WWUHIWHHHNINIWINIHWIWNIIHHUWHHI

        

.b-r
01-h

3
6

    

  

._|
(I)

THE :1! 3-}

i 1390

Illlliliilll‘l'illllllllllllllilllliil'llllil

3 1293 02074 1751

 

LIERARY

Michigan State
University

 

 

 

This is to certify that the
thesis entitled
Self-Concept Organization and the Suppression of
Self-Relevant Thoughts

presented by
Jeanette M. Renaud

has been accepted towards fulfillment
of the requirements for

Master’s Psychology

degree in

 

 

OWN

k/Naior pro so:
Date 20‘05"??

 

0-7 639 MS U i: an Affirmative Action/Equal Opportunity Institution

PLACE 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

 

(€353 Q 832302

 

 

 

 

 

 

 

 

 

 

 

 

 

moo Mass-p.14

 

SELF-CONCEPT ORGANIZATION AND THE SUPPRESSION OF
SELF-RELEVANT THOUGHTS

By

Jeanette M. Renaud

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF ARTS
Department of Psychology

1 999

ABSTRACT

SELF-CONCEPT ORGANIZATION AND THE SUPPRESSION OF
SELF-RELEVANT THOUGHTS

By
Jeanette M. Renaud

The rebound effect associated with thought suppression has been found in
research focusing on the suppression of both novel stimuli and stereotypical
thoughts. However, research examining the suppression of self-relevant
thoughts has been less successful in demonstrating the effect. Kelly and Kahn
(1994) suggest that this may be because individuals have more experience
suppressing self-relevant thoughts and, thus, have developed a network of
successful distracters to use during times of suppression. The current study
examined the role of self-concept organization in affecting successful thought
suppression. As predicted, following suppression of self-relevant information
while using other aspects of one’s life as distracters, lower self-complexity was
related to greater rebound effects than was greater self-complexity. A number
of ancillary hypotheses and exploratory hypotheses were also examined, and

their implications are discussed.

Dedicated to Susanne Atherton

ACKNOWLEDGMENTS

This thesis could not have been completed without the help and guidance of
a few but very important people. In particular, I would like to thank Larry Messe
and Norb Kerr for serving on my master’s committee and providing me with
challenges and insight that I may have missed otherwise.

I would also like to thank my advisor and the chair of my master’s
committee, Allen McConnell. I appreciate not only your guidance on my
master’s thesis, but also for everything that you have done to help me to
improve my work over the past two years. Your guidance has been invaluable.

In addition, I would like to thank Dan Skiver for his unconditional love and

support during the past few years. Without it, I could not have completed this

thesis.

TABLE OF CONTENTS

LIST OF TABLES .............................................. vi
LIST OF FIGURES ............................................. vii
INTRODUCTION ............................................... 1
METHOD ..................................................... 18
RESULTS .................................................... 27
DISCUSSION ................................................. 37
APPENDICES ................................................. 51

REFERENCES ................................................ 72

LIST OF TABLES

Table 1. Descriptive Statistics for Primary Variables across All
Participants .................................................. 61

vi

LIST OF FIGURES

Figure 1. Interaction Between Self-Complexity and Distracter Type
for Number of Mentions ........................................ 63

Figure 2. Marginal Interaction Between Self-Complexity and Suppressers
Versus Expressers for Creative Task Preference .................... 65

Figure 3. Marginal Interaction Between Self-Complexity and Suppressers
Versus Expressers for Number of Student-Related Words Created ...... 67

Figure 4. Marginal Interaction Between Self-Concept Organization
and Positive Importance for Self-Judgment ......................... 69

Figure 5. Interaction Between Self-Concept Organization and Positive
Importance for Creative Task Preference ........................... 71

vii

INTRODUCTION

Throughout our daily lives, we are often faced with the desire to avoid
certain thoughts. For example, we may want to avoid thoughts about
relationship problems, professional failures, and social embarrassments.
Ironically, research has shown that attempts to suppress unwanted thoughts
are often met with undesired consequences. For example, Wegner, Schneider,
Carter, and White (1987) found that participants who attempted to suppress
thoughts of a white bear subsequently thought about a white bear more often
than those who had not attempted to suppress such thoughts. This
consequence of thought suppression is known as the rebound effect. In
particular, it refers to the tendency to think about a previously suppressed
thought more after initial constraints to suppress it are removed than when such
initial constraints were never in place.

Wagner’s (1994) model of mental control suggests that this effect is a result
of two processes that operate when one is engaged in thought suppression.
These processes work in tandem but have different goals and are referred to
together as “ironic processes.” The goal of the automatic monitoring process is
to scan consciousness for evidence of the unwanted thought, whereas the goal
of the controlled operating process is to search for thoughts that will distract the
mind from the unwanted thought. Ironically, each time the monitoring process
encounters evidence of the unwanted thought, it focuses the mind’s attention on

the unwanted thought, sabotaging the work of the controlled process.

Consequently, the monitoring process increases the accessibility of the
unwanted thought by priming it each time it is encountered. Through this
sporadic priming, the monitoring process ironically makes the unwanted thought
more accessible, making suppression more difficult and rebound more likely.

As an extension of Wegner’s ( 1994) model, Macrae, Bodenhausen, Milne,
and Jetten (1994) posit that the monitoring process itself can increase the
accessibility of unwanted thoughts during thought suppression. In particular,
these researchers suggest that the intention to suppress a thought initiates the
operation of both the operating process and the monitoring process, as
suggested by Wegnefs (1994) model. But as the monitoring works toward its
goal of scanning consciousness for the unwanted thought, it ironically primes
the unwanted thought at a low but continuous level, regardless of the work of
the operating process.

