I.
1

"’7 -—..—,—._ - .—
. . ',‘ Ir,
Vn

W, '“1. II

5-11.

 

u

.xa-

WW ‘m‘- .u. .y‘ _

.l.\-'v

 

 

. a ”.1411;
“A a '
Lhishsygsé“
314091 J .1.» .
I“

 

Jfitgv'
Niki?

'§'
1: ‘

‘13,?" '
‘11: prawn

~i ‘ ‘ ,
fig-vg‘lal'fu
53.? pi?

 

16.03:!
'1flfigqg‘-:~

‘I :‘a h‘,’ 1‘

‘i'tf’iu

9:! XE
.¢~%$'

- Kr-‘A «a.
.—-.-. ....
o

.'......

‘_‘ .--...-~
-u—v<'-- -
w- 4

an—s-‘
4-

W v
‘
-—-‘

.. .c—aw~ -- .»

”u...
Luau-—

“A...
"0..

~"-......‘-.‘

ka-c—o—
.W
.-

-_—u‘-
‘ -m
““-

< ‘0
453%
M2" .

“a-.. W'V' -
Wigs..-

u'hJ—‘i

 

 

h

.m.‘
.1532“: - ,..
.7; o
"J"
1 1 v- wf ~

i...

53 ~13" a

- .4...“

 

x
{I

:1
A”

'v- ‘9

 

.‘fiz.

. ... ‘

"....~..
. 4‘
"1'1...

"".'

«a

3%
3,...

,.—..m
. wuru.‘
' "A <—¢.:‘

7'
Ea"

THESIS

2

A
a .fi \ :m I“

v” a" .,

llllllllllllllllllllllHllllllllllllll‘llllllllllllllllllllll

3 1293 02058 204

This is to certify that the

thesis entitled

DOES IT PAY TO BE NICE? THE EFFECTS OF EXPECTANCY
VIOLATIONS, NORM TRANSGRESSIONS, AND RECIPROCITY
OPPORTUNITY ON ATTRACTION AND RECIPROCITY TOWARD

A FAVOR-CIVER

presented by

Amani C. Ei—Aiayli

has been accepted towards fulfillment
of the requirements for

M.A. Psychology

degree in

 

 

‘

W Xé/IJ.

Major professor

Date osj/m/oo

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

‘v‘i— '-

 

LIBRARY

Michigan State
University

 

 

 

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

 

DATE DUE DATE DUE DATE DUE
MN D 0 no

“6411912303

 

 

 

} M i 9 |2{0:0[3ߤ

I ififil394‘

 

 

 

 

 

 

 

 

 

 

 

 

 

DOES IT PAY TO BE NICE? THE EFFECTS OF EXPECTANCY VIOLATIONS,
NORM TRANSGRESSIONS, AND RECIPROCITY OPPORTUNITY ON
ATTRACTION AND RECIPROCITY TOWARD A FAVOR-GIVER
By

Amani G. El—Alayli

A THESIS

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

MASTER OF ARTS

Department of Psychology

2000

ABSTRACT
DOES IT PAY To BE NICE? THE EFFECTS OF EXPECTANCY VIOLATIONS,
NORM TRANSGRESSIONS, AND RECIPROCITY OPPORTUNITY ON
ATTRACTION AND RECIPROCITY TOWARD A FAVOR-GIVER
By

Amani G. El-Alayli

Previous research has found that when someone violates a norm and/or
expectation to provide a favor, the recipient tends not to like this favor-giver very much in
spite of the fact that the donor appears to have gone to great lengths to help (Clark &
Mills, 1979; Kiesler, 1966; Morse, Gergen, Peele, & van Ryneveld, 1977; Schopler &
Thompson, 1968; Worchel & Andreoli, 1974). The current investigation examined three
potential mechanisms that may account for this phenomenon: perceived unpredictability
of the donor, felt obligation to reciprocate, and psychological reactance due to threatened
freedom of reciprocity behavior. To test these mechanisms, expectancy, norm, and prior
knowledge of a reciprocity opportunity were manipuhted and their efiects on attraction
and reciprocity toward a favor-giver were examined. The unpredictability hypothesis
predicted perceived unpredictability of the donor to be negatively correlated with liking
and to mediate the effects of expectancy/norm violations on attraction. The felt obligation
and reactance hypotheses expected derogation of the violating favor-giver only to occur
when participants had no prior knowledge of an opportunity to reciprocate. Moreover,
the felt obligation hypothesis predicted highest reciprocity toward the violating favor-giver
whereas the reactance hypothesis predicted lowest reciprocity toward this expectancy and

norm violator. Partial support for each of the three mechanisms was obtained.

Copyright by

AMANI G. ELALAYLI

2000

ea $XIX$ 80
ca X®Ignm 0 {12s gex In)
© I would love to dedicate this thesis to my adorable niece and nephews,
Maggie, Zack, Sammy, Ryan, and Adam, for always being able to make me laugh! ©
05 X®gnm I &5g®x an)
as $X§X$ an)

iv

ACKNOWLEDGMENTS

I have many people to thank for the completion of this thesis. First I would like to
thank my undergraduate mentor, Dr. Robert Hymes, whose dynamic lectures first got me
excited about social psychology. He was a wonderful mentor and was very supportive in
helping me make the transition to graduate school.

I would also like to acknowledge my friends and colleagues, Deborah Abber, Keith
Niedermeier, and Matthew Gates. They have provided me with tremendous emotional
support and encouragement whenever it was needed. Special thanks goes to my good
friend and colleague, Peggy Chin, with whom I had the pleasure of beginning and
continuing my graduate career. Her support along the way was invaluable. Without these
fiiends, the adjustment to graduate school, the overcoming of obstacles, and the
completion of this thesis would have been much more dificult and much less enjoyable.

Lastly, I would like to thank my committee members, Drs. Lawrence Messé,
Norbert Kerr, and Allen McConnell, whose guidance and insightful comments left this
thesis, and my research abilities, much improved. My advisor, Dr. Messé, has been more
supportive than I could have ever expected an advisor to be. He was always there to
share his experiences and provide empathy, warmth, and encouragement whenever I was
faced with any concerns. His guidance and excellence in methodology have also helped
me to improve my research skills dramatically. He was also able to provide me with
constructive criticism while simultaneously increasing my confidence as a researcher.

Thanks, Larry!

PREFACE

As a student in a graduate course on group processes, I developed an interest in
the topic of social norms. I learned that when an individual is given a certain role title,
s/he behaves differently and is treated differently by others (e. g., Messé, Kerr & Stattler,
1992). Such findings led me to wonder how individuals might react to inconsistencies
between role expectations for another person and that person's actual behavior. More
specifically, I became interested in how individuals would react toward a person who
violates expectations and/or norms just to do them a favor. As a result, I turned to the
empirical and theoretical literature on this topic to see the extent to which such behaviors
had been studied. Although research in the area is not very extensive, under the guidance
of Dr. Lawrence Messé, I uncovered some interesting findings which led to the ideas

behind the current investigation.

TABLE OF CONTENTS

LIST OF TABLES ......................................................................................................... x

LIST OF FIGURES ..................................................................................................... .xii
CHAPTER 1

INTRODUCTION .......................................................................................................... 1

Importance of Norms and Expectations ................................................................... 1

Reactions Toward a F avor-Giver Who Violates Norms and/or Expectations ............ 4

Possible Mechanisms Underlying Derogation of a Violating Favor-Giver ............... 12

Unpredictability as a Negative Attribute ......................................................... 12

Felt Obligation to Reciprocate ....................................................................... 15

Psychological Reactance from Receiving a Favor ........................................... 21

Summary and Theoretical Considerations .............................................................. 29

Attraction Toward the F avor-Giver ................................................................ 29

Summary of Mechanisms ........................................................................ 29

Potential Moderators .............................................................................. 29

Sufficient Number of Violations .............................................................. 31

Threshold versus Additive Eflects ........................................................... 31

An Alternative Mechanism: Ulterior Motive Attributions ....................... 32

A Second Alternative Mechanism: Threat to Self-Esteem. ..................... 34

A Third Alternative Mechanism: Derogation by Association. ................. 36

Operation of More than One Mechanism. ............................................... 37

Reciprocity Toward the Favor-Giver ............................................................. 38

The Current Investigation ..................................................................................... 40
CHAPTER 2

METHOD .................................................................................................................... 43

Participants ........................................................................................................... 43

Overview .............................................................................................................. 43

Design .................................................................................................................. 43

Materials and Procedure ....................................................................................... 45

Introduction by Experimenter ........................................................................ 45

Work Task .................................................................................................... 46

Manipulation of Expectation. ......................................................................... 47

Manipulation of Norm. .................................................................................. 48

Rating and Payment ....................................................................................... 49

Manipulation of Knowledge of Opportunity to Reciprocate ........................... 49

Expectancy and Norm Manipulation Checks .................................................. 50

Measure of Supervisor Evaluation. ................................................................ 50

Introduction to Departmental Task ................................................................ 51

vii

Measure of Temporary Aflect ........................................................................ 51

Manipulation Check for Knowledge of Reciprocity ........................................ 52
Measure of Importance of Freedom of Reciprocity ........................................ 52
Measure of Reciprocity .................................................................................. 53
Measure of Attributions ................................................................................. 54
Measure of Belief in the Importance of the Norm of Reciprocity .................... 54
Measure of Intolerance of Ambiguity ............................................................. 54
Measure of Suspicion. ................................................................................... 55
CHAPTER 3
RESULTS ................................................................................................................... 56
Pilot Tests ............................................................................................................ 56
Believability of +1 Rating .............................................................................. 56
Construction of Norm of Reciprocity Beliefs Questionnaire ........................... 56
Scenario Studies to Pre-test Manipulations .................................................... 57
Manipulation Checks ............................................................................................ 58
Expectation Manipulation. ............................................................................. 58
Norm Manipulation. ...................................................................................... 58
Knowledge of Reciprocity Manipulation ........................................................ 59
Construction of Attraction and Unpredictability Factors ....................................... 61
Unpredictability as a Negative Attribute: Attraction Findings ............................... 62
Felt Obligation and Reactance: Attraction Findings .............................................. 68
Felt Obligation and Reactance: Reciprocity Findings ............................................ 72
Supplementary Findings ........................................................................................ 84
Temporary Afl‘ect .......................................................................................... 84
Attributions ................................................................................................... 84
Additional Findings ........................................................................................ 86
CHAPTER 4
DISCUSSION ............................................................................................................. 88
Summary of Main Findings ................................................................................... 88
Unpredictability Mechanism: Attraction Findings .......................................... 89
Felt Obligation and Psychological Reactance Mechanisms: Attraction
Findings ......................................................................................................... 90
Felt Obligation and Psychological Reactance Mechanisms: Reciprocity
Findings ......................................................................................................... 91
Potential Explanations for Unpredicted Results ..................................................... 93
CHAPTER 5
CONCLUSIONS AND IMPLICATIONS .................................................................. 104
APPENDIX A
Supervisor Record Sheets .......................................................................................... 108

APPENDD( B

Measures ................................................................................................................... 112
APPENDIX C
Factor Loadings from Principal Components Analysis ................................................ 127
APPENDIX D
Removal of Problem Data. ......................................................................................... 129
APPENDIX E
Analysis of Variance Results ...................................................................................... 132
REFERENCES ......................................................................................................... 135

LIST OF TABLES

Table 1.1 - Summary of Main Predictions .................................................................... 42
Table 2.1 - Listing of Tasks, Materials, and Manipulations ........................................... 44

Table 3. l - Attraction Toward the Favor-Giver as a Function of Expectation and
Norm with Unpredictability as a Covariate ................................................. 64

Table 3.2 - Attraction Toward the Favor-Giver as a Function of Expectation and
Intolerance of Ambiguity under Conditions of No Prior Knowledge ........... 66

Table 3.3 - Attraction Toward the Favor-Giver as a Function of Prior Knowledge
of an Opportunity to Reciprocate and Importance of Freedom of
Reciprocity with Unpredictability as a Covariate ........................................ 69

Table 3.4 - Attraction Toward the F avor-Giver as a Function of Expectation, Norm,
and Importance of Freedom of Reciprocity with Unpredictability as a
Covariate .................................................................................................. 70

Table 3.5 - Reciprocity Toward the Favor-Giver as a Function of Expectation and
Prior Knowledge of an Opportunity to Reciprocate ................................... 73

Table 3.6 - Reciprocity Toward the Favor-Giver as a Function of Norm and Prior
Knowledge of an Opportunity to Reciprocate ............................................ 73

Table 3.7 - Reciprocity Toward the Favor-Giver as a Function of Expectation, Norm,
and Prior Knowledge of an Opportunity to Reciprocate ............................. 76

Table 3.8 - Reciprocity Toward the Favor-Giver as a Function of Expectation and
Importance of Freedom of Reciprocity ...................................................... 78

Table 3.9 - Reciprocity Toward the Favor-Giver as a Function of Norm and
Importance of Freedom of Reciprocity ...................................................... 79

Table 3.10 - Reciprocity Toward the Favor-Giver as a Function of Expectation, Norm,
and Prior Knowledge of an Opportunity to Reciprocate for Recipients
Placing High Importance in Freedom of Reciprocity .................................. 81

Table 3.11 - Reciprocity Toward the Favor-Giver as a Function of Expectation, Norm,
and Prior Knowledge of an Opportunity to Reciprocate for Recipients
Placing Low Importance in Freedom of Reciprocity ................................... 82

X

Table A.l - Factor Loadings from Principal Components Analysis ............................ 127

Table A.2 - The Effects of Expectancy, Norm, and Knowledge of a Reciprocity
Opportunity on Attraction Toward a Favor-Giver .................................. 132

Table A.3 - Attraction Toward a Favor-Giver as a Function of Expectation and Norm
Without Unpredictability as a Covariate (AN OVA means) ....................... 133

LIST OF FIGURES

Figure l - Reciprocity Toward an Expectancy/Norm Violating Favor-Giver as a
Function of Recipients' Knowledge of an Opportunity to Reciprocate:
A Path Diagram. ...................................................................................... 103

Chapter 1

INTRODUCTION

How would you feel about a total stranger who violates a norm and an expectation
for the sole purpose of helping you? This thesis attempts to answer this question and to
explore some social psychological mechanisms that may be responsible for recipients'
reactions to such unusual favors. First, the importance of norms and expectations to our
society is discussed. Then the literature on reactions to expectancy and/or norm violating
favor-givers is reviewed. Following this, the potential mechanisms that may play a role in
eliciting these reactions are described extensively. And finally, the investigation that was

created to test these mechanisms is introduced.

Madame of Norms and Expggtations

"The capacity and the need to have, to make, to follow, and to enforce rules are of
cardinal importance for human social existence... For without rules, there can be neither
society nor culture..." (Fortes, 1983, p. 6). Do people conform, then, because they
understand the important role that social rules (i.e., norms) play in maintaining human
society?

Thibaut and Kelley (1959/1986) would probably have agreed with this suggestion.
They defined "norms" as behavioral rules that are accepted to some degree by most
everyone involved in a social context; They proposed that norms serve an array of
invaluable functions in our everyday lives. For example, norms provide a positive

alternative to the use of interpersonal influence in controlling behavior. Partly based on

research by Frank (as cited in Thibaut & Kelley, 1959/1986), the authors maintained that
norms encounter less resistance and produce even more economically and efliciently
certain consequences than would the application of interpersonal power. Some
subsequent research (e. g., Michelini & Messé, 1974) has also supported this notion. Also,
Wispé and Lloyd (as cited in Thibaut & Kelley, 1959/ 1986) found that both powerful and
weak individuals appear to benefit from the regularity and control that the mutually
acceptable rules provide. Therefore, people appear to profit more from and be less
reactive to conformity to norms than to the use of power tactics (cf. Michelini & Messé,
1974).

Thibaut and Kelley (1959/1986) also mentioned that norms are beneficial in that
they prescribe stable, automatic interaction sequences. Because norms are based on
consensus, most individuals will conform to them without the need for external control
mechanisms. Individuals may even feel intrinsic pleasure from conforming because norms
often reflect widely held values (e.g., see Hertel & Kerr, in press, as presented by Kerr,
2000). Consequently, norms tend to form automatic behavioral sequences, which
minimize conflict and uncertainty in interactions. They provide us with a sense of
predictability, which nukes us more comfortable living in our world. Automatic
interaction sequences also minimize process loss (Steiner, 1972) by relieving people of the
burden of constant vigilance, communication, and decision-making. In fact, it is suggested
that one reason people develop norms is to synchronize behaviors, thereby using their time
more efficiently and increasing productivity (Thibaut & Kelley, 1959/1986).

Another benefit of norms, as suggested by the independent research efforts of

Arsenain, Merei, and Wright in the 1940s (as cited in mm & Kelley, 1959/1986), is

that they increase group solidarity and security. Therefore, norms appear to help promote
unity, a characteristic that is highly regarded by society.

Considering all of the benefits of norms, it is easy to see why Fortes (1983) and
Thibaut and Kelley (1959/1986) concluded that these social rules are an important
mechanism for promoting the smooth fimctioning of society. If Thibaut and Kelley
(1959/1986) were correct in their assertions, then norms are, indeed, an integral part of
everyday life. They provide efficiency, productivity, predictability, regularity,
synchronicity, unity, and peace.

It is likely that these psychological benefits may accrue from other types of
expectations as well. For example, expectations that we have about another’s behavior
may be based on widely-held norms or on that person's previous behaviors. Either way,
fulfilled expectations produce the same benefits. Thus, it is no wonder that consistency of
behavior and conformin to norms are such powerfirl issues for most people. Perhaps
individuals follow norms and expectations, at least in part, because they believe in the
importance of such social mles and personal expectancies. Ifthis is the case, it is likely
that individuals might react negatively toward someone who violates norms or other types
of expectations. If so, would individuals even disparage the expectancy/norm violator

when they stand to benefit fiom the violations?

Reactions Toward a Favor-Giver Who Violates Norms and/or Expgtations
One of the goals of this research was to examine whether individuals might dislike

someone who violates a norm and an expectation just to do them a favor. Morse, Gergen,
Peele, and van Ryneveld (197 7) suggest that they would disparage this favor-giver, in
spite of the fact that they would benefit from his/her unusual behavior. This suggestion is
quite paradoxical, especially in the face of literature demonstrating that individuals appear
to be more attracted to those who provide them with help, especially when it is voluntary,
deh'berate, or costly (Gergen, Ellsworth, Maslach, & Seipel, 1975; Goranson &
Berkowitz, 1966; Greenberg & Frisch, 1972; Gross & Latane', 1974; Nemeth, 1970).
Unfortunately, as discussed below, Morse et a1. (1977) did not describe the conceptual
basis for this somewhat counterintuitive hypothesis. Nevertheless, their investigation did
find empirical support for their ideas.

Morse et a1. (1977) had students participate in a music-identification contest
during which their goal was to guess the names of songs and singers after listening to
short clips of music. In all conditions there was a ”quizmaster" who provided unsolicited
hints to the contestants to help them win a prize. Thus, the quizmaster always appeared to
do a favor for the participants. Morse et a1. manipulated expectations by having a
“supervisor" tell the students beforehand that their quizmaster had either usually provided
hints to contestants in past sessions (expectation confirmed condition) or rarely provided
hints (expectation violated condition). The quizmaster also either appeared to violate a
norm or not violate a norm. To do this, the supervisor either told the participants that it
was against the rules for the quizmaster to give hints in their session (norm violated

condition) or that it was not against the rules (norm not violated condition). Thus, the

expectation was based on consistency of behavior while the norm involved obedience to
legitimate authority.

The study yielded some interesting results. With regard to attraction,l Morse et
a1. (1977) found a significant interaction such that the participants were least favorable
toward the quizmaster when he violated both a norm and an expectation in providing the
hints. In other words, the quizmaster was liked the least when he violated both the
supervisor‘s rule against helping and participants' expectation that they would not get any
help because the quizmaster rarely provided it in the past. There were no difi‘erences in
ratings across the other three conditions. Therefore, Morse et al. found support for their
argument that a favor-giver would only be derogated when s/he behaves in both an
unexpected and countemormative manner to provide the favor. Because no main efi‘ects
were found, the authors concluded that neither an expectation violation nor a norm
violation alone would sufficiently lead to derogation of a favor-giver. Rather, they
asserted that both types of violations must be committed in order for derogation to
00011112

However, there may have been a potential confound in Morse et al.'s (1977) study

that reduced the reactivity of the separate norm and expectancy violation manipulations.

The authors stated that in South Afiica, where the study was conducted, hint-giving was a

 

1Consistent with most previous research on this topic (e.g., Morse et al., 1977), the terms
"attraction" and ”liking” are used throughout this thesis to refer to the favorability of the
nality ratings of the various favor-doers.

It is important to consider, however, that Morse et a1. (1977) do not define and
distinguish the concepts of norm and expectation beyond the different manipulations that
they used to operationalize these constructs. However, the present thesis distinguishes
between norms and other types of expectancies by using the term "expectation” to refer
only to personal expectations about specific individuals (e. g., based on that person' 5
previous behaviors) and the term ""nonn to refer to expectations about anyone who might
be in a particular situation where social rules are made salient.

S

typical feature of this type of record contest. As a consequence, participants may have
come to the study with an established normative expectation that the quizmaster would
provide them with hints. Perhaps confirmation of this initial informal norm prevented the
participants from viewing the favor-doer as especially unusual in the single-violation
conditions. After all, their initial expectation of help would have been confirmed in every
condition since hints were always provided. Thus, it could be that only a double violation
was suflicient to counteract their initial anticipation of receiving hints. In other words,
perhaps help was only perceived as purely unexpected or counternorrnative by the
participants in the double-violation condition. '

There is also another reason that hint-giving may have only been seen as unusual in
the double-violation condition. Perhaps when the quizmaster violated only the expectancy
or the nomr, this single violation was overlooked in the face of the normative or consistent
behavior that he also exhibited. When both expectation and norm were violated, however,
the quizmaster's hint-giving behavior may have appeared to be nothing but unusual.
Mthout any discounting information, the recipients may have been forced to attribute the
quizmaster's actions to some negative personality attributes. Therefore, the possibility
remains that a strong single expectancy or norm violation might be enough to cause
recipients of a favor to derogate their benefactor.

