 

 

 

 

W5 in.

\L. ‘xJ

20;

 

l LlBRARY
I Michigan State
University

 

 

This is to certify that the
dissertation entitled

EFFECTS OF PUT-DOWN HUMOR ON COHESION IN
GROUPS

presented by
CHRISTINE GOCKEL

has been accepted towards fulﬁllment
of the requirements for the

Ph.D. degree in Psychology

 

 

7aszvfp<ﬂk

Major Professor’s Signature
7/20/0 7
i 7

Date

MSU is an afﬁrmative-action, equal-opportunity employer

I
u--—.—-—.—.-n-.—-—-----—--.---—--—-a--.-.-——c-u—-—I---.--.—~—-—~—-'-—---_u-g-a-.-o--o--—.------—

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/C|RC/DateDue.indd-p.1

 

 

EFFECTS OF PUT-DOWN HUMOR ON COHESION IN GROUPS
By

Christine Gockel

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Psychology

2007

ABSTRACT
EFFECTS OF PUT-DOWN HUMOR ON COHESION IN GROUPS
By
Christine Gockel

Put-down humor is a type of humor that may have a particularly strong impact on
socio-emotional group processes. It always has a target, and the intent of the speaker is to
make fun of someone or something. Therefore, put-down humor always implies
something about the speaker’s attitude or relationship with a person or object.
Observational studies suggest that put-down humor can affect cohesion in a group, and it
is assumed to increase cohesion when it refers to outgroup members (as compared to
‘ingroup members).

Two lab studies were conducted to examine the underlying processes and
moderators of the put-down humor — cohesion relation. Participants were led to believe
that they were able to interact with other group members via headsets. In fact, the
contributions of other ostensive group members had been recorded to manipulate the
occurrence of put-down humor.

In Study 1, I found that while put-down humor did not increase participants’
attraction to other group members, it did increase the perception of cohesion among ghg
group members. Participants in the Put-down Humor condition regarded the other group
members to be a cohesive unit and perceived them to be in a more positive mood, to feel
more psychological safety, and to be more similar to each other than did participants in
the No Humor condition. Perceived positive mood and perceived similarity among other

group members mediated the relation between put-down humor and cohesion.

In Study 2, changes in the procedure helped participants to feel safer in the group
to begin with. But even under those circumstances, the pattern of results remained the
same: When hearing put-down humor, participants were not more attracted to the group
than when not hearing any humor. However, they perceived the other group members to
be a more cohesive unit.

Overall, these results suggest that put-down humor can create positive
impressions of other group members without increasing one’s own attraction to the
group. I discuss when and why put-down humor might also increase one’s attraction to

the group.

For Sebastian

iv

ACKNOWLEDGEMENTS

I would like to thank the following people who supported my work on this dissertation:
My parents Annette and Norbert for letting me cross the ocean, my advisor Norbert Kerr
for always taking me seriously during many humorous exchanges about the topic and for
being an invaluable source of inspiration, Gwen Wittenbaum for helping me ﬁnd my
passion, Rick DeShon and Deborah Kashy for instilling a love for statistics in me,
Elisabeth Brauner for making me believe that I could write a dissertation, Dan Gi gone for
introducing me to research about small groups, my fabulous undergraduate research
assistants Erin Berard, Kelli Craft, Daniel Facer, Aaron Holland, Vanessa Johnson, Ken
Kettler, and Melissa Visconti for their help with data collection, Giuseppe Lorusso for
numerous discussions about bootstrapping, Kamila Rosolova for early mornings in the
library, my fellow runners Stefano, Francesco, Aaron, Monica, Alfredo, Fernando, and
Michelle for providing me with reasons to leave my ofﬁce, and last but not least
Sebastian Schnettler for being the best husband on earth and for crossing the ocean with

me.

TABLE OF CONTENTS

LIST OF TABLES .................................................................................. vii
LIST OF FIGURES ................................................................................. ix
INTRODUCTION ........................................ ' ........................................... 1
STUDY 1 ............................................................................................. 18
Method ....................................................................................... 19
Results ....................................................................................... 33
Discussion ................................................................................... 58
STUDY 2 ............................................................................................. 62
Method ....................................................................................... 64
Results ....................................................................................... 73
Discussion ................................................................................... 94
GENERAL DISCUSSION ........................................................................ 98
APPENDICES ...................................................................................... 1 11
Appendix A: Student Proﬁles ........................................................... 111
Appendix B: Transcripts of Recordings ............................................... 115
Appendix C: Questionnaires ............................................................. 1 17
BIBLIOGRAPHY .................................................................................. 123

LIST OF TABLES

Table 1: Condition Means and Standard Deviations for the Manipulation Check
Items ................................................................................................... 34

Table 2: Summary of Items and Factor Loadings for the Four Selected Items from the
Team Psychological Safety Scale ................................................................ 37

Table 3: Summary of Items and Factor Loadings for the Perceived Intentions Items. . ...38
Table 4: Pearson Correlation Coefﬁcients for the Perceived Intentions Measures. . . . . ....39

Table 5: Summary of Items and Factor Loadings from the Group Environment
Questionnaire (Measuring Experienced Personal Cohesion) ................................. 42

Table 6: Pearson Correlation Coefﬁcients for the Experienced Personal Cohesion
Measures ............................................................................................. 44

Table 7: Pearson Correlation Coefﬁcients for the Perceived Group Cohesion
Measures ............................................................................................. 45

Table 8: Pearson Correlation Coefﬁcients between Measures of Own Thoughts and
Feelings and Measures of Others’ Thoughts and Feelings .................................... 47

Table 9: Means, Standard Deviations, and T-tests for the Experienced Personal
Cohesion Measures ................................................................................. 48

Table 10: Means, Standard Deviations, and T-Tests for the Perceived Group
Cohesion Measures ................................................................................. 49

Table 11: Means, Standard Deviations, and T-Tests for the Measures of Participants’
Own Thoughts and Feelings ....................................................................... 52

Table 12: Means, Standard Deviations, and T-Tests for Measures of Other Group
Members’ Thoughts and Feelings ................................................................. 53

Table 13: Mediation of the Effect of Put-down Humor on Perceived Group Cohesion
Through Perceived Positive Mood, Perceived Shared Experiences, Perceived

Psychological Safety, and Perceived Similarity ................................................. 56

Table 14: Condition Means and Standard Deviations for the Manipulation Check
Items ................................................................................................... 74

vii

Table 15: Summary of Items and Factor Loadings for the Items from the Group
Identiﬁcation Scale ................................................................................. 76

Table 16: Pearson Correlation Coefﬁcients for the Experienced Personal Cohesion
Measures ............................................................................................. 78

Table 17: Pearson Correlation Coefﬁcients for the Perceived Group Cohesion
Measures ............................................................................................. 79

Table 18: Pearson Correlation Coefﬁcients between Measures of Own Thoughts and
Feelings and Measures of Others’ Thoughts and Feelings .................................... 80

Table 19: Means, Standard Deviations, and T-tests for the Experienced Personal
Cohesion Measures ................................................................................. 82

Table 20: Means, Standard Deviations, and T-Tests for the Perceived Group Cohesion
Measures ............................................................................................. 83

Table 21: Means, Standard Deviations, and T-Tests for the Measures of Participants’
Own Thoughts and Feelings ....................................................................... 84

Table 22: Means, Standard Deviations, and T-Tests for Measures of Other Group
Members’ Thoughts and Feelings ................................................................. 86

Table 23: Multivariate and Univariate Analyses of Variance for Measures of Own
Thoughts and Feelings .............................................................................. 88

Table 24: Multivariate and Univariate Analyses of Variance for Measures of
Experienced Personal Cohesion ................................................................... 90

Table 25: Mediation of the Effect of Put-down Humor on Perceived Group Cohesion
through Perceived Positive Mood, Perceived Shared Experiences, Perceived
Psychological Safety, and Perceived Similarity ................................................. 92

viii

LIST OF FIGURES

Figure 1: Example of a Student Proﬁle Shown to Participants ........................................ 22

Figure 2: Mediation of Put-down Humor on Perceived Group Cohesion by Perceived
Positive Mood, Perceived Shared Experiences, Perceived Psychological Safety, and
Perceived Similarity. Displayed are unstandardized coefﬁcients (and their standard
errors) ................................................................................................. 57

Figure 3: Proﬁles of Five Other Ostensible Group Members as Background
Information for the Recorder Vote ................................................................ 67

Figure 4: Table for Recorder in Study 2 .......................................................... 68
Figure 5: Mediation of Put-down Humor on Perceived Group Cohesion by Perceived
Positive Mood, Perceived Shared Experiences, Perceived Psychological Safety, and

Perceived Similarity. Displayed are unstandardized coefﬁcients (and their standard
errors) ................................................................................................. 93

ix

INTRODUCTION

Humor seems to occur in every group to different degrees. It is prevalent in many
different social contexts (Foot, 1991), including groups. Wyer and Collins (1992) point
out that in almost all conversations at least one person tries to elicit laughter or responds
to something with amusement.

Humor has a major inﬂuence on the quality of an interaction (Wyer & Collins,
1992). In groups, it is alleged to lead to higher camaraderie and stronger bonds among
group members, to better teamwork, more creativity and innovations, and better problem
solving (Hudson, 2001; Stauffer, 1999). One type of humor with a particularly strong
impact on group processes is put-down humor. Such humor always has a target, and the
intent of the speaker is to make fun of someone or something (Terrion & Ashforth, 2002).
Through put-down humor, speakers deﬁne their relationship with a target. It is this
inherently social nature of put-down humor that makes it more likely to inﬂuence group
processes than other forms of humor. In these two studies, I examined how put-down
humor affects cohesion in groups.
Theories of Humor

To date, there is no clear or universally accepted deﬁnition of humor, although
there are two broad classes of humor theories. Motivational theories focus more on the
emotional aspect of humor, whereas incongruity resolution theories focus more on the
cognitive aspect.

The ﬁrst interesting subset of motivational theories are the arousal and arousal
reduction theories (Wyer & Collins, 1992). According to these theories, humor responses

reﬂect a reduction in or release of arousal. Freud (1928, 1960), for example, proposed

that humorous reactions are solely motivated by the need to release tension or arousal that
was sex- or aggression-related and that could not be expressed openly due to societal
norms. Berlyne (1969, 1971), on the other hand, described a more general theory of
arousal and arousal reduction. According to him, hearing a joke induces arousal above
and beyond the optimal level of pleasure. The punch line then reduces that arousal to
more pleasurable levels.

The second subset of motivational theories are the superiority or disparagement
theories (La Fave, Haddad, & Maeson, 1976; Levine, 1969; Zillman & Cantor, 1976).
According to these theories, humor is experienced when one perceives another person’s
shortcomings and fulﬁlls the goal of establishing or maintaining feelings of mastery or
control. These feelings, in turn, are pleasurable.

In contrast to these varied motivational theories, the incongruity resolution
theories explain humor as a cognitive process. In general, humor is experienced by the
sudden awareness of an incongruity between two objects or events. Suls (1972, 1983)
suggested that when hearing a joke, the listener activates certain schemata to interpret the
incoming information and builds expectations. New information does not conﬁrm the
expectations, but instead, the listener needs to refer to other schemata to be able to
interpret the new information. The reinterpretation then elicits humor. Thus, the
experience of humor involves two stages: realizing the incongruity and understanding it.
Types of Humor

Long and Graesser (1988) described several types of humor. In general, they
distinguish between jokes and witticism. Jokes are context-ﬂee humor and contain all

information that is necessary to understand the content. Witticisms, on the other hand, are

context-speciﬁc. In order to understand them, the listener must understand the topic of
the conversation. Examples of witticism are put-down humor, teasing, and ironic
comments. In the case of all Witticisms, a “play frame” is important, which indicates that
the comment is not meant to be offensive (Foot, 1991). A play frame is indicated by the
use of cues such as disclaimers or contextualization cues (like winks, laughs, or
exaggerated expressions) (Eisenberg, 1986). In other words, a play frame functions like a
meta-message that shows that the actual content is not to be taken literally. A play frame
can sometimes be ambiguous, though. Listeners might not pick up on the cues. Thus, if a
play frame is not clear enough, the comments can be regarded as hostile. Speakers
usually convey both jokes and witticism with the aim of eliciting humorous reactions
such as smiles and laughter in listeners, but they might also have less obvious goals like
creating stronger social bonds among their listeners.

In this study, I will focus speciﬁcally on put-down humor and its effect on the
group’s cohesion. Put-down humor is the “attempt to derive amusement at the expense of
something or someone” (Terrion & Ashforth, 2002, p. 59) — that is, there is always a
target or “butt” of such humor. In groups, put-down humor should generally have
stronger effects on group members and group processes than context-free jokes due to its
inherently social nature. Put-down humor implies something about the speaker’s attitude
toward the people or objects and how the speaker relates to them. In other words, through
such humor, the speaker self-discloses about him- or herself and deﬁnes his or her
relationship to other people or objects.

For a number of reasons, such as its usual spontaneity, its ephemeral nature, and

its contextuality, put-down humor in groups is extremely difﬁcult to study. Some

researchers (e.g., Terrion & Ashforth, 2002) have used naturalistic, ethnographic studies,
usually with small samples, and focused on idiosyncratic interactions that can only be
generalized with difﬁculty. Others (e. g., Tragesser & Lippman, 2005) have used self-
report methods. However, put-down humor may be difﬁcult to describe after it has
occurred, or participants might only focus on very extreme forms or instances of put-
down humor in their retrospective responses. To my knowledge, no experimental or lab
studies have been conducted examining the effects of put-down humor on group
processes. The present studies will use experimental participants who will act as both
observers of and participants in group interaction. This methodology permits me to
achieve some degree of experimental control while gaining insight into processes that
affect actual interactions which cannot easily be stopped or controlled. I also believe that
this method is valid in its own right because people heavily rely on their observations to
form impressions of individuals and groups (of, Aﬁﬁ & Burgoon, 2000).
Functions of Humor in Groups

Humor can serve several functions in groups. In general, when making humorous
remarks, people can say what they mean, the opposite of what they mean, or both. In
other words, humor can be both aggressive and supportive (Norrick, 2003). Speakers can
retract their humorous message any time it appears to be too offensive (“Hey, I was only
joking!”).

By using put-down humor, speakers can both present information about
themselves and seek information from other group members (of, Foot, 1991). Speakers
can share feelings or intentions that they might ordinarily not reveal publicly. For

example, they can express their liking or disliking of another group member by making

fun of him or her. At the same time, they can ﬁnd out if other group members share their
feelings or intentions by having introduced a new topic humorously and by observing the
others’ reactions.

Put-down humor can help to control the interactions in the group. The speaker can
attract other group members’ attention to him- or herself and might thus gain their
approval. Alternatively, the speaker can also subtly express the opinion that one group
members’ behavior was not in line with the group requirements and might thereby
change the behavior of the person who got off track.

Furthermore, several scholars have suggested that humor (potentially including
put-down humor) can directly serve the purpose of enhancing group identiﬁcation and
cohesion. Vinton (1989), for example, showed in a naturalistic study that in a small
organization of 13 employees teasing (which is similar to put-down humor, but always
directed at someone present in the conversation) served two purposes: First, it helped to
get the necessary work done. Teasing allowed higher-status employees to order their
subordinates in a pleasant way to ﬁnish a task. Second, teasing lessened status
differentials between employees by putting everyone on the same plane, eventually
improving cooperation. Vinton (1989) concluded that humor was used to create bonds
and facilitated the socialization of newcomers. Similarly, Duncan and Feisal (1989)
stated that at every workplace they examined, the most important reason why joking
behavior occurred was that it makes people feel they belong. However, these authors do
not offer any explanations for why this might be the case. The two studies, though, give

us a ﬁrst hint that put-down humor can inﬂuence cohesion in groups.

More support for the put-down humor — cohesion relation comes from Terrion
and Ashforth (2002). They observed as participants a leadership training program in a
Canadian police college, focusing on the development of put-down humor in this
temporary group. The researchers found that put-downs followed a certain temporal
pattern: Group members would ﬁrst put down themselves, then shared social identities
(i.e. the group as a whole), then external groups, and ﬁnally other group members.
Terrion and Ashforth concluded that this developmental sequence signaled increased trust
and inclusion in the group and helped create a common identity and cohesion.

Two other studies also indicate that humor is related to cohesion in groups — but
they do not examine put-down humor speciﬁcally. Fraley and Aron (2004) conducted a
lab experiment and had same-sex strangers participate in a series of interactions that
created a humorous experience (e. g., by asking participants to act out a television
commercial of their choice in a language made up by them) or did not create a humorous
experience (e. g., by asking participants to act out a television commercial in English).
Participants who shared a humorous experience felt closer to each other after the
interactions than did participants who had interacted without any humor. Robinson and
Smith-Lovin (2001) used event history techniques to analyze recorded lab group
discussions. Based on their results, the authors developed a theory of humor and status in
groups. This theory predicts, among other things, who will produce humor, who will be
successful in producing humor, or how producing humor changes the speaker’s
subsequent interactions within the group. More important for this study, though, is that
the theory also provides explanations of how and under which circumstances humor can

increase or undermine cohesion in a group. Robinson & Smith-Lovin (2001) hypothesize

that humor can strengthen affective ties within the group, i.e. increase cohesion, when the
humorous comment refers either to the group as a whole or to outsiders. They
hypothesize further that humor can undermine cohesion when the comment refers to the
self or another group member. This is an important distinction, which will be discussed in
more detail below.

Group Cohesion as an Effect of Put-down Humor

Before explaining when and why put-down humor increases cohesion in a group,
the concept of cohesion needs to be deﬁned clearly. At this point, there is no universal
deﬁnition of cohesion that all researchers agree upon (Dion, 2000). However, most
researchers agree on the fact that cohesion has two distinct aspects: a social and a task
aspect.

The social dimension of cohesion can be deﬁned as “the emotional bonds of
friendship, liking, caring, and closeness among group members” (MacCoun, 1993, p.
291). Some of the ﬁrst researchers to examine cohesion were Kurt Lewin and Leon
Festinger. Festinger and his colleagues (Festinger, Schachter, & Back, 1950) wrote about
cohesion as “the total ﬁeld of forces which act on members to remain in the group” (p.
164). This ﬁeld of forces can also be described as attraction between group members
(personal attraction) and attraction of the individual to the group as a whole (social
attraction) (Forsyth, 2006). Besides attraction, social cohesion can also be described as
perceived unity. When the group is cohesive, members report a sense of belonging to the
group, feel included in the group, and report that the group is a part of who they are
(Bollen & Hoyle, 1990; Forsyth, 2006). Furthermore, cohesive groups “stick together”,

have high levels of solidarity, and are uniﬁed (F orsyth, 2006).

The task dimension of cohesion, on the other hand, can be deﬁned as “the shared
commitment among members to achieving a goal that requires the collective efforts of
the group” (MacCoun, 1993, p. 291). Forsyth (2006) adds that task cohesion can also be
regarded as collective efﬁcacy and team spirit. Task cohesion, in general, should not be
as strongly inﬂuenced by put-down humor as social cohesion because of the social nature
of put-down humor. Thus, the focus of this study will be mainly on social cohesion.

Cohesion, no matter if social or task, can always be measured at two levels: at the
individual and at the group level. If cohesion is assessed at the individual level, one group
member can be asked, for example, how much he or she is attracted to speciﬁc other
group members or to the group as a whole, or how strong a sense of belonging he or she
experiences in the group. All of these questions refer to an individual’s personal
perception or experience. If cohesion is assessed at the group level, some collective
assessment is required. For example, the mean of group members’ ratings of attraction to
each other or to the group as a whole could be computed.

As outlined above, these studies will employ individual experimental participants
serving sequentially as both observers of and participants in a group interaction. Because
the behaviors of all the other group members other than the individual participant will be
created and controlled by the experimenter, I will only be able to ask one group member,
namely the participant, about cohesion. Thus, I will not be able to collect all group
members’ assessments of cohesion and aggregate them to deﬁne a group-level measure
of cohesion. I will, however, collect two cohesion judgments from the participants: First,
I will ask the participant how much he or she is attracted to the group (experienced

personal cohesion). Second, I will ask the participant how strongly he or she perceives

other group members, excluding the self, to be attracted to each other (perceived group
cohesion). Below, when I refer to a perceived group-level construct, it should be
understood that I do so in the latter sense, i.e. as a judgment by one member of the group,
and not as a summary of all group members’ judgments about the group.

When Does Put-down Humor Increase Cohesion?

Put-down humor will not increase cohesion in every situation. The goal of the
present two studies was to create conditions under which put-down humor might increase
cohesion and to examine the underlying process in more detail. I suggest that two
conditions are likely to be necessary for put-down humor to have positive effects on
cohesion: First, the target of the humorous comment needs to be someone outside the
group. Second, other group members need to react positively after the comment.

The target of humor. The necessary condition is that the target of the put-down
humor, in other words the person who is the focus of the humorous comment, the “butt”
of the joke, is someone outside the group. As mentioned above, Robinson & Smith-Lovin
(2001) hypothesized that cohesion increases when the humorous comment refers to
outsiders or to the group as a whole, but decreases when the comment refers to other
group members or to the self.

Boxer and Cortés-Conde (1997) suggest that an important part of every humorous
interaction is the “development of a relational identity among participants which leads to
a sense of membership in a group” (p. 276). Similar to Robinson and Smith-Lovin
(2001), they also mention that outcomes of a humorous interaction can lie anywhere from
bonding to “biting”, which is functionally similar to creating an increase or decrease in

cohesion. According to Boxer and Cortés-Conde (1997), the identity of the target of

joking can determine the outcome: A speaker can direct his or her comment at a
participant in the conversation, which has the potential of “biting”, or the speaker can
direct the comment at an absent other, which has the potential of bonding. Sometimes, in
very intimate relationships or very close groups, though, comments that are directed at
participants in the conversation can also serve the purpose of bonding (Terrion &
Ashforth, 2002).

Joking about an absent other is similar to put-down humor about an outgroup
member. According to Boxer and Cortés-Conde (1997), this kind of humor is not very
ambiguous. Speaker and listener both recognize the target as an outsider. Because they
have a common outgroup and are members of the same ingroup, they bond. Therefore,
put-down humor should increase cohesion in the group more when the target is an
outgroup member than when it is an ingroup member (at least for relatively new groups,
as in the present study).