Whereas initial studies investigated the effects of suppressing thoughts
about novel stimuli (e.g., white bears), more recent studies have investigated
the effects of suppressing stereotypical information about various targets. This
research has shown that stereotype suppression results in greater accessibility
and better recall for stereotypical information, but poorer processing of non-
stereotypical information (Macrae, Bodenhausen, Milne, & Wheeler, 1996). It
has also been shown that stereotype suppression can influence judgments of
and behavior toward those to whom the stereotypes may apply (Macrae et al.,

1994; cf., Monteith, Sherman, & Devine, 1998).

Although there may be times when one would like to suppress novel stimuli
and stereotypical thoughts about others, it seems that the thoughts one would
most often want to suppress are negative thoughts related to oneself. For
example, after getting a poor grade on an exam, a student may try to focus on
more pleasant thoughts of the upcoming weekend in an attempt to avoid
thoughts about the academic failure. In fact, Baumeister (1991) has suggested
that some individuals will take extreme measures to avoid focused attention on
themselves because such attention brings their negative qualities to mind. For
instance, he argues that alcoholism, masochism, and even suicide are acts
performed with the primary purpose of escaping self-awareness.

While such acts may distract oneself from self-focused attention, they also
tend to lead to very undesirable consequences. Therefore, it is important to
understand alternative processes by which people distract themselves from
negative self-relevant information and to understand the conditions under which
thought suppression may be more or less successful. It seems plausible to
think that what we focus our attention on while attempting thought suppression
may influence our ability to suppress unwanted thoughts and also the extent to
which these thoughts rebound.

For example, Wegner et al. (1987) found that focusing individuals on a
specific distracter (e.g., a red Volkswagen) during suppression resulted in an
attenuation of the rebound effect. Subsequent investigators (e.g., Wenzlaff,
Wegner, & Klein, 1991) have suggested that this focused distraction is not the

typical strategy individuals employ while attempting to suppress unwanted

thoughts. The more typical approach to suppression seems to be unfocused
distraction, in which individuals select things in the immediate environment or
accessible experiences and memories to distract themselves. Thus, individuals
tend to sample a variety of distracters, rejecting each and selecting a new one
each time the unwanted thought recurs. This suggests that the distracter used
could potentially facilitate or undermine suppression. More specifically,
selecting an effective distracter could result in more successful suppression of
unwanted thoughts for a longer period of time. Selecting a less effective
distracter, on the other hand, could result in more limited success for a shorter
period of time. For instance, trying to suppress a thought that is associated with
many other thoughts may undermine suppression, whereas trying to suppress a
thought that is not associated with many other thoughts should facilitate
suppression. And similarly, selecting a distracting thought that is associated
with many other thoughts may undermine suppression, whereas selecting a
distracting thought that is not associated with many other thoughts should
facilitate suppression. In this vein, the way in which information about the self is
cognitively organized may be related to the ability to suppress unwanted self-
relevant thoughts.

Interestingly, previous investigations involving the suppression of self-
relevant thoughts have been less successful in demonstrating the rebound
effect. For example, Kelly and Kahn (1994) had participants either suppress
one of their own personally intrusive thoughts or thoughts of a white bear. The

rebound effect was found for those suppressing thoughts of a white bear, but

not for those suppressing their own personally intrusive thought. These
researchers suggested that the failure to demonstrate the rebound effect with
personally intrusive thoughts may be due to the prior experience individuals
have in suppressing such thoughts. It may be that the individuals have
developed a network of distracter thoughts that are used whenever the
unwanted thought is encountered. Such a network of distracter thoughts could
involve the organization of the self-concept itself. With this in mind, the
following discussion will begin by focusing on self-complexity, which is one way
in which the organization of the self-concept can vary among individuals. In
particular, the potential role of self-complexity in the suppression of negative
self-relevant thoughts will be discussed and a number of research hypotheses
will be offered. Afterwards, a number of ancillary hypotheses associated with
thought suppression and self-complexity separately will be discussed. And
lastly, the role of another way in which the self-concept can vary among
individuals will be discussed. More specifically, the role of
compartmentalization in the suppression of negative self-relevant thoughts will
be explored, with a number of exploratory hypotheses paralleling the self-
complexity hypotheses being offered.

As mentioned above, one way in which the organization of the self-concept
can vary among individuals is in its complexity (Linville, 1985). Self-complexity
involves both the number of self-aspects and the amount of interrelatedness
among those self-aspects. Greater self-complexity is revealed by a greater

number of self-aspects that are more independent of one another. Lower

self-complexity, on the other hand, is revealed by fewer self-aspects that are
more interrelated with one another.

