Вы находитесь на странице: 1из 98

Achieving a Rewarding Meeting with an Artificial Chairman

Master Thesis1

Marie-Susanne Ekelin & David Samuelsson

Department of Computer and Systems Sciences


Stockholm University / Royal Institute of Technology

January 2003

Abstract
This thesis concerns how formal face-to-face co-located meetings may be supported by
intelligent technology. According to the hypothesis, replacing a human chairman with an
artificial chairman will not result in a less rewarding meeting.

A preliminary study was conducted to identify the requisites of a rewarding meeting, and find
out what place intelligent technology could have in a meeting situation and what meeting
tasks could be computerised. A rewarding meeting was defined as one that portrays a number
of properties related to efficiency; clear structure, suiting information, and compliance with
efficiency-related norms.

In the experiment, conducted using the methodological framework Wizard of Oz, a simulation
of meeting support based on intelligent technology was set up to incorporate the properties of
a rewarding meeting. The setting used for the experiment was the iLounge, a technologically
enhanced meeting room at K2Lab, Department of Computer and Systems Sciences,
Stockholm University. The experiment served to investigate meeting participants’ attitudes
towards being directed by an artificial chairman, Agnes, implemented to fulfil the requisites
of a rewarding meeting.

Replacing the human chairman with Agnes did not negatively affect the efficiency of the
meeting in terms of meeting structure and the perception of the chairman. However, Agnes’
abilities to perform typical human tasks, like listening, reasoning and communicating were
less successful. It is concluded that this type of intelligent meeting support should preferably
aim at enforcing the structure of the meeting, such as the agenda, rather than striving towards
totally replacing the human chairman.

Keywords: Intelligent meeting support, meeting efficiency, rewarding meeting, artificial


chairman, structure, iLounge, Wizard of Oz.

1
This thesis corresponds to 20 weeks of full-time work for each of the authors.
Table of content
1 INTRODUCTION......................................................................................................................................... 1
1.1 BACKGROUND.......................................................................................................................................... 1
1.2 PROBLEM STATEMENT ............................................................................................................................. 1
1.2.1 Aim.................................................................................................................................................. 2
1.2.2 Hypothesis....................................................................................................................................... 2
1.2.3 Delimitation..................................................................................................................................... 2
1.2.4 Clarification of Terms ..................................................................................................................... 2
1.3 METHODOLOGY ....................................................................................................................................... 2
1.3.1 Preliminary Study............................................................................................................................ 3
1.3.2 Experiment ...................................................................................................................................... 3
1.4 CHAPTER OUTLINE .................................................................................................................................. 3
2 THEORETICAL BACKGROUND............................................................................................................. 4
2.1 RESEARCH ON MEETINGS......................................................................................................................... 4
2.2 WORK GROUPS ........................................................................................................................................ 4
2.3 MEETINGS ................................................................................................................................................ 5
2.3.1 Face-to-Face Meetings vs. Virtual Meetings................................................................................... 6
2.3.2 Meeting Constituents....................................................................................................................... 6
2.3.3 Efficiency in Meetings .................................................................................................................... 7
2.3.3.1 Co-operation................................................................................................................................ 8
2.3.3.2 Communication ........................................................................................................................... 8
2.3.3.3 Goals ........................................................................................................................................... 9
2.3.3.4 Socialising ................................................................................................................................... 9
2.3.3.5 Leadership ................................................................................................................................. 10
2.3.3.6 Structure .................................................................................................................................... 11
2.3.3.7 Norms........................................................................................................................................ 11
2.4 AGENTS ................................................................................................................................................. 12
3 PRELIMINARY STUDY ........................................................................................................................... 15
3.1 PURPOSE ................................................................................................................................................ 15
3.2 RESEARCH EXECUTION .......................................................................................................................... 15
3.3 RESULTS ................................................................................................................................................ 15
3.3.1 Efficiency ...................................................................................................................................... 16
3.3.1.1 Structure .................................................................................................................................... 16
3.3.1.2 Suiting Information ................................................................................................................... 16
3.3.1.3 Efficiency-related norms ........................................................................................................... 16
3.3.2 Artificial Assistant ........................................................................................................................ 16
3.3.3 Socialising ..................................................................................................................................... 17
3.4 CONCLUSION PRELIMINARY STUDY ....................................................................................................... 17
3.5 CRITICISM OF PRELIMINARY STUDY ...................................................................................................... 17
4 EXPERIMENT............................................................................................................................................ 18
4.1 PURPOSE ................................................................................................................................................ 18
4.2 APPLYING THE RESULTS OF THE PRELIMINARY STUDY .......................................................................... 18
4.3 WIZARD OF OZ METHODOLOGY ............................................................................................................ 18
4.4 AGNES ................................................................................................................................................... 19
4.5 THE THREE RESEARCH CONDITIONS ...................................................................................................... 20
4.6 SUBJECTS ............................................................................................................................................... 21
4.7 MEETING SCENARIO .............................................................................................................................. 21
4.8 EXPERIMENT EXECUTION ...................................................................................................................... 21
4.9 PILOT STUDY ......................................................................................................................................... 23
4.10 DATA GATHERING ................................................................................................................................. 24
4.10.1 The Questionnaire ......................................................................................................................... 24
4.11 DATA PROCESSING................................................................................................................................. 24
4.12 RESULTS EXPERIMENT ........................................................................................................................... 24
4.12.1 Questions on Efficiency ................................................................................................................ 25
4.12.2 Questions on Structure .................................................................................................................. 29
4.12.3 Questions on Components of Structure......................................................................................... 30
4.12.4 Questions on Efficiency-related Norms ........................................................................................ 32
4.12.5 Questions on the Presentation Technology ................................................................................... 34
4.12.6 Questions on Trust and Acceptance .............................................................................................. 35
4.12.7 Questions on Attitudes Towards the Chairman / Agnes................................................................ 37
4.12.8 Questions on Attitudes Towards Agnes, Q 39-45. ........................................................................ 38
4.13 SUMMARY OF CONCLUSIONS ................................................................................................................. 39
4.13.1.1 Efficiency .............................................................................................................................. 39
4.13.1.2 Structure ................................................................................................................................ 39
4.13.1.3 Components of Structure....................................................................................................... 39
4.13.1.4 Efficiency-related Norms ...................................................................................................... 39
4.13.1.5 Presentation Technology ....................................................................................................... 40
4.13.1.6 Trust and Acceptance ............................................................................................................ 40
4.13.1.7 Attitudes Towards the Chairman / Agnes.............................................................................. 40
4.14 CRITICISM OF EXPERIMENT EXECUTION ................................................................................................ 40
4.14.1 Group Composition ....................................................................................................................... 40
4.14.2 The Scenario.................................................................................................................................. 40
4.14.3 Sample of Subjects ........................................................................................................................ 41
4.14.4 Drawbacks in Experiment Execution ............................................................................................ 41
4.14.5 Other.............................................................................................................................................. 41
5 DISCUSSION .............................................................................................................................................. 43
5.1 FUTURE WORK ...................................................................................................................................... 45
5.2 RELATED WORK .................................................................................................................................... 45
6 ACKNOWLEDGEMENTS........................................................................................................................ 46

7 REFERENCES............................................................................................................................................ 47

8 APPENDIX I – PRELIMINARY STUDY ................................................................................................ 51

9 APPENDIX II – COMPLETE EXPERIMENT RESULTS .................................................................... 67

10 APPENDIX III – AGNES’ COMMANDS ................................................................................................ 95


1 INTRODUCTION
This thesis concerns meetings, and how they may be supported by intelligent technology. The
work has been carried out with support from the Fuse group at K2Lab, a joint research centre
between Stockholm University and the Royal Institute of Technology, located at the
Department of Computer and Systems Science in Kista, Stockholm.

1.1 Background

Mankind is a social species and all recorded human cultures have social gatherings where
people meet regularly in associations, teams or clubs [Jay, 1976]. Following this, meetings
and work groups are highly integrated parts of corporate culture. As an example, in 1991 it
was estimated that as many as 11 million meetings occurred daily in the United States alone,
and that managers spent 30 to 80 percent of their time in meetings [Poole, 1991]. A lot of
money was spent pursuing meetings; the same year one survey found unproductive meeting
time in the United States to be a $37 billion annual waste [Smith, 1991]. Managing meetings
is a complex issue, and it is necessary to understand the forces that shape the meeting in order
to make the employment of meetings successful.

To eliminate unnecessary meeting time, it is important that the meeting support available is
efficient and facilitates working together in group settings. Different types of more or less
advanced meeting technology have been introduced with the purpose of improving and
facilitating meeting performance. Some have succeeded in adding value to the meeting
session, whereas others have tended to complicate meeting execution even more. Advanced
computerised meeting support has often focused on supporting geographically distributed
meetings, such as virtual meeting rooms [for example, Lantz, 2000]. However, equivalent
types of support for meeting execution in a co-located setting are less common. Attempts at
improving face-to-face meetings with the use of intelligent technology are also hard to come
by.

Combining a co-located setting with intelligent technology in order to optimise meeting


performance could prove an interesting area of study, and is the focus of this thesis.

1.2 Problem Statement

The basis for this thesis has evolved from an interest in group psychology and artificial
intelligence, and the possible merger of these fields. We found that this could be investigated
by looking at meetings. Meetings are often inefficient and resource-consuming. Improving
them could make for higher meeting quality, and save time and money.

Meeting literature often focuses on the success of a meeting in terms of the productivity and
the efficiency of the group. However, we believe that it is equally important to take heed of
social aspects such as group relationships and the wellbeing of the members, something that
according to Schwartzman [1989], has been largely neglected by researchers. Therefore, we
further believe that the individual perception of the meeting is important for a successful
meeting.

Since we found that merely studying literature could not fully describe the individual
perception of a meeting, we conducted a preliminary study that concentrated on looking at
different personal experiences of meetings, and identifying the requisites of a rewarding

1
meeting. The term rewarding is a prevailing concept throughout this thesis and is used to
reflect what makes a meeting successful in the eyes of the participant.

The definition of what constitutes a rewarding meeting formed the basis for an experiment. In
the experiment, a simulation of meeting support based on intelligent technology was set up to
incorporate the identified properties of a rewarding meeting.

1.2.1 Aim
The aim of this thesis is to investigate whether it will be possible to execute rewarding
meetings by implementing support based on intelligent technology.

1.2.2 Hypothesis
Replacing the human chairman with an artificial chairman will not result in a less rewarding
meeting.

1.2.3 Delimitation
This thesis only concerns formal face-to-face meetings. The term meeting refers to a
“communicative event involving three or more people who agree to assemble for a purpose
ostensibly related to the functioning of an organisation or group” [Schwartzman, 1989].

A formal face-to-face meeting refers to a meeting situation that follows some sort of
predefined and agreed upon structure. The meeting takes place with attendants seated around
a meeting table in a co-located setting, i.e. where all attendants are positioned at the same
geographical location, and thereby visible to each other. Prior to the meeting, attendants are
designated, summoned and briefed about the contents and objectives of the meeting.

The preliminary study focused on Information interchange meetings, i.e. meetings dedicated
to the sharing and gathering of information. However, due to difficulties in isolating this
particular type of meeting from other types of meetings, a broader scope was adopted in the
experiment, to also encompass meetings carried out for other purposes.

The setting used for the experiment was the iLounge, a technologically enhanced meeting
room, as will be described in chapter 4.8. This type of meeting room therefore differs
somewhat from rooms normally used for meetings.

1.2.4 Clarification of Terms


Throughout this thesis, subjects participating in the two studies conducted are referred to as
interviewees, group members, respondents or attendants. In the experiment, subjects are
divided into participants and chairmen. Chairmen are at times spoken of as moderators or
facilitators of the meeting.

1.3 Methodology

In order to test the hypothesis we needed to define a rewarding meeting, as well as investigate
how different levels of chairman computerisation affect the perception of a meeting. The
preliminary study examined the requisites of a rewarding meeting, and the experiment
focused on the role of the artificial chairman.

A qualitative approach was adopted in both the preliminary study and the experiment. The
intention was not to reach a result from which statistically valid conclusions could be drawn
but rather to focus on individual experiences of meeting situations.

2
1.3.1 Preliminary Study
The method used in the preliminary study served to gain some insight to how people relate to
and feel about formal face-to-face meetings in general. The preliminary study consisted of a
set of open ended interviews.

The intention of the preliminary study was to establish a definition of a rewarding meeting
that could be used throughout the rest of the thesis, and clarify what role intelligent
technology could have in such meetings and what tasks it would perform.

The sample of subjects consisted of volunteers acquainted with the research administrators.
Prerequisites for the interviewees were a familiarity with information interchange meetings,
and familiarity with the use of computers.

1.3.2 Experiment
In the experiment, the qualitative method served to investigate participants’ individual
experience perception of a particular meeting in the experimental setting. Questionnaires with
closed and open ended questions were used to assess the meeting.

The experiment was based on the findings from theory and the preliminary study, and served
to investigate meeting participants’ attitudes towards being directed by a computerised
meeting facilitator, an artificial chairman, implemented in order to make the formal meeting
rewarding.

Three research conditions, HuMod, CoMod and AgMod, were used in order to compare
different levels of chairman computerisation. There was no agent technology behind the
artificial chairman, instead, the experiment was carried out within the methodological
framework Wizard of Oz [Dahlbäck et al, 1993], a methodology that encompasses simulation
of a computer service in a way that makes the users unaware that they are interacting with a
simulation. The intention was to make participants believe that the chairman was based on
real agent technology. It was decided that such a simulation would suffice to successfully
execute the experiment. The chairman communicated through a text based interface.

The experiment was situated in the iLounge, a technologically enhanced meeting room
located at the Department of Computer and Systems Science at Stockholm university, Kista.

Before the main experiment commenced, two pilot studies were performed to try out the
different experimental conditions. These involved a number of employees from the Fuse
research group.

The sample of subjects consisted of volunteers, mainly students, who responded to


advertisements.

1.4 Chapter Outline

• Chapter 2 discusses relevant theoretical findings, mainly meeting and work group theory
and the agent concept.
• Chapter 3 presents the results of the preliminary study.
• Chapter 4 describes the execution of the experiment and results related to the hypothesis.
• Chapter 5 discusses and elaborates on the results related to the hypothesis.

3
2 THEORETICAL BACKGROUND
Literature studies were conducted in order to gain a deeper understanding of how meetings
function in general, and why they are so important to most organisations. Difficulties in
finding literature discussing the meeting as a concept led us to look into a related area; work
groups. Group work in organisations is often managed and controlled through the execution
of formal meetings. Therefore, in order to understand the meeting situation, the properties of a
work group must be considered, and parallels are drawn between meetings and what is
considered effective groups or teams.

Theory about agent technology is briefly looked into, so as to investigate possible ways of
implementing them in a meeting scenario.

2.1 Research on Meetings

However pervasive and influential in western culture, meetings have rarely been focused on
by researchers. Schwartzman [1989] claims this to be the result of neglect due to meetings
being so common and taken-for-granted in organisations that they have not been recognised
as important parts of organisational life. Researchers have mainly been using the meeting as a
tool and background structure for studying small group interaction and communication, or
studied the processes behind decision-making, ideology, power and conflict. The meeting as a
"communicative event that structures the group’s behaviour" has not been thoroughly
investigated [ibid].

Artificiality is a problem which most studies of meetings suffer from as a result of being
conducted in experimental settings where people, who are usually students, do not know each
other [ibid]. A common misconception is that meetings can be seen as blank slates, that apart
from being tools for work have no impact or effect on organisations, communities or
societies. Meetings should instead be examined as a part of the sociocultural setting within
which they are embedded; an organisation, a community, a society. In real life meeting
sessions, participants almost always have a previous history of working together as a group. A
meeting is almost always preceded by other meetings or some other kind of preparation, and a
meeting often debouches into new meetings or some kind of follow up [ibid].

Like Schwartzman, McGrath [1991] points out that the way groups are put together and
studied experimentally does not represent the natural construction of groups in everyday
affairs. The groups that are being studied are usually put together by the experimenter and
have no past or future as a group, and they are assigned relatively simple tasks. Real groups
usually have a history and possibly a future together, they are seldom isolated but exist within
a larger social context, such as a community, organisation or department. Activities are often
goal-directed and tasks are not always simple or assigned by someone else, they are decided
on by the group.

2.2 Work Groups

Today, the work group is the arena where most organisational work is carried out. Kyng
[1991] maintains that all work essentially is a social activity, and that the meaning of work
arises in co-operation. Jewell and Reitz [1981] compare the importance of groups to
organisations, to the importance that fire had to our ancestors; " If understood and used
correctly they are a remarkably powerful and flexible tool; if misunderstood or used

4
incorrectly they are a destructive force. Like fire, groups can be created or emerge
spontaneously and, like fire, groups are most dangerous when ignored". When in a positive
group climate, people working together can achieve more than just the sum total of their
individual effort [Phillips, 1973], but when dysfunctional, groups can affect efficiency and
well-being in a negative way [Myers, 1999].

There are many definitions of what constitutes a work group. In order to establish what is
referred to when discussing work groups, a brief overview of theoretical definitions follows.

Turner & Killian [1987] distinguishes a group from a collective, which is a set of individuals
that is large, unorganised and temporary, whereas the group on the other hand is "smaller,
more permanent, and has a definite infrastructure". Shaw [1981] defines a group as two or
more people who interact and influence each other. He argues that the one thing all groups
have in common is interaction between group members. McGrath [1984] also emphasises the
importance of interaction, claiming that "the essence of a group lies in the interaction of its
members". He defines work groups as “social aggregates that involve mutual awareness and
potential mutual interaction” and describes the work team as a deliberately designed social
unit.

The idea of mutual awareness is shared by Hackman [Granström, 1998], who believes that the
individuals have to view themselves, and be recognised by others, as a social unit to qualify as
a group. In addition to this, members must be depending on each other to complete tasks, have
objectives that are of importance for others, and belong to some kind of organisation.
Granström [ibid] sees the work group as a social system and concludes that members in a
group collaborate and perceive themselves as a group. Additionally, in his definition the work
group has a previous history together, works towards a tangible goal, and performs a formal
function within the boundaries of a bigger organisation. Schwartzman [1989] also focuses on
the purpose of the group, stating that "a group is a gathering of individuals involved in a
particular form of activity, for example a meeting".

It can be argued that the view of the work group is somewhat influenced by the environment
that the group operates in, and the functions it is expected to carry out, something that makes
general definitions rather vague. However, as can be seen, the definitions of a work group
share a number of basic characteristics; a social aspect encompassing members’ awareness of
each other and interaction with each other, the group as a part of a larger entity, and the notion
that the group produces some form of output. This will suffice to convey an image of the
work group that is needed for full understanding of this thesis.

2.3 Meetings

Meetings occur in all societies to provide direction, govern or regulate activity in some way.
The meeting session is a frequently applied setting for a work group to conduct its business.
Schwartzman [1989] defines the meeting as a “communicative event involving three or more
people who agree to assemble for a purpose ostensibly related to the functioning of an
organisation or group, for example, to exchange ideas or opinions, to solve a problem, to
make a decision or negotiate an agreement, to develop policy and procedures, to formulate
recommendations, and so forth”. An obvious goal of conducting meetings is to communicate
information throughout the organisation, both from the top down and from the bottom up
[Napier, 1993].

5
Although research has proven that informal and spontaneous meetings are very important for
efficient information exchange within organisations [Wiberg, 2001], traditional formal
meetings are still abundant and considered vital in most organisations. Meeting time may not
always be spent efficiently, but meetings still perform important functions within
organisations, and eliminating them altogether is not an objective. As pointed out by Jay
[1976], if a meeting is seen as a pooling of individual resources to reach an aggregated result,
a meeting is useful only if the people present perceive that the meeting is leading somewhere
they could not have got to individually. There are both individual and organisational reasons
for conducting meetings.

Schwarzman [1989] argues that through the meeting, the organisation can make itself visible
and apparent for its members. The meeting can also be used to let people “get things off their
chest”, encourage open communication between managers and workers or evaluate changes in
leadership style.

On an individual level, there are several personal and interpersonal needs that a well-
functioning meeting can satisfy [Napier, 1993]. It provides a chance for individual members
to reflect on their own actions, as well as examine and manifest their relationships to each
other in the current context [Weick, 1979]. Schutz [1958] focuses on three needs that most
people share: the needs for inclusion, control and affection. He claims that the possibility of
fulfilling these needs is an incentive for people to join groups. "Inclusion represents our need
to belong, be involved and be accepted. Control is related to feeling competent and confident.
Affection refers to the need to feel liked.".

Napier [1993] mentions other social needs that meetings serve to satisfy; the building of
personal relationships, the opportunity for fun and social awards, and the possibility to learn
from each other. Meetings can infuse a sense of mutual accomplishment, lead to a
psychological commitment to the ideas generated within the meeting, and a feeling amongst
group members that their ideas are sought after.

2.3.1 Face-to-Face Meetings vs. Virtual Meetings


Today, alternatives to the traditional face-to-face meeting in the form of support for virtual
meetings, such as chat rooms and web conferencing, are growing in popularity. Some studies
[Lantz, 2000] show that it is possible to work more in a more task oriented fashion in a virtual
world than in a face-to-face meeting. However, participants often complain about
misunderstandings and time-consuming proceedings due to the fact that such communication
more or less lacks the richness of face-to-face communication with gestures, facial
expressions and verbal clues [ibid].The loss in comparison to face-to-face meetings is not
only verbal or body lingual. When Lesbie et al [1996] compared text based communication to
verbal real time communication they found that face-to-face meetings produce 2.6 times as
many actions as meetings that were communicated electronically. The result was partly
explained by the human ability to speak and listen much faster than reading and writing.

2.3.2 Meeting Constituents


Typically, a meeting includes episodic, multiparty conversation and conventions that regulate
communication within the group. This communication is assumed to relate to the purpose of
the meeting and frame the behaviour that occurs within the meeting [Schwartzman, 1989].