In fact, some research is consistent with this speculation (Clark & Mills, 1979;
Kiesler, 1966; Schopler & Thompson, 1968; Worchel & Andreoli, 1974). For example,
Kiesler (1966) found that individuals liked a stingy partner even better than a generous
one when role expectations suggested that being stingy was appropriate. Kiesler had pairs

of subjects participate in a word game with instructions to either cooperate or compete

with each other in order to win a monetary prize. Each subject was led to believe that his
partner had won most of the money by the end of the session. At this point, the partner
(really the experimenter) either shared or did not share his winnings with the subject.
Kiesler found that participants liked the partner most when he appeared to behave in
accord with the initial ”role requirement." Thus, partners who were told to cooperate and
indeed shared their winnings later and those who were told to compete and did not share
their winnings later were liked the most. Those partners who behaved out of role,
however, were derogated even when the participant received a tangible reward as a result.
Thus, even though the contest was over, initial norms prevailed and transgressions led to
the derogation of a favor-giver.

SchOpler and Thompson (1968) also conducted an investigation that found a single
expectancy or norm violation to be suflicient in reducing attraction toward a favor-giver.
Participants received a rose fi'om the experimenter during either a formal or an informal
interview. The authors suggested that the favor was only inappropriate when provided
during the formal interview. Participants in this condition tended to like the experimenter
less than those who received the rose in the informal condition but it was not reported
whether this difference was statistically significant. Thus, once again, a single norm
violation may have sufficiently decreased liking of a favor-doer.

Worchel and Andreoli's (1974) study is also relevant. In this experiment, the
researcher in charge gave the subject a stack of papers to sort while a confederate was
sitting right next to her. In the favor condition, the confederate helped the participant sort
the papers. Earlier in the session, the participant was either informed that she would later

be cooperating with the confederate to win a monetary prize, that they would be

competing for a prize, or that they would not be interacting later. Participants who
received help liked the favor-giver the least when they expected future competitive
interaction with her. Interestingly, this occurred in spite of the fact that the confederate
was apparently not aware of the nature of the filture interaction. Thus, even a violation of
a norm that had not yet been established led to decreased liking of the favor—giver.

Clark and Mills (1979) also found a single norm violation to be suflicient in
generating negative reactions. They manipulated whether recipients of a favor prefened a
communal versus exchange relationship with their benefactors. In a communal
relationship, the norm is to help partners only in response to their needs. Tit-for-tat
exchange, however, is only appropriate in exchange relationships. Participants were
unmarried male college students who were led to believe that they would be working with
an attractive partner who was either single or married. It was expected that when
participants were paired with the married partner, they would prefer an exchange
relationship. But when they were paired with the single partner, they would prefer a
communal relationship. In all conditions, participants were urged to help their partner
with a task by giving her some of his materials so she could complete a project more
quickly. In the favor conditions, she (really the experimenter) reciprocated by passing him
one of her research participation extra credits. In the no—favor conditions, she simply
thanked the participant. Clark and Mills found that when the woman was married, the
reciprocal favor resulted in greater liking of the partner than when no favor was provided.
But when the woman was single, participants liked her less when she provided the favor
than when she did not. In this latter condition, the favor was in violation of the expected,

or at least desired, normative conditions.

Although these investigations support the possibility that a single expectancy or
norm violation may cause recipients of a favor to derogate a donor, there are a few studies
inconsistent with this suggestion as well (Brehm & Cole, 1966; Morse, 1972; Suls,
Witenberg, & Gutkin, 1981). However, the lack of a derogation effect in these studies
can be explained.

In the Brehm and Cole study, which is discussed in greater detail later, participants
in both the favor and no—favor conditions were told to base their impressions of the favor-
giver solely on his responses to questions that would be asked and to avoid being
influenced by any external factors. Participants were strongly urged to be accurate and
unbiased in the relevant conditions. Consequently, it is not surprising that they were
successful in preventing the favor fiom affecting their impressions of the favor-giver in any
way.

The Suls et al. investigation found that a noun violating altruistic act led to the
greatest liking of the favor-giver. However, this investigation used observers rather than
recipients as raters of the likability of the favor-giver. Clearly, observers are likely to
generate very difl‘erent reactions than recipients. Moreover, the effect was only found for
adults. Grade-school children actually liked the norm violating altruistic favor-giver less
than someone who was just returning a favor.

In Morse's (1972) study, participants worked on a puzzle, which was introduced as
an ”eye-hand coordination task" in order to win a monetary prize. The experimenter
either told them that she would provide help later (by showing a picture of the puzzle) or
would not provide help. Expectancies were then either confirmed or disconfinned. When

help was given, there was a trend for participants to like the favor-giver more when it was

unexpected than when it was expected. However, this difference was not statistically
significant. Moreover, there was also a trend for participants to see the experimenter as
less helpful in the unexpected conditions when a favor was provided than when one was
not. Thus, although there is a little evidence to the contrary, the literature appears to
provide greater support for the notion that a norm or expectancy violating favor-giver
would not be liked very much.

Overall, there is quite a bit of evidence that a single expectancy or norm violation
might be sufficient in causing derogation of a favor-giver. Morse et al. (1977) were the
first to suggest that both an expectancy and norm violation must occur in order for
derogation to take place. Unfortunately, the bases for their predictions and the
explanation of their results were not very descriptive. In introducing their predictions
regarding attraction, they simply stated that it seemed unlikely that either an expectancy
violation or a norm violation by itself would sufficiently produce derogation of an aid-
giver. They suggested that the two events must co—occur but did not indicate why this
might be the case.

In their discussion of the results, however, they briefly mentioned two possible
reasons for this view. The first was that individuals may generally diser expectancy
violators and norm violators but that these separate tendencies alone may be insuficient to
counteract the liking that individuals would typically experience towards someone who
does them a favor. However, Morse et al. (1977) did not articulate the reasoning behind
this threshold of disfavor notion. The second explanation provided by Morse et al. was
that, in the fiamework of Thibaut and Kelley (1959/1986), unexpected counternorrnative

behavior may be seen as a more negative outcome tlmn expected counternorrnative

10

behavior. They suggested that only the former may be sufficiently negative to lead to the
derogation of a favor-giver. Morse et al. did not elaborate on this idea either, however.
For example, they did not discuss the reasons that counternormative behavior may be seen
as a negative outcome. Therefore, their reasoning remains somewhat nebulous due to a
lack of discussion regarding underlying mechanisms.

The purpose of the present investigation was to conduct a conceptual replication
of Morse et al. (1977) and to further extend their work by testing some possible
mechanisms that may account for people's reactions to expectancy and norm violating
favor-givers. Three potential mechanisms afl'ecting attraction toward such benefactors are
considered. The first idea is the view of unpredictability as an undesirable attribute (e. g.,
Kiesler 1973). The other two are based on Gouldner’s (1960) "norm of reciprocity."
Specifically, the second mechanism involves the possible consequences of a feeling of
obligation to reciprocate a favor (U ehara, 1995). The third involves psychological
reactance experienced by the recipient due to expectations of reciprocity (Brehm, 1966;

Brehm & Brehm, 1981).

ll

Possible Mechanisms Undeming Demation of a Violating Favor-Giver
Unpredictability as a Negative Attribute
One potential mechanism that could account for the derogation of a violating
favor-giver is the perception that the favor-giver is unpredictable. As Kiesler (1973)
writes:
In considering the functioning of norms, theorists have made several assumptions:
(it) Persons have a need for predictability in social interaction; (b) norms filnction
to increase the predictability of behavior; and (c) norm violators are perceived as

unpredictable. Thus, disliking may result from a norm violator‘s unpredictability.
(p.354)

It seems reasonable to assume that this same logic would also apply in
understanding reactions to violators of other types of expectancies as well. For example,
expectancies based on someone's previous trends in behavior surely increase predictability,
probably to at least the same extent as normative standards do. Thus, one possible reason
that individuals may derogate a favor-giver who violates both a norm and an expectation is
that the person would be perceived as being highly unpredictable. People may simply
dislike unpredictability, regardless of whether its consequences are favorable or
unfavorable. Gergen and Jones (1963) found, in fact, that a (normal) stimulus person was
disliked as a function of his unpredictability and not the consequences of his unpredictable
behavior.

Kiesler ( 1966; 1973) made use of various theories to explain why individuals might
perceive expectancy/norm violators as being unpredictable and why people might dislike
unpredictable others. Thibaut and Kelly's (1959/1986) social exchange theory can provide
one potential explanation. According to this theory, much behavior is guided by a goal to

maximize benefits and minimize costs. According to Kiesler, the theory also implies that

12

irregularity resulting fi'om expectancy and norm violations is a potential cost. Thus, an
individual is likely to be less satisfied with an outcome and the person responsible for it
when violations occur.

Kiesler (1966; 1973) also discussed an alternative explanation based on Goflinan's
(1959) role-theory perspective. According to Gofiinan, norms are developed so that
individuals can come to a consensus about the definition of a situation. This allows them
to have a sense of predictability and control over each other. In violating a norm, then, a
person would be forcing others to reject their definition of the situation. These individuals
would then diser the person who was responsible for their lost sense of predictability.

In addition, Kiesler (1966) briefly suggested that Carlsmith and Aronson's (1963)
version of expectancy theory could provide a third potential explanation. According to
Kiesler's interpretation of this perspective, dissonance is aroused when an individual
encounters an unexpected event. prerceptions of the event itself cannot be easily
distorted, then the unpredictable person responsrhle for the event should be liked less.
Kiesler suggested that one example of an event that is difiicult to distort is receiving some
sort of unexpected tangible reward. Kiesler did not discuss the reasoning behind these
ideas, however.

There is some evidence that norm violators are perceived as unpredictable and are
thus liked less. Kiesler (1973) had participants listen to a tape-recorded conversation and
evaluate the personality of one of the speakers. This speaker was a college student who
was discussing her inability to concentrate on studying due to the fact that her boyfiiend
had not called her. She was either speaking to a dormitory counselor or to a faculty

member. In the former case, her behavior was appropriate, but in the latter, it was

13

counternorrnative. Participants perceived the student as behaving less appropriately and
liked her less when she was speaking to a professor than when she was speaking to a
counselor. This finding is consistent with the notion that perceived unpredictability is
responsible for the derogation of someone who violates a norm. Kiesler found the reverse
pattern when she set up a situation where unpredictability was preferred, but, in these
conditions only, participants were asked to avoid imposing any value judgments in
describing the student's traits. Perhaps, then, participants overcompensated for bias by
liking the norm violator more than the speaker who did not violate a norm.

However, overall the research provides little support for the suggestion that
unpredictability mediates the derogation of a norm and/or expectancy violator.
Unpredictability was not examined as a mediator in Kiesler's (1973) study because it was
inferred from the manipulations rather than measured. Moreover, in Kiesler's (1966) role
requirement study described earlier, unpredictability was measured but not found to be a
significant mediator of the effects of role inconsistencies on derogation. It is possible,
however, that because unpredictability was measured using only a single item, actual
mediation may have gone undetected due to a lack of sensitivity of the measure.
Thaefore, while there is no strong evidence in the literature for the mediation of
unpredictability, it is still a plausrhle mechanism to account for derogation. It is reasonable
to expect that derogation of a favor-giver may occur to the extent that the individual
behaves in an unpredictable manner.

Both Kiesler (1973) and Kiesler (1966) appeared to deal with norms, rather than
other types of expectations. However, following Kiesler's logic, perceived unpredictability

could potentially result from the violation of either a norm, an expectation, or both, as

14

long as the degree of unpredictability exceeds some threshold of tolerance. Thus, while
Morse et al. (1977) suggest that both a norm and an expectation violation must occur, the
unpredictability mechanism suggests that a single violation might sufice. As mentioned
earlier, there is a possrbility that the manipulations in Morse et al.'s study were not strong
enough to separately lead to the degree of unpredictability necessary to cause derogation
of the target. Thus, in Morse et al.'s study, both the expectation and the norm violation
may have been needed to exceed participants’ threshold of tolerance for idiosyncrasy.

Thus, the unpredictability notion provides one potential explanation for the reason
individuals tend to derogation expectancy and/or norm violating favor-givers. Whereas
the unpredictability notion only makes predictions regarding attraction toward a favor-
giver, the next two theories, felt obligation and psychological reactance, also make
predictions regarding reciprocity.
Felt Obligation to Reciprocate

As mentioned, there is a second mechanism that could potentially explain the
dislike of a noun and/or expectancy violating favor-giver. It involves feeling a sense of
obligation to reciprocate favors (see Fisher et al., 1981; Gouldner, 1960; Greenberg, 1980;
Uehara, 1995). Gouldner's (1960) "norm of reciprocity" is of primary relevance here.
Part of this norm is a universal moral obligation to return any favors that one receives.
Gouldner suggests that the degree of obligation that one feels depends on the perceived
value of the benefit received. Thus, when an individual receives a large favor, s/he would
likely feel a strong sense of duty to return the favor. This may occur, for example, when
the favor is perceived as out of the ordinary or unjustified. Feeling an uncomfortable

sense of indebtedness, the individual would surely be motivated to reduce perceptions of

15

obligation by reciprocating the favor. According to Uehara (1995), moral obligations to
return a benefit are at least as motivating as the desire to get back when you give. In fact,
an individual may feel so obligated to another person upon the receipt of a large favor that
s/he might even resent the favor-giver until reciprocity can take place.

This reasoning can explain why individuals might derogate someone who violates a
norm and/or expectation to do a favor for them. In essence, a favor is likely to be
perceived as greater when it appears that the benefactor violated a norm or expectation
violations in order to provide it. In engaging in at least one violation, it may appear that
the favor-giver really went out of his/her way to help. Thus, the recipient would feel a
very strong sense of obligation to reciprocate the favor. This could then result in
resentment and dislike of the favor-giver, particularly if the recipient sees no opportunity
to reduce his/her sense of obligation by paying the favor-giver back.

Uehara (1995) reviews various lines of research that provide support for the
experience of felt obligation and its consequences. First, she cites some evidence for the
existence and prevalence of the norm of reciprocity. Specifically, research has found
support for the existence of the norm among various types of individuals and in various
regions of North America (Antonucci, Fuhrer, & Jackson, 1990, Greenberg & Shapiro,
1971, Ingersoll-Dayton & Antonucci, 1988, Schrerher & Glidewell, 1978, Shumaker &
Jackson, 1979, Uehara, 1987, Wentowski, 1981, as cited in Uehara, 1995). There is also
research consistent with the notion that people feel a sense of obligation to reciprocate
when they receive favors from others. For example, individuals have been found to be
more likely to request help when they believe that they will have a chance to reciprocate

the favor than when they perceive no opportunity to do so (DePaulo, 1978, as cited in

16

Fisher et al., 1981; Greenberg & Shapiro, 1971, Morris & Rosen, 1973, as cited in Fisher
et al., 1981 and Uehara, 1995; Gross & Latané, 1974). Moreover, Castro (1974) found
that individuals are also more likely to request firture aid if they had been able to repay
benefactors in the past. It was also shown that help that is not as costly to the benefactor
is more likely to be sought out or accepted (Castro, 1974; DePaulo & Fisher, 1980, as
cited in Fisher et al., 1981). Thus, people have such strong beliefs in the norm of
reciprocity that they may not even want to request or accept favors unless they believe
that they will be able to reciprocate them.

Moreover, other research that Uehara discussed suggests that not only do we
avoid "overbenefiting" fi'om a relationship, but we also tend to perceive ourselves as
”underbenefiting," even if reality must be distorted in order to do so (Antonucci, 1985,
Beckman, 1981, O'Connell, 1982, 1984, Palo—Stoller, I985, Poulin, 1993, Weinberg,
1992, Wentowski, 1981, as cited in Uehara, 1995). Weinberg describes this phenomenon
as a "debt-reducing bias. " Consistent with this idea, there is also evidence that individuals
are more distressed when they are overbenefiting than when underbenefiting, even though
the latter is also an inequitable relationship (lngersoll-Dayton & Antonucci, 1988, Roberto
& Scott, 1984, Van Tilburg, 1992, as cited in Uehara, 1995). Thus, the felt obligation
mechanism may be more plausrhle than any negative feelings that may result from inequity
or inequality.

In addition, other research suggests that negative consequences may result from
feelings of obligation. For example, when an individual does not expect to have a chance
to reciprocate, s/he tends to feel frustration or other adverse feelings, including resentment

toward the favor-giver (Brehm & Brehm, 1981; Greenberg, 1980; Fisher, Nadler, &

17

Whitcher-Alagna, 1983, Greenberg & Shapiro, 1971, Shumaker & Jackson, 1979, as cited
in Uehara, 1995). Moreover, individuals have been shown to have more positive feelings
toward a favor-giver when they are provided with the opportunity to reciprocate than
when they have no such opportunity (Castro, 1974; Gross & Latané, 1974). Meeting
reciprocity obligations, and even the mere ability to reciprocate, have also been associated
with increased self-esteem, greater pride, greater life satisfaction, better emotional and
social firnctioning, and more positive personal well-being (Black, 1985, Hays, Clmuncey,
& Tobey, 1990, Israel, Hogue, & Gorton, 1983, Rock, 1987, Stevens, 1992, Wentowski,
1981, as cited in Uehara, 1995). Another investigation even found that a helper who
asked for something in return was liked more than one who did not (Gergen et al., 1975).
Therefore, there is clear support for the idea that felt obligation may generate negative
afl‘ective consequences and that these potentially negative reactions are reduced when
recipients reciprocate the favor, or at least believe that they are able to do so in the firture.

All of the above research is consistent with the notion that a sense of felt obligation
may be responsible for the derogation of an expectancy and/or norm violating favor-giver.
It is possible, then, that resentment toward a favor-giver occurs when an expectancy
and/or nonn is violated because the favor is seen as very big under these circumstances,
thus inducing suflicient felt obligation to trigger negative reactions.

It is important to note that participants in previous research were typically not
aware ahead of time that they would be given a chance to reciprocate. They may have
derogated the violating favor-giver out of fiustration or resentment fi'om drinking that they
would not be able to return the favor, thereby reducing their sense of obligation. Perhaps

if the violating favor-giver was evaluated afier participants were given a chance to

18

reciprocate, or if they merely knew ahead of time that they would have such an
opportunity, derogation would not have occurred. It is reasonable to expect that under
these conditions the recipients would have known that their sense of obligation was only
transient. Thus, there would no longer be any hostility toward the favor-giver. Under
these circumstances, attraction level might have even been the same, if not greater, toward
the violating favor-giver because the favor was so out of the ordinary.

In fact, in the study by Suls et al. (1981), uninvolved adult observers liked the
favor-giver more as favors became greater and more unexpected. As outside observers,
the raters in this case were able to observe the acts of kindness without being subject to
feeling obligated to the donor. Thus they were able to evaluate the favor-giver more
objectively. This circumstance might be somewhat similar to a recipient's experience if
s/he knows that feelings of obligation would only be temporary. As mentioned earlier,
there is some evidence that individuals are more favorable toward a favor-giver when they
are given a chance to reciprocate than when they have no such opportunity. Therefore,
according to felt obligation theory and research, attraction toward a norm and expectancy
violating favor-giver should differ as a function of whether or not the beneficent has
knowledge about a chance to reciprocate. Again, derogation should not occur when the
recipient has prior knowledge that s/he will be able to pay back in kind.

The felt obligation notion can also be applied to reciprocity behavior. As indicated
ealiier, Gouldner (1960) believes that felt moral obligation is positively correlated with
favor magnitude. As such, larger favors should result in greater felt obligation to

reciprocate and thus greater reciprocity.

l9

Tesser, Gatewood, and Driver (1968, as cited in Fisher et al., 1981) have
found, in fact, that perceived indebtedness increases with the value of the favor received.
Moreover, many other investigations have found reciprocity to increase with the size of
the favor (Berkowitz & Friedman, 1967, Fisher & Nadler, 1976, Greenberg & Bar-Tal,
1976, Kahn & Tice, 1973, Levanthal, Weiss & Long, 1964, Pruitt, 1968, Stapleton, Nacci,
& Tedeschi, 1973, as cited in Fisher et al., 1981; Gergen et al., 1975; Goranson &
Berkowitz, 1966; Greenberg & Frisch, 1972; Gross & Latané, 1974; Nemeth, 1970; Wllke
& Lanzetta, 1982). This set of results is clearly consistent with felt obligation theory.

However, few investigations have looked at the effects of favor-doer norm or
expectancy violations on reciprocity. As asserted earlier, it is likely that individuals would
perceive greater obligation towards a favor-giver when a norm and/or an expectation were
violated to provide the favor. Consequently, if an individual is confronted with someone
who does such a favor, s/he is likely to reciprocate a great deal more than if no violations
took place.

There is one study that is partially consistent with this suggestion. A reciprocity
measure was included in Morse et al.'s (1977) quiz show study. After completing the
measure of attraction toward the supervisor, contestants were surprised with an
opportunity to reciprocate by volunteering to help the quizmaster on an unrelated task.
When the experiment appeared to be over, the supervisor handed the participant a note
ostensibly written by their former quizmaster. It explained that the person who had been
their quizmaster was actually a medical student who badly needed help with his anatomy

project. In the note, the participants were asked to indicate how much time they would be

20

willing to devote to helping their former quizmaster by allowing him to measure various
parts of their bodies.

Based on felt obligation theory, one might expect the most reciprocity in the
double-violation condition. However, the only finding was an expectation main effect, but
the direction of this effect was consistent with felt obligation theory. When the quizmaster
violated an expectation based on his past actions, participants were more willing to spend
time on the anatomy project. Although felt obligation would also predict the same effect
for nonn violation or an interaction such that both violations must co-occur in order for
reciprocity to be high, neither of these patterns emerged. Thus, felt obligation theory is
well-equipped to explain Morse et al.'s attraction findings, but can only partially explain
their reciprocity findings.

The only other research on reciprocity toward a norm and/or expectancy violating
favor-giver is inconsistent with felt obligation theory. In fact, Brehm and Cole (1966),
Worchel et al. (1976), and Schopler and Thompson (1968) all found some evidence that
directly contradicts the pattern of reciprocity predicted by felt obligation. These studies,
which are described in the next section, found that under some conditions, a favor
involving some type of violation led to less reciprocity than either no favor at all or one
that was expected/normative. These results can be best explained by the third potential
mechanism, psychological reactance, as descnhed below.

