
ner ad complexity was significant on both ad attitude (F (4, 185) = 21.70, p < 0.001) and
131
Table 1: MANOVA and univariate ANOVA results
Banner ad type
Text Picture with text Picture
Main Main
effect: effect: ED × BC
16 ms 150 ms 500 ms 16 ms 150 ms 500 ms 16 ms 150 ms 500 ms ED BC interaction
Viewer ad 3.43 4.04 4.79 2.83 4.74 3.75 4.85 3.00 4.56 F= F= F=
attitude (0.80) (1.00) (0.73) (1.45) (0.88) (0.82) (1.00) (1.23) (0.92) 7.51*** 2.54+ 21.70***
Proportion 0.18 0.42 0.64 0.12 0.71 0.36 0.45 0.04 0.40 F= F= F = 5.48***
of net (0.33) (0.33) (0.57) (0.26) (0.64) (0.58) (0.53) (0.77) (0.68) 2.56* 0.85n.s.
favourable
brand
thoughts
Combined Wilks’ Wilks’ Wilks’

variables lambda lambda lambda
(dependent = = =
vector) 0.91*** 0.89*** 0.64***
N 20 20 22 20 22 24 23 20 23
Notes: MANOVA: multivariate analysis of variance; ED: exposure duration; BC: banner ad complexity. Standard deviations are in
parentheses.
*p < 0.10; **p < 0.05; **p < 0.01; n.s. (not significant): p > 0.10
brand attitude (F (4, 185) = 5.48, p < 0.001). In order to test our research hypotheses, we
conducted a trend analysis on the two dependent variables.
Ad attitude
There was a significant linearly increasing trend (F (1, 59) = 26.84, p < 0.001) when
respondents viewed the difficult-to-process (text only) banner ad, in comparison to the
significant quadratic (inverted-U) trend (F (1, 63) = 27.05, p < 0.001) that was observed
with the moderately difficult-to-process ad (picture with text), in which respondent
attitudes towards the target ad first increased then decreased. There was a significant
quadratic (U-shaped) trend for ad attitude when respondents viewed the easy-to-process
(picture only) banner ad (F (1, 63) = 36.11, p < 0.001), with attitudes first decreasing and
then increasing.
Brand attitude
There was a significant linearly increasing trend (F (1, 59) = 11.80, p < 0.001) when
respondents viewed the difficult-to-process (text only) banner ad, in comparison to the
significant quadratic (inverted-U) trend (F (1, 63) = 11.74, p < 0.001) that was observed
when respondents viewed the moderately difficult-to-process (picture with text) banner
ad. Trend analysis indicated a significant quadratic (U-shaped) trend for the easy-to-
process (picture only) banner ad (F (1, 63) = 4.63, p < 0.05). These findings supported the
research hypotheses.
132
Figure 1: Interactive effects of banner ad type and exposure duration on viewer

ad attitude
5
Text (highly complex)
4.5 Picture with text

(moderately complex)
Ad attitude
4
Picture (less complex)
3.5
2.5
16 ms 150 ms 500 ms
Exposure duration
Figure 2: Interactive effects of banner ad type and exposure duration on

proportion of net favourable brand thoughts
0.8
Proportion of net favourable brand thoughts
Text (highly complex)
0.6 Picture with text

(moderately complex)
Picture (less complex)