However, different types of meetings serve to fulfil different purposes, and the constituents of
a meeting differ somewhat depending on the purpose of the meeting. For example, an

6
information sharing meeting mainly serves to keep people informed of changes in the
organisation, and eliminate the sense of isolation that many people experience in the
workplace. Diagnostic or fact finding meetings on the other hand, serve to identify problems
by generating information that leads to a better understanding of difficulties within the
organisation [Napier, 1993].

When presented with a new problem or challenge, the group decision-making process that
ensues consists of three phases [Brewer and Crano, 1994], the first being Orientation. During
this phase the group tries to reach a consensus about the problem and its nature, and what
methods should be applied to solve it. Research has shown that groups that discuss decision
strategies early on in the process have a greater chance of succeeding in finding a solution
[Hackman et al, 1976 ].

The next phase is the Discussion. When the members of the meeting group have understood
the problem and has come up with different suggestions on how to address it, they start to
evaluate these suggestions to see which will be the most suiting.

The final Decision phase is reached when there are sufficient votes for a certain solution and
it is time to try to get the whole group to agree on a decision. Such an agreement is reached
through specific norms and rules, so called ”social decision schemes ” [Brewer and Crano,
1994]. These include voting for a majority decision, uniting different suggestions into a
compromise, delegation of the decision to a smaller group that makes a decision for the rest of
the members, and consensus formation where the discussion does not end until a unanimous
solution exists.

2.3.3 Efficiency in Meetings


The success of a meeting can be defined in more than one way, and literature covering this
area frequently refers to success as the productivity and efficiency of the group. Improving the
efficiency of work groups and meetings is a popular subject in meeting literature, likely due to
that fact that so much unproductive time is spent pursuing them. This means that when
discussing successful meetings, efficiency will be a primary concern. Different aspects of
researchers’ views on meeting success is discussed below.

Meeting efficiency is mediated by a range of factors; ineffective leadership, critical norms and
lack of goal to name but a few [Napier, 1993]. Lennéer [1998] states that the cohesiveness of
the group, along with a well-functioning structure and system of norms are vital for work
group success.

Napier [1993] also focuses on the importance of cohesiveness in the group, and its effect on
productivity; "If members spend their time strictly on business – the surface agenda – and
ignore interpersonal relationships and hidden agendas, misunderstandings can increase".
However, he also argues the need for structure and guidelines for process maintenance in the
group. In one study, when more structure was provided, members felt encouraged to risk more
and be more disclosing of feelings [Kernberg, 1980].

If efficiency is regarded as a high level of productivity, Hare [1992] suggests that apart from
the obvious need for appropriate information and sufficient time for task completion, a
functioning communication structure for group interaction, and a competent leader to co-
ordinate activities is very important. Hence, finding ways to further enhance group
communication and co-ordination could produce higher group performance and efficiency.

7
In short, group dynamics must be considered to ensure efficiency in the meeting.
Schwartzman [1989] suggests a number of important features related to this:

• Involvement of all members


• Having a meeting “facilitator” present
• Effective leadership
• Developing trust and shared responsibility
• Ability to resolve conflicts

Some of the factors that affect efficiency are more elaborately discussed below.

2.3.3.1 Co-operation
As mentioned, the actions of each member in an interdependent work group affect and
influence other members. This means that a high degree of personal responsibility is required
to ensure successful and efficient collaboration [Engleberg, 1997]. Collaboration implies that
the members of the work group share common goals, and co-operation implies that the
members of the group work together [Maher et al, 1999]. As soon as there is any
collaboration being initiated the need for co-ordination arises. Through co-operation people
become interdependent of each other and therefore have to co-ordinate their joint efforts
[Wiberg, 2001]. This interdependence indicates a need for a functioning social group
atmosphere, cohesiveness.

When one person strongly dominates and controls the group it is usually a sign that the group
is failing to function [Phillips, 1973]. However, group members that are too focused on co-
operation might start striving to please each other instead of working towards the goal. This
group may benefit from a stronger leadership.

2.3.3.2 Communication
Communication is a vulnerable issue, and effective interaction can be disturbed by one person
monopolising a discussion, or group members ignoring a group member’s comments
[Engleberg, 1997]. The style of interaction helps determine the communicative relationship
within the group. Listening in groups is a challenging activity, due to the multiplicity of
speakers, perspectives and goals. Misunderstandings and missed information often occur
when people are expected to listen, pay attention to everyone in the group, and at the same
time be able to respond to questions on the spot. There is a risk for cognitive overload, leading
to the inability to process all the information present.

Another reason for information overload could involve excessive information sharing
between group members, which may lead to members starting to ignore each others
information [McGrath, 1994]. Difficulties in attending to all members simultaneously, that
follow from inefficiencies in listening, often go unnoticed [Watson, 1996].

Open communication is hard to achieve and can easily be threatened by destructive group
behaviour. The communication pattern within the group is a crucial factor for predicting the
chances for successful co-operation. Lennéer [1998] claims that "communication is the most
important tool for co-operation and a good expression of the atmosphere in the group:
communication largely is the climate". According to Phillips [1973], by studying the
discussion process it is possible to guide members to more constructive styles of group
behaviour. Increased communication is generally believed to result in increased co-operation

8
[Short et al, 1976], although differences when conducting experiments and defining co-
operation make results inconclusive.

Jay [1976] claims that it is usually daily meetings where people work together on the same
project that are most likely to be successful, since people who work closely together
communicate better than strangers. In other meetings that are more irregular, occasional or
between people whose work is not directly related to each other, the discussion is easily swept
off course.

2.3.3.3 Goals
It is important that the reason for the meeting, and what the meeting is intended to achieve, is
communicated to the group [Jay, 1976]. Having a clear understanding of the goal of the work
group has been found to have a determining effect on the success of the group [Larson et al,
1989]. Goal orientation can help keep the group focused on its mission [Myers, 1999]. Even a
group without explicitly stated goals will develop goal norms to work towards, but group
members that have a clear understanding of the goal will have better control over their
behaviour [Phillips, 1973].

A meeting group can sometimes lose its direction and begin to struggle towards any kind of
consensus regardless of if it is relevant to the goal of the group. This is caused by members
being excessively solicitous of the feelings of others. It may then be necessary to threaten this
consensus to force the group back in focus. However, appointing group members to keep
track on destructive group events may trigger other group members to model their behaviour
after what they think the observers want, rather than to moderate their behaviour to keep
themselves as well as the group satisfied [ibid].

2.3.3.4 Socialising
The work climate of the group, and the way the group perceives it - in this case the meeting
session - has great impact on the groups' chances of achieving their goal. Lennéer [1998]
points out that a poor work climate often leads to a higher turnover rate, higher absence, and
lower production than in workplaces with a healthier climate. The experience that the group
members have of the meeting is a crucial factor when it comes to attaining efficiency and
wellbeing in the workplace.

Lennéer [1998] argues that over the course of time, everything that happens to the group –
changes, conflicts and crises – affects the way the group is functioning. Tuckman [1965] and
Fisher [1970] claim that groups go through different developmental stages. The first stages
are dedicated to exploring and testing personal relationships, discussing important issues and
resolving conflict, whereas in the latter stages members start working as a team, focusing their
energies on the task at hand. Fisher classifies decision-making groups in four stages;
orientation, when members get to know one another, conflict, when members express
opinions, emergence, when a search for solutions replaces argumentation, and reinforcement,
which is when group members agree upon a decision. Both Tuckman and Fisher acknowledge
that group members need time to get acquainted with one another and sort out differences in
opinion that otherwise could be a ground for conflicts.

Poole and Roth [1989] list a set of factors that can derail a group from this ideal sequence of
phases, and cause the group to move more slowly through the stages. These factors include
group size, the level of member co-operation, the clarity of goals, the history of conflict, the
availability of information and possible solutions, and the concentration of power in the
group.

9
It is probable that when a group has just formed, the social dimension may require more
attention than the task, whereas in later developmental stages there is less need to concentrate
on the social needs of the group members. However, to sustain efficiency, the group must find
a balance between fulfilling the personal needs of the members and accomplishing the goal of
the group [Engleberg, 1997]. If the group dedicates time to improving communication
processes in the initial stages of its life, and continues to strive towards a functioning
communication pattern, greater work efficiency will ensue [Napier, 1993]. Hall and Williams
[1970] point out that groups that have been trained in problem solving work more efficiently
together.

According to Schutz´s theory [1958] mentioned earlier, recognising and adapting to meeting
participants’ different needs could improve a group’s performance. This means promoting
feelings of inclusion, acceptance, control, confidence, and affection. [Engleberg, 1997].
Napier [1993] claims that the failure of a meeting is often not the result of inattention to the
tasks, but rather by poorly addressed interpersonal relationships. He sees a correlation
between the success of the meeting and the satisfaction of interpersonal-personal needs of the
group. He also suggests preventative measures that can be undertaken in order to focus on
group dynamics ensuring the success of the meeting:

• Stimulating participants through variation of activities


• Participants should feel utilised in a meaningful manner
• Individuals should experience some feeling of success
• Participants should feel personally responsible for the success of the meeting
• Participants should be given the opportunity to learn something new
• Participants should feel that the cause of the group is meriting
• Participants should feel challenged
• Members should enjoy themselves
• People should feel accepted
• There should be an interdependence between members

2.3.3.5 Leadership
The leader of the group, the chairman, facilitator or moderator, plays a vital role for the
success of the meeting. A trained meeting facilitator is preferred in many group settings, in
order to fully control and lead the meeting, whether it be a traditional or computerised
gathering. [McGrath, 1991].

Jay [1976] stresses the importance of the chairman, and sees leadership as an act of assisting
the group towards efficiency rather than leading it. It is the leader that sets the standards for
the type of meeting behaviour that is desired, and keeps the meting pointed toward the
objective. The leader must make it clear to the group what the expected outcome of the
meeting is, whether it be merely discussing a topic or deciding on future plans. It is the task of
the leader to make sure that everyone in the group has an equal chance of contributing to the
meeting. This involves “drawing out the silent members, controlling the talkative, protecting
the weak, encouraging the clash of ideas, watching out for suggestion-squashing, coming to
the senior people last and closing on a note of achievement” [ibid].

Jay states that most effective discussions have in fact two leaders, one team/social leader and
one task/project leader. If this is not possible, the social leadership must surpass the
importance of task leadership. Social leadership is the democratic leadership style, delegating

10
authority, building teamwork, mediating conflicts, and being supportive. Social leadership
helps prevent groupthink1, and helps members feel more satisfied with their participation in
making decisions [Spector, 1986]. Given control of their tasks, workers also feel more
motivated [Burger, 1987]. Task leadership is a more directive leadership style, organising
work, setting standards, focusing on goal attainment [Myers, 1999].

Related to this, Viller [1993] makes a distinction between two parts of a groups task; the
content, or what is to be achieved, and the process, how it is to be achieved. The process can
be subdivided into task behaviour that serves to carry out the assigned task, and maintenance
behaviour, which aims at securing the group as a cohesive unit. To successfully perform the
task, the group needs to heed to both these types of behaviour. Viller stresses the need for a
group facilitator that assists the group in achieving its objectives. The facilitator should be
someone who understands group processes and can help the group understand and resolve its
problems. The facilitator works as a central person, and may initially be all that is common to
everyone in the group, thus having a chance to set group norms and improve group cohesion.
Later on in the group process, as interpersonal relationships among group members evolve,
the role of the facilitator changes into that of an enabler who intervenes only when necessary.
Due to differences in power, the role of the facilitator as either a leader or an assistant affects
the status the facilitator will have in the group. Even though the dynamic aspect of the group,
with relationships changing throughout the group’s life, is one of the group’s advantages,
problems can occur.

2.3.3.6 Structure
Research has shown that intervening into the work group, by providing more structure, can
increase problem solving effectiveness, cohesion, risk taking and productivity [Napier, 1993].

Schwartzman [1989] claims that the meeting can benefit from improving its structure, by
measures such as:

• Thorough preparation of the meeting


• Following a strict series of steps
• Adhering to time frames
• Setting meeting priorities and goals
• Developing a structured agenda

Many researchers stress the importance of stating the purpose of the meeting in advance, in
order to make meeting participants aware of what the goal and expected outcome of the
meeting is. Jay [1976] argues that it is important to make sure the agenda is clear and not
ambiguous, as well as stating the reason for bringing each topic up, whether it be for
information, for discussion or for decision. Circulating the agenda a couple of days before the
meeting gives people time to prepare themselves.

2.3.3.7 Norms
The formation of norms, the rules of the group, is a factor that influences a group’s ability to
mature [Engleberg, 1997]. Norms can be defined as "sets of expectations held by group
members concerning what kinds of behaviour or opinion are acceptable or unacceptable, good
or bad, right or wrong, appropriate or inappropriate” [Andrews, 1996]. All groups develop
norms [Parks & Sanna, 1999]. When interacting with the group, new members internalise the
norms of the group through socialisation. Sometimes norm formation in the group can be
1
The tendency to suppress disagreement in order to keep the group in harmony

11
problematic, for example, the norms of interaction may contradict social rules by which
members have learned to live [Phillips, 1973]. However, deviating from group norms can be
constructive if the member who resists the norm is still working to promote the group goal
[Engleberg, 1997]. In fact, deviation can bring about better group performance and decision
making by forcing other members to defend positions, address important issues and explore
alternatives. This may lead to the group examining and ultimately changing non-functioning
norms.

Norms are meaningful when accomplishing group goals, since without them there would be
no agreed-upon way to organise and perform work. The interaction and communication
within the group is also heavily regulated by the norms of the group. It should thus be an
objective to create group norms that help build a supportive communication climate. This is
perhaps more easily achieved when norms are explicit and discussed openly within the group
setting. Implicit norms may be harder to change, since they are often the result of group
interaction and usually recognised only when violated. Changing destructive norms can also
be difficult due to the group pressuring the members to conform to the existing norms
[Engleberg, 1997]. According to Napier and Gershenfeld [Napier, 1993] this is especially
hard if any of the following factors are applicable:

• You want to continue your membership in the group


• You have lower status than the other members
• You strongly support the groups principles and goals
• You get along and like working with the other group members
• You may be punished for violating norms and/or rewarded for compliance

Since these factors are valid for many work groups operating in all types of organisations, the
difficulties of altering destructive norms should not be neglected.

2.4 Agents

Intelligence is a typical human characteristic often put in relation to computers, partly because
people often apply a certain amount of intelligence to computers and their applications.

Defining intelligence is a controversial issue and has been much debated. This study will not
discuss intelligence as a standalone concept but rather how it occurs in artificial agents, a
central concept in the field of artificial intelligence (AI). In this discipline the term
intelligence is generally understood as automated goal oriented behaviour that is flexible and
rational under changing environmental circumstances [Weiss, 1999]. Note that this should not
be perceived as a definition but more as a collection of generally accepted ideas.

Just as with a definition of intelligence, what constitutes an agent differs considerably


depending on who you ask and what theories you look into. According to Wooldridge [Weiss,
1999] there is no generally accepted definition of the term agent. However, Wooldridge
stresses the importance of such a definition so that the term does not lose its meaning, hence:

“An agent is a computer system that is situated in some environment, and that is capable of
autonomous action in this environment in order to meet its design objective” [Wooldridge and
Jennings, 1995].

12
This generous view points out the obvious determinant for every agent, its environment. The
environment can often be both computational and physical; an agent could need to
communicate with people, software and other agents (humans could also be regarded as
agents). There are situations when an agent can be useful on its own but more often it takes
part in a larger social system where the combined knowledge and resources are used to reach
some common or individual goal.

Bradshaw [1997] attempts to make a distinction between two related approaches to defining
agents: one that looks upon agenthood as the ascription made by someone, and another
focused on the objective description of characteristics that the agent possesses. To explain the
notion of agents as ascriptions we simply appreciate that people define agents subjectively, an
agent is not so much a sum of its attributes as it is an attribution of it’s observer. “/…/ one
person’s “intelligent agent” is another person’s “smart object” is tomorrow’s “dumb
program.”.

Obviously, this constructivistic way of reasoning can be applied in the attribution of all
phenomena in our environment but is perhaps more valid when trying to define abstract
concepts such as intelligence. A less philosophical and possibly more practical definition is
that of an agent as a sum of its characteristics and attributes, where some are commonly
regarded as typical for software agents. According to Etzioni and Weld [1995] and Franklin
and Gaesser [1996] these are:

• Reactivity: ability to selectively sense and act.


• Autonomy: being goal directed and proactive.
• Collaborative behaviour: being able to work in concert with other agents.
• Communication ability: ability to communicate with persons and other agents with
language resembling human-like dialogue rather than typical program-to-program
protocols.
• Inferential capability: Ability to act on abstract task specification using prior knowledge
of general goals and preferred methods to achieve flexibility, and to go beyond the
information given.
• Temporal continuity: persistence of identity and state over long periods of time.
• Personality: the capability of manifesting the attributes of a “believable” character such as
emotion.
• Adaptivity: being able to learn and improve with experience.
• Mobility: being able to migrate in a self-directed way from one host platform to another.

Since the agent flora is so diverse these characteristics cannot be found in every agent and the
lack of a clear definition mentioned above obviously makes for many possible agent
configurations. It would therefore be desirable to make a distinction and place agents in some
sort of hierarchy. Verhagen [2000] differentiates agents into four types with ascending
complexity; reactive agents, plan autonomous agents, goal autonomous agents, and norm
autonomous agents.

Reactive agents are only capable of stimuli-response actions, and have no personality-like
traits or means of adapting to their surrounding world.

Plan autonomous agents differ from reactive agents in the sense that they have some type of
goal-oriented behaviour. However, this goal is predefined and cannot be changed by the agent
itself. To achieve this goal or objective the agent is equipped with a set of actions that are

13
triggered according to what type of requests or environmental state affects the agent. Plan
autonomous agents are more autonomous than reactive agents because they can choose what
action to take to strive closer to their main objective. It has no other interests than to reach its
goal.

Goal autonomous agents are more aware of their situation than the above agent types. They
have the ability to evaluate requests made by other agents or environmental states, to judge if
these requests are in the interest of the agent, that is, if the result puts the agent in a better
state in terms of it’s goals. The agent itself decides with what and whom to react. Verhagen
[2000] calls this strategic reasoning.

The last and most complex form of agent in this hierarchy is the norm autonomous agent.
Here, the ability to choose which goals are legitimate to pursue according to a system of
norms is very important. The agent has the autonomy to create it’s own goals and objectives,
but also to judge other agents and their objectives and deny requests from these agents if it
finds them illegitimate. However, if the agent finds that its norms do not comply with the
social system it shares with other agents, it can change these norms accordingly. When the
norms change, the agent has to re-evaluate goal priorities, remove some goals and make new
ones.

14
3 Preliminary Study
A deeper understanding of what constitutes quality in a meeting was required in order to
investigate whether it would be possible to achieve rewarding meetings with the aid of
intelligent technology. To that end a set of interviews were conducted to investigate peoples’
experience of meetings in general, and attitudes towards implementing intelligent technology
to support meeting sessions in particular.

3.1 Purpose

The main purpose of the preliminary study was to gain understanding of the forces behind the
vast amount of meetings taking place in companies, and what people hope to achieve when
partaking in them.

This was done to establish a definition of a rewarding meeting that could be used throughout
the rest of this thesis, i. e., to pinpoint the features of a meeting that are fundamental when
deciding if a meeting has been rewarding to the participants or not.

Another central idea was to identify weak spots in the overall management of meetings, that
could benefit from the involvement of intelligent technology. The respondents' experience of
technology for meeting support was also investigated, with the intention of recognising flaws
in existing tools that are commonly used.

Different aspects of group dynamics such as co-operation, communication and collaboration,


were to be considered in the preliminary study, since these concepts are commonly cited in
literature as being important to meeting efficiency. This prompted questions related to norms
and roles to be asked.

Questions regarding the actual design of the implementation of intelligent technology were
also posed, by asking what place an artificial assistant could have in a meeting situation, and
what tasks of the meeting could be computerised. The term artificial assistant was used since
it was believed that assistant-related tasks would primarily be requested, such as taking
minutes, summing up the meeting’s key information and other administrative tasks.

3.2 Research Execution

Eleven interviews were conducted with people of mixed age and gender. The sample of
subjects consisted of volunteers acquainted with the research administrators. Prerequisites for
the interviewees were a familiarity with information interchange meetings, and familiarity
with the use of computers. Due to the need for respondents to be acquainted with the use of
computers, the professions of the interviewees were often similar. Consequently, most
respondents worked at companies placed within the IT-sector.

Fourteen qualitative questions were asked, related to the respondents' personal experience of
information interchange meetings.

3.3 Results

The information derived from the interviews served to point out and clarify what areas to
further investigate in a more controlled and less abstract manner. Throughout the interviews

15
certain issues were repeatedly brought up by respondents, indicating a common attitude
among interviewees towards what renders a meeting rewarding. All of these issues can be said
to relate to efficiency. For a full presentation of questions and results, see appendix I.

3.3.1 Efficiency
Prior to conducting the interviews, our expectation was that efficiency would probably prove
to be an important factor, since much of the literature available covering this area stresses this
argument. This was strikingly apparent within respondent behaviour. Even before it was
formulated in a specific question (see appendix I, question 3), there was a general consensus
among answers that efficiency was the primary requisite of a rewarding meeting. Important
parts of efficiency was the structure of the meeting, the properties of shared information, and
norms that can be related to efficiency.

3.3.1.1 Structure
Most respondents stressed the importance of an agenda and a chairman to uphold meeting
structure. Respondents’ wish for a set timeframe, a clear purpose or goal of the meeting and
everyone coming to the meeting prepared, also reflect the need for structure to make the
meeting efficient.

3.3.1.2 Suiting Information


Responses showed that the sharing of information is a delicate matter, frequently hampered
by lack of sufficient information, which leads to misunderstandings. Access to suiting
information, meaning that it is relevant, correct, sufficient and up-to-date, was regarded an
important aspect of efficiency.

3.3.1.3 Efficiency-related norms


The norms that were identified were frequently related to the efficiency of the meeting. The
most obvious norms were that everyone should come prepared to the meeting and partake
equally in the meeting. Coming prepared included reading information sent to you prior to the
meeting, and planning presentations in advance. Meetings where some of the attendants did
not partake in decision-making, or were ill prepared where perceived as less efficient.

