. .r
,.~wm.m...h...v

um.

s:
‘\ L
ha," ..

 

 

 

 

 

‘10:: , .0!—

.
a, infirm“.
11:51.” A.»

. A.
> x
maxi“. .2

.. "L in. . . .. . . r... ,
“3.13.. _ _ 7 . ‘ a Eufifi gayfié.

 

 

 

 

n1.

:4 cc 4
50§w33”1

LIBFMHUY
Michigan State
University

This is to certify that the
dissertation entitled

EXPLORING EXPLANATIONS FOR THE ILLUSION OF
GROUP PRODUCTIVITY

presented by

Emest Sunwoo Park

has been accepted towards fulfillment
of the requirements for the

Ph. D. degree in Psychology

 

 

 

Major Professor’s/Signature

December 9, 2003

 

Date

MSU is an Affirmative Action/Equal Opportunity Institution

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c;/ClRC/DateDue.p65-p.15

EXPLORIN G EXPLANATIONS FOR THE ILLUSION
OF GROUP PRODUCTIVITY

By

Ernest Sunwoo Park

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Psychology

2003

ABSTRACT
EXPLORING EXPLANATIONS FOR THE ILLUSION OF GROUP PRODUCTIVITY
By

Ernest Sunwoo Park

Although brainstorming groups are often used to generate ideas, there is little (if
any) empirical evidence to suggest that groups are actually more productive at idea
generation than the same number of individuals brainstorming alone. In fact, results from
numerous studies have found that interacting groups are actually much less productive in
comparison to the combined products of the same number of individual brainstormers
(known as nominal group members). It is possible that this seemingly irrational trend of
assigning groups for brainstorming purposes is perpetuated by an interesting effect, called
the illusion of group productivity. The illusion of group productivity refers to a
consistent finding where group members report feeling more satisfied with their personal
performance at the task relative to individuals who brainstormed alone, despite the fact
that group members are generally being outperformed. The purpose of the current study
was to examine possible explanations for this illusion effect, in particular, to examine the
potential role of mood influences on performance perceptions.

Participants were 231 female undergraduates from Michigan State University.
The design was a 2 (group: nominal vs. face-to-face) X 2 (mood: neutral vs. negative)
between-subjects factorial. Participants were asked to brainstorm on a topic either as a

member of a 3-person nominal or face-to-face group. Mood was then manipulated, and

participants were subsequently asked to make judgments about their brainstorming
performance.

Results show the illusion effect was again replicated. However, there was no
evidence for the hypothesis that mood mediates the relationship between group condition
and perceptions of performance. Other possible explanations besides mood were
examined and are discussed, as are potential implications and future research

possibilities.

To Norbert Kerr, Lawrence Messé, Gwen Wittenbaum and Richard Lucas for sharing
their knowledge and guidance over the years. And, to Kristy Dean, my family, and

friends for all of their support.

iv

TABLE OF CONTENTS

INTRODUCTION ............................................................................... l

NIETHOD ........................................................................................ 46
RESULTS ........................................................................................ 64
DISCUSSION .................................................................................... 92
APPENDICES .................................................................................. 105
REFERENCES ................................................................................. 129

Introduction

Group researchers can attest to the basic notion that groups differ from collections
of individuals. This principle is perhaps the foundation or building block for the
fascination that has inspired and enthralled group researchers over the years. Being a part
of a group, or working within group settings psychologically “feels” different than being
or working alone. In some sense, the fact that arousal increases in the mere presence of
others helps to support this claim (e. g., Zajonc, 1980a; Carver & Scheier, 1981), as do the
numerous surveys (e.g., Stroebe, Diehl, & Abakoumkin, 1992; Paulus, Dzindolet,
Poletes, & Camacho, 1993) which show that a majority of people favor working on tasks
in groups rather than by themselves. So for whatever reason, being in a group feels
different from being alone, and being in a performance group is generally more desired
than performing alone. The true significance and implications of this statement however
do not fully surface until one begins to consider and incorporate some of the ideas that
have been born out of the social cognition area.

One major contribution from social cognitive psychologists has been the
discovery that affective states influence judgments and information processing (e. g.,
Bruner, 1957; Zajonc, 1980b; Schwarz & Clore, 1983). To date, a wide range of research
has been devoted to understanding how and why feelings affect cognitions (see Eich,
Kihlstrom, Bower, Forgas, & Niedenthal, 2000; Forgas, 2001; and Martin & Clore, 2001
for reviews). So building on the assumption that the affective experience differs between
people working with others versus working alone, one can conclude that under some

circumstances group members are likely to process information differently than people

who are by themselves. Therefore, some important questions for group researchers to
examine are: l)“What are people feeling when they are in groups?” 2) “How do their
feelings affect their cognitions?” 3) “How do their cognitions affect their perceptions?”
and 4) “How do their perceptions affect their behavior?” Conveniently, to attempt to get
at the heart of many of these questions, one can turn to an interesting effect that has been
consistently found in small group research: “the illusion of group productivity.”

The illusion of group productivity effect

In laboratory studies on idea generation, face-to-face (ftf) groups are consistently
outperformed by the same number of individuals working at the task alone (e.g., Diehl &
Stroebe, 1987; Stroebe et al., 1992; Paulus et al., 1993). Despite the abundance of
empirical research illustrating this effect, the procedure of group brainstorming persists as
a popular and widely used technique for problem-solving and idea generation. So why is
group brainstorming so common if it generally yields sub-optimal results?

One possible explanation may follow from an interesting perceptual effect, a
finding which has been aptly dubbed by Diehl & Stroebe (1987), “the illusion of group
effectivity (productivity).” Past researchers (e. g., Diehl & Stroebe, 1987; Stroebe et al.,
1992; Paulus et al., 1993; Nijstad, 2000) have found that people have a persistent belief
that: 1) groups can stimulate creativity, 2) that relative to working alone, working in
groups can lead to enhanced personal performance and 3) that groups as a whole are
more effective and productive than the same number of people working alone (the latter
are commonly referred to as nominal groups). In line with these beliefs, participants who
work on a task in a group are generally more satisfied with their personal performance

than participants who work by themselves. These beliefs and overestimations in

performance are often referred to as the illusion of group productivity because although
ftf groups are generally not more productive than nominal groups people believe they are,
and although participants feel more satisfied when working in groups compared to
working alone, they are actually less productive in group settings (Nijstad, 2000). This is
an interesting and consequential effect because the feelings of satisfaction that are
experienced by ftf group members may: a) inadvertently lead to poorer performance
(Paulus et al., 1993) and, b) may contribute in part to the persistent use of task groups in
the work place.

Now unfortunately, to date, there have been several discrepancies in how the
illusion is defined. In the abstract of their 1992 paper, Stroebe et al. state that, “the
illusion of group effectivity refers to the belief, persistent despite contradictory empirical
evidence, that groups can stimulate creativity.” However, within the introduction of the
same paper, the authors refer to the illusion of group effectivity as a phenomenon where
people are more likely to embrace working in groups, “in spite of consistent empirical
evidence that people produce many more ideas when they work individually rather than
in groups.” This second conceptualization of the illusion effect differs from the first in at
least one significant way. The essence of the illusion as defined in their abstract (Stroebe
et al., 1992) is based on the notion that groups stimulate creativity. However, in the latter
conceptualization of the illusion, productivity and not creativity is what is emphasized.
Therefore, there seems to be some ambiguity as to whether the illusion effect refers to
perceptions of creativity versus productivity.

However, later in the paper when discussing their results, Stroebe et al. (1992)

suggest that they have “persuasive evidence for the operation of the illusion of group

effectivity,” because: 1) compared to participants who worked individually, group
members were more satisfied with their performance and felt more at ease in the
brainstorming session despite the usual paradoxical performance effect, and because 2)
participants who worked individually believed that they would have had many more ideas
if they had been in a group, whereas group members did not believe that they would have
done better individually. Given these conclusions offered by the authors, and considering
the proposed mechanism that the authors develop to explain the illusion effect (which
will be described subsequently), one can only reach the conclusion that the authors are
conceptualizing the illusion effect to be one of an overestimation of performance by
group members which is based on productivity, and not creativity. Admittedly, Stroebe
et al. would be likely to argue that the point is moot, because quantity has been found to
be highly correlated with quality (Diehl & Stroebe, 1987). However, I believe the
inconsistency should at least be noted because how concepts are operationalized can have
profound effects on how one chooses to measure the phenomenon.

To further complicate the matter, Paulus et a1. (1993) published a paper outlining
a series of studies on the illusion of group productivity. Paulus et al. defined this illusion
effect as one where members of ftf groups overestimate their perceptions of efficacy.
Paulus et al. demonstrated this effect by showing that participants in ftf brainstorming
groups consistently rated “the number of ideas (from very few to very many) that they
personally came up with” significantly higher than participants in nominal groups,
despite the usual performance differences that favored nominal group members. As one
can see, the illusion of group productivity bears a striking resemblance to the illusion of

group effectivity. The only apparent difference seems to be in how the illusion is

measured. In Stroebe et al. (1992), participants rated how satisfied they were with their
performance, whereas in Paulus et al. (1993) participants made ratings on the number of
ideas that they personally generated.
Now interestingly in 2000, Nijstad, a former student of both Paulus and Stroebe,
published a series of studies on brainstorming groups. In his report, Nijstad (2000) states:
Most participants working in the laboratory, as well as in organizations, believe
that group brainstorming is more effective than individual brainstorming (e.g.,
Stroebe et al., 1992; Paulus et al., 1993; Paulus, Larey, & Ortega, 1995). This is
true even after repeated experiences with group or individual brainstorming. In
line with this finding, participants who have worked in groups are generally more
satisfied with their performance than participants who have worked individually
(e. g., Diehl & Stroebe, 1987). Researchers have called these beliefs the ‘illusion
of group effectivity’ (Stroebe et al., 1992) or the ‘illusion of group productivity
(Paulus et al., 1993). (pg. 5)
From his synopsis, it is evident that Ni jstad assumes that the illusion of group effectivity
and productivity are interchangeable concepts, and he assumes that the illusion is both a
general belief in the superiority of groups, and an “unwarranted,” and increased sense of
satisfaction from ftf group members (or conversely, an “unwarranted” and decreased
sense of satisfaction from nominal group members).
Curiously, in his empirical studies on the illusion effect, Nijstad (as do Stroebe et
al., 1992) concentrates on the “inflated” ratings of satisfaction that are found among the
ftf group members, and does not again refer to the illusion in terms of a general belief in

the superiority of groups when proposing explanatory mechanisms that possibly underlie

the effect. Therefore, the most consistent (and preferred by the author of this study)
conceptualization of the illusion of group productivity or effectivity appears to be one

where members of ftf groups feel more satisfied with their personal performance

 

compared to members of nominal groups, in spite of the performance superiority of
members of nominal groups (oddly, although Paulus et al. (1993) never measured ratings
of satisfaction, Nijstad (2000) continually interprets the differences found in Paulus et al.
on the “rate the number of ideas generated” measure in terms of differences in
“satisfaction”). An additional advantage of using “satisfaction” as a measure rather than
the measure used by Paulus et al. (1993), is that satisfaction measures can be applied in
studies that use paradigms other than idea generation tasks. However, in the present
study, all of the various ways that have been used to define the illusion were considered.
Given that the illusion of group productivity has been consistently documented in
brainstorming studies and seems robust and replicable, the interesting psychological
question that remains to be answered is, “why are people working in groups more
satisfied with their performance relative to nominal group members when their actual
performance does not reach the standard of those in nominal groups?”
Summary of alternative explanations for the illusion of group productivity
Misappropriation of ideas. Several researchers have tried to explain the illusion
of group productivity. For example, Stroebe, Diehl, & Abakoumkin (1992) suggest that
the illusion of group productivity stems. from misappropriating ideas to an incorrect
source. In other words, they suggest that ftf group members may be more satisfied with
their performance at a brainstorming task because they may be inadvertently taking credit

for ideas that were brought up by others in their group. This explanation is based on two

assumptions: 1) individuals share a common motivation to see themselves and their own
performance in a positive light, and 2) people who generate ideas in groups (which refers
to “face-to-face” of ftf groups unless otherwise noted) have difficulty differentiating their
own ideas from those ideas generated by other group members. So, because group
members are motivated to perceive themselves as performing well and hence, their
imperfect recall is biased, they will be more likely to claim ideas that were originally
stated by others as their own. This act of misappropriating ideas may therefore lead to
enhanced perceptions of personal performance relative to nominal group members, who
cannot make such misappropriations.

Although the rationale for the misappropriation explanation is plausible, “direct”
empirical evidence supporting this hypothesis is sparse and relatively weak. In their
study (Stroebe et al., 1992), participants were first asked to take part in a brainstorming
task, either individually or in groups of four. They were given the usual brainstorming
rules (i.e., the more ideas the better; say anything that comes to mind; improve or
combine ideas already suggested; do not be critical, etc; see Osborn, 1957 for a more
detailed description) and were then asked to brainstorm on the issue of improving traffic
safety in their country. Participants were led to believe that the purpose of the study was
to compare the productivity of persons working individually and persons working in
groups. All participants were told that for the purpose of such comparisons, the ideas of
those working individually would be combined with ideas from the other three
participants in their session to form a “nominal” group product.

A clip-on microphone was attached to all participants so ideas could be accurately

traced to the original source. The brainstorming rules were reiterated, and participants

were then given 15 minutes to work on the task. After the task was completed,
participants filled out a post-experimental questionnaire that among other things, asked
them to estimate the number of ideas produced by their group (nominal or ftf) and the
percentage of those ideas that they themselves had generated. Participants were also
asked to state how satisfied they were with their personal performance, how at case they
had been during the task, and whether they would have had more ideas if they had
worked in a group (individual participants) or if they had worked alone (group
participants).

To score the performance of each group, first the ideas were transcribed from the
tape recordings. Each idea was written on a separate card, with the participant number
written on the back. Ideas that were generated more than once were not used when
calculating the total number of ideas produced by each group. Results from this study
show that the number of non-redundant ideas produced by nominal groups (M = 110.83)
was significantly higher than the number of ideas produced by ftf groups (M = 84.55).
Participants in nominal groups estimated the number of ideas produced by their group to
be M = 35.46, whereas those working in ftf groups estimated the total number of ideas to
be M = 22.67. This difference was statistically significant, illustrating nominal group
participants reported a substantially mgh_e_r estimate of their group’s productivity, in
apparent opposition to the illusion of group productivity]. In direct contradiction to the
misappropriation explanation, working condition had no effect on participants’ estimates

of what percentage of ideas they themselves had come up with (M nominal = 26.74%; M

 

I Admittedly, this is a curious finding. Perhaps, those participants who were assigned to work in ftf groups
came to the conclusion after working in their group, that obstacles such as production blocking (Diehl &
Stroebe, 1987) exist. If this is true, then one could assume that the belief in the superiority of groups
should diminish after one has some direct experience of being in a work group. Regardless, what is of

ftf: 24.48%). If anything, this finding demonstrates that nominal participants felt they
were generating more ideas than ftf group participants. For example, if nominal
participants believed their group came up with 35 ideas, than 27% of that total would be
approximately 9 ideas. On average, ftf group participants on the other hand estimated
that their group came up with 23 ideas; 25% of that total would be approximately 6 ideas.
It should also be noted that participants in both work conditions estimated that they
themselves had contributed to approximately 25% of the group product. This is a
reasonable and socially appropriate estimate considering that four person groups were
used, however again, this finding directly contradicts the misappropriation of ideas
explanation.

Interestingly, despite the lack of difference in percentage estimates, participants in
the ftf group condition were both significantly more satisfied with their personal
performance and enjoyed the experience more. In addition, whereas participants working
individually felt they would have had many more ideas if they had worked in a ftf group,
participants in the ftf group condition did not feel they would have had as many ideas had
they brainstormed alone. 2

So to this point in my review, it appears that there is support for the existence of

the illusion of group productivity in certain regards, viz. participants in ftf groups were

 

importance for this paper is why people have a greater sense of satisfaction after working in a ftf group
compared to a nominal group.

2 The wording from the Stroebe et al. (1992) is somewhat vague; it is unclear as to whether the participants
felt that their performance in a nominal group setting would equal their performance in a face—to—face
setting, or whether their performance in a nominal group setting would be inferior to their performance in a
face-to-face setting. It should also be noted that from their report, the exact nature of the questions that
were proposed to the participants could not be determined. In other words, from reading their article, it is
unclear as to whether the participants were presented with Likert-scale questions (e.g., “how enjoyable was
the experience?”) or with dichotomous-scale questions (e.g., “in which condition would you enjoy the
experience more?”). To clarify these issues, I contacted the authors multiple times, however, they never
responded.

statistically more satisfied with their personal performance, despite the absence of any
empirical justification for such perceptions. In addition, and in line with the illusion
effect, Stroebe et al. (1992) found that nominal group participants believed they would
have been more productive working in a ftf group.

To further test their proposed misappropriation explanation, Stroebe et al. (1992)
had their participants come back two weeks after their initial brainstorming session. Prior
to working on the task, participants were told during Session 1, that the purpose of the
study was to compare the productivity of persons working individually versus individuals
working in groups. As a cover story, they were then informed that “because experience
had shown that people had many ideas only after the end of a brainstorming session, they
would be asked to return for a second session where they would be shown the ideas
produced by their group and would be asked to list the ideas they had had in the
meantime.” Eighty-six of the original 92 participants returned for Session 2.

Upon arrival, participants were given the total set of non-redundant ideas
generated by their nominal or ftf group, and were asked to assign each idea to one of the
following categories: 1) suggested by me; 2) suggested by another group member but
had also occurred to me; and 3) suggested by another group member and had not
occurred to me. It was emphasized that the “also occurred to me” category should be
used only if the idea had occurred to the participant during the first session. Compared to
participants in the nominal group condition, ftf group members assigned a significantly
lower percentage of ideas to the categories of “suggested by me” and “suggested by
another group member and had not occurred to me.” Participants in the ftf group

condition however, assigned a significantly higher proportion of ideas to the category of

10

“suggested by another group member but had also occurred to me.” When participants in
ftf groups were asked their reasons for not saying an idea, the typical response was that a
similar idea had already been suggested. Nominal group participants on the other hand,
typically stated that they had thought that the idea was “not original enough to be
mentioned.”

To test their assumption that ftf group members would be less accurate than
members of nominal groups in their recall or appropriation of ideas, Stroebe et al. also
measured the number of “hits” and “misses” by each participant. A “hit” was assigned
for the ideas that participants actually suggested and identified as having suggested or for
those ideas that the participants correctly identified as not having suggested themselves.
“Misses” were assigned for ideas that participants actually suggested but later did not
identify as having suggested, or for those ideas that participants erroneously identified as
having been suggested by themselves. Results show that there was no difference between
conditions in bit rates for ideas that they had not suggested, but there was a significant
difference in hit rate for the ideas that they had suggested. More specifically, participants
who worked in ftf groups claimed fewer ideas as their own, even though they did suggest
the ideas in the initial session. Note that this finding also directly contradicts the
misappropriation explanation offered by Stroebe et al. (1992).

However, Stroebe et a1. argued they had support for their misappropriation
hypothesis because they did in fact find that members of ftf groups had less confidence in
the accuracy of their assignment of ideas. In addition, when considering the proportion
of ideas that participants claimed occurred to them, regardless of whether or not they

suggested them, ftf group members asserted that 61% of the total set of ideas occurred to

11

them while nominal group members claimed that only 47% of the total ideas occurred to
them. From their report, it is unclear whether this difference is significant3.

Stroebe et al. suggest that the failure to directly support their hypothesis was
perhaps due to the fairness norms or self-presentation (e.g., modesty) norms that operate
in face-to-face groups. In other words, when the source of an idea is not confidently
known, or when the idea is redundant, group members are more likely, relative to
nominal group members, to give credit to fellow group members rather than to oneself.
The authors argue that because these norms are not relevant to the categorization of
“suggested by another but also occurred to me,” the results show higher estimates for this
category by members of real groups. Unfortunately, Stroebe et al. (1992) do not provide
any direct evidence to show these norms affected the way participants made their
categorizations.

So, Stroebe et al. (1992) demonstrated the illusion of group productivity by
showing that ftf participants were more satisfied with their personal performance
compared to nominal participants despite actually performing worse, and by showing that
most participants felt that they would have been more productive (or at least equally
productive) under ftf group conditions. These are fascinating findings given that the
empirical results on actual performance contradict these perceptions. Although Stroebe
et al. (1992) in many ways pioneered the research on this interesting phenomenon, the
empirical evidence in support of their misappropriation explanation is mixed at best.

A Social Comparison Explanation. Another explanation for the illusion of group

productivity comes from Paulus et al. ( 1993), who explain the illusion using social

 

3 A test of significance was not conducted because I could not tell from the article, how many participants
were in each condition. All that was reported was that 86 of the original 92 participants returned.

12

comparison theory. These researchers argue that when people are in novel situations or
performing novel tasks, they are uncertain about their relative ability level.
Conveniently, one inherent aspect of being in a group setting is the provision for people
to have the opportunity to compare their own performance to that of others.
Extrapolating from findings from Paulus and Dzindolet (1993) which show that
brainstorming group members exhibit a tendency towards uniformly low levels of
performance, Paulus et al. (1993) argue that social comparison processes should lead
most group members to discover that their own performance level is similar to or better
than others’. Consequently, group members may perceive their own performance more
favorably than is actually warranted. Paulus et al. argue that this process is analogous to
the Goethals and Darley (1987) claim that, “groups of similar people. . .with relatively
modest abilities can maintain a positive self-evaluation based on their comparisons with
the restricted range of abilities that is represented within their group.”

Therefore, Paulus et al. suggest that when an individual can compare her or
himself to others, because she/he is likely to find that her/his performance level is
equivalent to or better than others (if the socially defined level of performance is below
the mean as Paulus et al. suggest), the person will be quite satisfied with her/his
performance. In addition, reductions in feelings of uncertainty which result from the
opportunity to make ability comparisons may increase perceptions of satisfaction. It
should be noted, however, that Paulus et al. (1993) did not explicitly argue that the
reductions in uncertainty about one’s relative ability level is what leads to increases in

satisfaction. Rather, it is actually in Nijstad’s (2000) summary of Paulus et al.’s (1993)

13

social comparison explanation where this argument is explicitly made. In his synopsis of

the social comparison explanation, Nijstad (2000) states:
Their central assumption is that people are motivated to compare their
performance with that of others (social comparison). This is generally the only
possible way to evaluate performance, because no objective standards are present.
Group members, who have the opportunity to compare their performance with
that of fellow group members, will usually find that their performance is quite
similar (also see Paulus & Dzindolet, 1993), which will reduce uncertainty and
result in high levels of satisfaction. Individuals, who do not have this opportunity,
will be insecure and may presume that they have not done particularly well. (pg.

121)

In line with Stroebe et a1. (1992), Paulus et al. (1993) also noted that having the
opportunity to compare oneself to others may increase the probability of
misappropriating ideas which may increase perceptions of performance. So Paulus et al.
argue that the ability to compare oneself to others, which is found only in ftf groups and
not nominal groups, is essential for boosting one’s perceptions of performance.
However, the exact details of this relationship are left quite vague.