0.4
0.2
0
16 ms 150 ms 500 ms
Exposure duration
Overall, the results for each of the dependent variables supported the predictions.
Highly complex (text only) banner ads generated increasingly favourable effects for
the target brand and ad as the duration of exposure to priming banner ads increased.
Moderately complex (picture with text) banner ads produced an inverted-U-shaped effect
as the exposure duration to priming banner ads increased. Less complex (picture only)
banner ads generated a U-shaped effect as the exposure duration to priming banner ads
increased.
133
General discussion
The findings of this research support the proposed theories and are consistent with
known processing fluency effects. Consistent with previous research on banner advertis-
ing (Briggs & Hollis 1997; Cho et al. 2001; Yoon & Lee 2007), the present study shows
that banner advertising generates desirable effects for advertising and brand judgements.
Although increasing the duration of exposure enhances processing fluency under some
conditions, our findings suggest that longer exposure durations do not always lead to more
favourable effects. Instead, exposure duration and ease of banner ad processing jointly
influenced attitudes towards the ad and brand judgements among the participants in this
study. The optimal duration of exposure appears to depend upon the banner ad complex-
ity. When the banner ad was difficult to process, increasing the duration of exposure
caused a linear increase in respondent attitudes towards the target brand and ad. However,
when the priming banner ad was moderately difficult to process, increasing the duration
of exposure first increased and then decreased respondent attitudes towards the target ad
and brand (an inverted-U pattern). When the banner ad was easy to process, increasing
the duration of exposure first decreased and then increased respondent attitudes towards
the target ad and brand (a U pattern). In addition, the results in the moderately difficult-
to-process and easy-to-process conditions indicate that, when exposure duration increases,
the processing fluency effect is reduced. This finding is consistent with the processing
fluency theory. In general, such findings might be caused by respondent awareness of
fluency sources.
The findings also suggest that banner ads activate perceptual and semantic processing.
Specifically herein, the types of processes activated were determined by the banner ad
complexity and the duration of exposure for viewers. Perceptual processing was activated
at subliminal levels of exposure duration (e.g. 16 milliseconds), whereas semantic process-
ing was activated at supraliminal levels (e.g. 500 milliseconds). The effects of perceptual
and conceptual fluencies on ad and brand attitudes were additive. That is, even when per-
ceptual fluency effects became weaker for respondents in this study, as long as conceptual
fluency occurred, the fluency effects together enhanced the affective judgements towards
target brands and ads among respondents.
The results of this research contribute to our understanding of processing fluency
effects in several ways. From a theoretical point of view, this research supports the supposi-
tion that processing fluency induced by perceptual or conceptual priming enhances evalu-
ations of any target ad subsequently experienced by viewers. Based on previous research
(Whittlesea 1993; Reber et al. 1998), this study used both a priming phase and a testing
phase when presenting banner ads to respondents. It was expected that the priming ban-
ner ads would facilitate the perceptual or conceptual processing of subsequently presented
target ads, and thereby lead to more favourable attitudes among the viewers. The findings
support the prediction.
Previous research demonstrates that stimulus characteristics (e.g. complexity) interact
with the number of exposures to a stimulus to determine the degree of processing fluency
among viewers ( Janiszewski & Meyvis 2001). Likewise, the present research identifies
134
boundary conditions for processing fluency effects via exposure duration and ease of
banner ad processing. Longer exposure duration to a priming banner ad, which leads to
increased processing fluency, does not always lead to more desirable attitudes towards
subsequent target ads. The findings demonstrate how the content of priming banner ads
(complexity) influences processing fluency and viewer evaluations of subsequent target ads
and advertised brands in the context of web advertising. In addition, despite discrepant
views on banner ad effects, this study shows that banner ads can influence ad attitude even
in cases when users do not click through to visit product websites (Mitchell & Valenzuela
2005).
From a practical perspective, this research has important implications for web advertis-
ing strategies, particularly for designing banner ads and establishing optimal ad exposure
duration. This research suggests that, in order to generate favourable effects (e.g. positive
ad and brand attitudes among consumers), advertisers should consider the complexity of
an ad when designing banner ad content. The content of a banner ad, in terms of visual
and verbal information, influences the banner ad complexity for viewers and thus impacts
viewer evaluations of the target ad and brand.
This research used three common types of internet banner ad (i.e. picture only, a
picture and text, and text only) in order to manipulate the banner ad complexity.
Consistent with previous research (McQuarrie & Mick 2003), visual cues were more eas-
ily processed than visual plus verbal cues. In turn, visual plus verbal cues were more easily
processed than strictly verbal cues. The pictures presented in the priming phases of our
research were the same as the pictures presented in the testing phases. Accordingly, very
brief exposure to a strictly visual priming banner ad is more likely to enhance perceptual
fluency and the evaluation of a subsequently presented target ad. This finding is consist-
ent with the idea that the effects of perceptual fluency are likely to be strongest when the
perceptual features encoded during an initial presentation of a stimulus exactly match the
perceptual features available during subsequent presentation of the stimulus (Roediger
1990). Unlike with strictly visual information, consumers may need more time and more
elaboration on verbal information in order to activate semantic analysis. Thus when pre-
senting only verbal cues in priming banner ads, advertisers should increase the initial brief
duration of exposure for viewers in order to generate favourable viewer attitudes towards
target ads in subsequent exposures.
The very brief priming phase exposure durations of 16, 150 and 500 milliseconds in
this research might not reflect a real-life situation due to current industry unit standards.
Regardless, the processing fluency effects produced by presenting banner ads in a priming
phase do influence consumer evaluations towards target ads shown in the testing phase.
Thus practitioners may want to determine how to advertise their brands and products in
similar settings. For instance, advertisers may present a priming banner ad at a website
portal, where viewers visit and click on headline links of interest. The advertisers would
then present a second target ad on corresponding web pages showing news or articles in
greater detail. When browsing on web portals, viewers usually focus on the main content
and click away from the portal to the next web page of interest quickly. User exposure to
priming banner ads on website portals could be very brief, similar to the brief exposure
135
durations utilised in the context of our research. Moreover, according to the Rich Media
Creative Guidelines, the IAB Ad Sizes Working Group recommends a maximum
exposure duration of 18 frames per second (IAB 2009b), which allows each frame to be
presented to viewers for approximately 55 milliseconds. The three exposure durations
in this research allow approximately one frame, three frames and ten frames per second,
respectively. The present results can therefore provide a guideline for content planning in
animated banner ads.
Incidental exposure and short exposure duration to banner ads in an online environ-
ment present challenges to advertisers developing banner ads. Specifically, advertisers need
to determine what types of information should be included in banner ads at different
exposure durations. Although exposure duration is not a variable that is often controlled
in the current online environment, the findings of the present study suggest that it may be
one of the critical elements influencing the effectiveness of online banner advertising. The
interactive advertising industry is now facing creative challenges with existing standards,
ad units and technologies. Many practitioners do not believe that existing standards best
facilitate brand building or encourage innovation and creativity in online advertising (IAB
2009b). This research provides an alternative direction for the future.
Practitioners may want to investigate the possibility of controlling the exposure dura-
tion of priming banner ads. The development of technology helps advertisers identify
and target potential consumers with relevant behavioural criteria, such as online browsing
interests. Advanced advertising technologies allow advertisers to target consumers using
more relevant and localised messages (Truong et al. 2011). Since this study demonstrates
that priming and testing phases together enhance processing fluency and ad evaluations,
practitioners are encouraged to take advantage of advanced technology in order to develop
similar settings in online environments. Given that the effectiveness of banner advertis-
ing is often undervalued by measuring click-through rates, which is a metric that does
not capture the full impact of how the viewing of an advertisement influences consumer
attitudes about ads and brands, the findings from this research demonstrate that banner
advertising is an important vehicle to enhance brand evaluations in line with previous
research (Mitchell & Valenzuela 2005). Our research suggests that scholars and practi-
tioners should continue developing and testing banner ads to optimise advertising effects
among consumers (Taylor 2009).
While our research investigated the advertising effects in a traditional web advertising
environment (i.e. a web portal), an increasing population of consumers have been attracted
by various new media formats (e.g. Facebook and Twitter). Thus, it would be important
for future research to explore how the observed effects in our study differ in the new
media formats. With the development of Web 2.0, which emphasises two-way com-
munication between users, consumers engage in online brand-related activities such as
viewing brand-related video, rating brands and writing brand-related articles (Muntinga
et al. 2011), electronic word-of-mouth and referral behaviour (Chatterjee 2011; Chu &
Kim 2011) on social network sites. The new media, which have characteristics such as
interactivity, customisation and social interaction, significantly influence online consumer
behaviour and deserve more research effort (Hill & Moran 2011).
136
Despite its contributions, this research has some limitations and leaves a number of
questions unanswered. First, the study used two fictitious brand names and logos in the
banner ads, about which respondents had no existing knowledge. Future research should
examine whether the same findings are obtained when consumers are familiar with the
advertised brands. Second, the stimulus complexity should be varied in ways other than
the ways utilised herein. In some situations, a visual stimulus could be more complex and
difficult to process than a verbal message would be ( Janiszewski & Meyvis 2001). In such
circumstances, the results observed in the picture-only condition in this research might
not be reproducible. In addition, the banner ad content in this study consisted primarily
of scenes involving nature (e.g. mountains and rolling hills), which were congruent with
the context of the website (e.g. weather and air quality). Previous research has shown that
congruity between an internet banner ad and website context generates more favourable
ad attitudes, whereas incongruity produces favourable effects on recognition and recall
(Newman, Stem & Sprott 2004; Moore et al. 2005). Empirical studies are required to
further investigate the possibility that the degree of congruity has a meaningful impact
on the effects of exposure duration and banner ad complexity for consumer ad and brand
attitudes, which is the focus of the current study (Yaveroglu & Donthu 2008). In fact, we
did not investigate the impacts of exposure duration or banner ad complexity on memory
measures (e.g. recognition and recall). Future research may want to include these measures
to determine whether the effects on memory are similar to the effects on attitude among
consumers. This research did not include a control group and measure perceptual and
conceptual fluency. Future research could include a control group as well as perceptual
and conceptual fluency measures in the experimental design in order to provide evidence
of the existence of priming effects and additional support for the process.
Further, future research could also examine the impact of individual differences on the
effects observed in this study. For instance, the aesthetic preferences of novices and experts
(Dahlen 2001; Reber et al. 2004), as well as varying consumer information processing ori-
entations (Childers et al. 1985), may influence the banner ad complexity and evaluations
of target brands and ads (Dahlen 2001; Reber et al. 2004). Finally, repeated exposure to the
same banner ad may be considered intrusive by some consumers, and thereby have nega-
tive effects on consumer ad and brand attitudes. The optimal display frequency of banner
ads needs to be empirically determined (Ying et al. 2009) in future research.
References
Anand, P. & Sternthal, B. (1990) Ease of message processing as a moderator of repetition
effects in advertising. Journal of Marketing Research, 27, 345–353.
Barone, M.J., Miyazaki, A.D. & Taylor, K.A. (2000) The influence of cause-related marketing
on consumer choice: does one good turn deserve another? Journal of the Academy of
Marketing Science, 28, 248–262.
Berlyne, D.E. (1970) Novelty, complexity, and hedonic value. Perception and Psychophysics, 8,
279–286.
Biederman, I., Mezzanotte, R.J. & Rabinowitz, J.C. (1982) Scene perception: detecting and
judging objects undergoing relational violations. Cognitive Psychology, 14, 143–177.
137
Bornstein, R.F. (1989) Exposure and affect: overview and meta-analysis of research 1968–1987.
Psychological Bulletin, 106, 265–289.
Bornstein, R.F. & D’Agostino, P. (1992) Stimulus recognition and the mere exposure effect.
Journal of Personality and Social Psychology, 62, 545–552.
Bornstein, R.F. & D’Agostino, P. (1994) The attribution and discounting of perceptual fluency:
preliminary tests of a perceptual fluency/attribution model of the mere exposure effect.
Social Cognition, 12, 103–128.
Briggs, R. & Hollis, N. (1997) Advertising on the web: is there any response before click-
through? Journal of Advertising Research, 37, 33–45.
Brown, S.P. & Stayman, D.M. (1992) Antecedents and consequences of attitude toward the ad:
a meta-analysis. Journal of Consumer Research, 19, 34–50.
Carr, T.H., McCauley, C.C., Sperber, R.D. & Parmelee, C.M. (1982) Words, pictures,
and priming: on semantic activation, conscious identification, and the automaticity
of information processing. Journal of Experimental Psychology: Human Perception and
Performance, 8, 757–777.
Chatterjee, P. (2011) Drivers of new product recommending and referral behaviour on social
network sites. International Journal of Advertising, 30, 77–101.
Chen, Q. & Wells, W.D. (1999) Attitude toward the site. Journal of Advertising Research, 39,
27–38.
Cheng, J.M., Blankson, C., Wang, E.S. & Chen, L.S. (2009) Consumer attitudes and
interactive digital advertising. International Journal of Advertising, 28, 501–525.
Childers, T.L. & Houston, M.J. (1984) Conditions for a picture-superiority effect on consumer
memory. Journal of Consumer Research, 11, 643–654.
Childers, T.L., Houston, M.J. & Heckler, S.E. (1985) Measurement of individual differences in
visual and verbal information processing. Journal of Consumer Research, 12, 125–134.
Cho, C.-H., Lee, J.-G. & Tharp, M. (2001) Different forced-exposure levels to banner
advertisements. Journal of Advertising Research, 41, 45–56.
Chu, S.C. & Kim, Y. (2011) Determinants of consumer engagement in electronic word-of-
mouth (eWOM) in social networking sites. International Journal of Advertising, 30, 47–75.
Clifford, S. (2008) Web marketing that hopes to learn what attracts a click. New York Times.
Available online at: http://www.nytimes.com/2008/12/03/business/media/03adco.html?_
r=2&emc=eta1&pag (accessed 25 October 2009).
comScore.com (2008) E-marketing effectiveness benchmarks for the pharmaceutical industry.
Available online at: http://www.comscore.com/press/release.asp?press=2343 (accessed
23 July 2008).
Dahlen, M. (2001) Banner advertisements through a new lens. Journal of Advertising Research,
41, 23–30.
Dreze, X. & Hussherr, F.X. (2003) Internet advertising: is anybody watching? Journal of
Interactive Marketing, 17, 8–23.
Filloux, F. (2009) Measuring time spent on a web page. Monday Note. Available online at:
http://www.mondaynote.com/2009/05/24/measuring-time-spent-on-a-web-page/ (accessed
18 November 2009).
Gardner, M.P. & Houston, M.J. (1986) The effects of verbal and visual components of retail
communications. Journal of Retailing, 62, 64–78.
Hill, R.P. & Moran, N. (2011) Social marketing meets interactive media. International Journal
of Advertising, 30, 815–838.
Hirschman, E.C. (1986) The effect of verbal and pictorial advertising stimuli on aesthetic,
utilitarian and familiarity perceptions. Journal of Advertising, 15, 27–34.
IAB (2009a) IAB Internet Advertising Revenue Report. PricewaterhouseCoopers. Available
online at: http://www.iab.net/AdRevenueReport (accessed 19 November 2009).
138
IAB (2009b) IAB ad unit guidelines 2009 update. Available online at: http://www.iab.net/
ad_unit (accessed 19 November 2009).
IAB & MBI (1997) The IAB online advertising effectiveness study. Available online at: http://
www.intelliquest.com/search/results.asp?vpath=/resources/reports/reports.asp (accessed
15 January 2007).
Jacoby, L.L. & Dallas, M. (1981) On the relationship between autobiographical memory and
perceptual learning. Journal of Experimental Psychology: General, 110, 306–340.
Jacoby, L.L. & Whitehouse, K. (1989) An illusion of memory: false recognition influenced by
unconscious perception. Journal of Experimental Psychology: General, 118, 126–135.
Jacoby, L.L., Kelly, C.M. & Dywan, J. (1989) Memory attributions. In: Roediger, III, H.L. &
Craik, F.I.M. (eds) Essays in Honor of Endel Tulving. Hillsdale, NJ: Erlbaum, 391–422.
Janiszewski, C. & Meyvis, T. (2001) Effects of brand logo complexity, repetition, and spacing
on processing fluency and judgment. Journal of Consumer Research, 28, 18–32.
Kim, J. & McMillan, S.J. (2008) Evaluation of internet advertising research: a bibliometric
analysis of citations from key sources. Journal of Advertising, 37, 99–112.
Kim, M. & Lennon, S. (2008) The effects of visual and verbal information on attitudes and
purchase intentions in internet shopping. Psychology & Marketing, 25, 146–178.
Labroo, A.A. & Lee, A.Y. (2006) Between two brands: a goal fluency account of brand
evaluation. Journal of Marketing Research, 43, 374–385.
Labroo, A.A., Dhar, R. & Schwarz, N. (2008) Of frog wines and frowning watches: semantic
priming, perceptual fluency, and brand evaluation. Journal of Consumer Research, 34, 819–831.
Lee, A.Y. & Labroo, A.A. (2004) The effect of conceptual and perceptual fluency on brand
evaluation. Journal of Marketing Research, 41, 151–165.
Lee, M. & Youn, S. (2009) Electronic word of mouth (EMOW) – how eMOW platforms
influence consumer product judgment. International Journal of Advertising, 28, 473–499.
Lutz, K.A. & Lutz, R.J. (1977) Effects of interactive imagery on learning: application to
advertising. Journal of Applied Psychology, 62, 493–498.
MacKenzie, S.B., Lutz, R.J. & Belch, G.E. (1986) The role of attitude toward the ad as a
mediator of advertising effectiveness: a test of competing explanations. Journal of Marketing
Research, 23, 130–143.
McQuarrie, E.F. & Mick, D.G. (2003) Visual and verbal rhetorical figures under directed
processing versus incidental exposure to advertising. Journal of Consumer Research, 29,
579–587.
Mitchell, A. & Valenzuela, A. (2005) How banner ads affect brand choice without click-
through. In: Haugtvedt, C.P., Machleit, K.A. & Yalch, R.F. (eds) Online Consumer
Psychology. Mahwah, NJ: Lawrence Erlbaum, 125–142.
Moore, R.S., Stammerjohan, C.A. & Coulter, R.A. (2005) Banner advertiser-web site context
congruity and color effects on attention and attitudes. Journal of Advertising, 34, 71–84.
Muntinga, D.G., Moorman, M. & Smit, E.G. (2011) Introducing COBRAs: exploring
motivations for brand-related social media use. International Journal of Advertising, 30,
13–46.
Nelson, D.L., Reed, V.S. & Walling, J.R. (1976) Pictorial superiority effect. Journal of
Experimental Psychology: Human Learning and Memory, 2, 523–528.
Newman, E.J., Stem, D.E. & Sprott, D.E. (2004) Banner advertisement and web site congruity
effects on consumer web site perceptions. Industrial Management and Data Systems, 104,
273–281.
Paivio, A. (1969) Mental imagery in associative learning and memory. Psychological Review, 76,
241–263.
Peracchio, L.A. & Meyers-Levy, J. (1994) How ambiguous cropped objects in ad photos can
affect product evaluations. Journal of Consumer Research, 21, 190–204.
139
Peracchio, L.A. & Meyers-Levy, J. (2005) Using stylistic properties of ad pictures to