3.3.2 Artificial Assistant


We believed that it was primarily assistant-related tasks that would be requested of the
intelligent technology, more specifically taking minutes, summing up the meeting’s key
information and other administrative tasks. These were also found to be important issues to
the interviewees. However, since it was mainly tasks related to the agenda and the meeting
co-ordinator (i. e. the chairman), that were deemed critical for a rewarding meeting, the idea
of an artificial assistant was abandoned. The characteristics of a rewarding meeting were
tightly coupled to chairman tasks and the agenda, and the term artificial chairman was instead
adopted. Typical chairman tasks listed by the respondents involved:

• make sure that all issues are processed


• force the meeting to proceed
• make sure the meeting is going in the right direction
• have responsibility
• make sure the meeting follows the agenda
• serve as an administrative function; start the meeting, end the meeting, take notes, monitor
attendants’ right to speak
• make sure that issues dealt with are acted upon

16
• monitor the structure of the meeting
• make sure the meeting does not get out of hand

3.3.3 Socialising
As was also found in literature (see chapter 2.3.3.4), social aspects were seen as important to
strengthen the work group, and the respondents rated socialising as a purpose of meetings.
However, it was not regarded as an essential part of a rewarding meeting. This demonstrates
an interesting discrepancy. An explanation could be that the social side of meetings is seen as
important in a more abstract sense, that is, it is crucial to the general meeting situation in order
to strengthen the work group, but not as obvious in the formal meeting per se, where the
fulfilment of stated goals decides how rewarding the meeting is perceived. Some respondents
also said that such social issues were better handled outside the conference room in a less
formal situation. This could mean that socialising is more important to make work itself
rewarding, whereas the meeting mainly serves as an arena for productive achievements.
Therefore, it can be concluded that a socialising is not an aspect that is important enough to
qualify as a property of a rewarding meeting, when it comes to the specific meeting situation.

3.4 Conclusion Preliminary Study

Drawing from these results, that showed to be supported by findings in literature regarding
successful meetings described in chapter 2.3.3, a rewarding meeting is from here on defined
and referred to, as one that is efficient and thereby incorporates the following properties:

• Clear structure
- An agenda
- A clear purpose or goal of the meeting
- A set timeframe
- A moderator such as a chairman monitoring the meeting
• Suiting information
• Compliance with efficiency-related norms
- Everyone arriving at the meeting prepared for their task
- Equal participation by all members

3.5 Criticism of Preliminary Study

The occupations of the subjects indicated a possibility that the subjects might have earlier
experience in theories concerning meeting management from university or other education;
for example, two of them were CEOs. Additionally, all but one of the respondents had a
university degree. It is possible that these factors have influenced responses, and
generalisations from these results to other areas of research should be done with caution.

17
4 EXPERIMENT
Interviews in the preliminary study showed that meeting participants’ idea of a rewarding
meeting was to a major part grounded in efficiency, and its constituent parts; structure, suiting
information and compliance with efficiency-related norms. These findings were also
supported by literature. It was clear that a chairman and an agenda were central parts of the
structure, and that it was mainly tasks related to the agenda and the meeting co-ordinator (i. e.
the chairman), that were critical for a rewarding meeting. Therefore, the experiment focused
on the role of the chairman and the agenda.

4.1 Purpose

The objective of the experiment was to find out if a transition from human to computerised
control is feasible. Our hypothesis, that the replacement of the human chairman with an
artificial chairman will not result in a less rewarding meeting, was investigated by examining
meeting participants’ experience and acceptance of different levels of computerisation in a
formal face-to-face meeting.

4.2 Applying the Results of the Preliminary Study

It was believed that most of the requisites for an efficient meeting could be fulfilled by using
tighter meeting control in the form of an artificial chairman strictly following an agenda. To
execute the experiment we needed to incorporate the properties of a rewarding meeting into
the artificial chairman.

By adhering to the properties of a rewarding meeting, the artificial chairman would also
execute the typical chairman tasks described in chapter 3.3.2. This included setting the
timeframe, controlling and working through the agenda, and generally monitoring
discussions.

We believed that the processes of information sharing within the meeting session must be
supported in order to secure suiting information and thereby efficiency. In the experiment,
support for the sharing of information was implemented by using presentation technology
available in the iLounge. To further attain the criteria for suiting information in the meeting,
information about the meeting scenario was distributed to the participants prior to the
experiment. The technology (iTable, Tipple, Filestarter, and SMART Boards) used during the
experiment was meant to serve as tools for the information interchange, and capture attitudes
towards this new presentation technology. It was intended to make it possible to examine the
success of such support.

It was recognised that the social aspects of the meeting play a part in the perception of the
meeting, but characteristics were not explicitly applied to the chairman in order to manipulate
the social meeting environment. However, the possible social effects caused by an artificial
chairman solely dedicated to achieving the above described properties of efficiency, were still
an area of interest.

4.3 Wizard of Oz Methodology

In order to test the hypothesis, the subjects in the experiment had to be presented with an
artificial chairman. Also, to encompass the tasks such a chairman should perform, for

18
example dynamically distributing the right to speak, it had to convey a certain degree of
intelligence.

Instead of technically implementing an agent application, it was decided that a simulation of


an agent application would suffice to successfully execute the test. Hence, there was no agent
technology behind the execution of commands, rather, the agent application was simulated in
real-time by one of the research administrators.

The experiment was carried out within the methodological framework Wizard of Oz
[Dahlbäck et al, 1993]. The essence of this methodology is to simulate a computer service in a
way that makes the users unaware that they are not interacting with a simulation.

The reason this methodological framework was chosen was related to the nature of the
situation examined. Even though a lot of the functionality needed for the study was already
present in the iLounge, as described in chapter 4.8, to fully develop an automated application
with the necessary functions would have taken a considerable amount of time. Moreover, the
ambition of this thesis was not to build a solution to a problem, but rather to examine the
consequences of such a solution. If the same results can be achieved by simulation, there is no
need for complete construction and implementation.

The make believe artificial chairman was named Agnes for no particular reason other than to
make the subjects perceive the simulation as more believable.

4.4 Agnes

Although Agnes did not really exist other than in the eyes of the meeting attendants, a brief
discussion on where she would be placed in the agent hierarchy (see chapter 2.4) is
appropriate. This serves to put Agnes in relation to the previously described levels of
agenthood, in the case she would have been implemented.

Agnes’ main goal was to run an effective meeting. In the literature and the preliminary study
effectiveness was found to be the result of a firm chairman, a clear agenda, set timeframes etc.
Hence, these constituents of efficiency were the sub goals that in concert lead to the fulfilment
of the main goal. The meeting environment was monitored by Agnes by listening to the
meeting attendants and their individual contribution to the meeting. She only interrupted the
meeting if it was needed to further her sub goals, for example to move the meeting to the next
issue on the agenda. Communication with meeting attendants was indirect, by issuing a
command as a reaction to their behaviour. The way Agnes was acted out during the test, there
was no possibility for meeting attendants to ask her direct questions or get her to change her
goals, as they were previously set. However, adaptation to the attendants was made if this put
Agnes’ in a better goal state. For example, if the participants could not agree on a problem
solution, Agnes would signal that the decision should be postponed in order to at least reach a
timeframe goal, but if attendants a minute later suddenly seemed to come close to an
agreement she would issue a command to vote. By doing this she would complete more sub
goals by reaching a decision and eliminating the need for another meeting.

Agnes’ behaviour can therefore be tentatively compared to what Verhagen [2000] refers to as
a goal autonomous agent. Such an agent determines which environmental cues and requests
are in its “prevailing interest” [ibid], that is, get it closer to its goals. The agent itself decides
on what to act; Agnes told people not to interrupt others because equal participation was one

19
of her sub goals. It can not be argued that Agnes belonged to the more complex norm
autonomous agent, since she did not implement norms or put such a concept in relation to her
goals. Furthermore, she could not re-evaluate her goals in order to better fit the situation.

4.5 The Three Research Conditions

Three research conditions were stipulated that represent different levels of computerisation in
meeting management.

1. Human Moderation (HuMod); Chairman controlled meeting


A traditional meeting session moderated by a human chairman.
• No computerised assistance present.
• The chairman followed an agenda not presented to the other attendants.
• The agenda was predetermined by the research administrators and the chairman was told
to follow it strictly.

2. Concealed Moderation (CoMod); Hidden chairman operating through human


chairman
A meeting session where the human chairman was given directions by a computerised
chairman.
• No agenda was available to the chairman or the attendants.
• The commands given to the human chairman by the artificial chairman were
predetermined by the research administrators and the chairman was told to keep strictly to
them.
• The commands were presented to the chairman through a laptop computer, and were not
visible to the other meeting participants.
• The chairman executed orders without revealing their origin to the attendants.

3. Agnes Moderation (AgMod); Openly Agnes-controlled meeting


A meeting session where the human chairman was replaced by Agnes.
• No human chairman was present.
• No agenda was available to the participants.
• Agnes acted as chairman and conducted the meeting, giving orders directly to participants.
• The commands given to meeting attendants were predetermined by the research
administrators.
• The commands were presented to all the meeting participants through a wall-mounted
touch-screen (SMART Board). Each command was followed by a sound that intended to
attract participants’ attention.

The HuMod condition served as a control group, designed to resemble a typical formal face-
to-face meeting. The only difference between the CoMod and AgMod conditions was the
interface towards the participants. The same orders were given, in the CoMod condition
mediated by a human, and in the AgMod condition delivered directly by a seemingly
intelligent application, Agnes. This was done in order to identify differences in participants
experiences, differences that were related only to the perception of the chairman.

The reason for not making the agenda available to all participants was to emphasise the role
of the chairman, since s/he was the only one who knew how the meeting was to proceed. It

20
was evident when conducting the pilot studies that participants were less prone to attending to
the chairman if they were aware of the agenda.

4.6 Subjects

Each condition was examined using three groups of students, each group consisting of four
participants and one chairman, save the third condition where the chairman was replaced by
Agnes. This totals 42 persons across nine groups. The sample of subjects consisted of
volunteers who responded to advertisements. 11 of the subjects were female and 31 were
men. Almost all were students at Stockholm University in Kista, Sweden. They were all
familiar with using computers.

4.7 Meeting Scenario

The experiment simulated a formal face-to-face meeting situation, with an initial information
interchange phase, followed by a concluding discussion and problem solving phase. A
meeting scenario was prepared, where each participant was randomly given a specific role
within a housing co-operative. As stated by the scenario, the reason for conducting the
meeting was to decide how to make use of a shared space in the basement of the house. Each
subject was provided with an opinion on this issue and a set of arguments to support it. The
task given was to present the opinion and the arguments. Material was chosen in order to
evoke opinion and discussion, to make the meeting proceed without the risk of grinding to a
halt. The material each participant was to present had been distributed to some of the subjects
in advance, but not all due to some subjects applying late to participate in the experiment. The
scenario was the same in all conditions.

4.8 Experiment Execution

The experiment was situated in the iLounge, a technologically enhanced meeting room
located at the Department of Computer and Systems Science at Stockholm university, Kista
(see picture 1), and took place on the 7th, 8th, and 9th of October 2002.

Picture 1. The iLounge

21
Technology available in the iLounge was used to stage the experiment:

• SMART Board ™, a touch-screen computer mounted on the wall [www.smarttech.com]


• iTable, a touch-screen computer built into the meeting table [Croné and Jaensson, 2001].
• Tipple, a shared desktop application that presents services to the users [Werle and
Jansson, 2001].
• Filestarter, an application that enables users to share files [ibid].

The iTable was prepared in advance to hold pdf-files containing the arguments that
participants were instructed to present to each other. These files were visibly present to the
participants as file icons at the top of the iTable. When their turn to hold a presentation,
participants could transfer their file to one of the SMART Boards via the Filestarter operating
on Tipple, using direct manipulation, i. e. drag-and-dropping a file icon onto the Filestarter
area located on the SMART Board.

A short presentation of the iLounge and the technology therein was given to the participants
before the start of the meeting. They were instructed in how to use the iTable and the SMART
Boards to present their material, if they desired to do so. To make the situation resemble a
regular meeting, the visual presentation of the arguments was not mandatory. Participants
were also handed a print out of their pdf-file containing the arguments. They were also told
that the room contained a positioning system that could keep track of who was currently doing
a presentation. Four video cameras were positioned to overlook all activities in the room, and
subjects were informed that the meeting was being videotaped, but that the tapes would be
destroyed shortly after the experiment. The positioning system was mentioned to the
participants, with the intention of making the agent technology more convincing, even though
it was actually not in operation during the experiment.

The meeting lasted for about 15 minutes, with each group member, except for the chairman,
being allocated two minutes for sharing of the information. This was followed by five minutes
of discussion in an attempt to reach a joint decision. Since the sharing of information and the
meeting processes were the primary focus of the study, there was little focus on the actual
outcome of the meeting.

The research administrator controlling Agnes was situated in a room adjacent to the meeting
room and was surveilling the meeting via the video cameras and the microphones (see picture
2). Depending on the agenda and the current situation in the meeting, appropriate commands
were executed, so as to make participants believe there was real intelligent technology built
into the system. The administrator was constrained by choosing from a list of 12 predefined
commands (see appendix III). Agnes’ tasks were to open and end the meeting, assign the right
to speak, keep to timeframes, make sure participants stayed on topic and make sure everyone
participated. Commands were designed to reflect these tasks, and resemble the chairman
commands in the agenda used in the HuMod condition. When a situation occurred in the
meeting that demanded chairman moderation according to the tasks, for example someone
talking for too long or straying from topic, a command was issued. Commands were
predetermined phrases, such as “ Person X has the floor”, “Does person X have anything to
add?”, and “Round up the meeting”.

22
Picture 2. Surveilling the meeting and controlling Agnes.

The research administrator posing as Agnes presented her commands to the chairman (in
CoMod) and the participants (in AgMod), using the Microsoft NetMeeting ® application (see
picture 3). The application was manipulated so as to make it impossible for participants to
recognise it as NetMeeting.

Picture 3. Agnes' Interface.

4.9 Pilot Study

Before the main experiment commenced, two pilot studies were performed to try out the
different experimental conditions. These involved a number of employees from the Fuse
research group. Problems in the experimental design were hereby highlighted which led to
some changes.

23
4.10 Data Gathering

After each meeting session, participants were asked to fill out a questionnaire consisting of
approximately 40 questions. Since subjects’ perceptions of the meeting were destined to differ
depending on the condition they belonged to, four types of questionnaires were developed:
• one for chairmen of the HuMod condition
• one for chairmen of the CoMod condition
• one for participators in the HuMod condtion and the CoMod condition
• one for participators in the AgMod condition

There were closed questions with answers ranging from 1 (I totally disagree) to 5 (I totally
agree) on a Likert-scale, and open-ended questions asking subjects to elaborate on specific
questions.

4.10.1 The Questionnaire


To a large extent, the questionnaire developed on the key areas identified in the preliminary
study. Questions served to pinpoint differences in how these areas were perceived in the three
research conditions. Hence, areas were added to assess these differences. These were related
to attitudes towards the chairman and Agnes, as well as trust and acceptance issues.

The question areas were:

• Efficiency
• Structure
• Components of structure
• Efficiency-related norms
• Presentation technology
• Trust and acceptance
• Attitudes towards the chairman
• Attitudes towards Agnes

4.11 Data Processing

Most of the questions were identical for all groups, although some were specific for each
condition. The results for chairmen and participators were separately analysed. However,
some of the questions related to the experience of Agnes and were posed to all subjects who
came in contact with her. As mentioned, since so few groups of individuals were observed,
there was no intention to attain a statistically valid result, therefore evaluation was made
separately for each individual. Non-responses to questions are pointed out in cases where they
occur. Since some follow-up questions were only directed to subjects who agreed to the
particular question statement, these questions rendered a lot of non-responses.

4.12 Results Experiment

Due to the extensive amount of data, not all questions are presented in this section. Focus is
on the results directly related to the hypothesis; that replacing the human chairman with an
artificial chairman will not result in a less rewarding meeting. Hence, emphasis lies on
questions regarding efficiency and its constituent parts, and how participants’ experiences

24
differ depending on the chairman; the traditional human chairman, the controlled human
chairman, and Agnes.

Some results are complemented with graphs showing the distribution of answers among
meeting attendants. For a complete presentation of the results, and all the graphs for meeting
attendants as well as chairmen, see appendix II.

4.12.1 Questions on Efficiency

Q1. I perceived the execution of the meeting as efficient.

1. I percieved the execution of the m eeting as efficient

12
Number of subjects

10
8 HuMod
6 CoMod
4 AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
It is evident that participants of the AgMod condition were more inclined to perceive the
meeting as efficient than the participants of the other two conditions.

Chairmen
There seems to be little difference between chairmen’s perception of meeting efficiency in the
two conditions. None of the answers suggests any dissatisfaction with the efficiency.

Conclusions Q1.
The results suggest that the computerisation did not affect subjects perception of meeting
efficiency in a negative way. On the contrary, in the AgMod condition where the degree of
computerisation was the highest, the perception of efficiency was also the strongest.

Q2. I experienced that the commands of the chairman were followed.

2. I experienced that the commands of the


chairman/Agnes were followed
Number of subjects

12
10
HuMod
8
6 CoMod
4
2 AgMod
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

25
Participants
Participants in the CoMod condition were less convinced that the chairman’s orders were
followed than the other two conditions. All participants of the AgMod condition maintained
that Agnes’ commands were followed, whereas only six participants of the CoMod condition
experienced that the commands of the chairman were followed.

Chairmen
There is a slight indication towards chairmen of the HuMod condition being more content
with the way commands were followed.

Conclusions Q2.
It is interesting to note that all participants in the AgMod condition perceived that Agnes’
commands were followed, whereas CoMod participants were less convinced. There could be
many reasons for this, for example that Agnes was better at capturing the attention of the
participants than the human chairmen. It is also possible that the commands of the human
chairmen were not perceived as orders by the participants to the same extent as those made by
Agnes. Agnes could not respond to reasoning, which might have made the command more
obvious. The unfamiliar situation in the AgMod condition might have affected the behaviour
of the participants, which possibly made them more attentive to commands. However, it is
still possible that Agnes orders were followed because they were clear and precise. Even
though the same orders were given in the CoMod condition, the human characteristics of the
chairmen might have coloured their recitation of Agnes’ orders.

Q4. The meeting would have functioned better without the chairman.

4. The meeting would have functioned better without


the chairman
Number of subjects

12
10
HuMod
8
6 CoMod
4
2 AgMod
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Having a chairman was deemed valuable in all conditions by both chairmen and participants,
and there is very little reason to believe there is any difference between the groups related to
this question.

Conclusions Q4.
It is interesting to note that the participants in the CoMod condition regarded the chairman
important, since not everyone in the CoMod condition thought that commands from the
chairman were adhered to (see Q 2).

26
Q8. It happened on one or several occasions that the meeting became inefficient.

Q9. If you have answered 4 or 5 to the question above, at what point/s did you
experience the meeting as inefficient? What do you think was the reason for
inefficiency?

8. It happened on one or several occasions that the meeting


became inefficient
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
The results stemming from Q 8 show that the first two conditions, especially the CoMod
condition, experienced inefficiency more than the AgMod condition, where only 2 subjects
experienced inefficiency.

The answers to the follow-up question were very diverse in all three conditions, but some
related to insecurities about the scenario and structural weaknesses , for example “ I was
rambling due to poor background information/…/”, ”When we didn’t know how big the
common space [in the scenario] was”, and “Waiting for Agnes to give new instructions”.
Apart from this, there was no specific source for inefficiency revealed in Q9.

Chairmen
Chairmen of both conditions did not experience inefficiency in the meeting. Since only
subjects who experienced the meeting as inefficient were requested to answer the follow-up
question, there were no relevant answers.

Conclusions Q8 and Q9.


Participants in the AgMod condition seemed more content with meeting efficiency. This
could indicate that inefficiency was less frequent with Agnes as chairman.

Q28. I perceived it as the chairman/Agnes had control over the meeting.

Participants
Half of participants in the HuMod and CoMod conditions perceived it as the chairman /
Agnes had control over the meeting. Participants in the AgMod condition were somewhat
more enthusiastic than the others.

Chairmen
All chairmen agreed that the chairman (HuMod), or Agnes (CoMod) had control over the
meeting.

27
Conclusions Q28.
There is very little difference between the conditions, indicating that Agnes had as much (or
little) control as the human chairmen.

Q35. I believe that the chairman/Agnes made the meeting more efficient.

Q36. If you answered 4 or 5 to the question above, why do you think the chairman /
Agnes made the meeting more efficient?

35. I believe that the chairman/Agnes made the


meeting more efficient
Number of subjects

12
10
8 HuMod
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
Participants in general believed that the chairman / Agnes improved meeting efficiency. Only
3 respondents marked lower than 3 (neither nor) in Q 35 indicating negative responses to the
question, all of these belonging to HuMod.

In the follow-up question, the HuMod comments focused on the chairman bringing stability
and structure to the meeting; “ The meeting got started quickly because of the chairman.”,
“He saw to it that everyone spoke and that no one took over”. The comments in the CoMod
were similar; “Otherwise the meeting would have consisted of pointless bickering” and
“Opening the floor, telling when it is time to move on.”. Interestingly enough, the AgMod
comments expressed even more positivity, saying that “We made a unanimous decision
significantly faster than we would have done without Agnes”, “I did not need to TAKE the
floor on my own initiative, instead I was asked to take the floor”, “/…/,it was time efficient
and we made a good decision.”.

Chairmen
All chairmen agreed that the chairman (HuMod), or Agnes (CoMod) made the meeting more
efficient (as opposed to not having a chairman present) by providing more structure to the
meeting. The chairmen that moderated the meeting on command by Agnes were equally
content with Agnes’ ability to achieve efficiency (“I focused more on “managing the meeting”
and avoiding too loose discussions”), as the chairmen who moderated the meeting themselves
were content with their own ability to do so (“The chairman added structure to the meeting”).

Conclusions Q35 and Q36.