Psychological Reactance Fm Receiving a Favor

The third potentially relevant mechanism involves Brehm's concept of

"psychological reactance" (Brehm, 1966; Brehm & Brehm, 1981). According to this idea,

when an individual believes that one of his/her fiee behaviors has been threatened, s/he

21

feels an undesirable restriction of freedom.3 Brehm suggests that if freedom of the
restricted behavior is of some importance to the individual, then ”psychological reactance”
would be aroused. Reactance is defined as a strong desire to reestablish fieedom of
behavior. In order for fieedom to be directly restored, a person would have to engage in
the threatened behavior. Thus, an individual experiencing reactance would be highly
motivated to engage in the behavior(s) that s/he has learned s/he cannot or should not
perform. According to Brehm and Brehm, reactance may also include adverse emotional
consequences, such as feelings of resentment toward the person or object threatening
one's Mom of behavior.

A good example of reactance might be an interaction between a teenager and her
parents where the parents adamantly tell the teen that she is not allowed to go to a concert
with her fiiends. As a consequence, a once mildly attractive concert suddenly becomes a
must-do event for the teen. The concert becomes so enticing that the teen even sneaks
out of the house to attend it. This increased motivation to attend the concert would stem
from the daughters perception that her parents threatened her once free behavior to go
out with her fiiends. Because attending the concert was of some importance, the daughter
wanted to reestablish her sense of freedom by ultimately going to the concert. It is likely
that she also resented her parents, at least temporarily, for threatening this freedom.

In the current fiamework, the nonn of reciprocity (Gouldner, 1960) can limit one's
fieedom of behavior when one receives a favor from someone. The normative pressure to
reciprocate the favor would impose a restriction on an individual's fieedorn not to

reciprocate (Brehm & Brehm, 1981). According to Brehm and Brelun, normative

 

3 The term ”fi'ee behaviors” is used 100me by Brehm and Brehm (1981) to include such
things as the fieedom to act, think, believe, or feel certain things.
22

pressuresarequite strongsoitwouldtakeafairarnountofreactanceto motivatean
individual to violate the norm of reciprocity. Thus, it is likely that reactance would only
be aroused as a result of receiving large favors in situations where it is of some cost to
reciprocate. In other words, the freedom not to reciprocate must be somewhat important
to the recipient and the favor-giver must impinge upon this freedom to a great extent by
performing a substantial favor.

According to reactance theory, if these criteria are met, recipients would be
motivated to avoid reciprocating (Brehm & Brehm, 1981 ). Moreover, the favor-giver
who caused such discomfort would likely be derogated by the recipients. As noted in the
previous section, there is a host of literature demonstrating that individuals avoid
overbenefiting and that receiving favors often leads to negative feelings, even resentment
toward the favor-giver. Also mentioned earlier were the findings that inappropriate and/or
unexpected help leads to less liking of a favor-giver. In another study that is consistent
with reactance theory, Indian, Hanfinan, and Dembo (1947, as cited in Fisher et al., 1981)
found that injured and incapacitated individuals disliked their aid-givers to the extent that
the help they received restricted their freedom of behavior. Thus, there is support for the
suggestion that individuals may derogate a favor-giver who violates a norm and/or
expectancy because they would feel that the favor restricted their freedom of behavior.

This noted, psychological reactance theory, like felt obligation, can provide a
potential explanation for the attraction findings obtained by Morse et al. (1977). Perhaps
reactance, and thus target derogation, only occurred in the double-violation condition for
the same reason that felt obligation may have been greatest in this condition, namely,

because the favor may have been seen as greatest and most out ofthe ordinary. As a

23

consequence, participants may have felt much pressure to reciprocate and thus perceived a
greater threat to their freedom of behavior. They may have then derogated the favor-giver
because he was to blame for the arousal of their reactance.

Ifthis notion is correct, perhaps reactance occurred because participants in Morse
et al.'s (1977) study, and in other similar investigations, were not aware that they would
get the chance to reciprocate later. Had they known, they may have foreseen the
opportunity to avoid reciprocating the favor. Therefore, mere knowledge that they would
have been able to respond to the favor-giver would have allowed them to anticipate being
able to re-gain their lost freedom. Consequently, they may not have felt resentment
toward the benefactor because they would have been aware that s/he only restricted their
fi’eedom of behavior temporarily.

Thus, although for very different reasons, reactance would predict the same
pattern of favor-giver evaluations as that suggested by felt obligation theory. Both would
expect derogation of a favor-giver to occur only when a favor is perceived as large, as
when a norm and/or expectation is violated. Moreover, it would be consistent with both
theories to expect that dislike of a favor-giver would only take place when individuals do
not perceive an opportunity to reciprocate. Again, some evidence of this has been noted
earlier.

Dre to their similarity, it is important to compare and contrast psychological
reactance and felt obligation. They are similar in that they are both products ofthe nonn
of reciprocity. Moreover, they generate essentially the same predictions regarding
attraction toward a favor-giver. It is clear, however, that they differ in their most

important aspect: the motivational component. As noted, psychological reactance is a

24

strong urge to avoid reciprocity behavior whereas felt obligation promotes an urge to go
to great lengths to reciprocate. Thus, the two mechanisms involve completely difl'erent
motivations and subsequent reciprocity behavior. The predicted emotional experiences
arising from these very different motivations, however, are nearly identical. Feelings such
as distress, frustration, or resentment toward a favor-giver could logically arise from either
type of motivation if an opportunity to reciprocate freely is thwarted or unavailable.
However, these reactions are only secondary while the motivations toward certain
reciprocity behavior are the essence of felt obligation and psychological reactance.

In fact, Brehm (1966) and others suggest that the key indication of reactance
arousal is engagement in the once the behavior. In the current context, the behavior
would be less than appropriate reciprocity. Reactance theorists typically believe that it is
not essential for derogation to occur because it is not instrumental in the reduction of
reactance (Brehm & Cole, 1966; Worchel & Andreoli, 1974). The definition of reactance
refers strictly to the motivation to engage in certain behaviors that restore freedom. There
is no indication that certain affective consequences must occur as a result ofreactance.
They are simply common by-products.

Thus, the best indication of whether reactance or felt obligation results from
receiving a favor is the degree to which individuals are motivated to reciprocate the favor.
Felt obligation theory would clearly predict high reciprocity as a response to receiving
large favors, whereas the theory of psychological reactance would predict that individuals
would reciprocate the least when favors are large. Because a favor is perceived as quite
out of the ordinary when an expectation and/or norm violation takes place, feelings of

reactance and felt obligation would lead to very difl‘erent reciprocity behavior in this

25

circumstance. Iffelt obligation is experienced, an individual would be inclined to pay back
the violating favor-giver quite generously in order to relieve an uncomfortable feeling of
indebtedness. However, if reactance is experienced, participants would be more inclined
to restore their freedom of action by engaging in the threatened behavior, i.e., minimal or
no reciprocity.

Thus, reactance cannot explain the reciprocity findings obtained by Morse et al.
(1977). Reactance theory would have expected participants to reciprocate the least to the
quizmaster who violated both a norm and an expectation. However, not only did
reciprocity fail to decrease for the violation conditions, but expectancy violations even
generated more reciprocity. Thus, the reciprocity data fiom this study are clearly more
consistent with felt obligation. Therefore, although reactance theory can be used to
explain Morse et al.'s attraction findings, it fits the overall data quite poorly since it
predicts reciprocity findings antithetical to those obtained in the study.

However, some other investigations have found that individuals do avoid
reciprocating large favors under reactance-arousing conditions (Brehm & Cole, 1966;
Schopler & Thompson, 1968; Worchel, Andreoli, & Archer, 1976). Schopler and
Thompson (1968) found that individuals reciprocated inappropriate favors less than
appropriate favors. As mentioned earlier, in their study, the favor-giver gave the
participant a rose in either an appropriate or inappropriate situation. Following this nice
gesture, this "unemployed salesman" asked participants if they would be willing to hand-
wash a new blouse that was on the market in order to test it for durability. Participants
then indicated how many consecutive days they would be willing to do so. Those in the

inappropriate (formal) favor condition offered to wash the blouse much less than those in

26

the appropriate (informal) favor condition. This is precisely what reactance theory would
predict.

Brehm and Cole (1966) found that receiving a counternorrnative favor even led to
less reciprocity than receiving no favor at all, but only when importance of behavioral
fieedom was high. In this experiment, participants received a fiee soft drink fiom a
confederate posing as another participant. This was an especially inappropriate and
unexpected favor because the confederate did not even buy one for himself. Importance
of behavioral freedom was manipulated by telling the participants that it was either
important or not important to be able to report an accurate first impression of the
confederate. It was suggested that this freedom of evaluation was threatened when the
confederate provided the favor. Following this, participants evaluated the confederate and
were then given a chance to reciprocate the favor. The experimenter asked the participant
and the confederate ifone ofthem could sort some papers for ha. She then placed the
stack of papers in front of the confederate and observed how long it took the participant
to join in and help.

As expected, when importance was high, lower reciprocity took place when the
participant received a soft drink than when no favor was provided. However, as
mentioned earlier, derogation of the favor-giver did not occur as reactance theory might
predict. But, again, this is not surprising since participants in the high importance
conditions were urged to be unbiased in their evaluations. They were even told that
accuracy of first impressions was associated with success in life. Therefore, Brehm and
Cole's study may still be considered supportive of the tenets of reactance theory because

the measurement of attraction was invalid due to this methodological flaw.

27

In a similar study, Worchel et al. (1976) also found some evidence for decreased
reciprocity under reactance-arousing conditions. However, this effect only occurred when
a certain type of attribution was made for the favor. Worchel et al. manipulated
importance of behavioral fieedom in about the same way that Brehm and Cole did. Later,
the participant was asked to administer a picture-identifying task to the confederate. The
confederate appeared to win some money for her performance on the task. She then gave
halfof her winnings to the participant. Thus, in all conditions the confederate provided
this somewhat unusual favor. later in the session, the experimenter gave the confederate
and the subject each a stack of data sheets to sort. The confederate received a stack about
four times the size of that received by the subject. The reciprocity measure was whether
or not the participant offered to help the confederate.

Worchel et al. expected reciprocity to be less likely in the high importance
condition than in the low importance condition. Interestingly, they found that this effect
only emerged when participants were led to believe that the confederate's favor occurred
as a result of a special relationship between them (unique interaction attribution
condition). The study also supported reactance predictions of attraction toward the favor-
giver. Worchel et al. found that evaluations of the favor-givers were less favorable when
it was important to rate them accurately than when it was not important. This occurred
even when specific attributions for the favor were not induced. Interestingly, derogation
took place in spite of the fact that this study had the same methodological flaw as the
Brehm and Cole study. Overall, there is a fair amount of research that supports reactance

theory predictions.

28

Summary and Theoretical Considerations
Attraction Toward the Favor-Giver

Summary of Mechanisms. Three mechanisms have been proposed as possible
explanations for the reason(s) individuals might derogate someone who provides a favor
that is out of the ordinary, such as one that involves norm/expectancy violations.
Unpredictability, felt obligation, and reactance theories have all been discussed as potential
mechanisms involved in this process. The unpredictability mechanism suggests that
derogation occurs because the favor-giver is perceived as unpredictable and this
characteristic is considered to be a negative attribute. The felt obligation mechanism
suggests that the favor-giver is liked less because large favors are capable of producing
uncomfortable feelings of moral obligation to reciprocate. The reactance mechanism
suggests that the favor—giver is resented because s/he restricts the recipient's fieedom of
reciprocity behavior.

Potential Moderators. Norm and expectancy relevant behavior have been
discussed at length to be potential variables involved in the three processes. There are
other variables that may be involved as well. As mentioned earlier, knowledge regarding
any potential reciprocity opportunities may be a key factor. Ifreeipients are not aware of
any future opportunities to reciprocate a large favor, very negative reactions should result I
fiom either felt obligation or reactance. If the chance to reciprocate is anticipated,
however, individuals should not experience such negative reactions from these
mechanisms. Knowledge of a reciprocity opportunity does not have implications for the

unpredictability mechanism, however. This mechanism simply suggests that a favor-giver

29

would be disliked to the extent that she violates norms and/or expectations, regardless of
whether an opportunity to reciprocate is available or not.

If the unpredictability mechanism is operating, it should also be valuable to
consider individual differences in tolerance of ambiguous people or situations. Budner's
(1962) concept of "intolerance of ambiguity," then, should be relevant to the involvement
of unpredictability in reactions to a favor-giver. To the extent that perceived
unpredictability might mediate the relationship between norm/expectation violations and
derogation of a favor-giver, intolerance of ambiguity should play a role. Specifically,
individuals who are intolerant of ambiguity should derogate the violating favor-giver more
than those who are tolerant. Thus, intolerance of ambiguity may be an important
personality characteristic involved in such responses.

It is also likely that individual differences in the value placed on the norm of
reciprocity might play an important role. Although the reciprocity norm is strong and may
even be ubiquitous, individuals probably differ in the degree to which they value it.
Individuals' belief in the reciprocity norm should primarily have implications for felt
obligation. To the extent that felt obligation is involved, individuals placing high
importance in the reciprocity norm should derogate a favor-giver more under the relevant
circumstances than individuals who do not value the norm as much. It is also possible that
reactance may produce these same results as well.

Also mentioned earlier, the perceived importance of freedom of reciprocity
behavior is another factor to be considered. Its implications have already been discussed
with regard to reactance. Brehm (1966) and Brehm and Brehm (1981) consider it to be

an important instigator of reactance. Perceived importance may be a meaningful variable

30

involved in felt obligation as well, but for difi‘erent reasons. Specifically, if individuals feel
a moral obligation to reciprocate favors, then it is likely that they would prefer to have
fi'eedorn in their reciprocity behavior. Thus, theory would suggest that importance is a
cause of psychological reactance, whereas it may be a consequence of feelings of
obligation. Either way, greater importance placed on the fieedom of reciprocity should be
associated with greater derogation of the favor-giver when the favor is seen as large and
either reactance or felt obligation are elicited.

Suflicient Number of Violations. As mentioned earlier, it is also important to
consider the possibility that derogation of a favor-giver may occur when only one violation
takes place. Research in support of this possibility was discussed above. Only one
investigation (Morse et al., 1977) found that both an expectancy and a norm violation
were essential. Therefore, it is very possible that a single norm or expectancy violation
alone might be sufficient in generating perceptions of unpredictability, felt obligation, or
reactance.

Humbold versus Additive Eflects. This raises another issue, the issue of additive
versus threshold effects.4 A high threshold of tolerance effect for attraction toward the
favor-giver appeared to emerge in Morse et al.'s (1977) study. Specifically, attraction was
only low in the double-violation condition while the single and no violation condition
means did not differ. However, research that finds derogation to result from a single
violation implies that either a low threshold or an additive pattern is more plausible. In the
case of a low threshold effect, single and double violations should generate the same

degree of derogation. Thus, an interaction would result. Ifthe effects are additive,

 

" Brehm and Brehm (1981) also raise this unresolved issue with regard to reactance.
3 l

however, two main effects would emerge. In other words, each transgression would be
sufficient in leading to less liking of the favor-giver but a double-violation would generate
even more derogation.

An Alta-native Mechanism: Ulterior Motive Attributions The current
investigation focuses on testing three potential mechanisms that may account for the
derogation of a violating favor-giver. However, there are at least a few other plausible
meclmnisms. One of these is the suggestion that derogation only results when recipients
believe that the donor had an ulterior motive for providing such a large favor. It is
possible that a violating supervisor would only be disliked if recipients assume that the
favor is the result of an expectation to get something in return.

A study by Brounstein and Sigall (1977, as cited in Brehm & Brehm, 1981)
supports this ulterior motive explanation. In the study, a confederate provided a favor to
the participant. The researchers had the confederate's pay either appear to be dependent
on the participant's evaluation of the favor-giver or not dependent on it. Dependency had
no effect on liking when the favor was performed before the dependency manipulation.
However, when the favor was performed after the manipulation, high dependency led to
decreased liking. It is likely that in this latter condition, participants perceived an ulterior
motive while receiving the favor from someone who would later benefit from the
recipient's evaluation.

The perception of an ulterior motive could generate derogation via many
processes, such as attributing negative characteristics (e.g., selfishness) to the donor or
experiencing a higher degree of reactance or felt obligation. For example, the perception

of an ulterior motive might cause increased psychological reactance because the recipient

32

learns that the favor-giver is definitely trying to influence his/her behavior. Alternatively,
the perception of ulterior motive might increase felt obligation because one might feel
more obligated to reciprocate if it is clearly expected by the donor.

Although the ulterior motive mechanism is plausible, it is not compatible with most
previous research. An ulterior motive is unlikely to have been perceived in most of the
studies that have found derogation of a violating favor-giver. For example, in Worchel
and Andreoli's (1974) study, efforts were made to minimize any ulterior motive
attributions by leading participants to believe that the confederate did not know what type
of interaction they would have together later. Also, in Morse et al.'s (I977) quiz show
study, participants completed the measure of attraction before learning that the quizmaster
wanted something in return. However, in Kiesler’s (1966) study, one of the attraction
measures did follow the receipt of the favor so it is possible that recipients believed that
their partner provided the favor in order to receive a positive evaluation. The ulterior
motive mechanism is most likely to have been in operation in the Schopler and Thompson
(1968) study because the inappropriate favor was provided by a salesman, albeit an
unemployed one, introducing a new product.

Thus, although it is unlikely to have been responsible for derogation in most
studies, the perception of an ulterior motive causing negative reactions remains plausible.
In order to prevent the operation of this mechanism, the current investigation took efforts
to minimize the possibility that recipients might attribute the favor to an ulterior motive.

There is also a chance that other types of attributions for the favor may be
responsible for the derogation of a favor-giver. However, research on attributions of a

favor provided by an expectancy/norm violator is limited and very inconsistent, so no

33

predictions can be made (Worchel & Andreoli, 1974; Worchel et al., 1976). Nevertheless,
to explore some of these possibilities, the current investigation did include a
supplementary measure of attributions for the favor.

A Second Alternative Mechanism: Threat to Self-Esteem. Another potential
mechanism might be a threat to self-esteem that individuals may experience upon receiving
large favors (see Fisher et al., 1981). The receipt of help may generate perceptions of
failure, inferiority, and/or dependence, which could potentially lead to other negative
responses such as disliking the favor-giver. It is reasonable, then, that one might not like
someone who does not think highly of him/her.

Some past research is consistent with this notion. Some of the evidence for this
process comes from studies finding that individuals seek less help from physically
attractive (Nadler, 1979, as cited in Fisher et al., 1981) or similar donors (Nadler, Fisher,
& Streufert, 1976). Ifthreat to self-esteem were not an issue, these donor characteristics
should not matter much.

Also, other work has found that recipients accept less help on ego-involving tasks
than on non-central tasks (Gergen, Morse, & Kristeller, 1973, as cited in Fisher et al.,
1981). The authors suggest that individuals are uncomfortable receiving help on a task
that is important to them. Clearly self-esteem loss would be more of a threat when tasks
are relevant.

This avoidance of help on ego-involving tasks has been shown to particularly
occur for recipients high in self-esteem (Tessler & Schwartz, 1972; Wallston, 1976, as
cited in Fisher et al., 1981). Also, high self-esteem people seek help more frequently when

they are aware that many others also needed help on the same task (Nadler & Porat, 1978,

34

Tessler & Schwartz, 1972, as cited in Fisher et al.). Thus self-esteem does appear to be
involved and individuals with high self-esteem may engage in self-esteem maintenance
strategies by only accepting help when it is less threatening to do so.

Recipients have also been found to obtain more help and like favor-givers more
when help is offered, rather than when it must be requested (Broll, Gross, & Piliavin,
1974; Piliavin & Gross, 197 7). It is likely that recipients would experience greater threat
to self—esteem when help must be sought out by them. Moreover, when recipients have
greater perceptions of need, they tend to seek help less, have more negative attitudes
about receiving help, and have less attraction toward the source of their help (Calhoun,
Dawes, & Lewis, 1972, Franklin, 1975, Mikesell & Calhoun, 1971, as cited in Fisher et
al., 1981; Morse & Gergen, 1971). These reactions might occur as a result of
embarrassment stemming fiom their high nwd.

And, as mentioned above, recipients who are anonymous or who are not expecting
to meet their helper are more likely to request favors. It is possible that recipients feel less
threat to self-esteem when they do not have to face their benefactor. Thus, there is much
research consistent with the notion that perceived threat to self-esteem may affect
reactions to offers of help.

However, threat to self-esteem is not likely to have driven the derogation of the
violating favor-givers in most of the previous research discussed earlier. Given the nature
of the favor manipulations, the favors that were given (e.g., receiving an extra research
credit, receiving a flower, or receiving help on an easy paper sorting task) did not imply
that the recipient was substandard in any way (Clark & Mills, 1979; Sch0pler &

Thompson, 1968; Worchel & Andreoli, I974).

35

It is possible, however, that a threat to self-esteem mechanism was in operation in
the Morse et al. (1977) study. In that study, the contestants were probably not aware of
their performance relative to others. And more importantly, the quizmaster provided hints
to help the contestants answer the questions correctly. Although this took place in every
condition, participants might have been especially likely to perceive failure when the
quizmaster both broke the rules and behaved out of character to provide these contestants
with help. Perhaps the contestants believed that other previous participants did not need
the hints that they were given. Thus, the operation of a threat to self-esteem mechanism
may have occurred. However, it should also be noted that hint-giving was a standard
feature of these types of contests. One might argue that threat to self-esteem from the
expectancy and norm violating quizmaster would be less likely to occur because hints
were expected and normative to some extent. Nevertheless, the current investigation took
efforts to reduce the potential for self-esteem threat by informing participants in all
conditions that their performance was comparable to that of previous participants.

A Third Alternative Mechanism: Derogation by Association. Another potential
mechanism involves negative affective responses that may arise fiom having associated
with an expectancy or norm violator. This is particularly relevant for norm violations.
Perhaps a recipient might feel that the favor-giver made him/her an ”accomplice” in the
norm violations that took place in providing the favor.

In Morse et al.'s (1977) study, the contestant may have felt guilty or uncomfortable
as a result of perceiving him/herself as playing a role in the nrle-breaking. However, this
explanation cannot account for the derogation of an expectancy violator unless the

recipient does not want to be part of the donor's unpredictability. Overall, the derogation

36

by association mechanism may have been responsible for the quizmaster attraction findings
in Morse et al.

Due to the nature of the favors provided in the other investigations on reactions to
a violating favor—doer, the possibility that the recipients felt like an accomplice in a
negative act was minimal. This is simply because the favors in the violation conditions did
not involve breaking any specifically stated rules and were thus not illicit in any way.
Although it deserves mention, this is an unlikely explanation for the derogation of a
violating favor-giver in most previous research.