communicate with consumers. Journal of Consumer Research, 32, 29–40.
Reber, R., Schwarz, N. & Winkielman, P. (2004) Processing fluency and aesthetic pleasure: is
beauty in the perceiver’s processing experience? Personality and Social Psychology Review, 8,
364–382.
Reber, R., Winkielman, P. & Schwarz, N. (1998) Effects of perceptual fluency on affective
judgments. Psychological Science, 9, 45–48.
Roberts, J.A. & Pirog, S.F. (2004) Personal goals and their role in consumer behavior: the case
of compulsive buying. Journal of Marketing Theory and Practice, 12, 61–73.
Roediger, H.L. (1990) Implicit memory: retention without remembering. American Psychologist,
45, 1043–1056.
Roediger, H.L. & McDermott, K.B. (1992) Implicit memory in normal human subjects. In:
Boller, F. & Grafman, J. (eds) Handbook of Neuropsychology, Vol. 8. Amsterdam: Elsevier,
63–131.
Schacter, D.L. (1987) Implicit memory: history and current status. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 13, 501–518.
Seamon, J.G., Williams, P.C., Crowley, M.J., Kim, I.J., Langer, S.A., Orne, P.J. & Wisengrad,
D.L. (1995) The mere exposure effect is based on implicit memory: effects of stimulus type,
encoding conditions, and number of exposures on recognition and affect judgments. Journal
of Experimental Psychology: Learning, Memory, and Cognition, 21, 711–721.
Shapiro, S. (1999) When an ad’s influence is beyond our conscious control: perceptual and
conceptual fluency effects caused by incidental ad exposure. Journal of Consumer Research,
26, 16–36.
Shepard, R.N. (1967) Recognition memory for words, sentences, and pictures. Journal of Verbal
Learning & Verbal Behavior, 6, 156–163.
Smith, G.F. & Dorfman, D.D. (1975) The effect of stimulus uncertainty on the relationship
between frequency of exposure and liking. Journal of Personality and Social Psychology, 31,
150–155.
Smith, N.C. & Cooper-Martin, E. (1997) Ethics and target marketing: the role of product
harm and consumer vulnerability. Journal of Marketing, 61, 1–20.
Sojka, J.Z. & Giese, J.L. (2006) Communicating through pictures and words: understanding
the role of affect and cognition in processing visual and verbal information. Psychology &
Marketing, 23, 995–1014.
Stevenson, J.S., Bruner II, G.C. & Kumar, A. (2000) Webpage background and viewer
attitudes. Journal of Advertising Research, 40, 29–34.
Taylor, C.R. (2009) The six principles of digital advertising. International Journal of Advertising,
28, 411–418.
Truong, Y., McColl, R. & Kitchen, P. (2011) Practitioners’ perceptions of advertising strategies
for digital media. International Journal of Advertising, 29, 709–725.
Whittlesea, B.W.A. (1993) Illusions of familiarity. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 19, 1235–1253.
Winkielman, P. & Cacioppo, J.T. (2001) Mind at ease puts a smile on the face:
psychophysiological evidence that processing facilitation elicits positive affect. Journal of
Personality and Social Psychology, 81, 989–1000.
Yaveroglu, I. & Donthu, N. (2008) Advertising repetition and placement issues in on-line
environments. Journal of Advertising, 37, 31–43.
Ying, L., Korneliussen, T. & Gronhaug, K. (2009) The effect of ad value, ad placement and
ad execution on the perceived intrusiveness of web advertisements. International Journal of
140
Yoo, C., Bang, H. & Kim, Y. (2009) The effects of consistent ad series on consumer evaluations:
a test of repetition-variation hypothesis in a South Korean context. International Journal of
Yoon, H.S. & Lee, D.H. (2007) The exposure effect of unclicked banner advertisement. Cross-
Cultural Buyer Behavior, Advances in International Marketing, 18, 211–229.
Zajonc, R.B. (1980) Feeling and thinking: preferences need no inferences. American Psychologist,
35, 151–175.
Zinkhan, G.M., Locander, W.B. & Leigh, J.H. (1986) Dimensional relationships of aided
recall and recognition. Journal of Advertising, 15, 38–46.
About the authors

Kai-Yu Wang (PhD, University of Wisconsin-Milwaukee) is an associate professor
of marketing at the Department of Marketing, International Business and Strategy,
Goodman School of Business, Brock University. His research interests include consumer
psychology, marketing communication and WOM marketing in virtual communities. His
work has been published in Journal of Business Research, Industrial Marketing Management,
Service Industries Journal and Advances in Consumer Research.
Eric Shih is associate professor of marketing at SKK GSB, Sungkyunkwan University
in Seoul, Korea. His research interests include diffusion of technology and market-
ing managers’ strategic decision-making process. His work has appeared in Journal of
Marketing, Management Science, Journal of Consumer Psychology, Journal of Retailing and
Strategic Management Journal.
Laura A. Peracchio is Professor of Marketing at the University of Wisconsin-Milwaukee.
Laura received her PhD from Northwestern University. Laura’s areas of research inter-
est are focused on food and nutrition issues, visual persuasion, and language and culture.
Her work has appeared in the Journal of Consumer Research, Journal of Marketing Research,
Journal of Consumer Psychology, Journal of Public Policy and Marketing and Journal of
Advertising.
Address correspondence to: Eric Shih, Associate Professor, SKK Graduate School
of Business, Sungkyunkwan University, 53 Myungryun-dong 3-ga, Jongro-gu, Seoul
110-745, Korea.
Email: eshih@skku.edu.
141
Advertising and promotion budgeting
during volatile economic conditions
Factors influencing the level of decentralisation
in budgeting and its relations to budget size
and allocation
Yunjae Cheong
Hankuk University of Foreign Studies
Kihan Kim
Seoul National University
Hyuksoo Kim
St. Bonaventure University
This paper provides a perspective on traditional budgeting approaches and managerial pro-
cesses in the midst of the severe economic downturn during May 2009, in the United States.
Specifically, it examines the extent to which various company, brand and organisational factors
are related to the level of decentralisation in the budgeting for advertising and promotion.
Also, the influences of the level of decentralisation in the budgeting on the marketing budget
size and the advertising-to-promotion ratio were examined. It appeared that the company
size, company profitability, brand equity, brand price, brand type and the influences of the
marketing and the finance departments were associated with the level of decentralisation in
the budgeting, which, in turn, had significant impact on the actual budget size and allocation.
Introduction
The world currently faces its worst economic crisis since the Great Depression of the
1930s (Macleod 2009). During a recession, retail sales are slow, and advertising budgets
are often trimmed back, as firms face severe pressure to cut back marketing investments
(Willis & Willis 1993; Srinivasan et al. 2005). Several observers (Barwise 1999; Brenner
2009), however, suggest that firms that either maintain or increase their marketing spend-
ing during difficult economic times see significant benefits.
Budgeting research in advertising and marketing suggests that the specific methods
to set marketing budgets play significant roles in determining the budget size, and are
associated with various explanatory factors related to the firm and the product (Piercy

DOI: 10.2501/IJA-32-1-143-162
1987a, 1987b; Hung & West 1991). The prior research, however, was conducted when
the economy was good; we know little about whether the relationships between budgeting
methods and various explanatory factors reported in the literature are accurate reflections
of how firms act during a recession.
These issues are important for both managerial practice and theory (e.g. Bhargava 2011;
Zdravkovic & Till 2012). Business cycles are a common occurrence in the world’s major
economies. According to the National Bureau of Economic Research, there have been 11
recessions in the US in the 58 years since World War II – that is, a major recession every six
years or so (Srinivasan et al. 2005). Despite the severe effects of recessions on firms’ perfor-
mance, there is little empirical study to bridge the gap in our understanding of marketing
budgeting during those challenging times. Furthermore, there is a dearth of research that
generates insights to guide firms in their budget decisions when the economy slows down.
We attempt to examine advertising budgeting practices in the context of the severe
economic downturn in 2009, triggered by a liquidity shortfall in the US banking system.
Specifically, using a survey of 169 budgeting decision makers in US firms, this paper
examines the influences of company, brand and organisational factors on the degree of
decentralisation in the budgeting process at the brand level, and its relation to the adver-
tising and promotion budget size and budget allocation.
In this study, advertising spending indicates the investments in the ad creation and
media purchases, while promotion spending indicates all the other non-advertising mar-
keting expenses. Thus, the advertising and the promotion budgets are operationalised
together to comprise the total marketing budget. The following sections first review vari-
ous budgeting methods and processes, and their relations to the level of decentralisation
in the budgeting process. Then a series of hypotheses was developed, based on the review
of research in advertising and marketing budgeting.
Budgeting methods and level of decentralisation

There are several methods, or decision rules, on which many firms draw in making budget
decisions (Aaker et al. 1992). Previous research classifies such techniques into five distinct
categories. Table 1 describes the nine most common budgeting techniques by each of
the five categories – judgement, competitive, sales, measurement, and objective and task
(Hung & West 1991; Prendergast, West & Shi 2006).
The budgeting categories shown in Table 1 are systematically associated with one of
two processes. The first is the top-down process, where the initiative in budget decisions
is with senior management personnel and various types of committee, with little or no
input from, and minimum interaction with, those professionals who handle the day-to-
day marketing plans (Piercy 1987a, 1987b). On the other hand, the initiative lies with the
marketing department for the bottom-up process, which involves greater negotiation and
interaction between senior management and marketing managers; thus, those marketing
practitioners at a lower level of the hierarchy in an organisational structure have greater
influence on the final budget decisions (Hung & West 1991). The budgeting practices
identified in the previous research are explained by either the bottom-up or the top-down
144
Ad and promotion budgeting during volatile economic conditions
Table 1: Descriptions and distribution of budgeting methods
Category Method Description N %

Judgement Arbitrary Based on what is ‘felt’ to be necessary 77 46.1
Affordable An ‘all you can afford’ approach 96 57.5
Competitive Competitive/absolute The budget is set in line with the market share 8 4.8
Competitive/relative The budget is set in line with that of your closest rival 19 11.4
Sales % of last year’s sales Set % of sales from the previous financial year 33 19.8
% of anticipated sales Set % of the firm’s anticipated sales 28 16.8
Unit sales A fixed % of the unit price multiplied by the projected 9 5.4
sales volume
Measurement Marginal analysis Use of mathematical techniques to estimate the point 20 12.0
above which the incremental expenditures in marketing
are exceeded by the marginal revenue they generate
Objective/task Objective/task Spending is in accordance with what is required to 70 41.9
meet the marketing objectives
Note: Percentages do not add up to 100% because multiple methods could be used in making budget decisions
process, or some mixture of the two, such as bottom-up/top-down or top-down/bottom-

up (Piercy 1987a, 1987b; Belch & Belch 2001).
Overall, the literature suggests that the budgeting categories displayed in Table 1 get
increasingly interactive and bottom-up in process from the judgement methods to the
objective and task category (Prendergast et al. 2006). The conventional wisdom among
prior researchers has been that the more interactions with and inputs from the profession-
als actually involved in the planning and implementation of marketing strategies, the more
‘sophisticated’ the budget decisions are (Sheth & Sisodia 2002; Prendergast et al. 2006;
West & Prendergast 2009). Such a convention has emerged from the implicit assumption
among researchers that more data-driven methods provide more viable approximations
of the ideal marketing budget (Lynch & Hooley 1990), while judgement-orientated or
simple heuristic budgeting methods are crude and likely to waste resources (Prendergast
et al. 2006).
The conventional reasoning that the bottom-up process is more sophisticated and,
therefore, is a better practice of budgeting does make sense to some degree, as the bottom-
up process is more participative from the marketing department’s perspective, and, as
noted by Hulbert and Pitt (1996), good planning is generally participative, whereas the
opposite, such as authoritarian withholding of information and non-participative decision
making, is considered counterproductive (Prendergast et al. 2006). However, it should be
noted that a bottom-up process by itself does not necessarily confirm sophistication in the
process. Instead, it is the degree to which the power to influence budget setting is decen-
tralised, or is shared with lower-level marketing department personnel – referred to as the
level of decentralisation in the budget process – that plays a significant role in advertising
budgeting. The link between how much the budgeting process is decentralised and the
resulting budget size and/or the profitability of a firm is, then, an empirical issue; previous
145
research has generally shown that the level of decentralisation of the budgeting process
does affect the size of the marketing budget and the profitability of a firm.
From this perspective, we use the term ‘decentralisation’ to mean that the more decen-
tralised budgeting process connotes more participation from and interaction with those
involved in the actual planning and implementation of marketing strategies; the budget-
ing categories shown in Table 1 are conceptualised to vary in their level of decentralisa-
tion, such that the budgeting methods get increasingly decentralised from the judgement
methods to the objective and task category. In this study, the level of decentralisation in
budget decisions is considered an important factor influencing the size of the budget, as
well as the profitability of a firm.
Hypotheses and research question
Overview
As shown in Figure 1, three groups of factors – the company, brand and the organisational
factors – were identified as influencing the overall level of decentralisation in the adver-
tising budgeting, which, in turn, is predicted to influence the budget size in terms of the
overall marketing budget size, and the advertising-to-promotion ratio.
Factors influencing the level of decentralisation in budgeting

Company factors
Previous research indicates that the company size and the financial performance of a
company play important roles in the level of budgeting decentralisation. The relation-
ship between the company size and the budgeting methods has appeared in finance and
accounting research, which indicates that large firms tend to use more decentralised budg-
eting techniques than smaller firms (Istvan 1961; Klammer 1972; Fremgen 1973; Schall
et al. 1978). The rationale is that large firms have a greater vested interest in good practice
than smaller ones, thus are more likely to rely on market-driven information (Gilligan
Figure 1: Overview of relationships
Company factors
• Company size
• Profitability
Brand factors Degree of Budget size