The results suggest that computerisation did not affect subjects’ perception of the chairman’s
ability to achieve efficiency.

28
4.12.2 Questions on Structure

Q25. I perceived it as the chairman/Agnes strictly followed an agenda.

Participants
The differences between conditions were small when it came to whether subjects perceived
that the meeting was based on an agenda. A majority of participants in the HuMod and
AgMod conditions thought so. It was to a lesser extent noticed in the CoMod condition.

Chairmen
Two of the CoMod chairmen perceived that Agnes strictly followed an agenda, one did not.

Q26. If you answered 4 or 5 to the question above, did you perceive it as negative that
the chairman/Agnes strictly followed an agenda?

Q27. Can you develop your answer to the question above?

Participants
Of those applicable to answer Q26, in HuMod and AgMod one person found the enforcement
of an agenda as negative, and in CoMod two persons. The differences are therefore small.

In the follow-up question, it is worth noticing that all four comments made by CoMod
participants complained that the tight structure made the meeting “inflexible”, “stiff” and “too
formal”, while most (7 out of 10) comments from AgMod were positive saying that Agnes
“got the meeting going” and facilitated “quick decisions” when time was limited. The seven
answers from the HuMod condition were mixed with no general inclination.

Chairmen
None of the chairmen perceived it as negative that Agnes strictly followed an agenda. Two of
the chairmen did not answer Q27. The third thought that Agnes was “good to have as a
guiding line” but that s/he perceived it as “Agnes seemed to know who was talking” and
“interacted”, implying that Agnes did not actually follow the agenda.

Q31. I thought the meeting was affected in a negative way by the chairman’s/Agnes
directions.

31. I thought that the meeting was affected in a


negative way by the chairman's/Agnes' directions
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

29
Participants
Most participants in all conditions did not think the meeting was affected in a negative way by
the chairman’s /Agnes’ directions. As in Q 26 the CoMod had the most problems with
directions from the chairman/Agnes. The AgMod condition showed least dissatisfaction.

Chairmen
All of the chairmen were neutral on this item.

Conclusions on Structure, Q 25, Q 26, Q 27, and Q 31.


It seems that subjects in all conditions perceived that there was an agenda behind the meeting,
even though it was not presented to them. Also, most of them seemed pleased with the
agenda. There is a slight indication that the participants in the CoMod condition were less
satisfied. Maybe, since the chairman in the CoMod condition did not receive the agenda, s/he
could not mediate it to the participants in the same way as the HuMod chairmen, who where
aware of the agenda in advance, and Agnes in the AgMod condition, who acted out the
agenda.

4.12.3 Questions on Components of Structure


These questions relate to key components in the concept of structure identified in the
preliminary study; clarity in goal and timeframes. Questions regarding focus and discipline in
the execution of the meeting were asked since they are strongly related to the structure.

Q10. I thought the goal of the meeting was clear.

Participants
In both the HuMod and the AgMod condition, all subjects (12 of 12) thought the goal of the
meeting was clear. In the CoMod condition, 8 of 12 believed so.

Chairmen
All chairmen were more or less content with the clarity of the goal of the meeting.

Q11. I thought the timeframes of the meeting were clear.

Participants
The CoMod condition participants were slightly less inclined to believe the timeframes of the
meeting were clear than the other conditions. There was no difference between the HuMod
and AgMod conditions.

Chairmen
All chairmen were more or less content with the clarity of the timeframes of the meeting.

30
Q12. I thought the structure of the meeting was clear.

12. I thought the structure of the meeting was clear

12
Number of subjects

10
8 HuMod
6 CoMod
4 AsMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
Since only one person did not perceive the structure as clear in the AgMod condition, this can
be said to be the condition where the structure was most obvious to participants. Least
structure was perceived in the CoMod condition, where only 3 of 12 agreed to the question.

Chairmen
All chairmen were equally content with the clarity of the structure of the meeting.

Q13. I thought the meeting was focused.

13. I thought the meeting was focused

12
Number of subjects

10
8 HuMod
6 CoMod
4 AsMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
Meeting focus was most obvious to participants in the AgMod condition, and, again, least
obvious to CoMod participants.

Chairmen
All chairmen agreed that the meeting was focused.

Q14. I thought the meeting was disciplined.

Participants
A majority of the participants in the AgMod and HuMod conditions experienced meeting
discipline, whereas half of the CoMod participants did so.

Chairmen
All chairmen equally agreed that the meeting was disciplined.

31
Conclusions on Components of Structure Q 10 – 14.
There is a general tendency throughout all questions (Q 10 –14) that the AgMod condition
ranked highest in agreement to the question statements, and that the CoMod condition gave
the lowest estimates. The HuMod condition resembled the AgMod condition in their
assessments, but with a slightly less positive inclination.

It seems that goal, timeframes, clarity of structure, focus, and discipline were most obvious in
the AgMod condition and least obvious in the CoMod condition. However, conclusions
regarding clarity in timeframes can not be easily drawn, due to the fact that far from all
participants were beforehand aware of the meeting timeframes, while some knew exactly the
time they had at their disposal.

4.12.4 Questions on Efficiency-related Norms


Efficiency-related norms were studied by looking at the degree of equal participation in the
meeting. The other aspect of these norms, coming prepared to the meeting, was not studied
due to difficulties in controlling for this variable. As mentioned, some subjects applied late to
the experiment, and did not have the chance to read the material distributed prior to the
meeting.

Q5. All meeting attendants participated equally in the meeting.

5. All meeting attendants participated equally in the


meeting
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
The AgMod subjects seemed slightly more content with the degree of equal participation,
than the other participants.

Chairmen
All chairmen of the HuMod condition were content with the degree of equal participation. In
the CoMod condition, two were openly discontent.

32
Q6. The chairman / I as a chairman facilitated the participation in the meeting.

Q7. Can you develop on your answer to the question above?

6. The chairman / I as the chairman facilitated the


participation in the meeting
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
AgMod condition participants clearly felt pleased with the degree to which the chairman
/Agnes facilitated the participation in the meeting.

When asked to further develop on Q 6, the written answers in the HuMod condition showed a
spread in opinions about the chairman, because of the his/her way of conducting the meeting.
For example, “During our meeting, the chairman was relatively passive and did not influence
much” as compared to “Yes, by letting everyone speak! Otherwise some people will not say a
thing while others take command of the whole meeting”.

The CoMod condition was much more unified than the HuMod condition. Most (8 out of 12)
participants mentioned in their written answers that the chairman actively pointed out whose
turn it was to speak, although some questioned why the chairman only behaved as a
moderator and had no opinions of his/her own.

All but 2 participants in the AgMod condition thought that Agnes added a structure to the
meeting in a positive manner. However, there was a general agreement that Agnes pushed the
meeting forward only allowing the speaker a certain amount of time before passing the word
on to the next person. “You were appointed a certain amount of time for your arguments”, ”
We were held rather short so that everyone got to speak, which was good, the time given for
discussion was a bit short but we still reached a good solution.”, but “I think we as a group
were anxious to follow the rules of the game, we didn’t want to interrupt each other. It is
doubtful that we would have been as willing to co-operate had it been a real meeting” were
some of the reflections made.

Chairmen
Answers to Q 6 do not differ a lot between conditions. Judging by the open question, the
chairmen of the HuMod condition all seemed to believe that their contribution facilitated
equal participation in the meeting (“The chairman distributed time in a fair manner and
created opportunities for all to partake”), whereas the chairmen in the CoMod condition
seemed more critical of whether they achieved this. As one chairman in the CoMod condition
put it, “I perceived it as some participants dominated the discussion”.

33
Conclusions on Efficiency-related Norms Q 5, Q 6 and Q 7
Participants in the AgMod condition were more content with the equal participation, and were
positive in their judgements of Agnes as a facilitator of the meeting. The HuMod participants
were least satisfied with the chairman. CoMod participants were least content with equal
participation, but agreed that the chairman improved participation. The reason for the slight
difference between the AgMod and CoMod conditions might be that AgMod participants
immediately perceived Agnes’ commands to change speaker and acted on them, whereas the
chairmen in CoMod often waited to execute a “change speaker”-command until the person
currently speaking was finished, making the timeslots for participants less equally distributed.

When comparing the results to Q 5, it is evident that the two chairmen who were discontent
with the degree of equal participation still believed that they as chairmen improved the
participation, implying that the degree of equal participation would have been even smaller
had they not been there.

4.12.5 Questions on the Presentation Technology


These questions were asked since it was mentioned in the preliminary study that the
characteristics of the sharing of information were important to the perception of the meeting
as rewarding. Based on the fact that some participants did not use this technology, and when it
was used it did not always function properly, conclusions on these questions will only be
briefly discussed. It is important to remember that the chairmen themselves did not use any
presentation technology.

Q22. I thought the presentation technology (Tipple, iTable, SMART Boards) made the
meeting more efficient.

Participants
Only two of the participants in the CoMod condition thought the presentation technology
made the meeting more efficient, and were therefore much harder in judging the technology
than subjects in the other conditions, where half of the participants believed so.

Chairmen
Chairmen of the HuMod condition generally seemed more convinced that the presentation
technology made the meeting more efficient than the CoMod chairmen.

Q23. I thought the presentation technology (Tipple, iTable, SMART Boards) was better
than the presentation technology I normally use.

Participants
Again, participants in the CoMod condition were harder in judging the technology than
subjects in the other conditions.

Chairmen
Chairmen of the HuMod condition generally seemed more positive towards the presentation
technology, and the CoMod chairmen were more negative.

34
Q24. I thought the presentation technology (Tipple, iTable, SMART Boards) facilitated
the sharing of information.

Participants
This time, participants in the CoMod condition were only slightly harder in judging the
technology. The biggest difference is to the AgMod condition, where no one disagreed to the
question asked.

Chairmen
Again, the discrepancy between the conditions is large when it comes to the acceptance of the
presentation technology among the chairmen, showing more satisfaction in the HuMod
condition.

Conclusions on Presentation Technology Q 22, Q 23, Q 24


An interesting attitude identified was that participants in the CoMod condition were generally
more negative towards using the presentation technology in the particular scenario than the
other respondents. Chairmen were more positive than participants. It is probable that the
problems with the technology mentioned earlier has affected the results. The limitations in the
experiment design make it difficult to draw conclusions regarding functioning support for
sharing of information related to computerisation.

4.12.6 Questions on Trust and Acceptance


Direct questions regarding trust and acceptance towards the chairman were asked, in order to
find out if there are differences in the perception regarding trust and acceptance attitudes
towards the chairman / Agnes. They were posed to all subjects except chairmen of HuMod
conditions.

Q17. I felt respect for the chairman/Agnes.

17. I felt respect for the chairman/Agnes

12
Number of subjects

10
8 HuMod
6 CoMod
4 AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
Most participants in all conditions claim to have felt respect for the person in charge of the
meeting. Five people answered “I totally agree” in the AgMod condition, compared to two
people in each of the other two conditions.

Chairmen
All chairmen felt respect for Agnes.

35
Q18. I was reluctant to following orders from the chairman/Agnes.

Participants
There seemed to be no problem for the participants in either condition to follow the orders of
the chairman, whether it was human or artificial.

Chairmen
None of the chairmen were reluctant to following orders from Agnes.

Q20. I perceived it as negative not being allowed to talk as much as I wanted to (only
answer this question if you at any time during the meeting were interrupted by the
chairman/Agnes.

Participants
Only people who at any time during the meeting were interrupted by the chairman (in the
HuMod and CoMod conditions) or Agnes (in the AgMod condition) were requested to answer
this question. Hence, there were a lot of non-responses. Of those applicable to answer, 1 of 5
in the HuMod condition, 3 of 8 in the CoMod condition, and 6 of 9 in the AgMod condition
perceived it as negative not being allowed to talk as much as they wanted to.

First of all, answers to this question showed that the number of subjects who perceived
themselves to be interrupted increased with the degree of computerisation. Second, higher
degree of computerisation also seems to increase the participants discontent with being
interrupted.

Chairmen
The chairmen that answered (2 of 3) were neutral.

Q21. I perceived it as negative being encouraged to talk (only answer this question if you
at any time during the meeting were encouraged to talk by the chairman/Agnes.

Participants
Of those applicable to answer this question, none of 8 in the HuMod condition, 1 of 9 in the
CoMod condition, and 1 of 11 in the AgMod condition perceived it as negative being
encouraged to talk. Answers to this question show that the number of subjects who perceived
themselves to be encouraged to talk increased with the degree of computerisation. However,
as opposed to in Q 20, a higher degree of computerisation does not seem to increase the
participants discontent with being encouraged to talk. Rather on the contrary, more subjects in
the AgMod condition answered in disagreement to the question statement than in the other
two conditions.

Chairmen
Of the chairmen, none of the subjects perceived it as negative being encouraged to talk.

Conclusions on Trust and Acceptance, Q 17, Q 18, Q 20 and Q 21.


The results for participants may imply that Agnes was able to induce more respect than the
human chairmen, or at least a stronger sense of respect amongst those persuaded by her
authority in the first place.

36
The fact that more subjects perceived themselves to be interrupted / encouraged to talk with
an increased degree of computerisation might mean that the human chairmen did not interrupt
/ encourage participants to the same extent as Agnes did. It is also possible that subjects in the
AgMod condition talked more, thus being more disposed to the chance of being interrupted.
This, however, can not explain why they perceived themselves to be encouraged to talk more
often, which would only happen if they talked less than the others.

An alternative explanation could be that subjects did not perceive themselves as being
interrupted or encouraged to talk when it was done by a human, as when interruption or
encouragement was made by Agnes. It is possible that human chairmen were better at
disguising the commands, for example; saying “thank you, maybe we should move on”
instead of only saying “next speaker”. It could also be that it is more accepted to be
interrupted by a human, than by a computer, maybe because this is expected from a human
chairman but not from a text based computer interface. However, being encouraged to talk
was not perceived as negative in any condition.

4.12.7 Questions on Attitudes Towards the Chairman / Agnes

Q29. I perceived the chairman’s/Agnes requests as easy to understand.

Participants
Almost all participants in all conditions perceived the chairman’s /Agnes’ requests as easy to
understand. In fact, all AgMod subjects gave Agnes’ requests a “top grade” in clarity.

Chairmen
All chairmen perceived Agnes’ requests as easy to understand.

Q33. I would have liked to communicate with the chairman / Agnes in another manner.

Q34. If you answered 4 or 5 to the question above, how would you have preferred to
communicate with Agnes?

Participants
7 of 11 of the participants in AgMod wished for a different type of communication, but the
other conditions did not.

The few answers in the HuMod and CoMod conditions to Q34 did not directly relate to the
question posed (the communication with Agnes), but rather referred to ways of easing up the
structure of the meeting. The AgMod condition differed since it had no human chairman and
only one-way communication, therefore the results were different from the other conditions.
AgMod answers regarded technical ways of interacting with Agnes. The majority (6 out of 8)
wanted to communicate via voice commands with for example a possibility to ask for more
time. Only one person asked for a “/…/graphic 3D avatar/…/”.

Chairmen
None of the chairmen wanted to communicate with Agnes in another way.

37
Conclusions on Attitudes towards the Chairman / Agnes, Q 29, Q 33 and Q 34.
The AgMod participants were most content with the clarity of requests, but they still wanted
to change the way they communicated with Agnes. They did not like the idea of one way-
communication.

4.12.8 Questions on Attitudes Towards Agnes, Q 39-45.


The following questions were only posed to participants of the AgMod condition and
Chairmen of CoMod condition, since they concerned the perception of Agnes.

Q39. I perceived it as negative not being able to talk to Agnes.


Most participants (9 of 12) of the AgMod condition perceived it as more or less negative not
being able to talk to Agnes. Only one person did not see this as a problem, while two subjects
were neutral. Amongst the chairmen on the other hand, no one agreed that lack of ability to
talk to Agnes was problematic.

Q40. I believe that the meeting would have functioned better with a human chairman.

Q41. If you answered 4 or 5 to the question above, why do you believe that the meeting
would have functioned better with a human chairman?

4 of the 12 participants believed that the meeting would have functioned better with a human
chairman, whereas 4 did not believe so. It is therefore not possible to identify a specific
inclination of attitudes. Human qualifications like being able to “see and listen to
participants”, “collaborate with people” and “sum up people’s opinions” were desired. Three
participants implied that they missed the superior role that a chairman often has in a meeting,
“having the last say”, while one person actually liked the idea of not directing the attention to
an “omnipotent” person. One person stated that with a human chairman with the same
authority that Agnes had “more people would have felt satisfied after the meeting, and not as
stressed out”. One of the chairmen believed that the meeting would have functioned better
with a human chairman, saying that a human chairman would make the meeting “more natural
and less artificial”. Two chairmen were neutral.

Q42. I experienced it as negative receiving commands from Agnes.


10 of 12 subjects did not experience it as negative receiving commands from Agnes. None of
the chairmen perceived it as negative and one did not answer.

Q43. It was easy to follow Agnes’ commands.


All participants considered it easy to follow the commands given by Agnes. Of chairmen, two
thought it was easy while one did not.

Q44. It would have been easier to administer the meeting without Agnes.
9 of 12 participants did not agree that it would have been easier to administer the meeting
without Agnes. One person did agree, while two were neutral. Two chairmen were neutral,
one thought a meeting without Agnes would not have been easier to administer.

Q45. How do you believe that the technology behind Agnes works?
As a control question to detect if subjects believed that there was real agent technology
operating behind Agnes, it was asked how subjects believed that Agnes was technically
implemented. Only one of the participants and one of the chairmen guessed, correctly, that
Agnes was in reality operated by a person.

38
Conclusions on Attitudes towards Agnes Q 39-45.
Overall, subjects seemed pleased with Agnes as a chairman, however, they wished for ways
of enabling two-way communication with Agnes. Other human qualifications were also
requested by some, such as seeing, listening and summing up arguments and ideas. Basically,
the interaction was the part they were least satisfied with. They had no problem with receiving
commands from Agnes. In all, Agnes was deemed a valuable component in the management
of the meeting.

4.13 Summary of Conclusions

The most important experimental findings regarding each question area in chapter 4.10.1 are
presented in this section. Focus is on comparing the results of the CoMod and the AgMod
conditions.

4.13.1.1 Efficiency
Questions on efficiency indicate that the computerisation did not affect subjects perception of
meeting efficiency in a negative way. On the contrary, in the AgMod condition where the
degree of computerisation was the highest, the perception of efficiency was the strongest.
Occurrences of inefficiency was less frequent in the AgMod condition, and most common in
the CoMod condition.

All participants in the AgMod condition perceived that Agnes’ commands were followed,
whereas participants in the CoMod condition were the least convinced that the chairman’s
orders were followed. However, participants in the CoMod condition still regarded the
chairman important.

There is very little difference between the conditions regarding the chairman’s control of the
meeting, indicating that Agnes had as much (or little) control as the human chairmen. The
results suggest that computerisation did not negatively affect subjects’ perception of the
chairman’s ability to achieve efficiency.

4.13.1.2 Structure
A meeting agenda was not presented to the participants. Instead, the agenda was conveyed by
Agnes through her behaviour, to provide a structure for the meeting. Subjects in all conditions
perceived that there was an agenda behind the meeting, but CoMod did so to a lesser extent.
Most participants seemed pleased with the agenda. There is a slight indication that the
participants in the CoMod condition were less satisfied.

4.13.1.3 Components of Structure


On each of the components of structure; goal, timeframes, clarity of structure, focus, and
discipline, the AgMod participants were most positive. The HuMod condition resembled the
AgMod condition in their assessments, but with a slightly less positive inclination, while the
CoMod participants were least positive.

4.13.1.4 Efficiency-related Norms


Participants in the AgMod condition were more content with the equal participation, and were
positive in their judgements of Agnes as a facilitator of the meeting. The HuMod participants
were least satisfied with the chairman. CoMod participants were least content with equal
participation, but agreed that the chairman improved participation.

39
4.13.1.5 Presentation Technology
Participants in the CoMod condition were generally more negative towards using the
presentation technology in the particular scenario than the other respondents. Chairmen were
more positive than participants. It is probable that the problems with the technology
mentioned earlier has affected the results. The limitations in the experiment design make it
difficult to draw conclusions regarding functioning support for sharing of information related
to computerisation.

4.13.1.6 Trust and Acceptance


All participants reported to have felt respect for the chairman. The fact that AgMod answers
were most positive may imply that Agnes was able to induce more respect than the human
chairmen, or at least a stronger sense of respect amongst those persuaded by her authority in
the first place.

More subjects perceived themselves to be interrupted or encouraged to talk with an increased


degree of computerisation. Being encouraged to talk was not perceived as negative in any
condition.

4.13.1.7 Attitudes Towards the Chairman / Agnes


In all three conditions, the participants had no problem with receiving commands from the
chairman. All participants found chairman requests easy to understand. Every one of the
AgMod participants gave Agnes a “top grade”. Overall, AgMod subjects seemed pleased with
Agnes as a chairman, however, they wished for ways of enabling two-way communication
with Agnes. Other human qualifications were also requested by some, such as seeing,
listening and summing up arguments and ideas. Basically, the interaction was the part they
were least satisfied with. In all, Agnes was deemed a valuable component in the management
of the meeting. The conditions with a human chairman made no objections regarding the
communication.

4.14 Criticism of Experiment Execution

4.14.1 Group Composition


As mentioned in chapter 2.1, artificiality is a recurring problem when studying groups,
something that has not been controlled for in this study. In view of the fact that the subjects
had no prior experience of working together as a group, we can not argue that this
experimental setting is equal in all aspects to an actual meeting session. Studies of group
dynamics show that authentic workgroups change and develop over time, something that the
groups in this study have had no chance to do. Being part of a group makes people more
aroused, more stressed and more tense than they are as individuals [Myers, 1999]. These
kinds of tensions are especially common in new groups, when group members test out their
environment and observe each other's personalities [Crook, 1961].