In their first study, Paulus et al. presented participants with the traditional
brainstorming instructions (e. g., generate as many ideas as possible without concern for
their quality; say anything that comes to mind; combine and improve on ideas already
generated; and do not criticize) and then asked them to rate the extent to which they
would perform better in groups or alone on a hypothetical brainstorming task. So after

being given a brainstorming topic, participants were asked to answer, “If you had been

14

asked to brainstorm in a group with three other people on this problem, do you think you
personally would have generated more or less ideas than if you were brainstorming
alone?” Participants responded to this question and a similar question pertaining to the
quality of ideas, using a 9-point Likert-scale ranging from “many more ideas in a group”
to “many less ideas in a group.” They were also asked to explicitly indicate how many
ideas they thought they could generate in a 15 minute period if they were working alone,
as well as if they were working in a group with three other members. And, participants
were asked to estimate how many ideas a group of four other individuals at their
university could generate.

Consistent with previous findings which illustrate a bias in favor of groups,
Paulus et al. found that people expected to personally generate more ideas (and better
quality ideas) in a group (M = 14.32) compared to when working alone (M = 11.56).
Interestingly, results also showed that participants estimated that they would generate
more ideas alone than the number of ideas a group of four would generate divided by four
(M = 6.74).

In addition, in attempts to document a general bias in favor of groups, Paulus et
al. also categorized participants into those that favored group brainstorming (less than the
midpoint) and those that favored solitary brainstorming (greater than the midpoint).
Results showed that 65% of the participants thought they would generate more ideas in a
group, while in comparison, only 30% thought they would generate more ideas alone.
Fifty-one percent of the participants surveyed thought they would generate better-quality

ideas in a group setting, while only 24% thought they would do better alone.

15

In a follow up study (Experiment 3), Paulus et al. (1993) actually had participants
work at a brainstorming task. To examine the role of social factors in performance
perception, the experimenters manipulated the degree of information provided to the
participant concerning one’s partner. After an initial session of brainstorming alone, half
of the participants were paired with a partner and asked to continue brainstorming, but
this time with the opportunity of sharing ideas with one another via using a common
receptacle in which the written ideas were deposited. Participants in this “share”
condition used different color pens so that the source of each idea could be identifiable.
Using this common receptacle not only provided participants with an opportunity to look
over the ideas of one’s partner, but also provided the opportunity to monitor the pace at
which one’s partner was working.

The other half of the participants were also paired with a partner, however, they
were instructed to continue brainstorming without sharing ideas. In other words, they
were in the presence of a partner, but after they wrote their ideas down they deposited
them in an individual bin, taking away the opportunity to share ideas with one another".
In addition, after their initial individual brainstorming session, half of all of the
participants were told the number of ideas that they themselves generated and the number
of ideas that were generated by their fellow group member during the initial independent
brainstorming session. The experimenter presented this information aloud to the
participants at the end of the break period which followed the initial independent

brainstorming session. The other half of the participants were given no such information.

 

‘ From the article, is unclear as to whether participants in this condition could see their partner working,
and hence, gauge how quickly he/she was depositing ideas in the box

16

This performance feedback manipulation was orthogonal to the idea sharing manipulation
previously described.

So Paulus et al. suggest that because social comparison information leads to
enhanced perceptions of performance, providing performance feedback should lead to
positive perceptions. The sharing of ideas should have a similar positive effect on one’s
self-evaluation. Paulus et al. do not go on to explain how social comparison processes
would lead one to have rnisperceptions concerning not only their personal performance,
but their group’s performance as well. This is further support for the argument that the
illusion of group productivity should be conceived of in terms of “inflated” perceptions
of personal performance rather than “inflated” perceptions on a more global group level
(e.g., a general belief in the superiority of groups).

Results from Paulus et al. (1993) show that neither the sharing nor information
manipulations affected the overall performance of the brainstorming pairs. This finding
is not surprising considering that the performance of brainstorming pairs is generally no
different from that of nominal groups of two (e. g., Diehl & Stroebe, 1987; Mullen,
Johnson, & Salas, 1991). If the opportunity for social comparison increases evaluation
apprehension for group members and therefore motivates members to perform well, then
one would expect increased performance under conditions of social feedback and sharing.
Because these manipulations did not affect overall performance, it is difficult to argue
that the illusion effect is driven by a social comparison process that increases motivation
and effort. However, when participants were provided with feedback, and therefore
knew how well they and their partner performed during Session 1, they gave significantly

higher ratings on the number of ideas they personally generated during Session 2

l7

compared to those who received no performance feedback information (M feedback =
5.45; M nofeedback = 4.5).

Likewise those who were in the “share” condition, who were sharing a common
receptacle and were therefore provided with knowledge of the others’ productivity,
reported significantly higher ratings on the number of ideas they personally generated
during Session 2 relative to those who used individual receptacles and had no opportunity
to compare (M share = 5.5; M no share = 4.6). And, as Paulus et al. would predict,
participants who were given both feedback information and who were allowed to share
ideas, reported the highest ratings on the “number of ideas” measure relative to the other
three conditions (M feeback and share = 5.8; M no feedback_and no share = 4.1; M no
feedback and share = 5.2; and M feedback and no share = 5.1). Similarly, participants
who were neither given feedback nor the opportunity to share ideas reported the lowest
ratings for number of ideas generated. These results support the notion that that being
able to compare one’s performance to others is a contributing factor to the illusion of
group productivity, however it is still unclear as to why or how.

“Gestalt ” of the group product. Paulus, Brown, & Ortega (1999) suggest that one
basis for the illusion of group productivity may deal with the “gestalt” of the group
product. When people work together in groups, there is a general tendency to focus on
the group rather than on individual products. The group product may appear to be quite
significant due to its absolute size, and will inevitably be larger than that of individual
performers. Paulus et al. (1999) argue that:

...although group members should cognitively divide the performance by the

number of group members to determine the efficacy of group performance, they may not

18

appropriately perform this operation. As a result, the mere size of the group contribution
may inflate perceptions of both the efficacy of the group’s overall performance and the
individual’s judgement [sic] of their contribution. (pg. 162)

This explanation is consistent with the previously reported finding from Study 1
of Paulus et al. (1993), where participants reported thinking they would be more
productive working in a group (rather than individually), but also reported that they
would come up with more ideas brainstorming alone compared to the number of ideas a
group of four would generate divided by four. It is also consistent with work conducted
by Josephs, Giesler, & Silvera (1994). In their study, Josephs et a1. had participants take
part in a proofreading task. In one condition, the sheets of paper that the participants
were proofreading were attached to empty cardboard boxes or journals. In another
condition, nothing was attached to the sheets of paper. Throughout the task, participants
were instructed to stack all of the sheets that they successfully completed proofreading.
After the task was completed, participants made assessments of their performance. Those
who were in the cardboard box condition (and who’s stacks subsequently appeared
larger) rated their performance more favorably than those who had nothing attached to
their sheets, even though they did less work. Importantly, when the products were not
visible to the participants, no such effects occurred.

Experiment 2 of Paulus et a1. (1993) may indirectly speak to this “gestalt”
explanation. In this study, participants were asked to perform a brainstorming task either
alone or in same-sex groups of four. Half of the ftf groups were instructed to brainstorm
using one common microphone and a tape recorder, while participants in the other groups

(the other half of the ftf group condition as well as the individual condition) were

19

provided with their own individual microphones and recorders. The authors
hypothesized that: 1) participants working in groups would rate their performance more
favorably than those working in isolation, and 2) that participants working with a
common microphone would perceive their performance more favorably than those in the
ftf group/individual microphone condition.

The logic outlined by Paulus et a1. follows that if working on a common group
product affects perceptions of performance by leading members to focus on the group
product rather than their individual contributions, than participants in the common
microphone condition should view their performance more favorably. The authors note
that even though those in the group/individual microphone condition are also working on
a common group product, they should be more attuned to their own personal performance
than those in the common microphone condition. Because of an increased sense of
accountability by those in the individual microphone condition, the authors also predicted
that participants in the group/individual microphone condition would outperform those in
the common microphone condition.

In their study, participants in the ftf group conditions were seated around a
rectangular table, were given the traditional brainstorming instructions, and then
brainstormed in their groups of four for 25 minutes. For the common microphone
condition, a microphone was placed in the middle of the table. For the individual
microphone condition, a small microphone was attached to each participant. Those in the
alone condition, also had a small microphone attached to them, however, they

brainstormed in isolation.

20

After the brainstorming session was completed, participants were asked to answer
several questionnaires. Many questions focused on evaluations of performance, with key
phrases underlined to minimize potential confusion about the rating of individual or
group performance. All participants were asked, “How would you rate the number of
ideas you as an individual were able to generate? I was able to generate: (they responded
on a 9-point scale ranging from very many ideas to very few ideas)” A similar question
was also asked regarding the quality of ideas mentioned.

In addition, ftf group participants were asked, “How would you rate the number
of ideas your ggoup was able to generate? Our group was able to generate: (the same 9-
point scale ranging from very many ideas to very few ideas was used)” Again, a similar
question was asked to ftf group participants regarding the quality of ideas mentioned. th
group members were also asked, “If you had been asked to brainstorm plppp on this
problem, do you think you would have generated more or less ideas than you did M
individual in the group? I would have generated: (they responded using a 9-point scale
ranging from many more ideas to many less ideas)” Again, a similar question regarding
the quality of ideas was asked.

Participants in the alone condition were asked: “If you had been asked to
brainstorm in a group with three other people on this problem, do you think you M
individual would have generated more or less ideas? I would have generated: (they
responded using a 9-point scale ranging from many more ideas to many less ideas)” The
same question was asked to alone participants, however regarding the estimated quality

of their performance had they brainstormed in a group.

21

All participants were asked to indicate, “How many ideas do you think you a_sp_p
individual generated (open-ended)?” th group participants were also asked to answer,
“How many ideas do you think the ggoup (including yourself) generated (open-ended)?”
and “What percentage of the performance of this group was due to your efforts (with a
scale ranging from O to 100 in lO-point intervals)?”

At the end of the experiment, all of the tapes were transcribed and the number of
ideas from each brainstorming session was recorded. In the usual fashion, overlapping
ideas were counted only once when calculating the total number of ideas per session.
Results show that on average, nominal groups (M = 68.7) generated significantly more
ideas than the common microphone groups (M = 28.8). And, that performance did not
differ between individual microphone groups (M = 51.4) and common microphone
groups, or individual microphone groups and nominal groups.

Importantly, although the nominal groups were most productive (significant only
when compared to the common microphone group), participants in this condition gave
the lowest ratings on the “number of ideas produced” measure. When asked to answer
“How would you rate the number of ideas you as an individual were able to generate? I
was able to generate: (they responded on a 9-point scale ranging from very many ideas to
very few ideas)” participants in the alone condition (M = 3.7) gave ratings significantly
lower than either of the two ftf groups, who did not statistically differ from one another
(M individual microphone = 5.2; M common microphone = 4.5). A similar pattern of
findings appeared for the quality of ideas measure, however, the effect of condition was
only marginally significant (M alone 2 4.4; M individual microphone = 5.3; M common

microphone = 5.1). Interestingly, the pattern of findings for the estimated number of

22

ideas generated by individuals did not match either the actual or the rated performance
measures. Individual microphone participants reported generating the most ideas (M =
10.4), followed by alone participants (M = 7.1), with common microphone participants
reporting the worst performance (M = 4.8).

Similar to Stroebe et al. (1992), Paulus et al. found that ftf group participants
significantly underestimated their group’s performance (M individual microphone = 39.5;
M common microphone = 16.2) when these data were compared with the group’s actual
performance. However unlike Stroebe et al., participants from Paulus et al.
overestimated their own contribution to their group. Results show that when ftf group
participants were asked to estimate what percentage of the performance was due to their
own individual effort, participants took credit for 36% of the total product, which is
significantly greater than one-fourth of the group performance. Not surprisingly, male
participants took more credit for the group product relative to female participants (39.2%
vs. 32.9%).

Interestingly, no significant differences were found between conditions when
participants were asked to consider whether they would have performed better alone or in
a group. Recall, ftf group members were asked, “If you had been asked to brainstorm
plppp on this problem, do you think you would have generated more or less ideas than
you did as an individual in the group? I would have generated: (they responded using a
9-point scale ranging from many more ideas to many less ideas)” And, participants in
the alone condition were asked, “If you had been asked to brainstorm in a group with
three other people on this problem, do you think you as an individgaj would have

generated more or less ideas? I would have generated: (they responded using a 9-point

23

scale ranging from many more ideas to many less ideas)” Answers to these questions
were coded so that higher scores would reflect a preference towards working in groups.
The mean for the alone condition was 5.9, while the means for the ftf group conditions
were 4.8 and 5.2 (for the individual microphone condition and the common microphone
condition respectively). Although there was no effect for condition on this “number of
ideas I would have produced” measure, a comparison of the mean ratings among the
conditions revealed an effect for quality of ideas. Fisher’s LSD post-hoe comparisons
indicate that alone participants (M = 5.6) were more positive about generating hi gh-
quality ideas in groups compared to participants in the group/individual microphone
condition (M = 4.8). Participants in the group/common microphone condition (M = 5.0)
did not statistically differ from the other two conditions.

To determine the proportion of participants who favored group brainstorming
after having direct brainstorming experience, participants who reported scores above and
below the midpoint were compared for each of the three conditions. In the alone
condition, 66.7% of the participants felt they would have been more productive had they
worked in a group, while 23.1% reported feeling they would be more productive
brainstorming individually. In the individual microphone condition, 32.5% felt they
would have been more productive group brainstorming, while 40% favored individual
brainstorming. And in the common microphone condition, 40% believed they would be
more productive group brainstorming while 32.5% believed they would be more
productive brainstorming individually.

This pattern of results is striking because it implies that a_f_tpr_ direct ftf group

brainstorming experience, some participants reduce their general bias in favor of the

24

superiority of groups. Again, it may be likely that during their task experience, processes
such as evaluation apprehension or production-blocking become salient to ftf members;
processes most people would normally be unaware of prior to having a group
brainstorming experience. This is perhaps why participants in the alone condition still
overwhelmingly expected to generate more ideas under group conditions. This finding
contradicts the previously outlined statement from Nijstad (2000) in which he proposed
that, “most participants working in the laboratory, as well as in organizations, believe that
group brainstorming is more effective than individual brainstorming (e.g., Stroebe et al.,
1992; Paulus et al., 1993; Paulus, Larey, & Ortega, 1995). This is true even after
repeated experiences with group or individual brainstorming. (pg. 5)” Of course, there is
some evidence consistent with Nijstad’s claim, such as findings from Stroebe et al.
(1992) where participants continued to exhibit a bias in favor of brainstorming groups,
even after direct ftf brainstorming experience.

When conducting a log-linear analysis on group preference in terms of the
expected quality of ideas produced, Paulus et al. found a significant interaction between
group preference and condition. Results show that only in the alone condition was the
percentage of individuals favoring group brainstorming greater than that favoring
brainstorming alone. In the alone condition, 53.9% of the participants believed they
would generate higher quality ideas if working in a group, whereas only 18.0% believed
they would come up with higher quality ideas working alone. Fifteen percent of the
group/individual microphone participants felt they would generate higher quality ideas in
a group setting, whereas 32.5% felt they would generate higher quality ideas

brainstorming alone. In the common microphone condition, 25% of the participants felt

25

they would generate higher quality ideas in a group, while 37.5% felt they would
generate higher quality ideas alone.

Overall, Paulus et al. conclude that their results are generally consistent with
previous findings. In their study, nominal participants outperformed common
microphone participants, and the performance of individual microphone participants was
intermediate. As they predicted, they also found that individuals in ftf groups rated their
performance more favorably than isolated performers. However, contrary to their
prediction, they did not find that common microphone participants rated their
performance more favorably than individual microphone participants. In addition, the
proportion of participants favoring group brainstorming was greater than those favoring
individual brainstorming only for the nominal group. Again, this finding is interesting,
considering the claims of N ijstad concerning the negligible effect of direct experience on
the group bias, and considering the opposing finding from Stroebe et al. (1992). It is also
curious considering that in the survey study by Paulus et a1. (1993; Study 1), when asked
prior to any direct brainstorming experience, the majority of participants felt they would
be more productive if working in a ftf, and would come up with higher quality ideas if
working in a ftf group. Therefore, there appears to be some evidence that suggests that
after direct experience in ftf groups, participants become leg confident about the overall
efficacy of group brainstorming, but not in their individual performance ability.

The findings of this study indirectly refute the “gestalt” explanation proposed by
Paulus et al. If working in a group leads one to focus on the group product rather than
one’s own individual contribution, and if the size of the group product inflates the

perception of both the efficacy of the group’s overall performance and the individual’s

26

judgment of their contribution, then one should see a greater degree of the illusion in the
common microphone condition compared to the individual microphone condition. In
their study, Paulus et al. fail to show this. Admittedly, because having a common mike
does not produce a larger visible or detectable product, this is a rather weak test of the
gestalt explanation. However, if the gestalt explanation requires seeing a visible or
detectable product, than this explanation could not explain why the illusion occurs when
ideas are recorded via tape recorder.

Schemas about groups. In Paulus et al. (1999), the authors suggest another
interesting basis for the illusion of group productivity. Presumably individuals have
developed various schemas or stereotypes about the effectiveness of groups that are based
on their experience and cultural mores. Importantly, these expectations concerning group
productivity and teamwork are generally positive. The authors therefore suggest the
possibility that such pro-group schemas may influence judgment about group or team
performance.

This hypothesis is consistent with results from Study 1 of Paulus et al. (1993)
which found that most people believed they would be personally more productive in a
group setting compared to an individual setting. This explanation can also logically
explain why members of nominal groups consistently feel that they would be more
productive in a group setting. But Paulus et al. do not explicitly outline how this schema
or bias in favor of groups can precisely explain the presence of the illusion effect as
defined in this paper. Recall, the illusion has previously been conceptualized as a general
belief in the superiority of groups, as well as in terms of an “unwarranted” sense of

satisfaction in group member’s perceptions of their individual performance. The schema

27

explanation can help to explain the illusion when referring to it as a general bias in favor
of groups, however, it is somewhat unclear how schematic beliefs would lead to
inflations in satisfaction concerning one’s ppfiprLal performance? Admittedly, it may be
possible that participants with activated pro-group schemas also have expectations they
will perform well when working in groups. Therefore, their favorable expectations may
be driving their ratings of satisfaction. However if this line of reasoning were true, then it
is unclear how the results of Study 2 from Paulus et al. (1993) could be explained. In this
study, a large proportion of the ftf participants no longer expressed a pro-group bias after
having direct ftf brainstorming experience. Yet, they still reported feeling more satisfied
with their performance relative to nominal group members.

Search for Ideas in Associative Memory (SIAM). A more recent explanation
offered by Nijstad (2000) suggests that the illusion of group productivity stems more
from aspects dealing with cognitive activity (given that the act of generating ideas is a
cognitive activity, which can be described as a repeated search for ideas) rather than from
either a misappropriation or social comparison processs. Ni jstad correctly argues that
although the misappropriation and social comparison explanations are plausible, they
both fail to address several questions. For example, neither explanation can convincingly
explain the consistent finding that people feel facilitated in the presence of group
members (e.g., Stroebe et al., 1992; Paulus et al., 1993; Larey & Paulus, 1995).
Presumably, Ni jstad is referring to the commonly held belief that hearing ideas from
others stimulates creativity. According to Nijstad, it is unclear as to how memory

confusion or social comparison would result in this perception. He argues that both

 

5 It should be noted that his explanation is rather specific to brainstorming or comparable creativity tasks.
and may not explain similar misperceptions for other tasks.

28

explanations would suggest that the perception of facilitation is a consequence (rather
than a cause) of satisfaction, and that people erroneously attribute their high level of
satisfaction to this perceived facilitation. Ironically, Nijstad himself also neglects to
explicitly explain how his explanation (which will be outlined shortly) addresses this
finding as well.

He also points out that both explanations (misattribution and social comparison)
assume that the belief in the relative superiority of group brainstorming over nominal
brainstorming is a consequence of the repeated experience of satisfaction with one’s
performance after a group brainstorming session. He then raises the important and
insightful question of, “why should the experience of a group member that he or she has
done relatively well (as, for example, compared to the other group members) be
attributed to group interaction, and not to personal abilities?” In other words, although
social comparison and the act of misappropriating ideas could theoretically explain why
one would feel more satisfied with himself/herself, neither explanation explicitly outlines
how such processes would then transfer into a pro-group bias. Instead, Nijstad suggests
that it is because communication in brainstorming groups is perceived to be stimulating
and facilitating, that the impression of group brainstorming as being highly effective
results.

Nijstad also claims that the explanations offered by Stroebe et al. (1992) and
Paulus et al. (1993) do not address the particular finding that has been consistently
demonstrated in previous brainstorming studies (e.g., Diehl & Stroebe, 1991; Stroebe et
al., 1992; Paulus et al., 1995) in which group members consistently report enjoying their

task experience more than those who worked alone. How would the misattribution or

29

social comparison explanation explain why people enjoy the task more when they are in
groups? Nijstad argues, if one perceives communication in groups to be stimulating, then
this could explain why group members enjoy their experience more than individuals who
do not have the opportunity to communicate. This may be plausible, considering that
Diehl & Stroebe (1991) found their participants enjoyed the brainstorming task more
when they could overhear each other compared to when no communication between
members was possible. Nijstad concludes that, “while social comparisons and memory
confusion may play a role in the subjective experience of brainstormers, these
explanations must be supplemented by other explanations, taking into account
subjectively perceived facilitation as a consequence of communication. (pg. 122)”

Nijstad then proposes a cognitive approach to the effects of communication in
brainstorming groups. This explanation is one that stresses cognitive activity, more
specifically the process of searching for ideas in associative memory (SIAM). According
to his SIAM (Search for Ideas in Associative Memory) model, the search for ideas
involves two stages. In the first stage, problem relevant knowledge is activated in long-
term memory. This information is activated in the form of a fuzzy set of knowledge,
which is called an image. During the second stage, the knowledge contained in the image
is transformed or combined into (new) solutions for the respective problem.