Group members ’ reactions. The second necessary condition for put-down humor
to increase cohesion is that group members’ reactions after the humorous comment are
positive. The strongest sign of approval or disapproval of a joke or humorous comment is
group members’ laughter or lack of laughter. According to Norrick (2003), laughter
demonstrates both understanding and appreciation of the humorous comment. The lack of
laughter, or silence, is ambiguous because listeners either did not get the joke and/or do
not appreciate it. Thus, laughter signals appreciation, but lack of laughter may signal a
lack of appreciation. Terrion and Ashforth (2002) found in their study that the meaning of
certain humorous comments was sometimes ambivalent, but that the uniform reaction of

every group member, the laughing, made the comment seem benevolent. Thus, the

10

relation between put-down humor and cohesion in the group should be more positive
with than without a laughter reaction.

We can ﬁnd out more about the social functions of laughter by examining theories
of the evolution of laughter. Hayworth (1928) hypothesized that laughter signals to other
group members that the environment is safe and that they can relax. Ramachandran
(1998) also theorizes that a sudden change in expectations leads to laughter in one
person. This reaction informs conspeciﬁcs that expectations about lurking danger were
wrong and that the environment is indeed safe.

Thus, put-down humor about outgroup members followed by laughter should
increase cohesion. Laughter signals the group’s appreciation of the comment, that the
speaker had positive intentions, and that the situation is safe.

Hypothesis 1: Basic effects

Hypothesis 1.1: Put—down humor of outgroup members followed by
laughter increases experienced personal cohesion.
Hypothesis 1.2: Put—down humor of outgroup members followed by
laughter increases perceived group cohesion.

Wthoes Put-down Humor Inﬂuence Cohesion? The Underlying Processes

Several mediators might help explain the relation between put-down humor and
cohesion. These include mood, shared experiences, psychological safety, the perceived
intentions of the speaker, and the perceived similarity of group members. In the following
section, I will explain how each mediator might be inﬂuenced by put-down humor and
how it, in turn, might inﬂuence cohesion. Similar to the assessment of cohesion, I will

ask participants ﬁrst about their own thoughts and feelings and second about the thoughts

11

and feelings of the other group members. For example, I will ask participants about their
own mood and their perception of the mood of the other group members. Mediators
regarding the individual participant should predict experienced personal cohesion; and
mediators regarding the other group members should predict perceived group cohesion.
For example, an individual’s mood should predict how strongly he or she is attracted to
the group (= experienced personal cohesion); and the perceived mood of other group
members should predict how strongly they seem to be attracted to each other (= perceived
group cohesion).

Mood. Humor, in general, increases positive mood (e.g., Martin, 2001 , 2002) and
decreases negative mood in people. More speciﬁcally, research has shown that humor
increases hope and energy — indicators of positive mood — and decreases anxiety, tension,
and stress — indicators of negative mood (cf., Berk, 2001). Some even regard humor as
interpersonal emotion management (Francis, 1999). It is still an unanswered question
whether put-down humor in a group also has these beneﬁcial effects on group members’
mood. Put-down humor should theoretically increase positive mood and decrease
negative mood if the humor is not regarded as threatening to the self.

It has been shown that people in a positive mood form more positive impressions
of others (Forgas & Bower, 1987), like others better, and are even more willing to work
with them (Gouaux, 1971; Veitch & Grifﬁt, 1976). All these effects could be regarded as
indicators of cohesion. Thus, there is some evidence to suggest that put-down humor
changes mood, which in turn affects the cohesion in the group.

Put-down humor can not only change the individual’s own mood, it can also

change how the individual perceives and interprets the mood of his or her fellow group

12

members. If someone makes a put-down humorous comment about somebody else
outside the group, the individual’s own mood might become more positive and less
negative, and he or she might also perceive the mood of the fellow group members to
become more positive and less negative. This could in turn affect how cohesive the
individual regards the group to be.

Hypotheses 2: Mediation by mood

Hypothesis 2.1: The basic effect 1.1 is partially mediated by increases in
own positive mood.

Hypothesis 2.2: The basic effect 1.1 is partially mediated by decreases in
own negative mood.

Hypothesis 2.3: The basic effect 1.2 is partially mediated by increases in
the perceived positive mood of other group members.

Hypothesis 2.4: The basic effect 1.2 is partially mediated by decreases in
the perceived negative mood of other group members.

Shared experiences. Coser (1959, p. 173) regards humor as the “expression of the
collective experience of the participants”. Humor helps to translate individual
experiences into collective group experiences (Pogebrin & Poole, 1988). In other words,
through humor various issues, some of them important to the group, can be brought to the
attention of all group members and can be tackled. Alternatively, group members could
also realize through a humorous comment that they have the same perception of a
situation and share the same feelings and thoughts. Furthermore, if all group members
start laughing loudly alter a humorous comment, they show the same behavior, which

might make them come to the conclusion that they share the same feelings and thoughts

l3

on the issue. This realization could make them feel closer to each other and increase
cohesion in the group. Eventually, shared experiences facilitate group interactions and
help build a group culture (Fine & DeSoucey, 2005).

Schachter’s classic research on afﬁliation (1959) supports the claim that shared
experiences can lead to an increase in cohesion. Schachter (1959) showed in his well-
known studies that anxious people preferred to be in the company of similarly anxious
others and that one plausible and well-supported reason for this was the individual’s
desire to receive comfort and support from another person. Schachter hypothesized that
the presence of the other person who is in the same situation could have a direct anxiety-
reducing effect and reported experimental evidence supporting this hypothesis (e.g.,
Wrightsman, 1959). Other researchers working with animals (e.g., Carbajal and Orihuela,
2001; Westenbroek, Snijders, den Boer, Gerrits, Fokkema, and Ter Horst, 2005) also
pointed out that the shared presence of a same-species individual has stress-reducing
effects. This research might show that shared experiences can lead to an increase in
comfort, closeness, and eventually cohesion.

Hypotheses 3: Mediation by shared experiences

Hypothesis 3.1: The basic effect 1.1 is partially mediated by increases in
shared experiences with other group members.

Hypothesis 3.2: The basic effect 1.2 is partially mediated by an increase in
perceived shared experiences among other group members.

Psychological safety. Psychological safety is the belief that it is safe to take
interpersonal risks in the group (Edmondson, 1999). It is “a sense of conﬁdence that the

team will not embarrass, reject, or punish someone for speaking up” (p. 354). Put-down

l4

humor is risky — especially in newly formed groups. However, if group members start
making funny comments exclusively about people o_ut_s_ig§ the group, group members’
psychological safety should increase. They should feel more comfortable and believe that
they can also take some risks in the group when they hear put-down humor about
outgroup members. Furthermore, put-down humor about outgroup members might
distract them from their initial feelings of discomfort in the new group (F raley & Aron,
2004) because it shifts their focus from the current interaction to others outside the group
who, for some reason, deserve to be made ﬁrn of.

The laughter that accompanies put-down humor also signals that the situation is
not too dangerous or serious. According to one hypothesis about the evolution of laughter
(Hayworth, 1928), laughter is a signal to other group members that the environment is
safe. Thus, group members who witness amusement or laughter in others should perceive
their environment as safer and as less serious than at ﬁrst sight.

Psychological safety should increase cohesion because it alleviates concerns
about the impression that one has to create and about other’s reactions to one’s own
statements. It is not the same as cohesion, though, because cohesion can reduce one’s
willingness to disagree with others (Janis, 1982), but psychological safety does not
necessarily have the same effect.

Hypotheses 4: Mediation by psychological safety

Hypothesis 4.1: The basic effect 1.1 is partially mediated by increases in
psychological safety.
Hypothesis 4.2: The basic effect 1.2 is partially mediated by increases in

perceived psychological safety among other group members.

15

Perceived intentions of the speaker. A humorous comment usually catches
people’s attention. Humor is inherently targeted at other people in the interaction
(Francis, 1999). No one makes a joke and laughs loud when being alone (except perhaps
in anticipation of sharing it with others). Joking and laughing alone could even be
regarded as pathological. Humorous comments might therefore make listeners more
likely to form an attribution of the speaker’s intentions. Unless the put-down humor is
patently hostile, not presented within a play frame, and directed at group members,
hearing such humor should generally lead listeners to conclude that the speaker had
positive and not negative intentions. For example, she or he might have wanted to create
a pleasant atmosphere in the group or to make others laugh. Overall, the humorous
comment should lead listeners to conclude that the speaker cares about the feelings of
group members and regards the quality of group interactions as important. These positive
perceptions of the speaker’s intentions should in turn lead to higher cohesion in the
group.

Hypotheses 5: Mediation by perceived benevolent intentions of the speakers
Hypothesis 5.1: The basic effect 1.1 is partially mediated by an increase in
perceived benevolent intentions of the speakers.

Hypothesis 5.2: The basic effect 1.2 is partially mediated by an increase in
perceived benevolent intentions of the speakers.

Perceived similarity. Usually, listeners show a similar reaction after a humorous

comment. They might all laugh heartily, chuckle, or even show their disapproval through
being silent. Realizing that other group members show the same reaction as the individual

does should lead him or her to conclude that the others are similar to him or her, at least

16

in certain respects (e. g., their “senses of humor”). This similarity should, in turn, increase
attraction to the other group members (Byme, 1971) and increase cohesion.
Hypotheses 6: Mediation by perceived similarity
Hypothesis 6.1: The basic effect 1.1 is partially mediated by an increase in
perceived similarity to other group members.
Hypothesis 6.2: The basic effect 1.2 is partially mediated by an increase in

perceived similarity among other group members.

17

STUDY 1

In a lab experiment, participants were told that their fellow group members were
sitting in adjacent rooms and that all group members interacted via headsets. Participants
then listened to a highly-structured group discussion that had been pre-recorded to
manipulate the amount of put—down humor. They believed that the group discussion was
taking place naturally while they were listening and that they would later also contribute
to it. The discussion content in the two experimental conditions (Put-down Humor and
No Humor) was the same — except that when put-down humor was present, some
additional discussion content was presented in a joking manner. After listening to the ﬁrst
part of the discussion, participants answered several questionnaires and then continued

the discussion.

18

METHOD

261ng

In Study 1, I sought to obtain a reliable relationship between put-down humor of
outgroup members and cohesion and to examine the mediational pathway between these
two variables. More speciﬁcally, I tested whether any of the variables mood, shared
experiences, psychological safety, intentions of the speaker, and perceived similarity
mediated the hypothesized relation between put-down humor of outgroup members and
cohesion. The design of the study was a single-factor design: Put-down humor of
outgroup members (followed by common laughter) was present or absent in the group
discussion. The discussion content in these two conditions was the same — except that
when put-down humor was present some additional discussion content was presented in a
joking manner. In order to bolster the relation between put-down humor of outgroup
members and cohesion, group members’ reactions to the humor were always positive. In

other words, group members always showed their approval of the humorous comments
through laughter.
Participants

One hundred and eleven female college students participated in the study. Due to
technical failures, data from eight participants were not recorded. Out of the remaining
103 participants, seven (=6.8%) reported in the post-experimental suspicion check that
their fellow group members’ contributions had been recorded; their data were excluded
ﬁ'om all analyses. Thus, data from 96 participants were analyzed. Fifty-three participants
had been randomly assigned to the No Humor condition and 43 had been randomly

assigned to the Put-down Humor condition.

19

The average age of participants was 19.23 years (SD = 1.33, Min. = 18, Max. =
25). In terms of class rank, 33% were ﬁrst-year students, 41% sophomores, 18% juniors,
and 8% seniors. All participants were recruited through the MSU HPR Pool and received
one hour of credit for taking part in the study.

The gender of the participants and of the fellow group members (whose
contributions had been recorded) was always female. Previous research (e.g., Robinson &
Smith-Lovin, 2001) has shown that there are gender differences in reactions to humor and
in humor production. Because the MSU HPR Pool consists mainly of female students, I
only invited females to participate in this study.

A power analysis showed that 51 participants per condition were needed in order
to have an 80% chance of correctly rejecting the null hypothesis if the effect was
moderate (d = .5). Power was estimated with the computer program GPOWER (Faul &
Erdfelder, 1992) for studies using a one-sided t-test for independent means and testing
hypotheses at the .05 signiﬁcance level. Therefore, the number of participants in this
study was sufﬁcient to test the hypotheses.

Cover StoLy

 

The experimenter explained to participants that we were examining how groups
work together when their members can only communicate via headsets. Each participant
had to sit in a room by herself and was told that the other group members were likewise
in different rooms in the Psychology building. Participants were also told that a group
consisted of six members and that there were ﬁve more people participating in the study
at the same time besides her. All group members would be connected with headsets that

are attached to computers. Thus, the participant would be able to hear what the other

20

group members were saying through her headphone and to transmit messages to the
others by talking into her microphone. All instructions would be shown in detail on the
computer screen. In reality, there was only one actual participant per group (although up
to four participants were run simultaneously). All verbal statements by the other
ostensible group members had been recorded in advance in order to have control over the
discussion content.
The Task

The experimenter explained that this study was also about how groups form
impressions of people. The group’s task was therefore to examine student proﬁles and to
estimate where each student stood on ﬁve basic personality dimensions (viz. how
outgoing, ﬁiendly, reliable, moody, and open—minded the student was). Group members
would be able to see a student proﬁle consisting of a picture and supplemental
information (e. g., academic major, employment, hobbies, and an interesting fact) on their
computer screens (for an example, see Figure 1) Every group member would see the
same picture at the same time. After every member had a chance to examine the ﬁrst
student proﬁle closely in private, one person would publicly give her impression of the
student. The group member should ﬁrst brieﬂy describe the student in the proﬁle (e.g., “I
have Megan here. She is 20 years old and a sophomore”). Then, she would judge how
outgoing the student in the proﬁle was and use a four-point rating scale of 1(well below
average) to 2 (below average) to 3 (above average) to 4 (well above average) for her
judgment (e. g., “Well, Megan seems to be pretty outgoing because she is in a sorority. So
I’d give her a 3, above average, for outgoing”). Aﬁer the ﬁrst group member had

evaluated a student proﬁle, every group member would be provided with a second proﬁle

21

on her computer screen to examine privately, and a second group member would then

publicly give her impression of the student. Thus, the evaluation process would not be a

spontaneous discussion in which group members would be able to exchange their

impressions about the students in the proﬁles. It would rather be a highly structured

sequence of statements. The high amount of structure enabled me to manipulate the

content of the statements without (most) participants noticing that they were not

interacting with other group members but listening to recordings.

 

 

First name:
Age:
Major:
C_la_s_sz

Work:
Clubs/Memberships:

 

Favorite movie or TV show:

Hobbies:
Hometown:
Number of siblings:
Current residence:
Career goals:

Interesting fact:
Heroes:

 

Megan

20
psychology
sophomore

none
Gamma Phi Beta Sorority and Phi Beta Delta Honor Society
Clueless, Sex in the City, Desperate Housewives, Orange County
hanging out with friends, playing softball

Hastings, MI

2

I live in a sorority house.

go to grad school

I made the Dean’s List every semester so far.
my grandfather, my “sisters”, Ms Piggy

 

Figure 1: Example of a Student Proﬁle Shown to Participants

22

 

Procedure

For every session, up to four participants could sign up. Immediately at their
arrival, the experimenter led them into separate and adjacent lab rooms, asked them to
read and sign a consent form, and instructed them individually.

The experimenter told each participant that we were examining how impressions
are formed in groups when members can only communicate via headsets. Six group
members would sit in separate rooms throughout the psychology building and be
connected via headsets that were connected to computers. All instructions would be given
in detail on the computer screen, but if the participant had a question, she could ask the
experimenter any time.

The experimenter also brieﬂy explained the task to participants: “Each member of
your group will be given a student proﬁle to evaluate over the microphone, one after the
other. The evaluation process is not so much a discussion, as one group member will
have the chance to speak ﬁrst, then a second group member, then a third, and so on. The
order of the speakers will be chosen at random.” At this point, the experimenter presented
several folded paper cards to the participant and asked her to pick one. On the inside of
each card, there was a color mark. Unknown to participants, each color mark was red. So
the participant always picked a card with red color. The experimenter went on: “This is
your color. It will show up as a frame around your student proﬁle when you need to
speak.” Then the experimenter left the room and the participant was led through the rest
of the study by a computer program.

In order to familiarize them with the material in the study, participants were ﬁrst

asked to ﬁll out a short questionnaire about themselves (e. g., what their major was, what

23

kind of work they had). They were told that they needed to answer these questions to
acquaint themselves with the format of the student proﬁles, which they would later see
and evaluate, because the student proﬁles would contain exactly the same information.

After that, participants were told that the evaluation process would start and were
reminded not to talk back and forth while one group member was evaluating a student
proﬁle. Participants were also informed that all student proﬁles were ﬁ'om University of
Michigan students (a relevant outgroup for MSU students). The computer program
indicated that participants would have to wait several seconds to proceed to the ﬁrst
student proﬁle because all group members needed to start examining it at the same time.
After a few seconds, the ﬁrst student proﬁle (see Appendix A for this and all other
student proﬁles) showed up on the screen. The ﬁrst student proﬁle was framed with green
color, which indicated to participants that it was not yet their turn to talk (because they
needed to talk when a red frame showed up). The ﬁrst ostensible group member gave her
impression of the student (for a complete transcript of the statements, see Appendix B).
While participants listened to the group member’s statements, they could always see the
student proﬁle on the computer screen. The second and third student proﬁles (with a
yellow and blue frame, respectively) showed up afterwards and were evaluated by two
more ostensible group members.

After three group members had had the chance to speak, the evaluation process
was interrupted and participants were asked to answer some questions about the group.
At this point, all-mediating and dependent variables (described below) were assessed.

Then participants were informed that the way the evaluation process was

conducted would change. Whereas at the beginning speakers were able to hear other

24

group members’ reactions, after the break speakers would not be able to hear other group
members’ reactions anymore. The fourth student proﬁle with a red frame showed up on
the computer screen. Because participants had drawn a card with red color, it was now
their turn to talk. Participants evaluated the student proﬁle. Then the computer program
indicated that the last steps of the evaluation process, i.e. the last two speakers, had to be
skipped due to lack of time.

Finally, participants were asked to indicate how much they liked all four students
from the previous proﬁles. They were also asked what they believed the purpose of the
study to be, whether anything struck them as unusual or odd, and to answer some
manipulation check items (described in more detail below). They were then thoroughly
debriefed and dismissed.

Manipulation of the Independent Variable

In this study, put-down humor was manipulated by having the ﬁrst three
ostensible group members either use or not use put-down humor in their statements. In
the Put-down Humor condition, each ostensible group member used put-down humor
several times, whereas in the No Put-down Humor condition, no group member used put-
down humor (or any other form of humor). The discussion content in these two
conditions was the same as far as possible, except that in the Put-down Humor condition
group members made some brief humorous comments about the person in the proﬁle and
other group members responded with laughter. For example, one group member said
about a student who belongs to a sorority: “Her favorite shows are Clueless and she looks

a little clueless so she'll ﬁt right in there.” The ﬁrst speaker made four humorous

25

comments, the second speaker ﬁve, and the third speaker three. (The transcripts of all
statements can be found in Appendix B.)
Assessment of Potential Mediatirg Variables

The following mediating variables were assessed: mood, shared experiences,
psychological safety, the intentions of the speakers, and the perceived similarity of group
members. Participants were ﬁrst presented with the items assessing their own thoughts
and feelings and then with the items assessing the perception of their fellow group
members’ thoughts and feelings. The order of the variable assessment is described in
more detail below.

Mood: Participants’ own mood and the perceived mood of the other group
members were assessed with the PANAS (Watson, Clark, & Tellegen, 1988). The
PANAS consists of two scales that measure the primary dimensions of mood: positive
and negative affect. The scales are internally consistent and have been shown to have
good convergent and discriminant validity (Watson, Clark, & Tellegen, 1988). Also, the
short-term instructions (“Indicate to what extent you feel this way right now, that is, at
the present moment”) are sensitive to ﬂuctuations in mood, which were expected to
occur in this experiment.

Participants were presented with 10 words indicating positive mood (e. g.,
interested, active, excited) and 10 words indicating negative mood (e. g., nervous, upset,
hostile). First, they were asked to indicate for each word to what extent the word
described their own feelings and emotions at the current moment to assess their own
mood (1: very slightly or not at all — 5: extremely). After participants had answered all

questions about their own thoughts and feelings (i.e. their shared experiences,

26

psychological safety, etc.), they were asked to indicate to what extent each word from the
PANAS described the feelings and emotions of the average other group member at the
current moment to assess the perceived mood of the other group members.

Shared experiences: Participants were asked to rate three statements about how
much they shared the same experience with the other group members:

“When group members can interact face-to-face, they sometimes feel like they

share the same experiences. In this study, we are interested in ﬁnding out how

much this is the case for teleconferencing groups. Of course, we expect these
feelings to be different than in face-to-face groups, but one can still have some
sense of shared experience even when one is not interacting face-to-face. So
please indicate your agreement with each of the following statements:

0 I feel like I have shared some experience with the other group members.

0 I feel like we have started to have some shared experience together.

0 The discussion so far has created a shared experience for us.”

Participants were asked to indicate their agreement with each of these statements on a
scale from 1 (not at all) to 7 (very much).

Again, after participants had completed all items regarding their own thoughts and
feelings, they were also asked to rate the statements above about the average other group
member in order to assess perceived shared experiences among other group members.
For example, the ﬁrst statement was reworded to: “The average other group member has
shared some experience with the group.” Participants rated these statements on the same

scale.

27

Psychological safety: Participants’ psychological safety was assessed with two
measures: Edmondson’s (1999) Team Psychological Safety scale and ﬁve questions
designed speciﬁcally for this study.

Edmonson’s (1999) Team Psychological Safety scale contains seven items that
are supposed to assess different features of psychological safety. For this study, they were
slightly modiﬁed. An example item is: “It is safe to take a risk in this group.” Participants
rated all items on a scale from 1 (very inaccurate) to 7 (very accurate). Because a re-
wording of the items to assess perceived psychological safety among other group
members made these items very difﬁcult to understand, the questionnaire was not used to
assess perceived psychological safety among other group members.

Participants’ own psychological safety was also assessed with the following ﬁve
questions: “How safe do you feel in the group? How threatening is the group discussion
for you? How relaxed do you feel among the other group members? How comfortable do
you feel with the other group members? How much do you trust the other group
members?” Participants answered these questions on a scale from 1 (not at all) to 7 (very
much). Participants were also asked to answer these questions from the perspective of the
average other group member in order to assess perceived psychological safety among
other group members. For example, the ﬁrst question was reworded to “How safe does
the average other group member feel in the group?”