4.14.2 The Scenario


Due to the experiment design it was not possible for the chairman to answer any scenario
specific questions, like for example “How many persons besides us live in this building?”. At
times the chairmen solved these situations by improvising, but when the chairman did not
answer, or shrugged, the other attendants often seemed puzzled. This can be seen as a
shortcoming in the experiment design, lessening the ability to draw parallels to an authentic
meeting situation. It was however a deliberate choice to try to suppress opinions made by the

40
human chairmen, since it was thought more important to keep all conditions and the
behaviour of both human and artificial chairmen as similar as possible.

A lot of the subjects perceived that the scenario was unrealistic, and that the timeframes of the
meetings were too tight. The fact that the agenda was set up by someone other than the
chairman might have affected the behaviour of the chairman, and thereby the participants’
perception of the situation.

4.14.3 Sample of Subjects


Most of the subjects were computer science students, and are therefore not representative for
any larger group lessening the possibility to generalise from the results. Being computer
science students, participants might have volunteered to the experiment out of interest for the
iLounge environment, making them more positive in their judgements of Agnes.

The small number of subjects, especially in regard to the chairmen, since there were only six
of them, lessens the possibility of drawing conclusions based on the material. This is also the
case where the phrasing of the question leads to a large number of non-responses. The small
number of subjects also means that individual differences in meeting participants might have
had impact on the results, meaning that the results may be influenced by the personalities of
the subjects. This is particularly critical for chairmen, since their role was so important for the
execution of the meeting. Some chairmen were more dominant than others, and some
participants were more determined to stick to their opinion than others.

4.14.4 Drawbacks in Experiment Execution


Caution should be taken when analysing the results to questions related to the presentation
technology, since not all subjects used the presentation technology available, and when it was
used, it often led to technical breakdowns. There were at times problems getting files from the
iTable to open on the SMART Boards, instead an error message occurred on the screen. In
addition, since none of the subjects had earlier experience with the touch screen interface on
the iTable, a lot of times subjects dropped their file and had to start over again. These
breakdowns might have discouraged subjects from using the technology.

It is possible that the chairmen in the HuMod condition felt more secure. They had
beforehand been given all the information they needed to run the meeting, whereas the
chairmen of the CoMod condition received orders in real time. As a result, chairmen in the
HuMod condition phrased the commands themselves, whereas the chairmen in the CoMod
condition only repeated the commands on the computer in front of them. Even though the
same orders were given in the CoMod condition, the human characteristics of the chairmen
might have coloured their recitation of Agnes’ orders, thus making the CoMod condition
differ. It is possible that this uncertainty affected the meeting situation and the other
participants in a negative way, which might explain the overall negative attitude in the
CoMod condition.

4.14.5 Other
Since concepts like efficiency, structure, focus, discipline, etc. were not specifically defined
for the respondents, it is hard to draw conclusions as to whether subjects perceive them as
positive aspects. Based on the fact that most respondents in the preliminary study (appendix I)
did find them desirable, it is likely that these subjects also did so.

41
The AgMod condition differed from the other two conditions, primarily in that there was no
human chairman. To instead get instructions from what seemed to be an intelligent
application was an unusual situation for the subjects, and they might not have reacted as they
did, had they been accustomed to this way of conducting a meeting. Re-tests would be
necessary to investigate whether this is a novelty effect that might wear off with more
experience in this type of meeting.

The only person present during the AgMod condition that was not an active participant was
one of the experiment administrators. His role during the experiment was to brief participants
and answer questions, and during the experiment to observe and correct possible technical
problems. Also, his presence served to augment the illusion of Agnes’ abilities by implying
that there was no one directly controlling Agnes’ output. However, it is possible that his
presence to some extent made the participants behave and answer the questionnaire in a less
critical manner, answering what they thought the experiment administrator wanted to hear2.

We acknowledge that the outcome of a meeting is an important factor to make it rewarding.


The meeting scenario was of an artificial nature where participants had no personal
involvement, which would not be the case in a real life meeting. Therefore, we believe that
participants would have found it hard to evaluate whether the outcome of the meeting
contributed to the perception of the meeting as rewarding. Additionally, the focus of this
study is not on the processes of problem solving in a group. Therefore, the fact that the
opinions of the roles given differed so much that most groups never reached a joint decision,
probably did not severely affect the results.

2
Good-subject tendency

42
5 Discussion
Generally, the participants belonging to the AgMod condition were more positive towards the
way their meeting was carried out, as compared to the CoMod condition in particular, but also
the HuMod condition. They were more pleased with meeting efficiency in terms of equal
participation and two of the components of structure, namely focus and discipline. This means
that computerisation of the structure did not affect efficiency in a negative way. On the
contrary, the results show that focus and discipline in fact were highest in the AgMod
condition. It is also interesting to note that the AgMod participants were more positive to the
way the agenda was implemented than the other conditions. We believe that these positive
results regarding the AgMod condition were mainly caused by the tighter structure enforced
by Agnes.

CoMod participants ranked lower in almost all areas, they did not perceive that orders from
the chairman were followed, they were less content with the efficiency, and they were less
satisfied with the presentation technology. As mentioned, the CoMod and AgMod conditions
differed a lot from the kind of meeting settings that most people are used to today, which
might have affected the participants. However, if this were the only explanation to the
negative remarks, the differences between the last two conditions would not be as great as the
results show. The explanation might be that the participants in the CoMod condition
perceived the meeting as a normal meeting moderated by a chairman, when it was in fact
moderated by an artificial chairman. Thus, when the human chairman did not behave as they
expected, like not responding to direct questions, they were puzzled and distressed.

McGrath [1994] points out that “the more a group’s activity is being affected by closely held
information, with each member’s interaction being channelled by information available only
to that member, the more the individual group members are likely to feel that the others have
hidden agendas and are manipulating the group.” As a consequence, the group may
experience trust issues among group members and work less effectively. Relating this to the
experiment, the fact that the chairman in the CoMod condition seemed to have disclosed
information might have affected the CoMod participants.

Research shows that people unconsciously and automatically behave socially towards
computers in a number of situations [Nass et al, 1997]. In fact, users respond the same way to
computer personalities as they do to human personalities, and they even apply politeness
norms to computers, perhaps even expecting politeness in return [Friedman, 1997].
Interestingly, this is true even in situations where the subject interacting with the computer
states it wholly inappropriate to respond to computers in such a way, and the computer system
is not very sophisticated, using only text-based output. Any word-based interface may thus
have a personality in the eyes of the users, construed by the language, prompts, navigation,
level of interactivity and error messages used in the system [ibid]. This means that the
participants’ perceptions of Agnes’ qualities can be compared to the perceptions of qualities
in the human chairmen.

The view of the chairman/Agnes was not affected much by the computerisation; Agnes
seemed to infuse as much respect, was able to control the meeting to the same degree, and
was considered equally important to the meeting as the human chairman.

Judging from the results we conclude; replacing the human chairman with an artificial
chairman will not result in a less rewarding meeting.

43
The tendency to attribute human qualities to computers described above also applies if an
artificial agent uses natural language, making people expect the agent to behave in a much
more sophisticated manner than it was originally programmed to do [Friedman, 1997]. An
interface, such as Agnes’, should therefore be constructed so that people do not project human
qualities to the agent, when it can not fulfil the expectations. It could be argued that the reason
for the positive attitude among the AgMod participants can partly be related to their optimistic
view of Agnes’ capabilities.

The possible attribution of human traits on to Agnes may also explain the somewhat negative
attitude towards Agnes’ abilities to perform typical human tasks, like listening, reasoning and
communicating, among the AgMod participants. However, these negative remarks concerning
Agnes’ communication skills could also imply that the agent implementation made in the
experiment is not optimal, even though it did not negatively affect the efficiency.

Therefore, we propose that the best use for meeting support such as Agnes is more in line
with the initial approach formulated in the preliminary study, to act as a support for the human
chairman rather than replacing him/her.

This is very much in accordance with a suggestion by Viller [1993] that group facilitation is a
human task that should not be carried out by a computer system alone. However, he argues
that the role of facilitator can be augmented through the use of computer support, and some
facilitator tasks can be automated. Support for decision-making procedures is mentioned as
one role of the facilitator that is possible to automate, as well as monitoring of under and over
participation of group members. In synchronous co-located meeting support systems, the
facilitator’s main task is to monitor communication that could possibly be detrimental to the
group’s working together as a unit.

It should be remembered that the groups in the experiment differ from real organisational
meeting groups, as discussed in chapter 2.1. Applying the results on such groups may
therefore pose a problem. Real meetings can often be more dynamic, requiring a more flexible
structure and an ability to adjust to new situations, something that it is uncertain if Agnes can
achieve as chairman. There is a risk that participants’ respect for Agnes might decrease due to
her limited flexibility, leading to irritation and participants starting to ignore Agnes. This
further reinforces our belief that the agent should act as a support for a human chairman, who
would have the option to evaluate which of Agnes’ suggestions to follow. Additionally,
Agnes could allow for customisations related to specific group properties, making her more
flexible. For example, if Agnes were to be used in a brainstorming-meeting, it would be
convenient if she were more accepting of deviations from the meeting topic.

We have not studied the consequences of the potential novelty effect mentioned in chapter
4.13.5. It is possible that an increasing experience in using an artificial chairman such as
Agnes might affect participants’ attitudes. It is hard to conclude what attitude changes would
occur, but it is likely that they would be the same as discussed in the previous section
concerning real organisational groups. However, we believe that these negative attitudes are
less likely to occur if Agnes is used mainly as support for meeting structure in concert with a
human chairman.

44
5.1 Future Work

As stated, we believe that to further enhance efficiency and thereby the chance of achieving a
rewarding meeting, this type of intelligent meeting support should preferably aim at enforcing
the structure of the meeting, such as the agenda, rather than striving towards totally replacing
the human chairman. As mentioned, this type of meeting support can not easily replace
human qualities in interaction and communication. In more complex meeting situations the
artificial chairman could function more as a support for the human chairman, but in that case,
it should not be hidden as in the CoMod condition. All meeting attendants should be aware of
its presence, capabilities and limitations.

In future experiments looking at rewarding meetings, we believe that the experimental design
should also take the outcome of the meeting, and its quality, into evaluation.

5.2 Related Work

Apart from the ideas of Viller [1993], we have not been able to find any related research
similar to the work in this thesis.

45
6 ACKNOWLEDGEMENTS
We would like to thank Harko Verhagen, PhD, for supervising this thesis and Patrik Werle,
PhD student, for invaluable support, especially during the iLounge experiments. Also, we are
grateful towards the staff at the Fuse group that participated in the pilot study and provided us
with important feedback.

46
7 REFERENCES
Andrews, P. H, Group Conformity. In Small Group Communication: Theory and Practice (7th
ed.), R. S. Cathcart, L. A. Samovar, & L.D. Henman (Eds.), Brown and Benchmark, 1996.

Bradshaw J. M., Software Agents, The MIT Press, 1997.

Brewer M. B. and Crano W. D., Social Psychology. West Publishing Company, 1994.

Burger, J. M., Increased Performance With Increased Personal Control: A Self-presentation


Interpretation, In Journal of Experimental Social Psychology, 23, 1987.

Crook, R., Communication and Group Structure, In Journal of Communication, 11, 1961.

Croné, M. and Jaensson, T., Design of an Interactive Table, DSV, Stockholm University,
2001.

Dahlbäck, N., Jönsson, A., and Ahrenberg, L., Wizard of Oz Studies – Why and How,
Department of Computer and Information Science, Linköping University, 1993.

Engleberg, I. N. & Wynn, D. Working in Groups, Communication Principles and Strategies,


Houghton Mifflin Company, 1997.

Etzioni, O. and Weld D. S., Intelligent Agents on the Internet: Fact, Fiction and Forecast, In
IEEE Expert 10(4), 1995.

Fisher, B. A., Decision Emergence: Phases in Group Decision-Making. In Speech


monographs, 37, 1970.

Franklin S. and Graesser A., Is It an Agent or Just a Program? A Taxonomy for Autonomous
Agents, In Proceedings of the Third International Workshop on Agent Theories,
Architectures, and Languages, 1996.

Friedman, B., Human Values and the Design of Computer Technology, Cambridge University
Press, 1997.

Granström, Small Group Studies; Proceedings From a Conference on Group and Social
Psychology, Linköping University, 1998.

Hackman J. R., Brousseau K. R., and Weiss J. A., The Interaction of Task Design and Group
Performance Strategies in Determining Group Effectiveness. In Organizational Behaviour
and Human Performance, 16, 1976.

Hall, J. and Williams, M. S., Group Dynamics Training and Improved Decision Making. In
Journal of Applied Behavioral Science, 6, 1970.

Hare, P., Groups, Teams, and Social Interactions, Praeger, 1992.

Jay, A, 1976, How to Run a Meeting, In Harvard Business Review, 1976.

47
Jewell, L. N., & Reitz, J. H., Group Effectiveness in Organizations, Scott Foresman and
Company, 1981.

Kernberg, O., Regression in Groups: Some Clinical Findings and Theoretical Implications. In
Journal of Personality and Social Systems 2, 1980.

Kyng, M., Communications of the ACM, 34, Association for Computing Machinery, Inc,
1991.

Lantz, A., Meetings in a Distributed Group of Experts Comparing Face-to-Face, Chat, and
Collaborative Virtual Environment., CID, KTH, 2000.

Larson C. E. &. Lafasto, F. M. J., TeamWork: What Must Go Right / What Can Go Wrong,
Sage Publications, 1989.

Laughlin P. R., Collective Induction: Group Performance, social combination process, and
mutual majority and minority influence. In Journal of Personality and Social Psychology 54,
1988.

Lennéer, B. The Psychology of the Work Group, Stockholm Joint Industrial Safety Council,
1998.

Lesbie, L. Rhoades, J. & MacGrath, J., Interaction Process in Computer Mediated and Face-
to-Face Groups. In International Journal of Computer Supported Co-operative Work. Vol 4,
No. 2-3, Kluwer Academic Press, 1995.

Maher, M. L., Simoff, J. S., Cicognani, A., Understanding Virtual Design Studios, Springer,
1999.

McGrath, J., Groups Interaction and Performance Prentice-Hall, 1984.

McGrath, J., Time Interaction and Performance, In Small Group Research, Vol 22, Sage
Publications Inc., 1991.

McGrath, J., Groups Interacting with Technology : Ideas, Evidence, Issues, and an Agenda,
Sage, 1994.

Myers, D. G., Social Psychology, Boston McGraw-Hill College, 1999.

Napier, R. W. and Gershenfeld, M. K., Groups; Theory and Experience (5th ed,), Houghton
Mifflin Company, 1993.

Nass, C.I. et al, Computers are Social Actors. In Human Values and the Design of Computer
Technology, Friedman, B. (Eds) Cambridge University Press, 1997.

Parks, C. D. & Sanna, L. J., Group Performance and Interaction, 1999.

Phillips, G. M., Communication and The Small Group, Bobs-Merrill Company, Inc., 1973

Poole, M. S. & Roth, J., Decision Development in Small Groups, V: Test of a Contingency
Model. In Human Communication Research, 15, 1989.

48
Poole, M. S. Procedures for Managing Meetings: Social and Technological Innovation. In R.
A. Swanson & B. O. Knapp (Eds). Innovative Meeting Management. Minnesota Mining &
Manufacturing, 1991.

Schutz, W. C., FIRO: A Three-dimensional Theory of Interpersonal Behaviour. Holt, Rinehart


& Winston, 1958.

Schwartzman, H. B. The Meeting, Plenum Press, 1989.

Shaw, M. E., Group Dynamics: The Psychology of Small Group Behaviour, McGraw-Hill,
1981.

Short, J., Williams, E. and Christie, B., Communication, Modes and Task Performance, In
The Social Psychology of Telecommunications, John Wiley & Sons, 1976.

Smith, S. M. Managing Your Meeting for a “Bottom Line Payoff”. In R. A. Swanson & B. O.
Knapp (Eds). Innovative Meeting Management. Minnesota Mining & Manufacturing, 1991.

Spector, P. E., Perceived Control by Employees: A Meta-analysis of Studies Concerning


Autonomy and Participation at Work, In Human Relations, 39, 1986.

Tuckman, B.W., Developmental Sequence in Small Groups. In Psychological Bulletin, 63,


1965.

Turner, R. H., & Killian, L. M., Collective Behavior, Prentice-Hall, 1987.

Verhagen H., Norm Autonomous Agents, Stockholm University, 2000.

Viller, S., The Group Facilitator: A CSCW Perspective, In Readings in Groupware and
Computer-Supported Co-operative Work, Morgan Kaufmann Publishers, 1993.

Watson, K. W., Listener Preferences: The Paradox of Small-group Interactions. In Small


Group Communication: Theory and Practice (7th ed.), R.S. Cathcart, L.A. Samovar, & L.D.
Henman (Eds.), Brown and Benchmark, 1996.

Weick, K. E., The Social Psychology of Organizing, Addison-Wesley, Second Edition, 1979.

Weiss G., Multiagent Systems – A Modern Approach to Distributed Artificial Intelligence,


The MIT Press, 1999.

Werle, P. and Jansson, C. G., Active Documents Supporting Teamwork in a Ubiquitous


Computing Environment, KTH Centre for Wireless Systems, DSV, 2001.

Wiberg, M., In Between Mobile Meetings Exploring Seamless Ongoing Interaction Support
for Mobile CSCW, Umeå University, 2001.

Wooldridge M., Intelligent Agents, In Multiagent Systems – A Modern Approach to


Distributed Artificial Intelligence, Weiss G (Eds), The MIT Press, 1999.

49
Wooldrige M., and Jennings N. R., Intelligent Agents: Theory and Practice. In The
Knowledge Engineering Review 10(2), 1995.

Web references

Smart Technologies Inc., SMART Board,


http://www.smarttech.com/products/smartboard/index.asp , 2002-12-17.

50
8 APPENDIX I – PRELIMINARY STUDY
Questions

The following questions were asked.

1. What signifies a rewarding meeting?

2. What purposes does a meeting have?

3. Can you think of any specific situations where meetings are being inefficient? What is the
cause?

4. What technological equipment are you using to support the meeting?

5. Can you think of any specific situations when the technology is hindering the meeting
rather than facilitating it? What are those technologies?

6. How can you make sure that everyone present is partaking in the meeting?

7. Is it important to have someone lead the meeting? If so, why?

8. If you had an assistant present to help you during the meetings, what would you want it to
do?
a. Would it make a difference if the assistant were artificial? In what way?
b. Is there anything you would let a human assistant do that you would not accept from
an artificial counterpart?

9. How would you like to communicate with/give orders to the assistant?

10. How do you prefer to take in information during a meeting?

11. Do you in your workplace use meeting agendas or some other meeting structure?

12. By norms I refer to the unspoken rules for behaviour and opinions that the meeting group
practises on the individual meeting members, concerning what is good/bad,
correct/incorrect, acceptable/unacceptable (for example the attitude towards mobile
phones during meetings)
- Could you please mention some important norms present in your meeting
situations?
- Does it happen that these norms are broken, and if so, when?

13. Are the different roles within the meeting group well established, or does confusion arise
on who is going to do what (secretary, head of meeting, who is going to do what till next
time)?

14. Can you think of anything that would facilitate meetings?

51
Results and conclusions based on interviews

Below is an aggregate of the most common answers received during the interviews. Also
listed are isolated reflections expressed by participants that we believe can have implications
for future design of support for information exchange meetings. Each interview item is
followed by a short conclusion.

1. What signifies a rewarding meeting to you?


Some answers varied depending on if the interviewee was the one presenting the information
or receiving it. However there were also issues relevant to all or most meeting paricipants
when it comes to identifying a rewarding meeting session:

• The meeting is held relatively short


• A consensus is reached, not just general discussion
• Issues presented are important and relevant
• The meeting is leading somewhere, towards some form of change
• The meeting is structured and follows an agenda
• Participants are informed about the purpose of the meeting
• Everyone actively partakes in the meeting
• Everyone receives and understands the information given
• Social rewards are present, such as getting to know each other
• People listen to each other
• The meeting is carried out in a positive atmosphere
• Critique is taken in a positive manner
• Everyone gets to talk without being interrupted
• Everyone gets an equal share of the meeting timeframe

Some issues were relevant mainly to the presenter:

• Being well prepared


• Not forgetting important information
• That listeners ask questions
• Receiving feedback from listeners, as a receipt that shows that they understand
• That listeners understand

Finally, some issues were relevant mainly to the listener:

• Having access to illustrations of difficult items


• Getting enough information
• Getting the right information
• That information is pedagogically presented
• Being able to ask questions
• Learning something new

Conclusions question 1
Judging by the interviews conducted two main areas can be identified among the aspects that
determine how rewarding a meeting will be perceived by participants; efficiency and
sociality. Whereas the efficiency of the meeting is more concerned with the practical

52
execution and structure of the gathering, the sociality aspect reflects the atmosphere and the
character of the co-operation present in the group.

Efficiency
The characteristics most frequently identified as important determinants of rewarding
information interchange meetings were the quality of the information interchange and
structure of the meeting. These are highly relevant for the efficiency of the meeting.

The information interchange must ensure that the information given and received is correct,
relevant and sufficient. The quality of the information must be certifiable, which means that
the person providing the information and the person obtaining the information must be able to
control the quality of information with each other. The information provider achieves this by
prompting for feedback, whereas the information receiver satisfies this need by asking
questions.

The structure of the meeting was considered essential. Interviewees mentioned that the
meeting must not be too stretched out in time, since longer meetings tend to get out of hand
and become inefficient. Most persons interviewed stressed the importance of following an
agenda to keep the meeting on topic.

It is apparent that the efficiency of the meeting is the primary concern of the respondents, and
that the persons interviewed believe that the purpose and goal of the meeting are vital and
must be clearly stated in advance. This wish for efficiency is reflected in the need for
participants to come prepared to the meeting and bring relevant issues up to discussion.
Respondents also considered it important that the meeting would lead to some form of
resolution, result or change. As was expressed by one person; “a successful meeting is an
integrated part of the work process”. A meeting without a goal or purpose was deemed
useless.