• Brand type
decentralisation • Total marketing budget
• Brand equity
in budgeting • Adv-prom. ratio
• Brand price
Organisational factors
• Marketing department
• Finance department
146
1976; Patti & Blasko 1981; Hooley & Lynch 1985). Also, the firm size would create
economies of scale to pay out investments in budgeting expertise, so that larger companies
make considerable capital expenditures over time, and have more resources than smaller
firms to derive a well-formalised marketing budgeting process (Kim 1982). In addition,
it seems that accumulated experience, both over time and across divisions, would provide
a larger pool for accumulating best practices of budget decisions that would help firms
manage to be profitable. This is because profitability, if sustained in a large company, must
come from learning rather than chance, since the range of experience would provide too
much variability to support chance performance advantages.
Such a positive correlation between the company size or profitability, and the use of
decentralised budgeting methods has been found in a number of studies in advertising
and marketing research (Lynch & Hooley 1990; Hung & West 1991; Sangster 1993; Pike
1996; Graham & Harvey 2001). A similar supposition that companies that perform better
will use more decentralised methods than will poor performers has also been confirmed
in several studies in the advertising and marketing literature (Copeland 1979; Kim 1982;
Hooley & Lynch 1985; Hung & West 1991). Based on the discussion above, the follow-
ing hypotheses are formulated.
H1: Larger companies will use more decentralised budgeting methods than smaller
companies.
H2: Profitable companies will use more decentralised budgeting methods than the
companies experiencing losses.
Brand factors
Previous research suggests that the type of brand, the brand equity and the relative price
of the brand influence the level of decentralisation in the budgeting. With respect to the
type of brand, the literature suggests that an individual brand for a single product would
involve more decentralised budgeting methods than a corporate/family brand. A corpo-
rate/family brand is the brand of a company, or a brand used in more than one product
category (Aaker 1996; Keller 2003). In general, the marketing of individual product
brands tends to be more complex than the marketing of a corporate/family brand because
all the supporting marketing activities for an individual brand need to be customised to
meet the needs of various customer groups, whereas the marketing of a corporate/family
brand generally focuses on building one overarching image of a firm (Hatch & Schultz
2001; Keller 2003). This indicates that the marketing budgeting process of single product
brands requires more decentralised methods, as it tends to involve a greater number of
marketing tactics than do corporate/family brands.
With respect to the brand equity and the price of the brand, it is proposed that the
brands with high brand equity and high relative price are likely to use more decentralised
budgeting methods than low-equity and low-price counterparts. Brand equity is associated
with the valuation of a brand as an asset (Aaker 1991; Keller 1993, 2003). Thus, strong
brand equity indicates that the brand name brings in added value to the product, making
147
the firm likely to be financially successful (Ailawadi et al. 2003). Strong brands command
a high price premium (Park & Srinivasan 1994; Dias & Ryals 2002), and consumers loyal
to a strong brand require a larger price differential in favour of the rival brand before they
switch away from their favourite brands (Keller 1993; Agrawal 1996). As noted above, the
companies performing successfully in the market tend to employ decentralised budgeting
processes. Thus, the brands with high brand equity, as well as relatively high price, are more
likely to involve more decentralised budgeting methods than are weaker brands.
H3: Individual product brands will use more decentralised budgeting methods than
corporate/family brands.
H4: Brands with high brand equity will use more decentralised budgeting methods
than brands with low brand equity.
H5: Brands that charge higher prices relative to the industry average will use more
decentralised budgeting methods than the lower-priced brand counterparts.
Advertising and marketing research identifies several organisational factors that influence
the level of decentralisation in the budgeting. Companies where the marketing depart-
ment plays a significant role in the budgeting process tend to use more decentralised
budgeting methods than those companies not involving their marketing department
in the budgeting process (Piercy 1987a; Prendergast et al. 2006). When the marketing
department is involved in the budgeting process, this involves a number of lower-level
marketing employees, each with specialised knowledge in specific areas of marketing
budgeting, and thus it is more likely to be based on the bottom-up process. The bottom-
up approach, involving specialised marketers, would also help to create a more accurate
picture of how much each marketing plan needs in order to function effectively, because
their budgeting plan would consider various marketing factors such as the assessment
of consumers’ wants, examination of competitors’ capabilities, and the engagement of all
elements of the company in marketing activities. Thus, it is generally known that a budg-
eting plan giving more power to the marketing department, and thus more decentralised,
is more accurate (Hulbert et al. 1980; Piercy 1987a; Prendergast et al. 2006).
In contrast to the role of the marketing department, it is proposed that, the greater the
influence of the finance department, the more the budgeting process is likely to be based on
the top-down process (Piercy 1987b). The senior management of the finance department
tends to offer a portfolio of budgeting frameworks for a top-down approach with, compara-
bly, less marketing consideration (Piercy 1987a). Researchers have believed that marketing
budgets driven by top management, based on such tools as parity with competitors and/or
instrumentalism such as adjustment to the previous year’s spending levels, are less interactive
in the process (Sheth & Sisodia 2002) and do not provide managers with a comprehen-
sive strategy to guide the process (Low & Mohr 1999). Based on the discussion about the
organisational factors presented above, the following hypotheses are developed.
148
H6: The stronger the influence of the marketing department on budgeting deci-
sions, the more decentralised the budgeting methods will be.
H7: The stronger the influence of the finance department on budgeting decisions,
the less decentralised the budgeting methods will be.
The level of decentralisation in budgeting and budget size

There has been evidence that the level of decentralisation in budgeting methods is posi-
tively related to the size of the firm’s budget (Schall et al. 1978). In the context of mar-
keting budgeting, Piercy (1987a, 1987b) claimed that different budgeting processes are
associated with different budget outcomes. The bottom-up process tends to lead to larger
budgets, while the top-down process tends to lead to smaller budgets. Piercy (1987b)
claims that top-down budgeting causes lower marketing expenditures and ultimately
depresses profitability. Alternatively, when profitability is low, there may be a tendency
to introduce stricter top-down management and the influence of the finance department
over the budgets increases, which tends to introduce the logic of cutting marketing expen-
ditures, further depressing volume and profitability. In the study of UK manufacturing
firms, Piercy (1987b) reported that 48% of the firms involved in a bottom-up budgeting
process set high overall marketing budgets, whereas 62% of the firms involved in a top-
down budgeting process set low overall marketing budgets.
The degree of decentralisation in budgeting also influences how firms allocate the total
marketing budget to advertising and promotion. In so far as companies are keen to create
immediate sales improvements, they will be more likely to focus on sales promotion at the
expense of advertising, and also more likely to be susceptible to market sales fluctuations
(Low & Mohr 2000). This indicates that their budgeting plan will be less decentralised,
because their marketing spending would be subject to spontaneous changes when the
market is fluctuating. Therefore, it is predicted that the advertising-to-promotion ratio
gets lower as the budgeting methods get less decentralised. Based on the discussion above,
the following hypotheses are proposed.
H8: The more decentralised the budgeting methods, the greater the overall market-
ing budget size will be.
H9: The more decentralised the budgeting methods, the greater the overall adver-
tising-to-promotion ratio will be.
Product category and budgeting methods

In addition to the hypotheses proposed above, this study examines how product factors are
related to budget practices, as well as the relationships between the product categories and
the likelihood of success of the brands, by exploring the following two research questions.
RQ1: How are the product categories related to the budgeting methods?
149
RQ2: How is the profitability of the brands related to the budgeting methods?
Method
Sample and procedure

A web-based survey was conducted. The potential participants of the survey were recruited
from the panel members of ZoomPanelTM, managed by a professional market research firm,
MarketTools, Inc. ZoomPanelTM is an opt-in, online panel dedicated solely to supporting
market research. The panel consists of more than two million active members across the
US. Participating panellists earn points through a rewards programme, and are also given
chances to enter monthly and annual sweepstakes administered by MarketTools, Inc.
The sampling procedure was randomised at two stages. First, randomisation occurred
when the sample was pulled from the panellist database. This eliminated any sort of bias
based on a panellist’s ‘position’ on the panel. At the second stage, the sample was again ran-
domised prior to survey deployment, to ensure that there was no bias when the lists were
divided or when the surveys were deployed. In May 2009, a total of 500 invitations were
sent to the senior marketing and brand managers in the US, in a cross-section of large and
medium-sized companies. A screening question was used at the beginning of the survey, so
that the responses were collected only from those managers responsible for and knowledge-
able about making their firms’ marketing budget decisions. In particular, respondents were
asked to answer the following two questions: ‘Are you responsible for the budgeting practice
of your firm?’ and ‘Are you knowledgeable about the budgeting methods used in your com-
pany?’ Only those respondents who answered ‘Yes’ to both questions were included in the
data analysis. A total of 169 complete and usable responses was collected; thus the response
rate was 33.8%. The profile of the participants is presented in the ‘Findings’ section.
Personalised email invitations to participate in the survey were sent, along with a brief
explanation of the study and its importance; a clickable link to access the questionnaire was
provided. Invitations were sent individually, and each was personalised by the first name of
each potential participant in order to increase the response rate (Heerwegh 2005). Upon
clicking on the link to participate in the study, participants were taken to a page that served
as the consent form. In the consent form, the researcher informed the subjects that their
participation was strictly voluntary and that they could choose not to participate. The sub-
jects’ decision whether or not to participate in this study did not affect their current or future
relations with the professional survey research firm, if there were any such relations. Clicking
on ‘Next’ at the bottom of the page indicated their consent to participate. The questions for
the study were spread across a series of web pages for ease of presentation, with participants
clicking on ‘Next’ at the bottom of each page to progress through the study.
Measures
The questionnaire consisted of six sections. In the first section, participants were asked
to indicate their official title, gender, age and highest degree earned. Also, participants
150
indicated the total number of years involved in their firm’s marketing budget decisions.
Then, in the second section, participants specified in an open-ended question the brand
they were most responsible for in setting the marketing budget. This question was asked
to avoid any confusion in answering questions when the participant was involved in more
than one brand’s marketing budgeting (e.g. Low & Mohr 2000). Participants were also
asked to indicate the type of the brand by indicating either a corporate/family brand
or an individual brand. In the third section, the marketing budget size was recorded.
Participants were asked to indicate the total marketing budget size in a dollar amount, and
then to provide their budget allocation between advertising and non-advertising promo-
tion in terms of percentages that added up to 100%.
The fourth section of the questionnaire contained questions regarding the choice of the
specific budgeting methods. Respondents were given a list of the nine budgeting methods
displayed in Table 1 with the corresponding descriptions, and were asked to check all
that applied, as more than one budgeting method could be used in budget setting (Piercy
1987b; Hung & West 1991; Prendergast et al. 2006). Recognising the problem that a
majority of firms use more than one budgeting method, we used the same scoring system
that was used by Prendergast et al. (2006) and West and Prendergast (2009) to estimate
a single index measure of the overall decentralisation score for each brand. The overall
degree of decentralisation takes into account all the budgeting methods used by each firm,
so the respondents were allowed to choose all the budgeting methods that applied to the
respondent’s firm.
The greatest benefit of using such an index score is that it allows us to perform reliable
inter-brand comparisons, based on the level of decentralisation in the budgeting process;
it also makes it easy to relate each brand’s decentralisation level to explanatory variables,
given firms’ multiple budget method choices. In particular, each budgeting method was
given a weight, ranging from 1 to 5, according to the level of decentralisation, as follows:
judgement (affordable, arbitrary): 1; competitive (relative, absolute): 2; sales (last year’s
sales, anticipated sales, unit sales): 3; marginal analysis: 4; objective and task: 5. Thus,
greater scores indicate a more decentralised budgeting method. These weights were
allocated based on the previous studies (Piercy 1987b; West & Hung 1993; Prendergast
et al. 2006). The overall budgeting decentralisation scores were then calculated by adding
all the available decentralisation scores for each brand. The reliability of the weighting
system, and the concept of adding the weightings together to come up with the overall
index score, were determined appropriate in previous research (Prendergast et al. 2006),
based on the consultation with nine judges, and the satisfactory level of the proportional
reduction in loss (PRL) values, which are comparable to Cronbach’s alpha in classifying
qualitative data (Rust & Cooil 1994). However, as noted by Prendergast et al. (2006),
the weighting system is not without limitation, as the intervals between the budgeting
categories are assumed to be equal, when, in fact, there is no such guarantee in practice.
Therefore, consistent with Prendergast et al. (2006, p. 169), we suggest the differences be
treated as ‘broadly directional rather than precise’.
In the fifth section, a series of brand factors was measured. Consistent with Keller’s
(1993, 2003) conceptualisation, perceived brand equity was measured by the degree of
151
brand awareness and by three dimensions of brand image (uniqueness, favourableness