Perceived intentions of speakers: Participants were told that we were also
interested in the impressions that prior speakers created and in how positive and
benevolent their intentions seemed to be. Participants were therefore presented with

several statements about the intentions behind the other group members’ evaluations and

28

asked to indicate their degree of agreement with each statement on a scale from 1 (not at
all) to 7 (very much).
“The ﬁrst speakers you have heard so far...

...wanted to show how much they cared about the group interaction.

...intended to create a relaxing atmosphere.

...wanted to make the other group members laugh.

...wanted to amuse the other group members.

...wanted to make the other group members angry.

...wanted to insult the other group members.

...wanted to distract group members from the personality evaluation task.”
Because it was very difﬁcult to assess what the participant thought other group members
believed about the ﬁrst speakers’ intentions, we did not ask these questions from the
perspective of the average other group member.

Perceived similarity: Each participant was asked to indicate how similar she saw
herself to the other group members in the following categories: perception of the
situation, feelings in the situation, thoughts about the situation, behavior in the situation,
goals for this interaction, and sense of humor. (Example: “How similar are you to the
other group members in your perception of the situation?”) She was asked to indicate
degree of similarity on a scale from 1 (not at all) to 7 (very much). Each participant was
also asked how similar the other group members were in these categories. (Example:
“How similar are the other group members to each other in their perception of the

situation?”)

29

Assessment of the Dependent Variable

Cohesion: At the moment, there is no generally accepted, standard measure of
cohesion (Dion, 2000). Thus, I used two measures of cohesion that focus on slightly
different components of cohesion in order to assess it as thoroughly as possible.

The ﬁrst measure of cohesion was the 6-item Perceived Cohesion Scale (PCS) by
Bollen and Hoyle (1990). It measures two aspects of cohesion, namely Sense of
Belonging and Feelings of Morale with three items each. An item measuring Sense of
Belonging is “I feel a sense of belonging to this group” and an item measuring Feelings
of Morale is “I am happy to belong to this group.” Participants were asked to indicate
their agreement with each of the six statements on a Likert-type scale from 1 (strongly
disagree) to 7 (strongly agree) to assess experienced personal cohesion. Participants were
also asked to answer these items about the average other group member to assess
perceived group cohesion. For this purpose, the ﬁrst example item was be re-worded to
“The average other group member feels a sense of belonging to this group.” The second
item of these two scales — assessing experienced personal and perceived group cohesion —
was consistently not recorded due to a technical error. Thus, only ﬁve items were used in
the analyses.

The second measure of cohesion was the Group Environment Questionnaire
(GEQ) developed by Carron, Widmeyer, and Brawley (1985). Like most other cohesion
researchers, Carron and his colleagues distinguish between social and task cohesion.
Furthermore, they also distinguish between an individual component and a group
component of cohesion (see also Carron & Brawley, 2000). The individual component is,

for example, an individual’s attraction to the group, and the group component is, for

30

example, closeness of group members. Because the questionnaire had originally been
constructed to measure cohesion in sports teams and not in newly formed task groups,
one item (out of 18 items) was dropped and some others were modiﬁed for this study. For
example, “We all take responsibility for any loss or poor performance by our team” was
re-worded to “We all take responsibility for any poor performance by our group”.
Participants were asked to answer all items from their own perspective to assess
experienced personal cohesion and all items from the other group members’ perspectives
to assess perceived group cohesion.
Order of Variable Assessment

Participants were ﬁrst asked to answer all items about their own thoughts and
feelings and then all items about their fellow group members’ thoughts and feelings. The
exact order of the items was as follows: own mood (20-item PANAS), shared experiences
from own perspective (3 items), own psychological safety (7-item Team Psychological
Safety scale + 5 items), perceived intentions of speakers (7 items), perceived similarity to
others (6 items), experienced personal cohesion (6-item Perceived Cohesion Scale, 17-
item Group Environment Questionnaire), perceived mood of other group members (20-
item PANAS), perceived shared experiences among other group members (3 items),
perceived psychological safety among other group members (5 items), perceived
similarity among others (6 items), perceived group cohesion (6-item Perceived Cohesion

Scale, l7-item Group Environment Questionnaire). All items are listed in Appendix C.

31

Manipulaftion Checks

Participants were asked the following questions about the group interaction after

all mediating and dependent variables had been assessed and after they had evaluated a

student proﬁle as well:

Who were the people in the proﬁles (the choices being “girls on a dating

9, “

website,” “students from the University of Michigan, students from Michigan

9, 6‘

State University, girls in this group”)?

How much humor did the other group members use (1: none at all — 7: very
much)?

How much did group members laugh during the interaction (1: not at all — 7: very
much)?

If there was any humor in the group interaction, what or who was it about (about
unrelated issues, about each other, about other students, about themselves)?
How funny, do you think, were the comments by the other participants (1: not
funny at all — 7: very funny)?

How often, do you think, did you smile during this experiment (1: not at all — 7:
very much)?

How often, do you think, did you laugh during this experiment (1: not at all — 7:
very much)?

Was there anything in the study that seemed strange or unusual to you?

32

RESULTS
Manipulation Checks

When asked who the people in the proﬁles were, 88 out of 96 participants
answered correctly that the people were University of Michigan students. Thus, almost all
participants were aware of the fact that group members referred to outgroup members in
their statements. The data of the remaining eight participants were left in the data set to
have more power for the analyses.

Participants in the Put-down Humor condition indicated that other group members
used more humor (t (94) = 7.94, p < .001, Cohen’s d = 1.64), made funnier comments (t
(94) = 3.08, p < .01, d = .64), and laughed more (t (94) = 10.64 p < .001, d = 2.21) than
did participants in the No Humor condition. Means and standard deviations for the
manipulation check items are displayed in Table 1.

Participants in the Put-down Humor condition also realized that speakers made
humorous comments about other students (vs. unrelated issues, about each other, or about
themselves). Thirty-nine out of 43 participants in the Put-down Humor condition chose
the correct response option. Again, the data from the four participants choosing the
response option were not excluded to have enough power for the analyses.

Participants in the Put-down Humor condition also reported smiling and laughing
more during the experiment than participants in the No Humor condition (t (94) = 2.73, p
< .01 , d = .57 for smiling and t (94) = 2.99, p < .01, d = .61 for laughing). Overall, the

put-down humor manipulation can be regarded as successful.

33

Table 1

Condition Means and Standard Deviations for the Manipulation Check Items

 

 

Put-down Humor No Humor

Item M SD M SD
How much humor did the
other group members use? 5.44 1.01 3.62 1.20
How much did group
members laugh? 5.67 .99 3.08 1.33
How funny were the 4.00 1.41 3.09 1.45
comments?
How often did you smile? 4'33 1'48 3'40 1'78

3.79 1.81 2.72 1.70

How often did you laugh?

 

Note: All ratings were made on a scale from 1 to 7.

Factor Arglyses

In order to assess the dimensionality of the employed measures, I conducted
factor analyses. When the dimensionality of a measure was known, I ran an exploratory
factor analysis and ﬁxed the number of factors in advance. When the measure was new, I
ran an exploratory factor analysis without ﬁxing the number of factors to ﬁnd out of how
many underlying dimensions the measure was comprised.

Mood: Responses to the 20 items from the PANAS were submitted to a principal
axis factor analysis with varimax rotation. Two factors were expected to emerge: Positive
Mood and Negative Mood. Results showed that two factors with eigenvalues of 4.16 and
2.76 explained 34.57% of the variance and that all items loaded on the expected factors.

lterns assessing positive mood loaded on the factor Positive Mood and had factor

34

loadings above .51. Items assessing negative mood loaded on the factor Negative Mood
and had factor loadings above .32. I thus computed the two variables positive mood and
negative mood by averaging the respective items. The internal consistency of these items
was satisfactory: apes, mood = .86 and (1mg, mood = .76.

The same kind of factor analysis with the items assessing perceived positive mood
of other group members and perceived negative mood of other group members yielded
similar results: Two factors with eigenvalues of 4. l 9 and 2.81 explained together 35.33%
of the variance. The rotated factor loading matrix showed the same pattern as the matrix
for items assessing participants’ own mood. Therefore, I computed the two variables
perceived positive mood and perceived negative mood by averaging the respective items.
Their internal consistency was also satisfactory: a perceived p0,, mood = .87 and a perceived neg,
mood = .79.

Shared Experiences: Three items had been designed to assess shared experiences.
Responses to these items were submitted to an exploratory principal axis factor analysis
with varimax rotation. The eigenvalues-greater—than—l rule and the scree plot both
indicated that all three items loaded on one factor, which explained 83.1 1% of the
variance. All factor loadings were higher than .80. In order to compute the variable
shared experiences, responses to these three items were averaged (a = .90).

The same kind of factor analysis with three items assessing perceived shared
experiences among other group members yielded very similar results and also suggested
one factor (eigenvalue = 2.46, 82% of variance explained, factor loadings > .77). Thus,
responses to these three items were also averaged to compute the variable perceived

shared experiences among other group members (a = .89).

35

Psychological Safety: Responses to the seven items ﬁ'om Edmondson’s (1999)
Team Psychological Safety (TPS) scale were submitted to a principal axis factor analysis.
It was expected that one factor would underlie the items. However, one factor with an
eigenvalue of 1.42 only explained 21.22% of the variance. The initial communalities of
the items were generally low; estimates ranged from .41 to .16 with M = .29. Three out of
seven factor loadings were below .30: “If you made a mistake on this group, it would be
held against you.” had a factor loading of .06. “People of this group sometimes reject
others for being different.” had a factor loading of -.20. “It would be difﬁcult to ask other
members on this group for help.” had a factor loading of -.27. I thus decided to exclude
these three items and to run a second factor analysis with the remaining four items.
Results of this second factor analysis showed that one factor with an eigenvalues of 1.38
explained 34.56% of the variance and that all factor loadings (displayed in Table 2) were
larger than .50. Therefore, I averaged responses to these four items to compute the
variable psychological safety ( T PS). Because the internal consistency (a = .67) was still
relatively low, results including this variable have to be regarded with caution.

Psychological safety was also assessed with ﬁve new items. I ran an exploratory
principal axis factor analysis with varimax rotation on these ﬁve items. The eigenvalues-
greater-than-l rule and the scree plot both indicated that all items loaded on one factor
with an eigenvalues of 2.88 explaining 57.56% of the variance. All factor loadings were
larger than the absolute value of .55. In order to compute the variable psychological
safety (new scale), the second item was reverse scored and all items were then averaged.

The internal consistency of these items was good (a = .86).

36

Table 2
Summary of Items and Factor Loadings for the Four Selected Items from the Team

Psychological Safety Scale

 

 

 

Factor
Item Psychological Safety
Working with members of this group, my unique
skills and talents would be valued and utilized. .66
It is safe to take a risk in this group. .65
Members of this group would be able to bring up
problems and tough issues. .52
No one in this group would deliberately act in a
way that undermines my efforts. .50
Eigenvalue 1 .3 8
% of variance 34.56

 

Because I had two different measures of psychological safety — the four items
from the Team Psychological Safety scale (Edmondson, 1999) and ﬁve newly created
items —- I computed the correlation between these two measures. The correlation
coefﬁcient of .40 indicates that the measures are related, but that they do not measure the
same aspects of psychological safety. A closer look at the wording of the items indicates
that the four items from the TPS scale refer to work-related behavior and that the ﬁve
new items refer to one’s general feeling when in the presence of other group members. I
therefore decided to keep these two measures as separate measures of different aspects of
psychological safety.

The ﬁve newly created items assessing psychological safety had also been
rephrased to assess the perceived psychological safety among other group members. I ran
a principal axis factor analysis with these items and expected one factor to emerge.
Indeed, one factor with an eigenvalue of 2.09 explained 41.71% of the variance. The

absolute values of all factor loadings were higher than .24. Therefore, the ﬁve items were

37

averaged (after the second item had been reverse scored) to compute the variable
perceived psychological safety among other group members. The internal consistency of
these items was satisfactory (or = .73).

Perceived Intentions: An exploratory principal axis factor analysis with varimax
rotation was run on the seven newly created items that were supposed to assess the
perceived benevolent intentions of the ﬁrst speakers. I used the eigenvalues—greater-than-
1 rule and the scree plot to determine the number of factors. Three factors (with
eigenvalues of 1.90, 1.49, and 1.15) emerged and explained together 64.84% of the

variance. Factor loadings are displayed in Table 3.

Table 3

Summary of Items and Factor Loadings for the Perceived Intentions Items

 

 

 

Factor 1 Factor 2 Factor 3
Intention to Intention to Concern
Item Entertain Antagmize
The ﬁrst speakers wanted to make the
other group members laugh. .92 .06 .17
The ﬁrst speakers wanted to amuse the
other group members. .95 .12 .12
The ﬁrst speakers wanted to make the
other group members angry. .13 .61 -.17
The ﬁrst speakers wanted to insult the
other group members. -.18 .85 -.07
The ﬁrst speakers wanted to distract
group members from the (. . .) task. .19 .55 -.09
The ﬁrst speakers wanted to show how
much they cared about the group. .05 —.07 .73
The ﬁrst speakers intended to create a
relaxing atmosphere. .25 -.27 .72
Eigenvalue 1 .90 l .49 1 .1 5
% of variance 27.41 21.32 16.38

 

Note: Boldface indicates highest factor loadings.

38

The ﬁrst factor consisted of the two items “The ﬁrst speakers wanted to make the
other group members laugh” and “The ﬁrst speakers wanted to amuse the other group
members”. It was labeled Intention to Entertain. The second factor consisted of the three
items “The ﬁrst speakers wanted to make the other group members angry” “(. . .) wanted
to insult the other group members” and “(. . .) wanted to distract group members” and was
labeled Intention to Antagonize. The third factor consisted of the two items “The ﬁrst
speakers wanted to show how much they cared about the group” and “(. . .) intended to
create a relaxing atmosphere” and was labeled Concern. Because the results of the factor
analysis clearly indicate that the seven items do not measure the same construct, I
decided to compute three different variables (intention to entertain, intention to
antagonize, and concern) by averaging the items that loaded highest on a factor. The
internal consistencies (aintention in entertain = .95, aintention to antagonize = .68, acnncen. = .71) were
satisfactory, although the scales were extremely short. The correlations between the three
variables were low to medium (see Table 4) and also indicate that the variables do not

measure the same construct.

Table 4

Pearson Correlation Coefﬁcients for the Perceived Intentions Measures

 

Measure 1 2 3

 

1. Intention to Entertain --

 

2. Intention to Antagonize .11 --
3. Concern .27* -.26* --
Note: *p < .05

39

Perceived Similarity: An exploratory principal axis factor analysis with varimax
rotation was run on the six newly created items measuring perceived similarity to others.
One factor with an eigenvalue of 3.44 emerged, explaining 57.39% of the variance. The
factor loadings were all higher than .53. Because all items seemed to load on the same
factor, they were averaged to compute the variable perceived similarity to others. This
scale had high internal consistency (a = .88).

A principal axis factor analysis was run on the seven items assessing perceived
similarity among others. Similar results were found as for the previous factor analysis:
One factor with an eigenvalue of 2.96 emerged, explaining 49.41% of the variance. All
factor loadings were higher than .48. Thus, the seven items seemed to load on the same
factor and were averaged to compute the variable perceived similarity among others. It
had high internal consistency (a = .83).

Cohesion: A principal axis factor analysis was run on the items assessing
experienced personal cohesion from the Perceived Cohesion Scale (PCS; Bollen &
Hoyle, 1990). As expected, one factor with an eigenvalue of 3.60 emerged, explaining
72% of the variance. All factor loadings were larger than .77.. To compute the variable
experienced personal cohesion (PCS), all ﬁve items were averaged. The internal
consistency of this variable was very high (a = .93).

The same kind of factor analysis was run on the ﬁve items from the PCS
assessing perceived group cohesion, i.e., participants’ perception of the cohesion among
the rest of the group, excluding themselves. Again, one factor with an eigenvalue of 3.48

emerged, explaining 69.54% of the variance. All factor loadings were larger than .73.

40

Thus, the items were averaged to compute the variable perceived group cohesion (PCS).
They had high internal consistency (a = .92).

Another principal axis factor analysis with varimax rotation was run with the
items assessing cohesion from the Group Environment Questionnaire (Carron et al.,
1985). Because Carron and his colleagues (1985) distinguish ﬁrst between social and task
cohesion and second between an individual and a group component of cohesion
(attraction and group integration, respectively), I expected four factors to emerge. The
four emerging factors had eigenvalues between 1.87 and 1.23 and together explained
44.19% of the variance. An examination of the rotated factor loading matrix (Table 5)
revealed that items supposed to hang together (e.g., all items measuring social attraction)
did not necessarily load on the same factor. For example, the items “I would not enjoy
being part of the social activities of this group”, “I am going to miss the members of this
group when the experiment ends”, and “I could become friends with some people in this
group” are supposed to measure social attraction, the individual component of social
cohesion, but load on different factors. The rotated factor loading matrix seemed to
suggest three different factors: Task Cohesion, Social Cohesion, and Negativity in the
Group. Negativity in the Group subsumes all items that refer to group members’
unhappiness and dislikes. I therefore decided to compute the three variables (by
averaging the following items): experienced task cohesion (item 7, 9, 13; a = .73),
experienced social cohesion (item 2 reverse scored, 6, 8 reverse scored, 12, 14 reverse
scored, a = .60) and experienced negativity in the group (item 1, 3, 5, 11, a = .73). The
items for these variables had relatively low internal consistencies, though. I excluded two

items ﬁom the computations: One item could not be assigned to any factor due to similar

41

loadings on all four factors (item 4 “I could become friends with some people in this
group”) and one item was the only item comprising a Party factor (item 10 “Our group

members would party together.”), which was irrelevant for this study.

Table 5
Summary of Items and Factor Loadings from the Group Environment Questionnaire

(Measuring Experienced Personal Cohesion)

 

Factor 1 Factor 2 Factor 3 Factor 4
Task Negativity (Party) Social
Item Cohesion Cohesion

 

7. Our group is united in trying to
reach its goal. .73 -.14 .13 -.05

13. If members of our group had

problems during the task,

everyone would want to help .60 -.22 .00 .00
them so we can get back on track.

9. We all take responsibility for
any poor performance by our

.45 .04 .15 -.02
group.
1. I would not enjoy being part of
the social activities of this group. -.04 .72 -.20 .17
3. I am unhappy with my group’s
level of desrre to work on the ‘21 0 61 ‘03 .00
task.
5. I do not like the style of work
in this group. -.38 .61 -.10 .25
1 1. Our group members have
conﬂicting aspirations for the .12 .51 .14 .26

group’s performance.

 

Table 5 continues

42

Table 5 continued

 

 

 

Factor 1 Factor 2 Factor 3 Factor 4
Task Negativity (Party) Social
Item Cohesion Cohesion
10. Our group members would
party together. .21 -.07 .87 -.10
6. For me, this group could
become an important social group .46 -.06 .24 -.53
to which I belong.
2. I am not going to miss the
members of thrs group when the _.01 .02 _.13 .50
experiment ends.
12. Our group would like to spend
time together after the experiment .47 .01 .48 ”44
is over.
14. Members of our group are not
likely to strck together outsrde of -.00 .22 ‘07 .42
this expenment.
8. Members of our group would
rather go out on therr own than _.03 .18 .05 .33
get together as a group.
4. I could become friends with
some people in this group. .31 -.32 .35 -.23
Eigenvalues 1.87 1.80 1.30 1.23
% of variance 13.32 12.85 9.27 8.75

 

Note: Boldface indicates the item is assigned to that factor.

Because the analyses resulted in four different measures of experienced personal
cohesion — one from Bollen and Hoyle’s (1990) Perceived Cohesion Scale and three from
Carron et al.’s (1985) Group Environment Questionnaire — it is essential to determine
whether these measures tap into the same aspects of cohesion. If they correlated very

highly, they might simply measure the same thing and could be collapsed. But because

43

the correlations between them are all smaller than the absolute value of .36 (M absolute values
= .31, range absolute values = .27 — .36; see Table 6), they do not correlate very highly and

were analyzed separately.

Table 6

Pearson Correlation Coefﬁcients for the Experienced Personal Cohesion Measures

 

Measure 1 2 3 4

 

1. Experienced Personal Cohesion (PCS) --

2. Experienced Task Cohesion (GEQ) .34* --
3. Experienced Social Cohesion (GEQ) .33* .27* ~-
4. Experienced Negativity (GEQ) -.30* -.27* -.36* --

 

Note: *p < .05. PCS = Perceived Cohesion Scale. GEQ = Group Environment
Questionnaire

With the items measuring perceived group cohesion from the GEQ (Carron et al.,
1985), I ran another principal axis factor analysis. For this analysis, I excluded items 4
and 10 because they had not been proven useful for the computation of experienced
personal cohesion. I expected three factors to emerge due to the results of the previous
factor analysis. Three factors with the eigenvalues 2.48, 1.53, and 1.50 explained together
45.91% of the variance. All items loaded on the expected factors. I therefore computed
the following variables by averaging items: perceived task cohesion (item 7, 9, 13; a =

.60), perceived social cohesion (item 2 reverse scored, 6, 8 reverse scored, 12, 14 reverse

44

scored, 0. = .76), and perceived negativity in the group (item 1, 3, 5, l 1, a = .60). The
internal consistencies of these items are generally low.

Because the correlations between the different measures of perceived group
cohesion (see Table 7) are medium-sized (M absolute value, = .28, range absolute value, = .04 —
.57), they indicate that the measures detect different aspects of the construct and should

be analyzed separately, just like the measures of experienced personal cohesion.

Table 7

Pearson Correlation Coefﬁcients for the Perceived Group Cohesion Measures

 

Measure 1 2 3 4

 

1. Perceived Group Cohesion (PCS) --

2. Perceived Task Cohesion (GEQ) .37* --
3. Perceived Social Cohesion (GEQ) .57* .35* --
4. Perceived Negativity (GEQ) -.16 -.04 -.19 --

 

Note: *p < .05. PCS = Perceived Cohesion Scale. GEQ = Group Environment
Questionnaire

Overall, the results of the factor analyses suggest that most measures had good
psychometric properties, except for the Team Psychological Safety scale, which measures
psychological safety, and the Group Environment Questionnaire, which measures
cohesion. They both have low internal consistencies. Therefore, all results based on these

scales need to be regarded with some caution.