Sociality
It is interesting to note that many of the respondents identified other group members'
understanding of and participation in the meeting as crucial for if they themselves were to
perceive the meeting as rewarding. In most workgroups, a shared understanding and
interpretation of the context, issues and objectives of the workgroup is vital for efficient
collaboration within the group. The fact that so many respondents listed this as an important
item suggests that they have been victims for the detrimental effects of poor co-operation,
whether they are aware of this or not.

Interestingly, only few of the persons interviewed mentioned social aspects as vital when
identifying the meeting as rewarding. This coincides with the inclination for efficiency and
swiftness when executing the meeting. However, this does not mean that social relations are
unimportant, but rather - as stated by some of the respondents – that the formal meeting is not
the proper forum for social discussion. Interviewees would rather have the meeting carried out
rapidly and thereafter devote time solely for the tending of workgroup relationships.
Nevertheless, when asked about the purposes of conducting meetings (see question no 2),
nearly all respondents mentioned social aspects as a key purpose of organising group
gatherings. In fact, apart from interchange of information, this was the most frequently
mentioned reason for arranging meetings. However, judging by interviewees attitudes, the
meeting session is not the proper place for team building.

53
Norms
Based on answers given by respondents, there is reason to believe that irritation occurs when
the meeting loses focus of the main topic. This lack of focus can arise in different situations,
but respondents maintain that it can be counteracted by efforts such as keeping the meeting
structured and following an agenda. We believe that respect for other persons’ time is an
important, but not always consciously implemented, norm for meeting behaviour. When
spending someone else's time in an unproductive manner, it may be perceived by others as a
norm violation. Other interview answers that reflect this respect for others were the opinion
that all attendants should come to the meeting well informed about its purpose, and that the
person presenting information is well prepared.

Democracy
Respondents also emphasise group members' right to voice their opinions and be heard by
other members. Decisions should be made as a result of group consensus. At the same time,
all the respondents stressed the importance of appointing a chairman to the meeting (see
question no 7). Thus, the meeting must be a democratic arena, but at the same time highly
moderated by a single group member. The notion that all attendants should understand all
aspects of the meeting and be well informed can also be seen as a statement of democracy.
The reason this was seen as an important aspect was because it enhanced efficiency in the
workplace.

2. What purposes does a meeting have?

Not surprisingly, the sharing of information was mentioned by nearly all interviewees as the
most important reason to hold information interchange meetings. The answers reflect the top-
down, bottom-up and peer-to-peer passing of information throughout the organisation;

• Spreading of knowledge for educational purposes


• The employees receive information about the organisation such as policies and directions
• The leadership passes on information to the employees
• Group leaders receive opinions and feedback from group members
• Group members emphasise the opportunity to voice opinions
• Group members learn from each other
• General information exchange to assure that participants are in agreement

Socialising, getting to know each other and creating a sense of union within the group was
considered almost equally important as information interchange, and was mentioned by a
majority of interviewees. This was considered particularly important in a workplace where
face-to-face communication between co-workers is rare, which is often the case in most
respondents' line of work. Other purposes mentioned were:

• Finding problems
• Enhancing quality
• Reaching a goal
• Getting an overview of large quantities of data without having to do all the research
• Communicating through other signals than merely speech, presence

54
Conclusions question 2

Information interchange
Since this question concerned the purposes of information interchange meetings, it is not
surprising that all participants agreed that the main purpose of meetings was indeed
information interchange. But judging by the responses obtained, the meeting arena
encompasses many different types of information interchange relationships. With the help of
meetings, information is passed on throughout the organisation, both from the top down and
from the bottom up, as well as between co-workers. It is apparent that management uses
meetings as a way of informing employees about organisational decisions and policies,
whereas middle management sees the meeting as an opportunity to receive information,
feedback and opinions from the team. To employees, meetings are important as a forum for
voicing opinions and learning from each other, but also receiving information and keeping up
to date with what is going on in the organisation. Ensuring agreement between meeting
participants, educating and receiving education are other forms of information exchange
occurring in meetings. Respondents paint a picture of constant flow of information within the
organisation and an important way of achieving this is by conducting meetings.

Other purposes
As already mentioned, tending of social contacts was listed by almost all respondents as a
primary purpose of conducting meetings, even though this aspect was not considered
important as a determiner of rewarding meetings. The meeting was seen by some respondents
as a way of unifying the team and getting to know each other. Other reasons for meetings
were getting an overview over large quantities of data and identifying problems.

3. Can you think of any specific situations where meetings are being
inefficient? What is the cause?

When a meeting becomes inefficient this is mainly due to four reasons; lack of structure due
to meeting mismanagement, inadequacies in the information interchange, environmental
impediments and norm violations.

• The meeting gets out of hand, this happens because:


− Issues are irrelevant
− The chairman is not controlling the meeting
− People keep interrupting the speaker
− The meeting takes too much time
− The information given is not relevant to the meeting participants
− The meeting is not leading anywhere, this happens because
− The purpose of the meeting is not clear to the participants, or there is none
− People are unfocused or tired
− People are uninterested and their attention drift, this happens because
− People are not sure why they are there.
− The meeting does not have a structure, this happens because
− The person presenting is unprepared
− There is no agenda
− The presenter does not present the material in an easily understandable manner
− Participants do not get the chance to ask questions

55
− Information is hard to understand
− The meeting is not sufficient to cover the need for information
− People do not dare to ask questions
− The participants do not have enough information, this happens because
− People do not respect each others’ right to finish
− Men do not let women talk
− Participants have objectives of their own that clash with objectives of the group
− Failure of technical equipment
− Disturbing noise from technical equipment
− Participants do not get along

Conclusions question 3
Inefficiencies in the meeting room are primarily caused by breakdowns; when the meeting for
some reason is interrupted and therefor grinds to a halt. The factors that respondents listed as
responsible for inefficiencies and breakdowns where almost invariably the same as were
deemed crucial when deciding if a meeting is rewarding or not. This strengthens our reasons
to believe that our conclusion, that the efficiency of the meeting is the primary factor to
determine how rewarding the meeting is perceived by the participators, is indeed correct.

We believe that meeting room inefficiencies can be divided into four main areas; structural
meeting mismanagement, inadequacies in the information interchange, norm violations and
environmental impediments. Whereas the latter category is more concerned with issues not
directly related to the meeting per se, such as failure of technical equipment and disturbing
noises, it is mainly the former three categories that are of interest in this preliminary study.

Structural meeting mismanagement


The circumstances under which the meeting is taking place are highly critical for a successful
session. Frequent contributors to meeting inefficiency mentioned by almost all interviewees
are lack of structure, focus, goal and purpose. According to respondents, shortcomings like
these become evident when the meeting gets out of hand, the meeting is not leading
anywhere, and the structure and goal of the meeting is not clear.

The first example, when the meeting gets out of hand and meeting participants lose focus of
the issue at hand, can occur for a number of reasons. Situations mentioned by respondents
where for example if the chairman is not in control of the meeting, people are unfocused,
tired, or uninterested, participants have not been informed about the goal of the meeting, or
issues presented are irrelevant to the people present. This can occur when the people invited
to the meeting are not directly concerned with the topics addressed. Related to this is the
problem of lack of goal and purpose to the meeting. Again, this reflects respondents' emphasis
on the need for efficiency, since this was also frequently mentioned as central when
classifying the meeting as rewarding.

In order to control these issues, it was deemed important to appoint a chairman to the meeting,
clearly state goals and purposes in advance as well as using an agenda during the meeting.
The structure of the meeting could also improve considerably if the person in charge or
presenting a topic was properly prepared for the task.

Inadequacies in the information interchange


Information insufficiencies occur when meeting participants do not get enough information or
correct information. This could be either on behalf of the presenter who might not present the

56
information in an easily understandable manner or would not give listeners a chance to ask
questions. Sometimes the information itself could be hard to understand and further
explanations were not given.

Responsibility also falls on the listeners to overcome any reluctance towards asking questions
in order to receive proper information and avoid any misunderstandings. Being prepared prior
to the meeting about the issues being presented might also improve the understanding
considerably.

Information inadequacy also occurred when the meeting was not sufficient to cover the need
for information, something that possibly could improve if employing a stricter meeting
structure.

Norm violations
When people interrupt each other, have objectives of their own that are incompatible with
goals of the group or simply can not co-operate in the meeting room, norm violations occur
that make the meeting less efficient. Since norms can be hard to recognise for group members
themselves, it is likely that an observatory study might serve to identify more instances of
norm violations in meeting groups. However, that was not of focus in this study.

4. What technological equipment are you using to support the meeting?

Nearly all interviewees used whiteboard and overhead projectors frequently. About half of
them also used pen and paper for taking minutes and making sketches to show others.
Flipcharts, speakerphones, conference systems and videoconferences were used by some but
this was not part of the every day meeting routine. Some of the respondents used lap top
computers for taking minutes, or connecting to an overhead projector. Applications that were
scarcely mentioned were Microsoft PowerPoint ® for illustrations, Microsoft Outlook ® for
meeting support, Microsoft Word ® for taking minutes and Microsoft Project ®. Some
respondents expressed aversion towards using overhead projectors or other electronic devices
since they had negative experiences of such appliances.

Conclusions question 4
The reason whiteboards, flip charts, pencils and papers were cited as useful instruments, was
to a large degree due to their advantage of not being prone to malfunction and breakdowns,
which was regarded a substantial weakness and recurring obstacle in the use of technological
equipment (see question 5).

5. Can you think of any specific situations when the technology is hindering
the meeting rather than facilitating it? What are those technologies?

Group members familiarity with the tool utilised and the flexibility of the system were critical
for not hindering the meeting. Purely technological breakdowns were also mentioned.

• The presenter is not using the tool the way it is intended, this happens because:
− Unfamiliarity with the tool
− The presenter relies on the tool too much for support
− A telephone conference is not regulated enough and gets out of hand
− Spelling mistakes
− Technological devices are not interactive enough, they should not be used

57
− Technology steals time and energy from the meeting, it should not be used
− Illustrations that do not support the topic
− Too much simultaneous information
− Too many illustrations
− Too small text/illustration
− Projectors are sometimes hard to connect to the computer
− Getting kicked out of the conference system
− Technological equipment that is not available, technological breakdown
− Overhead projector missing
− Whiteboard pencils dried out
− Access to networks denied
− Hardware problems: power cord missing/too short, overhead projector will not start,
lamp broken
− Power shortage
− Too much noise from technological equipment
− People using keyboards (noise), surfing during the meeting

Conclusions question 5
The technological problems as mentioned by respondents were either directly linked to the
technology itself, or a consequence of lack of expertise in operating the device on the
information presenter's behalf, i. e. poor meeting management.

The first category contains difficulties such as power shortage, missing hardware, like cords
and light bulbs, disturbing noises from the apparatus, not getting access to networks and
servers, and the often considerable time delays that can occur as a consequence of having to
solve these problems on the spot.

As mentioned before (see question 3), the other category is more of interest to this study.
According to our interviewees, full mastery of the equipment is essential, otherwise utilising it
will be considered more of a hindrance to the meeting than an aid. This includes presenters
being unfamiliar with the usage of the tool and relying too much on the tool for support. In
situations where the tool is employed as a means of illustrating what is being said, it is
important that text and illustrations support the topic, are easily understandable, well-arranged
and not undersized, and that there is not an overload of simultaneous information.

The inflexibility of the equipment was a common cause for irritation; some respondents even
deter from using any form of technology since it was perceived to waste time, energy and
impede creativity. Since most respondents had experience of Microsoft PowerPoint being
employed as meeting support it is likely that this application is what they had in mind when
answering this question. At least two respondents expressed their disappointment with the
common employment of Microsoft PowerPoint due to the inflexibility and lack of
interactivity it causes when the presenter feels obliged to keep strictly to the written material.

6. How can you make sure that everyone present is partaking in the meeting?

Ways of ensuring participation by all meeting attendants range from encouraging to


dictatorial, and much of the responsibility is placed on the chairman.

• All participants are asked for their opinion

58
• Those that are silent are asked for their opinion
• The chairman makes sure everyone is active
• Focus on the agenda
• Have the agenda state that everyone must speak
• Encourage people to ask questions
• Keep prestige low so that everyone dares to ask questions
• Make sure participants have access to meeting material in time before the meeting to
ensure that they are well informed
• All participants should be familiar with meeting support technology used
• Employ rules for the meeting that state that everyone has to be active
• Make sure the meeting is interesting to the participants, so that they feel affected
• Invite the right people, only those directly concerned with the right knowledge
• Encourage discussion
• Delegate responsibility so that everyone must be active

Conclusions question 6
When asked about the factors that signify a rewarding meeting, respondents were anxious to
make sure that everyone present would participate and leave the meeting with sufficient
information. Answers show that this democratic aspect is best protected by stimulating
involvement; something that can be approached in diverse manners.

It was commonly believed that it was the responsibility of the chairman to instil activity in all
participants. Asking each and everyone for their opinion on each issue, encouraging questions
and discussion through enforcing a positive atmosphere to the meeting, as well as making the
meeting interesting for all participants through making sure the right people are invited, were
some examples.

More brute methods were also suggested, like employing rules that declare everyone's
obligation to state their opinions, delegating responsibility to ensure activity, and strictly
following the agenda.

It was considered fundamental for the democratic aspect that participants were familiar with
the meeting support technology and that meeting material was available in advance so that
everyone would have a chance to prepare themselves for the meeting.

It is important to note that regardless of how this is to be achieved, none of the respondents
questioned the need for ensuring involvement by all group members. As mentioned before
(see question 1), efficiency was of primary concern for the people interviewed, something that
can be promoted by active participation since this may speed up the meeting pace as well as
minimise the risk for misunderstandings.

7. Is it important to have someone lead the meeting? If so, why?

All respondents considered a dedicated chairman crucial for determining the success of the
meeting session. The bigger the meeting, the more important the chairman. The person who
summoned the meeting should run it. Typical chairman tasks were to:

• Make sure that all issues are processed


• Force the meeting to proceed

59
• Make sure the meeting is going in the right direction
• Have responsibility
• Make sure the meeting follows the agenda
• Serve as an administrative function; start the meeting, end the meeting, take notes,
monitor attendants’ right to speak
• Make sure that issues dealt with are acted upon
• Monitor the structure of the meeting
• Make sure the meeting does not get out of hand

Conclusions question 7
All respondents unanimously agreed that someone must have the overall responsibility for the
progression of the meeting, especially if the meeting group contains more than five attendants.
This is in many ways an administrative function; starting the meeting, making sure the agenda
is adhered to, monitoring everyone's right to speak, and ending the meeting. It is the task of
this meeting facilitator to make sure that the goal and purpose of the meeting are successfully
carried out and that all issues on the agenda are properly dealt with and acted upon. If
someone is talking for too long, or if the meeting loses focus, the facilitator should steer it
back on track. Thus, the undertaking of the facilitator is mainly to tend to the structure of the
meeting. This could mean that the role of the chairman is subordinate to the agenda and it is
the agenda and not the chairman per se that makes the meeting efficient. However, it is likely
that the respondents have never experienced the implementation of an agenda without the
occurrence of a chairman and therefore overestimate the significance of the chairman.

8. If you had an assistant present to help you during meetings, what would you
want it to do?

This question was divided into one main and two sub questions concerning assistants. The
reason for this was to first get a more general idea of what people would want a human
assistant to do for them during a meeting, and then ask them if these tasks would be different
if the assistant were to be of an artificial nature. We also wanted to keep the interviewees
from limiting their answers to what they thought would be possible to achieve with current
technology. The second sub question concerned the trust issues people might have when it
comes to computer based assistants versus human ones.

Again, the answers to these questions depended on the supposed role of the meeting attendee.
Most interviewees had the opinion that an assistant would be most helpful to the chairman or
someone presenting some sort of material and with less or no individual use for group
members taking a more passive role.

The interviewees listed the following tasks for a human assistant supporting an active
member of a meeting:

• Taking minutes
• Summing up the meeting
• Keeping track on whose turn it is to take the floor
• Preparing the meeting by doing research and summarising information.
• Distributing information to the other attendants, and making sure everybody has the
correct material

60
• Taking care of technical equipment during the meeting, such as changing slides, operating
the projector etc.
• General help and support during the meeting.

However, a significant number of the interviewees claimed that they would not feel
comfortable with an assistant. It would be wrong to delegate such “boring” tasks to someone
else, and it would be harder to process meeting information if someone else took minutes for
them. Again, it was mentioned that during meetings it is expected that everyone take an
active part and that an assistant not contributing creatively would disturb more than it would
help. It was also said that it would be good to have an assistant taking a more cognitively
active role commenting on ideas and adding new perspectives to stalemated discussions.

8.1. Would it make a difference if the assistant were artificial ? In what way?

The tasks suggested did not differ much from the ones mentioned above with some additional
ideas:

• Reminding the chairman what is left on the agenda


• In a discrete way reminding the chairman that it is time to move on and end a discussion
• A very fast and efficient search engine that served you the exact information at the
moment you needed it

The interviewees also made reservations about the abilities of artificial assistants. How would
an artificial assistant be able to see and interpret what it means when someone is raising her
arm? Would it be able to write a protocol without entering words and phrases without
significance? And how would it know which important points to put in a meeting summary?
The main point made was that an artificial assistant would lack the social and creative
qualities found in a human assistant, “it would not be able to tell jokes” as one interviewee put
it.

8.2. Is there anything you would let a human assistant do that you would not
accept from an artificial counterpart?

People were positive to letting a computer based assistant take care of basic routine work like
mailing information to meeting attendants and taking minutes if it had proven that it did its
job well. However, they would leave social tasks and more delicate matters like sending
information to the company president to humans.

Conclusion question 8
One of the purposes of this question was to find out if people make distinctions between
human and artificial assistants, and if they do, which ones. Even though the number of
persons interviewed is not sufficient to make statistical conclusions, there is a tendency that
both types are trusted as long as they fulfil the tasks laid upon them, with perhaps a small
inclination towards human assistants when it comes to very delicate matters. As mentioned
above, the majority of the respondents had some sort of connection to the IT-industry, and it
is therefore likely that they are confident around information technology. It is possible that the
same level of trust or confidence in computer driven assistants would have been different if
the respondents heired from other areas of work. This possible bias is however acceptable
considering that the environment, the iLounge, in which this thesis resides to some extent
requires prior knowledge and experience in the IT field.

61
Respondents were concerned about delegating routine tasks like taking minutes to a fellow
co-worker and would be happier if this was made by an artificial assistant. This was more the
case if the respondent had the role of chairman and as such had to concentrate on managing
the meeting and presenting information. When merely observing a presentation or listening,
the respondents were more prone to taking their own minutes to help them remember key
issues. They also raised questions about the creative and social abilities of an artificial
assistant. In spite of this, the most common and seemingly obvious suggestion was that an
assistant should take minutes and sum up the meeting’s key information. Also, a fairly large
amount of suggestions had to do with keeping up meeting structure and order, that is, making
the meeting more efficient and to the point. These suggestions revolved around making sure
everybody had the same documents and information, handling of technical equipment, and
keeping better track of meeting time, agenda and whose turn it is to take the floor.

Thus, the majority of the suggestions had to do with meeting management in general in order
to free people from routine and tiresome work. The few objections people had towards
artificial assistants had to do with creative and social work. A conclusion could be that that an
artificial assistant should be focused on relieving people of administrative tasks and leaving
the creative thinking and discussions to humans. On a short note, two of the interviewees
made some remarks about a possible use for an artificial assistant being able to listen and
evaluate suggestions, and by own means coming up with ideas. It is possible that future
development in meeting technology will include cognitive and creative properties, but
however intriguing, this lies beyond this thesis considering how the evolution of artificial
intelligence at this point has yet to arrive at such heights in reasoning and creative thinking.

9. How would you like to communicate with/give orders to the assistant?

The most common notion among the interviewees was that whichever mode of
communication was used it had to be discrete and short in nature to keep it from disturbing
the meeting. Suggested forms of input were:

• Through short verbal commands like “send all”


• Typed commands
• Pushing buttons for execution of predefined orders
• Pointing on a screen
• Writing
• Giving orders by pressing a pedal
• Feeding the assistant with a meeting plan beforehand to make it automatic or reactive to
simple gestures during the actual meeting.
• Some sort of pointing device

Some persons were reluctant to spoken commands because other group attendants could
misinterpret these words as directed to them. It would be difficult to make a natural
distinction between speech directed to the assistant and speech directed to other people
without making some signal, like pushing a button or making a gesture and then you might as
well just do it in a non verbal manner.

Conclusions question 9
When asked about the communication between the assistant and the meeting members there
was no real consensus about what type of modality to use. Some people favoured voice

62
commands and other written commands into some sort of interface, or the pressing of
command buttons. The consensus among the respondents lied however in the characteristics
of the communication; the assistant had to be told in such a manner that the interaction did not
interrupt the meeting process or create confusion on who was addressed; humans or assistants.
Hence, a possible solution, as one respondent pointed out, could be to take care of as much of
the communication directed from human to assistant before the meeting, like feeding it the
agenda and giving orders about what files to send and presentations to launch, so that the
majority of the assistant’s actions during the meeting are automated or triggered by simple
cues. Naturally, the communication from assistant to humans has to obey the same regulations
and be kept short and without too much disturbance.

10. How do you prefer to take in information during a meeting?

Almost every person interviewed took personal minutes even though it was common that a
summary of the meeting was sent to every attendant after the meeting. But the question
touched more on opinions on how to present information the best way:

• A well prepared speaker is essential


• PowerPoint slides should only be supporting the presentation, it should not be the
presentation, that is, they should only show keywords and phrases and relevant supporting
pictures.
• Pictures and video are good to prove a point and show examples from the real world, it
makes it easier to remember.
• Drawings and sketching also facilitate remembering important points.
• The presentation should be held central to all so that everyone focuses on the same thing
• If discussions are held after a presentation it is easier to remember what was brought up.
• Presentations should be short and simple.

Conclusions question 10
One of the main purposes of meetings is obviously to share and distribute information. We
therefore asked what people do to take care of and process that information during the
meeting. The answer was very simple; take minutes. Almost everyone found it more
rewarding to write down important points made, even though a lot of times there were
summaries of the meeting distributed some time afterwards. However, the respondents had a
lot to say about the other side of the coin, that is the importance of how the information is
presented. People were generally tired of long PowerPoint presentations with way too much
information crammed into every slide. Instead of leaning on technology the presenter should
be well prepared and only use technology like PowerPoint slides, pictures and film to
emphasise what is important. Rather than using pre-made slides the presenter should illustrate
arguments by drawings and sketches made the same moment they are used.