An important and highly plausible assumption of this model is that this search for
ideas will not always be successful (i.e., when a new idea cannot be found, or has already
been mentioned). Such unsuccessful searches are termed “failures,” while successful
searches are labeled as “hits.” As time goes on during a brainstorming task, after more

and more ideas have been mentioned, the more likely it is that an old idea will be

30

generated again (a failure). Presumably, failures have consequences for the subjective
experience of the individual who is attempting to generate ideas. In particular, the ratio6
between the number of hits and the number of failures is likely to be important. When
the number of failures is high relative to the number of hits, then the brainstorming
individual is apt to conclude that he or she is not performing well. This should lead to
lower ratings of satisfaction. In addition, if the number of failures is high, the session is
likely to be less stimulating, and ratings of enjoyment are likely to be low as well.
Therefore, in general, more failures should lead to lower satisfaction and enjoyment.
How does this relate to group brainstorming and the illusion of group
productivity? N ijstad proposes that ftf or other collaborative (e. g., computerized
brainstomring) group members will inherently experience fewer failures than nominal
group members. First of all, people who brainstorm individually will start more searches
than people who brainstorm in groups. This will eventually lead to more failures for
nominal brainstormers. In addition, when people brainstorm in groups (especially if they
don’t have anything to say), they have the luxury of sitting and listening to the other
group members generate ideas whereas individuals are forced to fill the entire session
themselves. Since group members have the opportunity to just sit and listen to those
around them, they reduce the number of cognitive searches they perform. Because group
members are most likely to sit and listen when they themselves cannot generate a new
idea, it is the number of potential failures they experience in particular that should be
reduced. Individuals who cannot come up with new ideas, on the other hand, are likely to

continue their attempts to generate ideas because to isolated individuals, “long silences

 

6 Nijstad does not specify whether he is referring to the current, or final ratio

31

can be both boring and painful.” Because individuals will not always be able to generate
hits during their periods of silence and are more likely to generate failures, unlike
interacting ftf group members, any “down” time is likely to be attributed to personal
failures. Consequently, the process of idea generation will not always seem easy for
individuals, while it should seem much easier for group members. This in turn, will lead
group members to evaluate the process more favorably, and will lead to higher ratings in
satisfaction and enjoyment.

In his study of brainstorming groups, Nijstad (2000) found some evidence to
support his conjectures. Under the traditional brainstorming guidelines, participants
brainstormed on a given topic (“how the number of tourists coming to Utrecht can be
increased”) as individuals, or either in groups of two, four, or six people. All groups
were homogeneous with regard to gender. After the brainstorming instructions were read
to participants, they were informed that the study would consist of two different tasks and
that the experiment would take a total of two hours. Unlike the previous brainstorming
studies that implemented a set period of time for the task, participants in this study were
informed that they had to decide themselves when to end the first task. Participants were
instructed to end the session only after they felt it was a good time to stop. They were led
to believe that everyone would eventually spend two hours in the laboratory because the
duration of the second task would be dependent on the time remaining after the first task.
In actuality this was not true, however, the instructions were presented to participants in
such a manner to reduce any effects of time constraints. However, it may be possible that
in designing the study as he did, Ni jstad unintentionally created a possible confound. For

example, if nominal group participants expected that the second task would be more

32

enjoyable than brainstorming alone, they may have quit the first task earlier, or lowered
their ratings of enjoyment of the nominal task in contrast to that of the anticipated second
task.

After all of the instructions were read, and after any questions were answered,
participants were divided into their groups and were then placed in separate rooms.

Every room was equipped with a table that contained a microphone (which was
connected to a recorder located in the experimenter’s room) that was placed in the middle
of the table. In addition, every room contained a small device (with a light and a button)
that was connected to the experimenter’s room. Participants were instructed to begin
brainstorming when the light turned on, and were instructed to push the button when they
felt it was a good time to stop.

After the task was completed, participants made assessments of their level of
enjoyment (using a 9-point scale ranging from “very much” — “not at all”), satisfaction
with their own performance (using a 9-point scale ranging from “not satisfied” - “very
satisfied”), and other various measures that were intended to represent the degree of
failures that occurred during the cognitive search process. For example, participants were
asked (using a 9—point scale) to state how difficult it was to generate ideas (“very
difficult”-“not at all”), how often it had happened that they were unable to come up with

9’ ‘6

a new idea (“very often never”), and how often they had internally generated ideas
which had already been mentioned (“very often”-“never”). These last three questions

were combined to form a composite measure of the relative level of failures in one’s

history of cognitive searches. Note, that the wording of the last question may be

33

perceived as inapplicable to nominal group participants. This issue will be addressed in
more detail later in this paper.

As expected, group members enjoyed the brainstorming session significantly
more than individuals (M individual = 6.50; M dyad = 7.33; M 4 person = 7.36; M 6
person = 7.28). A post-hoc test illustrated that individuals differed significantly from
group members while the three group conditions did not differ from one another. Group
members were also significantly more satisfied with their performance relative to
individuals (M individual = 5.15; M dyad = 7.17; M 4 person = 6.47; M 6 person = 6.50).
Again, a post-hoe analysis showed that individuals differed from group members while
the three group conditions did not differ from one another.

In support of the failures explanation, results also indicated that individuals
experienced significantly more failures than group members irrespective of group size.
On the failures index (higher scores indicate fewer failures), individuals scored 10.08 on
average, while dyads scored 12.94, 4 person groups scored 14.57, and 6 person groups
scored 13.78. A post-hoe test shows that individuals differed from group members, while
the three group conditions did not differ from one another.

According to Ni jstad’s predictions, the failures scale should mediate the relation
between group size and satisfaction and enjoyment. The difference between group
members and individuals in satisfaction and enjoyment should disappear (or at least be
reduced) when the analysis controls for failures. In accordance to Baron and Kenny
(1986), a series of regression analyses were performed. To suggest mediation, first, it
must be demonstrated that type of setting (e. g., group or individual) affects satisfaction

and enjoyment. Then, it has to be shown that type of setting affects failures (the

34

mediating variable). Finally, the direct effect of type of setting on satisfaction and
enjoyment should disappear when the analysis corrects for failures.

Nijstad proceeded to effect-code type of setting and entered this variable in a
regression of enjoyment and satisfaction. He found that the effect of type of setting on
enjoyment was significant (Beta = .247), as was the effect of type of setting on
satisfaction (Beta = 0.308). Type of setting was also shown to significantly affect the
mediator (failures) (Beta = 0.371). When type of setting and failures were
simultaneously entered in a regression of enjoyment, the effect of type of setting was no
longer significant (Beta = 0.146), while failures had a significant effect (Beta = 0.267).
Similarly, when type of setting and failures were entered in a regression of satisfaction,
the effect of type of setting was shown to be only marginally significant (Beta = 0.171),
while failures had a significant effect (Beta = 0.366). Therefore, Nijstad concludes that
his analyses suggest that the failures-scale mediated the effect of type of setting (group
vs. individual) on enjoyment and satisfaction.

In a test of alternative explanations, task enjoyment was tested as a mediator of
the effect of type of setting on failures and on satisfaction. The effect of enjoyment on
failures was significant (Beta = 0.244), however, enjoyment could not account for the
effect of type of setting on failures because this effect remained significant when
enjoyment was controlled for (Beta 2 0.312). In addition, enjoyment could not explain
the effect of type of setting on satisfaction. Despite the fact that the effect of enjoyment
on satisfaction was significant (Beta = 0.200), the effect of setting remained significant

even when enjoyment was controlled for7 (Beta = 0.426). Therefore, Nijstad argues that

 

7 It is unclear as to whether these effects were less strong, because no Sobel test was reported.

35

how much the participants enjoy the task cannot explain the effects of type of setting on
failures or satisfaction. Importantly for the purposes of the current study, it is not clear

whether Nijstad’s enjoyment measure is a measure of mood per se, or a measure of task
evaluation.

To test the alternative explanation proposed by Paulus et al. (1993), Nijstad also
conducted the following analysis. Theoretically, it is possible that group interaction led
to increased feelings of satisfaction (e.g., because group members could compare their
performance and find that they performed quite well). This, in retrospect, could have
caused a positive evaluation of the brainstorming session (e. g., high levels of enjoyment
and few failures). Therefore, Nijstad tested this assumption by using feelings of
satisfaction as a mediator between type of setting and both failures and enjoyment. His
results show that when failures were regressed on type of setting and satisfaction, the
result was a significant effect of satisfaction on failures (Beta 2 0.348). However, the
effect of type of setting on failures remained significant (Beta = 0.265). Satisfaction,
however, could account for the effect of type of setting on enjoyment, because the effect
of satisfaction on enjoyment was significant (Beta = 0.442), and the effect of type of
setting on enjoyment ceased to be significant when setting and satisfaction were
simultaneously entered in a regression on enjoyment (Beta = 0.111).

Given these trends, Nijstad argues that the most likely possibility is that failures
mediate the relationship between type of setting and satisfaction, and type of setting and
enjoyment. Unfortunately though, Nijstad did not conduct any analyses that could rule

out mood as a proximal mediator of the illusion of group productivity. Although he did

36

test whether his failures-scale mediated the relation between group size and satisfaction,
he did not test whether mood mediated the relation between failure and satisfaction.

Mood as an explanation for the illusion of group productivity. Although the
various theories that have been proposed to explain the illusion of group productivity are
generally plausible, only two (the social comparison explanation and SIAM) seem to
have any convincing empirical support. In addition, all of the outlined theories can
explain why the illusion of group productivity would exist in idea generation scenarios,
however, only the social comparison explanation offered by Paulus et al. (1993) can
theoretically address why the illusion effect would occur in a group context other than
that of an idea generation paradigm. Unfortunately though, the way in which Paulus et
a1. measured the illusion was in comparisons between conditions on the number of ideas
generated, which is also specific to brainstorming tasks.

Consider the fact that a study by Hinsz and Nickell (in preparation) found that
group members were more satisfied with their performance at a card-sorting task
compared to individuals who worked at the task alone, despite no empirical difference in
performance between the two conditions. Although not identical to the illusion but along
a similar vein, in a study by Park, Levine, Harms, & Ferrara (2002), interacting groups of
three to six members were compared in deception detection accuracy to individuals
working on the same task. There was no difference in deception detection ability when
comparing group judgments to judgments of individuals working at the task alone.
However, interacting group members were significantly more confident in their decisions
compared to individuals. So given that the illusion of group productivity has in a sense

been demonstrated even in tasks that do not involve idea generation (such as the card-

37

sorting task), then it must be presumed that there are still other more parsimonious
explanations besides the misappropriation and SIAM explanation for the effect. It may
be possible that social comparison processes can explain the illusion of group
productivity; however, it is also possible or even likely that other mechanisms may still
be involved.

As Paulus et al. (1999) have alluded, one factor that may be facilitating the
illusion of group productivity is mood. It has been consistently shown across all of the
previously outlined studies that group members generally enjoy their experience and tend
to be more satisfied with their performance when working in groups. In addition, most of
the participants in these studies stated a preference for working in groups rather than
alone. So if this is the case, then it is quite likely that those working in groups are
generally in better moods than isolated individuals.

Now although Ni jstad’s (2000) study appeared to rule out the role of task
enjoyment as a sole mediator for the illusion, his measures did not capture mood per se.
Although it is likely that the more one enjoys a task the more positive he or she will feel,
but this is not necessarily the case. In addition, because Ni jstad measured task enjoyment
and not mood, his analyses could not rule out the possibility that mood was the proximal
mediator for the illusion effect. Therefore, there does not seem to be enough evidence to
conclusively rule out the role of mood as an underlying factor behind the illusion of
group productivity. In fact, results from other studies indirectly bolster the likelihood
that mood does play a role in the illusion effect. For example, in a study by Forgas
(1990), participants were induced with either a positive, negative, or neutral mood, and

subsequently worked at a task in either an individual or group context. Participants were

38

asked to rate nine person categories along various dimensions. Forgas found that
individual judgments were biased in a mood-consistent manner relative to the control
condition. Interestingly though, the effect of being in a group was to accentuate the bias
of positive moods (such that positive participants made even more positive judgments),
but to attenuate the bias of negative moods (such that negative participants made less
extreme negative judgments). Similarly, previous research on mood in groups (Hinsz,
Park, & Sjomeling, 2003) demonstrated that positive moods were sustained longer after
induction when participants subsequently entered group settings compared to individual
settings, and negative moods were diminished more quickly after participants entered
group settings compared to individual settings. These effects are consistent with the
notion that non-threatening groups serve as a positive mood induction, and lends
credence to the presumption that positive perceptions will be intensified in group settings.
Furthermore, if performing in groups is generally preferred and if positive mood
is generally induced in such contexts, then it is also likely that many of the previous
brainstorming studies (e. g., Stroebe et al., 1992; Paulus etal., 1993; and Nijstad, 2000)
inadvertently induced an increase in positive mood in their ftf group participants (or
conversely a negative mood in their nominal participants). Because in these studies
participants were likely to deduce that there were two group conditions (nominal or ftf), a
process analogous to anticipatory interaction effects (Hinsz, 1992) may have affected
mood. For example, if participants in a study were aware that they could be assigned to
either a nominal or group condition, such as in Stroebe et al. (1992) where participants
were given directions for the task and were then separated into rooms where they worked

alone or in a group, mood for participants could potentially be altered. More specifically,

39

if participants enjoy group interactions and then get assigned to a nominal condition, they
may feel disappointed. Or, if they get assigned to a ftf group condition, they may feel
excited or relieved. Therefore, across studies it is plausible that mood differed between
the individual and group conditions. In addition, although the previous explanations for
the illusion effect have demonstrated some support for their claims (specifically the social
comparison and SIAM explanations), none of the previously offered explanations can
conclusively rule out the role of mood on participant perceptions.

So, how might mood contribute to the illusion effect? According to Clore,
Schwarz, and Conway (1994), “one of the most reliable findings regarding the interplay
of affect and cognition concerns the impact of mood on evaluative judgments.”

Targets have consistently been found to be evaluated more favorably when the judge is in
a positive, rather than negative mood (Forgas, 1992; Morris, 1989; Schwarz, 1990; see
Schwarz & Clore, 1988, for reviews). For example, participants rated consumer goods
(Isen, Shalker, Clark, & Karp, 1978), other people (Forgas & Bower, 1987), selected
activities (Carson & Adams, 1980), or past life events (Clark & Teasdale, 1982) more
favorably when in a more positive mood. Therefore, it does not seem unreasonable to
assume that mood will not only affect judgments of a target, but also judgments of
personal performance as well.

Research by Forgas & Bower (1987), Clark & Isen (1982), and Schwarz & Clore
(1996) all suggest that people are more likely to make mood-consistent judgments rather
than inconsistent judgments. Why? One explanation posits that people are using mood
as information. For example, if I am in a good mood after working in a group

(presumably because being a part of a group enhances positive mood), that positive mood

40

will lead me to infer that the experience must have been a positive one. If the experience
was not positive, then it does not make intuitive sense why I would be feeling so good.

In other words, it is conceivable that individuals undergo a misattribution process in
which they make causal associations between their current mood state and their current
experience, overlooking the fact that mood and the assessment of their experience may be
unrelated.

Mood-consistent judgments can also be explained by findings that show people
are more able to recall and recognize mood-consistent details better than inconsistent
details (Forgas & Bower, 1987). So, mood may affect judgments by increasing the
likelihood that one will think about or recall information that is mood-consistent rather
than inconsistent. In the brainstorming context then, it is plausible the illusion occurs
because those working in groups are in a more positive mood, and are therefore more
likely to think about and recall positive instances in which they and/or others performed
well rather than those negative instances in which they could not generate ideas. If mood
has differential effects on what people recall, then one could predict that mood would
also affect people’s retrospective judgments about their personal performance.

Another way in which positive mood may lead to the illusion of group
productivity is through the increased likelihood of heuristic processing. Bodenhausen,
Kramer, and Stisser (1994) found that when participants were induced with a positive
mood, they were more likely to exhibit stereotypic judgments. In study 1, Bodenhausen
et al. had participants act as student members of a peer disciplinary review panel. After
reviewing a case, participants were then asked to report the likelihood of the accused

student’s guilt. In this study, half of the participants were given a case in which the

41

accused student was a member of a group stereotypically associated with the alleged
offense. Results of their study show that when given this information, participants
produced significantly more stereotypic judgments (i.e., were more likely to assume
guilt) when they were induced with a happy mood relative to a neutral mood.

Similarly, a study by Bless, Clore, Schwarz, Golisano, Rabe, and Wolk (1996)
found that participants in a positive mood were more likely to rely on scripts or heuristics
when making recall judgments. In their study, participants were presented with a tape
recorded story entitled, “Going out for dinner.” The story included information that was
typical, atypical, or irrelevant with respect to participants’ restaurant script. After
listening to the story, participants were then asked to judge whether an item had
previously been presented to them in the story or not. Results from their study show that
participants who were in a more positive mood were more likely to recall a “typical” item
as having been previously presented, even though it was not. This finding supports the
assumption that happy people are more likely to rely on general knowledge structures
than sad people.

In addition, previous research demonstrates that people do have scripts
concerning groups (e.g., Paulus et al., 1993; Hertel & Kerr, 2001). For example, Hertel
and Kerr were able to successfully prime participants with social scripts revolving around
“loyalty” to one’s group. So if people have a general belief or script that groups are more
effective or productive than individuals, as Paulus et al. (1993) presume and have shown,
then positive mood should lead to an increased reliance on those expectations when
making judgments about performance. Because participants are in a performance setting,

those scripts concerning group performance should be most accessible.

42

Because there are several theoretical reasons for why mood might affect
performance judgments or perceptions, it is important to better understand the possible
contributing role of mood to the illusion of group productivity. Psychological
mechanisms that give a central role to mood might also have greater generality because
they can also be applied to group contexts other than idea generation tasks.

To reiterate, the general line of reasoning that I am proposing is as follows.
Because most people enjoy interactive group experiences more than isolated individual
ones, they will be in a more positive mood when they are in a group setting compared to
when they are alone. Mood in turn, will affect how evaluative judgments are made, such
that perceptions will be mood congruent rather than incongruent. Therefore, when
participants are asked to assess their performance, people in ftf groups will be more
satisfied with their performance because their positive mood will influence their
judgments.

To empirically test this notion, the present study manipulated both working
condition (nominal vs. face-to—face) and mood (neutral vs. negative). By creating
negative and neutral mood conditions, the potential impact of mood on perceptions of
performance could be examined directly. For the purposes of efficiency, a positive mood
condition was not used in this study. It was expected that participants in the neutral
conditions, particularly those in the ftf group condition, would have a more positive
mood than those in the negative mood conditions. Therefore, the impact of positive
mood could still be assessed without experimentally inducing positive mood. Given that
most people enjoy group experiences, the ftf/neutral mood condition should essentially

represent a positive mood condition. One advantage of this design is that by using a

43

neutral condition for comparison rather than an induced positive mood condition, several
of the previous findings on the illusion of group productivity could be replicated, and the
relative validity of alternative explanations further assessed. The only effect that would
be expected if a positive mood condition were actually included would be an even
stronger degree of the illusion of group productivity. Although this effect could serve as
evidence to bolster my hypothesis, for the sake of efficiency, this positive mood
condition was not used in the current study.

So, the design of this study is a 2 (condition: nominal vs. face-to-face group) X 2
(mood: neutral vs. negative) between—subjects factorial. The hypotheses are as follows:
1) replicating the findings of Paulus et al. (1993), prior to any actual brainstorming
experience in the study, the majority of participants will demonstrate a pro-group
preference and report a belief that ftf brainstorming groups are more productive than
nominal brainstorming groups. Similarly, most participants will report (prior to any
direct experience) that they personally, would be more productive in ftf group settings
compared to nominal group settings. 2) Nominal groups will outperform ftf groups at the
brainstorming task; 3) despite this outcome, a main effect for mood and a main effect for
group are expected. More specifically, ftf participants are expected to feel more satisfied
with their performance, as are neutral mood participants. Admittedly, it is unclear how
the mood manipulation will combine with the work condition in determining current
mood state. Regardless, a mood explanation would predict that any mood effect, whether
induced by work condition or other situational factors, would be paralled by a
comparable illusion effect. For example, if ftf participants induced with a negative mood

feel less negative than nominal participants induced with a negative mood, ratings of

44

satisfaction should reflect this difference in mood state if the mood explanation is correct.
And finally, 4) after their brainstorming session, participants in the nominal group
conditions will report that they would have personally performed better if they had

worked in a ftf group rather than a nominal group.

45

Method

Participants

Participants were two hundred and thirty-one female undergraduate students who
were all enrolled in introductory psychology courses at Michigan State University. All
participants voluntarily signed up to take part in this study in exchange for course credit.
Design

The design of this study was a 2 (work group condition: nominal vs. ftf group) X
2 M: neutral vs. negative) between-subjects factorial. Participants were asked to
brainstorm in either nominal or ftf work groups of three people, and were then asked to
answer various questions concerning their brainstorming experience.

After the brainstorming task was completed, participants were asked to participate
in a second “unrelated” study. In this supposedly unrelated study, the mood induction

took place. Participants were either induced with a neutral mood (in which they were

 

asked to write about and reflect on a mundane or everyday event from their life), or they
were induced with a rigative mood (in which they were asked to write about and reflect
on a sad event from their life).
Procedure

Groups were randomly assigned to a condition prior to the arrival of the
participants for a given session. However, if less than three people showed up for a
session, the participants were automatically randomly assigned to one of the two nominal
group conditions. In such cases, the nominal group was completed when another session
with less than three people occurred. For example, if only two participants signed or

showed up, they were assigned to a nominal group condition. If another session occurred

46

where only two participants signed up or showed up, one of the participants would be
used to complete the previous nominal group of two, while the other participant would be
used to complete the next nominal group of two. When less than three participants
arrived, they were led to believe that other sessions were also being conducted elsewhere,
and that they were still in a nominal group of three despite the fact less than three people
were present during their session.

Prior to the arrival of the participants, all materials were unobtrusively labeled on
the bottom right corner of the back page with the participants’ subject number so that
data could be tracked to the appropriate source. The index cards that would be used
during the brainstorming task were assembled into stacks 1 1/2” high. This was done so
that all participants would have the same size stack of index cards, thus, judgments of
performance would not be differentially affected by the amount of index cards that
remained after the task was completed. When participants were tested in ftf groups, they
were given a single stack of index cards three times the height of the ones given to
nominal group participants.

All participants were first asked to meet at a predetermined location outside of the
lab room where the brainstorming task was to take place. Once they all arrived, they
were escorted into the lab room and were asked to take off their watches and to put them
either in their pockets or on a table in the lab room. They were also asked to leave all
their coats and belongings on the table in the lab room as well. Participants were then
asked to have a seat at one of the three cubicles. The cubicles were formed by putting a
removable divider (which was partitioned into three separate sections) that stood

approximately 4’ high on top of a long table. Participants were all seated next to each

47

other facing in the same direction. When the divider was in place, participants could only
see each other from a side view, and definitely could not see what was in front of the
person next to them. These cubicles therefore prevented the participants from seeing how
the others answered questions, and prevented participants from being able to monitor
each other’s performance during the brainstorming task.

Once participants were seated in their cubicle, they were given a consent form to
read over and sign. Everyone who arrived agreed to take part in the study. After the
consent forms were signed and collected, the experimenter announced to the participants
that they would be working on a series of tasks. They were told that the tasks that they
would be working on were actually from two unrelated studies. Participants were led to
believe that because each of the studies did not require a lot of time, the experimenters
combined the studies for the sake of efficiency. This deception was necessary so that
participants would not psychologically connect the mood induction to the brainstorming
task in any way.