Research on self-complexity has suggested that it is related to differences in
affective responses to life events. In particular, the affective-extremity
hypothesis associated with self-complexity suggests that greater self-complexity
is related to more moderate affect in response to life events. For instance,
Linville (1985) found that following either positive or negative feedback about an
important aspect of one’s life, individuals greater in self-complexity reported
more moderate affect than individuals lower in self-complexity who reported
more extreme affect (positive and negative, respectively). Linville (1985)
argues that these differences in affective reactions to positive and negative
events by individuals who vary in self-complexity occur because of affective
spillover. Because greater self-complexity involves a larger number of
independent self-aspects, there is less affective spillover among different self-
aspects when emotional events impinge upon one’s life. Thus, a relatively
smaller proportion of self-aspects may be affected at any one time and,
therefore, a small proportion of the self is implicated during emotional episodes.
For example, consider a woman who sees her roles of mother, wife, doctor, and
friend as independent of one another. If she has a bad day at work, the
negative affect related to her work is less likely to “spill over" into her roles of
mother, wife, and friend. Individuals lower in self-complexity, on the other hand,
have fewer self-aspects that are more interrelated with one another. Thus, a

greater proportion of their self-aspects may be affected by positive and negative

life events, which, in turn, should magnify the intensity of the affect experienced.
In this case, if a woman who works with her husband in a family business has a
bad day at work, the negative affect related to her work aspect is more likely to
“spill over” into her roles of wife and mother. The same would be true for
positive events as well.

An important issue to consider is how the organization of the self-concept
may be related to the ability to moderate one’s affective responses to life
events. One way in which individuals can moderate the effects of negative
thoughts about the self is to focus attention away from the self. For example,
Dixon and Baumeister (1991) found that self-complexity is a moderator in the
relationship between failure and escape from self-awareness. In particular,
they found that following negative self-relevant feedback, individuals lower in
self-complexity attempted to reduce self-awareness faster than individuals
greater in self-complexity. This may be because a greater proportion of one’s
self-concept is affected for those lower in self-complexity than for those greater
in self—complexity. Consequently, those lower in self-complexity have a greater
need to focus attention away from the self following negative feedback than do
those greater in self-complexity.

Although there may be instances when it is possible to moderate the effects
of a negative event by focusing attention away from the self, there may be times
when such an escape is not possible or desirable. In these cases, individuals
may try to focus on other aspects of themselves in order to avoid negative self-

relevant thoughts associated with a particular self-aspect. Individuals greater in

self-complexity, by definition, have more potential internal distracters (i.e., self-
aspects) that are relatively independent of one another than do individuals
lower in self-complexity. Thus, suppressing negative information about the self
while focusing on internal distracters (i.e., other self-aspects) should be a
relatively easier task for individuals greater in self-complexity than for
individuals lower in self-complexity. External distracters, on the other hand,
should be equally available to individuals both higher and lower in self-
complexity, resulting in no difference in rebound when the distracter used is
external to the self. Therefore, it is predicted that those lower in self-complexity
will exhibit greater rebound following suppression of negative self-relevant
information only when focusing on internal distracters.

Interestingly, it may be that the ability to suppress negative self-relevant
information about a particular self-aspect underlies much of the affective-
extremity hypothesis discussed previously. Because individuals greater in self-
complexity have more internal distracters available that are relatively
independent of one another, they may be better able to moderate their affective
responses to negative information by focusing on other unrelated self-aspects
that do not implicate the feedback-relevant domain. And because individuals
lower in self-complexity have fewer internal distracters available, they are less
able to moderate their affective responses, resulting in more extreme affective
experiences. In cases where feedback about the self is negative, the ironic
monitoring process should increase the relative accessibility of negative

information for the individual. Individuals lower in self-complexity focusing on

internal distracters will be more likely to prime more self-aspects related to the
negative information due to the greater number of associative links leading back
to the to-be-avoided self-aspect. Therefore, it is predicted that following
suppression of negative self-relevant information while focusing on internal
distracters, those lower in self-complexity will reveal self-judgments greater in
negativity.

Further, this accessibility of negative self-relevant information may lead to an
avoidance of tasks related to the feedback domain. As Dixon and Baumeister
(1991 ) found, following negative feedback, those lower self-complexity tended
to put forth less effort into a subsequent task. Because those lower in self-
complexity may be less able to escape negative self-relevant information when
using internal distracters, they may have a greater need to avoid similar,
potentially negative, situations. Hence, it is predicted that following suppression
of negative self—relevant information while focusing on internal distracters, those
lower in self-complexity will be more likely to avoid performing a task similar to
that for which negative self-information was provided.

And finally, the accessibility of this negative self-relevant information may
also make it more difficult to alter one’s thoughts after finding out that the
negative self-information was not accurate. Previous research on such belief
perseverance effects (e.g., Ross, Lepper, & Hubbard, 1975) has demonstrated
that participants’ beliefs about information given to them in experimental studies
often tends to persist even after they have been told that the information was

bogus. The current research will explore the possibility that some individuals

are more susceptible to such effects than others. Thus, it is predicted that
following suppression of negative self-relevant information while focusing on
internal distracters, those lower in self-complexity will manifest a greater
perseverance of negative beliefs even after being told that the negative self-
information was not accurate.

In addition to the hypotheses above, a number of ancillary hypotheses
separately involving self-complexity and thought suppression will be tested. As
mentioned previously, Macrae et al. (1994) suggest that the process of thought
suppression itself increases the accessibility of the to-be-suppressed thoughts.
The methods to be employed in the current study allow for a test of this
hypothesis. Based on research by Macrae et al. (1994), it is first expected that
participants who previously suppressed thoughts of their failure on the task
(which, purportedly, is diagnostic of their academic aptitude) would reveal
greater accessibility of student-related thoughts than would participants who did
not previously suppress such thoughts. Based on the logic of the above
hypotheses for self-complexity, it is also expected that this relationship will be
moderated by self-complexity and the type of distracter used during
suppression. More specifically, it is expected that following suppression of
negative self-relevant information while focusing on internal distracters, those
lower in self-complexity will demonstrate greater accessibility of student-related
thoughts.