and the strength of association), all on a 5-point scale, as follows: ‘Relative to the average
industry brand’, ‘How high is the brand awareness level?’ (low awareness–high awareness);
‘…, your brand is regarded as …’ (not distinguishable–unique); ‘…, how favourable are
customers’ perceptions of your brand?’ (unfavourable–favourable); ‘…, how strongly held
is the image of the brand in customers’ minds? (weakly held–strongly held). Participants’
responses to the four items were tallied to generate the index score of brand equity.
Meanwhile, the relative price was measured by the following 5-point scale: ‘Relative to
the industry average, your brand’s retail selling price is …’ (very low–very high) (Low &
Mohr 2000).
In the sixth section, the influence of the marketing and the finance departments were
measured by asking ‘How much influence does the … have over the marketing budget?’ (a
great deal–quite a bit–some–none) on a 4-point scale, respectively (Piercy 1987b). Finally,
each participant was asked to indicate whether his or her company was profitable, just
breaking even, or experiencing losses last year (Prendergast et al. 2006).
Findings
Profile of respondents
The median number of employees of the companies was 20,000, with a range from 50
to 1,000,000. The participants were marketing, service or brand managers (47%); vice-
presidents or general managers, and/or marketing directors (35%); finance directors,
financial analysts or financial managers (7%); marketing assistants (5%); with the remain-
der presidents (3%) and general directors (3%). The respondents had an average of 7.8
years’ experience in making marketing budget decisions, with a range from 1 to 35 years.
The median age of the respondents was 43.7 years old, with a range from 30 to 64. The
sample comprised approximately 40% males and 60% females. A total of 91% had college
degrees or higher. Also, the average number of respondents per company was 1.04 (169
respondents from 161 companies), indicating that the influence of multiple responses
from a same company was minimal.
Usage of budgeting methods

As shown in Table 1, around half of the respondents indicated that they had experience
with the two judgement methods. A total of 46% of the respondents indicated that they
used the arbitrary method, and 58% of the respondents used the affordable method. As
described above, the two judgement methods are considered to be the least decentralised
budgeting methods. Interestingly, approximately 42% of the respondents indicated that
they had used the objective and task method – the most decentralised budgeting method.
With respect to the sales-related methods, approximately 20% of the respondents indi-
cated that they used the percentage-of-last-year’s-sales method, and 17% indicated that
they used the percentage-of-anticipated-sales method. The remaining budgeting methods
152
were used by less than 12% of the respondents. Only about 5% of the respondents indi-
cated that they had used the absolute competitive budgeting method, which appeared to
be the least frequently used budgeting method.
Company factors (H1 and H2)

It was proposed that company size (H1) and the financial performance of the company
(H2) would be positively associated with the degree of decentralisation in the budgeting.
To test these hypotheses, a series of independent sample t-tests was performed, based on
the median split of the companies. Companies with number of employees greater than
the median number – 20,000 – are considered large firms, whereas firms with fewer than
20,000 employees were operationalised as small firms. Table 2 presents the results. As
expected, the degree of decentralisation in the budgeting was significantly higher for large
companies than for small companies (Mlarge = 6.51 vs Msmall = 4.59), t = 2.64, p < 0.01.
Table 2 also shows that the companies that were profitable or breaking even last year
were associated with significantly higher levels of decentralisation than those companies
experiencing losses (Mprofitable = 5.81 vs Mloss = 3.74), t = 2.76, p < 0.01. These findings sup-
port H1 and H2.
Table 2: Comparison of budgeting decentralisation scores by company, brand

and organisational factors
Independent variables N Mean SD t/F

Company factors
Company size No. of employees ≥20,000 51 6.51 4.46 2.64**
No. of employees <20,000 114 4.59 4.26
Profitability Profitable or broken even 124 5.81 3.13 2.76**
Experienced loss 43 3.74 4.57
Brand factors
Brand type Single brand 101 5.99 4.59 2.67**
Family brand 66 4.20 4.20
Brand equity High equity 79 5.94 4.74 2.31*
Low equity 80 4.37 3.80
Relative price High price 22 7.36b 5.07 4.10*
Average price 93 5.29a 4.19
Low price 50 4.24a 4.01
Influence of marketing department High 77 6.27 5.07 2.82**
Low 87 4.39 3.37
Influence of finance department High 82 4.26 3.60 2.58*
Low 86 5.97 4.90
**p < 0.01; * p < 0.05; means with no common lowercase superscript differ at p < 0.05
153
Brand factors (H3, H4 and H5)

The degrees of decentralisation in the budgeting were examined across several brand fac-
tors, such as the type of brand, brand equity and the relative price of the brand. The results
are displayed in Table 2.
Consistent with the prediction of H3, setting marketing budgets for individual
brands involved more decentralised budgeting methods than for corporate/family brands
(Msingle = 5.99 vs Mfamily = 4.20), t = 2.67, p < 0.01. H4 predicted that more decentralised
budgeting methods would be used for brands with high brand equity than for the low-
equity brands. In order to examine this hypothesis, the budgeting decentralisation levels
were compared across the two groups of brands determined by a mean split based on the
overall brand equity scores (M = 13.43). In support of H4, more decentralised budget-
ing methods were used when the brand equity was high rather than low (Mhigh = 5.94 vs
Mlow = 4.37), t = 2.31, p < 0.05. In addition, consistent with the prediction of H5, the
level of decentralisation in the budgeting was higher when the relative price of a brand
was high rather than low (Mhigh = 7.36 vs Mmid = 5.29 vs Mlow = 4.24), F = 4.10, p < 0.05.
Organisational factors (H6 and H7)

The relationships between organisational factors and the degree of decentralisation in the
budgeting process were examined. In order to examine these hypotheses, companies were
mean split into two groups based on the influence of the marketing department and the
finance department. The results are shown in Table 2.
Consistent with H6, the overall budgeting decentralisation level was higher when
the influence of the marketing department was high rather than low (Mhigh = 6.27 vs
Mlow = 4.39), t = 2.82, p < 0.01. H7 was also supported, as evidenced by the lower budget-
ing decentralisation level when the influence of the finance department was high rather
than low (Mhigh = 4.26 vs Mlow = 5.97), t = 2.58, p < 0.05.
Budget size (H8 and H9)

The last two hypotheses addressed the relationships between the level of decentralisation
in budgeting methods, and the size of the marketing budget in terms of the total budget
size and the advertising-to-promotion ratio. A positive relationship was predicted between
the budgeting decentralisation level and both the total marketing budget size (H8) and the
advertising-to-promotion ratio (H9). In order to test these hypotheses, a mean split was
performed based on the degrees of decentralisation (M = 5.0). This mean split grouping
was then used as the independent variable in a MANOVA analysis with the two correlated
dependent variables: total marketing budget size and the advertising-to-promotion ratio.
As shown in Table 3, the multivariate effect of the decentralisation level on the two
budget size variables was significant (Wilks’ lambda = 0.90), F = 6.68, p < 0.01, with sig-
nificant univariate relationships for each of the two dependent variables. It appears that a
higher degree of budgeting decentralisation is significantly associated with a larger total
154
marketing budget size (Mhigh = $20 million vs Mlow = $2,000), F = 7.91, p < 0.01, and a
relatively larger allocation of marketing budget to advertising rather than to sales promo-
tion (Mhigh = 82.5% vs Mlow = 25.3%), F = 4.86, p < 0.05. Hence, H8 and H9 are supported.
Table 3: Effects of decentralisation scores on budget size and allocation
Grouping by Univariate effects

decentralisation scores Multivariate effects Total budget size Adv-prom. ratio
Wilks’ lambda F F (1, 123) F (1, 123)
0.90 6.68** 7.91** 4.86*
High decentralisation $20 million 82.5%
Low decentralisation $2,000 25.3%
*p < 0.05, **p < 0.01
Product categories and budgeting methods (RQ1 and RQ2)

RQ1 and RQ2 asked about the relationships between product categories and budgeting
methods, and the relationships between budgeting methods and the profitability of the
brands. Table 4 shows the distributions of the nine budgeting methods by 15 generic product
categories. The products were classified based on Advertising Age’s typical product classifica-
tion scheme. It appeared that Education Products (13.0%) were the most popular product
using the arbitrary budgeting method, followed by Sporting Goods (11.7%) and General
Service (10.4%). With respect to the affordable method, Home Supplies & Cleaners (11.5%),
Apparel and Jewellery (10.4%) and Insurance (9.4%) were the three most popular product
categories. The competitive budgeting methods were distributed too thinly across the 15
product categories examined to generate a meaningful interpretation. This is due to the small
number of brands adopting competitive methods – only eight – and 19 brands adopted the
competitive/absolute and the competitive/relative methods, respectively. Automobile seems
to dominate the percentage of sales method. In particular, Automobile accounted for 18 and
21% of the brands adopting the percentages of last year’s and the anticipated sales methods,
respectively. Meanwhile, Financial Services (25%) and Telecommunication Services (20%)
together accounted for approximately 45% of the brands using the marginal analysis method.
Finally, Financial Services (12.9%) and Automobile (12.95%) were the two most popular
product categories using the objective-and-task method.
Table 4 also shows the distributions of the nine budgeting methods by the profitability
of the brands. Overall, the two most decentralised budgeting methods – the marginal
analysis and the objective-and-task methods – were used by greater percentages of prof-
itable brands than non-profitable brands, whereas the two least decentralised budget-
ing methods – the arbitrary and the affordable methods – were associated with greater
percentages of non-profitable than profitable brands. For example, 44% of the profitable
brands used the objective-and-task method, when only 34% of the non-profitable brands
used the same budgeting method.
155
156
Table 4: Budget practices by product categories and profitability
Judgement Competitive % of sales

Last year’s Anticipated Marginal Objective/
Arbitrary Affordable Comp./abs. Comp./rel. sales sales Unit sales analysis task
n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%)
Product categories
Apparel and jewellery 8 (10.4) 10 (10.4) 1 (12.5) 1 (5.3) 3 (9.1) 2 (7.1) 0 (0.0) 0 (0.0) 2 (2.9)
Automobile 3 (3.9) 6 (6.3) 1 (12.5) 2 (10.5) 6 (18.2) 6 (21.4) 1 (11.1) 2 (10.0) 9 (12.9)
Education 10 (13.0) 7 (7.3) 1 (12.5) 1 (5.3) 1 (3.0) 3 (10.7) 0 (0.0) 0 (0.0) 3 (4.3)
Financial services 3 (3.9) 6 (6.3) 0 (0.0) 1 (5.3) 2 (6.1) 2 (7.1) 1 (11.1) 5 (25.0) 9 (12.9)
Food, beverages and candy 3 (3.9) 4 (4.2) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 2 (22.2) 2 (10.0) 6 (8.6)
General services 8 (10.4) 6 (6.3) 1 (12.5) 1 (5.3) 2 (6.1) 1 (3.6) 1 (11.1) 1 (5.0) 5 (7.1)
Home supply and cleaners 5 (6.5) 11(11.5) 1 (12.5) 2 (10.5) 2 (6.1) 2 (7.1) 0 (0.0) 0 (0.0) 1 (1.4)
Insurance 5 (6.5) 9 (9.4) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0)
Media 5 (6.5) 7 (7.3) 0 (0.0) 1 (5.3) 3 (9.1) 2 (7.1) 0 (0.0) 1 (5.0) 3 (4.3)
Medicine and remedies 2 (2.6) 4 (4.2) 0 (0.0) 2 (10.5) 1 (3.0) 2 (7.1) 2 (22.2) 1 (5.0) 7 (10.0)
Movies, video and music 5 (6.5) 7 (7.3) 1 (12.5) 2 (10.5) 2 (6.1) 2 (7.1) 0 (0.0) 1 (5.0) 3 (4.3)
Personal care 1 (1.3) 1 (1.0) 0 (0.0) 0 (0.0) 0 (0.0) 1 (3.6) 1 (11.1) 1 (5.0) 5 (7.1)
Retail 3 (3.9) 3 (3.1) 0 (0.0) 3 (15.8) 4 (12.1) 2 (7.1) 1 (11.1) 2 (10.0) 6 (8.6)
Sporting goods 9 (11.7) 9 (9.4) 0 (0.0) 1 (5.3) 2 (6.1) 2 (7.1) 0 (0.0) 0 (0.0) 2 (2.9)
Telecom, internet services 3 (3.9) 3 (3.1) 1 (12.5) 1 (5.3) 2 (6.1) 1 (3.6) 0 (0.0) 4 (20.0) 8 (11.4)
Other 4 (5.2) 3 (3.1) 1 (12.5) 1 (5.3) 3 (9.1) 1 (3.6) 0 (0.0) 0 (0.0) 1 (1.4)
Profitability
Profitable N (%)a 55 (44.4) 68 (54.8) 6 (4.8) 14 (11.3) 25 (20.2) 24 (19.4) 7 (5.6) 17 (13.7) 55 (44.4)
Non-profitable N (%)b 22 (51.2) 28 (65.1) 2 (4.7) 5 (11.6) 8 (23.5) 4 (11.8) 2 (5.9) 3 (8.8) 11 (34.9)
Total no. of brands 77 96 8 19 33 28 9 20 70
Notes: aPercentage based on the total number of profitable companies; bpercentage based on the total number of non-profitable companies; Comp./abs. (Competitive/absolute), Comp./rel. (Competitive/
relative); Total no. of companies: number of companies used each budgeting method
Post-hoc regression analysis