The experimenter told the participants that the first task that they would be
working on was an idea generation task, in other words, a brainstorming task.
Participants were explicitly told that the purpose of the brainstorming task was to come
up with as many ideas as possible (see Appendix A for a verbatim copy of the
experimental script and instructions). They were specifically told that gu_antity and not
gr_rali_ty was desired in such type of tasks. The experimenter proceeded to explain that
brainstorming could be conducted in either nominal or face-to-face group conditions.
Participants were told that brainstorming in nominal groups occurred when individuals

brainstormed on a given task in isolation. They were told that individuals in nominal

48

groups do not communicate with one another, instead they simply brainstorm on the topic
alone. They were further told that these are called brainstorming “groups” because at the
end of the session, all of the ideas that are generated by each of the individuals are all
added together to form a group product. Participants were told that ideas that were
mentioned multiple times, in other words, by more than one individual, would only be
counted once when calculating the group product. The experimenter explicitly explained
that the group product for a nominal group would be the total number of non-redundant
ideas that were generated by all of the individuals who brainstormed in isolation.
Furthermore, consistent with Osbom’s (1957) traditional brainstorming instructions,
participants were told that in nominal groups, combining and improving on ideas that
have already been generated was encouraged, and again, quantity and not quality was
what was emphasized in brainstorming tasks.

After the experimenter outlined nominal group brainstorming, she/he asked the
participants if they had any questions about nominal group brainstorming or how the
product of a nominal group would be assessed. If any questions were asked, the
experimenter clarified the dynamics of nominal group brainstorming until all the
participants sufficiently understood the nature of nominal group brainstorming.

The experimenter then proceeded to explain how brainstorming was traditionally
conducted in ftf groups. Participants were told that in ftf group brainstorming, members
are allowed and encouraged to talk and interact with one another. The experimenter also
explained that in ftf groups, one member is assigned the role of the recorder for the
group. The role of the recorder is to record all of the ideas that are being generated by the

group. The experimenter stressed to the participants that the recorder is also allowed to

49

generate and contribute ideas, with the only difference being that the recorder has the
responsibility of writing down each of the ideas that are mentioned by the other members.
Participants were told that the group product for ftf groups would be all of the ideas that
are generated by the group at the end of the session. The experimenter explained that
ideas that are mentioned more than once are only counted once. So, similar to nominal
groups, the group product for a ftf group would be the total number of non-redundant
ideas that are generated by everyone in the group.

Following Osbom’s (1957) suggestions, participants were told that in ftf group
brainstorming tasks, criticism of ideas is prohibited. In addition, similar to the directions
for nominal groups, combining and improving on ideas that have already been generated
is encouraged, and again, quantity rather than quality is what is emphasized. After the
dynamics of ftf group brainstorming were explained, the experimenter asked the
participants if they had any questions concerning ftf group brainstorming or how the
product of a ftf group is assessed. Again, if any participants had questions, the
experimenter addressed their questions until everyone sufficiently understood the nature
of ftf group brainstorming. All participants, regardless of condition, were informed about
both types of brainstorming contexts (nominal and ftf) so that pre-task questions (see
Appendix B) that assessed expectations concerning nominal and ftf group brainstorming
could be answered.

After nominal and ftf group brainstorming was outlined by the experimenter and
understood by the participants, the experimenter handed each of the participants their first
questionnaire (see Appendix B for a full copy of the pre-task questionnaire). This

questionnaire was answered individually by all of the participants, while they were seated

50

in their individual cubicles. The purpose of this questionnaire was to assess general
beliefs and preferences about brainstorming groups pr_io_r to any direct brainstorming
experience. To answer the questions, participants were asked to imagine that the
president of their university wanted a council of students to generate ways to increase
student satisfaction concerning campus life. They were told to imagine that they, along
with a number of other students were asked to come up with as many ideas as possible to
increase satisfaction among students. Given this, participants were instructed to indicate
which type of group would be more likely to generate the most ideas. Participants
answered this question using a 9-point Likert-scale, with the anchors being nominal
group and face-to-face group. Using the same scale, participants were also asked to rate:
1) in which type of group they personally would generate more ideas, and 2) in which
type of group they would prefer to work in. Several other questions were included in this
measure to assess participants’ expectations about their work under various group
conditions (nominal vs. ftf).

After participants completed their first questionnaire, the experimenter collected
their answers and informed them that they would be working on a brainstorming task.
Participants in the nominal condition were told that the study was designed to examine
how people brainstorm under nominal conditions. To avoid any potential negative affect
stemming from disappointment, participants were not told that the study design also
included a ftf group condition, but instead were led to believe that the study was about
examining nominal groups. After participants learned they would be brainstorming in a

nominal group, they were asked if they had any questions about nominal groups. If

51

participants had any questions, or needed clarification, the experimenter addressed their
concerns at this time.

After all questions were answered, the experimenter proceeded to hand out a pre-
measured stack of index cards and a large envelope to each of the participants.
Participants were told that they were to write down one idea per card, and were told that
throughout the task they would be allowed to look over their own cards in order to
combine ideas or to improve on ideas that had previously been written down.

Participants were asked to express their ideas clearly, but were told not to worry about
proper grammar or writing in complete sentences. When the time for the task was up,
participants were told that they would be asked to place all of their completed index cards
into the folder that was provided to them.

After the instructions were understood, participants were given their
brainstorming topic. They were told to think about all of the practical benefits and
difficulties that would arise if everyone were to have an extra thumb on each hand. The
experimenter again stressed that quantity and not quality would emphasized, and that
they were free to combine and improve on ideas that had already been generated (Osborn,
1957). After these directions were given, the experimenter proceeded to turn on a video
camera that was situated directly in front of the participants, and then left the room for a
period of 8 minutes.

Participants assigned to the ftf group conditions were given slightly different
instructions. These participants were told that the study concerned how people
brainstorm in ftf groups. They were asked if they had any questions regarding ftf groups.

If participants had any questions, they were answered by the experimenter. After all

52

questions were addressed, the experimenter removed the barrier that separated the
participants, and placed it aside. After the temporary barrier was removed, participants
were able to see one another, although they still sat side by side. The experimenter then
handed a pre-measured stack of index cards (three times as large as one of the stacks
given to a nominal group participant) and a folder to the participant in the middle seat and
told her that she had randomly been chosen to be the group recorder. She was told that it
would be her responsibility to write down all of the ideas that were generated by her
group. Again, it was stressed that she could also contribute ideas, but that it ultimately
was her responsibility to record all of the ideas.

Similar to the instructions given to nominal group members, these participants
were told to write down one idea per card. They were also told that during the task, they
could look over any of the cards to combine ideas or to improve on ideas. They were
also told that they could communicate with one another during the task. th group
participants were also asked to express their ideas as clearly as possible, however to not
worry about proper grammar or writing in complete sentences. When time for the task
was up, participants were told they would be asked to place all of the cards into the folder
that they were given. Participants in the ftf group brainstorming conditions were given
the same brainstorming topic as those in the nominal group conditions. They were also
told to emphasize quantity and not quality, were told that criticism of ideas was
prohibited, and were also told to feel free to combine and improve on ideas that had
already been generated (Osborn, 1957). Once the topic and directions were provided to

the group, the experimenter turned on the video camera and left the room for 8 minutes.

53

After the 8 minute period, regardless of condition the experimenter came into the
lab room and told the participants to finish up the idea that they were working on and to
then stop working at the task. The experimenter then stopped the video recording, and
proceeded to tell the participants that they would be working on another short task. At
this time, participants were led to believe that they would be taking part in another study
that the experimenter agreed to run, and were then given another consent form to read
over and sign. This second consent form was used to increase the likelihood that
participants would not connect the brainstorming task to the following mood induction
task.

After the second consent form was signed and collected, regardless of condition,
participants were moved to another lab and were each escorted to an individual room
where they were isolated from the other participants. In each of the rooms, materials for
the mood induction were already in place. Participants from the same session all
received the same mood induction instructions. In other words, participants in the same
ftf or nominal group all received either the neutral or negative mood induction. After the
participants were all seated, they were asked to look over the sheet of paper in front of
them. They were all told to read over and follow all of the instructions that were
provided. Once the instructions were read and understood, participants began working at
the task.

Participants in the neutral mood condition were informed that the Department of
Psychology was developing a questionnaire to assess life events in a systematic and
reliable manner. They were told in order to construct such a questionnaire it was

necessary to first collect a large sample of positive, negative, and neutral life events from

54

which to choose the appropriate materials. Therefore, it was their task to recall, re-
experience and to write about only the mundane (or everyday) events of their previous
day (see Appendix C for both mood inductions). Participants in the neutral mood
condition were told that the events that they were to write about should not be
particularly positive or negative, but rather neutral. The instructions further gave
participants examples of appropriate topics to write about, such events as going to class,
or going to the store to buy groceries. Before writing, participants were instructed to take
a moment to think about the events that they would describe on the paper provided to
them. They were told that because no one besides the experimenter would read their
descriptions, and because their descriptions wouldn’t be matched to their identity in any
way, it would be appreciated if they answered as truthfully as possible. In actuality,
participants’ answers could be matched to their subject number because each participant
was escorted into a particular room systematically by the experimenter, in a manner such
that her identity and answers could be paired. Participants were told that they could have
as much paper as necessary, and were told that they would have ten minutes to complete
this task. A period of ten minutes was used for this mood induction for all conditions
(similar to Bodenhausen et al., 1994), to minimize the chance that the participants would
finish the task early, and to minimize inadvertently induced with feelings of boredom.
Instructions for participants in the negative mood condition were similar to those
of the neutral mood condition with the following exception. Rather than asking
participants to recall and write about an everyday life event, participants in the negative
mood condition were asked to recall, re-experience, and write about a personal life event

that made them sad. Before writing about the sad life event, they were instructed to think

55

about how they felt at the time of the event, were asked to focus on concrete and vivid
aspects of the event rather than on an abstract or objective assessment of it, and were also
instructed to provide an account of how the event happened to come about (Strack,
Schwarz, & Gschneidinger, 1985). Prior to completing the mood induction task,
participants were reinforced with notions that their answers for the task would be kept
confidential.

After the ten minute period of writing, participants were given a version of the
PANAS (Watson, Clark, & Tellegen, 1988) to assess their current mood state (See
Appendx D). Instructions asked participants to rate on a 1 to 5 Likert-scale (“very
slightly/not at all” to “extremely”), how they were feeling at the present moment along
various positive and negative affect dimensions. Participants were also asked to answer
an additional question about their general current mood state using a l to 9 point Likert-
scale (“very negative” to “very positive”).

After the PAN AS was completed, all participants were led back to the original lab
room where they first completed the brainstorming task and were seated in their original
chairs. For participants in the ftf group conditions, the removable barrier was put back
into place so that the remaining questionnaires could be completed in isolation. Once
participants were seated, they were instructed to “switch gears” and to take some time to
think about the brainstorming task that they previously completed. Participants were
asked to close their eyes, take a deep breath, and think about how they performed at the
brainstorming task. They were asked to reflect on their brainstorming experience for a

period of two minutes. During this two minute period of reflection, the experimenter left

the room. After the reflection period was over, the experimenter returned to the lab and

56

handed each participant a packet of questionnaires that were to be answered individually.
Participants were explicitly instructed not to interact or communicate with one another
while answering this series of questionnaires.

The first post—task questionnaire (see Appendix E) was designed to assess
perceptions of performance at the brainstorming task. The first question asked
participants to indicate how satisfied they were with their personal performance at the
task, using a 1 to 9 Likert-scale ranging from “not at all satisfied” to “extremely
satisfied.” The second question was an open-ended question, which asked participants to
explain why they chose the satisfaction rating that did. The next question, similar to the
measure used in Paulus et al. (1993), asked participants to rate the number of ideas that
they personally were able to generate using a l to 9 Likert-scale ranging from “very few
ideas” to “very many ideas”. And the fourth question (open-ended) asked participants to
estimate how many ideas they thought they generated during the brainstorming task.

In this questionnaire, participants were also instructed to imagine that they had
been asked to work on the same brainstorming problem under the other type of group
condition (nominal or face-to-face), and were then asked to answer whether they thought
they as individuals, would have generated more or less ideas (using a 1 to 9 Likert—scale
with 1 being “many less ideas”, 5 being “the same amount of ideas,” and 9 being “many
more ideas”). In other words, participants in the nominal group conditions were asked to
imagine that they worked on the same task, but in a face-to-face group and vice versa for
those in the ftf group conditions. They were then asked to answer whether they would
have been more or less productive under face-to-face group conditions, compared to the

nominal group condition in which they had previously performed and vice versa for those

57

in the ftf group conditions. Participants were also asked to answer “which type of group
do you think would come up with more ideas?” using a 1 to 9 Likert-scale (with 1 being
“nominal group would come up with many more ideas,” 5 being “both groups would
come up with the same amount of ideas,” and 9 being “face-to-face group would come up
with many more ideas”).

Similar to Stroebe et al. (1992), participants in the ftf group conditions were
asked: 1) “Out of the total number or non-redundant ideas from your face-to-face group,
what percentage do you think were suggested by you?” 2) . .what percentage do you
think were suggested by others (and not mentioned by you), but also occurred to you?”
and, 3) . .what percentage do you think were suggested by others, and did not occur to
you?” Participants in the nominal group conditions were asked similar questions,
however, they were asked to make assessments concerning their nominal group and were
not asked the second question listed above concerning ideas that were suggested by
others but also occurred to the participant.

Similar to Nijstad (2000), all participants in the ftf group conditions were asked to
answer the following three questions using 9 point Likert-scales: 1) “During the
brainstorming task, how difficult was it to generate ideas? (“not at all difficult” to
“extremely difficult”), 2) “During the brainstorming task, how often were you unable to
come up with a new idea? (“not very often” to “very often”), and 3) “During the
brainstorming task, how often had you internally generated an idea which had already
been mentioned? (“not very often” to “very often”).” However, unlike Nij stad (2000),
participants in the nominal group conditions were only asked the first two questions in

this series of three measures, because at the time of questionnaire construction for this

58

study, the third question did not seem applicable to the nominal group brainstorming
experience. Initially it was unclear how nominal participants would respond to that third
question because it was originally interpreted as meaning, “How often had you internally
generated ideas which had already been mentioned by others?” Because it was not until
after the data were collected that I realized the question could be in reference to
“mentioned by oneself,” the measure was excluded from the questionnaire provided to
nominal group participants. Therefore, the test of the SIAM explanation for this study
could only be done using a composite measure that was comprised of only two of
Nijstad’s three original questions. This omission in retrospect should have been avoided,
however, it does seem understandable in some regards especially considering that several
of the reviewers of the current study also interpreted the term “had already been
mentioned” as being inapplicable to nominal brainstormers.

To test the social comparison explanation provided by Paulus et al. (1993), all
participants were asked to answer how they thought they personally, relative to other
students at MSU, performed at the brainstorming task. Participants used a l to 9 Likert-
scale ranging from “not well at all” to “extremely well” to answer this question. A
follow up question was also administered, which asked participants to rate from 1 to 9,
how certain they were of their answer (from “not certain at all” to “extremely certain”).

All participants were also asked to provide ratings concerning the level of effort
that they put into the task, how motivated they were to perform well, the amount of
pressure they felt to come up with ideas, the level of difficulty they had concentrating
while trying to generate ideas, and how much enjoyment they felt during the task.

Participants were also asked to rate how creative they felt their own ideas were, and how

59

they felt about the quality of their own ideas. Participants were also asked to answer
whether they thought the quality of their ideas would have been better or worse if they
had been in the other type of working condition.

After participants completed the performance assessment questionnaire, they
completed a questionnaire designed to assess their general preferences and beliefs about
the various type of work groups (nominal and ftf) (See Appendix F). Using a 1 to 9
Likert-scale, participants were asked to rate whether they disagreed or agreed with the
claim that in brainstorming tasks, face-to—face groups generate more ideas than nominal
groups (with 1 being “I disagree completely” and 9 being “I agree completely”). They
were also asked to provide confidence ratings for their answer. Using the same scale as
mentions above (“I disagree completely” to “I agree completely”), participants were also
asked “For most types of tasks, face-to-face groups are more likely to do a better job than
nominal groups.”

In this questionnaire, participants were also asked to estimate how many ideas
they would produce in a 5 minute period, if they worked in a face-to-face group, and also
in a nominal group (open-ended). All the measures that asked participants to make
comparisons between nominal and face-to-face groups were counterbalanced to minimize
any order effects. For example, with this series of questions, half of the participants were
asked to first estimate how many ideas they would produce if working in a ftf group, and
then were asked to estimate how many ideas they would produce if working in a nominal
group. The other half of the participants were first asked how many ideas they would
produce if working in a nominal group, and then were asked to estimate how many ideas

they would produce if working in a ftf group.

60

To get a sense of the participants’ general beliefs about working groups, all
participants were asked the following using a continuous 9 point scale: 1) “Regardless of
the nature of the task, I am generally more satisfied with my personal performance when
I work. . .(“1 = alone to 9 = in a face—to-face group”), 2) “For problem-solving tasks, I am
more satisfied with my personal performance when I work. . .(same scale as above)” 3)
“In general, I am more creative when I work. . .(same scale as above)” and 4) “I enjoy
myself more when I work. . .(same scale as above)”

After completing the general beliefs questionnaire, all participants were given a
measure to assess their level of extraversion (See Appendix G). Participants were
provided with ten statements, and were instructed to rate how accurately each statement
described them. They were told to provide ratings based on how they are generally in the
present, not how they wish to be in the future. They were also instructed to describe
themselves honestly, and in relation to other people they know who are the same sex, and
roughly the same age. Participants were provided with a 5 point scale to make their
assessments, with 1 being “very inaccurate”, 3 being “neither inaccurate nor accurate,”
and 5 being “very accurate.” Participants were asked to give ratings on the following
statements: 1) “Am the life of the party,” 2) “Don’t talk a lot,” 3) “Feel comfortable
around people,” 4) “Keep in the background,” 5) “Start conversations,” 6) “Have little to
say,” 7) “Talk to a lot of different people at parties,” 8) “Don’t like to draw attention to
myself,” 9) “Don’t mind being the center of attention,” and 10) “Am quiet around
strangers.”

After the extraversion measure was completed, participants completed a

manipulation check measure (See Appendix H) which asked them (open-ended) to write

61

and explain in detail if: 1) there was anything unusual about-the first part of the study
(where they brainstormed), and 2) the second part of the study (where they wrote about a
life experience). Participants were also asked to explain what they thought was the
purpose of the study.

After the manipulation check was completed, the experimenter collected all of the
questionnaires. Once the forms were collected, the experimenter announced to all
participants that they would be receiving an email sometime in the near future.
Participants were told that the subject heading for the email would say “MSU Psychology
Study.” They were also told that the email would contain a quick survey concerning the
study. After the announcement was made, participants were debriefed and given credit
for their participation.

After a period of fourteen days (similar to the length of time between session 1
and 2 from Stroebe et al. (1992)) from the initial date of their experiment session, all
participants received an email from the experimenter. The email message contained the
following paragraph:

“A few weeks ago, you participated in a study called “Idea Generation Task.” In

this study you were asked to brainstorm about the benefits/difficulties that would

arise if everyone were to have an extra pair of thumbs. Listed below are all of the
ideas that were mentioned by your group. We are interested in people’s memory
concerning their brainstorming experience. So, next to each idea, please answer
by using the following number system:

1 = I thought of this idea and mentioned it

2 = I thought of this idea but did not mention it
3 = I did not think of this idea, nor did I mention it”

The email message listed all of the ideas that were generated from the participants’
group, in a random order. Participants were asked to provide one of the three ratings

shown above next to each one of the ideas that were listed, and were then asked to reply

62

and send the email back to the experimenter. Using a l to 9 point Likert-scale (ranging
from “not at all satisfied” to “extremely satisfied”), participants were also asked to rate
how satisfied they were with their personal performance at the brainstorming task that

they completed two weeks before.

63

RESULTS
Manipulation and Suspicion Checks

The last questionnaire that was completed by participants asked them to make
note of any unusual aspects concerning the study. The most common response noted by
participants dealt with the nature of the brainstorming topic. Some participants stated
that they could not understand why they were asked to brainstorm on an odd topic such as
how the world would change if people had additional thumbs. Because the participants
who felt the topic was odd were randomly distributed across conditions, and because the
nature of the voiced suspicion should not have affected the results of the study, data from
all participants were included in the analyses. Participants were also asked to state what
they perceived was the purpose of the study. Importantly, none of the participants
believed the purpose was to examine the role of mood on perceptions of performance.

A manipulation check was also performed to assess the effectiveness of the
experimental mood induction. To test the relative effectiveness of this manipulation, the
PAN AS was administered immediately after the time for the experimental induction
ended. To properly analyze the results, first, intraclass correlations for data from
participants in the ftf group conditions were calculated for each of the twenty—one items
on the mood assessment questionnaires. The purpose behind these calculations was to
identify the occurrence of any violations of independence (Kashy & Kenny, 2000).
Following the suggestions outlined in Kashy & Kenny (2000), a liberal test (alpha = .20)

of nonindependence was used.

 

8 Intraclass correlations were calculated for all post-brainstorming task measures to assess for violations of
independence. Group was used as the unit of analysis when violations of the independence assumption
occurred. Therefore, all subsequent measures that used individual as the unit of analysis showed no
evidence of interdependence.

1101

det

cal

Of the twenty-one dependent measures, eight showed evidence of
nonindependence. These eight items were: upset, strong, irritable, alert, ashamed,
determined, attentive, and current mood state. The effect of the mood induction was
calculated for these eight measures, using group as the unit of analysis. In this study,
there were a total of 37 three-person groups in the neutral mood conditions, and a total of
40 three-person groups in the negative mood conditions.

Results from 2 X 2 ANOVA tests show the following significant and relevant
findings. A main effect for mood (M neutral = 1.35; M negative = 2.32) in the expected
direction was found on the “upset” item, F (1, 73) = 39.37, p < 0.001. There was no main
effect for group (M nominal = 1.92; M ftf = 1.74), F (1, 73) = 1.34, p = 0.25, and the
interaction of group and mood failed to reach significance, F (1, 73) = 1.37, p = 0.25 (See
Appendix 1, table 1 for condition means). On the 9-point “current mood state” question,
negative mood participants reported feeling significantly less positive than those in the
neutral mood condition (M = 5.21 and M = 5.99, respectively), F (1, 73) = 9.76, p =
0.003. The main effect for group was not significant (M nominal = 5.43; M ftf = 5.77), F
(1, 73) = 1.91, p = 0.17, nor was the interaction effect. Participants in the negative mood
conditions (M = 2.0) also reported feeling more irritable than neutral participants (M =
1.66), F (1, 73) = 5.68, p = 0.02 and more ashamed (M negative = 1.54; M neutral =
1.15), F (l, 73) = 9.96, p = 0.002. The main effect for group (M nominal: 1.46; M ftf =
1.24) on the ashamed measure was marginally significant, F (1, 73) = 3.11, p = 0.08. A
main effect for mood was found on the “strong” item (M neutral = 2.24; M negative =
2.67), however in the opposite direction as would be expected, F (1, 73) = 5.21, p =

0.025. One plausible post-hoe explanation that could account for this finding would be if

65

participants in the negative mood condition wrote about negative life experiences in
which they asserted themselves (e.g., ended a relationship). The main effect for group (M
nominal = 2.29; M ftf = 2.63) was marginally significant, F (1, 73) = 3.53, p = 0.06.
There were no significant main effects or interactions on the items “alert,” “determined,”
or “attentive.”