Secondly, because self-esteem is an important construct in assessing self-

relevant affect, many researchers have attempted to discover a relationship

10

between self-complexity and self—esteem. For instance, Campbell, Scratchley,
and Chew (1991) found a positive relationship between them, suggesting that
greater self-complexity is related to higher self-esteem. Woolfolk, Novalany,
Gara, Allen, and Polino (1995), on the other hand, found a negative relationship
between them, suggesting that greater self-complexity is related to lower self-
esteem. And finally, Niedenthal, Setterlund, and Wherry (1992) found no
relationship between them at all. The relationship between self-complexity and
self-esteem will also be examined in the current study. Based on the
contradictory findings in the literature, no directional prediction will be made.

The third ancillary hypothesis involves the affective-extremity hypothesis
associated with self-complexity. As mentioned previously, this hypothesis
suggests that greater self-complexity is related to more moderate affect in
response to life events, whereas lower self-complexity is related to more
extreme affect in response to similar life events. In an examination of this
hypothesis, Linville (1985) found that those lower in self-complexity reported
greater change in mood and self-evaluation following self-relevant feedback
that did those greater in self-complexity. The methods employed in the current
study allow for a replication of thisfinding.

And finally, the conceptual process of thought suppression and subsequent
rebound suggests that the more an individual tries to suppress a thought, the
more that individual subsequently ends up thinking about it, regardless of the
organization of the self-concept. As described by Wegner et al. (1987), “...the

person who is most successful in carrying out the suppression may eventually

11

be the most susceptible to the resulting obsession” (pp. 11-12). This suggests
that there should be an inverse relationship between initial suppression and
subsequent expression within individuals, such that the less an individual
mentions the unwanted thought during suppression the more that individual
tends to mention it during subsequent expression. Interestingly, previous
investigations (i.e., Wegner et al., 1987) have not found such correlations for
suppression groups. The current research will examine the correlations
between the suppression and the expression periods for the suppression
groups.

In addition to self-complexity, information about the self can also vary in
terms of its content. In particular, the contents of information about the self can
be either positively or negatively valenced. Showers (1992a) has examined the
extent to which positive and negative information about the self is
compartmentalized among one’s self-aspects. Specifically, individuals with
compartmentalized self-information tend to have either predominantly positive
or predominantly negative information about the self contained within different
self-aspects. Individuals with self-information that is more evaluatively
integrated, on the other hand, tend to have both positive and negative
information mixed within their different self-aspects.

Investigating the correlates of compartmentalization, Showers (1992a,
1992b) found that compartmentalized organization is related to more extreme
positive or negative affective responses. The valence of the affect typically

experienced by the individual is mediated by differential importance, which

12

refers to the relative importance individuals place on particular self-aspects
(Pelham & Swann, 1989). The product of compartmentalization and differential
importance is used to differentiate between those who are positively
compartmentalized and those who are negatively compartmentalized. In
particular, positive compartmentalization refers to an organization in which the
individual perceives self-aspects that are predominantly positive as more
important, whereas negative compartmentalization refers to an organization in
which the individual perceives self-aspects that are predominantly negative as
more important. Consequently, those with positive compartmentalization tend
to have more positive self-evaluations and more positive affect in general,
whereas those with negative compartmentalization tend to have more negative
self-evaluations and more negative affect in general (Showers, 1992a).
Showers (1992b) suggests that the effects of compartmentalization can best
be understood by conceptualizing the self-concept as an associative network
model of memory. Such a model suggests that when one item of information is
activated, other items that are strongly associated with it in memory are also
more likely to be activated. In terms of compartmentalization, this suggests that
when a particular self—aspect is activated, all of the information within that self-
aspect is likely to be activated. For negative-compartmentalized individuals,
this suggests that when a negative self-aspect is activated the attributes (which
should be predominantly negative) within that self-aspect will also be activated.
And because these individuals have relatively more negative self-aspects that

are important, negative self-aspects and the contents therein will be activated

13

more often, leading to more negative self-evaluation and more negative affect
being experienced in general. On the other hand, positive—compartmentalized
individuals tend to have few, if any, negative self-aspects that are important.
Thus, their positive self-aspects and the contents therein will be activated more
often, leading to more positive self-evaluation and more positive affect being
experienced in general. Finally, individuals who are more evaluatively
integrated tend to have self-aspects that are neither predominantly positive nor
predominantly negative. Therefore, they tend to have relatively moderate self-
evaluation and experienced affect.

Each of the following hypotheses involving compartmentalization parallels
those previously described involving self-complexity. Because the development
of theory related to compartmentalization is more recent and less explored in
the literature, the current work will examine these hypotheses in an exploratory
fashion.

The first exploratory hypothesis involves the potential role of
compartmentalization in thought suppression. In particular, individuals with
compartmentalized self-aspects have a relatively high degree of positive or
negative attributes contained within different self-aspects, so it may be more
difficult for them, as opposed to more evaluatively integrated individuals, to
suppress thoughts related to a particular self-aspect. More specifically, those
who have many negatively compartmentalized self-aspects, as opposed to
those who have more evaluatively integrated self-aspects, may find it more

difficult to suppress negative self-relevant thoughts related to a particular

14

self-aspect. This is because they have fewer positive things, either within the
to-be-avoided self-aspect or with other self-aspects in general, on which to
focus on in order to distract themselves from the negative unwanted thought.
Those who have many evaluatively integrated self-aspects, on the other hand,
may find it less difficult to suppress negative self-relevant thoughts because
they have other less negative things (i.e., the positive attributes within their self-
aspects) on which to focus. And those who have many positively
compartmentalized self-aspects may find it relatively easy to suppress negative
self-relevant thoughts because they have many positive self-aspects and
attributes therein on which to focus.