A post-hoc regression analysis was performed with the company, brand and organisational
factors as the predictors of the degree of decentralisation. This analysis was performed to
examine the impacts of the various independent variables together, as opposed to examin-
ing the impacts of each factor separately.
Specifically, a hierarchical regression analysis was performed with the two company
factors (size and profitability) entered as independent variables in the first block. The
three brand factors (brand type, brand equity and brand price) and the two organisational
factors (the influences of the marketing department and finance department, respectively)
were entered as independent variables in the second and third block, respectively. The
results of Model 3, shown in Table 5, confirm the hypotheses predicting the relationships
between the company, brand and organisational factors, and the degree of decentralisation
in the expected directions. The greater the company size (b = 0.40, p < 0.01) and company
profitability (b = 0.46, p < 0.05), the more decentralised the budgeting method was. Also,
individual brands (b = 0.35, p < 0.05), as opposed to family brands, and the brands with
greater brand equity (b = 0.36, p < 0.05) and higher prices (b = 0.40, p < 0.05) were associ-
ated with more decentralised budgeting methods. Finally, the greater the influence of the
marketing department (b = 0.27, p < 0.05), but the less the influence of the finance depart-
ment (b = –0.26, p < 0.10), the more decentralised the budgeting methods. Meanwhile,
the product categories shown in Table 4 were also entered as independent variables in a
separate regression model predicting the degree of decentralisation, but none of the prod-
uct categories reached the significance level at 95% and were not reported.
Table 5: Influences of the company, brand and organisational factors on the

degree of decentralisation
Model 1 Model 2 Model 3

Independent variables b (S.E.) b (S.E.) b (S.E.)
Company factors
Company size 0.37** (0.12) 0.38** (0.12) 0.40** (0.12)
Company profitability 0.49* (0.28) 0.44* (0.28) 0.46* (0.28)
Brand factors
Brand type 0.34* (0.20) 0.35* (0.20)
Brand equity 0.37* (0.33) 0.36* (0.33)
Brand price 0.39* (0.21) 0.40* (0.21)
Organisation factors
Marketing department 0.27* (0.23)
Finance department –0.26a (0.20)
Adjusted R2 0.21 0.42 0.50
F-value 15.18** 20.43** 9.95**
*p < 0.05, **p < 0.01, ap < 0.10
157
Discussion
This is one of the few studies, if any, in the past two decades to examine marketing budg-
eting practices in the US, and perhaps the first to do so in the midst of a severe economic
downturn. This study is especially valuable in that an increasing number of studies focuses
on the returns on firms’ investments in advertising (e.g. Broadbent 2008; van den Putte
2009; Eisend & Langner 2010; Taylor 2010), while researchers have recently shifted their
attention from the actual practice of setting advertising and promotion budgets.
The findings suggest that the degree of decentralisation in budgeting varies as a
function of the company, brand and organisational factors in the marketing practices.
In particular, the evidence presented in this paper indicates that companies engaged in
more decentralised budgeting processes are more likely to perform better in the market,
as indicated by the size of the company, and its profitability and brand equity. It was also
found that the marketing budgets for individual brands and the brands with relatively
high prices were associated with more decentralised budgeting methods. In terms of the
organisational factors, the influences of the marketing department and the finance depart-
ment appeared to have positive and negative influences on the degree of decentralisation
in the budgeting processes, respectively. Finally, the level of budgeting decentralisation was
positively related to the total marketing budget size, and more allocation of the marketing
budget to advertising in relation to promotion.
The overall findings are consistent with the previous research examining the relation-
ships between the budgeting methods and company, brand and organisational factors, as
well as the overall budget size and allocation (e.g. Piercy 1987a, 1987b; Hung & West
1991; Graham & Harvey 2001). In this sense, the findings of this study are important in
that they provide support for traditional budgeting approaches and managerial processes
during a severe economic downturn. When the economy slows down or is in the middle
of a severe recession – as when this study was performed – businesses tend to cut back
and are reluctant to spend (Picard 2001). During recessions, firms do not act normally,
and with respect to budget setting, this indicates an inflated power of top management
over the lower-level professionals in firms, such as marketing managers. For most busi-
nesses, this translates into pulling back on discretionary spending; typically, one of the
first line items to go is advertising and promotion (Brenner 2009). It might be that top
management believes that investments in advertising do not respond to the market as they
usually do, so the management expects less-than-ordinary returns on their advertising
investments, justifying the top management intervention to cut marketing investments.
The findings of this study, however, suggest that firms using more decentralised
budgeting methods are likely to invest more in advertising and promotion, even during
recessions, and are generally more profitable than firms using less decentralised, more top-
down approaches in budgeting. These findings suggest that firms should be cautious in
cutting back marketing investments, based on top-down approaches, simply because the
sales are expected to decrease during a recession. Instead, this study suggests that firms
see the changed economic environment as an opportunity to strengthen their business by
careful budget setting and the implementation of marketing plans (Srinivasan et al. 2005).
158
This is consistent with Tikoo and Ebrahim’s (2010) notion that investors perceive that an
increased emphasis on advertising can enable firms to stem the earnings erosion that can
potentially occur during an economic downturn.
As noted above, to our knowledge no other academic research has systematically exam-
ined the nature of a firm’s advertising and promotion budgeting methods in the middle of
severe economic downturn. Another contribution of this study to the literature on advertising
budgeting is that we introduce the construct of the level of decentralised, rather than using
the term ‘sophistication’, as in the previous research. This approach resolves much confusion
that has been caused by implicitly assuming that more bottom-up budgeting methods are
more sophisticated and are better practices, when, in fact, it is the decentralised level, or the
degree to which the power to influence the final budget decisions is devolved to the lower-
level professionals in an organisational structure, that plays an important role in budgeting.
Despite the merits discussed above, the findings presented in this study are limited
in several ways. First, it examined company, brand and organisational factors. However,
there are additional factors worth investigating. As indicated by Hung and West (1991),
the selection of the advertising budgeting method is a complex issue not easily distilled
to a few factors. For example, previous research indicates that such factors as product life
cycle, purchase frequency, market share and the number of customers can affect budget-
ing methods (e.g. Lilien & Little 1976). Future researchers are encouraged to incorporate
additional factors not examined in this study, in their study of marketing budgeting.
Second, company profitability and brand equity was measured, based on self-reported
measures. This could pose a problem, as many managers have inflated the profitability
and the perceptions of their brands. Unfortunately, due to the restriction in the data, the
current study could not examine other objective measures of company and brand perfor-
mance such as market share. This limitation should be taken into consideration when the
findings are interpreted. Future researchers are encouraged to adopt alternative measures
of brand equity and cross-validate the findings reported in this study. Third, we claim that
the findings of the study reflect the recession period, as the survey was performed during
a recessive year. Although the hypotheses drawn from the previous research performed
during non-recessive periods were confirmed, the current study bears a limitation in that
it could not make a direct comparison to the findings of the surveys performed during the
non-recessive years. Such a limitation was due to the lack of recent studies on budgeting
methods, and the differences in the sample and methodology employed in the analyses.
For example, the most recent studies on advertising budgeting methods are Prendergast
et al. (2006) and West and Prendergast (2009), which were performed in China and
the UK, respectively; thus, it was difficult to make direct comparisons to the results of
our study, which was performed in the US market. Future researchers are encouraged
to replicate the current study during a normal economic condition in the US. Lastly, as
described earlier, the method of weighting individual budgeting techniques by the degree
of decentralisation is not perfect. Although the method introduced by Prendergast et al.
(2006), and adopted in this study, is the best available method of deriving a single index
score of the overall budgeting decentralisation, a more accurate method of weighting the
decentralisation level for each budgeting technique should be developed.
159
Acknowledgement
This article was supported by the Hankuk University of Foreign Studies Research Fund
of 2012.
References
Aaker, D.A. (1991) Managing Brand Equity. New York: The Free Press.
Aaker, D.A. (1996) Building Strong Brands. New York: The Free Press.
Aaker, D.A., Batra, R. & Myers, J.G. (1992) Advertising Management. Englewood Cliffs, NJ:
Prentice-Hall, Inc.
Agrawal, D. (1996) Effects of brand loyalty on advertising and trade promotions: a game
theoretic analysis with empirical evidence. Marketing Science, 15(1), 86–108.
Ailawadi, K.L., Lehmann, D.R. & Neslin, S.A. (2003) Revenue premium as an outcome
measure of brand equity. Journal of Marketing, 67(4), 1–17.
Barwise, P. (ed.) (1999) Advertising in a Recession: The Benefits of Investing for the Long Run.
Henley-on-Thames, UK: NTC Publications.
Belch, G.E. & Belch, M.A. (2001) Advertising and Promotion: An Integrated Marketing
Communications Perspective. New York: McGraw-Hill.
Bhargava, R. (2011) Spend shift: how the post-crisis values revolution is changing the way we
buy, sell and live. International Journal of Advertising, 30(1), 191–192.
Brenner, B.K. (2009) The value of increasing investment in employee benefits during an
economic downturn. Journal of Financial Service Professionals, January, 29–31.
Broadbent, T. (2008) Does advertising grow markets? International Journal of Advertising, 27(5),
745–770.
Copeland, T.E. (1979) Liquidity changes following stock splits. Journal of Finance, 34(1), 115–141.
Dias, S. & Ryals, L. (2002) Options theory and options thinking in valuing returns on brand
investments and brand extensions. Journal of Product and Brand Management, 11(2), 115–128.
Eisend, M. & Langner, T. (2010) Immediate and delayed effects of celebrity attractiveness and
expertise. International Journal of Advertising, 29(4), 527–546.
Fremgen, J.A. (1973) Capital budgeting practices: a survey. Management Accounting, 54 (11), 19–25.
Gilligan, C. (1976) How advertising budgets are arrived at. Advertising Quarterly, 46, 32–35.
Graham, J.R. & Harvey, C.R. (2001) The theory and practice of corporate finance: evidence
from the field. Journal of Financial Economics, 60(2–3), 187–243.
Hatch, M.J. & Schultz, M. (2001) Are the strategic stars aligned for your corporate brand?
Harvard Business Review, 79(2), 128–134.
Heerwegh, D. (2005) Effects of personal salutations in e-mail invitations to participate in a
web survey. Public Opinion Quarterly, 69(4), 588–598.
Hooley, G.J. & Lynch, J.E. (1985) How UK advertisers set budgets. International Journal of
Hulbert, J. & Pitt (1996) Exit left-centre stage? The future of functional marketing. European
Management Journal, 14(1), 47–60.
Hulbert, J.M., Brandt, W.K. & Richers, R. (1980) Marketing planning in the multinational
subsidiary: practices and problems. Journal of Marketing, 44(3), 7–15.
Hung, C.L. & West, D.C. (1991) Advertising budgeting methods in Canada, the UK and the
US. International Journal of Advertising, 10(3), 239–250.
Istvan, D.F. (1961) The economic evaluation of capital expenditures. Journal of Business, 34(1), 45–51.
Keller, K.L. (1993) Conceptualizing, measuring, and managing customer-based brand equity.
Journal of Marketing, 57(1), 1–22.
160
Keller, K.L. (2003) Strategic Brand Management: Building, Measuring, and Managing Brand
Equity. Upper Saddle River, NJ: Prentice Hall.
Kim, S.H. (1982) An empirical study on the relationship between capital budgeting practices
and earnings performance. Engineering Economist, 27(3), 185–196.
Klammer, T. (1972) Empirical evidence of the adoption of sophisticated capital budgeting
techniques. Journal of Business, 45(3), 387–397.
Lilien, G.L. & Little, J.D.C. (1976) The advisor project: a study of industrial marketing
budgets. Sloan Management Review, 17(3), 17–31.
Low, G.S. & Mohr, J.J. (1999) Setting advertising and promotion budgets in multi-brand
companies. Journal of Advertising Research, 39(1), 67–79.
Low, G.S. & Mohr, J.J. (2000) Advertising vs sales promotion: a brand management
perspective. Journal of Product & Brand Management, 9(6), 389–414.
Lynch, J.E. & Hooley, G.J. (1990) Increasing sophistication in advertising budget setting.
Macleod, C. (2009) Global economy and adspend prospects. International Journal of
Advertising, 28(1), 187–189.
Park, C.S. & Srinivasan, V.S. (1994) A survey-based method for measuring and understanding
brand equity and its extendibility. Journal of Marketing Research, 31(May), 271–288.
Patti, C.H. & Blasko, V.J. (1981) Budgeting practices of big advertisers. Journal of Advertising
Research, 21(6), 23–29.
Picard, R.G. (2001) Effects of recessions on advertising expenditures: an exploratory study of
economic downturns in nine developed nations. Journal of Media Economics, 14(1), 1–14.
Piercy, N.F. (1987a) Advertising budgeting: process and structure as explanatory variables.
Piercy, N.F. (1987b) The marketing budgeting process: marketing management implications.
Journal of Marketing, 51(October), 45–59.
Pike, R. (1996) A longitudinal survey on capital budgeting practices. Journal of Business Finance
& Accounting, 23(1), 79–92.
Prendergast, G., West, D. & Shi, Y. (2006) Advertising budgeting methods and processes in
China. Journal of Advertising, 35(3), 165–176.
Rust, R.T. & Cooil, B. (1994) Reliability measure for qualitative data: theory and implications.
Sangster, A. (1993) Capital investment appraisal techniques: a survey of current usage. Journal
of Business Finance & Accounting, 20(3), 307–332.
Schall, L.D., Sundem, G.L. & Geijsbeek, W.R. (1978) Survey and analysis of capital budgeting
methods. Journal of Finance, 33(1), 281–287.
Sheth, J.N. & Sisodia, R.S. (2002) Marketing productivity: issues and analysis. Journal of
Business Research, 55(5), 349–362.
Srinivasan, R., Rangaswamy, A. & Lilien, G.L. (2005) Turning advertising into advantage:
does proactive marketing during a recession pay off ? International Journal of Research in
Marketing, 22, 109–125.
Taylor, C.R. (2010) Measuring return on investment from advertising: ‘holy grail’ or necessary
tool? International Journal of Advertising, 29(3), 345–348.
Tikoo, S. & Ebrahim, A. (2010) Financial markets and marketing the tradeoff between R&D
and advertising during an economic downturn. Journal of Advertising Research, March, 50–55.
Van den Putte, B. (2009) What matters most in advertising campaigns? The relative effect of
media expenditure and message content strategy. International Journal of Advertising, 28(4),
669–690.
West, D. & Prendergast, G.P. (2009) Advertising and promotions budgeting and the role of
risk. European Journal of Marketing, 43(11/12), 1457–1476.
161
West, D.C. & Hung, C.L. (1993) The organizational budgeting processes of top advertisers in
Canada, the UK and the USA. Journal of Euromarketing, 2(3), 7–22.
Willis, J. & Willis, D.B. (1993) New Directions In Media Management. Boston, MA: Allyn &
Bacon.
Zdravkovic, S. & Till, B.D. (2012) Enhancing brand image via sponsorship: strength of
association effects. International Journal of Advertising, 31(1), 113–132.
About the authors