Operation of More Than One Mechanism. Additional theoretical issues also
deserve some attention. For instance, although each of the primary mechanisms discussed
can sufficiently account for benefactor derogation, it is also possrble that more than one
mechanism can operate simultaneously. For example, it may be that derogation often
results from both felt obligation and the target's perceived unpredictability. Sirnilariy,
derogation may occur as a result of both reactance and perceived unpredictability. It is
unlikely that individuals would simultaneously experience both reactance and felt
obligation, however, due to their incompatible motivations. Nevertheless, the other two
combinations are very possible.

It is also possible that various combinations of the primary and alternative
mechanisms can co-occur or even be more directly related to each other. For example, it
was mentioned that there is a possibility that perceptions of an ulterior motive might cause
increased psychological reactance or felt obligation. Of course, the independent operation
of more than one mechanism is also plausible. The current research focuses only on the

initial three mechanisms described while attempting to minimize the operation of any

37

alternative mechanisms. Therefore, only simultaneous operation among these three
mechanisms is tested in the current investigation.

Thus, in considering attraction toward an expectancy/norm violating favor-giver,
one should take all of the following into account: potential mechanisms involved;
knowledge regarding a reciprocity opportunity; intolerance of ambiguity; norm of
reciprocity beliefs; importance of free reciprocity behavior; single versus double violation
effects; additive versus threshold effects; and the possibility of simultaneous operation of
more than one mechanism.

Reciprocity Toward the Favor-Giver

Only two of the three theories can make predictions about reciprocity. The
unpredictability perspective has no basis for making any predictions regarding reciprocity.
It is possible that unpredictability might be negatively correlated with reciprocity but this
not need to be shown for the theory to be supported. Reciprocity behavior is a critical
outcome, however, in determining whether felt obligation or reactance mechanisms are at
work. One assumption is that a favor will be perceived as larger when a favor-doer
violates a norm and/or expectancy. Felt obligation theory would suggest that the greater a
favor is, the more individuals will seek to firlfill their obligation to reciprocate. Thus,
when a favor-giver breaks rules or expectations in providing the favor, greater reciprocity
should be observed. Reactance, however, would predict the opposite pattern. larger
favors should generate the least reciprocity, due to the recipient's desire to reestablish
fieedom of behavior.

As previously mentioned, mixed results on reciprocity are evident in the literature.

Moreover, none of the mechanisms can completely account for the pattern obtained by

38

Morse et al. (1977). Thus, it is important to examine the operation of both felt obligation
and reactance simultaneously with the appropriate related variables discussed above in

order to gain a clearer picture of which process(es) take place.

39

The Current Investigation

The purpose of the current investigation was to conduct a conceptual replication
of Morse et al.'s (1977) study with some extensions that permitted examination of the
three potential underlying mechanisms described above. To do this, participants were told
that they would participate in a ”business-setting simulation” during which they would be
assigned the role of ”worker. " They were firrther informed that each worker had a
”supervisor" (really the experimenter) who was in charge of paying the worker for his/her
work product. The supervisor always appeared to pay the participant very generously.
Expectations and norms regarding payment were manipulated in a manner similar to that
used by Morse et al.

Moreover, an additional manipulation regarding participants' knowledge of a
fiiture opportunity to reciprocate was included. Upon receiving the supervisor's generous
payment, participants were either told nothing about a chance to reciprocate or were told
that they would later get the chance to pay the supervisor for participation in an unrelated
task.

Following the supervisor’s apparent allocation, participants were asked to evaluate
the supervisor. Then they were all given the opportunity to reciprocate the favor by
paying the supervisor for the unrelated task. Attraction and reciprocity toward the favor-
giver were the major dependent variables, as in previous research.

Various additional measures were included to further explore the implications of
the three mechanisms for evaluations of a favor-giver. To firrther test the unpredictability
mechanism, items relating to unpredictability were included in the supervisor evaluation

and participants' intolerance of ambiguity was measured. To fiirther test felt obligation

40

and reactance, a questionnaire assessing norm of reciprocity beliefs was included along
with a measure of importance of the freedom of reciprocity. Additionally, measures of
temporary affect and favor attributions were included for exploratory reasons.

The major predictions made by each of the three mechanisms are summarized in
Table 1.1. However, for simplification, these predictions assume that the double-violating
favor-giver is perceived differently than the single-violating favor-giver (i.e., an additive
effect or a high threshold effect). As discussed above, the low threshold explanation may

suggest otherwise.

41

Table 1.1

 

 

SummaLLof Main Predictions
Attraction Reciprocity Additional
Mechanism Predictions Predictions Predictions
Unpredictability lowest attraction in no predictions «highest perceived
as a Negative double-violation unpredictability in
Attribute conditions, regardless double-violation
of prior knowledge conditions
of reciprocity
«unpredictability as
a significant mediator
of effects on attraction
«findings should
especially hold for
participants high
on intolerance of
ambiguity
Felt Obligation lowest attraction in most reciprocity «findings should
double-violation, double-violation especially hold for
no prior knowledge conditions participants who
condition highly value norm
of reciprocity
-perhaps especially
true for participants
who place high
importance in free-
dom of reciprocity
Psychological lowest attraction in least reciprocity in «findings should
Reactance double-violation, double-violation especially hold for
no prior knowledge conditions participants who
condition place high import-
ance in fi'eedom of
reciprocity

42

Chapter 2

METHOD

Participants
Participants consisted of 300 undergraduate students enrolled in an introductory
psychology course at Michigan State University. They participated in this study as partial

firlfillment of one of their course requirements.

M!

Participants were informed that they would be participating in a business
simulation study during which they would play the role of worker. About half-way
through the study, each worker received a generous payment from the supervisor for a
work task that was completed. This generous payment favor either confirmed or violated
a payment expectation (when one was given) and was either in accord with or against a
payment rule that the supervisor was instructed to follow (when one was given).
Participants were later given a chance to evaluate the supervisor and reciprocate the favor.
Some participants had prior knowledge of this opportunity to reciprocate, while others did

not know ahead of time.

Luisa
The basic design was a 3 (expectation) x 3 (normative standard) x 2 (knowledge of
opportunity to reciprocate) between-subjects factorial. The expectation manipulation had

three levels: strict, generous, or no payment expectations. For the norm manipulation,

43

participants were led to believe that the supervisor was instructed to follow a strict
payment rule, a generous payment rule, or no rule at all. “fith regard to the third
independent variable, participants either had or did not have prior knowledge of their
subsequent opportunity to reciprocate. In all conditions, participants received a generous
payment from the supervisor. Thus, the expectation and norm conditions included
violation, confirmation, and no information control conditions. Refer to Table 2.1 below

for a listing of the manipulations and measures in the order that they were presented to the

participants.

Table 2.]

Listing of Tasks, Materials, and Manipulations
Creativity task
Expectation manipulation
Norm manipulation
Rating and payment from supervisor
Knowledge of reciprocity opportunity manipulation
Manipulation checks for expectancy and norm
Measure of supervisor evaluation (attraction and unpredictability)

”$9999.“?

Measure of temporary afi‘ect

9. Manipulation checks for knowledge of reciprocity opportunity
IO. Measure of importance of freedom of reciprocity

ll. Measure of reciprocity

12. Measure of attributions

13. Measure of belief in the norm of reciprocity

14. Measure of intolerance of ambiguity

15. Measure of suspicion

Materials and Procedure

Introduction by Ercperimenter. Participants arrived in groups of five and were
welcomed by one of three female experimenters who introduced the study to them. They
were told that they would be participating in a ”business simulation” and that they would
be playing the role of "workers" in a workplace. They were further told that each worker
would be assigned to his/her own "supervisor.” Participants were informed that the
supervisors were actually other undergraduate students who volunteered to participate in a
number of these sessions for extra credit. They were led to believe that these supervisors
were waiting in a nearby room. In actuality, there were no supervisors.

The experimenter told the participants that they had been randomly assigned to be
part of a 2—person, separate work space, business simulation. In doing so, they would be
working in worker-supervisor pairs but would not meet or have any face-to-face contact
with their supervisor. Specifically, each worker was to perform his/her duties in a separate
cubicle in one room, while the supervisors sat in separate cubicles in a difi‘erent room to
perform their supervisory duties. To justify this worker-supervisor separation, participants
were told that the goal of that particular simulation was to minimize bias on the part of the
supervisor. In order to sound more convincing, the experimenter mentioned that other
types of simulations were also being conducted where the workers interact with each other
or with their supervisors.

Participants were fiirther informed that, in order to make the business simulation
more realistic, they would even get paid for their work. They were told that their
supervisors would be using raflle tickets as payment and that participants could win up to

three prizes of $25 each with the raflle tickets earned during the study. A random drawing

45

was to be held at the end of the semester to award these prizes.

Afier signing a statement of informed consent, they were given more information
about the study. The experimenter told the participants that the main purpose of the study
was to examine the capacity to be creative in different work settings. As the workers,
then, they would be asked to perform a creativity task, which involved drawing three
pictures. Their end products would then be slipped in an envelope and taken to their
respective supervisors. Participants were told that each supervisor’s role was to rate
his/her worker‘s project on creativity and then pay the worker in rafie tickets. The
supervisor would use a -5 (very negative) to +5 (very positive) scale to report his/her
subjective reaction to the work and then decide how much to pay the worker.

Work Task. The researcher handed the worker three sheets of paper with random
markings on them (e.g., one with an incomplete circle with a diagonal line crossing
through it). She instructed the participants to incorporate the markings on each page into
as creative a picture as possible. They were told that they would have ten minutes to
complete all three. The experimenter explained to the participants that there were no right
or wrong answers because it was a completely subjective task. The experimenter
continued:

Your supervisor's going to rate your work simply based on his/her own opinions

of what's creative-and we've noticed that that varies a lot depending on which

supervisor you have... Because creativity is so subjective, you probably won‘t be
able to predict how high of a rating score you'll receive. This is also true in many

real business settings, like advertising agencies, where workers may complete a

project and have no idea how clever their boss will think it is.

This extensive explanation was included in an attempt to reduce any pay expectations that
workers may form on the basis of their perception of the quality of their work or their

creative ability.

46

The researcher then gave the drawing materials to the workers and told them that
she would be leaving the room so they could complete the tasks in private. Participants
were instructed not to talk to each other at this time or during any other point in the
session. To firrther minimize contact, they were completely separated by walls and
curtains. Upon leaving the room, the researcher led the participants to believe that s/he
was going to make sure all ofthe supervisors were in the other room waiting.

Alter ten minutes had passed, the experimenter returned and asked the participants
to finish up their pictures and put them into an envelope labeled with their worker number.
The experimenter then collected the envelopes and appeared to deliver them to their
supervisors, when in fact, she just dropped them off in another room. She then returned
and said that it would take a few minutes for the supervisors to rmke their ratings and
assign payment. She said that she would then go pick up an envelope for each one of
them from their assigned supervisor. In the envelope would be a slip of paper with a
number from «5 to +5 indicating their creativity rating. Also enclosed in the envelope
would be the number of rafie tickets that they were awarded by their supervisor. While
they appeared to be waiting for the supervisors to evaluate the pictm'es, the expectation
and norm manipulations were introduced.

Manipulation of Expectation. The experimenter introduced the supervisor
expectancy manipulation first. The researcher handed each of the participants a
"Supervisor Record Sheet," which ostensibly was a record of the ratings and, in the
experimental conditions, the number of raffle tickets that their own supervisor had
assigned to previous workers (see Appendix A for examples of these record sheets). The

researchers explained that it would be the job of the workers to record the rating and

47

payment they were about to receive on the record sheet. Participants were then asked to
look over the sheet so that they fully understand the information that would be requested
of them. This was done mainly to call participants' attention to the expectancy
information.

Participants in the experimental expectation conditions received one of two
versions of this record sheet. Each illustrated that the supervisor previously gave
creativity ratings from -2 to +2 for the efforts of the ten workers previously assigned to
them. One of the versions, however, showed that the supervisor had assigned anywhere
fiom eight to ten raffle tickets in the past (generous expectation condition), while the
other showed that the supervisor had assigned only one to three rafile tickets in the past
(strict expectation condition). The researcher went on to say ”We are noticing that some
supervisors are strict in terms of worker pay (giving out only a couple of the rafile tickets)
while others are generous (giving out most of the ten available tickets)" Participants were
then told that they could glance over the last column of their supervisor’s “Record Sheet”
if they wanted to see how strict or generous their own supervisor had been in past
sessions. For the no expectation control, participants received a similar supervisor record
sheet, but their version only had past creativity ratings on it without any payment
information indicated. Each participant was presented with a new record sheet with all of
the numbers circled in pencil in different handwriting in order to make it more realistic.

Manipulation of Norm. In the no norm condition, participants were told “There
aren’t any mles about how strict or generous your Supervisor should be in paying you; so
your Supervisor can pay you with as little or as many of the available raffle tickets as he or

she wants to.” In the strict and generous norm conditions, they were told:

48

This particular work session will be one in which your Supervisor will not be
allowed to assign payment fi'eely. Rather, to make sure we simulate a particular
kind of company atmosphere, your supervisor has been instructed to be strict
(generous) in assigning you raffle tickets for your work. Company regulations in
this simulation dictate that your supervisor allocate rame tickets specifically within
the range of about 1 to 3 (8 to 10) raffle tickets for your payment. It is completely
against company rules for your supervisor to be generous (strict) in paying you.

It was further explained that although the supervisor’s payment decisions would be made

in private, as they are in many real business settings, s/he has been explicitly instructed to

follow company rules.

Rating and Payment. Following the expectation and norm manipulations, the
researcher left the room once again and returned with the rating and payment envelopes.
Participants opened their envelopes in private and read a folded slip of paper indicating in
all cases, that their creativity rating was +1. Each participant also found a generous
payment of nine raflle tickets in his/her envelope. They were asked to transfer this
information onto the record sheet and set their tickets aside. The experimenter then
collected the record sheets.

Manipulation of Knowledge of Opportunity to Reciprocate Following this,
knowledge of reciprocity was manipulated. Participants in the prior knowledge condition
were told:

By the way, although your supervisors are not aware of this, the tables will be

turned later and you're actually going to get to assign payment to them for their

participation in another task. I'll give you the hill details about this later. I just
thought I‘d mention it now because this is the first session that we're going to do
this. Your supervisors don't even know about it yet. You'll be the first
participants to get the chance to pay them like they've paid you. Both you and
your supervisor are going to fill out a departmental survey later and you'll be in
charge of assigning payment to both yourself and the supervisor.

The experimenter then told the participants that this opportunity would come after

completing some of the written measures. Individuals in the no prior knowledge condition

49

were simply told that they and their supervisors would be filling out a departmental survey
later in the session.

Expectancy and Norm Manipulation Checks. Participants were then asked to
complete a brief questionnaire consisting of three manipulation check questions. They
were asked to use a Likert-type scale fiom 1 (disagree very much) to 7 (agree very much)
to express their agreement with the statements. The two statements of interest were “The
supervisor behaved in accord with his/her past behavior” and “The supervisor followed the
instructions s/he was given precisely as directed.” These were used to assess expectancy
and norm manipulation effectiveness, respectively (refer to Appendix B for all written
measures).

Measure of Supervisor Evaluation (Attraction and Unpredictability). The
researcher then introduced the supervisor evaluation measure. She explained:

This information is useful because one of the goals of this research is to examine

differences in supervisors' characteristics. Keep in mind that your ratings will not

be used to judge your supervisor in any way. We're just interested in your
perceptions of your supervisor‘s personality and working style. Your supervisors
don't even know that they're being rated by their workers. In fact, they'll never
even know that they've been evaluated or what scores they receive. We don't want
them to know now that they'll be evaluated partly because we're afraid that might
sway the decisions they make when performing their duties.

The experimenter then passed out the evaluation measure. It consisted of 22 sets
of bipolar adjectives. Each pair was presented in lO-point semantic differential format
with 0 attached to one adjective and 9 attached to its opposite. Five of these items were

used by Morse et al. (197 7) in their composite measure of attraction: unlikable—likable,

mean-kind, helpfirl-unhelpful, unfriendly-fiiendly, and good-had.s Sixteen additional items

 

5 Morse et al. (1977) also used the dimension, good-looking—ugly, but this was
eliminated in order to prevent confusion and/or suspicion in the participants due to the fact

that they had not seen their supervisor.
SO

were included to get a more complete picture of participants' perceptions of their
supervisors, including perceptions of their supervisors' unpredictability and related
characteristics. These additional adjective pairs were: straightforward-manipulative,
dishonest-honest, easy-going-temperamental, mature-immature, annoying-pleasant,
submissive-dominant, hostile—calm, stubbom-yielding, obedient-disobedient, predictable-
unpredictable, impulsive—restrained, rational-irrational, inconsistent-consistent, typical-
unusual, independent-conforming, and conventional-rebellious. While administering the
evaluation measure, the experimenter informed the participants that she would be leaving
to go administer the ”departmental task” to the supervisors. She then pretended to do so.
Introduction of Departmental Task. Following participants' completion of the
supervisor evaluation, the experimenter told them that she was going to stop the business
simulation study for a few minutes to administer the departmental survey she mentioned
earlier. She told participants that the ethics committee for the psychology department was
interested in collecting data about how participants feel while they are participating in
psychological research. The researcher explained that the psychology department supplied
some of the researchers with this brief questionnaire and asked them to administer it in the
middle of their experimental sessions. She apologized for the interruption and then
administered the ”departmental survey,” which really consisted of a measure of temporary
afl‘ect, manipulation check questions for knowledge of reciprocity, and a measure of
importance behavioral fieedom.
Measure of Temporary Afl'ect. This measure was a version of Diener and
Emmons's (1984) temporary mood state questionnaire. It consisted of nine mood

adjectives for which participants were to indicate on a 7-point scale, from ”not at all” to

5]

"extremely much," the degree to which they were currently experiencing each mood. The
moods consisted of happy, joyful, pleased, enjoyment/hm, unhappy, depressed,
fi'ustrated, worried/anxious, and angry/hostile.

Manipulation Check for Knowledge of Reciprocity. The supposed departmental
questionnaire also addressed questions regarding payment in the current experiment in
order to tap knowledge of reciprocity. Participants were asked to answer a couple of
questions regarding their knowledge of their subsequent reciprocity opportunity.
Specifically, they were asked if they would be receiving any compensation, who would be
in charge of such compensation, and whether they would have a say in anyone else's pay.
The question of interest was, “Will you have any say in how much anyone else will be paid
for their participation in this departmental survey?” They could respond with answer
choices, ‘yes,” “no,” or “don’t know.”

Measure of Importance of Freedom of Reciprocity. The theory of psychological
reactance (Brehm, 1966; Brehm & Brehm, 1981) suggests that reactance will only occur if
fi'eedom of the threatened behavior is of some importance to the individual. Therefore, a
measure assessing participants’ perceived importance in their fi'eedom to receive and have
control over the raffle tickets was included as the third and final portion of the
“departmental questionnaire.” It consisted of three seven-point items ranging from
“extremely unimportant” to “extremely important.” These were, “How important is it for
you to receive payment/compensation for completing this departmental task?” and “How
important is it for you to have a say in how much payment/compensation you receive for

completing this type of departmental task?” and “How important is it for you to have a say

52

in how much payment/compensation others will receive for competing this type of
departmental task?”

Measure of Reciprocity. Following their completion of the departmental
questionnaire, participants were given the chance to pay themselves and their supervisor
for filling it out. The experimenter told them:

As I said, the supervisors are filling this out too even though it's not part of their
commitment to this research project. Because it is something extra for them to do,
we'd like to pay the supervisors for their participation in this departmental task.

It's also something extra for you, so we think you should be able to decide if you
would like some extra payment of this departmental task also.

The psychology department gave us a little extra money for collecting this data,
but not much, so we decided to just add it to the money you can win with the raflle
tickets and then assign payment in rafile tickets again. This is actually the first and
only session where your particular supervisors will be earning raffle tickets too.
We thought that since your supervisor decided how much payment you would
receive during the business simulation study, that it would only be fair if you were
in charge of assigning payment for the departmental task. So, I'm going to give
you nine blank raffle tickets and ask you to decide how many of the nine you
would like to give to your supervisor and put them in this envelope marked
'supervisor pay.’ Although your supervisors weren't allowed to do this, you can
keep whatever raffle tickets remain after you've paid the supervisor. . .. There aren't
any rules for how you should assign payment so you're completely flee to allocate
the tickets however you want.

The reason participants were told that they can keep any remaining rafile tickets was to
make this reciprocity be somewhat costly as it was in Morse et al. (1977). Another reason
was to make the reciprocity decision more important to the participants.

Participants were also told that the supervisors would not be receiving the payment
envelopes until after the experiment was over in case it might afi‘ect their behavior during
the business simulation. In reality, this was said to reinforce the idea that the supervisors
never had any knowledge about getting paid themselves. When the participants finished

paying the supervisors, the experimenter collected the supervisors' envelopes and set them

53

aside. She then apologized for the interruption and appeared to resume the business
simulation study. Participants were then asked to complete a final packet of
questionnaires that was comprised of the measures described below.

Measure ofAttributions. This measure was a free-response question asking the
participants to write down why they think their supervisor gave them the specific amount
of payment they had received. This question was intended to tap participants’ payment
attributions and other thoughts. It was also helpfirl in providing an indication of
participants’ suspicion.

Measure of Belief in the Importance of die Norm of Reciprocity. The next
questionnaire assessed the extent to which participants valued the norm of reciprocity on a
general level. The questionnaire pitted "repaying those who have been nice to you”
against four other virtues. For each pair of virtues, participants were asked to indicate
which was more important and to what degree. They did this by choosing one of ten
points, with one virtue anchored at “0" and the other at "9."

Measure of Intolerance of Ambiguity. The next questionnaire was Budner's
Intolerance of Ambiguity Scale (1962). It asked participants to indicate on a 7-point
Likert-type scale from 1 (strongly disagree) to 7 (strongly agree) how much they agreed
with 16 general statements regarding their tolerance/mtolerance of ambiguous situations.
Some examples are, "The sooner we all acquire similar values and ideals, the better,”
”People who fit their lives to a schedule probably miss most of the joy of living," and
"Often the most interesting and stimulating people are those who don't mind being

difi‘erent and original." The latter two are reverse-scored items.

54

Measure of Suspicion. As a final task, participants were asked to respond to
some suspicion items. The most important item was, “Please describe, in your own
words, what you believe was the purpose of the study in which you have just
participated. "

Following this, participants turned in their completed measures and were asked to
fill out and turn in their raflle tickets. They were informed that they would be notified if
they won a monetary prize. At that time, participants were thanked, debriefed, and

dismissed.