As opposed to either focused or unfocused distraction, using one’s own self-
aspects as distracters during suppression can be thought of as semi-focused
distraction. Individuals are focused on themselves but will probably select
different self-aspects, rejecting each and selecting a new one each time the
unwanted thought recurs. Although it may be reasonable to suggest that
individuals with negative compartmentalization or evaluative integration may
select a particular positive self-aspect on which to focus in times of negativity,
this assertion is not consistent with previous findings. For example, as
mentioned previously, Showers (1992a, 1992b) found that negative-
compartmentalized individuals tend to have more negative self-evaluation and
more negative affect in general. This suggests that these individuals are not

able to focus on a particular positive self-aspect as a means of avoiding

15

negative self-thoughts. If they could, they would not exhibit the negative affect
observed in previous research.

Thus, those with many important negative self-aspects (i.e., negative
compartmentalization), as compared to those with many important positive self-
aspects (i.e., positive compartmentalization), may be subject to greater rebound
following suppression of negative self-relevant information while focusing on
internal distracters. This will result because their available negative self-
aspects will be less likely to distract them from the unwanted thought for a long
period of time. Each important self-aspect (and the contents therein)
encountered is likely to be negative and will, therefore, be a less effective
distracter. Consequently, negatively compartmentalized individuals will tend to
sample from their various self-aspects more often than positively
compartmentalized individuals. And because the monitoring process primes
unwanted thoughts during suppression, the to-be-avoided self-aspect will be
more accessible than other self-aspects. Hence, it is more likely that the
unwanted self-aspect will be encountered during each distracter search. In
sum, the greater frequency of searches and the greater accessibility of the to-
be-avoided self-aspect together should result in a greater likelihood of selecting
the self-aspect associated with the unwanted thought more often, resulting in
greater rebound for negatively compartmentalized individuals.

Therefore, it is expected that negative compartmentalization will be related
to greater rebound following suppression of negative self-relevant thoughts

while focusing on internal distracters. Alternatively, positive

16

compartmentalization will be related to less rebound following suppression of
negative self-relevant thoughts while focusing on internal distracters. And,
evaluative integration will be related to relatively moderate rebound following
suppression of negative self-relevant thoughts while focusing on internal
distracters.

Furthermore, research has shown that the reactions of compartmentalized
individuals tend to be more extreme than evaluatively integrated individuals.
The ability to suppress negative self-relevant thoughts may underlie these
different affective reactions.

As previously discussed for those lower in self-complexity, negative-
compartmentalized individuals should be more likely to prime more negative
than positive information about the self during suppression while focusing on
internal distracters. This accessibility may lead to more negatively biased
judgments of the self. So, it is expected that following suppression of negative
self-relevant information while focusing on internal distracters, negative-
compartmentallzed individuals will reveal self-judgments greater in negativity.

In addition, this greater accessibility of negative self-relevant information
may lead to an avoidance of tasks related to the feedback. Therefore, it is
expected that following suppression of negative self-relevant information while
focusing on internal distracters, negative-compartmentalized individuals will be
more likely to avoid performing a task similar to that for which negative self-

inforrnation was provided.

17

And finally, the accessibility of this negative self-relevant information may
also make it more difficult to change one’s beliefs about the information after
finding out that the negative information was not accurate. Therefore, it is
expected that following suppression of negative self-relevant information while
focusing on internal distracters, negative-compartmentalized individuals will
manifest a greater perseverance of negative beliefs even after being told that
the negative self-information was not accurate.

METHOD
Participants and Overview

Ninety-eight introductory psychology students at Michigan State University
participated in the study for extra credit. The study consisted of two sessions.
During the first session, participants completed the self-complexity,
compartmentalization, and positive importance measures via a computer
program on the university’s network server.

During the second session (which took place in the lab), participants were
first informed of the tasks to be completed during the experiment. It was
explained that by signing the consent form, they were agreeing to allow their
voices to be audio taped and used as data for the study. Upon providing
consent, they began the experimental session by completing the initial mood
and self-esteem measures. They then completed and received negative
feedback on an analytical task, which presumably assessed academic success

in college. Following this, they completed the mood and self-esteem measures

18

for a second time to examine the extent to which the negative feedback had an
impact on their mood and self-esteem.

Participants then read instructions on how to report their stream of
consciousness. They performed this task in three separate 5 min periods.
During the first period, all participants were asked to verbally express their
ongoing thoughts, without filtering them in any way. The instructions explicitly
stated that they were to express any and all thoughts even if the thoughts
involved the feedback from the analytical task and how the feedback may be
related to their academic life. For the second period, two-thirds of the
participants were asked to suppress the negative feedback from the analytical
task and their academic life in general. One-half of these suppression
participants did so while focusing on other self-aspects of themselves, whereas
the other one-half did so while focusing on a white bear. The remaining one-
third of the participants expressed their ongoing thoughts, including thoughts
about the feedback and how the feedback may be related to their academic life.
For the third period, all participants were asked to express their thoughts again,
including thoughts about the feedback from the analytical task and how this
feedback may be related to their academic life.