45

Correlations between Measures of Own Thoughts and Feelings and Others’ Thoughts and
Feelings

As explained in the methods section, participants were asked to answer questions
both from their own perspective and from the perspective of the average other group
member. The correlations between the resulting two measures could potentially be very
high for two reasons: First, participants might not be able to distinguish clearly between
their own thoughts and feelings and their fellow group members’ thoughts and feelings.
Actually, the main base for inferring what their fellow group members think and believe
in a certain moment could simply be their own thoughts and feelings. Second, there is
likely to be a high amount of common method variance because the wording of the
respective items is very similar and the items are assessed with the same method at
almost the same time. However, the correlations between measures of own and others’
thoughts and feelings were, on average, medium in size (M = .44, range .12 — .65, see
Table 8) and low enough to indicate that two scales referring to one construct — but from
different perspectives — do not entirely measure the same thing, but slightly different

aspects.

Did Put-down Humor Affect Experienced Personal Cohesion?

To ﬁnd out if put-down humor had the expected positive effect on experienced
personal cohesion, I conducted independent-samples t-tests on all measures of
experienced personal cohesion. As can be seen in Table 9, put-down humor did not have

the expected positive effect. Participants in the Put-down Humor condition felt about the

46

same amount of personal cohesion (assessed with the PCS), social cohesion, and

negativity (both assessed with the GEQ) as participants in the No Humor condition.

Table 8
Pearson Correlation Coefﬁcients between Measures of Own Thoughts and Feelings and

Measures of Others’ Thoughts and Feelings

 

 

Measures Pearson’s r
own positive mood and
perceived positive mood of other group members .65*
own negative mood and
perceived negative mood of other group members .49*
shared experiences and
perceived shared experiences among other group members .47 *

own psychological safety and
perceived psychological safety among other group members .43*

perceived similarity to others and
perceived similarity among others .12

experienced personal cohesion (PCS) and

 

perceived group cohesion (PCS) .36*
experienced task cohesion (GEQ) and
perceived task cohesion (GEQ) .65*
experienced social cohesion (GEQ) and
perceived social cohesion (GEQ) .47*
experienced negativity in group (GEQ) and

merceived negativity among others (GEQ) .33*
Note: *p < .05

Participants in the Put-down Humor condition even felt signiﬁcantly l_e_s_s experienced

task cohesion than participants in the No Humor condition, t(94) = -2.66, p < .01. They

47

did not feel that the group was striving towards the same goal and that group members

were supportive and willing to take responsibility. Thus, hypothesis 1.1 is not supported.

Table 9

Means, Standard Deviations, and T—tests for the Experienced Personal Cohesion

 

 

Measures
Measure Condition M SD t-value p-value Cohen’s d

Experienced personal Put-down Humor 3.26 1.13

cohesion (PCS) No Humor 3.48 1.16 -.88 .38 -.19
Experienced task Put-down Humor 3 .48 1.08

cohesion (GEQ) No Humor 4.10 1.18 -2.66 .009 -.55
Experienced social Put-down Humor 3.26 .97

cohesion (GEQ) No Humor 3.17 1.04 .46 .65 .09
Experienced Put-down Humor 3.17 1 .12

negativity (GEQ) - No Humor 2.96 1.04 .96 .34 .19

 

Notes: All ratings were made on a scale from 1 to 7. PCS = Perceived Cohesion Scale.
GEQ = Group Environment Questionnaire. All degrees of freedom = 94.
Did Put-down Humor Affect Perceived Group Cohesion?

To ﬁnd out if put-down humor had the expected positive effect on perceived
group cohesion (i.e., participants’ perception of the cohesion among the rest of the
group), I conducted another set of independent-samples t-tests on all measures of
perceived group cohesion. As displayed in Table 10, put-down humor had the expected
positive effect on perceived group cohesion (assessed with the PCS) and — to a lesser
extent — on perceived social cohesion (assessed with the GEQ; recall that these two
measures were the most highly correlated of all perceived cohesion measures, r = .57).
Participants in the Put-down Humor condition felt that their fellow group members were

a more cohesive unit than participants in the No Humor condition, t(94) = .3.34, p = <

48

.01. Participants in the Put-down Humor condition also tended to perceive slightly more
social cohesion, t(94) = 1.74, p < .10, and slightly less negativity among their fellow
group members than participants in the No Humor condition, t(94) = 1.85, p < .10.
However, they did not perceive their fellow group members to have higher task cohesion
than participants in the No Humor condition. To summarize, put-down humor from
fellow group members seemed to indicate to participants that the others liked the group
and fellow group members (even beyond the experimental context) and were happy to be
in the group, that they were less negative about several aspects of their group, but that
they did not feel particularly committed to the task. Thus, hypothesis 1.2 is partially

supported.

Table 10

Means, Standard Deviations, and T-Tests for the Perceived Group Cohesion Measures

 

 

Measure Condition Mean SD t-value p-value Cohen’s d

Perceived group Put-down Humor 4.41 1.08

cohesion (PCS) No Humor 3.63 1.20 3.34 .001 .68
Perceived task Put-down Humor 3.86 1.04

cohesion (GEQ) No Humor 4.10 1.02 -1.14 .26 -.23
Perceived social Put-down Humor 3.79 1.08

cohesion (GEQ) No Humor 3.39 1.13 1.74 .09 .36
Perceived negativity Put-down Humor 2.90 .83 ,

(GEQ) No Humor 3.21 .83 -1.85 .07 -.37

 

Notes: All ratings were made on a scale from 1 to 7. PCS = Perceived Cohesion Scale.
GEQ = Group Environment Questionnaire. All degrees of freedom = 94.

49

Did Put-down Humor Affect Participants’ Own Thoughts and Feelings in the Gromg?

The put-down humor manipulation did not affect most of the variables assessing
participants’ own thoughts and feelings (see Table 11). It did not increase participants’
positive mood, their sense of shared experiences, their psychological safety, and their
feelings of similarity to other group members. It did, however opposite to expectations,
increase their negative mood, t(94) = 1.96, p = .05.

Participants in the Put-down Humor condition believed that the other group
members had a stronger intention to entertain (t(94) = 8.6, p < .001) than did participants
in the No Humor condition. Participants in the Put-down Humor condition did not,
however, believe that their fellow group members showed more concern than did
participants in the No Humor condition, but rather assumed that their fellow group
members’ intention was to antagonize others, t(94) = 2.02, p < .05. Although the obvious
goal of fellow group members might have been to entertain each other, they were
perceived as lacking regard and as being motivated to antagonize others.

In order to ﬁnd out why participants in the Put-down Humor condition had higher
scores on negative mood than participants in the No Humor condition, I computed scores
for speciﬁc negative emotions according to the PANAS-X (Watson & Clark, 1994). Fear
is the average of participants’ scores on the items distressed, scared, nervous, jittery, and
afraid. Hostility is the average of participants’ scores on the items upset, hostile, and
irritable. Guilt is the average of participants’ scores on the items guilty and ashamed. As
can be seen in Table 1 1, participants in the Put-down Humor condition did not experience
more fear and guilt, but they did experience more hostility than participants in the No

Humor condition (t (94) = 2.16, p < .05). The humorous comments of the (ostensible)

50

fellow group members seemed to have upset and irritated participants in the Put-down
Humor condition.

Overall, it seems that participants in the Put-down Humor condition felt
uncomfortable and disconnected from the other group members. Thus, hypotheses 2.1,

2.2, 3.1, 4.1, 5.1 and 6.1 were all not supported.

Did Put-down Humor Affect Perceptions of Others’ Thougﬂs and Feelings?

The put-down humor manipulation affected how participants evaluated their
fellow group members’ thoughts and feelings (see Table 12) — but not unifome
according to the expectations. Participants in the Put-down Humor condition regarded
their fellow group members to have slightly more positive emotions, to share slightly
more experiences among each other, to feel more psychological safety among each other
(t(94) = 2.27, p < .05), and to be more similar to each other (t(94) = 2.27, p < .05) than
did participants in the No Humor condition. Participants in the Put-down Humor
condition did not, however, believe that their fellow group members were in a less
negative mood than did participants in the No Humor condition.

Overall, put-down humor seems to have had some positive effects on the
perception of other group members’ thoughts and feelings. They were regarded as similar
to each other and displayed psychological safety in their actions. However, hypotheses
2.4 and 5.2 were clearly disconﬁrrned. Hypotheses 2.3, 3.2 4.2 and 6.2 will be tested in

more detail.

51

Table 11

Means, Standard Deviations, and T-Tests for the Measures of Participants’ Own

 

 

Thoughts and Feelings
Measure Condition Mean SD t-value p—value Cohen’s (1

positive mood Put-down Humor 2.23 .58

No Humor 2.31 .80 -.54 .59 -.11
negative mood Put-down Humor 1.55 .45

No Humor 1.38 .37 1.96 .05 .4-1
fear Put-down Humor 1.69 .62

No Humor 1.54 .60 1.22 .22 .25
hostility Put-down Humor 1.51 .62

No Humor 1.28 .37 2.16 .03 .45
guilt Put-down Humor 1.27 .43

No Humor 1.16 .44 1.21 .23 .25
shared experiences Put-down Humor 3.23 1.44

No Humor 3.38 1.34 -.53 .60 -.11
psychological safety Put-down Humor 3.47 1.01
(TPS) No Humor 3.72 .98 -1.24 .22 -.25
psychological safety Put-down Humor 4.24 1.23
(own scale) N0 Humor 4.38 1.09 -.61 .55 -.12
intention to entertain Put-down Humor 6.16 .88

No Humor 4.20 1.27 8.60 <.001 1.79
intention to Put-down Humor .83
antagonize No Humor .69 2.02 .05 .42
Concern Put-down Humor 3.93 1.39

No Humor 3.83 1.40 .35 .73 .07
similarity Put-down Humor 4.03 1.23

No Humor 4.19 .87 -.77 .44 -.15

 

Notes: TPS = Team Psychological Safety scale. Mood ratings were made on a scale from
1 to 5. All other ratings were made on a scale from 1 to 7. All degrees of freedom = 94.

52

Table 12

Means, Standard Deviations, and T-Tests for Measures of Other Group Members’

 

 

Thoughts and Feelings

Measure Condition Mean SD t-value p-value Cohen’s d
perceived positive mood Put-down Humor 2.54 .61
0f other group members No Humor 2.29 .74 1.76 .08 .37
perceived negative mood Put-down Humor 1.43 .45
0f other group members No Humor 1.40 .34 .48 .63 .08
perceived shared ex- Put-down Humor 4.06 1.15
Peﬁeﬂces among Others No Humor 3.58 1.37 1.85 .07 .38
perceived psychological Put-down Humor 4.98 .83
SﬂfeW among others No Humor 4.60 .78 2.27 .03 .47
perceived similarity Put-down Humor 4.97 .86
among others No Humor 4.58 .79 2.27 .03 .47

 

Notes: TPS = Team Psychological Safety scale. Mood ratings were made on a scale from
1 to 5. All other ratings were made on a scale from 1 to 7. All degrees of freedom = 94.

Mediational Analysis

I hypothesized that the thoughts and feelings of other group members would
mediate the effect of put-down humor on perceived group cohesion. Because only four of
the potential mediating variables (perceived positive mood, perceived shared experiences,
perceived psychological safety and perceived similarity among others) differed between
conditions (but not necessarily signiﬁcantly), I tested whether these four variables might
mediate the relation between put-down humor and perceived group cohesion. The
correlations between the four potential mediating variables ranged between .27 and .54.

I included all of these variables in a multiple mediation model depicted in Figure
2. This model shows both the direct effect of put-down humor on perceived cohesion and

the indirect effects via the four mediators perceived positive mood, perceived shared

53

experiences, perceived psychological safety, and perceived similarity. It is important to
include all potential mediating variables in the model (instead of testing each one
separately) to ﬁnd out if the whole set of variables can collectively mediate the relation
between put-down humor and perceived cohesion. One can, of course, also still test to
what extent each variable mediates the relation. Each speciﬁc indirect effect is the ability
of one variable to mediate the effect when all other mediators are included in the model
(i.e., controlled for). An indirect effect is deﬁned as the product of the two
unstandardized coefﬁcients connecting put-down humor with perceived cohesion through
that one speciﬁc mediator. The total indirect effect is the sum of all speciﬁc indirect
effects.

In order to test the effects, I bootstrapped them using the SPSS version of a macro
supplied by Preacher and Hayes (Preacher & Hayes, under review; see also
www.quantpsy.org). Bootstrapping is a nonparametric resampling procedure that is very
well suited for testing hypotheses about mediation because the sampling distributions of
indirect effects are sometimes not normal (MacKinnon, Lockwood, & Williams, 2004).
The advantage of bootstrapping is that is makes no assumptions about the sampling
distributions of coefﬁcients and indirect effects (the product of the coefﬁcients).
MacKinnon, Lockwood, Hoffman, West, & Sheets (2002) showed that bootstrapping led
to more accurate Type I error rates and greater power for ﬁnding indirect effects than
other methods — partly because fewer inferential tests are involved than in other methods
(e. g., the causal steps strategy, Baron & Kenny, 1986; the Sobel test, Sobel, 1982, 1986).
Another advantage is that bootstrapping places more emphasis on the direction and the

size of the effect instead of signiﬁcance levels.

54

The point estimates and the 95% bias—corrected conﬁdence intervals of the
indirect effects are listed in Table 13. Point estimates are equivalent to the product of
unstandardized coefﬁcients that would have been found using the causal steps method
(Baron & Kenny, 1986). Traditionally, they would have been tested for signiﬁcance
using the Sobel test (1982). (As can be seen in the table, none of the estimates for indirect
effects are signiﬁcantly different from zero when using the Sobel test, which has been
shown to have too little power for small samples like this one.)

The interpretation of the bootstrapped estimates is that all four variables together
mediated the effect of put-down humor on perceived cohesion. The total indirect effect
was .41. Again, the total indirect effect is the sum of all speciﬁc indirect effects. The total
and direct effects of put-down humor on perceived cohesion are .79, p < .01 and .3 8, p <
.10, respectively. These two values are the unstandardized regression coefﬁcients and
were not bootstrapped. When all four mediators are taken into account, the effect of put-
down humor is reduced from .79 to .38 and becomes non-signiﬁcant. The difference
between these two effects is the total indirect effect (.41, see above). The 95% biased-
corrected conﬁdence interval (CI) for the total indirect effect is .08-.71. Because this CI
does not contain 0, we can claim that the total indirect effect is different from 0.

It is not only essential to test whether there is an overall indirect effect. It is also
important to test individual mediators in the model. Unstandardized coefﬁcients and their
standard errors are shown in Figure 2. The directions of all coefﬁcients are, as predicted,
positive. Participants who heard put-down humor in the group perceived their fellow
group members to have more positive emotions, shared experiences, to feel safer, and to

be more similar to each other than participants who did not hear put-down humor. A

55

closer examination of each speciﬁc indirect effect shows that only perceived positive
mood and perceived similarity can be regarded as mediators, though, because their
conﬁdence intervals do not contain 0. The other two variables do not contribute
signiﬁcantly to the indirect effect above and beyond the effects of perceived positive

mood and perceived similarity.

Table 13
Mediation of the Effect of Put-down Humor on Perceived Group Cohesion Through
Perceived Positive Mood, Perceived Shared Experiences, Perceived Psychological

Safety, and Perceived Similarity

 

 

 

Bootstrapping
Product of Coefficients BC 95% CI

Indirect Effects Point Estimate SE Z Lower Upper
Perceived positive mood .14 .09 1.61 .01 .40
Perceived shared experiences .07 .06 1.18 -.01 .24
Perceived psy. safety .09 .07 1.38 -.03 .38
Perceived similarity .11 .07 1.57 .002 .36
TOTAL .41 .16 2.59 .08 .71

 

Notes: BC = bias-corrected. 5000 bootstrap samples. Z = value for the Sobel test (1982)

56

Perceived
positive mood _57 (.15)

  
 

.24 (.14)
.48 (.26)

/-" Perceived shared .13 (.09)
experiences \

 

 
 

Put-down Humor -38 (20) A Perceived Group
' Cohesion
\ Perceived psy. /
safety .23 (.14)
Perceived .30 (.14)
Similarity

Figure 2: Mediation of Put-down Humor on Perceived Group Cohesion by Perceived
Positive Mood, Perceived Shared Experiences, Perceived Psychological Safety, and

Perceived Similarity. Displayed are unstandardized coefﬁcients (and their standard

errors).

57

DISCUSSION

The goal of this study was to examine the effects of put-down humor on cohesion
in groups. More speciﬁcally, I created conditions under which put-down humor was
hypothesized to increase cohesion and examined the underlying process. I assumed that
the following two conditions were necessary and sufﬁcient for put-down humor to
increase cohesion: First, the target of the humorous comment is someone outside the
group (an outgroup member) and second, the comment is appreciated by group members,
what they signal through laughter.

All variables were assessed at two levels. First, from the participant’s own
perspective: Participants answered questions about their own thoughts and feelings
(Example: How safe do you feel?) Second, from the perspective of the other group
members: Participants also answered questions about their fellow group members’
thoughts and feelings (Example: How safe does the average other group member feel?)

The participants who heard put-down humor did not feel more attracted to the
group than participants who did not hear any humor. In other words, participants in the
Put-down Humor condition did not experience more personal cohesion than participants
in the No Humor condition. In fact, they even felt less task cohesion. The put-down
humor did also not affect participants’ own thoughts and feelings besides their negative
mood. Participants who heard put-down humor were in a more negative mood than
participants who did not hear any humor. They felt especially upset, irritated, and hostile.

However, hearing put-down humor made participants believe that the o_tl;e_r group
members (excluding the participant) were more attracted to the group (compared to

participants not hearing humorous comments and reactions). Participants in the Put-down

58

Humor condition regarded their fellow group members to be a more cohesive unit, even
beyond the experimental context, but also to be less committed to the task than
participant in the No Humor condition. Participants in the Put-down Humor condition
judged their fellow group members to be in a more positive mood, to share more
experiences, to feel safer, and to be more similar to each other than did participants in the
No Humor condition. Furthermore, the variables perceived positive mood and perceived
similarity among other group members were found the mediate the relation between put-
down humor and cohesion in a model containing four potential mediators.

To summarize, although put-down humor was presumed to lead to increased
personal attraction to the group, it did not. Moreover, the participants in the Put-down
Humor condition seemed to feel very different from how they perceived the feelings of
their fellow group members. This observation is also reﬂected in the low correlation
coefﬁcient between the measure of similarity with other group members and similarity
mg other group members. The put-down humor had no or a negative effect on own
feelings and feelings about the group, but made those who expressed it seem to be feeling
more positively emotionally and about the group.

In this study, participants anticipated speaking and being heard by other group
members, three of whom had been successfully joking during their turns. It is possible,
therefore, that I did not ﬁnd the expected effect on experienced personal cohesion
because participants in the Put-down Humor condition experienced heightened anxiety.
Such anxiety may not be inherent in overbearing put-down humor of outgroup members
in one’s group. Rather, such anxiety may have been aroused by the particular

experimental procedures used in Study 1. In this experimental paradigm, participants

59

anticipated having to perform the same task that three fellow group members had
previously undertaken. In the No Humor condition, participants heard three ostensible
fellow group members talk in a fairly direct, sober fashion, with practically no reactions
from the others in the group. In the Put-down Humor condition, on the other hand, the
three ostensible fellow group members made several put-down comments about the
stimulus persons, and these comments were reacted to with hearty laughter. So, in the
Put-down Humor condition, participants may have perceived additional risks that were
absent from the No Humor condition: 1) They might not meet the emerging
norm/standard of being humorous in the group. That is, they might have been afraid that
they would not be as ﬁrnny as the ﬁrst speakers. 2) Their apparently highly judgmental
fellow group members would evaluate them negatively (as they had the outgroup
members). It was not my expectation or intent to show that group members are more
strongly attracted to groups which put higher performance demands upon them. Eys,
Hardy, Carron, and Beauchamp (2003) showed that athletes who perceived their
cognitive anxiety as debilitative had lower perceptions of group cohesion than athletes
who perceived their cognitive anxiety as facilitative. In my study, participants probably
experienced a certain amount of debilitative anxiety as well, which could have undercut
the hypothesized relationship between put-down humor and cohesion. This interpretation
is consistent with other features of the data: In the Put-down Humor condition (as
compared to the No Humor condition), participants’ positive moods did not improve,
they did not feel they shared more experiences with the other group members, and they
did not feel more psychological safety. My original hypothesis assumed that the

perceived demands and risks of participation were identical for those who did or did not

60

hear put-down humor. However, my original paradigm may well have failed to achieve
such equivalence.

Study 1 clearly showed that there are some boundary conditions for the personal
cohesion effect. In this study, as l have argued above, I probably produced experimental
conditions that suppressed the effect. Therefore, these conditions will be altered in Study

2 to ones less likely to suppress the hypothesized effect.

61

STUDY 2

In Study 2, I altered several elements of Study 1’s experimental procedures to
reduce participants’ potential anxiety and evaluation apprehension. In order to decrease
the participant’s fear of not meeting an apparent high-humor standard, she always talked
ﬁrst in the group. At the beginning of the study, the participant listened to an example
recording so that she would know how to handle the personality evaluation task. Then
she was (supposedly) randomly chosen to be the ﬁrst group member to evaluate a student
proﬁle. Because at that point, the participant had not heard any other group member talk
and make humorous comments, there was no humor standard when the participant
started.

In order to decrease the participants’ fear of being evaluated negatively by others,
she was always evaluated positively by her fellow group members at the beginning of the
study. The participant was asked to ﬁll out a short proﬁle about herself (with, e.g.,
information about her major and career goals) at the beginning of the study and told that
all other group members would be able to examine it. She was then able to see the
proﬁles of her fellow (ostensible) group members and asked to pick the person who
seemed the most reliable. All other group members would be asked to do the same. After
that, the participant was told that the other group members had voted her to be the most
reliable person in the group and that her task would be to record all evaluations of group
members. Thus, in Study 2, the participant received a positive evaluation from her fellow
group members and attained a higher status position in the group.

In Study 2, all other aspects of the procedure remained the same. To measure

group identiﬁcation, a construct that is closely related to cohesion, I added the Group

62

Identiﬁcation Scale (Hinkle, Taylor, Fox-Cardamone, & Crook, 1989) because the Group

Environment Questionnaire exhibited poor psychometric qualities in Study 1.

63

METHOD
De_si&

The design of the study was a single-factor design, the same as in Study 1: Put-
down humor of outgroup members (followed by common laughter) was present or absent
in the group discussion.