For the remaining single item measures from the mood assessment questionnaire,
individual was used as the unit of analysis in the following 2 X 2 ANOVA tests. Again,
only significant and relevant findings are reported. A main effect for mood was found
such that negative participants felt significantly more distressed than those in the neutral
conditions (M = 2.34 and M = 1.73 respectively), F (1, 227) = 16.78, p = 0.001 (See
Appendix 1, table 2 for condition means). Interestingly, a main effect for group was also
found illustrating that nominal participants (M = 2.17) felt more distressed than ftf
participants (M = 1.88), F (1, 227) = 4.14, p = 0.043. Negative mood participants
reported feeling more guilty than neutral mood participants (M = 1.64 and M = 1.24
respectively), F (1, 227) = 10.43, p = 0.001; more scared (M = 1.57 and M = 1.19
respectively), F (l, 227) = 12.60, p = 0.001; more hostile (M = 1.54 and M = 1.14
respectively), F (1, 227 ) = 9.78, p = 0.002; less enthusiastic (M = 1.70 and M = 2.16
respectively), F (l, 227) = 10.83, p = 0.001; and more afraid (M = 1.47 and M_= 1.13
respectively), F (l, 227) = 10.99, p = 0.001. There was a marginal main effect for group
on the nervous measure (M nominal = 1.79; M ftf = 1.54), F (1, 227) = 3.31, p = 0.07,
such that nominal participants were reporting more nervousness than ftf participants.
And there were no significant main effects or interaction effects on the items:

“interested,” “excited,” “proud,” “inspired,” “jittery,” or “active.”

66

Two composite measures were also created to assess the relative effectiveness of
the mood manipulation. One composite measure consisted of the ten positive affect
items, while the other consisted of the remaining ten negative affect items. Using the
positive affect composite measure as the dependent variable, no differences between the
two mood conditions were found (M neutral = 23.54; M negative = 23.05), F (1, 227) =
0.22, p > 0.05, nor was any group effect found (M nominal = 23.59; M ftf = 23.02), F (l,
227) = 0.271, p > 0.05. Because there was evidence of nonindependence for the negative
affect composite measure, group was used as the unit of analysis in the 2 X 2 ANOVA.
With group as the unit of analysis, a significant difference was found in the expected
direction, such that those in the negative mood conditions felt significantly more negative
than those in the neutral mood conditions (M = 18.18 and M = 13.97 respectively), F (1,
73) = 28.02, p < 0.01. Interestingly, a significant main effect was also found for the
group variable (M nominal = 16.85; Mftf= 15.28), F (l, 73) = 3.85, p = 0.05,
demonstrating that nominal participants felt more negativity than ftf participants. There
was no interaction effect between group and mood on the negative composite measure.
Because differences were found on many of the negative but not positive measures
(which makes sense because the mood induction manipulated negative mood and not
positive mood), there is evidence to support previous claims (e. g., Costa & McCrae,
1980; Diener & Emmons, 1985; Watson & Tellegen, 1985) that negativity and positivity
are orthogonal dimensions rather than opposite ends of the same continuum. Given this,
then the results imply that mood differences between the group conditions under the
contexts of this study are essentially occurring along the negativity dimension rather than

the positivity dimension. In other words, nominal participants feel more negative than ftf

67

participants, however, ftf participants feel no more positive than nominal participants.
Despite the group main effect on the negative composite measure (and the distress
measure), it should be noted that the mood manipulation effects were stronger and more
frequent than the work condition effects. Hence, at this point, it appears that current
mood state was more strongly determined by the recent mood manipulation than the prior
work condition, although some traces of influence from the latter still existed.

To further specify what the participants in the negative mood conditions were
feeling after the mood induction, t-tests were conducted to see what the participants were
feeling most strongly. Results show that means for the items “upset” and “distress” were
significantly higher than any of the means for the other negative affect ratings. The mean
for “distress” (M = 2.33) did not significantly differ from the mean for “upset” (M =
2.32), t (119) = 0.05, p > 0.05. Given all of these results, it appears that the mood
manipulation successfully induced those in the negative mood conditions with negative
affect, in particular with feelings of distress and succeeded to make participants feel
upset. Conveniently, these are the kinds of moods that would presumably facilitate a
misattribution process concerning performance assessments, and therefore offer a valid
test of the mood explanation.

General beliefs and expectations concerning groups

Previous studies (e.g., Stroebe et al., 1992; Paulus et al., 1993) have found that
prior to any immediate brainstorming experience, most people generally believe that ftf
brainstorming groups are more productive than nominal groups, and that they personally
would be more productive at a brainstorming task if they were in a ftf group compared to

a nominal group. For example, Stroebe et al. (1992) found that 80% of their participants

68

thought that a person working in a ftf group would be more productive than an individual
working in a nominal group. Paulus et a1. (1993) showed that individuals expected to
generate more ideas (M = 14.32) in a ftf group setting, compared to in a nominal group
setting (M = 11.56), p < 0.05 (one-tailed).

Results from this study also illustrate a biased belief in favor of ftf brainstorming
groups. Prior to the brainstorming task, when participants were asked to answer the
following question using a 9-point scale, “Which type of group is more likely to generate
the most ideas (1 = nominal group, 5 = same, and 9 = ftf group)?” the average response
was 6.79 (N = 229). This mean response was significantly higher than the scale
midpoint, t (228) = 9.67, p < 0.01, which suggests that on average participants believed
that ftf groups would generate more ideas. Results from a 2 X 2 ANOVA show no main
effect for group or mood, which demonstrates that as expected, there were no condition
differences on this pre-task measure.

Using the same 9-point scale, participants were also asked to answer, “In which
type of group would you personally generate more ideas?” The mean response was 5.50,
which was also significantly higher than the scale midpoint, t (229) = 2.32, p = 0.021.
Again, 2 X 2 ANOVA results showed there was no significant interaction or main effects.
Along similar lines (using a 9-point scale ranging from “not many ideas at all” to “very
many ideas”), participants were asked to answer the following questions: 1) if you were
given 10 minutes to complete the brainstorming task, how many ideas would you,
personally, come up with under nominal group conditions, and 2) if you were given 10
minutes to complete the brainstorming task, how many ideas would you, personally,

come up with under face-to-face group conditions. There were no interactions or main

69

effects for either of these measures. For the first measure, the grand mean response was
5.78 (SD = 2.37). For the second measure, the grand mean response was 5.73 (SD =
1.35). Results from a paired sample t-test indicate that these two means did not differ, t
(227) = -0.300, p = 0.74. Given that there has been a substantial amount of evidence
from this study as well as previous studies (e.g., Stroebe et al., 1992; Paulus etal., 1993)
suggesting that a pro-group biases exists, particularly before direct group brainstorming
experience, this finding was unexpected and suggests that perhaps the framing of the
question affects the nature of the responses. For whatever reason, it seems that the
illusion effect becomes more consistent and pronounced when the framing of the question
is broad (e. g., ratings of satisfaction) rather than specific (e. g., how many ideas do you
think you would come up with).

For the item, “In which type of group would you prefer to work in?” the mean
response was 6.44, which was also significantly higher than the scale midpoint, t (229) =
7.38, p < 0.001. No main effects or interactions were found. Therefore, analyses from
these pre-task measures generally show: 1) there were no mood condition effects prior to
working on the brainstorming task, 2) prior to direct brainstorming experience,
participants expected that ftf groups would be more productive than nominal groups, 3)
participants expected they personally, would be more productive in ftf groups instead of
nominal groups, and 4) participants had a preference for working in ftf groups over
nominal groups.

Performance and perception results
Consistent with findings from previous studies (e.g., Diehl & Stroebe, 1987;

Stroebe, et al., 1992; Paulus et al., 1993), nominal groups outperformed ftf groups at

70

approximately a 1:1.5 ratio (see Table 3, Appendix I for results by condition). The mean
number of nonredundant ideas generated by nominal groups (n = 39) was 38.40 (SE =
1.34), while the mean number of ideas generated by ftf groups (n = 38) was 23.05 (SE =
1.36). Redundant ideas, which were only found among nominal groups, were only
counted once towards the overall group product. Ideas were labeled as redundant only if
they suggested identical concepts, and were not labeled redundant if they were variations
of one another. For example, if two group members submitted the idea “will be harder to
play instruments,” their nominal group would be credited with one idea. However, if one
member suggested “will be harder to play instruments,” and another member suggested
“will be harder to play the trumpet,” the group was credited with two ideas. 2 X 2
ANOVA results of the number of ideas at the group level show this main effect was
significant, F (1, 73) = 65.09, p = 0.001. Because performance was completed prior to
the mood manipulation, no main effect for mood or interaction was expected, and neither
was found.

Interestingly, when participants were instructed to assess their perceptions of their
performance at the brainstorming task, results differed depending on how the question
was asked. Similar to Stroebe et al. (1992) and Nijstad (2000), when participants were
asked to rate how satisfied they were with their personal performance at the task,
participants from ftf groups were significantly more satisfied than participants from
nominal groups (M = 6.14, SE = 0.166 and M = 5.40, SE = 0.164 respectively), F (1, 227)
= 10.08, p = 0.002. This result from the 2 X 2 ANOVA confirms the interesting pattern
where participants in ftf groups report feeling more satisfied, despite having performed

worse relative to those who brainstormed individually (see Appendix I, Table 4 for

71

condition means). The main effect for mood and the interaction effect were not
significant, F (1, 277) = 1.16, p = 0.28, and F (1, 227) = 0.03, p = 0.85, respectively.

To directly test whether mood mediated the relationship between satisfaction and
group, a series of regression analyses were conducted in accordance with the guidelines
suggested by Baron and Kenny (1986). When using the single item, “upset,” from the
PAN AS, the test of mediation could not be performed because although the relationship
between group and satisfaction were significant (Beta = 0.205) the relation between the
“upset” measure and group condition, and “upset” and perceptions of satisfaction were
not significant (Beta = -0.096 and 0.075 respectively). Similarly, when using the PANAS
item “distressed,” the test of mediation could not be performed because “distressed” was
not significantly correlated with group (Beta = -0.123, p = 0.06), or satisfaction (Beta = -
0.067, p = 0.30). When using the ten-item negative mood composite measure from the
PANAS, group and negative mood scores were not significantly related (Beta = -0.184, p
= 0.10), nor was the relationship between negative composite scores and satisfaction
(Beta = 0.706, p = 0.51).

Because mood was manipulated in this study, the mood condition (neutral vs.
negative) was also simultaneously entered into the following regression analyses with
group, to parcel out any influence the mood condition could have had on the associations
between group and the mediator and dependent variable. When conducting the
regression analyses in this manner, the following results were found. When distress was
used as the mediator, and was regressed on group condition and mood condition that
were entered simultaneously, a significant association was found between group and

distress (Beta = -0.127, p = 0.047) such that participants in the nominal groups reported

higher levels of distress. When satisfaction was regressed on group and the mood
condition, the relationship between group and satisfaction was significant (Beta = 0.206,
p = 0.002). However, when satisfaction was regressed on group, mood condition, and
distress, the relationship between distress and satisfaction was not significant (Beta = -
0.052, p = 0.44), so the prerequisite conditions for the test of mediation were not present.
Similar results were found when the PANAS negative composite measure was used as a
potential mediator between the relationship of group and satisfaction. When the negative
composite measure was regressed on group and the mood condition, the relationship was
significant (Beta = -0.191, p = 0.05), again demonstrating that nominal participants felt
more negative than ftf participants. The relationship between group and satisfaction was
significant, however, when satisfaction was regressed on group, mood condition, and the
negative composite scores (all entered simultaneously), the relationship between negative
composite scores and satisfaction was not significant (Beta = 0.112, p = 0.41).

To test the mood hypothesis with participants that most resembled those from
previous studies where mood was not manipulated, tests of mediation were performed
selecting only those participants in the neutral mood conditions. For the following
analyses, distress was tested as a potential mediator between the relationship of group and
satisfaction. Although the correlations between group condition and satisfaction (Beta =
0.187, p = 0.049) and group condition and distress (Beta = -0.215, p = 0.023) were
significant, the association between satisfaction and distress was not significant (Beta = -
0.141, p = 0.14) when satisfaction was regressed on group and distress (which were
entered simultaneously). When group and distress were entered simultaneously, the

relationship between group and satisfaction was also no longer significant (Beta = 0.157,

73

p = 0.104). Even though the relation between satisfaction and distress was not significant
(p = 0.14), a Sobel test was conducted to examine whether the reduction in the strength of
the association between group and satisfaction was significant. Results indicate the
reduction was not significant (Sobel 2 test statistic = 1.24, p = 0.21).

Again selecting only participants in the neutral mood conditions, the PAN AS
negative composite measure was tested as a potential mediator between group and
satisfaction. When the negative composite measure was regressed on group, the
association was only marginally significant (Beta = -0.307, p = 0.06) (in the expected
direction). When satisfaction was regressed on group, the relationship was significant
(Beta = 0.187, p = 0.049). When satisfaction was regressed on the negative composite
measure and group (which were entered simultaneously), the relationship between
satisfaction and negativity was significant (Beta = 0.38, p = 0.026), as was the
relationship between group and satisfaction (Beta = 0.369, p = 0.029). So, in this set of
analyses using only neutral participants, distress was related to satisfaction however in
the opposite direction as predicted. In other words, results show that the more distressed
participants were, the more satisfied they reported feeling. Interestingly, when the effect
of negativity was controlled for, the relationship between group and satisfaction actually
seemed to increase, but not significantly (Sobel 2 test statistic = -1.48, p = 0.14).

To examine the multiple hypotheses concerning satisfaction ratings, the two
independent variables were combined to form one four-level variable so that an ANOVA
and subsequent contrasts could be conducted. First, the One-Way ANOVA was
significant, F (3, 227) = 3.75, p = 0.012. Similar to the 2 X 2 ANOVA results, a main

effect for work condition was found, t (227) = -3.175, p = 0.002, however no main effect

74

for mood and no group x mood interaction was found; t (227) = 1.075, p = 0.28 and t
(227) = 0.185, p = 0.853 respectively. In the closest approximation of and replicating the
findings from Stroebe et al. (1992) and Nijstad (2000), a planned contrast shows
ftf/neutral participants (M = 6.24) were more satisfied with their performance compared
to nominal/neutral participants (M = 5.54), t (227) = — 2.073, p = 0.039. Interestingly,
ftf/negative participants (M = 6.03) were not significantly more satisfied than
nominal/neutral participants, t (227) = -1.50, p = 0.14. This at best, indirectly suggests
that simply being in a ftf group will not necessarily lead to the illusion effect, and that
other factors, such as the presence of negative moods could minimize the magnitude of
the illusion.

However this conclusion may be erroneous because when post-hoe comparisons
on negativity were conducted between ftf/negative and nominal/neutral participants,
results show that ftf/negative participants actually felt more upset (M = 2.32) than
nominal/neutral participants (M = 1.53), F (1, 37) = 11.80, p = 0.01, and scored higher on
the negative composite index (M ftf/negative = 17.48, M nominal/neutral = 14.82), F (l,
37) = 4.74, p = 0.036. 80, despite feeling more negative than nominal/neutral
participants, ratings of satisfaction did not differ between ftf/negative and
nominal/neutral participants. If the absence of negativity was a determining factor in
producing the illusion effect, then nominal/neutral participants should have been more
satisfied than ftf/negative participants. Furthermore, post—hoe analyses also show that
ftf/negative and nominal/negative participants did not differ in any of the measures of
negativity, yet ftf/negative participants (M = 6.03) were more satisfied than

nominal/negative (M = 5.25) participants, t (227) = - 2.42, p = 0.016.

75

Basis of satisfaction ratings. After participants were asked to make satisfaction
ratings, they were asked (open-ended question) to write out why they chose the response
that they did. To analyze the results for this question, answers were number coded in
accordance with the following classification system. Responses were designated to one
of the following categories: 1) “came up with many ideas,” 2) “came up with
good/creative ideas,” 3) mentioned both “many and good/creative ideas,” 4) “couldn’t
come up with many ideas,” 5) “couldn’t come up with good/creative ideas,” 6) mentioned
both “couldn’t come up with many ideas and with good/creative ideas,” 7) “other,” and 8)
“was the recorder for the group.”

For nominal participants, the most frequent response (24.8%) was “came up with
many ideas,” followed by (22.2%) “couldn’t come up with many ideas (presumably from
participants who were not particularly satisfied)” Therefore, 47% of the total responses
were in reference strictly to quantity. A similar pattern was seen among ftf participants;
the most frequent response (31.6%) was “came up with many ideas,” followed by
(23.7%) “other” and then (13.2%) “couldn’t come up with many ideas.” Therefore,
44.8% of the responses from ftf participants referred strictly to quantity. Conversely,
26.5% of the responses from nominal participants made reference to strictly to quality,
while 13.2% of responses from ftf participants made reference strictly to quality.

Post-task responses from participants imply that ratings of satisfaction were not
driven primarily from ratings of quality. Interestingly however, post-hoe analyses offer
evidence to the contrary. After the brainstorming task and mood induction procedure,
participants were asked to assess, “How creative do you think your ideas were from the

brainstorming task (using a 1 to 9 Likert—scale ranging from “not at all creative” to

76

“extremely creative”)?” They were also asked to rate, “How do you feel about the
quality of your ideas from the brainstorming task (using a 1 to 9 Likert-scale ranging
from “very poor quality” to “extremely high quality”)?9” A 2 X 2 ANONA was
conducted on ratings of creativity. There was no main effect for mood (M control =
6.46; M negative = 6.02), however, a main effect for group was found, F (l, 137) = 8.75,
p = 0.004. Participants from ftf groups (M = 6.8, SE = 0.35) rated their ideas as being
more creative than participants from nominal groups (M = 5.68, SE = 0.152). There was
no interaction between group and mood.

Similarly, a 2 X 2 AN OVA was conducted on ratings quality. Again, there was
no main effect for mood (M control = 5.54; M negative = 5.49), however, a main effect
for group was found, F (1, 137) = 17.11, p = 0.001. Participants from ftf groups (M =
6.22, SE = 0.31) rated their ideas as being higher in quality than participants from
nominal groups (M = 4.81, SE = 0.14). Again, no interaction between mood and group
was found.

Do determine whether perceptions of creativity and quality influenced ratings of
satisfaction (despite the experimental instructions which informed participants that the
goal of the task was to come up with as many ideas as possible), creativity and quality
were tested separately as potential mediators between the relationship between group and
ratings of satisfaction. First, when creativity was regressed on group, a significant
relationship was found (Beta = 0.22, p = 0.01). Group influenced creativity such that ftf
participants felt more creative. Next, when satisfaction was regressed on group, the

relationship was significant (Beta = 0.205, p = 0.002). Again, indicating that ftf

 

9 Due to a clerical error, data from only 24 ftf participants was collected on these two questions. Data from
117 nominal participants was collected

77

participants felt more satisfied with their performance. When satisfaction was regressed
on group and ratings of creativity (which were entered simultaneously), the relationship
between satisfaction and creativity was significant (Beta = 0.346, p = 0.001), and the
relationship between satisfaction and group ceased to be significant (Beta = 0.118, p =
0.14). Results from the Sobel z-test show that the influence of group on satisfaction was
significantly reduced when controlling for ratings of creativity, Sobel 2 statistic = 2.27, p
= 0.02.

Because it is theoretically possible that feelings of satisfaction influenced ratings
of creativity, satisfaction was entered as a potential mediator between group and
creativity. To test this alternative explanation, satisfaction was first regressed on group.
This relationship was significant (Beta = 0.205). When creativity was regressed on
group, the relationship was significant (Beta = 0.22). And, when creativity was regressed
on group and satisfaction, the relationship between creativity and satisfaction was
significant (Beta = 0.34), and the relationship between group and creativity remained
significant (Beta = 0.15, p = 0.05). Therefore, it seems more likely that ratings of
creativity partially mediate the relationship between group and satisfaction, rather than
satisfaction mediating the relationship between group and creativity.

Similarly, a test of mediation was conducted using ratings of quality as a potential
mediator between group and satisfaction. When quality was regressed on group, a
significant relationship was found (Beta = 0.33, p = 0.001) such that ftf group members
felt their ideas were higher in quality. When satisfaction was regressed on group, the
relationship was significant (Beta = 0.205, p = 0.002). When satisfaction was regressed

on group and quality, the relationship between quality and satisfaction was significant

78

(Beta 2 0.398, p = 0.001), however, the relationship between group and satisfaction was
no longer significant (Beta = 0.065, p = 0.43). Results from a Sobel z-test show that the
influence of group on satisfaction was significantly reduced when controlling for ratings
of quality, Sobel z statistic = 3.12, p = 0.002.

Again, because it is theoretically possible that feelings of satisfaction could
influence assessments of quality, satisfaction was tested as a potential mediator between
group and quality. When satisfaction was regressed on group, the relationship was
significant (Beta = 0.205, p < 0.05). When quality was regressed on group, the
relationship was significant (Beta = 0.326, p < 0.05). And, when quality was regressed
on satisfaction and group, the relationship between satisfaction and quality was
significant (Beta = 0.370, p < 0.05), and the relationship between group and quality
remained significant (Beta = 0.254, p = 0.001). Therefore, it seems more likely that
assessments of quality influence ratings of satisfaction, rather than feelings of satisfaction
influencing ratings of quality.

These tests of mediation suggest that assessments of creativity and quality each
partially mediate the relationship between group and satisfaction. These patterns of
results are in line with findings from previous literature (e.g., Stroebe et al., 1992) that
show that participants expect to generate better quality ideas in ftf group settings. A test
of mediation using ratings on the number of ideas generated was not conducted because
there were no differences between groups on this measure. These results all imply that
the illusion of group productivity may be driven by participants’ expectations that they
will be more creative and come up with better quality ideas when working in a ftf group.

Despite the experimental instructions that emphasize quantity, participants seem to be

79

 

 

 

basing
may ii
of per
than 5
numh
many
with ll
chose
consic
ratings

diretti

perfor
the Sn
menr
gTOUp
WhCT
from
and j
mood

COndr

Palllcj

basing their ratings of satisfaction on their assessments of quality and creativity. This
may in part speak to why the illusion effect is most likely to be found when perceptions
of performance are assessed using broad measures such as ratings of satisfaction, rather
than specific measures which are designed to assess quantity (such as ratings of the
number of ideas or open-ended questions which prompt participants to estimate how
many ideas they themselves had generated). Although these conclusions are inconsistent
with the analyses from the open-ended responses from participants concerning why they
chose the satisfaction rating that they did, the inconsistencies could be explained by
considering that participants may not be able to accurately report why they chose the
ratings the did, and by considering that participants may have simply been reiterating the
directions in which they were asked to come up with as many ideas as possible.