SS

Chapter 3

RESULTS

Pilot Tests

Believability of +1 Rating. A pilot test was conducted in order to be sure that
participants in the actual study would believe that they deserved a +1 rating for creativity
on the incomplete figure drawing tasks. Forty-four participants worked on the same
drawing tasks as those in the actual study and then estimated the creativity score that they
believed they would get from a supervisor based on those drawings. On a scale from -5
(very negative evaluation) to +5 (very positive evaluation), the average was +1.29. Thus,
the closest round number would be +1, making that the optimal choice for believable false
feedback.

Construction of Norm of Reciprocity Beliefs Questionnaire. A pre-test was
used to construct the questionnaire assessing individuals' beliefs regarding the importance
of the norm of reciprocity. Perceived importance of various virtues was assessed to find
some that were similar in importance to repaying others. These were then all pitted
against "repaying those who have been nice to you” and participants were asked to rate
their relative importance. The reason participants were not just asked to rate the
importance of this reciprocity virtue on its own is that ceiling efi‘ects may have been a
problem. With the virtue anchored against similarly important virtues, a more sensitive
and valid measure was expected. The final measure is in Appendix B.

However, the questionnaire may have been invalid in spite of these efforts because

it produced no effects, including no correlation with actual reciprocity. Thus, the details

56

concerning the construction, hypotheses, and analyses regarding this questionnaire will not
be discussed further.

Scenario Studies to Pretest Manipulations. Scenario versions of both the quiz
show study (Morse et al., 1977) and the creativity task business simulation study were
constructed. These were used to make sure that the expectation and norm manipulations
for the business simulation were both effective and at least as strong as those used by
Morse et al. The expectation manipulation was effective for both scenarios and was
stronger for the business simulation. The norm manipulation in the business simulation,
however, was both too weak and less effective than that in the quiz show scenario. A
second version of the business scenario was created using stronger manipulations. The
manipulation checks revealed that these changes were successful. Thus, these newly-

constructed stronger manipulations were then used in the live study.

57

Manipulation Checks in the Present Study
Expectation Manipulation. An ANOVA revealed that the expectancy

manipulation was successful, E (2, 287) = 133.58, p < .0001. On average, those in the
expectancy confirmation condition (where the supervisor had given all generous payments
in the past), agreed with the statement, "The supervisor behaved in accord with his/her
past behavior. " Those in the expectancy violation condition disagreed with the statement,
on average. The means were 6.22 and 2.96, respectively, on a Likert-type scale ranging
fiom 1, disagree very much, to 7, agree very much. The mean score for the no expectancy ‘
group was 5.83. A Tukey post-hoc test revealed that the mean in the expectancy violation
condition was significantly different from that of both the expectancy confirmation
condition and the control (ps < .0001). Interestingly, the post-hoe test also found that
those in the expectancy control condition did not differ from those in the expectancy
confirmation condition (p = .19). Thus, the supervisor was perceived as behaving
consistently even in the absence of information regarding his/her consistently generous
behavior.

Norm Manipulation. An ANOVA also revealed a successful norm manipulation,
E (2, 287) = 51.03, p < .0001. On average, those in the norm confirmation condition
(where the supervisor obeyed a generous payment rule) agreed with the statement, "The
supervisor followed the instructions s/he was given precisely as directed.” Those in the
norm violation condition disagreed with the statement, on average. The means were 5.88
and 3.68, respectively, on a Likert-type scale ranging from 1, disagree very much, to 7,
agree very much. The mean score for the no expectancy group was 5.71. A Tukey post-

hoc test revealed that participants in the norm violation condition agreed with the

58

statement significantly less than those in either the norm confirmation condition or in the
control (ps < .0001). Once again, those in the norm control condition did not differ from
those in the norm confirmation condition on the manipulation check. Thus, the supervisor
was perceived as behaving in a normative manner even in the absence of information
regarding the operating norm.

Because the participants in the no norm and no expectation control conditions
perceived the supervisor as being normative and consistent, respectively, data fiom these
participants are not very informative. Therefore, for the sake of interest and practicality,
the focus of the subsequent AN OVAs and simple effects tests will be on comparing the
violation and confirmation conditions only.6 However, all effects were tested within the
fiamework of the larger design using the pooled error term across all conditions. Also,
contrasts with the appropriate control groups were performed where relevant.

Knowledge of Reciprocity Manipulation. The effectiveness of the knowledge of
reciprocity manipulation was assessed by examining participants' responses to the item
asking whether they would have any say in determining anyone else's pay during the
session. Responses were coded as a 3-point scale, with 1 = "yes," 2 = "don't know," and 3

= ”no." This variable was then entered as a dependent measure in an ANOVA. The

 

6 The confirmation conditions were chosen over the controls for two reasons. First, these
conditions are more analogous to the violation conditions, thus allowing for more
appropriate comparisons. The reason for this is that both the violation and confirmation
conditions include information regarding the supervisor's previous behavior and both
include a normative standard presented by the experimenter. The control conditions,
however, alter the situation more drastically by presenting no expectancy and norm
information at all. The second reason for using the confirmation conditions is that they
provide increased confidence in understanding how the participants perceived the
supervisor because the situations were constructed in order to elicit certain perceptions.
59

analysis revealed a significant difference between the prior knowledge and no prior
knowledge conditions, 13 (l, 256) = 16.74, p < .0001.

The mean difference was quite small, however. Means for the no knowledge and
knowledge groups were 2.85 and 2.51, respectively. Although more participants
answered "no” in the no knowledge condition than in the prior knowledge condition, the
majority of participants in both groups responded with a ”no." Thus, there were many
participants in the prior knowledge condition who incorrectly indicated that they would
not have a say in anyone else's pay during the session. However, these errors were
probably largely due to a misinterpretation of the manipulation check question.
Participants appeared to be confirsed about this item because some of them left it blank or
asked the experimenter in charge for help at this point in the study. Thus, the

manipulation may have been stronger than would be suggested by participants' responses.

Construction of Attraction and Unpredictability Factors
The likability and unpredictability factors were constructed based on the results of

a Principal Components Analysis with Varimax rotation with the number of factors
specified as two.’ All 21 supervisor evaluation items were submitted to the analysis. The
likability and unpredictability factors were the chief factors extracted, with eigenvalues of
5.19 and 4.60, respectively. The likability factor consisted of ten bipolar items with
likable, kind, helpful, fiiendly, good, honest, easy-going, mature, pleasant, and calm on the
positive poles. These items had factor loadings ranging from 0.42 to 0.82 after rotation.
The internal consistency, assessed using Cronbach's alpha, was 0.87 for this attraction
factor. The unpredictability factor consisted of seven items, namely, unpredictable,
disobedient, unusual, rebellious, inconsistent, irrational, and impulsive on the high
unpredictability poles. These items had factor loadings ranging fiom 0.61 to 0.82.
Cronbach's alpha for these seven items was 0.85, also indicating high reliability of the

measure. Refer to Appendix C for the rotated factor loadings for both factors.

 

7 To maximize power, control groups were included in this analysis and in the internal
consistency and correlational analyses.
or

 

The first set of hypotheses to be tested are those related to Kiesler's (1973) ideas
and the unpredictability mechanism. It was predicted that unpredictability would be
perceived as a negative attribute and that individuals would derogate a favor-giver who
violates norms/expectations because that person is perceived as unpredictable.

The first hypothesis to be tested was the suggestion that a favor-giver might be
perceived as more unpredictable when s/he violates an expectation and/or a norm to
provide the favor. To test this, a norm by expectation ANOVA was conducted with
unpredictability as a dependent variable. Problem data were excluded from all analysis.3

Two main effects were revealed. As predicted by the hypothesis, those in the
expectancy violation condition perceived the supervisor as more unpredictable than those
in the expectancy confirmation condition, E (l, 232) = 25.00, p < .0001. The means were
4.77 and 3.50, respectively, on a lO-point scale. Similarly, participants in the norm
violated condition rated the supervisor as more unpredictable than those in the norm
followed condition, E (1, 232) = 15.93, p < .0001. These means were 4.60 and 3.62,
respectively. No interaction was revealed, suggesting that the efl‘ects of these violations
are additive. Consistent with the manipulation check results, contrast analyses revealed
that the analogous expectancy control (M = 3.40) and norm control (M = 3.75) conditions
did not difi‘er from the expectancy and norm confirmation conditions, respectively (ps =

.99 and .45).9

 

8 Because of the highly complicated and deceptive nature of the study, data from many
participants had to be dropped fiom all analyses. Refer to Appendix D for details
regarding these exclusions.
9 All contrasts for all analyses involved using F -tests but only 95 were given for the sake of
simplicity.

62

As indicated earlier, the unpredictability hypothesis would firrther suggest that the
favor-giver would be liked less if s/he violates an expectation and/or a norm. An
expectancy by norm by knowledge of reciprocity AN OVA was performed with supervisor
attraction as the dependent measure. An analogous AN COVA with unpredictability as a
covariate was also performed to test for mediation. Unpredictability emerged as a
significant covariate, _F_' (l, 232) = 4.24, p < .05, but it did not eliminate any of the
significant findings that emerged in the original ANOVA In fact, including
unpredictability as a covariate increased the F—values. Because the same pattern of results
emerged for both analyses and unpredictability accounted for some of the variance, the
results fiom the ANCOVA are discussed. Refer to Tables A-2 and A-3 in Appendix E
for the ANOVA results.

The ANCOVA revealed some significant results, but all were antithetical to the
unpredictability hypothesis as well as the felt obligation and reactance hypotheses. A main
effect for expectancy revealed that participants liked their supervisor more when s/he
violated an expectation M = 6.17) than when s/he confirmed an expectation to be
generous M = 5.53), E (l, 226) = 7.60, p < .01. A main effect was also revealed for
norm, 13(1, 226) = 5.81, p < .05, such that participants rated their supervisor more
favorably when s/he broke a rule in order to be generous M = 6.11) than when s/he
followed a rule to be generous M = 5.57). These main effects were qualified by a
significant expectancy by norm interaction, E (l, 100) = 5.30, p < .05. Adjusted cell
means are shown in Table 3.1. The mean in the no expectancy, no norm group was 5.32.
All means were on a lO—point scale from 0 to 9 with higher scores indicating more positive

evaluations of the supervisor.

63

Simple effects tests to filrther explore this interaction revealed that it was due to
greater liking of the supervisor in any of the violation conditions than in the double-
confirrnation condition (ps s .001). None of the violation conditions differed significantly
from each other. Thus, the pattern reveals a low threshold effect, such that increased
liking of the supervisor occurred to the same extent regardless of whether an expectation
violation, a norm violation, or both took place.

Further analyses using contrasts revealed that attraction in the double-confirmation
condition did not difl‘er fiom that in the no expectancy, no norm control. Means for the
single- and double-violation conditions were all higher than the control, but one of these

difi‘erences was only marginally significant (23 = .02 to .07). Thus, a generous supervisor

Table 3.1

Amcfion Toward the Igor-Giver fig Function of Expectation and Norm with
UnpredicMability as a Covariate

 

 

 

Expectation
Norm Confirmation . Violation
Confirmation 4.95 .° 6.12 r,’
Violation 6.00 r,‘ 6.22 r,‘

 

Me. Attractionwasbasedontenitems,eachrangingfrom0to9. Themeansforthe
average score of these ten items are presented in the table. High scores indicate more
attraction. Means that do not share subscripts differ significantly at p < .05. Means
marked with a ' are significantly different from reciprocity in the no expectation, no norm
control M = 5.32) at p < .05. A ° indicates a marginally significant difference from the
control.

was liked more when there was at least one violation of an expectation or a norm to be
generous than when no information was given.

Thus, the patterns revealed by the simple effects and contrast analyses directly
contradict the unpredictability hypothesis, the felt obligation and the reactance hypotheses.
Although derogation did not take place in the expected cells, other evidence for the
operation of each of the three mechanisms was still investigated.

Other than the expected derogation, the unpredictability hypothesis also predicted
a negative correlation between unpredictability and likability to emerge. As predicted, a
significant, albeit small, negative correlation was revealed, (r = -0.13, p < .05). However,
in light of the evidence that perceived unpredictability was overlooked when participants
received a large favor (i.e., in the violation conditions), correlational analyses were
performed for the violation and no violation groups separately. For all conditions where
at least one violation occurred, no correlation emerged (r = .02, p_ = .80). However, for
those conditions where no violations took place, a highly negative correlation was
revealed (r = -.43, p <.0001). Thus, it appears that participants disliked a supervisor who
was perceived as unpredictable unless they benefited from his/her unpredictability.

Another prediction made by the hypothesis was that individuals scoring high on
Budner’s Intolerance of Ambiguity Scale (1962) would especially derogate the
expectancy/norm violating favor-giver. In other words, this individual difference was
originally expected to moderate the hypothesized derogation of the target. Although
derogation did not take place in the appropriate conditions, intolerance of ambiguity was

65

A median split was used to divide high and low scorers on the scale and this
dichotomous variable was entered into the expectation by norm by knowledge AN COVA
as another factor. Because unpredictability was not a significant covariate in this model, it
was removed fi'om the analysis. The only significant result involving intolerance of
ambiguity was its three-way interaction with expectancy and knowledge, 13 (l, 210) =
5.33, p < .05. To examine this interaction more closely, separate ANOVAs were
performed for each knowledge condition. A significant expectancy by intolerance of
ambiguity simple interaction only emerged for participants who had no prior knowledge of
a reciprocity opportunity, E (1, 103) = 4.71, p < .05. Means for this no knowledge group

are presented in Table 3.2.

Table 3.2

AttrJaction Toward the [Ava-Giver air; Function of ExpectatioL and Intolerance of
Ambigliity under Conditions of No Prior Kflwleggg

 

 

 

 

Intolerance of Ambiguity
Expectation Low High
Confirmation 5.33 . 5.43 .
Violation 6.46 5' 5.38 .

 

1% High scores indicate more attraction. Means that do not share subscripts differ
significantly at p < .05. Means that do not share subscripts differ significantly at p < .05.
Means marked with a ° are significantly different from reciprocity in the no expectation
control M = 5.04) at p < .05.

As seen in the table, increased liking toward the expectancy violating supervisor
disappears for people high on intolerance of ambiguity. Simple effects tests reveal that the
mean in the low intolerance, expectancy violation condition was significantly greater than
the other three (ps < .05). But this pattern did not emerge for norm nor did it emerge
under conditions of prior knowledge. Also, it was expected that high intolerance

individuals would like the violator less than the confirrner, but this did not occur.

67

Felt Obligation and Reactance: Attraction Finding;
One of the major predictions of both the felt obligation and the reactance

hypotheses was that the supervisor would be disliked most if she violated a norm and an
expectation to help when the recipients were not aware that there would be an opportunity
to reciprocate. As indicated in the previous section, derogation of the doubleviolator was
not borne out. In fact, the opposite pattern emerged. Moreover, the three-way AN OVA
and ANCOVA described above did not reveal the predicted expectation by norm by
knowledge interaction. , The only significant or marginally significant efi‘ects or
interactions for likability were those indicated in the previous section, namely expectation
and norm main effects qualified by an expectation by norm interaction. As mentioned, all
of these patterns run counter to the felt obligation and reactance hypotheses as well as the
unpredictability hypothesis.

To finther test the reactance hypothesis, perceived importance of the freedom of
reciprocity (i.e., having control over the raffle tickets) was included in the analysis.
Reactance theory predicts that when fieedom of behavior is considered important, more
reactance will be aroused. In this case, that would result in greater dislike of the
supervisor in the violation conditions. Importance of fi'eedom of behavior was measured
by taking the average of the three importance measures described earlier. A Principal
Components analysis with Varimax rotation revealed that all three items loaded on the
same factor, with loadings from .76 to .90. Cronbach's alpha for these items was 0.80,
indicating high internal consistency. A median split was performed on the importance

variable to make it dichotomous.

68

Subsequently, an expectation by norm by knowledge by importance AN COVA
was conducted for the attraction measure. Once again, unpredictability was entered as a
covariate. The findings involving importance were a significant knowledge by importance
interactiorn, E (1, 208) = 4.53, p < .05, and a significant expectation by norm by
importance interaction, 15 (1, 208) = 4.69, p < .05.

Means for the knowledge by importance interaction are presented in Table 3.3.
Simple effects tests revealed that the knowledge by importance interaction emerged due to
less liking of the supervisor in the no prior knowledge, high importance condition than in
the others (ps <.05). The remaining means were not significantly different from each
other. Thus, individuals who highly valued their freedom to allocate tickets and had no
prior knowledge of any reciprocity opportunity were least attracted to the supervisor.

This pattern could represent a threshold of reactance or felt obligation pattern. In other

Table 3.3

Attraction Towgd thfiavor-Giver a_s_g Function of Prior Knowledge of an Opgommg
to Reciprocate and Importgnce of Freedom of Recgnrocity with Unpredictability as a
Covariate

 

 

 

Importance of Behavior
Knowledge Low High
No prior krnowledge 6.01 r, 5.37 .
Prior knowledge 5.81 r, 6.05 r,

 

Note. High scores indicate more attraction. Means that do not share subscripts differ
significantly at p < .05.

69

words, it appears that both high importance and no prior knowledge were needed in order
to exceed the threshold of reactance and/or felt obligation to lead to derogation.

As mentioned, an expectation by norm by importance interaction also emerged.
Adjusted means are shown in Table 3.4. Simple efl'ects tests were conducted for the high
and low importance goups separately. Within the low importance goup, participants
liked the double-violating supervisor the most, but one of these differences was only

marginally significant (ps = .005 to .06). No other significant differences emerged but the

Table 3.4

Attrgction Toward the F gym-Giver a_s_a Function of Expectation Norm. 8_Il:d IIQwrtance
of Freedom of Reciprocity with Unpredictability as a Covariate

 

 

 

 

 

Expectation
Norm Confirmation Violation
Low Importance
Confirmation 5.34 . 5.90 .
Violation 5 .96 . 6.64 5'
High Importance
Confirmation 4.36 .‘ 6.28 .'
Violation 5.97 r, 6.10 5'

 

N_o_t_e_. High scores indicate more attraction. Within the low importance goup, means that
do not share subscripts differ significantly at p s .06. In the high importance goup, means
that do not share subscripts differ significantly p < .05. Means marked with a ' are
significantly different from reciprocity in the no expectation, no norm control M = 5.25)
at p < .05.

70

trend of the means suggests that participants tended to like the double-confirming
supervisor the least with an increase in liking for each violation committed. Contrary to
the reactance and felt obligation predictions, participants who placed high importance in
their fieedom to allocate tickets liked the supervisor less when s/he confirmed both a norm
and an expectation to help than when one or more violations took place (ps < .005). The
single- and double-violation cells did not differ significantly.

Another simple effects test revealed that high importance participarnts liked the
double-confirming supervisor even less than their low importance counterparts (p < .05).
There was a marginally significant difference between high and low importance
participants' ratings of the double-violating supervisor (p = .07). Thus, within the double-
confirnning and double-violating conditions only, the patterns are consistent with reactance
and felt obligation. However, as a whole, these data do not support the reactance
hypothesis because derogation did not take place in the double-violation condition for
either the low or high importance groups.

The adjusted mean in the no norm, no expectancy control was 5.25. Contrast
analyses revealed that the low importance participants liked only the double-violating
supervisor significantly more than the supervisor in the control (p < .005). High
importance participants liked only the expectancy violators more than the supervisor in the
control (ps < .05). They liked the doubleconfirming supervisor significantly less than the

supervisor in the control (p_ < .01).

7]

Felt Obligation and Reactance: Reciprocity Findingg
Unlike the unpredictability hypothesis, the felt obligation and reactance hypotheses

made predictions regarding reciprocity as well as attraction. Concerning reciprocity, the
felt obligation hypothesis expected participants to reciprocate the most toward the
supervisor when the favor was perceived as large, i.e., when the supervisor violated an
expectation and/or a norm to do the favor. The reactance hypothesis predicted the
opposite pattern; that participants would reciprocate the least when violations were
committed. It further suggested that reciprocity for large favors would be especially low
for individuals who placed high importance in their fi'eedom to allocate rafile tickets.

All expectation by norm by krnowledge by importance of behavioral freedom
ANOVA was performed to simultaneously examine these hypotheses. Reciprocity was
measured as the number of raffle tickets awarded to the supervisor subtracted by the
number of tickets kept for oneself. Thus, the possible range of reciprocity values is -9 to
+9 with higher numbers indicating higher reciprocity. Contrary to the hypotheses, there
were no significant main effects for expectation and norm nor an interaction between the
two. However, expectation by knowledge, norm by knowledge, and expectation by norm
by krnowledge irnteractions did emerge. With regard to importance of behavioral freedom,
the results revealed a main efi‘ect of importance, and expectation by importance, norm by
importance, and expectation by norm by krnowledge by importance interactions. Details
regarding all of these findings are presented below.

As mentioned, an expectation by krnowledge and a norm by knowledge interaction
merged, E (1, 204) = 6.45, p < .05 and E (l, 204) = 14.50, p < .0001. As seen in Tables
3.5 and 3.6, it appears that there is a reactance pattern for participants with no prior

knowledge of an Opportunity to pay back the supervisor. In other words, reciprocity was
72

Table 3.5

Reciprm Tom the Falor-Giver as a Function of Exgectation and Prior Knowledgg
of_an ijortunig to Reciproca_tr_3

 

 

 

Expectation
Knowledge Confirmation Violation
No prior knowledge +5.46 .r, +4.42 .
Prior knowledge +4.43 . +5.88 r,’

 

Egg Reciprocity was measured as the number of tickets given to the supervisor
subtracted by the number of tickets kept for oneself. Thus, lnigher numbers indicate
greater reciprocity. Means that do not share subscripts differ significantly at p < .05.
Means marked with a ' are significantly different fi'om the analogous no expectation
control M = +4.24) at p < .01.

Table 3.6

Recipfofl'tv Towid the Favor-Giver as a Function of Norm and Prior Knowledge of an

Memo Reciprocate

 

 

 

 

Norm
Knowledge Confirmation Violation
No prior krnowledge ' +6.35 r, +3.78 .’
Pn‘or knowledge +3.97 .‘ +6.33 .,

 

 

 

NOte. Reciprocity was measured as the number of tickets given to the supervisor
si-llbtracted by the number of tickets kept for oneself. Thus, lnigher numbers indicate
gi‘eater reciprocity. Means that do not share subscripts differ significantly at p < .05.
eans marked with a ° are significantly different fiom the analogous no norm control M
= $5.30) at p < .05.