Participants

Participants in this study were 139 female college students who were recruited
through the MSU HPR Pool and received one hour of credit for taking part in the study.
Sixty-one percent of participants were in their ﬁrst year of study at MSU, 24% were
sophomores, 11% juniors, and 4% seniors.

Due to technical failures, data of four participants were not recorded. A total of 21
(= 15.1%) participants realized that the voices of their fellow group members had been
recorded. This might have been due to the fact that participants needed to speak before all
other group members and did not receive any feedback from them — unlike the later
speakers who all received feedback. The data of the participants who ﬁgured out the set-
up of the study were excluded from the analyses.

T wenty-two (= 15.8%) participants expressed some suspiciousness about the
voting process for a recorder and did not fully believe that the other group members
voted them to be the recorder. Instead, they indicated in their answers to the open-ended
manipulation check question (Was there anything in the study that seemed strange or
unusual to you?) that the computer program told every group member they were voted to
be the recorder (which was, in fact, the case). However, the data of these participants

were not excluded from the analyses because they still believed that they were members

64

of an ad-hoc group and collaborating with their fellow group members. The remaining
number of participants for the analyses was thus 114, with 57 participants each in the No
Humor and in the Put-down Humor condition.

As for Study 1, a power analysis with the computer program GPOWER (F aul &
Erdfelder, 1992) had shown that 51 participants per condition were enough to have an
80% chance of correctly rejecting the null hypothesis if the effect was moderate (d = .5;
power was estimated for studies using a one-sided t-test for independent means and
testing hypotheses at the .05 signiﬁcance level).

Cover Stog

The cover story was the same as in Study 1: The experimenter told participants
that we were interested in how groups work together when group members could only
communicate via headsets. The participant believed that she could communicate with ﬁve
other group members, but in fact listened to recorded statements.

The Task

The task was also the same is in Study 1: Participants were told that each group
member was supposed to examine a proﬁle from a University of Michigan student and to
judge how outgoing, ﬁiendly, reliable, moody, and open-minded the student in the proﬁle
seemed to be. Each group member would evaluate one student proﬁle, one after the other.
Procedure

The procedure in this study resembled the one from Study 1. Upon arrival at the
lab, the experimenter led participants into separate rooms and told them individually what
the goal of the study and their task was. Then the participant had to draw one out of

several cards presented by the experimenter and after opening it, found out that her card

65

was marked with a red dot. The experimenter explained that this color would later
determine when the participant had to speak.

After this, the participant started reading the instructions on the computer screen.
She was informed that each group needed one member to record everyone’s evaluations.
This recorder would be determined through popular vote. The participant read: “All
group members need to answer a few questions. The answers will be shown to everyone.
Then every group member votes who she thinks would be the best recorder in this
group.”

The participant then typed her answers to the following questions: “What is your
major? What is your class level? If you have a job, where do you work? Where do you
currently live (e.g., on campus, off campus)? What are your career goals? What is an
interesting fact about you? Who is your hero?” The participant then clicked on a button to
send off her answers to all other group members. In fact, the answers were not sent to
anyone else, though.

After a few seconds of waiting, answers of ﬁve other ostensible group members
were displayed on the screen (see Figure 3). The participant was asked to study these
answers carefully because she would later need to choose who would be the most reliable

recorder. Nobody could vote for themselves.

66

30> Succeed 05 com cowmﬁuomﬁ casewxoem we E09802 95.5 036530 .550 aim me 8595 ”m Semi

 

 

 

 

 

 

A-.. meohmg 854 “Baum
ﬁaonm scam ”Baum m. you» mqnmuhzﬁ
435% E “San Em m ”8an wannabes magmas“.
€me 6”.an ”28m .528 owes a 8 33588 cm 2889 umﬁom hobo
madame co ”00:8me “Ego an: 530 henna“: Ego
rueﬁ Hanna mmm e33
gnaw ”$30 .583. ”$30 50> a we
$205.3; “8.82 manganese. ”SEE
c 20353 W ZOaMHm v ZOmMmm
gong ”moobm
Human 53m “808$ Egan £3830 ”moobm S038
.305 “52 H “Home wenmuhg n5 2: Sam :8 Se wannabe: amﬂ a nooom BEE ”8m.“ maumogﬁ
ESE m 0838. imam beau amazon €89 «Rom nacho Bog reg “mﬁom .950
uﬁE ﬁrs engage “3838 “850 «885% SE53? Hooﬁuaﬂ “Ego 885$ “882%: “Ego
use: “x83 FEE E ”€03 page 05 E “can; m mm 383
ucoﬁoﬁom umﬂo Baa ”3&0 agave.“ SR0
$22.6th ”3.82 .3038ka “8.8—2 83039. SE2
M 20333 ~ 203mg u 203mg

 

.azeaw .59». i .8289. £35: 3.2.: .2: 2— 3.3%
2—3 :8 =5..— 3 >5 En: inn—59.5 maﬁa.— 32: 25:53 once—m

67

After reading the answers, the participant needed to vote for one person to be the
recorder. She was subsequently informed that group members had voted Person 4 to be
the recorder “because Person 4 seemed to be the most reliable”. We made sure each
participant would know she was Person 4 by putting a poster that read “You are Person
4” on the wall in front of her and by indicating in the proﬁles that she was Person 4 (see
Figure 3). The participant ﬁrrther read: “Your task is to record all evaluations. Please tum
over the white sheet that is lying on your desk and write the evaluations in the appropriate
places.” The participant then turned around a sheet that was lying in front of her. It
contained the following instructions: “Please record the ratings of your group members in
the following table. You only need to write down the numbers that correspond to the
ratings, but NO comments.” The participant also read that group members need to use the
following scale for their ratings: 1 (well below average), 2 (below average), 3 (above
average), 4 (well above average). The participants further read: “For example, if someone
says ‘Victoria is above average in moodiness’, write down 3. If a group member does not
make a rating, leave the appropriate space blank.” At the bottom of the page was the table

that the participant in her role as recorder needed to ﬁll out:

 

How How How How How open-
outgoing? ﬁiendly? reliable? moody? minded?

 

Juﬁe

 

Ashley

 

 

Megan
Kim

 

Katie

 

 

 

 

 

 

 

Victoria

 

Figure 4: Table for Recorder in Study 2
The participant (and, ostensibly, the other ﬁve group members) then listened to an

example evaluation that was supposedly from an earlier study in order to become familiar

68

 

with the task. She saw a student proﬁle on her screen and heard how the evaluator gave
her impression of the student and judged how outgoing, friendly, reliable, moody, and
open-minded the person in the proﬁle seemed to be. There was no put-down humor in
this recording.

After listening to the example evaluation, the participant read that now members
of her group would start evaluating the student proﬁles. One group member would
evaluate the ﬁrst student proﬁle, another group member the second one, and so on. The
participant was reminded that she picked a card with a color (red) at the beginning of the
study and that it would be her turn to talk when this color showed up as a frame around a
student proﬁle.

Then the evaluation process started. The ﬁrst student proﬁle showed up on the
screen with a red frame, which meant that the participant needed to talk ﬁrst. While the
participant was speaking into the microphone, background noise that included people
breathing and moving chairs played in her headphones so that she would believe other
group members were listening to her. However, the participant did not receive any kind
of verbal feedback from the other group members. But all subsequent speakers received
very explicit feedback from the others in the form of laughter (because their statements
and the laughter had been recorded). Thus, some participants might have rightfully
concluded that the statements of their fellow group members had been recorded. After the
participant had evaluated the ﬁrst student proﬁle, three other proﬁles showed up on the
screen and were evaluated by three other ostensible group members.

After these four evaluations, the task was interrupted and the participant was

asked to answer questions about the group so that we could assess all potential mediating

69

and dependent variables. The computer program subsequently indicated that the last two
group members would not be able to evaluate a student proﬁle because time for the
experiment was running out. At the end, the participant was asked to indicate how much
she liked the four students from the previously evaluated proﬁles. She was also asked
what she thought the purpose of the study to be, whether anything had struck her as
unusual, and to answer some manipulation check items. Finally, the participant was
debriefed and thanked.
Manipulation of the Independent Variable

Put-down humor was manipulated like in Study 1: The ﬁrst three ostensible group
members whom the participants heard either used or did not use put-down humor in their
statements.
Assessment of Potential Mediating Variables

All potential mediating variables were assessed with the same questions as in
Study 1.
Assessment of the Dependent Variable

Cohesion was again assessed with the 6-item Perceived Cohesion Scale (PCS;
Bollen & Hoyle, 1990) and the 17-item Group Environment Questionnaire (GEQ; Carron
et al., 1985), which had been modiﬁed for this study. Group identiﬁcation was also
assessed in this study because it is closely related to cohesion and because the GEQ had
proven to have rather poor psychometric properties in Study 1. It was measured with
Hinkle et al.'s (1989) version of the Group Identiﬁcation Scale (GIS). This questionnaire

contains nine items, which participants needed to rate on a scale from 1(strongly

70

disagree) to 7(strongly agree). Example items are: “I identify with this group. I am glad
to belong to this group. I do not consider the group to be important (reverse scored)”
Order of Variable Assessment

Parallel to Study 1, participants were asked to answer all items about their own
thoughts and feelings ﬁrst and all items about the thoughts and feelings of their fellow
group members after that. The order of the items was as the same as in Study 1 with the
insertion of the Group Identiﬁcation Scale after the cohesion measures: own mood (20-
item PANAS), shared experiences from own perspective (3 items), own psychological
safety (7-item Team Psychological Safety scale + 5 items), perceived intentions of
speakers (7 items), perceived similarity to other group members (6 items), experienced
personal cohesion (6-item Perceived Cohesion Scale, 17-item Group Environment
Questionnaire), own identiﬁcation with the group (9-item Group Identiﬁcation scale),
perceived mood of other group members (20-item PANAS), perceived shared
experiences among other group members (3 items), perceived psychological safety
among other group members (5 items), perceived similarity among other group members
(6 items), perceived group cohesion (6-item Perceived Cohesion Scale, 17-item Group
Environment Questionnaire), perceived identiﬁcation of others with the group (9-item
Group Identiﬁcation Scale). All items are listed in Appendix C.
Manipulation Check_s_

At the end of the study, participants were asked the same questions as in Study 1
to ﬁnd out if the manipulation had been successful. They were also asked: “Who was the

recorder in this group?” (with the response options “I was the recorder” and “Someone

71

else was the recorder”) to check if they had realized that they had been voted to be the

recorder in their group.

72

RESULTS
Manipulation and Cover Story Checks

One hundred nine out of 114 participants correctly identiﬁed the students in the
proﬁles as students from the University of Michigan. Almost all participants were thus
aware that the statements they heard were about outgroup members. One hundred thirteen
participants also indicated that they were the recorder in the group. Only one person
indicated that someone else was the recorder, 38 (1) = 110.03, p < .001. No data was
discarded to have enough power for the analyses.

The put-down humor manipulation can again be regarded as successful:
Participants in the Put—down Humor condition stated that the other group members used
more humor (t(112) = 7.97, p < .001, d = 1.49), laughed more (t(112) = 11.08, p < .001, d
= 2.07), and made funnier comments (t(112) = 2.73, p < .01, d = .51) than did participants
in the No Humor condition. Participants in the Put-down Humor condition also reported
smiling and laughing more (t(112 = 2.64, p < .01, d = .49, t(112) = 3.32, p < .01, d = .62
respectively) than participants in the No Humor condition. All means and standard
deviations from the manipulation check items can be found in Table 14.

Fifty six out of 57 participants in the Put-down Humor condition indicated that
the humorous comments were about the students in the proﬁles (vs. unrelated issues,
about each other, or about themselves). Almost all participants were aware that the
humorous comments referred to outgroup members. Again, no data was discarded to have

enough power for the analyses.

73

Table 14: Condition Means and Standard Deviations for the Manipulation Check Items

 

 

Put-down Humor No Humor

Item M SD M SD
How much humor did the
other group members use? 5.93 .92 4.44 1.07
How much did group
members laugh? 6.05 1.03 3.82 1.12
How ﬁrnny were the
comments? 4.49 1.55 3.74 1.40
How often did you smile? 4.84 1.84 3.98 1.63

' o

HOW Oﬁe" “1 you laugh' 4.53 1.84 3.44 1.65

 

Note: All ratings were made on a scale from 1 to 7.

Comutation of Variables

For all scales, I ran the same kinds of factor analyses as in Study 1. The results
turned out to be very similar to the ones from Study 1. I therefore computed the following
variables in exactly the same way: own positive mood (or = .88), own negative mood (or =
.80), shared experiences from own perspective (a = .88), own psychological safety (a =
.84), perceived intention to entertain (a = .94), perceived intention to antagonize (a =
.74), perceived concern (a = .59), perceived similarity to other group members (a = .90),
experienced personal cohesion (PCS, a = .92), perceived positive mood of other group
members (a = .90), perceived negative mood of group members (a = .81), shared
experienced among other group members (a = .91), perceived psychological safety
among other group members (a = .78), perceived similarity among other group members
(a = .92), perceived group cohesion (PCS, a = .93). The reliabilities of all scales were

satisfactory.

74

For the Team Psychological Safety scale (Edmondson, 1999) and for the Group
Enviromnent Questionnaire (Carron et al., 1985), however, a different factor structure
emerged. I still decided to compute the variables like in Study 1 in order to be able to
compare results across studies. The reliabilities of these scales are extremely low, though:
For psychological safety (TPS) u = .45, for experienced task cohesion a = .58, for
experienced social cohesion a = .59, for experienced negativity a = .67, for perceived
task cohesion a = .42, for perceived social cohesion a = .53, for perceived negativity a =
.66. Results including these scales need to be regarded with caution and are only reported
to present the complete set of results.

The Group Identiﬁcation Scale (GIS; Hinkle et al., 1989) that had not been used
in Study 1 was added in Study 2. Responses to the nine questions were submitted to a
principal axis factor analysis with varimax rotation. The eigenvalues-greater—than-l rule
and thescree plot indicated that two factors with eigenvalues of 2.62 and 1.7 9 emerged.
They explained together 48.98% of the variance. Items and factor loadings are displayed
in Table 15.

The ﬁrst factor consisted of items that indicated identiﬁcation with the group and
was therefore labeled Identiﬁcation. The second factor consisted of items that indicated a
lack of identiﬁcation and was therefore labeled De-identification. Because the results of
the factor analysis indicate that Identiﬁcation and De-identiﬁcation are different
constructs, I decided to compute two variables (experienced identiﬁcation and
experienced de-identiﬁcation) by averaging the appropriate items. Their internal
consistencies (aidemjﬁcau’on = .83 and ade_idemiﬁcation = .73) were acceptable and their

correlation was relatively low (r = -.20).

75

Table 15

Summary of Items and Factor Loadings for the Items from the Group Identiﬁcation Scale

 

 

 

Factor 1 Factor 2
Item Identification De-identification

I am glad to belong to this group. .87 -.17

. -.02
I identify with this group. 77
I think this group will work well .76 —.23
together.
I see myself as an important part .72 -.07
of this group.
I feel strong ties to the members .37 .001
of this group.
I feel uneasy with the members of -.03 .81
this group.
I feel held back by this group. '03 '68
I do not ﬁt in well with the other -. 16 .57
members of this group.
I do not consider the group to be -.11 .50
important.

Eigenvalue 2.62 1 .79
% of variance 29.14 19.84

 

Note: Boldface indicates highest factor loadings.

The same kind of factor analysis was run with the nine items from the GIS
measuring perceived identiﬁcation of others with the group. Again, two factors with
eigenvalues of 2.88 and 2.18 emerged and explained together 56.22% of the variance.
The rotated factor loading matrix showed the same pattern as the one in Table 15. Thus,
the two variables others ’ perceived identification (or = .86) and other ’s perceived de-

identiﬁcation (or = .77) were computed. They did not correlate at all (r = -.003).

76

Because in Study 2 the Group Identiﬁcation Scale had been added, four more
dependent variables were examined overall: experienced identiﬁcation, experienced de-
identiﬁcation, others’ perceived identiﬁcation, and others’ perceived de-identiﬁcation. It
is again essential to show how these variables correlate in order to ﬁnd out if they
measure the same underlying constructs.

As can be seen in Table 16, experienced personal cohesion (as measured by the
PCS), experienced task cohesion, experienced social cohesion (both measured by the
GEQ), and experienced identiﬁcation (as measured by the GIS) all correlate fairly highly
(.43 < r’s < .74). It seems that these measures tap into similar constructs, namely the
positive aspects of cohesion and identiﬁcation. They seem to show how much
participants like the group. Experienced negativity (measured by the GEQ) and
experienced de-identiﬁcation (measured by the GIS) also correlate highly (r = .65), but
do not correlate highly with the other variables (r’s < .24). They also seem to tap into the
same construct, which could be the negative aspects of cohesion and identiﬁcation. They
seem to show how much participants dislike the group. Although this pattern emerged, I
decided to analyze the variables separately because some of their reliabilities were very
low and in order to be able to compare results with those from Study 1.

A different pattern emerged for the measures of perceived cohesion and perceived
identiﬁcation, though (see Table 17). The variables did not correlate as highly as the ones
for experienced cohesion and experienced identiﬁcation. The only high correlation
coefficients could be found between perceived group cohesion (as measured by the PCS)
and perceived identiﬁcation (as measured by the GIS) on the one hand and between

perceived negativity (as measured by the GEQ) and perceived de-identiﬁcation (as

77

Table 16

Pearson Correlation Coefﬁcients for the Experienced Personal Cohesion Measures

 

Measure 1 2 3 4 5 6

 

1. Experienced Personal Cohesion (PCS)

* --
2. Experienced Task Cohesion (GEQ) '59

. . . .43* .29* --
3. Expenenced Socral Cohesron (GEQ)

- =I= - a: _ __
4. Experienced Negativity (GEQ) '22 '24 '18

a: :Is at: _ a: __
5. Experienced Identiﬁcation (GIS) '74 '67 ‘51 '22

- * - IF _ * _ * __
6. Experienced De-identiﬁcation (GIS) '21 '19 '12 '65 ~20

 

Notes: *p< .05. PCS = Perceived Cohesion Scale. GEQ = Group Environment
Questionnaire. GIS = Group Identiﬁcation Scale.

measured by the GIS) on the other hand. These measures could tap into two equivalent
constructs, namely now much the other group members seem to like and to dislike the
group. The reason for the low correlation coefﬁcients between the other variables could
either be their low reliabilities or that group members are better able to distinguish
between different kinds of cohesion when they judge others. There is no direct evidence

for either of these interpretations, however.

78

Table 17

Pearson Correlation Coefﬁcients for the Perceived Group Cohesion Measures

 

Measure 1 2 3 4 5 6

1. Perceived Group Cohesion (PCS)

* _-
2. Perceived Task Cohesion (GEQ) '42

* " -—
3. Perceived Social Cohesion (GEQ) '34 ~05

.. _ * __
4. Perceived Negativity (GEQ) '1 1 -02 23

31! 31‘ 3|: _ __
5. Perceived Identiﬁcation (GIS) '81 ‘48 '30 -08

- - * * _ __
6. Perceived De-identiﬁcation (GIS) '17 '03 '26 '58 -003

 

Notes: *p < .05. PCS = Perceived Cohesion Scale. GEQ = Group Environment
Questionnaire. GIS = Group Identiﬁcation Scale.

Correlations between Measures of Own Thoughts and Feeling and Others’ Thorﬂts and
filings

The correlation coefﬁcients between participants’ own thoughts and feelings and
their perceptions of their fellow group members’ thoughts and feelings were relatively
high (M = .55, range: 39-66, see Table 18). For each measure, they were about as high
as or higher than in Study 1. Especially the correlation coefﬁcients for similarity (r = .12
in Study 1 and r = .48 in Study 2) and experienced cohesion measured with the PCS (r =
.36 in Study 1 and r = .59 in Study 2) were much higher than in Study 1. Participants in
this study either perceived themselves to be more similar to their fellow group members,
which is supported by the higher correlation coefﬁcient between similarity to others and
similarity among others, or they did not distinguish very well between their own thoughts

and feelings and those of their fellow group members. However, the correlation

79

coefﬁcients are still low enough to conclude that two scales referring to one construct —
but from different perspectives — don’t measure exactly the same thing. Instead, they

seem to measure different, but related, constructs.

Table 18
Pearson Correlation Coefﬁcients between Measures of Own Thoughts and Feelings and

Measures of Others’ Thoughts and Feelings

 

 

Measures Pearson’s r
own positive mood and
perceived positive mood of other group members .66*
own negative mood and
perceived negative mood of other group members .48*
shared experiences and
perceived shared experiences among other group members .44*

own psychological safety and
perceived psychological safety among other group members .39*

perceived similarity to others and
perceived similarity among others .48*

experienced personal cohesion (PCS) and
perceived group cohesion (PCS) .59*

experienced task cohesion (GEQ) and
perceived task cohesion (GEQ) .63*

experienced social cohesion (GEQ) and
perceived social cohesion (GEQ) .53*

experienced negativity in group (GEQ) and

 

perceived negativity among others (GEQ) .62*
experienced personal identiﬁcation with the group and

perceived identiﬁcation with others .66*
Note: *p < .05

80

Did put-down humor affect experienced personal cohesion?

Put-down humor did not affect participants’ experienced cohesion (see Table 19).
Participants in the Put-down Humor condition felt the same amount of personal cohesion,
task and social cohesion, and negativity as participants in the No Humor condition. They
also identiﬁed as strongly with the group as participants in the No Humor condition.
Participants in the Put—down Humor condition tended to de-identify more strongly with
their group, though, than participants in the No Humor condition. To sum, put-down
humor did not have the expected positive effect on experienced cohesion. If anything, it
made participants feel uneasy and feel like they did not ﬁt into the group. Thus,

hypothesis 1.1 was not supported.

Did put-down humor affect perceived gout) cohesion?

Put-down humor had the expected positive effect on perceived group cohesion
(see Table 20). It increased participants’ overall sense of perceived group cohesion (as
measured by the PCS, t(112) = 3.81, p < .001), their perceived social cohesion (t(112) =
2.16, p < .05), and the perceived identiﬁcation of others with the group (t(112) = 2.90, p
< .01). Put-down humor did not affect participant’s sense of perceived task cohesion,
perceived negativity, and perceived de-identiﬁcation of others from the group. To
summarize, put-down humor seemed to indicate to participants that their fellow group
members were a cohesive unit, especially in regards to their social interactions, and that
the others strongly identiﬁed with the group. Thus, hypothesis 1.2 can be partly

supported.