An alternative measure to ratings of satisfaction. Unexpectedly, when measuring
performance perceptions similar to Paulus et a1. (1993), a different pattern of results from
the satisfaction ratings was found. When Paulus et al. (1993) asked participants to “rate
the number of ideas you as an individual were able to generate,” participants in the ftf
group condition gave higher ratings than participants in the nominal group condition.
When participants in this study were asked the same question, no differences were found
from the 2 X 2 ANOVA between the two group conditions (M nominal = 5.59, SE = 0.16
and M ftf = 5.58, SE = 0.16), F ( 1, 227) = 0.002, p = 0.962. Similarly, no main effect for
mood or group x mood interaction was found, p > 0.05 (see Appendix I, Table 5 for
condition means).

The following question was included in the e mail message that was sent to

participants two weeks after their brainstorming session: “How satisfied were you with

80

your personal performance during the brainstorming exercise?” Participants answered
this question using a 1 to 9 Likert-scale that ranged from “not at all satisfied” to
“extremely satisfied.” If mood is the primary factor behind the illusion of group
productivity, then there should be no differences between conditions on this measure
(given the assumption that mood would be similar across conditions two weeks after the
experimental session)”. If there are no differences between conditions, it is somewhat
unclear how alternative explanations, in particular the misappropriation explanation,
could account for any lack of difference on ratings of satisfaction. Admittedly, if no
differences in satisfaction are found between conditions this would only indirectly
support the mood explanation. It is possible that participants will not have perfect
recollection of social comparison information, therefore analyses on delayed ratings of
performance perception may not speak to the validity of the social comparison
hypothesis. However, if the misappropriation explanation is correct and the illusion is
due to memory confusion, then one would expect the illusion to be even stronger after an
extended period of time. Since participants are given all of the ideas from their group in
the body of the e mail message, a misappropriation explanation would predict that
participants would inadvertently take credit for ideas that were not their own, and hence,
ratings of satisfaction should be elevated. But because it is speculative to make
conclusions based on null results and because possible confounds concerning who replied

to the message may exist, the findings from this analysis should be interpreted cautiously.

 

’0 It is possible that a memory of mood could exist which would produce an effect, and it is possible that
participants could simply remember and reiterate the satisfaction rating that they provided earlier.
However, because the mood induction was not particularly extreme, and because the e mail was sent two
weeks after the session, it seems rather unlikely that either situation would occur.

81

One hundred and eighteen participants responded to the e mail message (51.08% of the
original sample). More specifically, 54.39% (n = 31) of the nominal/neutral participants
replied, 50% (n = 30) of the nominal/negative participants replied, 44.44% (n = 24) of the
ftf/neutral participants replied, and 55% (n = 33) of the ftf/negative participants replied.
Results from the 2 X 2 AN OVA show there were no main effects for group or mood (p =
0.18 and 0.60 respectively) and no interaction effect (see Appendix I, Table 6 for
condition means).
Test of previous explanations

Social comparison. Recall, according to Paulus et a1. (1993), the illusion of group
productivity is in part driven by social comparison processes. They argue that because ftf
group members are likely to find that they are similar to or better than others in ability,
they will perceive their own performance more favorably than is actually warranted.
Note that in Paulus et al. (1993), there is no mention of uncertainty leading to insecurity,
this was later added in a summary of the social comparison explanation from Nijstad
(2000). Because Ni j stad has collaborated closely with Paulus, it may be possible that
what Nijstad is claiming is what Paulus et al. actually intended. If this is indeed the case,
then the social comparison explanation can be tested by comparing certainty ratings in
ability across conditions.

In the current study, participants were asked to answer the following question
using a 1 to 9 point Likert-scale (1=”not well at all” and 9 = “extremely well”): “How
well do you think you personally, relative to other students at MSU, performed at the

brainstorming task?” After providing an answer, participants were then asked a follow

82

up question which stated (using a 1 to 9 point Likert-scale from “not certain at all” to
“extremely certain”) “How certain are you of your answer (to the previous question)?”
2 X 2 ANOVA results show that participants in the ftf group conditions (M = 5.63, SE =
0.12) did not differ in their assessment of relative performance compared to participants
in the nominal group conditions (M = 5.39, SE = 0.12), F (1, 227) = 1.99, p = 0.16 (See
Appendix I, Table 7 for condition means). This seems to contradict the explanation
offered by Paulus et al. that would suggest that ftf group participants would feel more
able than nominal group members. Participants in the two group conditions also did not
differ in their certainty ratings (M ftf = 5.66, SE = 0.19; M nominal = 5.55, SE = 0.18), F
(1,227) = 0.17, p = 0.68, which also contradicts the social comparison explanation as
interpreted by Nijstad (See Appendix I, Table 8 for condition means). No significant
main effects or interactions were found for either 2 X 2 ANOVA.

SIAM. N i jstad (2000) proposed that the illusion was due to a relative increase in
failures during the search for ideas in associative memory among nominal participants
compared to participants working in ftf groups. Nij stad found support for his conjecture
by comparing nominal and ftf participants on his “failures” index. Recall, his failures
index included the following three questions which were answered using a 9-point scale:

9, 6‘

“rate how difficult it was to generate ideas (very difficult-not at all), rate how often it
had happened that you were unable to come up with a new idea (very often-never)” and
“rate how often you had internally generated ideas which had already been mentioned
(very often-never)” Although this three item composite measure seems valid as a

failures index for the participants in the ftf condition, it was unclear at the time of

questionnaire construction for this study how the same three items could be applicable to

83

participants in the nominal condition. In particular, it was unclear how nominal
participants would be able to answer the last question that asked them to rate how often
they generated ideas that had already been mentioned. I originally presumed that since
participants in the nominal condition worked individually, they would most likely
respond “never” to this question. If this was the case, then it could be possible that the
difference between group conditions in Nijstad’s failure index was due in part to this one
question that was not applicable to nominal participants and that his three item index was
functionally just a restatement of the work condition treatment.

So in the current study, the same three questions were asked to ftf participants,
however, for nominal participants, only the first two questions were asked. The third
question was not included in the questionnaires for nominal participants because it was
presumed to not be applicable, and presumed that participants would not know how to
respond to such a question. Therefore, the failures index for the current study consisted
of only the first two items that were asked by Nijstad. Using this failures index, the
results of Nijstad were not replicated. 2 X 2 ANOVA results show no main effect for
group, F (1, 227) = 0.86, p = 0.36 (See Appendix I, Table 9 for condition means). Tests
of mediation were not conducted because the relationship between scores on the failures
index and group were not significant (Beta = -0.063, p = 0.34). Similarly, when selecting
only participants in the neutral mood conditions, tests of mediation were not conducted
because scores on the failures index and group were not related (Beta = -0.02, p = 0.83).

Ni j stad could argue that the failure to replicate his effect was due to the fact that
the same three items were not included in the current analysis. However, if no difference

was found using the two-item composite measure (M nominal = 10.75 vs. M ftf_ = 10.36),

84

then presumably what was driving his effect were the responses from the omitted
question. If this is the case, then it is odd that Nijstad never used or justified that single
item as his measure of failures, and therefore is damaging for his theoretical explanation
of the illusion effect.

Misappriation of ideas. Similar to the procedure used in Stroebe et al. (1992),
participants in this study were presented with a list of all of the ideas that were generated
by their group. Unlike the participants in Stroebe et a1. (1992) who returned to the lab in
person, participants in this study were given this list via e mail, two weeks after their
brainstorming session. Participants were instructed to look over the list of ideas, and then
next to each idea, respond by entering a number which corresponded with one of the
following options: “I thought of this idea and mentioned it,” “I thought of this idea but
did not mention it,” or “I did not think of this idea, nor did I mention it.” Recall, a total
of 118 participants responded to the email message.

In the Stroebe et al. (1992) study, participants in the ftf groups assigned a
significantly lower percentage of ideas to the categories “suggested by me,” and
“suggested by another group member and had not occurred to me.” Only the first result
was replicated in the current study. The percentage data from the current study were
calculated by dividing the number of items in each category by the total number of ideas.
Using this procedure ensured that the category percentages for each participant summed
to 100%. Participants in the ftf group conditions (n = 57) reported that 36.62% of the
ideas were “mentioned by me,” while participants in the nominal group conditions (n =
60) reported that 43.27% of the ideas were “mentioned by me” (See Appendix I, Table 10

for condition means). 2 X 2 ANOVA results show a main effect for group, illustrating

85

that the previous finding was significant, F (l, 114) = 6.84, p = 0.01. Participants in the
ftf group conditions also reported that 36.9% of the ideas were “suggested by another
group member and had not occurred to me,” while participants in the nominal group
conditions reported that 44.66% of the ideas were “suggested by another group member
and had not occurred to me.” Again, 2 X 2 ANOVA results show this main effect for
group was significant, F (1, 114) = 6.54, p = 0.012 (See Appendix I, Table 11). In the
Stroebe et al. (1992) study, participants in the ftf group condition reported a significantly
higher percentage to the category “suggested by another group member but had also
occurred to me,” compared to those in the nominal condition. This finding was also
replicated in the current study (See Appendix I, Table 12). Participants in the ftf group
conditions reported that 25.53% of the ideas were “mentioned by another group member
but had also occurred to me,” while nominal participants reported that 12.03% of the
ideas were “mentioned by another group member but had also occurred to me.” 2 X 2
ANOVA results show the main effect for group was significant, F (1, 114) = 30.96, p =
0.01.

Stroebe et al. (1992) argue that they have support for their misappropriation
explanation, because if one compares the two group conditions in terms of the proportion
of ideas that participants claimed to have had during their brainstorming session,
regardless of whether they mentioned the idea or not, ftf group participants report that
61% of the ideas at least occurred to them, while only 47% of the total ideas at least
occurred to nominal participants. This effect was replicated in the current study. After
combining the percentages from the first two categories that were outlined in the e mail

(“. . .mentioned by me” and “mentioned by another but occurred to me”), participants in

86

the ftf group conditions reported that 62.15% of the total ideas at least occurred to them,
while only 55.30% of the total ideas at least occurred to nominal group participants, t
(116) = 2.20, p = 0.03. Therefore, if one agrees with the validity of the approach used by
Stroebe et al. (1992), then the results of this study provide evidence in support of the
misappropriation explanation.

However, other possible tests of the misappropriation explanation seem to be less
supportive. After the mood induction task (which occurred after the brainstorming task),
participants were asked to estimate how many uses they personally generated (open-
ended). Results from a 2 X 2 ANOVA show a main effect for group in direct
contradiction to the misappropriation explanation, F (1, 221) = 28.34, p = 0.001, such
that nominal group participants (M = 12.14) estimated a higher number of ideas than ftf
participants (M = 7.82) (See Appendix I, Table 13 for condition means)". The main
effect for mood was marginally significant, F (l, 221) = 3.28, p = 0.07, such that negative
participants (M = 10.71) estimated a higher number of ideas than control mood
participants (M = 9.24). The interaction between group and mood was not significant, p
= 0.69. Although these results suggest there is no evidence from direct tests of the
misappropriation explanation, it is true that these results could reflect the “modesty”
norm process proposed by Stroebe et al. (1992). In other words, ftf participants could be
claiming a fewer number of ideas in accordance with social norms. And if the normative
response is to claim little credit, then it could be reasonable that negative mood
participants would be less likely to be led reflexively by normative behavior (e.g., Hertel

et al., 2001), therefore taking more credit than neutral mood participants.

 

” 2 participants from the nominal/neutral, l participant from the nominal/negative. 2 participants from the
ftf/neutral and l participant from the ftf/negative did not answer this question,

87

Since the brainstorming sessions were videotaped, ideas from ftf group
participants could be traced back to their original source. One research assistant viewed
the ftf sessions (since nominal participants deposited their ideas in their own individual
envelopes, a tally of how many ideas each nominal participant generated could be kept)
and recorded who said what idea. Prior to the arrival of participants, all materials were
labeled with subject numbers. Subject numbers were always designated in a systematic
fashion, such that the far right seat (when facing the participants) was always the lowest
number of the session. Subject number always increased from right to left. Therefore,
when viewing the tapes, as long as the session number was known (and therefore the
three possible subject numbers were known), the subject number of each participant
could be deduced based on their position at the table. Ideas were credited to whoever
vocalized the idea first. When variations of an idea were mentioned, whoever suggested
the variation received credit for the idea. There were no instances in which the same idea
was mentioned simultaneously by multiple group members. 2 X 2 ANOVA results show
a main effect for group on the average number of ideas that were mentioned by each
participant, F (1, 227) = 122.30, p = 0.001, such that nominal group participants (M =
13.91) generated significantly more ideas than ftf participants (M = 7.65). Because this
task occurred prior to the mood induction, no mood effect was expected, and no effect
was found, p = 0.64 (See Appendix I, Table 14 for condition means).

By taking the number of ideas that were actually generated by each participant,
and subtracting that number from the number that the participant estimated she generated,
a test of the misappropriation explanation could be conducted. 2 X 2 ANOVA results

show a main effect for group, F (1, 221) = 7.39, p = 0.007, such that nominal participants

88

(M = -1.76) actually came up with more ideas than they estimated, while ftf group
participants (M = 0.17) were rather accurate in their assessments. This again, directly
refutes the misappropriation explanation and to some degree, the modesty explanation
which was offered by Stroebe et al. to explain their results. No significant main effect for
mood was found, p = 0.11, and no interaction was found (See Appendix I, Table 15 for
condition means).
Additional Analyses

Expectations about performance in groups. After the brainstorming task was
completed, participants were asked to hypothetically imagine that they were to
brainstorm on the same problem, but in a different group setting. For example, nominal
group participants were asked to imagine brainstorming in a ftf group, while ftf group
participants were asked to imagine brainstorming in a nominal group. Participants were
then asked to assess whether they would have generated more or less ideas relative to the
amount of ideas that they had just come up with. When ftf group participants were asked
to answer using a 9-point scale, “If you were asked to brainstorm in a nominal group
(consisting of the same number of people as in your ftf group) on the same problem, do
you think you as an individual would have generated more or less ideas (1 = many less, 5
= same amount and 9 = many more)” the mean response was 4.63. This value was
significantly lower than the scale midpoint, t (113) = -2.3, p = 0.02, which shows that ftf
participants felt that they would be more productive in a ftf group setting. Comparisons
between the two ftf group conditions illustrate that mood had no effect on responses to
this question (M ftf/neutral = 4.76 and M ftf/negative = 4.52), F (1, 112) = 0.59, p = 0.45.

When nominal group participants were asked a similar question using the same scale, “If

89

you were asked to brainstorm in a ftf group (consisting of the same number of people as
in your nominal group) on the same problem, do you think you as an individual would
have generated more or less ideas?” the mean response was 5.83, which was significantly
higher than the scale midpoint, t (116) = 6.39, p < 0.01. Again, comparisons between the
two nominal group conditions show that mood did not affect responses to this question
(M nominal/control = 5.75 and M nominal/negative = 5.90), F (1, 115) = 0.313, p = 0.58.
Overall, these results suggest that even after brainstorming, participants in both group
conditions felt they would be more productive in ftf group settings. A 2 X 2 ANOVA
was not performed on these measures because nominal and ftf group participants were
asked separate questions that were specific only to their group condition, furthermore, the
scales that were used could not be transformed to make the responses across group
conditions compatible.

Extraversion measure. After participants completed the mood induction and their
performance assessment measures, they were also asked to complete an extraversion
scale. The ten items on this scale were combined to form a composite index of
extraversion (some items were recoded so that higher scores were more indicative of
being more of an extravert). The median score for this composite measure was 35.
Participants were classified as low in extraversion if they scored below 35 (n = 98), while
participants were classified as high in extraversion if they scored above 35 (n = 112).
Participants who scored 35 (n = 21) were not included in the analyses. Exploratory
analyses were conducted to examine whether extraversion, which has been associated
with feelings of positivity (Lucas & Fujita, 2000; Church & Burke, 1994), is related to

perceptions of satisfaction. Only participants in the neutral mood conditions were used

90

for the following analyses. First of all, participants high in extraversion (M = 2.88)
reported feeling more interested than participants low in extraversion (M = 2.44), F (1,99)
= 6.37, p = 0.013; more enthusiastic (M high = 2.47, M low = 1.90), F (l, 99) = 6.60, p =
0.01; more alert (M high = 3.36, M low = 2.54), F (1, 30) = 6.72, p = 0.015; and more
active (M high 2 2.63, M low = 2.08), F ( 1, 99) = 6.48, p = 0.012. Participants high in
extraversion also rated their current mood state as more positive (M = 6.43) compared to
participants low in extraversion (M = 5.85), F (1, 30) = 3.48, p = 0.07, however the
difference was only marginally significant. When high and low extraverts were
compared on scores from the positive PAN AS composite measure, the difference was
significant (M high 225.59, M low = 21.66), F (1,99) = 6.10, p = 0.015, thereby
replicating previous findings which illustrated the association between extraversion and
positivity. Interestingly, high and low extraverts did not differ along any of the negative
affect measures. Despite the differences in positivity, high and low extraverts in the
neutral mood condition did not differ in ratings of satisfaction (M high = 6.14, M low =

5.70), F (1, 99) = 1.40, p = 0.24.

91

 

DISCUSSION

In this study, as in the previously outlined studies (e. g., Stroebe et al., 1991;
Paulus et al., 1993; Nij stad, 2000), nominal groups outperformed ftf groups at the
brainstorming task. Also consistent with previous research, participants in the ftf groups
felt more satisfied with their personal performance, illustrating that this is quite a robust
effect. So why does the illusion occur? There is no direct evidence from this study to
suggest mood affects or mediates the relationship between group condition and
perceptions of satisfaction. Results did show that nominal group participants did feel
more distressed and negative than participants in the ftf group conditions, however
satisfaction was not associated with these mood measures.

So why was there no support for the mood explanation? One reason may have
been because participants did not assess perceptions of performance until approximately
five minutes after they completed the mood assessment questionnaire. It is therefore
possible that the mild negativity that participants in the nominal groups or negative mood
conditions were feeling was no longer potent enough to influence perceptions of
performance at the time of judgment. Perhaps, if a manipulation was included to sustain
negativity until the time of judgment, distress would be negatively related to satisfaction.
So for example, if negative mood inducing music was present when participants in the
negative mood condition assessed their performance (and neutral mood inducing music
was played for neutral participants), the likelihood that participants in the negative
conditions would feel negatively would be increased. In addition, it is quite possible that
participants in the negative mood conditions were able to attribute their negative mood to

the mood induction task, and therefore did not attribute their feelings of negativity to

92

 

their brainstorming performance. Again, a more indirect and diffuse mood manipulation
(e. g., background music) might provide a test of this possibility.

It could also be possible that perceptions of satisfaction are influenced by the
mood of participants during the task, rather than during the time of judgment. If we
presume that participants feel less negative while they are in a ftf group situation
compared to working alone, than this could explain why ftf group members reported
feeling more satisfied. Apart from inducing mood after the task rather than during the
task, one possible reason the test of mediation did not appear significant could be because
the mood was assessed approximately 15 minutes after brainstorming was completed.
Therefore, even if nominal participants felt more negative immediately after the task than
ftf participants, their negative mood may have diminished or dissipated by the time mood
was assessed which would reduce the relationship betWeen satisfaction and negativity.

This could be why the test of mediation using only neutral participants failed to
reach significance. However, the hypothesis that satisfaction is influenced by mood
during the time of the task rather than mood during the time of judgment making would
have to be based on the assumption that participants are recalling how they felt while
they worked at the task. To be consistent with the pattern of results, this would have to
be true unless, participants assess their performance during or immediately after the task
and mood influences judgments at that point in time. Then, it would have to be presumed
that when participants are later asked to assess their performance, they anchor around the
initial assessment that they made right after brainstorming. Because it is possible that
participants naturally assess their performance after completing a task, or that they recall

how they were feeling, the role of mood state during the task (as well as during the time

93

of judgment) should be explored. This could be done in a study much like the current
study if mood was also assessed immediately after the brainstorming session.

Could the illusion be driven by positivity amongst interacting group members,
rather than negativity by nominal group members? This speculation is unlikely because
results of this study show that groups did not differ along measures of positivity, yet, ftf
group members still reported feeling more satisfied. Furthermore, participants high in
extraversion reported feeling more positive, yet they were no more satisfied with their
performance compared to participants low in extraversion.

Comments on other eflects and previous explanations

Failure to replicate the illusion using the “rate the number of ideas ” measure.
Unexpectedly, in this study the four conditions did not differ on the Paulus et al. (1993)
measure of, “How would you rate the number of ideas that you as an individual were able
to generate?” This was a curious finding because in Paulus et al., ftf group members
gave higher ratings on this measure than nominal members. Because the instructions for
this study emphasized quantity as a basis for performance evaluation numerous times, it
was presumed that in the mind of the participant, satisfaction ratings would be driven by
how many ideas the participant came up with. In this study, although the ratings of
satisfaction were significantly positively correlated with the “number of ideas generated
measure” (Pearson ’s r = 0.48, p = 0.017), in retrospect, it is quite plausible however
though, that participants were basing their satisfaction rating on multiple dimensions, or
dimensions other than assessments of productivity. For example, despite knowing the
instructions, participants may have been basing ratings of satisfaction on not only how

many ideas they came up, but also on the perceived quality or creativeness of the ideas

94

generated, or any mood which couldn’t be simply attributed otherwise (e.g., to the mood
manipulation). In fact, post-hoe tests of mediation suggest that participants’ ratings of
satisfaction were indeed influenced by their assessments of creativity and quality.

Since “number of ideas generated” was positively correlated with satisfaction, it is
likely that satisfaction ratings were in part driven by perceptions of productivity. So, why
were differences between the nominal and ftf group participants present in satisfaction
ratings but not in the “number of ideas generated” measure? In agreement with Paulus et
al. (1993), it is possible that social comparison processes led to such effects. Recall,
participants in both type of work conditions had similar ratings on how many ideas they
felt they generated. Therefore as Paulus et al. would suggest, the work condition that
participants were in most likely affected how they interpreted these productivity
assessments. From observing members in their interactive group, ftf group participants
probably reached the conclusion that they performed either similarly to or better than
others. This type of conclusion could very well lead to higher ratings of satisfaction.
Nominal group participants on the other hand, would have no basis for comparison and
would therefore have no information that could serve to bolster their perceptions of
satisfaction. So, a social comparison process explanation could explain why there were
no differences in the “number of ideas generated” measure while there were differences
in satisfaction. However, if one interprets the social comparison explanation as
summarized by Ni jstad, then there is no support that shows the illusion effect can be
completely explained by social comparison processes because there were no differences
between conditions in ratings of how certain participants were of their relative ability

level.

95

 

Another possibility could be that ftf participants in this study gave lower ratings
on the “number of ideas” measure due to the operationalization of the fairness/modesty
norm proposed by Stroebe et al. (1992). If this is the case, then why didn’t ftf
participants in the original Paulus et al. (1993) study also lower their ratings in
accordance with such norms? It could be because Paulus et al. used both males and
females as participants, while the current study only recruited females. Previous research
has shown that women have traditionally been socialized to express positive emotions
about the successes of others, while being modest about their own achievements
(Stoppard & Gruchy, 1993; Rudman, 1998). So, female participants should be most
likely to adhere to such norms. Interestingly, Stroebe et al. (1992) only used female high
school students in Germany as participants.