11. Do you in your work place use meeting agendas or some other meeting
structure?

All interviewees had some sort of agenda during formal meetings. This was in most cases put
together by the person who summoned the meeting and sent out in advance. One person used
a multi-user calendar system in his work place to book meetings and send out agendas and
additional information. Only a few had more standardised types of agendas and often only for
certain types of meetings such as board meetings. Most meetings were of a more informal
character even though an agenda was put together. However, the presence of an agenda was

63
not a guarantee for efficiency, it often happened that discussions strayed from the issue and
stole valuable time. Some interviewees stressed the importance of a moderator as a
compliment to the agenda to keep up meeting efficiency.

Conclusion question 11
Agendas were commonly used by the respondents to facilitate meeting structure, even though
the agenda itself seldom sufficed to keep the meeting from getting sidetracked. It was agreed
upon that an agenda is imperative but that a moderator or chairman to keep discussions to the
point was almost equally important. The responses to this question furthered the impression
that structure and guidance are cornerstones of an efficient and rewarding meeting.

12. By norms I refer to the unspoken rules for behaviour and opinions that the
meeting group practises on the individual meeting members, concerning what
is good/bad, correct/incorrect, acceptable/unacceptable (for example the
attitude towards mobile phones during meetings). Could you please mention
some important norms present in your meeting situations?

The most frequent answer to this question was “common sense” and “civil manners”. Some
did not mind phones being turned on as long as people stepped outside to talk and that the call
was important. Other norms where:

• Act professionally, you are invited to the meeting because you posses some sort of
professional knowledge.
• Arrive to the meeting on time
• Let other people finish talking before speaking
• Listen and be attentive
• Let everyone take part in the meeting
• Show respect for the meeting and it’s members
• Come prepared to the meeting, read the information sent to you in advance to save
valuable time during the meeting
• Stay until the meeting is over on the agreed time

12.2 Does it happen that these norms are broken, and if so, when?

There were no severe punishments for breaking the meeting norms (in one meeting group you
had to buy everyone else beer if your phone rang). However, people tend to get irritated and
the result of the meeting suffers. One interviewee mentioned that a result of norm breaking is
that you do not get to hear and learn other people’s perspectives and ideas. Another reflection
was that norm violations tend to occur when discipline gets lower. This happens when the
meeting discussions get out of hand and sway too much from the agenda and the attendants
get bored.

Conclusion question 12
To find out more about the different symbols (cultural signs) and ways of acting during
meetings, questions about norms were asked. The definition of norms chosen was “unspoken
rules for behaviour and opinions that the meeting group (or company) practises on the
individual members, concerning what is good/bad, correct/incorrect and
acceptable/unacceptable”. The attitude towards mobile phones during meetings was taken as
an example. The respondents had some difficulty identifying the norms present in their
meetings, and they often took the given phone example as a typical meeting norm. Judging

64
from the responses, the most obvious norms during the meeting situation do not differ much
from those found in other group situations. This is reflected by the most frequent answers;
common sense and civil manners. People should be treated with respect by arriving on time,
listening to what others have to say and letting them speak till they are finished.

More meeting oriented norms also reflected these values. It was expected that respect was
shown for the other’s time and effort by sharing professional knowledge and preparing for the
meeting in advance. There were no immediate and obvious penalties for breaking norms, but
some respondents claimed that breaking norms had consequences on the quality of the
meeting. People get irritated and bored when someone makes a poorly prepared presentation
because it is wasting their time. Norm breaking also results in poorer creativity during the
meeting. One interesting remark was that norm breaking tends to occur more often when the
discipline drops due to irrelevant discussions far from the agenda and poor order keeping.
Hence, improved structure could perhaps decrease norm breaking.

13. Are the different roles within the meeting group well established, or does
confusion arise on who is going to do what (secretary, chairman, who is going
to do what till next time)?

In all cases it is usually the person calling the meeting that becomes the chairman, and if no
one is explicitly appointed secretary it is often this person that takes minutes. It is also
common that the attendants take turns from meeting to meeting to keep the protocol and write
a summary. In other cases choosing secretary can consume time and be a source for irritation
since it is not a popular task. A few of the interviewees mentioned that problems might arise
when it comes to professional roles. Even though a member has a title, other group members
have different opinions on the capabilities of that person, they may overestimate or
underestimate the person. This can make people unsure of their role in the group and they
might hesitate to come to follow-up meetings. Another problem is when the wrong people are
invited to the meeting, for example out of politeness. These persons are not as familiar with
the purposes of the meeting as other meeting members and therefore ask unnecessary
questions and take up valuable time from more constructive discussions.

Conclusions question 13
Norms influence how people behave and react to stimuli and hence help create and manage
different roles within a group. As these two phenomena are related the respondents were
asked to what extent roles in meetings are established and if problems arise when these roles
are not clear. This question was primarily targeted towards the typical meeting roles like
chairman and secretary but there were also remarks about how the meeting attendants'
professional roles outside the meeting influence the meeting progress.

The general idea was that the chairman almost exclusively was the person calling the meeting
and there were seldom any arguments about this role. It was often the task of the chairman to
take minutes during the meeting, but if a secretary was required this appointment could cause
some disturbance and unnecessary waste of time. The reason given for this was that taking
minutes and making summaries was not considered a desirable task (as mentioned above, this
was the reason some respondents did not want to delegate this task to a human assistant). A
person's professional role was sometimes challenged if other meeting attendants questioned
suggestions based on that persons competence. This resulted in that people whose competence
got questioned hesitated to attend more meetings in the matter since they were insecure
whether their contributions really made a difference.

65
Another problem with roles during meetings is the lack thereof, that is when some attendants
actually lack the professional competence or importance needed, and are invited to a meeting
out of politeness or routine. As a result, the meeting gets disturbed by unnecessary questions,
and discussions do not reach the desired level of detail. It is clear that meetings would benefit
from more knowledge of the different roles present, not only to locate the competence but
also to gain a better meeting structure and efficiency during discussions and decision making.

14. Can you think of anything that would facilitate meetings?

A lot of different suggestions were mentioned:

• A more competent moderator to steer the discussions back on track, and move on when it
is time to move on
• Something that eliminates disturbances and keeps people focused on the meeting
• More meeting rooms available
• Less criticism and more constructive criticism
• Better presentations
• Less obvious technology that does not steal the focus from the discussion
• Keep the meeting from being too focused on details instead of the big picture
• The context of the meeting, it should be equal with no symbols of position or power
• That everyone comes prepared to the meeting.

Conclusions question 14
When addressed with question number fourteen concerning things that would facilitate
meetings, not mentioning assistants, the answers were of a much more abstract character and
more directed toward correcting human weaknesses. Most of the answers concerned
improving human characteristics during the meeting. A topic mentioned here and in other
places during the interviews was the lack of chairmen as trained presenters and moderators to
keep meetings interesting yet focused. Another typical human weakness in need of
improvement that also was mentioned several times was arriving to meetings more or less
unprepared. These suggestions on meeting improvement are most likely very difficult to carry
out through some technical solution, since they aim to correct phenomena like motivation and
experience. If you are not interested you are not likely to prepare yourself for a meeting and
make a difference during discussions even if you have the best technology to guide and
support you.

An interesting reflection made by one of the respondents was the importance of the contextual
properties of the meeting. In order for the meeting to be rewarding for all participants there is
a need to keep the status of the people involved at an equal level. If, for example there is a
more comfortable chair assigned to the chairman it could signal prominence over other
members of the group hindering them to speak freely or holding back relevant opinions. To
keep constructive discussions going there has to be some equality within the group. Another
remark about the meeting context was how an inspiring milieu promotes creative thinking.

66
9 APPENDIX II – COMPLETE EXPERIMENT RESULTS
Here are the complete results from the experiment questionnaires. The results are presented in
the following order:

1. General questions posed to all subjects


2. Questions posed to all subjects except HuMod chairmen
3. Questions posed to AgMod participants and CoMod Chairmen
4. Questions posed to HuMod Chairmen
5. Questions posed to AgMod participants

1. General questions posed to all subjects

Q 1. I perceived the execution of the meeting as efficient.

1. I percieved the execution of the m eeting as efficient

12
Number of subjects

10
8 HuMod
6 CoMod
4 AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
It is evident that participants of the AgMod condition were more inclined to perceive the
meeting as efficient than the participants of the other two conditions. 10 of 12 did so, 4 of
them answering “I totally agree” to the specific question. Only 5 of 12 in the HuMod and
CoMod conditions perceived the meeting as efficient, and none of them gave it a “five”, i.e.
full rating.

1. I perceived the execution of the m eeting as efficient.

3
Number of subjects

2
HuMod
CoMod
1

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
There seems to be little difference between chairmen’s perception of meeting efficiency in the
two conditions. None of the answers suggests any dissatisfaction with the efficiency, however
two of the chairmen of the CoMod condition were neutral.

67
Q2. I experienced that the commands of the chairman were followed.

2. I experienced that the commands of the


chairman/Agnes were followed
Number of subjects

12
10
HuMod
8
6 CoMod
4
2 AgMod
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
Participants in the CoMod condition were less convinced that the chairman’s orders were
followed than the other two conditions. All participants of the AgMod condition maintained
that Agnes’ commands were followed, whereas only six participants of the CoMod condition
experienced that the commands of the chairman were followed.

2. I experienced that the com m ands of the chairm an w ere


follow ed.
Number of subjects

2 HuMod

1 CoMod

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
There is a slight indication towards chairmen of the HuMod condition being more content
with the way commands were followed; two of three answered that they “totally agreed” that
their orders were followed. However, only one of all six chairmen was dissatisfied with the
adherence to commands. This happened in the CoMod condition.

Q3. The meeting would have functioned less well without the chairman.

3. The meeting would have functioned less well


without the chairman.
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

68
Participants
In the HuMod condition, only 5 of 12 believed that the meeting would have functioned less
well without the chairman. 10 of 12 in the CoMod condition believed so, which is interesting
since so few of them thought that commands from the chairman were adhered to. In the
AgMod condition, 6 of 12 seemed to think that Agnes was important for the meeting.

3. The m eeting w ould have functioned less w ell w ithout the


chairm an.
Number of subjects

2 HuMod
1 CoMod

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
Two chairmen of the CoMod condition and one of the HuMod condition valued their
contribution to the functioning of the meeting. Only one (in the HuMod condition) thought the
chairman did not improve meeting performance. The others were neutral.

Q4. The meeting would have functioned better without the chairman.

4. The meeting would have functioned better without


the chairman
Number of subjects

12
10
HuMod
8
6 CoMod
4
2 AgMod
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
In the HuMod and CoMod conditions, 1 of 12 believed that the meeting would have
functioned better without the chairman.. In the AgMod condition, 2 of 12 subjects believed
so. Apparently, having a chairman was deemed valuable in all conditions, and there is very
little reason to believe there is any difference between the groups related to this question.

69
4. The m eeting w ould have functioned better w ithout the
chairm an.
Number of subjects

2 HuMod
1 CoMod

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
None thought that the meeting would function better without a chairman, even if two of the
CoMod and one of the HuMod chairmen were less confident about this, choosing the neutral
option (“neither nor”).

Q5. All meeting attendants participated equally in the meeting.

5. All meeting attendants participated equally in the


meeting
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
6 of 12 participants in the HuMod condition believed that all meeting attendants participated
equally in the meeting. In the CoMod condition 4 of 12 believed so. The AgMod subjects
seemed most content with the degree of equal participation, 9 of 12 answered positively.

5. All m eeting attendants participated equally in the m eeting.

3
Number of subjects

2
HuMod

1 CoMod

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
All chairmen of the HuMod condition were content with the degree of equal participation. In
the CoMod condition, two were openly discontent.

70
Q6. The chairman/ I as a chairman, facilitated the participation in the meeting.

Q7. Can you develop your answer to the question above?

6. The chairman / I as the chairman facilitated the


participation in the meeting
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
6 out of 12 participants in the HuMod felt that the chairman indeed facilitated the
participation. However, 3 respondents were neutral and the remaining 3 disagreed. This
spread is not present in the CoMod condition where a total of 9 subjects complied with the
statement by answering 4 (“I agree”) , and only 1 person marking a 2 (I disagree) in
disagreement. The AgMod condition goes even higher with 10 subjects in agreement, 5 out of
those in total agreement.

When asked to further develop on Q 6, the written answers in the HuMod condition showed a
spread in opinions about the chairman, because of the his/her way of conducting the meeting.
For example, “During our meeting, the chairman was relatively passive and did not influence
much” as compared to “Yes, by letting everyone speak! Otherwise some people will not say a
thing while others take command of the whole meeting”.

The CoMod condition was much more unified. Most (8 out of 12) participants mentioned in
their written answers that the chairman actively pointed out whose turn it was to speak,
although some questioned why the chairman only behaved as a moderator and had no
opinions of his/her own.

All but 2 participants in the AgMod condition thought that Agnes added a structure to the
meeting in a positive manner. However, there was a general agreement that Agnes pushed the
meeting forward only allowing the speaker a certain amount of time before passing the word
on to the next person. “You were appointed a certain amount of time for your arguments”, ”
We were held rather short so that everyone got to speak, which was good, the time given for
discussion was a bit short but we still reached a good solution.”, but “I think we as a group
were anxious to follow the rules of the game, we didn’t want to interrupt each other. It is
doubtful that we would have been as willing to co-operate had it been a real meeting” were
some of the reflections made.

71
6. I, in the role of chairm an, facilitated the participation in the
m eeting.

3
Number of subjects

2 HuMod

1 CoMod

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
Answers to Q 6 do not differ a lot between conditions. The answers to the closed question
indicate that the hidden moderation was more successful in achieving equal participation.
However, judging by open questions, the chairmen of the human moderated condition all
seemed to believe that their contribution facilitated equal participation in the meeting (“The
chairman distributed time in a fair manner and created opportunities for all to partake”),
whereas the chairmen in the condition of hidden moderation seemed more critical of whether
they achieved this. As one chairman in the CoMod condition put it, “I perceived it as some
participants dominated the discussion”. When comparing the results to Q 5, it is evident that
the two chairmen that were discontent with the degree of equal participation still believed that
they as chairmen improved the participation.

Q8. It happened on one or several occasions that the meeting became


inefficient.

Q9. If you have answered 4 or 5 to the question above, at what point/s did you
experience the meeting as inefficient? What do you think was the reason for
inefficiency?

8. It happened on one or several occasions that the meeting


became inefficient
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
The results stemming from Q 8 show that a slight majority within the first two conditions
experienced inefficiency on one or several occasions. This was especially true for the CoMod
condition. The HuMod condition showed an almost exact composition of answers as in Q 6, a
result from which it is difficult to draw conclusions.

The third condition differ with only 2 subjects experiencing inefficiency and a total of 7
persons saying they did not experience inefficiency.

72
The answers to the follow-up question were very diverse in all three conditions, but some
related to insecurities about the scenario and structural weaknesses , for example “ I was
rambling due to poor background information/…/”, ”When we didn’t know how big the
common space [in the scenario] was”, and “Waiting for Agnes to give new instructions”.
Apart from this, there was no general source for inefficiency revealed in Q9.

8. It happened on one or several occasions that the m eeting


becam e inefficient.
Number of subjects

2 HuMod
1 CoMod

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
Chairmen of both conditions did not experience inefficiency in the meeting. Since only
subjects who experienced the meeting as inefficient were requested to answer the follow-up
question, there were no relevant answers.

Q10. I thought the goal of the meeting was clear.

10. I thought the goal of the meeting was clear

12
Number of subjects

10
8 HuMod
6 CoMod
4 AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
In both the HuMod and the AgMod condition, all subjects (12 of 12) thought the goal of the
meeting was clear. In the CoMod condition, 8 of 12 believed so, 2 did not and 2 were neutral.

73
10. I thought the goal of the m eeting w as clear.

3
Number of subjects

2
HuMod
CoMod
1

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
All chairmen were more or less content with the clarity of the goal of the meeting. There is no
reason to believe there is any difference between the conditions, related to this question.

Q11. I thought the timeframes of the meeting were clear.

11. I thought the timeframes of the meeting were clear

12
Number of subjects

10
8 HuMod
6 CoMod
4 AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
6 of 12 In the HuMod condition, 5 of 12 in the CoMod condition and 8 of 12 in the AgMod
condition, believed that the timeframes of the meeting were clear. When looking at how many
subjects overtly answered negatively to this question (2 of 12 In the HuMod condition, 6 of 12
in the CoMod condition and 4 of 12 in the AgMod condition), it can be concluded that the
CoMod condition participants were slightly less inclined to believe the timeframes of the
meeting were clear than the other conditions.

11. I thought the tim efram es of the m eeting w ere clear.

3
Number of subjects

2
HuMod
CoMod
1

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

74
Chairmen
All chairmen were more or less content with the clarity of the timeframes of the meeting.
There is no reason to believe there is any difference between the conditions, related to this
question.

Q12. I thought the structure of the meeting was clear.

12. I thought the structure of the meeting was clear

12
Number of subjects

10
8 HuMod
6 CoMod
4 AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
9 of 12 participants in the HuMod condition, 3 of 12 in the CoMod condition, and 10 of 12 in
the AgMod condition thought the structure of the meeting was clear. Since only one person
did not perceive the structure as clear in the AgMod condition, this can be said to be the
condition were the structure was most obvious to participants. Least structure was perceived
in the CoMod condition.

12. I thought the structure of the m eeting w as clear.

3
Number of subjects

2
HuMod

1 CoMod

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
All chairmen were content with the clarity of the structure of the meeting. There is no reason
to believe there is any difference between the conditions, related to this question.

75
Q13. I thought the meeting was focused.

13. I thought the meeting was focused

12
Number of subjects

10
8 HuMod
6 CoMod
4 AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
7 of 12 in the HuMod condition, 3 of 12 in the CoMod condition, and 11 of 12 in the AgMod
condition thought the meeting was focused. It can therefore be assumed that meeting focus
was most obvious to participants in the AgMod condition, and least obvious to CoMod
participants.

13. I thought the m eeting w as focused.

3
Number of subjects

2
HuMod
CoMod
1

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
All chairmen agreed that the meeting was focused. There is no reason to believe there is any
difference between the conditions, related to this question.

Q14. I thought the meeting was disciplined.

14. I thought the meeting was disciplined

12
Number of subjects

10
8 HuMod
6 CoMod
4 AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

76
Participants
8 of 12 in the HuMod condition, 6 of 12 in the CoMod condition, and 11 of 12 in the AgMod
condition thought the meeting was disciplined. It can therefore be assumed that participants in
the AgMod condition experienced most meeting discipline.

14. I thought the m eeting w as disciplined.

3
Number of subjects

2
HuMod
CoMod
1

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
All chairmen agreed that the meeting was disciplined. There is no reason to believe there is
any difference between the conditions, related to this question.

Q15. I thought the atmosphere in the meeting group was positive.

15. I thought the atmosphere in the meeting group


was positive
Number of subjects

12
10
HuMod
8
6 CoMod
4
2 AgMod
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
9 of 12 in the HuMod condition, 10 of 12 in the CoMod condition, and 10 of 12 in the AgMod
condition thought the atmosphere in the meeting was positive. There seems to be little reason
to believe there are any differences in the experience of meeting atmosphere related to the
different conditions. However, 3 people in the HuMod condition, and 6 in the AgMod
condition answered “I totally agree” to this question. In the CoMod condition, no one did.

77
15. I thought the atmosphere in the meeting group was
positive.

3
Number of subjects

2
HuMod
CoMod
1

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
All chairmen, save one who was neutral, felt that the atmosphere in the meeting was positive.
There is little reason to believe there is any difference between the conditions, related to this
question.

Q16. The atmosphere in the group was decisive for how meaningful I perceived
the meeting.

16. The atmosphere in the group was decisive for


how meaningful I perceived the meeting
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
7 of 12 in the HuMod condition, 6 of 12 in the CoMod condition, and 10 of 12 in the AgMod
condition thought that the atmosphere in the group was decisive for how meaningful they
were to perceive the meeting. Many respondents of the CoMod and HuMod conditions were
neutral. It is hard to see any apparent reasons to believe there are any differences between the
conditions related to this question.

16. The atmosphere in the group was decisive for how


meaningful I perceived the meeting.

3
Number of subjects

2
HuMod
CoMod
1

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

78
Chairmen
Most chairmen believed that the atmosphere in the group was decisive for how meaningful
they perceived the meeting. One person of the HuMod condition did not consider this an
important aspect.

Q22. I thought the presentation technology (Tipple, iTable, SMART Boards)


made the meeting more efficient.

22. I thought the presentation technology made the


meeting more efficient
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
6 of 12 in the HuMod condition, 2 of 12 in the CoMod condition, and 6 of 12 in the AgMod
condition thought that the presentation technology made the meeting more efficient.
Participants in the CoMod condition were much harder in judging the technology than
subjects in the other conditions, 6 of 12 answered “I totally disagree” to this question, as
opposed to 2 of 12 in the HuMod condition and 1 of 12 in the AgMod condition. It is hard to
draw any conclusions from these results, since not all participants used the technology
available. There were also some breakdowns in technology, that might have discouraged
participants from using it.

22. I thought the presentation technology made the


meeting more efficient.

3
Number of subjects

2
HuMod
CoMod
1

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
Chairmen of the HuMod condition generally seemed more convinced that the presentation
technology made the meeting more efficient. It is important to remember that the chairmen
themselves did not use any such technology. In some of the sessions, presentation technology
was not utilised at all, and technology breakdowns occurred in almost all meetings.

79
Q23. I thought the presentation technology (Tipple, iTable, SMART Boards) was
better than the presentation technology I normally use.