81

Table 19

Means, Standard Deviations, and T-tests for the Experienced Personal Cohesion

 

 

Measures

Measure Condition M SD t-value p-value Cohen’s d
Experienced personal Put-down Humor 3.83 1.40
cohesion (PCS) No Humor 3.75 1.28 .29 .77 .06
Experienced task Put-down Humor 3.93 1.17
cohesion (GEQ) No Humor 4.11 1.12 -.85 .40 -.16
Experienced social Put-down Humor 3.46 1.07
cohesion (GEQ) No Humor 3.32 .86 .75 .45 .14
Experienced Put-down Humor 2.93 1.13
negativity (GEQ) No Humor 2.83 .91 .50 .62 .10
Experienced personal Put-down Humor 3.87 1.20
identiﬁcation (GIS) No Humor 3.82 1.01 .20 .84 .05
Experienced personal
De-identiﬁcation Put-down Humor 2.89 1.23
(GIS) No Humor 2.54 .81 1.78 .08 .37

 

Notes: All ratings were made on a scale from 1 to 7. PCS = Perceived Cohesion Scale.
GEQ = Group Environment Questionnaire. GIS = Group Identiﬁcation Scale. All degrees
of freedom = 112.

Did put-down humor affect participants’ own thoughts and feelings in the group?
Put-down humor did not affect participants’ own thoughts and feelings as
expected (see Table 21). Participants in the Put-down Humor condition felt about the
same amount of positive mood, negative mood, and shared experiences as participants in
the No Humor condition. They even felt slightly less psychological safety and less similar
to others than participants in the No Humor condition. However, participants in the Put-
down Humor condition (compared to those in the No Humor condition) believed that the
ﬁrst speakers in the group intended to entertain (t(112) = 9.07, p < .001) and that they

were more concerned about the group (t(112) = 2.05, p < .05). On the other hand, they

82

Table 20

Means, Standard Deviations, and T-Tests for the Perceived Group Cohesion Measures

 

 

Measure Condition M SD t-value p-value Cohen’s d
Perceived group Put-down Humor 4.64 1.31
cohesion (PCS) No Humor 3.80 1.00 3.81 < .001 .72
Perceived task Put-down Humor 4.25 .98
cohesion (GEQ) No Humor 4.26 .96 -.07 .95 -.01
Perceived social Put-down Humor 3.69 1.04
cohesion (GEQ) No Humor 3.33 .77 2.16 .03 .39
Perceived negativity Put-down Humor 3.16 1.09
(GEQ) No Humor 2.89 .76 1.52 .13 .29
Perceived
identiﬁcation of Put-down Humor 4.44 1.13
others (GIS) No Humor 3.91 .83 2.90 .004 .53
Perceived de-
identiﬁcation of Put-down Humor 2.97 1.23
others (GIS) No Humor 2.79 .83 .91 .36 .17

 

Notes: All ratings were made on a scale from 1 to 7. PCS = Perceived Cohesion Scale.
GEQ = Group Environment Questionnaire. GIS = Group Identiﬁcation Scale. All degrees
of ﬁ'eedom = 112.
also believed that the ﬁrst speakers intended to antagonize others with their comments
more (t(112) = 3.10, p < .01). Overall, it seems as though participants in the Put-down
Humor condition (as compared to participants in the No Humor condition) regarded the
statements of the other ostensible group members ambivalently: They did not assume that
the speakers had bad intentions, but they still felt like they made other group members
angry and insulted them. The slightly negative perception of the speakers’ comments
might have made participants feel less safe in the group and somewhat distant from the
group. Thus, hypotheses 2.1, 2.2, 3.1, 4.1, 5.1, and 6.1 were all not supported.

More speciﬁc emotions (fear, hostility, and guilt) were computed as in Study 1.

Unlike in Study 1, though, where participants in the Put-down Humor condition had felt

83

Table 21

Means, Standard Deviations, and T-Tests for the Measures of Participants’ Own

 

 

Thoughts and Feelings
Measure Condition Mean SD t-value p-value Cohen’s (1

positive mood Put-down Humor 2.45 .76

No Humor 2.58 .69 -.92 .36 -.18
negative mood Put-down Humor 1.44 .39

No Humor 1.39 .48 .56 .58 .1 1
fear Put-down Humor 1.42 .41

No Humor 1.45 .55 -.27 .79 -.06
hostility Put-down Humor 1.58 .72

No Humor 1.42 .62 1.30 .20 .24
guilt Put-down Humor 1.25 .47

No Humor 1.21 .54 .46 .65 .08
shared experiences Put-down Humor 3.70 1.41

No Humor 3.67 1.27 .14 .89 .02
psychological safety Put-down Humor 3.93 .97
(TPS) No Humor 4.07 .84 -.83 .41 -.15
psychological safety Put-down Humor 4.52 1.24
(own scale) No Humor 4.91 .98 -1.84 .07 -.35
intention to entertain Put-down Humor 6.26 .96

No Humor 4.28 1.34 9.07 <.001 1.70
intention to antagonize Put-down Humor 1.81 .89

No Humor 1.35 .70 3.10 .002 .57
concern Put-down Humor 4.57 1.19

No Humor 4.17 .89 2.05 .04 .38
similarity Put-down Humor 4.04 1.21

No Humor 4.42 .91 -1.89 .06 -.35

 

Notes: Mood ratings were made on a scale from 1 to 5. All other ratings were made on a
scale from 1 to 7. TPS = Team Psychological Safety scale. All degrees of freedom = 112.

84

more hostility than participants in the No Humor condition, no signiﬁcant differences in
emotions emerged in this study. Participants in both conditions felt about the same

amount of fear, hostility and guilt.

Did put-down humor affect perceptions of others’ thoughts and feelipgs?

As expected, put-down humor inﬂuenced participants’ perceptions of their fellow
group members’ thoughts and feelings (see Table 22). Participants in the Put-down
Humor condition regarded their fellow group members to be in a slightly more positive
mood, to share slightly more experiences among each other, and to be more similar to
each other (t(112) = 2.10, p < .05) than participants in the No Humor condition.
However, participants in the Put-down Humor condition did not perceive their fellow
group members to be in a less negative mood and to feel safer among each other than
participants in the No Humor condition. Overall, put-down humor seems to have had the
expected positive effect on some of participants’ perceptions of their fellow group
members’ thoughts and feelings. Hypotheses 2.4 and 4.2 were not supported. Hypotheses

2.3, 3.2, and 6.2 will be tested in more detail.

Did the New Cover Story have the Intended Effect? Congrarison of Study 1 with Study 2
Study 2 differed in two important ways from Study 1: First, the participant talked
before all other ostensible group members. It was assumed that this change in procedure
would make the participant less anxious (because there would be no pressure to be as
funny as the previous speakers) and that she would feel safer in the group (because she

was less likely to be evaluated negatively by other group members). If the intended

85

Table 22

Means, Standard Deviations, and T-Tests for Measures of Other Group Members’

 

 

Thoughts and Feelings

Measure Condition M SD t-value p-value Cohen’s d
perceived positive mood Put-down Humor 2.68 .79
0f other group members No Humor 2.43 .63 1.84 .07 35
perceived negative mood Put-down Humor 1.37 .45
0f other group members No Humor 1.39 .39 -.29 .77 -.05
perceived shared ex- Put-down Humor 4.26 1.39
Periemes among OthCTS No Humor 3.78 1.21 1.94 .06 .37
perceived psychological Put-down Humor 5.02 .94
safety among Others No Humor 4.87 .95 .85 .40 .16
perceived similarity Put-down Humor 5.06 1.07
among others No Humor 4.67 .92 2.10 .04 .39

 

Notes: Mood ratings were made on a scale from 1 to 5. All other ratings were made on a
scale from 1 to 7. All degrees of freedom = 112.

effects were achieved, one should see a decrease in the emotion fear and an increase in
psychological safety from Study 1 to Study 2.

Second, the participant was voted to be the recorder in the group. This should
make her feel more positive because she had a special role with higher status. One should
see an increase in positive mood from Study 1 to Study 2. Taken together, these two
changes in procedure were also assumed to increase the mean level of experienced
personal cohesion.

In order to test if these procedural changes had the intended effect on the
participant’s own thoughts and feelings and on her experienced personal cohesion, I ran
two 2(Study: 1 vs. 2) x 2(Put-down Humor: Yes vs. No) MANOVAs. For the ﬁrst
MANOVA, all measures of participants’ own thoughts and feelings were the dependent

variables. Results are displayed in Table 23. Neither the factor Study nor the factor Put-

86

down Humor nor the interaction was signiﬁcant. Thus, the procedural changes did not
have a strong global effect on participants’ own thoughts and feelings. However, the
univariate ANOVAs reveal that, as intended, the factor Study had a signiﬁcant effect on
participants’ positive mood, fear, shared experiences, psychological safety (as measured
with the TPS and with my own scale), and on perceptions of others’ concern. Participants
in Study 2 were in a more positive mood (M = 2.51 (SD = .72) vs. M = 2.28 (SD = .71), d
= .32), experienced less fear (M = 1.43 (.48) vs. M = 1.60 (.61), d = -.31), felt they shared
more experiences (M = 3.68 (1.33) vs. M = 3.32 (1.38), d = .27), felt more psychological
safety (as measured by the TPS: M = 4.00 (.91) vs. M = 3.60 (1.00), d = .42 and by the
own scale: M = 4.71 (1.13) vs. M = 4.32 (1.15), d = .34), and felt that other group
members were more concerned about the group interaction (M = 4.37 (1.07) vs. M = 3.88
(1 .39), d = .40) than participants in Study 1. Thus, the procedural change evidently had
the intended effect on participants’ own thoughts and feelings.

In the second 2(Study: 1 vs. 2) x 2(Put-down Humor: Yes vs. No) MANOVA, all
measures of experienced personal cohesion were the dependent variables. It was
expected that the factor Study would have an overall effect on the dependent variables.
Results are displayed in Table 24. The factor Study did not have a signiﬁcant global
effect on the full set of dependent variables, but the factor Put-down Humor did (F
(4,203) = 2.49, p < .05). The interaction was also not signiﬁcant.

More speciﬁcally, the factor Study affected experienced personal cohesion scores
on the Perceived Cohesion Scale (PCS; Bollen & Hoyle, 1990) (F (1,206) = 5.38, p <
.05). Participants in Study 2 felt more experienced personal cohesion than participants in

Study 1 (M = 3.79 (so = 1.34) vs. M = 3.38 (SD = 1.26), respectively, d = .32). None of

87

woven

.530 98 5:38: 58m ﬁrs 3256 z @383

Embeds 2: E woes—2: Lo: 33 coo—2 oZEwoZ .oom J u E BethD 63 .2 N we 0353232
.ozmuﬁm 33:?— Set vowecoeow one? mouse m ButaZHE—z ”$on

 

 

 

ﬁm; 9..— E. Tm.— omA co. mm: mm. cm. om. .m. mmE
m. w. m. moo. c. N. N. _. v._ m. mu. _ m x m
av 58::
in Na *9 inf w.~ NN m._ and n. o; 322 ~ 850?“?—
h. .m.w 0; m. and ...wd no;q _. mg sow ...w.m mm.m _ Amv beam
No. 3 I I .d .d S D H ..d .d m to 8.28m
m m m m .m. m m. m. m m m.
H 0 U ..u 9 H u. r. m.
e a w. m. K. me D. H A
.W. m u u we we mm ,M m
m. m m m m w o
o w H o n w. m
m... m. m m S w
n no U. S (\ 9
S O m. 0 S
m. u W.
n mt
039333 23.83232

 

mwesoom can mﬁwsoﬁn :30 mo 83302 8m 8593 we Embed/w oﬁte>ED use 0353292

mm 03mm.

88

the other measures of experienced personal cohesion were affected by the factor Study,
though, which might be due to their low reliabilities.

Furthermore, the factor Put-down Humor affected scores on experienced task
cohesion (as measured by the GEQ; Carron et al., 1985) (F (1, 206) = 6.40, p < .05).
Participants in the Put-down Humor condition experienced less task cohesion than
participants in the No Humor condition (M = 3.74 (1.15) vs. M = 4.12 (1.14),
respectively, (1 = -.33). This result needs to be regarded with caution due to the low
reliability of the measure in Study 2. None of the other measures of experienced personal
cohesion were affected by the factor Put-down Humor. To the extent that Study 2 is
viewed as a conceptual replication of Study 1, this analysis further bolsters the conclusion
that when put-down humor had effects on experienced personal cohesion, it was a
negative effect.

Overall, these results show that the procedural changes were successful in making
participants in Study 2 feel less anxious and more positive in the group situation than
participants in Study 1. Participants in Study 2 also felt slightly more attracted to the
group. However, these changes were not sufﬁcient to turn put-down humor from a

neutral/negative factor into a positive one.

Mediational Analyses

I hypothesized that the thoughts and feelings of other group members would
mediate the effect of put-down humor on perceived cohesion. In Study 2, only three of
the potential mediating variables differed between conditions — and not always

signiﬁcantly. These were perceived positive mood of other group members, perceived

89

Table 24

Multivariate and Univariate Analyses of Variance for Measures of Experienced Personal

 

 

 

Cohesion
Multivariate Univariate
Experienced Experienced Experienced Experienced
personal task cohesion social cohesion negativity
Source df F cohesion (PCS) (GEQ) (GEQ) (GEQ)

Study (S) 1 1.50 5.38“ 2.11 1.68 1.72
Put-down l 2.49* .18 6.40* .71 1.12
Humor (P)

S x P 1 .51 .68 1.92 .02 .16
MSE 1.71 1.30 .98 1.10

 

Notes: Multivariate F ratios were generated from Pillai’s statistic. PCS = Perceived
Cohesion Scale. GEQ = Group Environment Questionnaire.

Multivariate df = 4, 203. Univariate df = l, 206.

* p < .05

shared experiences among others, and perceived similarity among others. Perceived
psychological safety did not differ between conditions, unlike in Study 1. In order to be
able to compare results with Study 1 and because perceived psychological safety might
still contribute to the overall mediating effect, I included the latter variable in the multiple
mediation model (see Figure 5). This model shows both the direct effect of put-down
humor on perceived cohesion and the indirect effects via the four mediating variables.
The correlations between the four potential mediating variables ranged between .14 and
.58.

In order to test the effects, I bootstrapped them (for an explanation of this
procedure, see Study 1). All four variables together mediated the effect of put-down

humor on perceived cohesion. The total indirect effect was .39. This total effect is the

90

sum of all speciﬁc indirect effects, which in turn are the products of the unstandardized
coefﬁcients. The total and direct effects of put-down humor on perceived cohesion are
.83, p < .001, and .44, p < .01, respectively. These estimates are all very similar to the
ones from Study 1. However, when all four mediating variables are taken into account,
the effect of put-down humor is reduced from .83 to .44 and remains signiﬁcant. In Study
1, the effect had been reduced to non-signiﬁcance. Again, the difference between these
two effects is the total indirect effect (.39, see above). The 95% bias-corrected conﬁdence
interval for the total indirect effect is .05-.74. This CI does not contain 0 and therefore
indicates that the total indirect effect is actually different from 0. Thus, taken as a
collective, the four variables perceived positive mood, perceived shared experiences,
perceived psychological safety, and perceived similarity partially mediated the effect of
put-down humor on cohesion.

The point estimates and the 95% bias-corrected conﬁdence intervals of all
speciﬁc indirect effects are displayed in Table 25. Unstandardized coefﬁcients and their
standard errors are shown in Figure 2. The directions of all coefﬁcients are positive, as
predicted. Similar to Study 1, participants who heard put-down humor from their fellow
group members concluded that the others were in a more positive mood, shared more
experiences, felt safer, and were more similar to each other than did participants who did
not hear put-down humor. Only the three variables perceived positive mood, perceived
shared experiences, and perceived similarity can be regarded as mediators, though,
because their conﬁdence intervals do not contain 0. Perceived psychological safety did
not contribute signiﬁcantly to the effect above and beyond the effects of the other three

variables.

91

Table 25
Mediation of the Effect of Put-down Humor on Perceived Group Cohesion through
Perceived Positive Mood, Perceived Shared Experiences, Perceived Psychological

Safety, and Perceived Similarity

 

 

 

Bootstrapping
Product of Coefﬁcients
BC 95% CI
Indirect Effects Point Estimate SE Z Lower Upper
Perceived positive mood .13 .08 1.74 .004 .33
Perceived shared experiences .12 .07 1.72 .009 .32
Perceived psy. safety .06 .07 .84 -.067 .20
Perceived similarity .08 .05 1.58 .001 .26
TOTAL .39 .17 2.29 .05 .74

 

Notes: BC = bias-corrected. 5000 bootstrap samples. Z = value for the Sobel test (1982)

92

 
 

Perceived
positive mood

  
   

.54 (.11)

 
 

 

.25 (.13)
.47 (.24)
/—————-—V Perceived shared .25 (.07)
experiences \, .
44 ( 14) Percerved Group
Put-down Humor ° ' g Cohesion
\ Perceived psy.
safety .38 (.08)

   

     

.15(.l8)

    

Perceived .21 (.09)

Similarity

Figure 5: Mediation of Put-down Humor on Perceived Group Cohesion by Perceived
Positive Mood, Perceived Shared Experiences, Perceived Psychological Safety, and
Perceived Similarity. Displayed are unstandardized coefﬁcients (and their standard

errors).

93

DISCUSSION

In Study 2, certain conditions that were hypothesized to suppress the effect of put-
down humor on experienced personal cohesion were removed. Participants were asked to
talk ﬁrst in the student evaluation task, were voted the most reliable person in the group
by the rest of the members, and had the task of recording everyone’s evaluations. A
comparison between studies showed that these procedural changes actually made
participants feel less anxious in the group situation, as intended. On average, participants
in Study 2 were in a more positive mood, felt less fear, shared more experiences with
their fellow group members, and felt safer than participants in Study 1. However, the
procedural changes did not alter the null relationship between put-down humor and
experienced personal cohesion. The results in Study 2 closely resembled those from
Study 1. Put-down humor inﬂuenced neither participants’ own thoughts and feelings nor
their attraction to the group. But as in Study 1, it did inﬂuence how participants regarded
their fellow group members. They were perceived more positively and to be more
attracted to the group.

The pattern of results across the two studies clearly suggests that simply
witnessing the exchange of out-group directed put-down humor and laughter among
fellow group members is not sufﬁcient to bolster one’s own attraction to the group, nor is
such a reaction made more likely by reducing one’s social anxiety (viz. by reducing one’s
performance pressure and evaluation apprehension). In the general discussion, I will
consider further which conditions might be necessary and sufﬁcient to observe a positive

effect of put-down humor on personal cohesion.

94

It is noteworthy that the procedural changes in Study 2 led to a much higher
suspicion rate than in Study 1. In Study 2, 15.1% of participants realized that the
statements of their fellow group members had been recorded, whereas in Study 1, only
6.8% of participants had realized that. In Study 1, participants were explicitly told before
they started talking that they would not be able to hear any verbal feedback from their
fellow group members. Their fellow group members would be able to hear them, but not
vice versa. In Study 2, however, the participant talked ﬁrst and heard some background
noise, but no reactions from other group members. Then the second group member talked
(which meant, in fact, that the recording started playing). Participants could hear what the
group member said and how other group members reacted to the speaker’s comments. In
the No Humor condition, some group members cleared their throats, opened and closed
bottles and shufﬂed papers. In the Put-down Humor condition, group members reacted
with hearty laughter to all humorous comments by the speaker. In both conditions, the
background noise in the speaker’s recordings might have shown the attentiveness of other
group members. Participants might have wondered why all other group members
received some kind of feedback, but why they did not receive any. This might have
contributed to suspiciousness. It might also have made them feel somewhat distinct and
excluded from the group. This feeling of exclusion, of not being in the same situation as
the others, could offer another explanation of why the effect of put-down humor on
personal cohesion did not occur in Study 2 — an important point which will be elaborated
further in the general discussion.

One problem that occurred in Study 2 was the low internal reliability of the Team

Psychological Safety scale (Edmondson, 1999) and the Group Environment

95

Questionnaire (Carron et al., 1985). The ﬁrst was used to measure psychological safety
and the later to measure cohesion. In Study 1, the results of factor analyses showed that
the factor structures of both measured were different than expected, based on prior scale
development reports. Thus, I created composite scores of items that correlated highly
under the present conditions. In Study 2, the results of equivalent factor analyses turned
out to be rather different from those in Study 1. In order to be able to compare variables
across studies, I computed all variables in Study 2 like in Study 1. However, the variables
based on items from the resulting two measures turned out to have very low internal
consistencies. Thus, all results including these measures need to be regarded with
caution. The reason for the varying factor structures of both measures might be that the
instruments had been designed to measure characteristics in existing teams and not in ad-
hoc lab groups. Some of the items (for example, “Members of this group would be able to
bring up problems and tough issues”), although speciﬁcally adapted to these studies,
might not have made a lot of sense to participants and have led them to answer them
arbitrarily.

In the mediational analyses in both studies, four mediating variables were
included in a model with perceived group cohesion as the dependent variable. These were
perceived positive mood, perceived shared experiences, perceived psychological safety,
and perceived similarity among other group members. Recall that in Study 1, perceived
positive mood and perceived similarity were found to be signiﬁcant mediators in this
model. In Study 2, the same variables plus perceived shared experiences were found to

be signiﬁcant mediators. The overall pattern of results was quite similar across studies.

96

The lower bounds of the CPS were very similar to each other, only the interval for shared

experiences included 0 in Study 1, but not in Study 2.

97

GENERAL DISCUSSION

When peOple use put-down humor, they make fun of a speciﬁc target — a person
or object. Put-down humor should have a strong inﬂuence on group processes and
particularly on cohesion because of its inherently social nature. I hypothesized that put-
down humor would increase cohesion in a group if it referred to outgroup members and if
other group members showed their appreciation of the comment through laughter. In
Study 1, I found that these conditions were not sufﬁcient to increase personal cohesion.
However, participants regarded their fellow group members to be a more cohesive unit
when the group interaction contained put-down humor as compared to no humor. In
Study 2, I found that even after making participants feel safer in the group, put-down
humor did still not increase personal cohesion, but made other group members seem to
feel good in the group and to be more attracted to it.