Yunjae Cheong is an assistant professor in the School of Journalism & Mass
Communication at Hankuk University of Foreign Studies. She previously was an assist-
ant professor in the Department of Advertising and Public Relations at the University of
Alabama. Her research interests focus on advertising media planning, evaluating advertis-
ing media spending efficiency, and the assessment of advertising agencies from various
perspectives. Her work has appeared in such journals as the Journal of Advertising, Journal
of Advertising Research and the International Journal of Advertising, among others. She
would like to acknowledge that this article was supported by the Hankuk University of
Foreign Studies Research Fund of 2012.
Kihan Kim (PhD, The University of Texas at Austin) is an associate professor in the
Department of Physical Education at Seoul National University (SNU). He holds a PhD
from the Department of Advertising at The University of Texas at Austin, an MA from
the School of Journalism at The University of Missouri-Columbia, and a BA from the
Department of Physical Education at SNU. His research interests include sports com-
munication and media, sports advertising and public relations, health communication and
sports marketing communications. He previously was an assistant professor in the School
of Journalism at Middle Tennessee State University, and in the School of Communication
at University of Hartford. His work has appeared in such journals as the Journal of
Advertising Research, Journal of Sport Management, Journalism & Mass Communication
Quarterly and the International Journal of Advertising, among others.
Hyuksoo Kim (PhD, The University of Alabama) is an assistant professor in the
School of Journalism & Mass Communication at St. Bonaventure University. He holds
a PhD from the College of Communication and Information Sciences at The University
of Alabama, an MA from the Department of Marketing Communication at Emerson
College, and a BA from the Department of Communication Studies at McMaster
University. His main research interest lies in consumer behaviour/psychology with a focus
on their use of digital media. His work has been published in various journals including
Journal of Advertising Research, CyberPsychology, Behavior & Social Networking, Journal of
Marketing Communications and Computers in Human Behavior.
Address correspondence to: Kihan Kim, Associate Professor, Global Sport Management,
Department of Physical Education, Seoul National University, 71-410, 599 Gwanak-ro,
Gwanak-gu, Seoul, 151-742, Korea.
Email: kihan@snu.ac.kr
162
Book Reviews
Edited by Stephanie O’Donohoe
The University of Edinburgh
Spending Advertising Money in the planning dynamics. The overview of more

Digital Age – How to Navigate the traditional channels and the principles of
Media Flow media strategy and planning are, I believe,
Hamish Pringle and Jim Marshall (2012) thorough, robust and insightful, and my
London and Philadelphia, PA: Kogan industry experience suggests that they
Page, £29.99 offer the right direction here.
ISBN 978-0749463052 The depth of the authors’ experience in
DOI: 10.2501/IJA-32-1-163-164 media is evident in their writing. For any-
one seeking insight into media dynamics
The current media landscape moves at and the process of media planning, who
such a pace, fuelled by digital technology does not possess years of hard-earned
and media, that one could suggest that any experience, I believe this book is well
attempt to document this and its dynam- worth buying and reading.
ics would almost certainly be doomed to More seasoned media practitioners,
failure. in particular those who focus on digital,
At the very least it represents a monu- however, may prefer to wait for a deeper
mental challenge and one not to be taken and fresher attempt to document what
lightly. So with this in mind the first thing perhaps is the undocumentable world of
I must do is applaud Hamish Pringle and digital media in the digital age. For me the
Jim Marshall for at least attempting to do book does not provide a particularly good
so, and for doing so in such a thorough and examination of specifically digital media,
detailed manner. either in terms of their component chan-
But this then obviously begs the ques- nels, or more importantly, their interaction
tion: have they succeeded in this task, or with more traditional media channels and
is their effort already consigned to failure the real world context in which they exist.
and out of date? And more importantly, Therefore on the NO side, sadly for me it
should readers purchase this book? doesn’t work as well as it could or should as
In short I think the answer to both an authoritative digital media text.
the above questions is yes and no in equal That became the key issue for me in
measure. Before you accuse me of sitting reviewing this book. If it did not present
on the fence in the most spectacular way, itself as a guide to media in the digital
let me articulate on what I mean. age by way of its very title, but simply
In support of the YES answers, the positioned itself as a contemporary guide
book is extremely thorough in its exami- to media (including digital), then I think
nation of media agencies, channels and it would be more convincing; perhaps the

© 2013 Advertising Association 163
Published by Warc, www.warc.com
key message here is not to judge the book familiar strange – to interrogate assump-
by its title! tions and taken-for-granted practices.
To finish where I started, however, Other books have examined advertising
above all, the authors are to be applauded through an anthropological lens, offering
for their courage in attempting to docu- insightful accounts of the everyday work-
ment the complex media world of the 21st ings of agencies in Japan (Moeran 1996),
century. Mr Pringle and Mr Marshall, I Trinidad (Miller 1997), India (Mazzarella
salute you both. 2003), Sri Lanka (Kemper 2001) and across
a range of cultural settings (Malefyt and
James Harris Moeran 2003). This book’s account of gen-
Managing Director of Digital Services eral agency structure and practice may
IPG Mediabrands – EMEA & not offer a great deal to readers who are
G14 markets well-versed in this literature or steeped
in advertising practice; indeed, the litera-
ture on advertising practice seems a little
Advertising and Anthropology: light, with notable absences including clas-
Ethnographic Practice and Cultural sic contributions by Alvesson (1994, 2001),
Perspectives Hackley (2000), and Nixon (2003). The
Timothy De Waal Malefyt and Robert J. chapter on advertising emotions could also
Morais (2012) have been developed to incorporate insights
London: Berg, £19.99 from literature on behavioural economics,
ISBN 978-0857852021 consumer tribes and brand communities.
DOI: 10.2501/IJA-32-1-164-165 Occasionally, the authors seem to take
certain practices for granted rather than
For advertising practitioners, or market- acknowledging their cultural specificity; for
ing academics interested in advertising example, they seem to assume that focus
agency practice, reading Advertising and groups are undertaken in dedicated research
Anthropology might feel a little like eaves- facilities with one-way mirrors. This may
dropping; this book is written by anthro- well be the norm in America, but in the UK
pologists, primarily for anthropologists. It at least focus groups often take place in the
is not primarily an academic text, however; living rooms of ordinary homes.
between them the authors have spent four These quibbles aside, Malefyt and
decades working in advertising agencies; as Morais offer some interesting observations
Brian Moeran puts it in his elegant pref- about the goals, roles and practices involved
ace, this experience has transformed them in client meetings, and the use of ritual in
from participant observers to observant response to liminal phases of agency work.
participants. The aim of the authors is to Where the book really comes into its own,
encourage other anthropologists to con- however, is in its treatment of research.
sider agency careers. The core of the book Although digital research (including net-
lies in nine chapters, five devoted to under- nography) receives relatively little atten-
standing advertising agencies and four to tion, Advertising and Anthropology offers a
applying anthropology in that context. fascinating account, with many examples,
A key role of anthropology is to make the of the range of ethnographic and other
164
Book Reviews
methods that anthropologists can draw remind us of the value of thoughtful