Immediately following the third period, all participants completed a word
completion task to measure the accessibility of student-related thoughts. The
participants then indicated how well they would perform on a similar analytical
task if given another opportunity. They then rated both the extent to which they

would like to perform another analytical task and the extent to which they would

19

like to perform a task that assesses creative abilities. Following this, they were
debriefed about the true nature of the study. During this debriefing, they were
told that the negative feedback was completely false. Afterwards, they were
asked to estimate the number of items that they think they actually answered
correctly on the analytical task in order to assess belief perseverance.
Self-Complexity and Compartmentalization Measures

In the first session, participants performed a trait sort task similar to that
used by Showers (1992a). They completed this task via a computer program in
which they were presented with 40 different traits, half of which were positive
and half of which were negative. They sorted the traits into groups that
described different and important aspects of themselves. They also provided a
label to describe each of the groups they formed. For example, a participant
may have sorted “intelligent,” “diligent,” and “focused” into one group and
labeled it as “student.” This task is comparable to that used by Linville (1985,
1987) for assessing self—complexity.

The statistical measure _l-_l, developed by Scott (1969) and used by Linville
(1985, 1987), was calculated to obtain a self-complexity score for each
participant. Scott’s 5 takes into account the number of self-aspects generated
and the interrelatedness among the traits within those self-aspects. The

following formula is used to calculate Scott’s fl:
fl= logzn-lzinilogznifln.
where n is the total number of traits available to the participant (40 in this case)

and ni is the number of traits that occur within each particular group

20

combination (i) across the self-aspects constructed by the participant. Scott’s L-l_
can be understood as an index of the minimal number of particular
combinations of traits needed to reproduce a participant’s whole sort. Appendix
A demonstrates the calculation of u for a participant in the current study.
Following Showers (1992a), a phi coefficient was calculated to obtain a
compartmentalization score for each participant. The phi coefficient is based on
the chi-square statistic and takes into account the distribution of positive and
negative traits among each participant’s self-aspects. It is based on a
contingency table in which the columns represent the self-aspects generated
and two rows represent the number of positive and the number of negative
traits within each self-aspect. Mathematically, phi was computed as the square
root of the chi-square statistic divided by the total number of traits within each
participant’s sort. Phi can range from 0 (perfectly integrated) to 1 (perfectly
compartmentalized). More generally, phi compares the ratio of positive to
negative traits within each self-aspect to the overall ratio of positive to negative
traits for the participant’s whole sort. If the ratio across all self-aspects matches
the overall ratio for the whole sort, the number of positive and negative traits
(i.e., the rows of the contingency table) is independent of the self-aspects (i.e.,
the columns of the contingency table). Such a match would result in a phi of
zero, which would be indicative of an evaluatively integrated sort. If, on the
other hand, the ratio across all self-aspects does not match the overall ratio for

the whole sort, the number of positive or negative traits is not independent of

21

the self-aspects. This would result in a phi closer to one, which would be
indicative of a more compartmentalized sort.
Aspect Importance and Valence Measures

Participants also rated the importance of each self-aspect on a scale that
ranged from 1 (somewhat important) to 7 (extremely important). In addition,
they rated how the self-aspect generally makes them feel on a scale that
ranged from -7 (very negative) to +7 (very positive). To distinguish between
positive- and negative-compartmentalized organization, previous investigators
have used the differential importance of positive and negative self-aspects.
This index is the Pearson correlation coefficient of participants’ valence and
importance ratings across all of their self-aspects. Higher correlations indicate
that positive self-aspects are rated as more important than negative self-
aspects. Because a differential importance score cannot be calculated when a
participant has just one self-aspect, the present research used a slightly
different approach to distinguish between positive- and negative-
compartmentalized organization. More specifically, a positive importance score
was calculated for each participant by first multiplying the ratings of valence by
the ratings of importance for each of their self-aspects. Then the mean product
across all of their self-aspects was computed. Larger positive values indicate
that positive self-aspects were rated as more important than negative self-

aspects.

22

Mood Measure

During the second session, participants were run individually in the lab.
They first completed the Positive and Negative Affect Schedule (Watson, Clark,
& Tellegen, 1988). This measure includes 1 0 positive mood adjectives (e.g.,
“proud”) and 10 negative mood adjectives (e.g., “upset”) based on 5-point
scales ranging from 1 (very slightly) to 5 (extremely). The items were presented
on a computer one at a time. Participants were told to answer the items in
terms of how they felt right at that very moment.
Self-Esteem Measure

Participants also completed Rosenberg’s (1965) self-esteem scale. This
scale includes 10 items (e.g., “I take a positive attitude toward myself” and “At
times I think I am no good at all”) that are based on a 4-point scale ranging from
1 (strongly disagree) to 4 (strongly agree). Items were summed such that larger
scores indicated greater self-esteem. Theoretically, the scale can range from
10 (low self-esteem) to 40 (high self-esteem). Again, the items were presented
on a computer one at a time. After the last item was presented, a bogus error
message appeared on the computer screen. When the participant called for the
experimenter, the experimenter acted surprised and told the participant that the
computers were old and somewhat unreliable.
Anamical Task

While the experimenter worked on solving the computer problem, the
participants were asked to move to another computer to solve 24 moderately

difficult analogy items taken from past Graduate Record Exams. To emphasize

23

the importance of its feedback, they were told that this task is often used to
predict success in college. After completing the task, the participant’s score
was presented on the computer screen, with an indication that the score was in
the bottom 10% of all students who have previously completed the task.