Admittedly, it is unclear as to why no differences between groups were found in
this study using the Paulus et al. measure (“rate the number of ideas generated”), unless
one accepts the operation of a fairness/modesty norm amongst female participants. The
illusion effect has been consistently shown across studies only on the satisfaction variable
(e. g., Stroebe et al., 1992; Nijstad, 2000; and the current study). In the outlined studies,
participants also express a belief that ftf groups would outperform nominal groups, and
that they personally would perform better in ftf groups, however this is consistently
shown only when asked prior to brainstorming experience. So if anything, what this
suggests is that how the illusion is measured may be relevant, and if one decides to test
for the illusion effect, for whatever reason it may be more likely to be found with ratings
of satisfaction rather than with ratings of the “number of ideas generated,” (particularly if

using participants who are socialized to be modest).

96

 

SIAM. In this study, no support was found for the SIAM explanation proposed by
Nijstad (2000). This lack of support should be accepted with some caution however,
because the same measures that were used in the N i jstad failures index were not used in
this study. The failures index for this study was comprised of two of the original
measures, however the third (“Rate how often you had internally generated ideas which
had already been mentioned (very often-never)”) was purposefully omitted when the
materials for the current were designed. The rationale behind this omission was that it
was unclear whether the question was meaningful to nominal participants. In other
words, the third question was originally interpreted as meaning, “How often had you
internally generated ideas which had already been mentioned by others?” But it is at
least possible that the question could be in reference to “mentioned by oneself.” Even
though only two of the original three items were included in the failures index, if the
SIAM explanation were correct then failures should have still mediated the relationship
between group condition and satisfaction, unless the effect of failures all hinged on the
third item which was omitted. Regardless of this omission, Nijstad’s hypothesis still does
not seem to be the strongest or most predictive explanation available because again, his
explanation is only sufficient when examining the illusion effect in idea generation
contexts.

Misappropriation explanation. The results from the current study replicated
many of the original findings of Stroebe et a1 (1992). th group participants compared to
nominal participants, claimed a fewer percentage of the group’s ideas were “mentioned
by me,” however, reported that a higher percentage of ideas at least occurred to them.

Although these results appear to contradict the misappropriation explanation, Stroebe et

97

al. do provide a reasonable hypothesis for why support for the misappropriation

explanation cannot be found using “direct” measures. In their defense, Stroebe et al.

(1992) argue that:
For members of real groups, the main difference between claiming to have
suggested an idea and claiming merely to have had that idea without verbalizing it
is that the first claim is exclusive and the second is not. Members of real groups
who claim to have suggested an idea imply at the same time that other members
did not suggest it. Like any distribution task, this type of distribution is likely to
be affected by fairness considerations. . .For members of nominal groups, the
claim to have verbalized an idea is much less exclusive. These subjects
brainstormed individually and can therefore assume that the same idea might have
been reported by several members of their nominal group. . .The failure to find an
overestimation effect for subjects’ own contribution to the group product for
members of real groups is therefore likely to be due to the differential operation of
a fairness norm. Whereas members of real groups were restrained from making
this claim, members of nominal groups were not. . .Faimess considerations are less
relevant to the claim by subjects that an idea suggested by another subject also
occurred to them even though they did not verbalize it. . .Thus, when both groups
were unrestrained by fairness norms, members of real groups showed a much
stronger overestimation effect than members of nominal groups. (pg. 648)

So, Stroebe et al. seem to be suggesting that ftf group members are taking sole credit for

a fewer percentage of contributions than nominal members because when the same ideas

(or even similar ideas) are mentioned by multiple members, ftf group members are more

98

 

1F

reluctant to take credit for the ideas due to normative pressures. Importantly, they also
argue that when ftf group members are unrestrained from normative pressures, such as
when they are asked to think about how many ideas occurred to them, they overestimate
their contributions.

Especially because participants in the Stroebe et al. study and the current study
were all females, the likelihood of being influenced by fairness norms cannot be ruled
out. In addition, several of the results from this study could also indirectly be explained
by the differential Operation of fairness/modesty norms. For example, when participants
were asked to estimate how many ideas they personally had generated during the
brainstorming task, ftf group participants estimated a fewer number of ideas than nominal
participants. Negative mood participants also estimated generating more ideas than
neutral mood participants. The operation of fairness norms would predict both of these
results (the latter would be predicted given that negative mood reduces the likelihood of
normative or scripted behavior). Furthermore, no differences between groups were found
in this study when using the “rate the number of ideas” measure used by Paulus et al.
(1993), despite the fact that Paulus et al. found that ftf group members gave higher
ratings than nominal members. The differences between these studies on this measure is
difficult to explain unless one incorporates the operation of a fairness norm, and
considers the fact that participants in the current study were all female while Paulus’
participants were either male or female.

However, the fairness/modesty norm mechanism for the misappropriation
BXplanation should be accepted with caution because to date, researchers have not

explicitly shown that the norm is in operation in the experiment context. In addition, it

99

 

must be assumed that a fairness norm would be influential and in operation, despite self-
enhancement motives, even when participants know: 1) they will not see their fellow
group members again; 2) their fellow group members will not see their responses; and 3)
they will not see the experimenter again.

Social Comparison. Although the current study cannot rule out the role of social
comparison information, this study did find that no differences between groups existed
when participants were asked to assess how certain they felt about their ability estimates.
This contradicts the implied role of social comparison processes as outlined by Nijstad.
However, because Paulus et al. gave no mention to the influence of uncertainty on
perceptions of satisfaction, it is still possible that social comparison processes influence
performance perceptions, just not via uncertainty in ability leading to lower assessments
of performance.

Future directions and concluding remarks

The results of the current study do not support a mood explanation or
unequivocally any of the other previous explanations for the illusion of group
productivity. However, the role of mood in generating the illusion of productivity should
not be entirely discounted because it seems likely that this effect would exist in contexts
other than idea generation tasks. A mood explanation would be a parsimonious theory to
explain the effect across contexts, something only a social comparison explanation could
do as well (to date). Furthermore, a mood explanation should not be ruled out at this
point because the placement of the mood induction may matter, something that has yet to
be explored. Future research should attempt to vary the placement of the mood induction

(e. g., before or during the task) to examine, for example, if mood during the time of

100

 

performance is what influences judgments. A study in which the task is varied could also
be conducted to explore the relative generalizability of the illusion effect. Results from
such a study should shed light on the strength of some of the previously proposed
explanations.

Post-hoe analyses from this study show that assessments of creativity and quality
each partially mediate the relationship between group and satisfaction. These are the first
results to date to hint that the illusion may be driven by perceptions along these
dimensions. Because people often feel that working in groups stimulates creativity (e.g.,
Nijstad, 2000), it is possible that participants in ftf groups are reporting more satisfaction
because they expect that their ideas will be more creative and of higher quality. Although
they are specifically instructed to emphasize productivity (i.e., quantity), it is possible
that they are inherently considering quality in their assessments of performance as well.
Future research should attempt to better understand the potential role of ratings of quality
and creativity towards the illusion effect, and should attempt to investigate whether
“illusions of group creativity” exist, and if so, why.

Because the misappropriation and social comparison explanations could not be
ruled out from the results of the current study, both explanations need to be scrutinized in
more detail. One possible method of examining the misappropriation explanation would
be do conduct a study using participants who will not be as likely to adhere to
fairness/modesty norms (e. g., males), or to manipulate the norms experimentally. If the
same pattern of results occur, despite the manipulation of norms (for example), then it

would be difficult to accept the misappropriation explanation given that ftf participants

101

 

take less credit for ideas rather than more credit. This would be especially true if the
illusion is found to exist in contexts other than idea generation tasks.

A direct test of the social comparison explanation could also be conducted by
designing a study in which the level and nature of social comparison information is
varied. For example, participants could be asked to work on an electronic brainstorming
task. In one condition, participants could be instructed to generate as many ideas as
possible while being provided with no social comparison information. In a second
condition, participants would work on the same task, however, they would be provided
with information on their computer which informs them how many ideas they have come
up with, as well as how many ideas others in their session had come up with. The nature
of this information could be manipulated such that some participants would learn they
were performing much worse than others, some would learn they were performing about
equally, while others would learn they were performing much better than others. By
manipulating the nature of the information that is provided to participants, we can better
attempt to understand how social comparison information might influence perceptions of
performance. For example, if providing information, even if it indicates poor
performance, leads to higher ratings of satisfaction compared to participants who are
given no information, than support for the hypothesis that uncertainty about one’s ability
level leads to negative perceptions would be supported. However, if this pattern does not
occur, than it must be assumed that providing information alone is not sufficient to lead
to illusory perceptions, and that feelings of uncertainty are not what lead people to
perceive their performance negatively. One advantage of having participants working

individually at the task is that any potential effect of social comparison information can

102

be isolated from any other potential influences that usually reside in group contexts.
Therefore, if differences are found, then the implication is that the illusion is driven by
the information that is generally gained in group contexts, and not by other factors that
may be present while people interact or work in groups.

The purpose of this study was to both examine mood as a potential explanation
for the illusion effect, and to further test the explanations that have been previously '
offered. The results of this study suggest that mood at the time of judgment is not a t-

feasible explanation for the illusion effect. This study also found no support for the

 

failures explanation offered by N ijstad. The social comparison explanation cannot be
ruled out, however, it is unlikely that social comparison processes lead to the illusion via
the reduction of uncertainty in ability. Similarly, the misappropriation explanation
cannot be discounted, however, again there was no direct evidence to support this
hypothesis. So, explanations for the illusion of group productivity cannot be offered
conclusively from this study, but results'do shed light on what factors are unlikely to
explain the effect. Furthermore, the notion that assessments of quality and creativity
mediate the illusion effect introduces an optimistic avenue to pursue in the future.

There are several interesting findings from the current study that need to be
examined in more detail in the future. Returning to the initial question that was posed in
the introduction about the nature of groups, why do participants in ftf groups feel less
distressed? Participants whose mood was not manipulated (neutral mood participants)
felt less distressed and less negative overall after working in a group compared to
working alone. Interestingly, participants in ftf groups did not feel more positive, rather

they felt less negative. Why? Are these affective differences due to the fulfillment of

103

particular social needs (e. g., need to belong and need for affiliation) or do they result
from aspects of the group interaction (e. g., social validation)? Would merely identifying
with a group, without social interaction, also decrease negative affect?

Why people feel differently when they are in groups compared to being alone is
still a fascinating mystery. The current study has in some ways addressed what people
feel when they are in groups, and has in some ways addressed how those feelings lead to
interesting outcomes. Hopefully in the near future, many of the questions that have been
proposed can be answered so that not only a better understanding of group dynamics can

be obtained, but also a better understanding of humans as social animals.

104

 

Appendix A: Experimenter Script

Experim enter Script
Before the study:

1) set up dividers in room 441

2) check CONDITION SHEET to determine which condition (FTF vs. Nominal) (Negative vs. Control)
should be run

3) In rm 441, place consent form 1 and pens in each cubicle. In room 439, place pens and appropriate
storytelling instructions in each cubicle

4) label subject number, group number, date and session time on back of each consent form and all of the
materials

5) make sure you have pre-measured stacks of index cards

After sutiects arrive:

I) escort subjects to Rm 441. Ask them to take off their watches and to put them in their pockets or bags.
They can leave their coats and belongings on the table next to the door.

2) ask them to have a seat in one of the cubicles, and ask them if they want to participate. If so, ask them
to please sign the consent form

3) If only 1 or 2 participants arrive, run them in a nominal condition. Make sure to change all of the labels
on their materials, and make sure to make a note in the log book

4) After the consent forms are signed, collect them and put them in the folder labeled “COMPLETED
CONSENT FORMS."

5) Tell the participants the following:

“Today you will be working on a series of tasks. The tasks that you will be participating in are
actually from two unrelated studies. Because each of these studies do not require a lot of time, the
experimenters have combined them for the sake of efficiency.”

“The first task that you will be working on is an idea generation task. This procedure is commonly
known as brainstorming. The purpose of brainstorming is to come up with as many ideas as
possible. It is actually quantity, and not quality that is desired in brainstorming tasks.

Brainstorming can be conducted in several ways. Some people design brainstorming tasks to be done
by nominal groups. Brainstorming in nominal groups is when individuals brainstorm on a given
topic in isolation. In other words, the individuals in nominal groups do not see each other or
communicate with one another. They simply brainstorm on the topic alone. They are called groups
because at the end of the brainstorming session, all of the ideas that are generated by each of the
individuals are all added together to form a group product. Ideas that are mentioned multiple times,
in other words, by more than one individual, are only counted once. So, the group product for a
nominal group is the total number of non-redundant ideas that are generated by all of the individuals
who brainstormed in isolation.

In nominal groups, combining and improving on ideas that you have already mentioned is
encouraged. Again, it is quantity rather than quality which is emphasized.

Does anyone have any questions about what a nominal group is or how the product of a nominal
group is assessed?”

105

 

“Brainstorming can also be conducted using face-to-face groups. In FTF group brainstorming,
group members are allowed and encouraged to talk and interact with one another. In FTF groups,
one member is assigned the role of the recorder. So this person records all of the ideas that are being
generated by the group. The recorder is also allowed to generate and contribute ideas, the only
difference between the recorder and the other group members is that the recorder has the
responsibility of writing down each of the ideas that are mentioned.

At the end of the brainstorming session, all of the ideas that are generated by the group are counted
up. Ideas that were mentioned more than once are only counted once. Therefore, the group product
is the total number of non-redundant ideas that are generated by everyone in the group.

In FTF groups, criticism of ideas is prohibited. Just like nominal groups, combining and improving
on ideas that are already mentioned is encouraged. Again, it is quantity rather than quality which is
emphasized.

Does anyone have any questions about what a FTF group is or how the product of a F'I‘F group is
assessed?”

“If there are no (more) questions at this time, please answer the following questionnaire.”
Experimenter passes out “PRE-TASK QUESTIONNAIRE.”

When participants complete this form, experimenter collects it and puts it in the folder marked
“COMPLETED PRE-TASK Q.”

Experimenter starts video camera.

“Now you are going to be asked to brainstorm on a given task.”

FOR NOMINAL CONDITION:

“This study concerns how people brainstorm in nominal conditions. Previously, I defined what a
nominal group is. Does anyone have any questions about nominal groups at this time?”

Experimenter hands out pre-measured stacks of index cards and large envelope to participants.

“1 am handing each of you a set of index cards and a folder. You are to write down one idea per
card. You may look over your own cards to combine ideas or to improve on ideas that you already
wrote down. When writing your ideas, express them clearly, but do not worry about proper
grammar, or writing in complete sentences. When time is up, place all of your completed index cards
in the folder. Remember, only write down one idea per card.

When I tell you to begin, you will brainstorm on the following topic: Your topic is to think about all
of the practical benefits and difficulties that would arise if everyone were to have an extra thumb on
each hand.

Remember, quantity and not quality is what is emphasized. And, feel free to combine and improve
on ideas that you already mentioned. If there are no further questions, you may begin brainstorming

now.”

Experimenter leaves the room and times for 8 minutes. After eight minutes is up, the experimenter says,
“FINISH UP THE IDEA THAT YOU ARE WORKING ON NOW AND THEN STOP” and then

106

 

collects each of the folders which contains the index cards. Collect the folders from left to right so that it is
possible to track which subject came up with which set of ideas. Place the folders down in order.

FOR FTF CONDITION:

“This study concerns how people brainstorm in FTF group conditions. Previously, I defined what a
FTF group is. Does anyone have any questions about FTF groups at this time?”

Experimenter removes the barriers and hands out a stack of index cards and a large envelope to the
participant sitting in the middle.

To the participant in the middle: “You have randomly been chosen to be the group recorder. It is your
responsibility to write down all of the ideas that are generated by your group. You may also
contribute ideas, but remember, it is ultimately your responsibility to record the ideas from your
group.

You are to write down one idea per card. You all may look over the cards to combine ideas or to
improve on ideas. You may also communicate with one another during the task. When writing your
ideas, express them clearly, but do not worry about proper grammar, or writing in complete
sentences. When time is up, place all of your completed index cards in the folder. Remember, only
write down one idea per card.

When I tell you to begin, you will brainstorm on the following topic: Your topic is to think about all
of the practical benefits and difficulties that would arise if everyone were to have an extra thumb on
each hand.

Remember, quantity and not quality is what is emphasized. Criticism of ideas is prohibited. And,
feel free to combine and improve on ideas that you already mentioned. If there are no further
questions, you may begin brainstorming now.”

Experimenter leaves the room and times for 8 minutes. After eight minutes is up, the experimenter says,
“FINISH UP THE IDEA THAT YOU ARE WORKING ON NOW AND THEN STOP” and then
collects each of the folders which contains the index cards.

ALL CONDITIONS:

Experimenter stops the video camera.

“Now, you will be working on another short task. This task is part of another study which we have
agreed to run, so please look over and sign the following consent form for this next task.”

Experimenter hands out Consent form 2. Waits for participants to sign it, and then collects it and puts it in
the envelope labeled Completed Consent forms.

When all consent forms are completed, experimenter says, “Ok, now we will move on to the next study,
so please come with me at this time.”

Experimenter escorts participants to room 439. Assigns one participant to each cubicle (and one in the
hallway). Directions for appropriate condition of STORYTELLING EXERCISE should be in place.

107

 

Experimenter says to all participants: “In front of you, there should be a sheet of paper. Please read all
of the instructions on the sheet of paper, and follow the directions. When you are done reading the
instructions, you may begin working at the task.”

Experimenter times for 9 minutes. During this 9 minute period, experimenter goes back and labels the
index cards. For the FTF group, an index card with the subject numbers and session time is put on the top
of the set of ideas. A rubber band is put on the stack of ideas, with the subject number label on top, and
then the ideas are put into the folder labeled “INDEX CARDS.”

For the nominal group, a separate subject label card which includes the subject number and session time, is
attached to each set of cards. The experimenter also puts the dividers in room 441 back up.

After the nine minute period, the experimenter goes to room 439 with an empty folder labeled
“COMPLETED ST” and collects the storytelling forms. Do not look over the forms, or it may make the
participants feel awkward. When all of the forms are collected, the experimenter asks the participants to
fill out the mood assessment questionnaire. After the mood assessment questionnaire is complete, the
experimenter collects them and puts them in the folder labeled “MA” and then asks the participants to
come back to room 441 and asks them to sit where they sat before.

The experimenter then says the following: “OK, now I’d like you to switch gears and take some time to
think about the first brainstorming task that you all previously completed. Close your eyes, take a
deep breath, and think about how you performed at the task. Keep reflecting on your experience
until I tell you to stop.”

Experimenter leaves the room and comes back after 2 minutes. When the 2 minute period is up,
experimenter hands out the final questionnaire packet which includes, the

performance assessment form, the global beliefs form, the extraversion measure, and then the
manipulation check.

After handing out the packet, experimenter says, “While answering the following questionnaires, please
do not interact or communicate with one another.”

After participants all finish, the experimenter collects the forms, and then says the following:

“In the near future, you will be sent an e mail message concerning this study. The subject heading
will be labeled, “MSU Psychology Study.” This message will contain a very quick survey, which
should take no longer than two minutes to complete. When you get this message, please take the time
to answer the survey and send it back electronically. When all of the data is collected, one
participant will be randomly chosen to win 50 dollars. Only participants who send back the survey
will be eligible, so please make sure we have your e mail address, which should be on your consent
form, and please make sure you send back the message after it is sent to you.”

Experimenter hands out the debriefing form and gives participants credit.

108

 

Appendix B: Pre-task Questionnaire

Imagine that the President of the University would like a council of students to generate ways to
increase student satisfaction concerning campus life. Imagine that you, along with a select
number of other students are asked to come up with as many ideas as possible.

1) Which type of group is more likely to generate the most ideas? (place a check in the
space provided to indicate your answer)

nominal group: _ : _ : _ : _ : _ : _ : _ : _ : _:face-to-face
neither

2) How confident are you in your answer? (please circle a number)

1 = not at all confident
2
3
4
5
6
7
8
9 = extremely confident

3) In which type of group would you personally generate more ideas? (place a check in
the space provided to indicate your answer)

nominal group: _ : _ : _ : _ : _ : _ : _ : _ : _:face-to-face
neither

4) How confident are you in your answer? (please circle a number)

1 = not at all confident
2
3
4
5
6
7
8
9 = extremely confident

5) In which type of group would you prefer to work in? (place a check in the space
provided to indicate your answer)

109

nominal group: _ : _ : _ : _ : _ : _ : _ : _ : _:face-to-face
neither

6) Please explain why you chose your answer for the previous question?

7) How do you think you personally, relative to other students at MSU, would perform
at the brainstorming task under nominal group conditions (e.g., working individually and
then adding up your performance with the performance of all of the others’ in your
nominal group)

1 = not well at all
2
3
4
5
6
7
8
9 = extremely well

8) If you were given 10 minutes to complete the brainstorming task, how many ideas
would you, personally, come up with under nominal group conditions?

1 = not many ideas at all
2
3
4
5
6
7
8
9 = very many ideas

9) Out of the total number of non-redundant ideas from a nominal group (e.g., all of the

ideas suggested, counting ideas mentioned more than once ONLY once), what percentage
do you think would be suggested by you?

 

10) Out of the total number of non-redundant ideas from a nominal group, what
percentage do you think would be suggested by others, and would not have occurred to
you?

 

110

 

11) Under nominal group conditions, what percentage of the total ideas from all group
members do you think would be redundant (mentioned more than once)?

 

12) How do you think you personally, relative to other students at MSU, would perform
at the task under face-to-face group conditions? (please circle a number)

1 = not well at all
2
3
4
5
6
7
8
9 = extremely well

13) If you were given 10 minutes to complete the brainstorming task, how many ideas
would you, personally, come up with under face-to-face group conditions?

1 = not many ideas at all
2

3

4

5

6

7

8

9 = very many ideas

14) Out of the total number of non-redundant ideas from a face-to-face group (e.g., all of

the ideas suggested, counting ideas mentioned more than once ONLY once), what
percentage do you think would be suggested by you?

 

15) Out of the total number of non-redundant ideas from a face-to-face group, what
percentage do you think would be suggested by others (and not mentioned by you), but
would have also occurred to you?

 

16) Out of the total number of non-redundant ideas from a face-to-face group, what
percentage do you think would be suggested by others, and would not have occurred to
you?

 

17) Under face—to-face group conditions, what percentage of the total ideas from all
group members do you think would be redundant (mentioned more than once)?

lll

 

Appendix C: Mood Induction Instructions

Neutral instructions:

The Department of Psychology is currently developing a questionnaire to assess life events in a systematic
and reliable manner. In order to construct this questionnaire, we need to first collect a large sample of
positive, negative, and neutral life events from which to choose appropriate materials. So at this time, we
would like you to recall, re-experience, and write about the mundane (or everyday) events of the previous
day. In other words, write about the normal, everyday events that occurred yesterday. The events that you
describe should not be particularly positive or negative, but rather neutral. For example, you may choose
to write about going to class, or going to the store to buy groceries. Please take a moment now to think
about the events that you will describe. Because no one except for the experimenter will read your
description, and because your description will not be matched with your personal identity in any way, we
would appreciate it if you would answer as truthfully as possible. Please use as much space (including the
back of the paper) as you feel necessary, and write as neatly as possible. You will have about ten minutes
to complete this task. You may begin now.