73

lower when there was an expectation or norm violation than when there was a
confirmation. However, for those who did have prior lmowledge, a felt obligation pattern
emerged such that participants reciprocated more when there was a violation than when
there was a confirmation.

Simple effects tests revealed that the expectancy by krnowledge interaction was due
to statistical significance of the felt obligation pattern for participants with prior
knowledge. These participants tended to reciprocate more when the supervisor was
unexpectedly generous than expectedly generous (p < .05). The reactance pattern in the
no knowledge condition did not reach significance (p = .16). In other words, reciprocity
was not significarntly less toward a supervisor who violated an expectation than toward
one who confirmed an expectation.

Contrast analyses were completed to examine differences between these mearns and
the no expectancy control M = +4.24). They revealed that reciprocity in the expectancy
violation, prior krnowledge condition was significantly geater than reciprocity in the
control condition (9 < .01 ). The remaining tlnree means did not differ significantly fiom
the control.

Simple efi‘ects tests to exannine the norm by knowledge irnteraction found that
participants who had prior krnowledge reciprocated more toward a supervisor who
violated a norm than toward one who behaved in a normative manner (2 < .005).

Individuals without prior krnowledge reciprocated less to the supervisor when s/he violated
a norm than when s/he followed the rule (p_ < .05).
Contrast analyses were performed to compare these means to the no norm control

CM = +5.30). They revealed that reciprocity in the norm violation, no prior krnowledge

74

condition was significantly less than reciprocity in the control (p < .05). Reciprocity in the
norm confirmation, prior knowledge condition was also significantly lower (p_< .05).
Based on the overall analyses for these two interactions, it appears that the felt obligation
pattern of reciprocity occurs under conditions of prior knowledge of reciprocity whereas a
reactance pattern appears under conditions of no prior knowledge, with the former being
slightly more robust.

As mentioned, there was also an expectancy by norm by krnowledge interaction, E
(1, 204) = 3.82, p = .05. Some interesting patterns emerged. As seen in Table 3.7, an
additive reactance pattern seems to occur under conditions of no prior knowledge. In
other words, reciprocity decreased with each violation that was committed by the
supervisor when participants did not know ahead of time that they would have a chance to
reciprocate. However, when participants did have prior knowledge of the opportunity,
they appeared to reciprocate more when the supervisor violated either a norm, an
expectation, or both. This latter pattern is consistent with the high threshold of felt
obligation notion and can also be seen in Table 3.7. With just one violation the supervisor
exceeded the participarnts' threshold for feeling obligated to reciprocate, leading to higher
reciprocity in any of the violation conditions.

Simple effects tests were performed to examine these patterns more closely. First,
the simple main effects of expectation arnd norm were examined within the no prior
knowledge conditions. Consistent with the additive reactance pattern, under conditions of
no prior knowledge, reciprocity was lower in the face of a norm violation than a norm

confirmation (2 < .05). The mean difference for expectancy violation and confirmation did

75

not reach significance, however (p = .16). Thus, the additive reactance pattern is
supported with the exception of the expectation main efi‘ect falling short of significance.
The felt obligation pattern under conditions of prior knowledge of an opportunity
to reciprocate was also tested using simple effects analysis. Both expectancy and norm
simple nnain effects emerged. Participants reciprocated more when the supervisor was

unexpectedly generous than when s/he was consistently generous (p < .05). There was

Table 3.7

Reciprocity Towgd the Favor-Giver a_§ a Function of Expectation, Norm, and Prior
Knowledge of in Onportupity to Reciprocatg

 

 

 

 

Expectation
Norm Confirmation Violation
No Prior Knowledge
Confirmation +7.00 r,‘ +5.76 .r,
Violation +4.33 . +3.08 .
Prior Knowledge
Confirmation +1.92 .‘ +5.63 .,
Violation +6.60 b +6.11 b

 

Holt; Reciprocity was measured as the number of tickets given to the supervisor
subtracted by the number of tickets kept for oneself. Thus, higher numbers indicate
geater reciprocity. Within the no prior knowledge condition, means that do not share
subscripts differ significantly at p < .05. Also, those marked with a ‘ are significantly
different from the analogous no knowledge, no expectancy, no norm control M = +2.90)
at p < .05. \Vithin the prior knowledge conditiorn, means that do not share subscripts
differ significantly at p < .005. And those marked with a ° are significantly different from
the prior knowledge, no expectancy, no norm control M = +7.36) at p < .0001.

76

also greater reciprocity toward the supervisor when s/he broke a rule to be generous than
when s/he followed a rule to be generous (p < .005). Consistent with the threshold of felt
obligation pattern, there was geater reciprocity when the supervisor committed one or
more violations than when s/he was consistent and normative in his/her generosity (ps <
.005).

Contrast arnalyses compared the means in no krnowledge condition to the mean in
the analogous no prior knowledge, no expectation, no norm control M = +2.90). Only
the norm confirmation means were significantly different fi'om the control (ps < .05).
Contrast analyses also compared the means in the prior knowledge condition to the mean
in the prior krnowledge, no expectation, no norm control M = +7.36), which was
considerably higher than the mean in the no krnowledge control. Results indicated that
only the mean in the double-confirmation cell was significantly different fi'om the control
in the prior knowledge condition (p < .0001).

The ANOVA also revealed a significant main efi‘ect of importance, E (1, 204) =
11.36, p < .001. Consistent with the reactance notion, individuals who placed more
importance in their freedom to assign raffle tickets reciprocated less M = +3.88 tickets)
than those who placed little importance in this fieedom of behavior M = +6.38). No
other main effects were significant

The ANOVA also revealed a marginally significant expectation by importance
interaction, E (l, 204) = 2.90, p = .09, and a significant norm by importance interaction, E
(1, 204) = 5.51, p < .05. As seen in Tables 3.8 and 3.9, these patterns are opposite to
whatreactancetheorywould predict. Inlookingatthemeans, itappearsthatreactance

(decreased reciprocity in the face of a violation) was exhibited only by participants who

77

did not value their fieedom of behavior in allocating raffle tickets. Contrary to reactance
theory, a felt obligation pattern (increased reciprocity in the face of a violation) was
evident for participants who perceived their freedom of this behavior to be important.

Simple effects tests conducted on the expectancy by importance interaction
revealed that both the felt obligation pattern under high importance and the reactance
pattern under low importance did not reach significance (p_§ = .14, .34). Refer to the table
for the pattern of significant differences across the four means.

Simple effects tests for the norm by importance interaction found the felt
obligation pattern under high importance to be marginally significant (p = .08). Once
agairn, the reactance pattern for low importance participants was not significant (p = .12).

Refer to the table for other differences between the means.

Table 3.8

Reciprocity Toward the Favor-Giver as a Function of Expectation gaglrnpgrtance of
Freedom of Reciprocity

 

 

 

 

Expectation
Importance Confirmation Violation
Low +6.29 .‘ +6.51 r.
High +3 .09 . +4.41 ..,

 

Mg Reciprocity was measured as the number of tickets given to the supervisor
subtracted by the number of tickets kept for oneself. Thus, higher numbers indicate
geater reciprocity. Means that do not share subscripts differ significantly at p < .05.
Means marked with 8 ° are significantly different from the analogous no expectation
control M = +4.24) at p < .005.

78

Table 3.9

Reciprocity Towgd the FQ/or-Giver afl Function of Norm and Impoflnce of Freedom
of Recr'procriv

 

 

 

Norm
Importance Confirmation Violation
Low +6.83 c" +5.93 1,,
High +2.92 .‘ +4.58 .,

 

Ellie; Reciprocity was measured as the number of tickets given to the supervisor
subtracted by the number of tickets kept for oneself. Thus, higher numbers indicate
geater reciprocity. Means that do not share subscripts differ significantly at p s .08.
Means marked with a ' are significantly different from the analogous no norm control M
= +5.30) at p < .005. A ° indicates a marginally significant difference from the control.

Contrasts for the expectancy by importance cells found that only the low
importance, expectancy confirmation mean was significantly different than the no
expectancy control mean of +4.24 (9 < .005). Similarly, contrasts for the norm by
importance cells found that the low importance, norm confirmation mean was marginally
significantly geater than the no norm control mean of +5.30 (9 < .05). Contrast analyses
also found that the mean of the high importance, norm confirmation condition was
significarntly lower than the control (p < .05)

The final interaction revealed by the four-way ANOVA was a significant
expectation by norm by knowledge by importance interaction, E (1, 204) = 15.68, p <

.0001. To analyze this four-way interactiorn, separate expectancy by norm by knowledge

79

ANOVAs were performed for each level of importance. The means in the high and low
importance goups are presented in Tables 3.10 and 3.11, respectively. Caution must be
taken in interpreting these results due to the small sample sizes in some of the cells, as
indicated in the tables. The three-way interaction was only significant in the lnigh
importance goup, F (1, 90) = 18.67, p < .0001.10 The pattern of the means is consistent
with the previous suggestion that felt obligation occurs under conditions of prior
knowledge while reactance occurs under conditions of no prior knowledge.

For the high and low importance goups, simple effects tests were performed for
each knowledge condition. The threshold of felt obligation pattern under conditions of
prior krnowledge was supported for those placing high importance on their fieedom of
behavior. More reciprocity took place in the single- and double-violation conditions than
in the double-confirmation condition (ps < .0001). Thus, only a single violation was
needed for the participants to feel obligated enough to reciprocate somewhat generously.
The remaining tlnree means did not differ significantly (ps = .16 to .66). Reciprocity was
considerably low in the double-confirmation condition for high importance participants
with prior knowledge M = -5.00 tickets). This was same amount of reciprocity that took
place in the high importance, no expectation, no norm, no knowledge control M = -5.00).
Contrast analyses revealed that the remaining three means in the prior knowledge
condition were significantly geater than the control for high importance participants.

The trend of the means indicates that high importance participants tended to
exhibit a reactance pattern, but this pattern was not significant. Also, it was not additive,

as before. Ornly one violation was needed to elicit a reactance response. Reciprocity

 

1° The analyses revealed that for the low importance goup, F (1, 114) = 1.89.
80

Table 3.10

Reciprocity Toward the Fayor-Giver as a Function of Expectation, Norm, grid Prior

Knowledgg of an Qpportunig to Reciprocate for Recipients Placing High Immrtance in
Freedom of Reciprocity

 

 

 

 

 

 

Expectation
Confirmation Violation
Norm M 9 M E
No Prior Knowledge
Confirmation +5.67 r,’ 6 +1.67 ..,‘ 3
Violation +0.50 .° 4 +3.60 .r: 10
Prior Knowledge

Confirmation -5.00 . 5 +5.60 .,‘ 10
Violation +7.50 .3 8 +4.82 .3 11

 

1M3, Reciprocity was measured as the number of tickets given to the supervisor
subtracted by the number of tickets kept for oneself. Thus, higher numbers indicate
geater reciprocity. Witlnin the no prior knowledge condition, means that do not share
subscripts differ significantly at p $.05. “fithin the prior knowledge condition, means that
do not share subscripts differ significantly at p < .0001. Means in either goup marked by
a ' are significantly different from reciprocity in the high importance, no expectation, no
norm, no krnowledge control M = -5.00) at p < .05. A ° indicates a marginally significant
difl‘erence fiom the control.

81

Table 3.1 l

Reciprocity Towgm the Favor-Giver g a Function of Expectation, Norm, and Prior

Knowlgge of an Qppgrtunity to Reciprocate for Recipients Placing Low Immrtance in
Freedom of Reciprocity

 

 

 

 

 

 

 

 

Expectation

Confirmation Violation

Norm M B M B
No Prior Knowledge
Confirmation +8.60 r, 5 +7.13 t, 9
Violation +5.73 .5 11 +0.50 . 2
Prior Knowledge

Confirmation +6.25 .r, 8 +5.67 . 4
Violation +5.57 .5 7 +8.14 r, 6

 

_N_o;e_. Reciprocity was measured as the number of tickets given to the supervisor
subtracted by the number of tickets kept for oneself. Thus, higher numbers indicate
geater reciprocity. Within the no prior knowledge condition, means that do not share
subscripts differ significantly at p < 05. Within the prior knowledge condition, means that
do not share subscripts differ significantly at p < .05. None of the means are significantly
different fiom reciprocity in the low importance, no expectation, no norm, no knowledge
control M = +5.33) at p < .05.

82

appeared to be lowest in the single- and double-violation conditions. However, only one
of these means was significamly lower than the double-confirmation mean. This was the
mean reciprocity toward the supervisor who only violated the norm. Contrast analyses
comparing the no prior knowledge means with that in the control revealed that three of
them were significantly higher (ps < .05) and one was marginally so (9 = .08).

For low importance participants, one might expect less felt obligation. However,
reciprocity was quite high in the pridr knowledge, double-violation condition M = +8.14).
This mean was only significantly different fiom one of the other three, though. Moreover,
the mean in the doubleconfirmation condition was even slightly higher than the single-
violation means, although not significantly so. None of the means were significantly
different from the low importance, no expectation, no norm, no prior knowledge control
(M = +5.33).

In examining the low importance, no prior krnowledge means, a high threshold of
reactance pattern appears to have emerged in spite of participarnts' disinterest in receiving
rafle tickets. These participants reciprocated the least to the double-violator but one of
thesedifl‘erenceswasjust short ofnnarginal significance(p= .097). Also,theespecially
small in in the double-violation cell strongly reduces confidence in these results. Once

again, contrasts revealed none of these means to be significantly different from the control.

83

Supplementagy Finding

Temporary Affect. It was hypothesized that participarnts might experience a
decrement in mood due to a sense of reactance or felt obligation when the supervisor
violated a norm and/or expectation. This speculation was not supported for positive nor
negative mood. Positive mood was measured as the average of the happy, joyful, pleased,
and enjoyment/fun items. Cronbach's alpha for these items was .88, indicating high
internal reliability. Negative mood was measured by averaging all of the negative mood
items, namely, unhappy, depressed, frustrated, worried/anxious, angy/hostile. Cronbach's
alpha for these items was .85.

An expectation by norm by knowledge ANOVA was performed for each
dependent measure. The first ANOVA revealed only an expectancy main effect, E (1,
103) = 11.80, p = .001, such that participants with a supervisor who violated an
expectation experiernced more positive mood than those who had a supervisor who was
consisterntly generous. Only a knowledge main efl‘ect emerged for negative mood, E(1,
103) = 5.36, p < .05. Interestingly, participants experienced more negative mood when
they had prior knowledge of an opportunity to pay back the supervisor. Both of these
findings are contradictory to predictions based on either reactance or felt obligation.

Attributions. The flee-response attribution data were coded by rating the
participarnts' opinions regarding any reasons that they believed they received such a
generous paymernt. Responses were coded with regard to their fit into each of four
categories of attributions: situational attributions, recipient attributions, supervisor

attributions, or unique interaction/relationship attributions.

84

These categories were defined in a manner similar to those used by Worchel et al.
(1976). A response was coded as a situational attribution if the respondent suggested that
some pressures or aspects of the particular situation led to the payment behavior (e.g.,
"The supervisor was just following the rules” or "We only had ten minutes to draw the
pictures so it was only fair to get paid a lot”). A response was coded as a recipient
attribution if the participant suggested that it was his/her personality or manner of
completing the task that caused the supervisor's generous payment to occur (e.g., "I'm a
very creative person" or "The supervisor could tell that I worked hard on the task"). A
response was coded as a supervisor attribution if the participant believed that it was
something about the supervisor that made him/her behave generously (e.g., "The
supervisor is a generous person” or ”The supervisor was in a good mood"). A response
was coded as a unique interaction/relationship attribution if the participant believed that
there was something particularly special about his/her relationship with the supervisor that
led to the generous paymernt (e.g., "My pictures weren't that creative but maybe the
supervisor saw something special in them” or ”We have the same taste in art"). To gain a
more sensitive measure of attributions and to take into account the presence of mixed
attributions, each participant's overall response was rated to the extent that it had themes
of each attribution on a scale ham "0," not at all suggested as a reason, to '4," suggested
as being a very importarnt reason for the supervisor's behavior.

Three-way ANOVAs were performed for each of the four attributions. The first
was an expectancy by norm by knowledge ANOVA with situational attribution rating as a

dependent measure.ll As one might expect, a significant norm main effect emerged, E (1,

x

" Control conditions were not included in these analyses.
85

101) = 8.93, p < .005, such that participants rated situation as being a more important
reason for the supervisor's behavior when the rule to be generous was confirmed than
when it was violated. There was also a significant norm main effect on supervisor
attributions, E (l, 101) = 7.84, p < .01, as indicated by a subsequent ANOVA Consistent
with the previous finding, participants rated the supervisor's internal characteristics as
being less responsible for his/her generosity when the norm was confirmed than when it
was violated. Recipient attributions were only afl‘ected by knowledge of a reciprocity
opportunity, If (1, 101) = 7.69, p < .001. Specifically, participants believed themselves to
have a greater impact on the supervisor's behavior when they were in the no prior
knowledge condition than when they were initially aware of their chance to reciprocate.
Interestingly, all participants had reciprocated by the time they filled out the attributional
questionnaire but differences as a fimction of knowledge still emerged. No other main
efl‘ects or interactions were revealed, including no condition effects on the unique
interaction attribution, and no expectancy effects on any of the attributions made. Thus,
derogation of the double-confirming supervisor probably did not result as a function of the
generation of different types of attributions because no expectancy by norm interactions
were revealed.

Additional Findings. An unpredicted correlation between liking and reciprocity
toward the favor-giver emerged (r = 0.29, p < .0001). Therefore, increased reciprocity
toward the favor-giver was associated with liking of that person. Liking was also
examined as a covariate in an ANOVA with reciprocity as a dependent measure to see if it

accounted for any of the findings. It emerged as a significant covariate, F (l, 183) =

86

12.47, p < .0001 but those interactions that were originally significant only became slightly

more SO.

87

Chapter 4

DISCUSSION

Summag of Main Findings

The initial goal of this investigation was to attempt to replicate the finding that
individuals tend to derogate favor-givers who violate an expectation and/or norm to
provide a favor. However, the reverse pattern was found in the current study. Individuals
actually liked a favor-giver more when s/he violated a norm or an expectation than when
she behaved in an expected and/or normative manner. While most previous research has
found the opposite, and thus predicted, pattern, there were two previous investigations
that found that increased liking resulted when either unexpected or counternorrnative
favors were provided (Morse, 1972; Suls et al., 1981). Only one of these, though,
measured the reactions of the recipients, themselves (Morse, 1972). The present study,
however, was perhaps the first investigation to (unexpectedly) find that expectancy
violations and norm violations independently generated increased liking of a favor-giver.

Another major goal of the current study was to test three potential mechanisms

that could be involved in the predicted derogation of the favor-giver. These were
unpredictability as a negative attribute, felt obligation, and psychological reactance. The
first mechanism involves the possibility that an expectancy/norm violator might be
perceived of as unpredictable and that unpredictability is viewed as a negative attribute.
The felt obligation and reactance mechanisms were based on Gouldner‘s "norm of
reciprocity” (1960). The felt obligation mechanism suggests that derogation might occur

because individuals would resent a favor-giver who provides a big favor that causes them

88

to have strong feelings of obligation to reciprocate. When this mechanism is in operation,
individuals have a strong desire to reciprocate. The reactance mechanism suggests that
derogation might result fiom the experience of psychological reactance arising from a
restriction in one's freedom not to reciprocate as a result of receiving a big favor. When
reactance is experienced, individuals have a strong urge to withhold reciprocity. Contrary
to the predictions of all three of the mechanisms, a favor-giver who violated a norm or
expectation was actually liked more than one who behaved expectedly and normatively.
However, as mentioned earlier, derogation need not occur if reactance or felt obligation
mechanisms are in operation. Although the expected derogation did not occur in the
violation conditions, evidence for the three mechanisms was still examined. Some findings
were consistent with the ideas behind each of the mechanisms.

Unpredictability Mechanism: Attraction Findings Three findings in particular
supported the unpredictability hypothesis. First, a supervisor who violated an expectation
or a norm was, in fact, seen as more unpredictable than one who did not. Second, there
was a small but significant negative overall correlation betwear unpredictability and liking.
However, contrary to the hypothesis, the tendency to dislike an unpredictable favor-giver
was overlooked when recipients received a large favor as a result of this unpredictability.
In other words, in the single- and double-violation conditions, there was no correlation at
all between unpredictability and liking. In the no-violation conditions, however, a
moderate-sized negative correlation emerged. The third piece of evidence for the
unpredictability hypothesis was a finding involving participants' intolerance of ambiguity.
It was predicted that individuals scoring high on intolerance would especially derogate the

expectancy and/or norm violator. Participants scoring high on intolerance of ambiguity

89

did not derogate the violating favor-giver, but they did refiain from increasing their
evaluation of the supervisor when s/he violated an expectation, as low scorers did.
However, this pattern only occurred under conditions of no prior knowledge. Thus, the
overall support for the unpredictability notion was quite weak.

Felt Obligation and Hydrological Reactance Mechanisms: Attraction
Findings. The felt obligation and reactance hypotheses also received very minimal
support regarding their predicted effects on attraction toward the favor-giver. Both of
these notions led to the prediction that derogation would at least occur in the double-
violation, no prior knowledge condition. As mentioned earlier, however, this three-way
interaction did not emerge. Only the expectancy by norm interaction emerged and the
means were in the opposite direction than would be predicted by both felt obligation and
reactance. This was even true for individuals who placed high importance in their freedom
of reciprocity behavior. However, consistent with reactance and felt obligation
predictions, there was a marginally significant trend for high importance individuals to like
the double-violating supervisor less than low importance individuals did. Interestingly,
this pattern of greater derogation by high importance recipients was only significant for
reactions to the double-confirming supervisor. Another finding consistent with both felt
obligation and reactance was that under conditions of no prior knowledge, participants
who placed high importance in their fieedom to reciprocate liked the supervisor less than
those who did not value this fieedom. Overall, however, the support for these two

mechanisms was quite weak with regard to attraction toward the favor.doer.

Felt Obligation and Hydrological Reactance Mechanism: Reciprocity
Findings. The reciprocity data were more encouraging for the felt obligation and
reactance hypotheses. The unpredictability mechanism did not have implications for
reciprocity behavior. However, reciprocity was the key variable in distinguishing the
operation of felt obligation from that of reactance. Partial support for each mechanism
was obtained. Prior knowledge of an opportunity to reciprocate played an unexpected
and crucial moderating role. It determined whether recipients of the bigger favor would
exhibit reactance or felt obligation patterns of reciprocity behavior.