The participants were then told that their responses to the previous mood
and self-esteem items were lost due to the prior computer error. Because of
this purported error, they were asked to complete the items a second time
based on how they felt right at that very moment. Thus, this second
administration served as the post-feedback (Time 2) measures of mood and
self-esteem.

Suppression Ability and Rebound Measures

Participants then read instructions adapted from Pope (1978) and used by
Wegner et al. (1987) on how to report their stream of consciousness. These
instructions ask participants to continuously verbalize their ongoing thoughts
without filtering them in any way. Similar to the procedure used by Wagner and
Gold (1995), participants verbally reported their thoughts during three 5-min
periods by themselves in a private room while being audio recorded. During the
initial expression period, all participants expressed their ongoing thoughts,
including those about the feedback provided on the analytical task and how this
feedback may be related to their academic life.

During the suppression period, two-thirds of the participants were asked to
suppress the negative feedback information provided by the score on the

analytical task and their academic life in general. Thus, the student self-aspect

24

served as the to-be-suppressed self-aspect. This self-aspect was selected
because participants were college students, and, therefore, their student self-
aspect should be important to them. Previous research (e.g., Linville, 1985;
Niedenthal et al., 1992) has found that feedback related to intelligence and
scholastic performance has affective consequences for college student
participants. One-half of the suppression participants suppressed their student
self-aspect while focusing on another aspect of themselves (internal
suppressers), whereas the other one-half suppressed their student self-aspect
while focusing on a white bear (external suppressers). The internal suppresser
participants read the following instructions, which were based on those used by
Wegner et al. (1987):
For the second five-minute period, please verbalize your thoughts as you
did before, with one exception. This time, try not to think about the
feedback you were given on the analogy task or anything else related to
your academic life, but mention it if you do. Instead, think about one or
more of the other groups that you described in the first experiment listed
in this envelope (they were given a few moments to look at the labels of
the groups they had described in the first session of the experiment).
Again, remember, gm} think about the feedback or your academic life,
but mention it if you do.
The external suppresser participants read similar instructions, but were told to
think about a white bear instead. The remaining participants (expressers)

performed a task similar to the initial expression period in which they verbally

25

expressed all of their ongoing thoughts, including thoughts about the feedback
given on the analytical task and how this feedback may be related to their
academic life, without filtering them in any way. Assignment to this between-
subjects manipulation (internal suppresser, external suppresser, or expresser)
was randomly determined. For the final period, all participants reported their
thoughts, including thoughts about the feedback given on the analytical task
and how this feedback may be related to their academic life. Appendix 8
contains the verbatim instructions given for each period.

The rebound measures were based on the number of times student-related
thoughts were mentioned by the suppressers (as compared to the expressers)
during the final expression period. An additional index of rebound included the
duration of mentions of the previous to-be-suppressed thoughts.

Accessibility Measure

All participants were then given a word completion task to measure the
accessibility of student-related thoughts. This task consisted of 27 items that
could be filled in with letters that would create either student-related or non-
student-related words. They were asked to complete the task as quickly as
possible to ensure that their responses reflected accessible concepts. Two
examples of items are “smart” and “pull” (see Appendix C for the complete list
of items). Each item that was completed as a student-related word was given a
score of one. These scores were summed, with larger scores indicating greater

accessibility of student-related thoughts.

26

Self-Judgment Measure

Participants then indicated how well they thought they would perform on a
similar analytical task if given another opportunity on a scale that ranged from 1
(very poorly) to 7 (very well).

Subseguent Task Preference

They then rated both the extent to which they would like to complete another
analytical task and the extent to which they would like to complete a different
task that assesses creative abilities unrelated to academic abilities on 7-point
scales that ranged from 1 (not at all) to 7 (very much).

Belief Perseverance

Participants were then debriefed about the study. During this debriefing,
they were told that the feedback from the analytical task was completely false
and in no. way related to their actual performance or to their potential success
as a college student. Following this disclosure, participants indicated the
number of items (from O to 24) that they thought they actually answered
correctly on the analytical task.

RESULTS

Table 1 (Appendix D) presents descriptive statistics for the primary variables
used in the inferential analyses.

The audio tapes were analyzed for evidence of rebound effects by two
independent judges who were unaware of the experimental hypotheses. Both
the number of mentions and the duration of mentions of the previous to-be-

suppressed thoughts were assessed by the judges. Numbers of mentions were

27

counted if they occurred in one of two ways. First, if a student-related thought
(i.e., an unwanted thought) occurred between two non-student-related thoughts,
it was considered a mention. Secondly, if a 5 s or longer pause occurred
between two student-related thoughts, two mentions were counted. The
duration of mentions was simply measured as the number of seconds the
participant expressed each student-related thought. These values were
summed separately for each of the three periods. lnterjudge reliability was
assessed by examining the correlation of the measurements between the two
judges. Strong positive correlations were found for both the number of
mentions, [ = .63, p < .001, and the duration of mentions, [ = .83, p < .001,
indicating good interjudge reliability. The measurements of the two judges were
then averaged.
Tests of PrimaLy Hypotheses

Each of the primary hypotheses was tested by conducting a multiple
regression analysis with LI, two contrast vectors (one comparing the two
suppresser groups to the expresser group, the other comparing the internal
suppressers to the external suppressers) and their interactions with H
regressed onto the corresponding dependent variable. A significant interaction
between H and the second contrast vector will demonstrate support for each of
the predictions.