Both studies have some obvious limitations: Participants are individuals who
believe themselves to be members of an ad-hoc group. Furthermore, all channels of
communication between the participant and other (ostensible) group members are
restricted due to the experimental procedure. It is thus difﬁcult to generalize the results to
existing groups and teams whose members interact face-to-face. However, this procedure
allowed me to manipulate put-down humor, to control for all other variables, and to ﬁnd
out how individual group members react when hearing put-down humor in a group
context. For this purpose, it was necessary to sacriﬁce naturalness. How much
naturalness we can sacriﬁce in the long run to study such an ephemeral phenomenon like
humor remains to be seen. In any case, these two studies are the beginning of what I hope

will be an interesting program of research about the effects of humor in groups.

98

Why did put-down humor not increasgersonal cohesion? Potential necessary and
sufﬁcient conditions

The experimental conditions created in Study 1 and 2 were not sufﬁcient for put-
down humor to increase personal cohesion. Because observational studies have
suggested, though, that put-down humor can increase cohesion (e.g., Terrion & Ashforth,
2002), the question arises as to just which conditions are necessary and/or sufﬁcient for
the effect to occur. Theory and research about humor markers, shared reality,
communication maxims, and group development suggest four additional conditions that
might be necessary.

Condition 1: The play frame of the put-down needs to be obvious.

According to Grice (1975), one can distinguish between on-record and off-record
communication. On-record comments can be taken literally and are direct and honest.
Off-record comments, however, cannot necessarily be taken literally and are indirect. All
humorous comments, among them put-down humor, can be considered off-record
because their content need not be taken literally. They are usually delivered with a play
frame. Communicators indicate that their comments are off-record by employing certain
markers. These can be verbal (e.g., exaggerations, changes in tone of voice) or nonverbal
(e.g., facial expressions, gestures). In the case of teasing, communicators have been
observed to use unusual vocalizations, a singsong voice, and unusual facial expressions to
indicate that they did not mean the comment seriously (Abrahams, 1962). Some of them
also laughed before or after the tease (Drew, 1987) — probably to indicate the play frame.

Such off-record markers signal that the comment should be taken in jest. In the

case of put-down humor, off-record markers include the humorous nature of the comment

99

and laughter. Without them, the comment would only be a put-down and regarded as a
hostile act. The presence and the nature of such off-record markers could strongly
inﬂuence the perception of the put-down humor. If there are many and obvious markers,
the humorous nature of the put-down should be more obvious. However, if there are few
and less obvious markers, the put-down could be taken more literally.

In this study, participants could not see nonverbal off-record markers of the
sources of the put-down humor because they were connected to the other (ostensible)
group members via headsets. They could only hear the voices of the other group
members. Furthermore, participants did not know the voices of their fellow group
members. Therefore, it might have been difﬁcult for them to pick up vocalizations as
unusual. It could also be the case that the speakers recorded for this study did not use
many audible off-record markers (e. g., a teasing tone of voice) to indicate the humorous
nature of their comments. In any case, participants might not have heard enough off-
record markers to realize that the put-down humor should have been taken in jest.
Participants might have heard the put-down, but not the humor. One indication of this is
participants’ answers to the manipulation check question “How funny were the
comments?” in the Put-down Humor condition. The average rating on a seven-point scale
was a 4.00 in Study 1 and 4.49 in Study 2, which is near the mid-point of the scale and
certainly not high — considering the fact that the manipulation was crafted to be very
strong. These results, in combination with the results from the other manipulation check
items, could also indicate that participants regarded the comments as humor attempts, but
did not ﬁnd them ﬁrnny. This, in turn, could mean that it was not put-down humor, but

put-down humor attempts that were manipulated. Another indication of too few off-

100

record markers is that participants in the Put-down Humor condition attributed a stronger
intention to antagonize to their fellow group members than participants in the No Humor
condition.

To summarize, participants in these studies might have perceived few and less
obvious off-record markers of put—down humor and have taken the comments more
literally. In a future study, the off-record markers and the resulting play ﬂame of the

humorous comments could be made more obvious or explicitly manipulated. Such

‘4‘
1. an

an

research could help establish whether having a high-level of off-record markers is E
necessary before put-down humor can have a positive effect on personal cohesion.

Condition 2: A shared reality needs to be created between group members.

According to Hardin and Higgins (1996), cognitive processes are shaped by social
activities. Furthermore, all social activities require some sort of social consensus.
Individuals can only interact with each other in a “shared reality”. If they share their
individual experiences with each other, subjective experiences can become objective and
a shared reality emerges.

Hardin and Higgins (1996) posit that a shared reality is a necessary requirement
for effective communication. Another requirement is that communicators take the
perspective of each other, in other words try to experience the situation as it is
experienced by others. If communicators take each others’ perspectives, they can verify
each others’ experiences. Usually, communicators wait for listeners to verify their
experiences. They phrase their utterances in a way that elicits feedback and wait for
listeners to show them that the message is understood. These signs of understanding do

not have to be whole words or sentences; often, simple “uh-huh”s and head nods sufﬁce.

101

If the communicator does not receive any feedback from listeners, he or she stops
transmitting information and tries to ﬁnd out why the listeners are not verifying his/her
experience. In the early stages of group development, the goal of establishing a shared
reality might be primary. Shared reality also seems to be necessary for groups to function
successfully.

In my studies, it was relatively difﬁcult for participants to create a shared reality
with other group members. In Study 1, participants listened to others talk and then
answered questions about the group. Before participants themselves talked, they were
told explicitly that the experimental procedure would not allow them to hear the reactions
of their fellow group members. Thus, they never received any feedback from others
throughout the study. In Study 2 also, participants did not receive any feedback from their
fellow group members although here they believed that this could have been possible.
Thus, participants were essentially left alone with their own subjective experience of the
situation. From their perspective, there was no consensus about the situation, or at least
no consensus that included them. Communication was not possible and thus a universally
shared reality in the group could not be created. This could also mean that participants
did not even feel like they were part of any group because their communications to and
experiences in the group were not validated. In summary, if an important initial goal of
the group was to create a shared reality, the experimental procedures made it difﬁcult or
impossible to achieve this goal. Instead of becoming a part of the group, the participant
remained apart from it.

In order to test if a common group experience would be sufﬁcient to create a

shared reality, I conducted a follow-up study (Study 3). Before participants engaged in

102

the personality evaluation task, they went through a team-building exercise with other
participants. Participants in the same session (i.e., up to four actual participants plus
confederates to have six members in the group) were asked to stand in a circle, facing
each other. Then the experimenter threw a balloon into the air. The group’s task was to
throw the balloon from one person to another so that the balloon would never touch the
ground. After some time, the experimenter added another balloon into the game, then a
third one, and so on, until the group had to keep ﬁve or six balloons in the air. This team
building exercise had been chosen because pilot testing indicated that it was fun for
participants, easy to conduct, and did not take up much time. Irnportantly, during the
whole team building exercise participants (and confederates) could see but not speak to
each other. This was done to prevent recognizing that the voices heard during the
following personality evaluation task were different from the voices heard in the team
building exercise. However, participants believed that the other group members in the
team building exercise would be the ones they would later talk to. I found that this
exercise signiﬁcantly increased overall levels of personal cohesion in both conditions, but
that the inclusion of such a team building phase was still not sufﬁcient for put-down
humor to increase personal cohesion.

Although these ﬁndings are quite tentative, they add to the earlier ﬁndings by
suggesting that the pre-existing level of cohesion in the group does not moderate the
relationship between put-down humor and personal cohesion (at least not within the
range of pre-existing cohesion levels examined here). In Study 2, a boost in cohesion was
achieved via positive evaluations from fellow group members and from a relatively safe

task procedure (in which the participant was not made as apprehensive about meeting

103

others’ expectations). In Study 3, a boost was achieved via shared experiences in a team
building exercise. Of course, this still leaves open the possibility that some other type of
shared experience may be necessary to create a shared reality between group members.

In future research, participants could be allowed to create a shared reality with
other group members so that they deﬁnitely feel like a member of the group. A shared
reality might be created by other group members’ explicitly acknowledging the presence
and the comments of the participant. Some form of communication between the
participant and other group members may well be necessary to create a shared reality.

Condition 3: Put-down humor needs to be relevant to the group or its task.

Grice (1975) described how people usually interact by describing the “cooperative
principle”. It states that communicators should make contributions that are required and
use accepted means of an interaction for their contributions. In other words, statements
should follow the purpose of a conversation if the communication is to be effective. The
speakers in the present Put-down Humor condition may have violated the cooperative
principle to a certain extent because they made some negative remarks that did not serve
the purpose of the conversation: The put—down humor may not have been seen as ﬂowing
from or assisting the group’s work on the personality evaluation task.

According to Grice (1975), the cooperative principle can further be divided into
four maxims: First, communicators do not say what they believe to be false (Truth).
Second, communicators make their contribution as informative as necessary for the
current conversation (Information). Third, communicators are relevant (Relevance).
Fourth, communicators are brief and avoid ambiguities (Clarity). As we all know from

experience, these maxims do not always describe how people usually communicate.

104

Instead, they describe the assumptions that listeners normally make about speakers’
statements. If the listener realizes that an utterance is not consistent with these maxims,
he or she will look for other meanings.

In these studies, speakers in the Put-down Humor condition (i.e., the recording
that the participant heard) may have violated at least two of the Gricean maxims: First,
the humorous comments of the speakers may not have been seen as highly relevant in the
situation. The task was to evaluate student proﬁles, and the put-down humor did not help
the evaluation process. It rather appeared to distract from the task, which is shown in the
lower ratings of task cohesion by participants of the Put-down Humor condition
compared to participants in the No Humor condition. Second, the humorous comments
were prolix and ambiguous. This might have led participants to look for other meanings
in the messages and to conclude eventually that the speakers were mean-spirited. That is,
the speakers who used put-down humor (i.e., the recorded statements) may have been
viewed as impolite, particularly for a ﬁrst conversation. People who don’t know each
other normally don’t insult strangers. If anything, they would rather say something
positive about others (Brown & Levinson, 1987). So if there was no apparent reason in
the situation to use to put-down humor, one plausible alternative conclusion could be that
something about the speakers themselves must have led them to use it. To summarize,
put-down humor may need to be more relevant for the group or its task to have positive
effects on personal cohesion.

In a future study, the identity of the outgroup members and their distinctiveness

from ingroup members should be more salient or manipulated. This could establish

105

whether the relevance of outgroup-member directed put-down humor moderates the
relation between put-down humor and personal cohesion.

Condition 4: Group members need to go through a speciﬁc sequence of different
kinds of put-down humor. '

As mentioned earlier, observational research suggests that put-down humor of
outgroup members can increase cohesion. Terrion & Ashforth (2002) concluded that put-
down humor of outgroup members increased trust and cohesion in a group of police
officers. However, this kind of put-down humor occurred relatively late in the
development of the group (after group members had started to put down themselves and
shared social identities). It could well be that put-down humor of outgroup members only
increases cohesion after the group has gone through several stages of put-down humor.
Group members might have to go through a sequence of different kinds of put-down
humor together before put-down humor of outgroup members increases personal
cohesion. Perhaps some (or all) group members need to put down themselves and the
whole group ﬁrst before they can use put-down humor about outgroup members for the
beneﬁt of the ingroup’s cohesion.

Conclusion: The following conditions might be necessary for put-down humor of
outgroup members to increase personal cohesion: The play frame and the relevance of the
put-down humor is clear to group members, and group members have a shared reality and
have already used put-down humor about themselves and shared social identities.
Clearly, these conditions were not satisﬁed in the present studies. However, participants
might well have perceived their fellow group members as if all these conditions had been

satisﬁed for them. That means the play frame of the put-down humor appeared to be clear

106

to other group members, which they indicated through their laughter. They acted as if
these comments were relevant to the task and as if they, as a group, had a shared reality.
They might even have appeared as if they had had past experience with each other.
Anecdotally, few participants left these studies asking the experimenter if the other group
members already knew each other or if they had been sitting in the same room. Thus, the
conditions might have been apparent in the rest of the group (excluding the participant),

but not in the full group (including the participant).

 

Can the lack of an effect be due to the gender of the participants?

All participants in these studies were women. This was mainly due to practical
considerations because most of the participants in psychology experiments at Michigan
State University are women. Unfortunately, data on put-down humor and gender is very
limited. If we assume that status and power inﬂuence how much put-down humor 3
person uses and that women are stereotypically of lower status than men, then men
should be more likely to use put-down humor than women. Men’s put-downs should also
be more hostile and contain fewer off-record markers according to this reasoning. Some
studies have suggested that men are more likely to tease than women (e.g., Lampert,
1996). Teasing is in many ways equivalent to put-down humor of an ingroup member.
But to date, no studies have contrasted how men and women deliver and react to put-
down humor. Put-down humor could be much more normative among all-male groups
than among all-female groups and mixed-gender groups. Men might also use put-down
humor more often than women to create bonds with each other (Tannen, 1990). However,
there is not enough evidence to support these speculations. To summarize, because put-

down humor might be more common in all-male groups, it might have a stronger effect

107

. ..
.w -

 

on personal cohesion in those groups. Future research needs to show if this is actually the
case.
Future research

In future research, it will ﬁrst be necessary clarify the nature of put-down humor
and to ﬁnd out what the essential ingredients of put-down humor are. In order to gain a
deeper understanding of the concept, I hope to conduct surveys or observational studies
to ﬁnd out how people use put-down humor in existing groups. More speciﬁcally, I will
examine what about the put-down humor makes it a put-down, how speakers indicate the
humorous nature of the comment, and how listeners indicate their appreciation of the
comment. Interviews with speakers could also reveal which goals they pursue when using
put-down humor in a group context. A better understanding of the concept will help
differentiate put-down humor from non-humorous put-downs and from other forms of
humor, which will eventually aid in the formulation of a general theory about the effects
of humor in groups.

As mentioned above, in future research it will also be necessary to test the
importance of several conditions to ﬁnd out when put-down humor of outgroup members
can increase personal cohesion. With the current paradigm, I plan to test in the near
future whether creating a shared reality is a moderating variable. Participants will be able
to communicate with other group members. As Hardin and Higgins (1996) suggest, it is
not necessary to exchange complete sentences to start building a shared reality; instead,
verbal feedback of the form “uh-uh” might be enough for participants to feel
acknowledged by others. This could lead them to develop a common understanding of the

situation and to feel like members of the same group as others. More speciﬁcally, in my

108

next study, I plan to have confederates listen to the participant’s statements and give her
feedback (through the headset) so that she realizes others hear her, understand her, and
have a similar perception of the situation. Furthermore, confederates will be instructed to
laugh at every humor attempt the participant makes so that the participant receives the
same reaction from group members as all other speakers.

If I can establish some conditions under which put-down humor will affect
personal cohesion, I plan to examine two potential moderators of the relation. First, I’d
like to test how the target of the humorous comment (an ingroup vs. an outgroup
member) moderates the relation. 1 hypothesize that outgroup targets will have a stronger
positive effect on cohesion than ingroup targets. Second, I’d like to systematically
explore the effect of laughter on the relation between put-down humor and personal
cohesion. I hypothesize that common laughter should increase cohesion to a stronger
extent than milder forms of appreciation or no appreciation of the humorous comment. In
the future, it will also be important extent this program of research to other, milder forms
of humor and compare their effects with those of put-down humor.

In the future, I also plan to extent the current model of put-down humor and
cohesion in groups. In these two studies, I explored how put-down humor might
inﬂuence cohesion. However, the relation between put-down humor and cohesion is
unlikely to be unidirectional; cohesion also inﬂuences the production of put-down humor.
The role of put—down humor is likely to be quite complex and embedded in dynamic
personal and interpersonal processes. Such complexities are fascinating to study, but
could not be examined with the current lab paradigm. I hope to conduct longitudinal

studies with existing groups to capture the interplay between put-down humor and

109

cohesion over time. For example, I could study groups of students completing a project
over one semester and record who uses put-down humor about whom, how listeners
react, and how the comment changes group members’ experiences in the group and their
interactions with each other. Such observational studies could help me in formulating a
richer theory about the effects of put-down humor in groups.

Applications

In her study about teasing at the workplace, Vinton (1989) observed that teasing
was used by employees for two reasons: to get the work done and to lessen status
differentials. She concluded that humor created bonds between employees and helped to
accomplish tasks. Although my studies focus on one speciﬁc kind of humor, put-down
humor, and Vinton examined all kinds of humor, my ﬁndings seem to contradict Vinton’s
to some extent. If anything, the put-down humor about outgroup members did not help to
get the work done. Participants who heard put-down humor tended to feel less task
cohesion and also perceived slightly less task cohesion among their fellow group
members. This observation is not to say that Vinton’s observations are invalid. They
simply help pointing out that the effects of put-down humor clearly depend on a variety
of situational factors.

There is no simple recipe for humor that one can follow when intending to
increase the cohesion in a group. Humor is a double-edged sword. Sharing put-down
humor about others outside the group can undercut the attraction of some group members
to the group. Unless all group members can be a part of the humor experience, humor

might not bond, but bite.

110

APPENDIX A

 

 

STUDENT PROFILES
First name: Ashley
Age: 21
Major: Sociology
Class: Junior
Work: bartender, babysitting

 

Clubs/Memberships:

Favorite movie or TV show:

Hobbies:
Hometown:
Number of siblings:
Current residence:
Career goals:

Interesting fact:
Heroes:

Democracy for America and my church
America’s next top model, MTV Real World
Talking on the phone, reading magazines

Ann Arbor, MI

5

off campus

Teach elementary students, preferbly 2nd grade
will get married next spring

Hillary Clinton

 

 

This proﬁle was evaluated by the ﬁrst speaker.

111

 

 

 

First name:
Age:
Major:
glass:

Work:
Clubs/M berships:

 

Favorite movie or TV show:

Hobbies:
Hometown:
Number of siblingg:
Current residence:
Career goals:

Interesting fact:
Heroes:

 

Megan
20

psychology
sophomore

none
Gamma Phi Beta Sorority and Phi Beta Delta Honor Society
Clueless, Sex in the City, Desperate Housewives, Orange County
hanging out with friends, playing softball

Hastings, MI

2

I live in a sorority house.

go to grad school

I made the Dean’s List every semester so far.
my grandfather, my “sisters”, Ms Piggy

 

This proﬁle was evaluated by the second speaker.

112

 

{- First name: I

‘f Age:

Major:

jf Class:

Work:
Clubs/memberships:

 

Favorite movie or TV show:

Edition

m:
Hometown:

' Number of siblings:
H. Current residence:

>_: Career goals:

: Interesting fact:

. Heroes:

   

22
Biosystems Engineering
Senior

I volunteer at Hurley Medical Center, Flint; part time at Subway
Biomedical Engineering Society
Muppet Show, Grey’s anatomy, Extreme Makeover: Home

dancing, traveling, hanging out with friends

Flint, MI

none

I share a house with ﬁiends.

Have a job that combines engineering and medicine
Have been to Europe three times

my uncle, he’s a paramedic

 

This proﬁle was evaluated by the third speaker.

113

 

 

 

 

First name:
Agg:
Major:
ﬁes:

Work:
Clubs/memberships:

 

Favorite movie or TV show:

1%:
Hometown:
Number of siblings:
Current residence:
Career goals:

Interesting fact:
Heroes:

 

Julie

20

Environmental Geosciences
Junior

part-time at Best Buy

Magicians Club

Friends, Will & Grace, Harry Potter of course
bike riding, yoga, shopping

Pittsburg, PA

3

off campus

Complete BS and teach high school science
volunteer at the Humane Society occasionally
Harry Potter

 

This proﬁle was evaluated by the participant.

114

 

APPENDIX B

TRANSCRIPTS OF RECORDINGS

Put-down Humor is printed in bold.

First Proﬁle: Ashley

Ok. . .um. . .well, we have Ashley and she's 21, she's a junior and her major is

sociology. . .um. . .I don't know, she has a lot of straps on her shoulders. It’s kind of
hippie-ish, I think. She's a bartender and a babysitter, so uh that's a nice combo. . .um. . .I
don't know if she's babysitting the drunks at the bar? (laughs)...anyway...um let’s
see, so... Outgoing... How outgoing is this person? She talks on the phone as a hobby.
And reading magazines? I don't know, she probably reads like US magazine or
something like that when she's babysitting. So I mean I kinda picture her sitting at

home. . .um. .. so I'm going to give her like a below average but I'll give her a 2 because
she does work at the bar. Um. . .how ﬁiendly is this person? She's not smiling in her
picture. . .so. . .ahem. . .um but she is a babysitter, so I guess that she has some kind of
interpersonal skills, so we'll give her, um, a 2 but I still think she's below average because
she's not smiling in the picture. How reliable? Ahem, babysitters are always reliable, so,
urn, it looks like she wants to be a teacher. You have to be able to trust a teacher. So, um,
I'll give her a 3 on that. Um. .. how moody is this person? She doesn't look very moody,
she really doesn't look like she has a lot of. . .um. . .ﬂuctuation in her personality at

all. . .ha-ha...if you look at the picture so...ha. Um we'll give her a two on that also.
Um. . .and how open minded? Well, she is the democracy for America; she's also
involved in the church. . .um...so I don't really know what to think about that. But she
does work at the bar, so I'm sure she's seen a lot of different things. . .like, people
throwing up and probably the same thing for the kids that she babysits and what
not, so. . .uh we'll give her a 3, I guess, on that. But she's probably gonna have to drop
the bar job and become a housewife pretty soon I would imagine. Um... since she's
getting married right out of graduation, it's like she went to school to get her misses
degree (laughs). Um, she likes Hilary Clinton so... we'll go ahead and give her a 3, she
might have some kind of hidden personality... we can't see. Okay, that's it. . .I mean.

Second Proﬁle: Megan
Alright, I have Megan and she’s 20. She’s a sophomore and her major is Psychology. She

is in the Phi Beta Delta honor society, which is a little intimidating, she looks a little
intimidating. I think she’s a bouncer at the bar that Ashley works at. (laughs)

115

And. . .uh. . .she’s also a member of the Gamma Phi Beta sorority. Um, her favorite

shows are Clueless, and she looks a little clueless so she’d ﬁt right in there. Ha—ha.