Negative instructions:

The Department of Psychology is currently developing a questionnaire to assess life events in a systematic
and reliable manner. In order to construct this questionnaire, we need to first collect a large sample of
positive and negative life events from which to choose appropriate materials. So at this time, we would
like you to recall, re-experience, and write about a personal life event that made you particularly sad.
Please take a moment now to think about the sad life event that you will describe. When you write about
the life event, think about how you felt at the time of the event, and please try to focus on concrete and
vivid aspects of the event rather than on an abstract or objective assessment of it. Because we are
interested in collecting in-depth information, please include in your account, how the event happened to
come about. Because no one except for the experimenter will read your description, and because your
description will not be matched with your personal identity in any way, we would appreciate it if you
would answer as truthfully as possible. Please use as much space (including the back of the paper) as you
feel necessary, and write as neatly as possible. You will have about ten minutes to complete this task. You
may begin now.

112

 

Appendix D: Mood Assessment (PANAS)

We would like you to please answer the following questions. No one besides the
experimenter will see your answers, so please answer as honestly as possible.

The following scale consists of a number of words that describe different feelings and
emotions. Read each item and then mark the appropriate answer in the space next to that
word. Indicate to what extent you feel this way right now, that is, at the present moment.
Use the following scale to record your answers.

1 = very slightly or not at all
2 = a little

3 = moderately

4 = quite a bit

5 = extremely

_interested __irritable
_distressed _alert
_excited _ashamed
_upset _inspired
_strong _nervous
_guilty _determined
_scared _attentive
_hostile __jittery
_enthusiastic _active
_proud _afraid

Now, please answer the following questions using the following 1-9 point scales:

Please rate your current mood state
1: very negative 5=neutral 9=very positive

113

Appendix E: Post-Brainstorming Task Questionnaire (Nominal Version)

The following questions concern the Brainstorming Task (the “thumbs” problem) that you
previously completed:

1) Using a 1 to 9 scale, please indicate how satisfied you are with your performance at the
Brainstorming Task? (circle a number)

1 = not at all satisfied
2
3
4
5
6
7
8
9 = extremely satisfied

2) Please explain why you chose your answer for question #1.

3) Using a 1 to 9 scale, please rate the number of ideas you as an individual were able to generate.
I was able to generate: (circle a number)

1: very few ideas
2
3
4
5
6
7
8
9 = very many ideas

4) In the space provided, please indicate how many uses you think you as an individual generated
during the Brainstorming Task? (open-ended)

114

 

5) If you had been asked to brainstorm in a face-to-face group (consisting of the same number of
people as in your nominal group) on the same problem, do you think you as an individual would
have generated more or less ideas? I would have generated: (circle a number)

1: many less ideas

= the same amount of ideas

OO\IO\UI-hb-IN

9 = many more ideas

6) On a brainstorming task, which type of group do you think would come up with more ideas?
(circle a number)

1: nominal group would come up with many more ideas

2

3

4

5 = both groups would come up with the same amount of ideas
6

7

8

9

= face-to-face group would come up with many more ideas

7) Out of the total number of non-redundant ideas from your nominal group (e. g., all of the ideas
suggested, counting ideas mentioned more than once ONLY once), what percentage do you think
were suggested by you?

 

8) Out of the total number of non-redundant ideas from your nominal group, what percentage do
you think were suggested by others, and would not have occurred to you?

9) How much effort did you put into the brainstorming task? (circle a number)

1: no much at all
2
3
4
5
6
7
8
9 = very much

115

10) How much pressure did you feel to come up with ideas? (circle a number)

1: not much at all
2
3
4
5
6
7
8
9

11) How difficult was it to concentrate while trying to generate new ideas? (circle a number)

 

1: not difficult at all
2

3

4

5

6

7

8

9 = very difficult

12) How do you think you personally, relative to other students at MSU, performed at the
brainstorming task? (circle a number)

1: not well at all
2
3
4
5
6
7
8
9 = extremely well

13) How certain are you of your answer (to the previous question)?

1: not certain at all
2
3
4
5
6
7
8
9 = extremely certain

ll6

14) How much did you enjoy the Brainstorming task? (circle a number)

1: not much at all
2
3
4
5
6
7
8
9 = enjoyed very much

15) During the brainstorming task, how difficult was it to generate ideas? (circle a number)

= not at all difficult

 

l
2
3
4
5
6
7
8
9 = extremely difficult

16) During the brainstorming task, how often were you unable to come up with a new idea?
1: not very often
2
3
4
5
6
7
8
9 = very often

17) How creative do you think your ideas were from the Brainstorming Task? (circle a number)
1: not creative at all
2
3
4
5
6
7
8
9

= extremely creative

117

18) How you feel about the quality of your ideas from the Brainstorming task? (circle a number)

1: very poor quality
2
3
4
5
6
7
8
9 = extremely high quality

19) If you had been asked to brainstorm in a face-to-face group (consisting of the same number of

people as in your face-to-face group) on the same problem, do you think you as an individual
would have generated better or worse quality ideas? I would have generated: (circle a number)

= much worse quality ideas

1
2
3
4
5 - the same quality ideas
6
7
8
9

= much better quality ideas

20) Relative to other students at MSU, how well do you think you performed at the
Brainstorming Task? I expect that I performed better than of the students.

a) 10-19%
b) 20-29%
c) 30-39%
(1) 40-49%
e) 50-59%
1) 60-69%
g) 70-79%
h) 80-89%
1) 90-99%
j) 100% (I performed better than any other MSU student)

21) How motivated were you to perform well during the brainstorming task?

1: not at all motivated
2
3
4
5
6
7
8
9 = extremely motivated

118

 

Appendix F: Group Beliefs Questionnaire

This questionnaire is designed to assess people’s general working styles and preferences. The
term “interactive group” (also known as a face-to-face group) which you will see below, refers
to a group in which the members talk and interact with one another. The term “nominal group”
refers to a group in which members work individually and do not communicate. The group
product for a nominal group is generated by summing up all of the individual contributions from
the nominal group members. Please take your time and answer every question. Thank you.

1) When it comes to brainstorming (i.e., generating ideas), interactive groups (also known as
face—to-face groups) generate more ideas than nominal groups (where members work individually
and whose individual products are added together). (circle a number)

1 2 3 4 5 6 7 8 9
I disagree completely (1) I agree completely (9)

2) How confident are you in your answer (to the previous question)? (circle a number)

1 2 3 4 5 6 7 8 9
not at all confident (l) extremely confident(9)

3) If I worked on a brainstorming task (i.e., the thumbs problem) in an interactive groap, I would
probably produce ideas in a 5 minute period. (Fill in the blank)

4) Excluding myself, the average member of my interactive group would probably produce
ideas in a 5 minute period. (Fill in the blank)

 

5) If I worked on a brainstonning task (i.e., the thumbs problem) in a nominal ggoup, I would
probably produce ideas in a 5 minute period. (Fill in the blank)

6) Excluding myself, the average member of my nominal group would probably produce
ideas in a 5 minute period. (Fill in the blank)

7) Regardless of the nature of the task, I am generally more satisfied with my personal
performance when I work

1 2 3 4 5 6 7 8 9
alone (1) in an interactive group(9)

119

 

8) For problem-solving tasks, I am more satisfied with my personal performance when I work

 

1 2 3 4 5 6 7 8 9
alone (1) in an interactive group (9)

9) For most types of tasks, interactive groups are more likely to do a better job than nominal
groups. (circle a number)

1 2 3 4 5 6 7 8 9
I disagree completely (1) I agree completely (9)

10) I feel anxious when working with other people. (circle a number)

1 2 3 4 5 6 7 8 9
I disagree completely (1) I agree completely (9)

11) If the task is important to me, I generally prefer to work

1 2 3 4 5 6 7 8 9
alone (1) in an interactive group (9)

12) If the task is important to me, I generally do a better job when I work

1 2 3 4 5 6 7 8 9
alone (1) in an interactive group (9)

13) If the task is apt important to me. I generally prefer to work

1 2 3 4 5 6 7 8 9
alone (1) in an interactive group (9)
14) If the task is apt important to me, I generally do a better job when I work
1 2 3 4 5 6 7 8 9
alone (1) in an interactive group (9)
15) I enjoy myself more when I work

1 2 3 4 5 6 7 8 9
alone (1) in an interactive group (9)

120

 

16) In general, 1 am more creative when I work

1 2 3 4
alone (1)

121

8 9
in an interactive group (9)

 

 

Appendix G: Extroversion Measure

On the following pages, there are phrases describing people's behaviors. Please
use the rating scale below to describe how accurately each statement describes
you. Describe yourself as you generally are now, not as you wish to be in the
future. Describe yourself as you honestly see yourself, in relation to other
people you know of the same sex as you are, and roughly your same age. So
that you can describe yourself in an honest manner, your responses will be kept
in absolute confidence. Please read each statement carefully, and then in the
space provide, write in an answer that corresponds to the number on the scale.

Response Options

 

: Very Inaccurate

: Moderately Inaccurate

: Neither Inaccurate nor Accurate
: Moderately Accurate

: Very Accurate

UlhbJNt-t

1. __ Am the life of the party.

2. __ Don’t talka lot.

3. __ Feel comfortable around people.

4. __ Keep in the background.

5. _ Start conversations.

6. __ Have little to say.

7. _ Talk to a lot of different people at parties.
8. __ Don’t like to draw attention to myself.

9. _ Don’t mind being the center of attention.
10. __ Am quiet around strangers.

122

 

Appendix H: Manipulation Check

Today you took part in several tasks.

1) Was there anflhing unusual about the first part of the study (where you brainstormed)?
If so, please explain in detail.

2) Was there anflhing unusual about the second part of the study (where you wrote about
a life experience)? If so, please explain in detail.

3) What do you think was the purpose of this study?

123

Table 1

Appendix 1: Condition means

Mean Ratings of Upset (Group level) By Condition

 

 

 

 

Condition n Mean Rating SE
Nom-neutral 19 1 .53 0. l4
Nom-negati ve 20 2.32 0. 18
th-neutral 18 1.17 0.08
th-negative 20 2.32 0.18
Table 2

Mean Ratings of Distress By Condition

Condition n Mean Rating SE
Nom-neutral 57 1.95 0.15
Nom-negative 60 2.40 0. 16
th-neutral 54 1.50 0.12
th—negative 60 2.25 0. 15
Table 3

Mean Number of Nonredundant Ideas Generated By Condition

 

Condition n Ideas per group SE
Nom-neutral 19 38.00 1.91
Nom—negative 20 38.80 1.87
th—neutral 1 8 22.00 1 .97
th—negative 20 24. 10 1 .86

 

124

 

 

Table 4

Mean Ratings of Personal Performance Satisfaction By Condition

 

 

 

Condition n Mean Rating SE
Nom-neutral 57 5.54 0.23
Nom-negative 60 5.25 0.23
th-neutral 54 6.24 0.24
th—negative 60 6.03 0.23
Table 5

Mean Ratings of Number of Ideas Generated By Condition
Condition n Mean Rating SE
Nom-neutral 57 5.49 0.23
Nom-negative 60 5.68 0.22
th-neutral 54 5.52 0.24
th-negative 60 5.63 0.22

 

Table 6

Mean Ratings of Personal Performance Satisfaction After 2 Weeks

 

Condition n Mean Rating SE
N om-neutral 31 5.52 0.28
Nom-negative 30 5.40 0.32
th-neutral 24 5.67 0.35
th-negative 33 6.12 0.34

 

125

Table 7

Perceived Ability Relative to Other MS U Students

 

 

 

Condition 11 Mean Rating SE
Nom-neutral 57 5.47 0. 17
Nom-negative 60 5.30 0. 17
th-neutral 54 5 .70 0.18
th-negative 60 5.56 0.17
Table 8

Certainty Ratings For Perceived Ability

 

 

Condition n Mean Rating SE
Nom-neutral 57 5.54 0.26
Nom-negative 60 5.57 0.26
th—neutral 54 5.44 0.27
th-negative 60 5.88 0.27
Table 9

Two-item Failures Index Scores

 

Condition n Mean Rating SE
Nom-neutral 57 10.60 0.42
Nom-negative 60 10.90 0.41
th—neutral 54 10.48 0.43
th—negative 60 10.24 0.41

 

126

 

Table 10

Percentage of Ideas Thought of By Me and Mentioned by Me

 

 

Condition n Mn PerceLtage SE
Nom-neutral 3 1 43 .39 2.00
Nom-negative 30 43.14 2.15

th-neutral 24 35.69 2.89

th-negative 33 37 .29 3.14

Table 11

Percentage of Ideas Thought of By Others and Did Not Occur to Me

 

 

Condition n Meapr Percentage SE
Nom-neutral 3 1 44.95 2.07
N om-negative 30 44.37 2.66
th—neutral 24 37.77 3.60
th-negative 33 36.27 3.41
Table 12

Percentage of Ideas Mentioned By Others But Also Occurred to Me

 

Condition n Mean Percentage SE
Nom-neutral 31 12.99 1.87
Nom-negative 30 1 1.05 1.76
th—neutral 24 26.36 3.20
th-negative 33 24.93 2. 80

 

127

Table 13

Estimated Number of Ideas Personally Generated (Open-Ended)

 

 

Condition n Mean Number SE
Nom-neutral 55 l 1 .56 0.89
Nom-negative 59 12.71 1.07

th-neutral 52 6.92 0.50
th—negative 59 8.71 0.61

Table 14

Actual Number of Ideas Personally Generated

 

 

Condition n Mean Number SE
Nom-neutral 57 14.02 0.56
N om-negative 60 13.82 0.68
th—neutral 54 7.28 0.49
th—negative 60 8 .02 0.51
Table 15

Actual Number of Ideas Generated Subtracted From Estimated

 

Condition n Mean Number SE
N om-neutral 55 -2.38 0.78
N om-negative 59 -1. 14 0.90
th—neutral 52 -0.35 0.51
th-negative 59 0.68 0.54

 

128

References

Baron, R. M., & Kenny, D. A. (1986). The mediator-moderator variable distinction in
social psychological research: Conceptual, strategic, and statistical
considerations. Journal of Personality and Social Psychology, 51, 1173-1182.

Bless, H., Clore, G. L., Schwarz, N., Golisano, V., Rabe, C., & Wolk, M. (1996). Mood
and the use of scripts: Does a happy mood really lead to mindlessness? Journal
of Personality and Social Psychology, 71, 665-679.

Bodenhausen, G. V., Kramer, G. P. & Siisser, K. (1994). Happiness and stereotypic
thinking in social judgment. Journal of Personality and Social Psychology, 66,
621-632.

Bruner, J. S. (1957). On perceptual readiness. Psychological Review, 64, 123-152.

Carson, T. P., & Adams, H. E. (1980). Activity valence as a function of mood change.
Journal of Abnormal Psychology, 89, 368-377.

Carver, C. S., & Scheier, M. F. (1981). Attention and self-regulation: A control theory
approach to human behavior. New York: Springer-Verlag.

Church, A. T., & Burke, P. J. (1994). Exploratory and confirmatory tests of the Big Five
and Tellegen’s three- and four- dimensional models. Journal of Personality and
Social Psychology, 66, 93-114.

Clark, L. F., & Isen, A. M. (1982). Towards understanding the relationship between
feeling states and social behavior. In A. Hastorf & A. M. Isen (Eds), Cognitive
social psychology (pp.73-108). New York: Elsevier.

Clark, L. F., & Teasdale, J. D. (1982). Diurnal variation in clinical depression and
accessibility of memories of positive and negative experiences. Journal of
Abnormal Psychology, 91, 87-95.

Clore, G. L., Schwarz, N., & Conway, M. (1994). Cognitive causes and consequences of
emotion. In R. S. Wyer & T. K. Srull (Eds), Handbook of social cognition (2“d
ed., pp. 323-417). Hillsdale, NJ : Lawrence Erlbaum Associates.

Costa, P. T., & Macrae, R. R. (1980). Influence of extraversion and neuroticism on
subjective well-being: Happy and unhappy people. Journal of Personality and
Social Psychology, 38, 668-678.

Diehl, M., & Stroebe, W. (1987). Productivity loss in brainstorming groups: Toward
the solution of a riddle. Journal of Personality and Social Psychology, 53, 497-
509.

129

Diehl, M., & Stroebe, W. (1991). Productivity loss in idea-generating groups: Tracking
down the blocking effect. Journal of Personality and Social Psychology, 61, 392-
403.

Diener, B., & Emmons, R. A. (1985). The independence of positive and negative affect.
Journal of Personality and Social Psychology, 47, 1105-1117.

Eich, E., Kihlstrom, J. F., Bower, G., H., Forgas, J. P., & Niedenthal, P. M. (2000).
Cognition and Emotion. New York: Oxford Press.

Forgas, J. P. (1990). Affective influences on individual and group judgments. European
Journal of Social Psychology, 20, 441-453.

Forgas, J. P. (1992). Affect in social judgments and decisions: A multi-process model.
In M. P. Zanna (Ed.), Advances in experimental social psychology (vol. 25, pp.
227-275). San Diego, CA: Academic Press.

Forgas, J. P. (2001). Handbook of affect and social cognition. Mahwah, NJ: Lawrence
Erlbaum Associates.

Forgas, J. P., & Bower, G. H. (1987). Mood effects on person perception judgments.
Journal of Personality and Social Psychology, 53, 53-60.

Goethals, G. R., & Darley, J. M. (1987). Social comparison theory: Self-evaluation and
group life. In B. Mullen & G. R. Goethals (Eds), Theories of group behavior
(pp.21-48). New York: Springer-Verlag.

Hertel, G. & Kerr, N. L. (2001). Priming in-group favoritism: The impact of normative
scripts in the minimal group paradigm. Journal of Experimental Social
Psychology, 37, 316-324.

Hinsz, V. B. (1992). Social influences on the goal choices of group members. Journal
of Applied Social Psychology, 22, 1296-1317.

Hinsz, V. B., & Nickel], G. S. (2003). Group shifts in goal decisions and the perceived
benefits of working in groups. Manuscript in preparation.

Hinsz, V. B., Park, E. S., & Sjomeling, M. (2003). Group interaction sustains positive
mood and diminishes negative mood. Manuscript in preparation.

Isen, A. M., Shalker, T. E., Clark, M. S., & Karp, L. (1978). Affect, accessibility of
material in memory, and behavior: A cognitive loop? Journal of Personality and
Social Psychology, 36, 1-12.

Josephs, R. A., Giesler, R. B., & Silvera, D. H. (1994). Judgement [sic] by quantity.
Journal of Experimental Psychology: General, 123, 21-32.

130

Kashy, D. A., & Kenny, D. A. (2000). The analysis of data from dyads and groups. In
H. T. Reis & C. M. Judd (Eds), Handbook of research methods in social
psychology (pp. 451-477). Cambridge, UK: Cambridge University Press.

Larey, T. S., & Paulus, P. B. (1995). Social comparison and goal setting in groups.
Journal of Applied Social Psychology, 25, 1579-1596.

Lucas, R. E., & Fujita, F. (2000). Factors influencing the relation between extraversion
and pleasant affect. Journal of Personality and Social Psychology, 79, 1039-
1056.

Martin, L. L., & Clore, G. L. (2001). Theories of mood and cognition: A user's
guidebook. Mahwah, NJ: Lawrence Erlbaum Associates.

Morris, W. N. (1989). Mood: The frame of mind. New York: Springer-Verlag.

Mullen, B., Johnson, C., & Salas, E. (1991). Productivity loss in brainstorming groups:
A meta-analytic integration. Basic and Applied Social Psychology, 12, 3-23.

N i j stad, B. A. (2000). How the group afiects the mind: Effects of communication in
idea generating groups. Utrecht, Netherlands: ICS dissertation series.

Osborn, A. F. (1957). Applied imagination. New York: Scribner.

Park, E. S., Levine, T. R., Harms, C., & Ferrerra, M. (2002). Group and individual
accuracy in deception detection. Communication Research Reports, 19, 99-106.

Paulus, P. B., Brown, V., & Ortega, A. H. (1999). Group creativity. In R. E. Purser &
A. Montuori (Eds), Social Creativity Volume 2 (pp. 151-176). Cresskill, NJ:
Hampton Press.

Paulus, P. B., Dzindolet, M. T. (1993). Social influence processes in group
brainstorming. Journal of Personality and Social Psychology, 64, 575-586.

Paulus, P. B., Dzindolet, M. T., Poletes, G., & Camacho, L. M. (1993). Perception of
performance in group brainstorming: The illusion of group productivity.
Personality and Social Psychology Bulletin, 19, 78-89.

Paulus, P. B., Larey, T. S., & Ortega, A. H. (1995). Performance and perceptions of

brainstormers in an organizational setting. Basic and Applied Social Psychology,
1 7, 249-265.

Rudman, L. A. (1998). Self-promotion as a risk factor for women: The costs and
benefits of counterstereotypical impression management. Journal of Personality
and Social Psychology, 74, 629-645.

131

Schwarz, N. (1990). Feelings as information: Informational and motivational functions
of affective states. In E. T. Higgins & R. Sorrentino (Eds), Handbook of
motivation and cognition: Foundations of social behavior (Vol. 2, pp. 527-561).
New York: Guilford Press.

Schwarz, N ., & Clore, G. L. (1983). Mood, misattribution and judgments of well-being:
Informative and directive functions of affective states. Journal of Personality and
Social Psychology, 45, 513-523.

Schwarz, N., & Clore, G. L. (1988). How do I feel about it? Informative functions of
affective states. In K. Fiedler & J. Forgas (Eds), Affect, cognition, and social
behavior (pp.44-62). Toronto: Hogrefe International.

Schwarz, N., & Clore, G. L. (1996). Feelings and phenomenal experiences. In E. T.
Higgins & A. Kruglanski (Eds), Social psychology: A handbook of basic
principles. New York: Guilford Press.

Stoppard, J. M., & Gruchy, C. D. G. (1993). Gender, context, and expression of positive
emotion. Personality and Social Psychology Bulletin, 19, 143-150.

Strack, F., Schwarz, N., & Gschneidinger, E. (1985). Happiness and reminiscing: The
role of time perspective, affect, and mode of thinking. Joumal of Personality and
Social Psychology, 49, 1460-1469.

Stroebe, W., Diehl, M., & Abakoumkin, G. (1992) The illusion of group effectivity.
Personality and Social Psychology Bulletin, 18, 643-650.

Watson, D., Clark, L. A., & Tellegen, A. (1985). Toward a consensual structure of
mood. Psychological Bulletin, 98, 219-235.

Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief
measures of positive and negative affect: The PAN AS scales. Journal of

Personality and Social Psychology, 54, 1063-1070.

Zajonc, R. B. (1980a). Compresence. In P. B. Paulus (Ed) Psychology of group
influence (pp. 35-60). Hillsdale, NJ: Erlbaum.

Zajonc, R. B. (1980b). Feeling and thinking: Preferences need no inferences. American
Psychologist, 35, 151-175.

132

   

iiiiiiiijiininiiniti