Specifically, when individuals had no prior knowledge of a reciprocity opportunity,
a reactance pattern emerged at the time of reciprocity. In other words, expectancy and/or
norm violations led to decreased reciprocity. However, when individuals did have prior
knowledge of an opportunity to reciprocate, they showed a felt obligation response. They
reciprocated more in the face of one or more violations. The trend of the means
suggested that the reactance efi‘ect was additive. A single violation was sufficient in
causing decreased reciprocity, but both a norm and expectancy violation combined led to
the least reciprocity. Felt obligation, however, produced a low threshold effect such that
either a single or double violation led to an urge to reciprocate generously in the prior
knowledge condition. Thus, the critical variable in determining whether reactance or felt
obligation will be experienced appears to be whether or not the recipient of a favor knows
ahead of time that s/he will have a chance to reciprocate.

Importance of behavioral freedom also played a substantial role in determining
reciprocity behavior. Consistent with reactance and contrary to felt obligation, a main

effect of importance revealed that participants reciprocated less when they placed high

91

importance in their freedom of reciprocity behavior than when they did not. However,
consistent with felt obligation, high importance participants reciprocated more when an
expectation or norm was violated than when one was confirmed. This last finding directly
contradicts reactance theory.

An expectation by norm by knowledge by importance interaction also yielded some
support for both mechanisms. Specifically, a felt obligation pattern emerged for both high
and low importance participants with prior knowledge. However, consistent with felt
obligation ideas, the pattern only reached significance for participants placing high
importance in fi'eedom of reciprocity. The reactance pattern for individuals with no prior
knowledge was also evident for both high and low importance individuals. For high
importance individuals, a single violation tended to be sufficient in eliciting a reactance
response. However, for low importance individuals, a double violation was needed to
elicit reactance. Surprisingly, though, only the latter pattern reached significance.

Overall, the reciprocity data provide some convincing evidence for the occurrence of felt
obligation under conditions of prior knowledge and reactance under conditions of no prior

knowledge.

92

Potential Exglanationg for Unpredicted Results

Because the results did not completely support any of the mechanisms, various
unpredicted findings deserve some explanation. First, there are several possrble reasons
why individuals did not derogate the violating supervisor. These are the possibility that
(1) the material value of the favor was not high enough, (2) the violations were not great
enough, (3) more interaction with the supervisor was needed, (4) the perception of an
ulterior motive was necessary, or (5) a threat to self-esteem was necessary.

With regard to the first explanation, perhaps the favor was simply not worth
enough to the recipients. If the favor were bigger (more tangibly rewarding), maybe the
recipients would have been more likely to either perceive the supervisor as being
unpredictable or to experience sufficient felt obligation or reactance. For example, if a
supervisor in a similar study violated an expectation and a norm by paying a worker $40
cash instead of a typical and mandatory payment of less than $10, different results may
have emerged. Perhaps this very valuable award would make the favor surprising enough
to cause the recipients to perceive the supervisor as being very unpredictable. This
increased unpredictability could then be sufficient enough to cause recipients to dislike the
violating supervisor in spite of the nice favor s/he provided. Similarly, perhaps this
exceptionally high payment would be suflicient enough to cause recipients to feel so
obligated to reciprocate that they would even resent the favor-giver given that no
opportunity to reciprocate is apparent. Likewise, perhaps only a very tangible reward
would elicit a feeling of restriction of free behaviors. However, this cannot explain why

the potentially moderate-sized favor in the current investigation not only failed to produce

93

derogation, but actually led to increased liking of the violating favor-giver. Therefore,
this explanation is plausible, but it may not be very likely.

The second potential explanation is a related issue. It suggests that the favor may
not have been perceived as large enough because the expectancy and norm violations that
took place were not big enough. For example, if a good fiiend who has been stingy his
entire life is all of a sudden very generous, this is quite a big expectancy violation. Or, if
someone breaks a law to steal a television set to give as a gift, this is a substantially large
norm violation. Either of these events should be more likely to activate any of the
mechanisms for derogation than the events that took place in the current investigation.
However, it must be noted that the expectancy and norm violations in the current study
did produce strong effects on the dependent variables. Because of this and because their
manipulation checks confirmed successfirl manipulation, this explanation may also not be a
very likely one.

The third explanation suggests that participants might have responded difi‘erently
had they had more interaction with the supervisor. For example, it is possrble that having
direct contact with the supervisor when s/he provided the favor might have elicited greater
discomfort than indirect contact because of felt obligation or reactance. Ifthe recipient
never even sees the favor-giver, s/he might not feel as obligated or as restricted because
both the favor and the supervisor are less threatening. Also, the salience of the violations
themselves may be greater if firture interaction is anticipated.

Some research is consistent with these suggestions. Kiesler, Kiesler, and Pallak
(1967) suggest that norm violators are disliked more when future interaction with them is

anticipated than when it is not. Moreover, Brehm and Brehm (1981) suggest that

94

anticipation of future interaction does elicit greater psychological reactance. And previous
research has found that the anticipation of meeting a donor results in a decreased
likelihood for a person to seek help (Nadler, 1979, as cited in Fisher et al., 1981) and more
defensive attributions for having received it (W orchel & Andreoli, 1974). Other research
has found that recipients who were not anonymous were also be more reluctant to seek
help (Nadler & Porat, 1978, Shapiro, 1978, as cited in Fisher et al., 1981).

Noteworthy as well is the fact that most of the studies that have found derogation
of a norm/expectation violating favor-giver employed procedures in which participants
actually had and/or expected to have some direct contact with that person (e.g., Kiesler,
1966; Schopler & Thompson, 1968; Worchel & Andreoli, 1974). While the remaining
two studies that found derogation did not involve face-to-face interaction, they still
involved more personal interaction than the current study. The recipients of the favor in
Clark and Mills's study (1979) believed that they were watching their partner on video.
The contestants in the Morse et al. (1977) study had vocal interaction with their
quizmaster over an intercom system. Thus, perhaps personal interaction or the
anticipation of such interaction facilitates derogation.

Participants in the current investigation were told that they would never meet their
supervisor. Thus, in all conditions, direct and future interaction were not anticipated.
Perhaps participants would have felt more obligation or reactance if they had direct
interaction with the supervisor or if they were told that they would be interacting later.
Under those conditions, derogation may have resulted fiom the expectancy and/or norm
violations. However, the reciprocity data in the current study suggest that felt obligation

and psychological reactance were experienced. One might argue, then, that additional

95

efforts to elicit these experiences (e.g., through the anticipation of firture interaction) were
not necessary in the current study. Perhaps, then, it was not insufficient arousal of
obligation or reactance that led to the unexpected attraction findings.

The fourth and fifth explanations involve two of the alternative potential
mechanisms described earlier. The fourth explanation presents the possibility that the
perception of an ulterior motive drives the derogation of a violating favor-giver. The
current experiment, unlike most previous research, took great efi‘orts to minimize the
chance that recipients would perceive an ulterior motive for the supervisor’s generosity.
For example, participants were told that their supervisor was not aware that s/he would be
evaluated or assigned payment by the participant. These efforts were assumed to be
successfirl because responses to the free-response attributions measure suggest that
participants did not attribute the favor to any ulterior motives. Perhaps if the supervisor's
motives were left ambiguous, an ulterior motive would have been automatically suspected.
Consequently, "good” attributions for the favor may have been made. Consistent with this
notion, there is some evidence in the current study that suggests that some recipients
simply attributed the favor to the supervisor's kindness or generosity.

Because the current study controlled for this alternative potential mechanism and
did not find derogation, there is a chance that it may be responsible for some of the
research that does find derogation. Thus, future research should examine the mechanism
more systematically. Ifresearch finds that ulterior motive attn‘butions do generate
derogation, then it should also address whether this efi‘ect is mediated by increased

psychological reactance or felt obligation.

96

The fifth explanation is the suggestion that derogation results fiom a threat to self-
esteern that individuals might experience upon receiving large favors (see Fisher et al.,
1981). As mentioned earlier, receiving a favor may result in the perception of failure,
inferiority, or dependence, which could then cause negative afi‘ective responses such as
derogation of a favor-giver. The potential for this type of threat was minimized in the
current investigation, however. Participants should not have perceived feelings of failure
because all participants received a moderate creativity score as compared to previous
participants (as noted on the ”Supervisor Record Sheet"). Moreover, the favor that was
provided did not involve helping the participants with the actual creativity task. lfit did, it
might have suggested to them that the favor-giver perceived inadequacies in the worker's
ability. Also, the creativity task may not have been very ego-involving to most
participants. Thus, the current investigation made efi‘orts to reduce the potential for a
threat of self-esteem to take place and derogation of the violating favor-giver did not
occur. Perhaps, then, a threat to self-esteem is a necessary requisite for derogation to
occur, whether alone or in combination with any other.

Although these last two explanations suggest that some alternative mechanisms
may have played a role, the other explanations suggest that some extraneous factors may
have interfered with the operation of the original three mechanisms. In other words, not
enough unpredictability was perceived or not enough felt obligation or psychological
reactance was aroused because of some interfering variable. There is some evidence, in
fact, that the violating supervisor may not have been perceived of as being unpredictable
enough. Although higher than in the no violation conditions, unpredictability ratings in the

violation conditions were actually near the midpoint of the scale.

97

Also, it is possible that with stronger feelings of felt obligation or reactance,
derogation may be more likely to take place as it has in previous research. Unfortunately,
the extent to which participants may have experienced felt obligation and reactance in the
current study cannot be determined. There are several reasons, though, that this
explanation may not be a likely one. First, as mentioned earlier, felt obligation and
reactance were sufficiently aroused to afi‘ect reciprocity behavior in the current
investigation. Thus, recipients did show evidence of these experiences. Moreover, Brehm
and Brehm (1981) suggefi that it takes greater amounts of reactance to cause failure to
reciprocate than to elicit derogation. Consistent with this, Morse et al. (1977) found
derogation but did not find reactance effects on reciprocity. Because failure to reciprocate
occurred in the current investigation (in the no knowledge condition) and derogation did
not, it cannot be a matter of degree of arousal if derogation happens more easily. The
same might apply to felt obligation as well.

Moreover, if importance in the freedom to reciprocate is suggested to be
associated with increased obligation and reactance, then derogation of the violating favor-
giver should at least have occurred in the current study for high importance recipients.
However, this was not the case in the current investigation. Most of the data even directly
opposed this spewlation. Also, as mentioned earlier, both felt obligation and reactance
are primarily motivations with affective components that are only secondary in nature.
Thus, reciprocity behavior is a greater indication of their operation than is derogation of a
favor-giver. Although plausible, it is not very likely, then, that derogation did not occur

because of insuflicient arousal of felt obligation or psychological reactance.

98

Perhaps, then, perceived unpredictability, felt obligation, and reactance sometimes
result in liking, rather than derogation, of someone who provides a big favor. Although
there is some literature that suggests that derogation does occur, there is also some
literature that suggests that, if anything, greater liking may result fiom these mechanisms.
As mentioned earlier, Morse (1972) and Suls et al. (1981) found increased liking toward a
violating favor-giver. Moreover, other research has found that when help is costly to the
favor-giver, the recipient tends to perceive him/her as being more caring, concerned, and
generous (Fisher & Nadler, 1976, Nalder & Fisher, 1978, as cited in Fisher et al., 1981),
and, perhaps consequently, likes him/her more (Gergen et al., 1975). Greater liking of a
favor-giver also results when help is apparently deliberate or voluntary (Goranson &
Berkowitz, 1966; Gross & Latané, 1974; Nemeth, 1970).

These results do not suggest that the three mechanisms were not involved,
however. In fact, some of these same investigations also found evidence for feelings of
obligation and/or reactance. Research has found that voluntary help, costly help, and high
amounts of help do, in fact, result in greater feelings of obligation to reciprocate
(Berkowitz & Friedman, 1967, Fisher & Nadler, 1976, Greenberg & Bar-Tal, 1976, Kahn
& Tice, 1973, Levanthal, Weiss & Long, 1964, Pruitt, 1968, Stapleton, Nacci, &
Tedeschi, 1973, as cited in Fisher et al., 1981; Gergen et al., 1975; Goranson &
Berkowitz, 1966; Greenberg & Frisch, 1972; Gross & Latané, 1974; Nemeth, 1970; Wilke
& Lanzetta, 1982). Thus, perhaps mechanisms such as felt obligation ofien produce
positive or neutral reactions to a favor-giver along with the expected motivational
response. Future research should be directed at identifying and exploring any moderating

variables that may account for the inconsistencies in the literature. Aside fi'om the

99

potential variables mentioned here, some research also suggests that various individual
difi‘erences may help determine whether receiving large favors is viewed favorably or
unfavorably (Fisher et al., 1981; Gergen, 1974). Thus, there are many possible reasons
that may explain why derogation of the violating favor-doer did not take place in the
current study.

Another point that deserves attention is the lack of an interaction between
knowledge of reciprocity and the norm and expectancy manipulations for attraction. One
reason, though perhaps not a likely one, is that knowledge did not play a role because not
enough unpredictability, felt obligation, or reactance was aroused to elicit derogation.
Had derogation of the violating favor-giver occurred, it is possible that knowledge of
reciprocity would have played a role moderating this response. Another possible reason is
based on a previous investigation that found that having no opportunity to reciprocate
resulted in negative consequences only for high self-esteem recipients (Nadler, 1986).
However, self-esteem was not measured in the current investigation. It is also possible
that the knowledge of reciprocity manipulation was not strong enough in the current
study. The experimenters noticed that some participants seemed preoccupied with filling
out their rafile tickets when the knowledge manipulation was introduced. Moreover, the
manipulation check did not reveal a strong awareness of the knowledge condition by
participants. However, prior knowledge was involved in some highly significant
interactions affecting both attraction and reciprocity, so the manipulation must have
affected many participants.

Another finding that deserves firrther examination is the fact that unpredictability

and liking were only negatively correlated in the no violation conditions. It was expected

100

that this negative correlation would be especially high in the violation conditions but it was
actually nonexistent in these circumstances. Again, perhaps this was because the
violations that took place were not great enough. The reason the correlation emerged in
the no violation conditions is because of people's "natural" tendency to dislike
unpredictable others. In the violation conditions, this tendency was simply overlooked in
the face of receiving the seemingly larger favor. Had the favor been even more unusual,
unpredictability may not have been so overlooked.

Some of the major findings of this investigation demonstrated evidence for both
felt obligation and reactance effects on reciprocity. As mentioned, felt obligation patterns
only emerged when participants knew ahead of time that they would get the chance to
reciprocate. Psychological reactance effects only happened when they did not have this
knowledge.

A plausible explanation for this interesting pattern is that once recipients learn of a
(big) favor granted by a norm/expectancy violating benefactor, they experience
psychological reactance. When they are not aware of an upcoming reciprocity
opportunity, reactance is maintained because the recipients see no opportunity to restore
freedom of behavior. When they do know that they will be able to reciprocate, however,
reactance is dissipated as a result of an expectation of being able to reduce it in the near

future. '2

 

'2 The idea that reactance may dwindle with the passing of time stems from research
conducted by Crano & Messé (1970). Just as reactance is suggested to diminish over
time, these researchers found cognitive dissonance to diminish over time. Participants
were asked to write a counterattitudinal essay in order to receive a small or large
monetary award. When issue attitude was assessed immediately after writing the essay,
participants displayed the typical reverse incentive response. Those in the small reward
condition had significantly more attitude change in the direction of the essay. Because of
dissonance arousal, they attributed their essay-writing to their attitude rather than to their
101

Moreover, felt obligation may be aroused when the recipients are aware of a
reciprocity opportunity because there is more time to think about how much they owe the
favor-giver. Obligation then builds up over time. The recipients in this situation have time
to make a rational decision regarding how much reciprocity the favor-giver deserves.
Consequently, they reciprocate the most toward a favor-giver who goes out of his/her way
to provide the favor. However, when the favor-giver is simply being consistent and
following the rules, the recipients rationalize that they do not really owe the supervisor
much, if anything. This explains why reciprocity in the current investigation was especially
low in the double-confirmation cell when participants had prior knowledge. These
relationships are depicted in Figure l on the following page. However, this path diagram

is based on post-hoe ideas and should be tested empirically in the future.

 

insuficient payment. However, when they performed a distracter task and attitude was
not assessed until well after the manipulation was imposed, participants recipients
receiving the large payment displayed the most attitude change. It was suggested that
cognitive dissonance diminished over time for participants with insufficient pay
justification, leaving only reinforcement effects for having received money for writing the
essay. Similarly, perhaps reactance dissipated over time in the current study when
participants had prior knowledge of a reciprocity opportunity causing felt obligation to be
fiee to increase as time went on.

102

recipient receives
favor from expectancy
and/or norm violator

~lv

recipient perceives
favor as being large

~l¢

recipient experiences
psychological reactance

 

2

 

8

reactance is maintained reactance immediately
for recipients with dissipates for recipients
£9 prior knowledge with prior knowledge
recipient pays little back recipient has time to
to violating favor-giver think about how much
to restore freedom s/he owes favor-giver

4v

recipient gives violating
favor-giver much in return

Fggg' r_e l. Reciprocity Toward an Expectancy/Norm Violating Favor-Giver as a Function
of Recipients' Knowledge of an Opportunity to Reciprocate: A Path Diagram

103

Conclusions and Implications

In sum, the current investigation found that recipients of a favor liked their
benefactor more when s/he violated a norm or expectation to provide it. The study also
found that individuals do not like an unpredictable other unless they can benefit from their
unpredictability (see also Kiesler, 1973). This research also revealed that when recipients
of a favor know that they will later get the chance to reciprocate, they feel more obligated
to reciprocate favors that are out of the ordinary. Consequently, to the extent that
someone breaks a rule and behaves out of character to provide a favor, the recipient pays
back the favor-giver. But, when recipients do not know that they will have a chance to
reciprocate later, they experience increased psychological reactance as the favor becomes
more unusual. This results in lower reciprocity toward a favor-giver who violates a norm
and/or an expectancy than toward one who confirms them. As discussed in the previous
section, other findings consistent with felt obligation and reactance were also obtained.

This research has implications for many domains. It is important to understand
when individuals will respond negatively to receiving help or favors because aid-giving is a
necessary part of society. As Fisher et al. (1981) put it, "receiving help is an integral
component of the educational process, the health care system, the psychotherapeutic
process, the welfare system, and the recurrent patterns of day-to-day life" (p. 367). When
recipients experience negative reactions to receiving help or favors, they will probably not
benefit as much in these domains. For example, they might not learn as much in school or
at work; follow doctors' orders effectively; engage in enough disclosure in therapy, seek

financial help when it is needed; or be able to effectively engage in successfirl relationships

104

with fiiends, family, acquaintances, or others. Clearly, it is important to understand how
to minimize the potential for negative responses to receiving benefits from others.

Research in the area of reactions to expectancy and/or norm violations shows
promise for understanding some of the situations in which help is unappreciated or
aversive. Future research is needed, however, to illuminate more fully the role of any
potential moderators and mediators that may play a role in these negative reactions. Given
the rather complicated pattern of findings observed in the present investigation, as well as
across past studies, it is likely that future research on reactions to a favor-giver will
continue to generate interesting results. It is to be hoped that these efforts will ultimately
yield compelling and definitive evidence identifying the processes and mechanisms

underlying people's responses to the ”kindness of strangers. "

105

APPENDICES

106

APPENDIX A

107

APPENDIX A

Supervisor Record Sheets
(Expectancy Confirmation Condition)
Supervisor 0 4
Condition : 2 persgrhsegmte work space
Directions to participant: Circle the rating and the number of
raffle tickets your Supervisor gave you.

SUPERVISOR RECORD SHEET

 

 

 

 

 

 
    

M gquective Ratigg Number of Raffle
Number from -5 (very Legative reactiogL Tickets Awarded
to +5 (very positive reactiog)

1. -5-4-3-2-10+1@3+4+5 1234567@—0
2. -5-4-3-2Q0+1+2+3+4+5 123456700
3. -5-4-3-2-10@o2+3+4+5 123456700
4. -5-4-3-2-10+1.3+4+5 12345678@
5. -5-4-3-2-1@+1+2+3+4+5 1234567@910
6. -5-4-3-2-10@+2+3+4+5 1234567@10
7. -5-4-3-2-10+1@+3+4+5 12345673963
3. -5.4-3-2-1r@+2+3+4+5 1234567Q10
9. .54-3@-10+1+2+3+4+5 1234567010
10. -5-4-3-2-1 +1 2+3+4+5 1234567®10
11. -5-4-3-2-10+1+2+3+4+5 12345673910
12. -5-4-3-2-10+1+2+3+4+5 12345678910
13. -5-4-3-2-1o+1+2+3+4+5 12345678910
14. -5-4-3-2-10+1+2+3+4+5 12345673910

15.

-5-4-3-2-10+1+2+3+4+5

108

12345678910

(Expectancy Violation Condition)

Supervisor 0 4

Condition : 2 person. segarate work sgace
Directions to participant: Circle the rating and the number of

raffle tickets your Supervisor gave you.

Worker
Number

SUPERVISOR RECORD SHEET

Subiective Rating

from -5 (very negative regctiogL

to +5 mlpositive reactiog)

Number of Raffle
Tickets Awarded

 

!"

990399.959.»

H
9

11.

12.

13.

14.

15.