So, that’s about it, she lives in a sorority house. So, how outgoing is she? I would

probably give her a 2. She is in a sorority, but, ha, again the whole bouncer image is

just kinda throwin’ me off. She looks a little intimidating; I’d be a little scared to

run into her walking around campus. Uh, so I’d have to give her a 2 on that. How

friendly? Again, she lives in a sorority so she’s probably pretty ﬁiendly, and probably

gets to know people pretty well and she does say that her hobbies are hanging out with

ﬁiends and playing softball, so probably a 3. Yeah, we’ll go with a 3 on that. Uh,

reliable? I don’t know. Um, I guess a 3, above average ‘cause she’s in an honor society

but, so I guess you’ve got to be pretty reliable to be in that, and to live in the sorority

house. I don’t know, be pretty reliable. She wants to go to grad school, that’s pretty high

ambitions. And, she’s made the dean’s list every semester so far, which is another r-
scary thing about her ‘cause she doesn’t Iook very “dean’s list” material. So we’ll
say that. Moody? She doesn’t strike me as very moody, she looks like she has one mood,
which is just like scary. So I don’t really think she is that moody, she doesn’t strike me w
as someone who’s constantly up and down. Um, so, I would say below average, I guess

that is like less moody than anything else. And how open minded? She doesn’t strike me

as very open minded, she seems very, she has a sor. . ., you know she’s in a sorority, and

she’s in a honor society, and she hangs out with friends, but she doesn’t seem like she

really branches out or has a lot of variety. So, I’d give her a 2 for open-mindedness.

Yeah, that’s it.

Third Proﬁle: Kim

Ok, um, guess I’m gonna go. 80, Kim here, she’s 22, Bio-systems engineering senior.
Hanging out with friends, I don’t know, I think that she’s probably below average for
outgoing. And how friendly is she? Maybe if she smiled in her picture I would think
she’s a little friendlier. . .but. . .probably below average for friendly too. Umm, I love that
her favorite T.V. show is The Muppet Show, by the way. Ha, I didn’t even know the
Muppet show was still on...for real. Let’s see, she volunteers at a Medical Center,
like, I don’t know, I thought working was something you get paid for, seriously, not
volunteering. Okay, how reliable is she? Um engineering, oh well, to be engineering and
to study medicine you probably have to be pretty reliable, so she’s probably above
average for reliable. Urn, by the looks of her, that’s probably all she does actually. Goes
to school, studies, and doesn’t do a whole lot of anything else. And, oh let’s see, um, she
wants to have a job that combines engineering and medicine? I don’t know like
what? Designing the bottles for prescriptions? (laughs)...or something? Um, how
moody is she? She looks ridiculously moody in her picture. She’s like, “Please don’t take
this picture of me.” So, I’d say she’s well above average for moody. How open minded?
Um. . .I don’t have a clue. How about below average?

 

116

APPENDIX C

QUESTIONNAIRES

Items are listed in the same order as in the studies.

Mood (PANAS, Watson, Clark, & Tellegen, 1988)
Answer options: 1: very slightly, 2: a little, 3: moderately, 4: quite a bit, 5: extremely
Items:

Right now, I feel interested.
Right now, I feel distressed.
Right now, I feel excited.
Right now, I feel upset.

Right now, I feel strong.
Right now, I feel guilty.
Right now, I feel scared.
Right now, I feel hostile.
Right now, I feel enthusiastic.
Right now, I feel proud.
Right now, I feel irritable.
Right now, I feel alert.

Right now, I feel ashamed.
Right now, I feel inspired.
Right now, I feel nervous.
Right now, I feel determined.
Right now, I feel attentive.
Right now, I feel jittery.
Right now, I feel active.
Right now, I feel afraid.

Shared Experiences

Answer options: 1: not at all — 7: very much
0 I feel like I have shared some experience with other group members.
0 I feel like we have started to have some shared experience together.
0 The discussion so far has created a shared experience for us.

Psychological Safety (Team Psychological Safety scale, Edmondson, 1999)
Answer options: 1: very inaccurate — 7: very accurate

0 If you made a mistake on this group, it would be held against you.
0 Members of this group would be able to bring up problems and tough issues.
0 People of this group sometimes reject others for being different.

117

It is safe to take a risk on this group.

It would be difﬁcult to ask other members on this group for help.

No one on this group would deliberately act in a way that undermines my efforts.
Working with members on this group, my unique skills and talents would be
valued and utilized.

Psychological Safety (Own scale)
Answer options: 1: not at all -— 7: very much

How safe do you feel in the group?

How threatening is this group discussion for you?

How relaxed do you feel among other group members?

How comfortable do you feel with the other group members?
How much do you trust the other group members?

Perceived Intentions of Speakers
Answer options: 1: not at all — 7: very much

The speakers you have heard so far wanted to show how much they cared about
the group interaction.

The speakers you have heard so far intended to create a relaxing atmosphere.
The speakers you have heard so far wanted to make the other group members
laugh.

The speakers you have heard so far wanted to amuse the other group members.
The speakers you have heard so far wanted to make the other group members
angry.

The speakers you have heard so far wanted to insult the other group members.
The speakers you have heard so far wanted to distract group members from the
personality evaluation task.

Perceived Similarity to Others
Answer options: 1: not at all — 7: very much

How similar are you to the other group members in: your perception of the
situation?

How similar are you to the other group members in: your feelings in the situation?
How similar are you to the other group members in: your thoughts about the
situation?

How similar are you to the other group members in: your behavior in the
situation?

How similar are you to the other group members in: your goals for this
interaction?

How similar are you to the other group members in: your sense of humor?

Experienced Personal Cohesion (Perceived Cohesion Scale, Bollen & Hoyle, 1990)
Answer options: 1: strongly disagree - 7: strongly agree

I feel a sense of belonging to this group.

118

I feel that I am a member of this group. (not recorded in Study 1 due to technical
error)

I see myself as part of this group.

I am enthusiastic about this group.

I am happy to belong to this group.

This group is one of the best groups in this study.

Experienced Personal Cohesion (Group Environment Questionnaire, Carron,
Widmeyer, & Brawley, 1985)
Answer options: 1: strongly disagree — 7: strongly agree

I would not enjoy being part of the social activities of this group.

I am not going to miss the members of this group when the experiment ends.

I am unhappy with my group's lever of desire to work on the task.

I could become friends with some people in this group.

I do not like the style of work in this group.

For me, this group could become an important social group to which I belong.
Our group is united in trying to reach its goal.

Members of the group would rather go out on their own then get together as a
group.

We all take responsibility for any poor performance by our group.

Our group members would party together.

Our group members have conﬂicting aspirations for the group's performance.
Our group would like to spend time together after the experiment is over.

If members of our group had problems during the task, everyone would want to
help them so we can get back on track again.

Experienced Identification with the Group (Group Identification Scale, Hinkle,
Taylor, Fox-Cardamone, & Crook, 1989)

Note: only used in Study 2

Answer options: 1: strongly disagree — 7: strongly agree

I identify with this group.

I am glad to belong to this group.

I feel held back by this group.

I think this group will work well together.

I see myself as an important part of this group.

I do not ﬁt in well with the other members of this group.
I do not consider the group to be important.

I feel uneasy with the members of this group.

I feel strong ties to the members of this group.

Perceived Mood of Other Group Members (PANAS, Watson, Clark, & Tellegen,

1988)

Answer options: 1: very slightly, 2: a little, 3: moderately, 4: quite a bit, 5: extremely

Right now, the average other group member feels interested.

119

Right now, the average other group member feels distressed.
Right now, the average other group member feels excited.
Right now, the average other group member feels upset.
Right now, the average other group member feels strong.
Right now, the average other group member feels guilty.
Right now, the average other group member feels scared.
Right now, the average other group member feels hostile.
Right now, the average other group member feels enthusiastic.
Right now, the average other group member feels proud.
Right now, the average other group member feels irritable.
Right now, the average other group member feels alert.

Right now, the average other group member feels ashamed.
Right now, the average other group member feels inspired.
Right now, the average other group member feels nervous.
Right now, the average other group member feels determined.
Right now, the average other group member feels attentive.
Right now, the average other group member feels jittery.
Right now, the average other group member feels active.
Right now, the average other group member feels afraid.

Perceived Shared Experiences Among Other Group Members
Answer options: 1: not at all — 7: very much

The average other group member has shared some experience with the group.
The average other group member has started to have some shared experience with
the group.

The discussion so far has created a shared experience for the average other group
member.

Perceived Psychological Safety (Own Scale)
Answer options: 1: not at all — 7: very much

How safe does the average other group member feel in the group?

How threatening is the group discussion for the average other group member?
How relaxed does the average other group member feel in the group?

How comfortable does the average other group member feel in the group?
How much does the average other group member trust the group?

Perceived Similarity Among Other Group Members
Answer options: 1: not at all — 7: very much

How similar are the other group members in their perception of the situation?
How similar are the other group members in their feelings in the situation?
How similar are the other group members in their thoughts about the situation?
How similar are the other group members in their behavior in the situation?
How similar are the other group members in their goals for this interaction?
How similar are the other group members in their sense of humor?

120

Perceived Group Cohesion (Perceived Cohesion Scale, Bollen & Hoyle, 1990)
Answer options: 1: strongly disagree — 7: strongly agree
0 The average other group member feels a sense of belonging to this group.
0 The average other group member feels that she is a member of this group. (not
recorded in Study 1 due to technical error)

0 The average other group member sees herself as part of this group.

0 The average other group member is enthusiastic about this group.

0 The average other group member is happy to belong to this group.

0 The average other group member thinks that this group is one of the best groups

in this study.

Perceived Group Cohesion (Group Environment Questionnaire, Carron, Widmeyer, r
& Brawley, 1985) 1
Answer options: 1: strongly disagree — 7: strongly agree L

o The average other group member would not enjoy being part of the social ‘

activities of this group.

0 The average other group member is not going to miss the members of this group
when the experiment ends.

0 The average other group member is unhappy with the groups level of desire to
work on the task.

0 The average other group member could become friends with other group
members.
The average other group member does not like the style of work in this group.

For the average other group member, this group could become and important
social group to which she belongs.

o For the average other group member, this group is united in trying to reach its

goal

0 The average other group member would rather go out on her own, then get
together as a group.

0 The average other group member takes responsibility for any poor performance
by the group.

The average other group member would party with this group.
The average other group member thinks that there are conﬂicting aspirations for
the group's performance.

0 The average other group member would like to spend time with the group after
the experiment is over.

0 If members of this group had problems during the task, the average other group
member would want to help them so the group can get back on track again.

0 The average other group member believes that the group is not likely to stick
together after the experiment.

Perceived Identification of Others with the Group (Group Identification Scale,

Hinkle, Taylor, Fox-Cardamone, & Crook, 1989)
Note: only used in Study 2

121

Answer options: 1: strongly disagree — 7: strongly agree

The average other group member identiﬁes with this group.

The average other group member is glad to belong to this group.

The average other group member feels held back by this group.

The average other group member thinks this group will work well together.

The average other group member sees herself as an important part of this group.
The average other group member does not ﬁt in well.

The average other group member does not consider the group to be important.
The average other group member feels uneasy with the members of this group.
The average other group member feels strong ties to the members of this group.

Manipulation Check Items

In your own words, what do you think this study is about?

How was the interaction with the other participants?

Was there anything in the study that seemed strange or unusual to you?

Who were the people in the proﬁles? (Answer options: girls on a dating website,
students from the University of Michigan, students from Michigan State
University, girls in this group)

Who was the recorder in this group? (Answer options: I was the recorder.
Someone else was the recorder.) (only used in Study 2)

How much humor did the other group members use? (1: none at all — 7: very
much)

How much did the group members laugh during the interaction? (1: not at all — 7:
very much)

If there was any humor in the group interaction, what or who was it about?
(Answer options: about unrelated issues, about each other, about the students in
the proﬁles, about themselves)

How funny, do you think, were the comments by other participants? (1: not funny
at all — 7: very funny)

How often, do you think, did you smile during this experiment? (1: not at all — 7:
very much)

How often, do you think, did you laugh during this experiment? (1: not at all — 7:
very much)

122

BIBLIOGRAPHY
Abraharns, R. D. (1962). Playing the dozens. Journal of AmericaLFolﬁklore, 75, 209-220.

Aﬁﬁ, W. A. & Burgoon, J. K. (2000). The impact of violations on uncertainty and the
consequences for attractiveness. Human Communication Research, 26, 203 -233.

Baron, R. M. & Kenny, D. A. (1986). The moderator-mediator variable distinction in

social psychological research: Conceptual, strategic, and statistical considerations.

Journ_al of Personal and Social Psychology, 51, 1173-1182.

Berk, R. A. (2001). The active ingredients in humor: Psychophysiological beneﬁts and
risks for older adults. Educational Gerontology, 27, 323-339.

Berlyne, D. E. (1969). Laughter, humor, and play. In G. Lindzey & E. Aronson (Eds),
Handbook of social psychology (Vol. 3, pp. 795 - 852). Reading, MA: Addison-
Wesley.

Berlyne, D. E. (1971). Aesthetics and psychology. New York: Appleton-Century.

Bollen, K. A. & Hoyle, R. H. (1990). Perceived cohesion: A conceptual and empirical
examination. Social Forces, 69, 479-504.

Boxer, D. & Cortés-Conde (1997). From bonding to biting: Conversational joking and
identity display. J oumal of Pragmatics, 27, 275-294.

Brown, P. & Levinson, S. C. (1987). Politeness. New York: Cambridge University Press.

Byme, D. (1971). Can Wright be wrong? Let me count the ways. Representative
Research in Social chhology, 2, 12-18.

 

Carbajal, S. & Orihuela, A. (2001). Minimal number of conspeciﬁcs needed to minimize
the stress response of isolated mature ewes. Journ_al of Applied Animal Welfare
Science 4, 249-255.

 

Carron, A. V. & Brawley, L. R. (2000). Cohesion: Conceptual and measurement issues.
Small Group Research, 31, 89-106.

Carron, A. V., Widmeyer, W. N., & Brawley, L. R. (1985). The development of an
instrument to assess cohesion in sport teams: The Group Environment
Questionnaire. Journal of Sport Psychology, 7, 244-266.

Coser, R. L. (1959). Some social functions of laughter. Human Relations, 12, 171-182.

Dion, K. L. (2000). Group cohesion: From “ﬁelds of forces” to multidimensional
construct. Group Dynamics, 4. 7-26.

123

 

Drew, P. (1987). Po-faced receipts of teases. Linguistics, 25, 219-253.

Duncan, W. J. & Feisal, J. P. (1989). No laughing matter: Patterns of humor in the
workplace. Organizational Dynamics, 17, 18-30.

Edmondson, A. (1999). Psychological safety and learning behavior in work teams.
Administrative Science Quarterly. 44, 350-3 83.

Eisenberg, A. R. (1986). Teasing: Verbal play in two mexicano homes. In: B. B.
Schieffelin & E. Ochs (Eds), Languages socialization across cultures (pp. 183-
198). Cambridge: Cambridge University Press.

Eys, M.A., Hardy, J ., Carron, A.V., & Beauchamp, MR. (2003). The relationship
between task cohesion and competitive state anxiety. J ourn_al of Sport and
Exercise Psycholagy, 25, 66-76.

Paul, F. & Erdfelder, E. (1992). GPOWER: A priori,post-hoc, and compromise mower
analyses for MS-DOS (Computer program). Bonn, Germany: Bonn University,
Dept. of Psychology.

Festinger, L., Schachter, S., & Back, K. (1950). Social pressures in informal groups. New
York: Harper.

Fine, G. A. & De Soucey, M. (2005). Joking cultures: Humor themes as social regulation
in group life. Humor 18, 1-22.

 

Foot, H. (1991). The psychology of humor and laughter. In: R. Cochrane & D. Carroll
(Eds), Psychology and social issues: A tutorial text (pp. l-l3). London, New
York: F almer Press.

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

Forsyth, DR. (2006). Group dynamics (4th ed.). Belmont, CA: Thomson/Wadsworth.

Fraley, B. & Aron, A. (2004). The effect of a shared humorous experience on closeness
in initial encounters. Personal Relationships, 1 1, 61-78.

Francis, L., Monahan, K., & Berger, C. (1999). A laughing matter? The uses of humor in
medical interactions. Motivation and Emotion, 23, 155-174.

Freud, S. (1928). Humour. International Journal of Psychoanalysis, 9, 1-6.

Freud, S. (1960). Jokes and their relation to the unconscious. New York: Norton.
(Original work published 1905)

124

Gouaux, C. (1971). Induced affective states and interpersonal attraction. Journal of
Personalitj and Social Psychology, 20, 37-43.

Grice, H. P. (1975). Logic and conversation. In P. Cole & J. C. Moran (Eds), Smtax and
semantics III: Speech acts (pp. 41-58). New York: Academic Press.

Hardin, C. D. & Higgins, E. T. (1996). Shared reality: How social veriﬁcation makes the
subjective objective. In E. T. Higgins & R. M. Sorrentino (Eds), Handbook of
motivation and cognition: The interpersonal context (V 01. 3). New York:
Guilford.

Hayworth, D. (1928). The social origin and function of laughter. Psychological Review,
35, 367-384.

Hinkle, S., Taylor, L. A., Fox-Cardamone, D. L., & Crook, K. F. (1989). Intragroup
identiﬁcation and intergroup differentiation. A multicomponent approach. British
Journal of Social Psychology, 28, 305-317.

Hudson, K. (2001, July-August). Transforming a conservative company — One laugh at a
time. Harvard Business Review, 79, 45-53.

Janis, I. L. (1982). Groupthink. (2nd ed.). Boston: Houghton-Mifﬂin.

La Fave, L., Haddad, J ., & Maeson, W. A. (1976). Superiority, enhanced self-esteem and
perceived incongruity humor theory. In A. J. Chapman & H. C. Foot (Eds),
Humor and laughter: Theory. research and applications. New York: Wiley.

Lampert, M. (1996). Gender differences in conversational humor. In D. I. Slobin, J.
Gerhardt, A. Kyratzis, & J. Guo (Eds), Social interaction, social context, and
langaage (pp. 579-596). Hillsdale, New Jersey: Erlbaum.

 

Levine, J. (1969). Motivation in humor. New York: Atherton Press.

Long, D. L. & Graesser, A. C. (1988). Wit and humor in discourse processing. Discourse
Processes, 11, 35-60.

MacCoun, R. J. (1993). What is known about unit cohesion and military performance. In:
Sexual Orientation and US. Military Personnel Policy: Options and Assessment
(pp. 283-331). Santa Monica, CA: RAND.

MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002).

A comparison of methods to test mediation and other intervening variable effects.
Psychologﬁrl Methods, 7, 83-104.

125

MacKinnon, D. P., Lockwood, C. M., & Williams, J. (2004). Conﬁdence limits for the
indirect effect: Distribution of the product and resampling strategies. Multivariate
Behavioral Research. 39. 99-128.

Martin, R. A. (2001). Humor, laughter, and physical health: Methodological issues and
research ﬁndings. Psychological Bulletin, 127, 504-519.

Martin, R. A. (2002). Is laughter the best medicine? Humor, laughter, and physical
health. Current Directions in Psychological Science, 11, 216-220.

Norrick, N. R. (2003). Issues in conversational joking. Journal of Pragmatics, 35, 1333-

 

1359. l
Pogebrin, M. R. & Poole, E. D. (1988). Humor in the brieﬁng room: A study of the r—
strategic uses of humor among police. J oumi of Contemporary Ethnography, l7, 3; ‘3
183-210. } *
.4:

Preacher, K. J. & Hayes, A. F. (under review). Asymptotic and resampling strategies for
assessing and comparing indirect effects in multiple mediator models.

Ramachandran, V. S. (1998). The neurology and evolution of humor, laughter, and
smiling: The false alarm theory. Medical Hypotheses, 51, 351-354.

Robinson, D. T. & Smith-Lovin, L. (2001). Getting a laugh: Gender, status, and humor in
task discussions. Social Forces, 80, 123-158.

Schachter, S. (1959). The psychology of afﬁliation. Stanford, California: Stanford
University Press.

Sobel, M. E. (1982). Asymptotic conﬁdence intervals for indirect effects in structural
equation models. In S. Leinhardt (Ed.), Sociological methodology 1982 (pp. 290-
312). Washington, DC: American Sociological Association.

Sobel, M. E. (1986). Some new results on indirect effects and their standard errors in
covariance structure models. In N. Tuma (Ed.), Sociologcal Methodology 1986
(pp. 159-186). Washington, DC: American Sociological Association.

 

Stauffer, D. (1999). Let the good times roll: Building a fun culture. Harvard Management
Update, 4, 4-6.

 

Suls, J. M. (1972). Two—stage model for the appreciation of jokes and cartoons:
Information-processing analysis. In J. H. Goldstein & P. E. McGhee (Eds), m
psychology of humor (pp. 81-100). San Diego, CA: Academic Press.

 

126

Suls, J. M. (1983). Cognitive processes in humor appreciation. In P. E. McGhee & J. H.
Goldstein (Eds), Handbook of humor research: Basic issues (V 01. 1, pp. 39-57).
New York: Springer-Verlag.

Tannen, D. (1990). You just don’t understand. New York: Random House.

Terrion, J. L. & Ashforth, B. E. (2002). From “I” to “we”: The role of putdown humor
and identity in the development of a temporary group. Human Relations, 55, 55-
88.

Tragesser, S. L. & Lippman, L. G. (2005). Teasing: For superiority or solidarity? T_h§
Journal of General Psychology, 132, 255-266.

Veitch, R. & Grifﬁt, W. (1976). Good news — bad news: Affective and interpersonal
effects. J oumal of Applied Social Psychology. 6, 69-75.

 

Vinton, K. L. (1989). Humor in the workplace: It is more than telling jokes. Small Group
Behavior, 20, 151-166.

 

 

Watson, D. & Clark, L. A. (1994) PANAS -X. Manual for the Positive and Negative
Affect Schedule — Expanded Form. The University of Iowa.

Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief
measures of positive and negative affect: The PANAS scales. Journal of
Personality and Social PsychologyL54, 1063-1070.

Westenbroek, C., Snijders, T. A. B., den Boer, J. A., Gerrits, M., Fokkema, D. S., Ter
Horst, G. J. (2005). Pair-housing of male and female rats during chronic stress
exposure results in gender-speciﬁc behavioral responses. Hormones and
Behavior, 47, 620-628.

 

Wrightsman, L. (1959). The effects of small-group membership on level of concern.
Unpublished doctoral dissertation. University of Minnesota.

 

Wyer, R. S., Jr. & Collins, J. E., II. (1992). A theory of humor elicitation. Psycholog'cal
Review 99, 663-688.

 

Zillman, D. & Cantor, J. R. (1976). A disposition theory of humor and mirth. In A. J.
Chapman & H. C. Foot (Eds), Humor and laughter: Theory, research and
applications (pp. 93-116). New York: Wiley.

127

 

U S U IIIIIII L BBBBBB

11111121111111111111111111

   

02956 4048