upon to generate advertising insights. The research, anthropological or otherwise,
authors make a strong case for the value of to the development of effective adver-
an anthropological perspective in agency tising and constructive agency-client
discussions and decision-making. Their relationships.
observations of clients, colleagues and con-
sumers also lead them to argue for greater References
dialogue between psychological and cul- Alvesson, M. (1994) Talking in organisations:
tural analyses of consumer behaviour and managing identity and impressions in an
advertising response. The arguments and advertising agency. Organisation Studies,
15(4), 535–563.
examples offered in relation to these points
Alvesson, M. (2001) Knowledge work:
are likely to be particularly appealing to ambiguity, image and identity. Human
account planners and practitioners respon- Relations, 54(7), 863–886.
sible for consumer insight, as well as Hackley, C.E. (2000) Silent running: tacit,
academics interested in the practice of discursive and psychological aspects of
advertising research. management in a top UK advertising
agency. British Journal of Management, 11,
The book concludes by reiterating the 239–254.
authors’ call for more anthropologists to Kemper, S. (2001) Buying and Believing:
work in the advertising industry. Echoing Sri Lankan Advertising and Consumers
arguments made as the discipline of plan- in a Transnational World. Chicago:
ning emerged in the UK from the 1960s, University of Chicago Press.
Malefyt, T. & Moeran, B. (2003) Advertising
the authors also argue for deeper integra-
Cultures. New York: Berg.
tion of anthropologists into the fabric of Mazzarrella, W. (2003) Shovelling Smoke:
agencies, so that they would act Advertising and Globalization in
Contemporary India. Durham, NC: Duke
…not so much as exotic outsid- University Press.
ers who work for corporations, but Miller, D. (1997) Capitalism: an Ethnographic
Approach. Oxford: Berg.
rather as insiders who understand Moeran, B. (1996) A Japanese Advertising
the business goals of an organisa- Agency: an Anthropology of Media and
tion from within and who contribute Markets. London: Curzon.
actively to the crafting and manage- Nixon, S. (2003) Advertising Cultures.
ment of business strategies. (p. 150) London: Sage Publications.
Clearly, the authors’ vision would serve Stephanie O’Donohoe

anthropologists well. It also serves to University of Edinburgh Business School
165
Ex libris
Winston Fletcher
DOI: 10.2501/IJA-32-1-166-168
Reprinted from International Journal of Advertising, 25(1), 2006.
Book Review Editor’s note: Nobody writes advertising blockbusters

This Ex Libris, originally published in 2006, any more. That’s not to say authors don’t
is reprinted here in tribute to its author, try to write blockbusters. They just don’t
Winston Fletcher, who died aged 76 in succeed. During the last 50 years there
September 2012. His passing has been greatly have been only four really huge sellers –
mourned and his contribution within and ‘Worldwide No. 1!’ as the airport book
beyond the British advertising industry has jackets like to claim in gold embossing
been truly remarkable. He set up two adver- – and they were all published during the
tising agencies, was a founder Chairman seven years 1957–1963. Those were the
of WARC, and served as President of the bubbling years when advertising began
Institute of Practitioners in Advertising (IPA), to boom again after the privations of the
Chairman of the Advertising Association Second World War. They were also the
and Chairman of the Advertising Standards years when commercial television first
Authority Board of Finance. Despite these found its feet as the greatest consumer
enormous demands on his time, he wrote hun- goods sales medium ever discovered. And,
dreds of papers and more than a dozen books as a consequence of those two historical
on topics ranging from the history of advertis- factors, they were years when the pub-
ing1 to making meetings fun, the work–life lic, politicians and social commentators
balance, and the story of Barnardo’s (one of grew increasingly worried about adver-
several charities which also benefited from his tising’s so-called inexorable power, and
involvement). He was President of the Royal feared that admen (there were very few
Institution for ten years, and his services to adwomen then) were going to conquer
the creative industries were rewarded with a the world. (How we wished …)
CBE in 2010. The four blockbusters were:
Winston Fletcher’s Ex Libris refers to
four classic books about advertising published 1. Vance Packard’s The Hidden Persuaders
between 1957–63; his own writing will (1957)
surely be read – and be resonant – for many 2. Martin Mayer’s Madison Avenue USA
decades to come. (1958)
3. Rosser Reeves’ Reality in Advertising
(1961)
1
Powers of Persuasion: The Inside Story of British 4. David Ogilvy’s Confessions of an
Advertising; reviewed by Tim Ambler (2008) in
International Journal of Advertising, 27(4), 668–670. Advertising Man (1963).
166
Book Reviews
All four were indeed global bestsellers, advertising, its progress was interrupted
at least three of them going into numerous by the war. More significantly, the sector
editions and countless languages, and sell- of market research we now call ‘qualitative’
ing well over a million copies. They were all was completely new, and was generally
written and published in New York, then called ‘motivational research’. This name
rapidly emerging as the world’s advertis- was deliberately chosen by those who car-
ing capital (much as London would like to ried it out, and were trying to sell it, to
think otherwise). And the first and last of imply that they could uncover consumers’
them remain extremely influential to this deepest psychological motivations, which
day. No recent advertising book has had would provide marketing people with the
anything remotely approaching their suc- ‘hidden’ power to ‘persuade’ people to do
cess or their influence. anything they wanted them to do. This
is twaddle of course, we know that now.
*** But it was symptomatic of a strand that
Even now, if you ask anyone outside the runs through the four books, and helps to
business if they have ever read a book explain their massive popularity.
about advertising, there is a fair chance Each of the authors believed, as did
they will name The Hidden Persuaders. Its almost everyone at that time, that there
very title has entered the language as a was some kind of Golden Key to effective
common phrase. And it is a mark of the advertising, and if you discovered what
book’s longevity that the idea the title it was you would be able to unlock the
encapsulates still dominates the thinking secrets of sales success. For Vance Packard
of those who are hostile to advertising. and The Hidden Persuaders the Golden
Whether they use the exact words or Key was motivational research. For Martin
not, advertising’s antagonists continue to Mayer, who was a good deal more subtle
believe that advertising is a means of ‘hid- – Packard was interested only in writing a
den persuasion’ – truly a paradoxical idea, money-making potboiler, and boy was he
since ads go out of their way not to be successful! – the Golden Key was added
hidden, but to have impact, to be publicly value. Mayer saw, correctly, that advertis-
seen and to be remembered. ing could and does add value to brands,
The paradox is in some measure which was then a radical new concept. He
explained as soon as you open the book. arrived at the right answer, but as he was
For although everybody thinks The Hidden a journalist rather than an advertising man
Persuaders is about advertising, it is not he did not quite understand the process by
really about advertising at all. It is about which value is added. Fully understanding
market research. And market research is, the process is not easy, even today.
relatively, hidden. It is usually carried out For Rosser Reeves, chief honcho at
privately, and its commercial findings – Ted Bates, which was then an immensely
and even many of its political ones – are successful global agency, the Golden Key
usually restricted and confidential. This was the Unique Selling Proposition (USP)
was even truer in the 1950s and 1960s. – another great phrase still widely used
Market research was then in its infancy. (and frequently misused). The USP phi-
It started only in the late 1930s and, like losophy held sway over the development
167
of most consumer campaigns until the the best single book about the ways adver-
1980s, when marketers began to focus on tising works ever to have been published.
brand images (and added value) instead of
concentrating exclusively on product spec- ***
ifications. But the USP approach to adver- In the decades since Confessions a host of
tising still has huge merit, it is just that it other advertising books have been writ-
applies only to certain types of consumer ten – quite a few by me! – but none has
goods, not to anything and everything as become a blockbuster. None has even
Reeves postulated. nearly made it. Not a single advertising
Just as Martin Mayer was more sub- book has achieved substantial sales for
tle than Packard, David Ogilvy was far over 40 years. Why not? Partly because,
more subtle than Reeves. He recognised between them, Packard, Mayer, Reeves
that there was no single Golden Key and Ogilvy answered most of the funda-
that would unlock all advertising’s secrets. mental questions about advertising, and
Instead he believed there were a host of, as since then we have just been refining their
it were, small Golden Keys, each of which answers further, tinkering at the edges. But
would reveal different solutions to different mostly because, as the years have passed,
advertising problems. Fully appreciating advertising has become so normal a part
the heterogeneous nature of advertising, of life that few people outside the busi-
he provided in Confessions a variety of ness believe it to be deeply mysterious
rules and guidelines for effective advertis- and worth reading books about. They’ll
ing, many of which hold true to this day. read articles in the media, and chat about
Being an instinctive copywriter Ogilvy advertising – briefly – to their friends. But
understood print advertising, of which he buy and read a book? It just isn’t that fas-
was a master, far better than television, cinating any more.
a visual medium he never fully grasped. Happily the four great blockbusters can
And he hated posters, because he felt they all still be obtained, with a little effort –
defiled the landscape. Nonetheless, despite and they are all well worth reading, albeit
its shortcomings, I believe Confessions to be in some cases with several pinches of salt.
168

IJASample

Загружено:

Сведения о документе

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

IJASample

Загружено:

Авторское право:

Доступные форматы

International Journal of

Volume 32 Number 1 2013

Constitutive marketing: towards understanding brand community formation 45

The differential effects of position, ad and reader characteristics on readers’  65

Rational and irrational influences on lucky draw participation 85

Communication using celebrities in the non-profit sector: determinants of  101

Advertising and promotion budgeting during volatile economic conditions:  143

The Advertising Association is not responsible for the opinions and

Hamish Pringle and Jim Marshall – Spending Advertising Money in

Timothy De Waal Malefyt and Robert J. Morais – Advertising and

Ex Libris – Winston Fletcher

Brian Till R. Dale Wilson

International Journal of Advertising, 32(1), pp. 7–12

The Hot Topics

How to get people to be attentive to digital ads

Enhancing attentiveness to television commercials in the digital video

New approaches in social media

Customized communications and related privacy issues

Measuring the impact of sponsorships and product placement

Advertising, promotion, nutrition and obesity

above), advertising to children (subject of a forthcoming special issue of IJA in 2014)

Advertising trends in emerging markets

Vedrashko, I. (2011) Game-based marketing: inspire customer loyalty through rewards,

Constitutive marketing: towards understanding brand  45

The differential effects of position, ad and reader characteristics  65

Rational and irrational influences on lucky draw participation 85

Communication using celebrities in the non-profit sector:  101

How banner ads can be effective: investigating the influences  121

Advertising and promotion budgeting during volatile economic  143

wisdom. It tells me what

is important and what

up and move forward.

Vanella Jackson, Worldwide CEO,

Hall & Partners

Admap_170x240mm_Advert_HiResPrintReady_GREY.indd 1 01.08.11 13:47

International Journal of Advertising, 32(1), pp. 17–44

evaluation (i.e. likeability) of the in-game ad placements. By examining the impact of

Player experiences as context characteristics

Specific player experiences

General emotions as player experiences

Relationships between specific and general player experiences

Manipulation of game difficulty

Impact of game difficulty on the effectiveness of in-game

Indirect effect: impact of player experiences

Effect on brand processing

Arousal: limited capacity model of mediated message processing hypothesis

Effect on brand evaluation

Direct effect of game difficulty

Effect on brand processing: limited capacity model of mediated

Effect on brand evaluation: mere exposure theory

Player experiences as mediators

Figure 1: Screenshots of the easy, medium and hard levels

Specif ic player experiences

providing initial confirmation that our

Pleasure Arousal Dominance

Competence Tension Challenge Immersion Flow Boredom

Effects of game difficulty on IGA recognition and brand likeability

Recproduct Recbrand Recbillboard Brandlike

Effects of player experiences on IGA recognition and brand likeability

*p < 0.05, **p < 0.01, ***p < 0.001

Discussion and conclusion

About the authors

International Journal of Advertising, 32(1), pp. 45–64

Models of brand community

Constitutive marketing: towards understanding brand community formation 45

The differential effects of position, ad and reader characteristics on readers’ 65

Rational and irrational influences on lucky draw participation 85

Communication using celebrities in the non-profit sector: determinants of 101

Advertising and promotion budgeting during volatile economic conditions: 143

Constitutive marketing: towards understanding brand 45

The differential effects of position, ad and reader characteristics 65

Rational and irrational influences on lucky draw participation 85

Communication using celebrities in the non-profit sector: 101

How banner ads can be effective: investigating the influences 121

Advertising and promotion budgeting during volatile economic 143

p < 0.05, p < 0.01, p < 0.001