23. I thought the presentation technology was better


than the presentation technology I normally use
Numeber of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
5 of 12 in the HuMod condition, 1 of 11 in the CoMod condition, and 5 of 11 in the AgMod
condition thought the presentation technology was better than the presentation technology
they usually use. Again, participants in the CoMod condition were harder in judging the
technology than subjects in the other conditions, 5 of 11 did not think the technology was
better, as opposed to 4 of 12 in the HuMod condition and 2 of 11 in the AgMod condition. It
is hard to draw any conclusions from these results, since not all participants used the
technology available. This might be the reason for the two non-responses. There were also
some breakdowns in technology, that might have discouraged participants from using it.

23. I thought the presentation technology was better


than the presentation technology I usually use.

3
Number of subjects

2
HuMod
CoMod
1

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
Chairmen of the HuMod condition generally seemed more positive towards the presentation
technology. The CoMod chairmen were more negative or, at best, neutral. It is important to
remember that the chairmen themselves did not use any such technology. In some of the
sessions, presentation technology was not utilised at all, and technology breakdowns occurred
in almost all meetings.

80
Q24. I thought the presentation technology (Tipple, iTable, SMART Boards)
facilitated the sharing of information.

24. I thought the presentation technology facilitated


the sharing of information
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
4 of 12 in the HuMod condition, 4 of 11 in the CoMod condition, and 8 of 11 in the AgMod
condition thought the presentation technology facilitated the sharing of information. This
time, participants in the CoMod condition were only slightly harder in judging the technology,
5 of 11 did not think the technology was better, as opposed to 4 of 12 in the HuMod
condition. The biggest difference is to the AgMod condition, where no one disagreed to the
question asked. It is hard to draw any conclusions from these results, since not all participants
used the technology available. This might be the reason for the two non-responses. There
were also some breakdowns in technology, that might have discouraged participants from
using it.
24. I thought the presentation technology facilitated
the sharing of information.

3
Number of subjects

2
HuMod
CoMod
1

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
Again, the discrepancy between the conditions is considerable when it comes to the
acceptance of the presentation technology, showing more satisfaction in the HuMod
condition. It is important to remember that the chairmen themselves did not use any such
technology. In some of the sessions, presentation technology was not utilised at all, and
technology breakdowns occurred in almost all meetings.

81
Q28. I perceived it as the chairman/Agnes had control over the meeting.

28. I percieved it as the chairman/Agnes had control


over the meeting
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
6 of 12 in the HuMod condition, 6 of 12 in the CoMod condition, and 8 of 12 in the AgMod
condition perceived it as the chairman / Agnes had control over the meeting. Participants in
the AgMod condition were more enthusiastic than the others, 5 of them answered “I totally
agree” to the question compared to 2 in the HuMod condition and 1 in the CoMod condition.

28. I perceived it as the chairman/Agnes had control


over the meeting.

3
Number of subjects

2
HuMod
CoMod
1

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
All chairmen agreed that the chairman (HuMod), or Agnes (CoMod) had control over the
meeting. There is no reason to believe there is any difference between the conditions, related
to this question.

82
Q35. I believe that the chairman/Agnes made the meeting more efficient.

Q36. If you answered 4 or 5 to the question above, why do you think the
chairman / Agnes made the meeting more efficient?

35. I believe that the chairman/Agnes made the


meeting more efficient
Number of subjects

12
10
8 HuMod
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
Only 3 respondent marked lower than 3 (neither nor) in Q 35 indicating negative responses to
the question, all belonging to HuMod. The other two conditions showed higher agreement,
to the statement that the chairman/Agnes made the meeting more efficient.

For those of the respondents that answered 4 (I agree) or 5 (I totally agree) to Q 35 there was
a follow-up question to explain why the meeting got more efficient with a chairman. Seven
HuMod participants answered, saying that the chairman brought stability and structure to the
meeting. Among the comments were “ The meeting got started quickly because of the
chairman.”, “He saw to it that everyone spoke and that no one took over”. The eight
comments in the CoMod were similar; “Otherwise the meeting would have consisted of
pointless bickering” and “Opening the floor, telling when it is time to move on.”. Interestingly
enough, the AgMod comments expressed even more positivity saying that “We made a
unanimous decision significantly faster than we would have done without Agnes”, “I did not
need to take the floor on my own initiative, instead I was asked to take the floor”, “…,it was
time efficient and a we made a good decision.”.

35. I believe that the chairman/Agnes made the


meeting more efficient.

3
Number of subjects

2
HuMod
CoMod
1

0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Chairmen
All chairmen agreed that the chairman (HuMod), or Agnes (CoMod) made the meeting more
efficient by providing more structure to the meeting. The chairmen that moderated the
meeting on command by Agnes were equally content with Agnes’ ability to achieve

83
efficiency (“There was more focus on “managing the meeting” and avoiding too liberate
discussions”), as the chairmen who moderated the meeting themselves were content with their
own ability to do so (“The chairman added structure to the meeting”). There is no reason to
believe there is any difference between the conditions, related to this question.

Q37. Can you think of anything that could have improved the meeting?

This question was posed in order to capture general attitudes towards the meeting execution.

Participants
8 subjects from the HuMod choose to answer. The most common improvement discussed in
this condition had to do with weaknesses in the scenario; “More facts about the prerequisites
of the room, the size, economic conditions, possibilities to invest, number of residents,/…/” ,
“If the file [pdf-file with arguments] had had better arguments I would have used it” and
“Maybe if we had more time to debate, I would have come up with more arguments”.

The subjects in the CoMod did not comment on the scenario to the same extent, instead they
had technological issues with tipple and the iTable. There were at times problems getting files
from the iTable to open on the SMART Boards, instead an error message occurred on the
screen. Also, since none of the subjects had earlier experience with the touch screen interface
on the iTable, a lot of times subjects dropped their file and had to start over again.

AgMod participant had comments about Agnes; “It should be possible to communicate with
Agnes in some way, either in speech or in writing”, ”Agnes could if it was desired by the
meeting attendants summarise the arguments and try to come up with different solutions”,
”Possibly if Agnes gave us some options like ”Do you want more time/…/”. However,
subjects in AgMod also mentioned technological mishaps and doubts concerning the scenario.

Chairmen
Two chairmen (one for each condition) thought the meeting could have benefited from
participants utilising the presentation technology more. More background information about
the scenario was also requested. One chairman in the CoMod condition would have liked to
be given a set of “alternatives to choose from /…/, instead of being completely controlled [by
Agnes]”.

Q38. Other comments

This question was posed in order to capture general attitudes towards the meeting execution.

Participants
4 participants in the HuMod condition made general comments. 2 were directed towards the
presentation technology, one criticised the scenario. One was just positive in general.

6 participants in the CoMod condition made general comments. Most answers focused on the
presentation technology (“It was not clear that we were supposed to use the iTable”, and “The
iTable was really good”). The lack of information about the scenario was mentioned by 2
subjects.

84
7 participants in the AgMod condition made general comments. 3 of them were overtly
positive, one said that it “seemed to be a good way to arrange discussions and conferences”.
Another said that “It [the situation] was really nice”. One person found it “satisfying to reach
a quick decision” but believed there to be “drawbacks of raising the tempo so considerably”.
Other negative comments were related to the lack of information about the scenario. One
person wanted to see a more personified Agnes, “so that you can relate to her as a real
chairman”.

Chairmen
There were few comments. One chairman of the CoMod condition thought that “Chairmen
that talk too much and find it hard to keep to the agenda would probably feel supported by
Agnes”.

2. Questions for all Subjects except HuMod Chairmen

Q17. I felt respect for the chairman/Agnes.

17. I felt respect for the chairman/Agnes

12
Number of subjects

10
8 HuMod
6 CoMod
4 AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
8 of 12 in the HuMod condition, 9 of 12 in the CoMod condition, and 10 of 12 in the AgMod
condition claim to have felt respect for the person in charge of the meeting. Five people
answered “I totally agree” in the AgMod condition, compared to two people in each of the
other two conditions. This may imply that Agnes was able to induce more respect than the
human chairmen, or at least a stronger sense of respect amongst those persuaded by “her”
authority in the first place. However, since this indication is small, one should be careful to
conclude there is any difference between the conditions related to this question.

Chairmen
All CoMod chairmen felt respect for Agnes.

85
Q18. I was reluctant to following orders from the chairman/Agnes.

18. I was reluctant to following orders from the


chairman/Agnes
Number of subjects

12
10
8 HuMod
6 CoMod
4
2 AsMod
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
There seemed to be no problem for the participants in either condition to follow the authority
of the chairman, whether it was human or artificial.

Chairmen
None of the CoMod chairmen were reluctant to following orders from Agnes.

Q19. If you answered 4 or 5 to the question above, why were you reluctant to
following orders from the chairman / Agnes.

Participants
Since so few were reluctant to following orders, there were just two comments made in the
following open question, both comments came from the CoMod condition, one was not
relevant and the other expressed doubts about the way the chairman conducted the meeting
(“/…/, she [the chairman] governed the meeting in a peculiar way.”).

Chairmen
Since none of the CoMod chairmen expressed reluctance to following orders, they did not
answer this question. However, one person mentioned waiting for “a good opportunity”
before carrying out the order.

Q20. I perceived it as negative not being allowed to talk as much as I wanted to


(only answer this question if you at any time during the meeting were
interrupted by the chairman/Agnes.

20. I percieved it as negative not being allowed to talk as


much as I wanted to
Number of subjects

12
10 HuMod
8
6 CoMod
4
2 AgMod
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

86
Participants
Only people who at any time during the meeting were interrupted by the chairman (in the
HuMod and CoMod conditions) or Agnes (in the AgMod condition) were requested to answer
this question. Hence, there were a lot of non-responses. Of those applicable to answer, 1 of 5
in the HuMod condition, 3 of 8 in the CoMod condition, and 6 of 9 in the AgMod condition
perceived it as negative not being allowed to talk as much as they wanted to. First of all,
answers to this question show that the number of subjects who perceived themselves to be
interrupted increased with the degree of computerisation. This can on the other hand be
circumstantial, meaning that it is possible that subjects in the AgMod condition talked more,
thus being more disposed to the chance of being interrupted. Secondly, a higher degree of
computerisation also seems to increase the participants discontent with being interrupted.
Note should be taken to the small number of responses, and the difficulties of drawing
exhaustive conclusions this leads to.

Chairmen
The CoMod chairmen that answered (2 of 3) were neutral.

Q21. I perceived it as negative being encouraged to talk (only answer this


question if you at any time during the meeting were encouraged to talk by the
chairman/Agnes.

21. I percieved it as negative being encouraged to talk

12
Number of subjects

10
8 HuMod
6 CoMod
4 AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
Of those applicable to answer this question, none of 8 in the HuMod condition, 1 of 9 in the
CoMod condition, and 1 of 11 in the AgMod condition perceived it as negative being
encouraged to talk. First of all, answers to this question show that the number of subjects who
perceived themselves to be encouraged to talk increased with the degree of computerisation.
This can on the other hand be circumstantial, meaning that it is possible that subjects in the
AgMod condition talked less, thus being more disposed to the chance of being encouraged to
talk. It is also possible that they actually were encouraged to talk more often. Perhaps the
meetings with a human chairman did not require as much involvement on behalf of the
chairman. However, as opposed to in Q 20, higher degree of computerisation does not seem to
increase the participants discontent with being encouraged to talk. Rather to the contrary,
more subjects in the AgMod condition answered “I totally disagree” to the question (7 of 11)
than in the CoMod condition (4 of 9) and in the HuMod condition (1 of 8). Note should be
taken to the small number of responses, and the difficulties of drawing exhaustive conclusions
this leads to.

Chairmen
None of the CoMod chairmen perceived it as negative being encouraged to talk. One was
neutral.

87
Q25. I perceived it as the chairman/Agnes strictly followed an agenda.

25. I percieved it as the chairman/Agnes strictly


followed an agenda
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
The majority of participants in the HuMod and AgMod conditions perceived that the meeting
had some form of structure/agenda that the chairman/Agnes enforced, 7 out of 12 in HuMod
and 9 out if 12 in AgMod. However, this was to a lesser extent noticed in the CoMod
condition, and no conclusions from this condition could be reached also due to the large
spread. As in the HuMod condition, 7 subjects expressed agreement but only 1 was neutral
and 4 claimed not to notice a strict agenda.

Chairmen
Two of the chairmen perceived that Agnes strictly followed an agenda, one did not.

Q26. If you answered 4 or 5 to the question above, did you perceive it as


negative that the chairman/Agnes strictly followed an agenda?

26. If you answered 4 or 5 to the question above, did


you percieve it as negative that the chairman/Agnes
strictly followed an agenda?
Number of subjects

12
10 HuMod
8
6 CoMod
4
2 AgMod
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
Since Q 26 only concerned subjects marking 4 (I agree) or 5 (I totally agree) on the previous
question there was a cut back in the number of answers. In each of the two first conditions
there were 7 responses and in the third condition 9 responses. In HuMod and AgMod one
person found the enforcement of an agenda as negative, and in CoMod two persons.

Chairmen
Two of the CoMod chairmen did not perceive it a s negative that Agnes strictly followed an
agenda. One did not answer, since s/he was not requested to answer this question.

88
Q27. Can you develop your answer to the question above?

Participants
There was also a follow-up question tied to Q 26 asking the subjects to further explain how
they felt about strictly using an agenda. Worth noticing is that all four comments made by
CoMod participants complained that the tight structure made the meeting “inflexible”, “stiff”
and “too formal”, while 7 out of 10 comments from AgMod ( 1 subject answered despite not
answering Q 26) were positive saying that Agnes “got the meeting going” and facilitated
“quick decisions” when time was limited. The seven answers from the HuMod condition were
mixed with no general inclination.

Chairmen
Two of the CoMod chairmen did not answer. The third thought that Agnes was “good to have
as a guiding line” but that s/he perceived it as “Agnes seemed to know who was talking” and
“interacted”, implying that Agnes did not actually follow the agenda.

Q29. I perceived the chairman’s/Agnes requests as easy to understand.

29. I percieved the chairman's/Agnes' requests as easy


to understand
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
10 of 11 in the HuMod condition (there was one non-response), and 9 of 12 in the CoMod
condition perceive the chairman’s requests as easy to understand. In the AgMod condition, all
twelve did. In fact, they all gave Agnes’ requests a “top grade” in clarity.

Chairmen
All subjects perceived Agnes requests as easy to understand.

89
Q30. I perceived the chairman’s/Agnes requests as natural.

30. I perceived the chairman's/Agnes' requests as


natural
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
9 of 11 in the HuMod condition (there was one non-response), 7 of 12 in the CoMod
condition, and 7 of 12 in the AgMod condition perceived the chairman’s or Agnes’ requests
as natural. 4 respondents each for the CoMod and the AgMod conditions were neutral,
indicating hesitation. Only one in the HuMod condition chose the middle option of “neither
nor”. The difference between the groups are so small that they are difficult to conclude from,
but if anything, the HuMod chairmen’s requests were deemed more natural.

Chairmen
All CoMod chairmen perceived Agnes requests as natural.

Q31. I thought the meeting was affected in a negative way by the


chairman’s/Agnes directions.

31. I thought that the meeting was affected in a


negative way by the chairman's/Agnes' directions
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
These results to a certain degree followed those of Q 26 although somewhat less inclined to
thinking that the chairman had a negative effect. As in Q 26 the CoMod had the most
problems with directions from the chairman/Agnes. Also, the AgMod condition showed the
least dissatisfaction of the three.

Chairmen
All of the CoMod chairmen were neutral on this item, answering 3 on the Likert scale.

90
Q32. If you answered 4 or 5 to the question above, in what way did you
experience that the meeting was affected in a negative way by the
chairman’s/Agnes directions?

Participants
Q 32 developed the answers from Q 31. Due to a mistake the question was asked differently
in the AgMod, where only subjects answering 4 (I agree) or 5 (I totally agree) to Q 31 were
prompted to comment. Therefore, the opinion that “Agnes forced the group to a quick
decision, when there was not enough ground to make one” is the only answer from this
condition. In order to make a comparison between all conditions, answers coming from
respondents not marking 4 or 5 were ruled out in the other two conditions. The CoMod had
two persons in this range but their comments differed. One felt that there was “No possibility
for dialogue, to meet and develop on arguments” while the other said “… yet I feel she wasn’t
dominant enough, but she tried.”.

Chairmen
Since only subjects who thought the meeting was affected in a negative way by Agnes’
directions were requested to answer this question, two of the subjects did not answer. One
person thought the directions “facilitated [the meeting] at times” but that “Agnes wanted to
end the meeting abruptly”.

Q33. I would have liked to communicate with the chairman / Agnes in another
manner.

33. I would have liked to communicate with the


chairman/Agnes in another manner
Number of subjects

12
10
HuMod
8
6 CoMod
4
AgMod
2
0
Totally Disagree Neither/Nor Agree Totally
disagree agree

Participants
This question asked whether the subjects would have favoured communicating with the
chairman/Agnes in another way, and if they did, Q 34 asked them to specify how. HuMod and
CoMod show similar results, both in the Likert scale and writing. Only 1 person in the
HuMod marked that s/he would have favoured other ways of communication, respectively 3
in CoMod condition. In the AgMod condition, 7 of 11 did.

Chairmen
None of the subjects wanted to communicate with Agnes in another way. One was neutral.

Q34. If you answered 4 or 5 to the question above, how would you have
preferred to communicate with Agnes?

91
Participants
Only one person answered in HuMod and three in CoMod and their suggestions had to do
with easing up the structure of the meeting. The AgMod condition differed since it had no
human chairman and only one-way communication, hence the results were different from the
other conditions. AgMod answers to Q 34 regarded technical ways of interacting with Agnes.
The majority (6 out of 8) wanted to communicate via voice commands with for example a
possibility to ask for more time. Only one person asked for a “/…/graphic 3D avatar/…/”

Chairmen
Since only subjects who would have liked to communicate with Agnes in another way were
requested to answer this question, there were no answers.

3. Questions posed to AgMod participants and CoMod Chairmen

These questions were only posed to participants of the AgMod condition and Chairmen of
CoMod condition, since these questions only oncerned the perception of Agnes.

Q39. I perceived it as negative not being able to talk to Agnes.

Most participants (9 of 12) of the AgMod condition perceived it as more or less negative not
being able to talk to Agnes. Only one person did not see this as a problem, while two subjects
were neutral.

Amongst the CoMod chairmen on the other hand, no one agreed that lack of ability to talk to
the assistant was problematic.

Q40. I believe that the meeting would have functioned better with a human
chairman.

4 of the 12 participants believed that the meeting would have functioned better with a human
chairman, whereas 4 did not believe so. It is therefore not possible to identify a specific
inclination of attitudes.

One of the chairmen believed that the meeting would have functioned better with a human
chairman. Two of them were neutral.

Q41. If you answered 4 or 5 to the question above, why do you believe that the
meeting would have functioned better with a human chairman?

Human qualifications like being able to “see and listen to participants”, “collaborate with
people” and “sum up people’s opinions” were desired. Three participants implied that they
missed the superior role that a chairman often has in a meeting, “having the last say”, while
one person actually liked the idea of not directing the attention to an “omnipotent” person.
One person stated that with a human chairman with the same authority that Agnes had “more
people would have felt satisfied after the meeting, and not as stressed out”.

Of the chairmen, one thought that a human chairman would make the meeting “more natural
and less artificial”.

92
Q42. I experienced it as negative receiving commands from Agnes.

10 of 12 participants did not experience it as negative receiving commands from Agnes, one
person did and one was neutral.

None of the chairmen perceived it as negative and one did not answer.

Q43. It was easy to follow Agnes’ commands.

All participants considered it easy to follow the commands given by Agnes.

Of the chairmen, two thought it was easy while one did not.

Q44. It would have been easier to administer the meeting without Agnes.

9 of 12 participants did not agree that it would have been easier to administer the meeting
without Agnes. One person did agree, while two were neutral.

Two chairmen were neutral, one thought a meeting without Agnes would not have been easier
to administer.

Q45. How do you believe that the technology behind Agnes works?

As a control question to detect if subjects believed that there was real agent technology
operating behind Agnes, it was asked how subjects believed that Agnes was technically
implemented. Only one of the participants and one of the chairmen guessed, correctly, that
Agnes was in reality operated by a person.

4. Questions for Chairmen of the HuMod condition

These questions were only posed to chairmen of CoMod condition, since they concerned the
perception of managing the meeting without support.

Q46. I thought it was hard to follow the agenda.

None of the chairmen thought it was hard to follow the agenda

Q47. I thought the agenda supported me in controlling the meeting.

Two of the chairmen thought the agenda supported them in controlling the meeting. One was
neutral.

Q48. I perceived it as negative having to strictly follow the agenda.

One of the chairmen perceived it as negative having to strictly follow the agenda. Two were
neutral.

93
Q49. If you answered 4 or 5 to the question above, in what way did you
perceive it as negative having to strictly follow the agenda?

Since only subjects who perceived it as negative having to strictly follow the agenda were
requested to answer this question, two of the subjects did not answer. One thought that “an
agenda is good, but it is easier to follow an agenda that you have created yourself”.

5. Questions for Participants of AgMod condition

These questions were only posed to participants of the AgMod condition, since they
concerned the perception of the meeting session where Agnes was visible to all.

Q50. Agnes making a sound was a good way of calling for attention.

All participants (12 of 12) thought that Agnes making a sound was a good way of calling for
attention.

94
10 APPENDIX III – AGNES’ COMMANDS
Commands

This is a list of the 12 possible types of commands issued by Agnes in the experiment:

1. Welcome! The meeting is declared open.

2. The parent has the floor.


The rock musician has the floor.
The hobby florist has the floor.
The gymnast has the floor.

3. Please end your presentation.


4. Do not interrupt the person who has the floor.
5. Discuss in order to reach a mutual agreement.
6. Try to reach a conclusion.
7. Please round off the meeting.
8. Vote to reach a conclusion.
9. Please return to the topic.
10. The meeting is now over.
11. The meeting must be adjourned.

12. Does the parent have anything to add?


Does the rock musician have anything to add?
Does the hobby florist have anything to add?
Does the gymnast have anything to add?

95

Вам также может понравиться