-5.4-3-2-10+1@3+4+5
-5 .4 -3 ago +1 +2 +3 +4 +5
-5 -4 -3 -2 -1 O®¢2 +3 +4 +5
-5-4-3-2-10+1@3+4+5
-5 .4 -3 -2 ugu +2 +3 +4 +5
-5 .4 -3 -2 -1 0@ +2 +3 +4 +5
-5-4-3-2-10+1@3+4+5
-5 .4 -3 -2 -1 @+2 +3 +4 +5
-5 .4 -3®-1 0 +1 +2 +3 +4 +5
-5-4-3-2-1 +1 2+3+4+5
-5-4-3-2-10+1+2+3+4+5
-5-4-3-2-10+1+2+3+4+5
-5-4-3-2-10+1+2+3+4+5
-5-4-3-2.10+1+2+3+4+5

-5 -4 -3 -2 -1 0+1 +2 +3 +4 +5

109

(3)45673910

a 45673910

0: 45678910

1@5673910

92345678910

    

(@345673910
12:)5673910
(3345673910
Cn345673910
{2b45678910
12345673910
12345673910
12345673910
12345673910

12345678910

(Expectancy Control Condition)

Supervisor # 5

Condition : 2 -gemon, separate work space
Directions to participant: Circle the rating and the number of

tame tickets your Supervisor gave you.
SUPERVISOR RECORD SHEET

Worker Subjective Ratng
Number from -5 (very neggtive reactiogj
to :0-5 (very positive reactiog)

1 -5-4-3-2-10+1 @+3 +4 +5
2 -5 -4 -3 -28 0 +1 +2 +3 +4 +5
3 54-3-24 0®+2 +3 +4 +5
4 -5 -4 -3 -2 -1 0 +1@ +3 +4 +5
5 -5 -4 -3 -2 ugu +2 +3 +4 +5
6 -5-4-3-2-10@+2+3+4+5
7 -5 -4 -3 -2 -1 0 +1 @+3 +4 +5
3 -5 -4 -3 -2 -1 0@ +2 +3 +4 +5
9 -5-4-3®-10+1+2+3+4+5
10 -5-4-3-2-10+1 +2 +3 +4 +5
11 -5 -4 -3 -2 -1 0 +1 +2 +3 +4 +5
12 -5-4-3-2-10+1 +2 +3 +4 +5
13 -5 -4 -3 -2 -1 0 +1 +2 +3 +4 +5
14 -5-4-3-2-10+1 +2 +3 +4 +5
15 -5-4-3-2-10+1 +2 +3 +4 +5

110

APPENDIX B

111

APPENDIX B

Measures
(Expectancy and Norm Manipulation Checks)

Please answer the following three questions regarding the Supervisor’s behavior.
For each item, indicate the number from the following scale that most accurately
describes your perception of the Supervisor’s actions.

1 = DISAGREE very much

2 = moderately disagree

3 = slightly disagree

4 = neither agree nor disagree
5 = slightly agree

6 = moderately agree

7 = AGREE very much

The Supervisor behaved as one might expect.

The Supervisor followed the instructions s/he was given precisely as directed.

The Supervisor behaved in accord with his/her past behavior.

112

(Measure of Supervisor Evaluation -- including Attraction and Unpredictability items)

This measure asks you to evaluate your Supervisor. The evaluation will provide us
with information about your Supervisor’s working style. Your ratings of the
Supervisor will be used strictly to help researchers understand the relationship
between personality factors and certain supervisor working styles. The evaluation
will not be used to judge the Supervisor in any way. With this in mind, please rate
your Supervisor on the following dimensions. For each item, circle the number from
0 to 9 that most accurately expresses your opinion of your Supervisor. You may not
have sufficient information to answer some of the questions with certainty. In these
cases, just try to guess, based on the Supervisor’s behavior, how you think s/he
might fall on the dimensions.

0 1 2 3 4 5 6 7 8 9
unlikable likable

mean kind

helpfirl unhelpful

unfiiendly fiiendly

0 l 2 3 4 5 6 7 8 9

continued on next page...

113

(Measure of Supervisor Evahration continued)

impulsive restrained

O l 2 3 4 S 6 7 8 9
obedient disobedient

rational irrational

0 l 2 3 4 5 6 7 8 9
easy-going temperamental

0 1 2 3 4 5 6 7 8 9
inconsistent consistent

typical unusual

mature immature

0 1 2 3 4 5 6 7 8 ' 9
conventional rebellious

annoying pleasant

0 1 2 3 4 5 6 7 8 9
submissive dominant
continued on next page...

114

(Measure of Supervisor Evaluation continued)

hostile calm

stubborn yielding

0 1 2 3 4 5 6 7 8 9
straightforward manipulative

115

(Departmental Task: Part I - Measure of Temporary Afi‘ect)

PBYBHDLDIY DEPARTMENT
MIDI-"IAN STATE UNIVERSITY

RESEARCH PARTIBIPATIDN SURVEY
1 998-2000

This is survey is part of a large scale research investigation being conducted by the ethics
committee of the psychology department at Michigan State University. This research
investigates participants’ reactions to research experiences. This questionnaire is being
administered in various research sessions being conducted by various experimenters. You
have been randomly selected to participate in this study. This data will provide the ethics
committee with information as to the mood state that participants generally experience
while involved in different types of research studies. This is got an evaluation of the
current study in which you are participating. If you have any further questions regarding
this departmental project, your experimenter can provide you with the appr0p1iate
contacts at the end of the session. Thank you for contributing to the university’s research.

 

PART 1: Please circle the appropriate mark to indicate how much you are feeling the
each of these emotions right now:

1. happy:
I l I I
not at all average extremely
2. joyfirl:
l I | |
not at all average extremely
3. pleased:
I l l I |
not at all average extremely

116

. enjoyment/hm:

not at all

. depressed:

notatall

7. fi'ustrated:

not at all

. worried/anxious:

I
not at all

. angry/hostile:

notat all

(Departmental Task: Part I continued)

average

average

average

average

average

average

extremely

extremely

extremely

extremely

extremely

extremely

 

117

(Departmental Task: Part H - Knowledge of Reciprocity Manipulation Checks)

PART 11: Please answer the following questions regarding research payment for your
participation in this departmental survey:

1. Will you be receiving any compensation (other than required course credit) for your
participation in this departmental survey? (please write ‘yes,’
‘no,’ or ‘don’t know’). Ifyou answered ‘yes’ to this question, please go on to answer
number 2.

2(3). Please check the type(s) of compensation you will be receiving:
cash payment
check payment
extra credit for a course
other: Please describe:

 

2(b). Will your current experimenter be in charge of deciding how much payment you will
receive for completing this task? (please write ‘yes,’ ‘no,’ or
don't know. ’)

 

2(c). If you answered ‘no’ to the question above, please indicate the person who will be
in charge of paying you for your completion of this departmental
survey:

 

2(d). Will you have any say in how much you will be paid for your participation in this
departmental survey? (please write ‘yes,’ ‘no,’ or ‘don’t
know’).

 

2(c). Will you have any say in how much anyone else will be paid for their participation in
this departmental survey? (please write ‘yes,’ ‘no,’ or
‘don’t know’).

 

2(1). Ifyou answered ‘yes’ to question 2(e), please indicate for whom you will be making
payment decisions (e.g., other participants in the study)

 

118

(Departmental Task: Part III - Importance of Behavioral Freedom)

3. How important is it for you to receive payment/compensation for completing this
departmental task?

I I I I I I I
extremely neutral extremely

unimportant important

4. How important is it for you to have a say in how much payment/compensation you
receive for completing this type of departmental task?

extremely neutral extremely
unimportant important

5. How important is it for you to have a say in how much payment/compensation others
will receive for completing this type of departmental task?

I I I I I I |
extremely neutral extremely

unimportant important

We appreciate your participation in this survey.
Please hand this completed questionnaire to your experimenter and continue
with the study in progress.

119

(Measure of Attributions)

Please write in your own words the reason(s) that you think the Supervisor gave you
the specific amount of payment that you received for completing the creativity task.
Please be as smific as possible.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

120

(Measure in the Belief of the Norm of Reciprocity)

Please answer the following questions regarding your opinions about certain human
virtues. You may consider all of these virtues as important, but we are interested in
learning which ones you believe are most important. Circle the number from 0 to 9
that most accurately illustrates the relative importance of the virtues below.

0
having concern
for family
members is
more important

0
waving those
who have been
nice to you is
more important

0
having respect
for one's elders
is more
important

0
being honest is
more important

0
having concern
for family
members is
more important

continued on next page...

121

9
being helpful is
more important

9
being honest is
more important

9
repaying those
who have been
nice to you is
more important

9
being helpfirl is
more important

9
having respect
for one's elders
is more
important

(Measure in the Belief of the Norm of Reciprocity continued)

0

repaying those who
for have been nice to you

0
being helpfirl is
more important

0
being honest is
more important

0
being helpfirl is
more important

0
having respect
for one's elders
is more
important

122

9
having concern
for family
members is
more important

9
having respect
for one's elders
is more
important

9
having concern
for family
members is
more important

9

repaying those
who have been
nice is more
important

9

being honest is
is more important

(Budner's Measure of Intolerance of Ambiguity)

Please answer the following questions regarding your general opinions. Use the
following scale to indicate how much you agree/disagree with the statements below:

1 = strongly disagm

2 = moderately disagree

3 = slightly disagree

4 = neither disagree nor agree
5 = slightly agree

6 = moderately agree

7 = strongly m

1. An expert who doesn't come up with a definite answer probably doesn‘t know
too much.

2. I would like to live in a foreign country for a while.

 

3. There is really no such thing as a problem that can‘t be solved.

 

4. People who fit their lives to a schedule probably miss most of the joy of living.

 

5. A good job is one where what is to be done and how it is to be done are
always clear.

 

6. It is more fun to tackle a complicated problem then to solve a simple one.

 

7. In the long run it is poss1ble to get more done by tackling small, simple
problems rather than large and complicated ones.

 

8. Chen the most interesting and stimulating people are those who don't mind
being different and original.

9. What we are used to is always preferable to what is unfamiliar.

 

10. People who insist upon a yes or no answer just don't know how complicated
things really are.

11. A person who leads an even, regular life in which few surprises or unexpected
happenings arise, really has a lot to be gratefirl for.

12. Many of our most important decisions are based upon insufiicient information.

continued on next page...

123

13.

14.

15.

16.

(Measure of Intolerance of Ambiguity continued)

I like parties where I know most of the pe0ple more than ones where all or
most of the people are complete strangers.

Teachers or supervisors who hand out vague assignments give a chance for
one to show initiative and originality.

The sooner we all acquire similar values and ideals the better.

A good teacher is one who makes you wonder about your way of looking at
things.

124

(Measure of Suspicion - including fillers)
This type of questionnaire is typically administered at the end of many research
experiments. Please answer the following standard questions regarding your
opinions about the study in which you have just participated.

1. Did the researcher in charge treat you in a courteous, professional manner?

Ifno, please explain:

 

 

 

2. Were the instructions that you were given presented in a clear and understandable
manner?

Ifno, please explain:

 

 

 

3. Was the purpose of this study explained to you in an honest and straightforward
manner?

If no, please explain:

 

 

 

4. Please describe, in your own words, what you behave was the purpose of the study in
which you have just participated:

 

 

 

 

 

 

 

125

APPENDIX C

126

APPENDIX C

Table A.l

Factor Loadings fiom Principgl-Components Angysis

 

 

 

Factor loading
Item: Attraction Factor Unpredictability Factor
likable (unlikable) .79 .07
kind (mean) .74 .17
unhelpfirl (helpfirl) - .52 .04
fiiendly (unfriendly) .76 .16
bad (good) - .72 -.03
unpredictable (predictable) - .04 .62
conforming (independent) - .31 -.43
honest (dishonest) .44 - .45
restrained (impulsive) - .17 -.67
disobedient (obedient) - .05 .73
irrational (rational) - .31 .61
temperamental (easy-going) - .73 - .02
consistent (inconsistent) .00 -.73
unusual (typical) .00 .78
immature (mature) - .42 .58
rebellious (conventional) .06 .82
pleasant (annoying) .82 .01
dominant (submissive) .11 .24
calm (hostile) .69 -.18
yielding (stubborn) .51 -.16
manipulative (straightforward) - .33 .40
Eigenvalues 5.19 4.60
Percent Variance Explained 24.71 21.89

 

Note: Variables with loadings of .50 or higher on a factor were selected for inclusion if
theoretically valid. Those with loadings of .40 to .49 were only included if they were
theoretically valid and they contributed to intanal consistency.

127

APPENDIX D

128

APPENDIX D

Removal of Problem Data

Data from 300 participants were collected for this investigation. However, 16% of
the data were dropped due to various problems. 2% of the data were excluded on the
basis of confirsion or disinterest. These participants were not able to follow the directions
and/or fill out the measures properly. For example, one participant did not speak English
and consequently was not able to follow directions. Another participant informed the
experimenter that she was not listening to the directions and later it became evident that
she did not even realize that the raflle tickets she received were hers to keep. Another 6%
of data were excluded due to problems during the sessions. Because the study was so
complicated, the experimenters made mistakes on occasion. For example, in two of the
sessions, the experimenter forgot to provide the payment to the workers at the appropriate
time. The workers themselves also contributed to problems occurring during the sessions.
Although participants were urged to remain quiet, they often blurted out comments when
receiving their payment, particularly during the violation conditions. Some said out loud
(in fiont of other workers) how much they were going to pay their supervisor back. Some
said out loud that they received more tickets than they were supposed to get or that their
supervisor was not consistent. Clearly these problems may have revealed the
manipulations or influenced other workers' reciprocity behavior. Due to the highly
deceptive nature of the study, another 8% of data were eliminated due to suspicion.
Participants' suspicion was most evident in their free response measures, particularly the

item included to assess such dubiousness. prarticipants indicated an awareness of the

129

manipulations or the dependent measures, their data were marked as suspicious. Their
data were also dropped if they doubted the existence of their supervisor. Overall, 16% of
data were eliminated due to confirsion, disinterest, suspicion, or experimental problems.

Only data fi'om the remaining 251 participants were included in all statistical analyses.

130

APPENDIX E

131

Table A.2

APPENDIX E

Analysis of Variance Results

The Efl‘ectsgf Expectancy, Norm, and Knowledg 0f Reciprocity Opportunity

on Attrgction Toward the Favor-Giver

 

 

Source sit“ 55. MS .E E0)
Expectation (E) 1 4.979 4.979 4.099 *“ 4.032 **
Norm (N) 1 3.922 3.922 3.229 * 3.176 *
Prior Knowledge (K) 1 1.265 1.265 1.041 1.024
E X N 1 7.673 7.673 6.316 ** 6.213 *"
E X K 1 0.954 0.954 0.785 0.772
NXK 1 1.360 1.360 1.120 1.101
E X N X K 1 0.428 0.428 0.352 0.347
Error 101 122.695 1.215
Pooled Error 227 280.355 1.235

 

_N_ot_g F (r) refers to the F re-computed using the pooled error fi'om all conditions
(including controls). In this case, the pooled error was unusually higher than

the error fi'om the experimental conditions only. F’s with p values < .05 are
marked with ". Those that are marginally significant (p S .09) are marked

with a *.

132

Table A.3

Attraction Towarrd the Favor-Givergg Function of Expectgtion g1 Norm Withor_r_t
[_Jppredictgbility as a Covariate (ANOVA megn_s_)

 

 

 

Expectation
Norm Confirmation Violation
Confirmation 5.08 .° 5.99 .,
Violation 5.99 b 5.90 1.

 

1:133 Attraction was based on ten items, each ranging from 0 to 9. The means for the
average score of these ten items are presented in the table. High scores indicate more
attraction. Means that do not share subscripts differ significantly at p < .01. Means that
share subscripts do not differ significantly (23 = .72 to .98). A ' indicates marginal
significant difl‘erence fiom the control (M = 5.47), p = .08.

133

REFERENCES

134

REFERENCES

Brehm, J. W. (1966). A theory of psychological reactance. New York: Academic
Press.

Brehm, S. S., & Brehm, J. W. (1981). Psychological reactance: A theory of
freedom 4a_nd Control. New York: Academic Press.

Brehm, J. W., & Cole, H. (1966). Effect of a favor which reduces freedom.
Joumal of PemrLahtygnd Socig Psychology. 3(4), 420-426.

Broll, L., Gross, A. E., Piliavin, I. (1974). Effects of offered and requested help
on help seeking and reactions to being helped. J ourngl of Applied Social Psycholggyg
4(3), 244-258.

Budner, S. (1962). Intolerance of Ambiguity. Joumgl of PersonalityJO, 29-50.

Carlsmith, J. M., & Aronson, E. (1963). Some hedonic consequences of the
confirmation and disconfirrnation of expectancies. J oumgl of Abnormal and Social

Pwhology, 2, 151-156.

Castro, M. (1974). Reactions to receiving aid as a fimction of cost to donor and
opportunity to aid. Journal of Applied Socjg Psychologygi, 194-207.

Clark, M. S., & Mills, J. (1979). Interpersonal attraction in exchange and
communal relationships. Jourpgl of Persongl_ity_and Socigl Psychology, 37(1), 12-24.

Crano, W. D., & Messé, L. A (1970). When does dissonance fail? The time
dimension in attitude measurement. Jouma_l of Personglity, 38(4), 493-508.

Diener, E., & Emmons, R. A. (1984). The independence of positive and negative
affect. Journal of Personality and Socigl Psychology, 47, 1105-1117.

Fisher, J. D., DePaulo, B. M., & Nadler, A (1981). Extending altruism beyond
the altruistic act: The mixed effects of aid on the help recipient. In J. P. Rushton, & R M.
Sorrentino (Eds), Altruism and helpingbehgayior: Socigl, persomgnd developmengal
W (pp. 367-502). I-Iillsdale, NJ: Lawrence Erlbaum Associates.

Fortes, M. (1983). Rules And the emergence of society London, England: Royal
Anthropological Institute.

Gergen, K. J. (1974). Toward a psychology of receiving help. Joummf Applied
Social chhology, 4(3), 187-193.

135

Gergen, K. J., Ellsworth, P., Maslach, C., & Seipel, M. (1975). Journal of
Personality apd Socgl ngchology, 31(3), 390—400.

Gergen, K. J ., & Jones, E. E. (1963). Mental illness, predictability, and affective
consequences as stimulus factors in person perception. Journal of Abnormal and Social

chhology, 67(2), 95-104.

Goflinan, E. (1959). The presentjation of self in everygiy life. Garden City, NY:
Doubleday.

Goranson, R. E., & Berkowitz, L. B. (1966). Journal of Personglity apd Socia_l
Psychology, 3(2)m 227-232.

Gouldner, A. (1960). The norm of reciprocity: A preliminary statement.
Americap Sociological Review, 25, 161-178.

Greenberg, M. S. (1980). A theory of indebtedness. In K. Gergen, M. S.
Greenberg, & R H. Willis (Eds), Social exchange: Advgnces in theory and research. 3-
26. New York: Plenum Press.

Greenberg, M. S., & Frisch, D. M. (1972). Effect of intentionality on willingness
to reciprocate a favor. Jourgal of Experimental Social Psychology,_8, 99-111.

Gross, A E., & Latané, J. G. (1974). Receiving help, reciprocation, and
interpersonal attraction. Journgl of Applied Social Psychology, 4, 210-223.

Kerr, N. L. (2000). Beyond Social Loafing: Motivation Gains in Groups.
Presidential Address presented at the meeting of the Midwestern Psychological
Association, Chicago, IL.

Kiesler, S. B. (1966). The effect of perceived role requirements on reactions to
favor—doing. 30011111 of Experimental Soci_a_l Psychology,_2, 198-210.

Kiesler, S. B. (1973). Preference for predictability or unpredictability as a
mediator of reactions to norm violations. Journal of Personality and Soci_a_l Psycholcgy,
£6), 354-359.

Kiesler, C. A, Kiesler, S. B., & Pallak, M. S. (1967). The effect of commitment
to firture interaction on reactions to norm violations. J ourngl, of Personalig, 35(4), 585-
599.

Messé, L. A, Kerr, N. L., & Sattler, D. N. (1992). ”But some animals are more
equal than others": The supervisor as a privileged status. In S. Worchel, W. Wood, & J.
A Simpson (Eds. ), Group processes and productivity. (pp. 203-223). Newbury Park:
Sage.

136

Michelini, R L., & Messé. (1974). Reactions to threat as a firnction of equity.
Sodom 37(3), 432-439.

Morse, S. J. (1972). Help, Likeability, and Social Influence. 1011:1131 of Applied
Social Psychology, 2(1), 34-46.

Morse, S. J., & Gergen, K. J. (1971). Material aid and social attraction. Journal
of Applied Sopiayl Psychology, 1, 150-162.

Morse, S. J., Gergen, K. J ., Peele, S., & van Ryneveld, J. (1977). Reactions to
receiving expected and unexpected help fiom a person who violates or does not violate a
norm. Joumpl of Experimental SOCflPsycholgv, 13, 397-402.

Nadler, A. (1986). Help seeking as a cultural phenomenon: Differences between
city and kibbutz dwellers. J0um_al of Personality afl Sogial Psychology, 51(5), 975-982.

Nadler, A, Fisher, J. D., & Streufert, S. (1974). The donor‘s dilemma: recipient
reactions to aid fiom fiiend or foe. Jourgl of Applied Social Psychology, 4(3), 275-285.

Nemeth, C. (1970). Effects of free versus constrained behavior on attraction
between people. Joumpl of Personalig and Social Psychology, 15, 302-311.

Pallack, M. S., & Heller, J. F. (1971). Interactive effects of commitment to future
interaction and threat to attitudinal freedom. J0um_a_l of Personality and Social

Pachology, 17, 325-331.

Piliavin, I., & Gross, A. E. (1977). The effects of separation of services and
income maintenance on AFDC recipients. Social Service Review, 51(3), 389-406.

Schopler, J ., & Thompson, V. D. (1968). Role of attribution processes in
mediating amount of reciprocity for a favor. Joumgl of Personality and Social

chhology, 10(3), 243-250.

Steiner, I. D. (1972). Group proces_s_gnd productivity. New York: Academic
Press.

Suls, J., \Vrtenberg, S., & Gutkin, D. (1981). Evaluating reciprocal and
nonreciprocal prosocial behavior. Developmental changes. Personalig and Social

chhology Bullg'p, 7(1), 25-31.

Thibaut, J. W., & Kelley, H. H. (1959/1986). The social pgchology of ggoups.
New Brunswick, NJ: Transaction, John Wiley & Sons, Inc.

Uehara, E. S. (1995). Reciprocity reconsidered: Gouldner’s 'Moral Norm of

Reciprocity' and social support. Journal of Social and Personal Relationships; 12(4), 483-
502.

137

Wilke, H., & Lanzetta, J. T. (1982). The obligation to help: factors affecting
response to help received. Europeryn Journal of Social Psychology, 12, 315-519.

Worchel, S., & Andreoli, V. A. (1974). Attribution of causality as a means of
restoring behavioral fieedom. J0um_al of Personaligg and Socigl Psychology, 28(2), 237-
245.

Worchel, S., Andreoli, V., & Archer, R (1976). When a favor is a threat to
fieedom: The effects of attribution and importance of freedom on reciprocity. Journal of
Persopglity, 44(2), 295-310.

 

138

"‘I'IIIIIIIIIIIIIIIIIIIIIIII