The first hypothesis predicted that those lower in self-complexity would
exhibit greater rebound following suppression of negative self-relevant

information only when focusing on internal distracters. The analysis involving

28

the number of mentions as the criterion revealed a significant main effect for the
second contrast vector (which compared the internal suppressers to the
external suppressers), pgt_a_ = .72, t(96) = 2.32, p < .05. This main effect was
qualified by the predicted interaction between H and the second contrast vector,
ppta = -.76, t(96) = -2.44, p < .05. Non-standardized b-weights using a range of
-/+ 2 SQ for self-complexity (i.e., Scott’s H) were used to graph this effect.
Figure 1 (Appendix E) illustrates how the relation between self-complexity and
the number of mentions of the to-be-suppressed thoughts varies as a function
of whether distracters were internal or external, as predicted. Analyses of the
slopes of each of the regression lines in Figure 1 were also conducted. The
slope for internal suppressers was significant, Ea = -.40, 1(31) = -2.35, p < .05,
revealing that participants lower in self-complexity focusing on internal
distracters showed stronger rebound than did those greater in self-complexity
focusing on internal distracters, as predicted. The slope of the regression line
for external suppressers was not significant, peg = .29, 1(29) = 1.61, as.
Contrary to predictions, the regression analysis involving the duration of
mentions as the criterion revealed no significant effects. No other effects were
significant.

The second hypothesis predicted that following suppression of negative
self-relevant information while focusing on internal distracters, those lower in
self-complexity would reveal self-judgments greater in negativity. Contrary to

the prediction, the interaction between H and the second contrast vector was

not significant, beta = -.09, 1(96) = -.27, n_s. No other effects were obtained.

29

The third hypothesis predicted that following suppression of negative self-
relevant information while focusing on internal distracters, those lower in self-
complexity would be more likely to avoid performing a task similar to that for
which negative feedback was provided. In order to determine if a composite
difference score (i.e., subtracting the creative task preference rating from the
analytical task preference rating) could be used as the dependent variable in
the following analyses, a correlational analysis was conducted between the
analytical task preference responses and the creative task preference
responses. No relationship between the two was found, [(96) = -.04, pp.
Therefore, multiple regression analyses were conducted on each of the two
task preference responses separately. Contrary to the predictions, the
analyses involving both the analytical task preferences and the creative task
preferences revealed no interactions between H and the second contrast
vector, pfis = .02 and -.20, 13(96) = .05 and -.66, g, respectively.
Interestingly, a marginal interaction between :1 and the first contrast vector
(suppressers versus expressers) was found for the creative task preferences,
pet_a = .55, 1(96) = 1.98, p < .06. The non-standardized b-weights of this effect
are plotted in Figure 2 (Appendix F). This pattern shows how the relation
between self-complexity and creative task preference varies as a function of
whether participants had suppressed or expressed student-related thoughts.
Analyses of the slopes of each of the regression lines in Figure 2 were also
conducted. The slope for suppressers was significant, peg = .30, 1(62) = 2.49,

p < .05, revealing that suppressers greater in self-complexity reported a

30

stronger preference for the creative task than did suppressers lower in self-
complexity. The slope for expressers was not significant, .b_et§ = -.16,
t(32) = -.89, p_s_.

The final hypothesis predicted that following suppression of negative self-
relevant information while focusing on internal distracters, those lower in self-
complexity would manifest a greater perseverance of negative beliefs even after
being told that the negative self-information was not accurate. Contrary to the
prediction, no interaction between H and the second contrast vector was found,
b_e_t§ = -.27, 1(96) = -.85, pp. No other effects were significant.

Ancillagy Tests

As discussed previously, it is argued that the process of thought suppression
increases the accessibility of the to-be-suppressed thought (Macrae et al.,
1994). Based on this research, it was first expected that participants who
previously suppressed thoughts of failure on the task and of their student self-
aspect would reveal greater accessibility of student-related thoughts than would
participants who did not previously suppress such thoughts. In addition, based
on the hypotheses in the current work, it was also expected that this
relationship would be moderated by self-complexity and the type of distracter
used during suppression. More specifically, it was expected that following
suppression of negative self-relevant information while focusing on internal
distracters, those lower in self-complexity would demonstrate greater
accessibility of student-related thoughts. A multiple regression analysis was

conducted with H, two contrast vectors (suppressers versus expressers and

31

internal versus external suppressers) and their interactions with H regressed on
the accessibility scores. As predicted, a marginal main effect for the first vector
(suppressers versus expressers) was found, 993 = .51, t(96) = 1.77, p < .08,
revealing that suppressers had greater accessibility for student-related
concepts than did expressers. Interestingly, this main effect was qualified by a
very marginal interaction between 11 and the first contrast vector, pe_ta = -.44,
t(96) = -1.52, p < .14. Non-standardized b-weights are displayed in Figure 3
(Appendix G) to illustrate how the relation between self-complexity and the
number of student-related words completed tends to vary as a function of
whether participants had suppressed or expressed student-related thoughts.
Analyses of the slopes for each of the regression lines in Figure 3 were also
conducted. The slope for suppressers was significant, ,b_et§ = -.36,

1(62) = -3.01, p < .01, revealing that suppressers lower in self-complexity
revealed greater accessibility of student-related concepts than did suppressers
greater in self-complexity. The slope for expressers was not significant,

bite = -.05, 1(32) = -.26, as. Contrary to expectations, distracter type did not
influence accessibility of student-related concepts.

Because previous investigations (Campbell et al., 1991; Niedenthal et al.,
1992; Woolfolk et al., 1995) examining the relationship between self-complexity
and self-esteem have revealed contradictory findings, this relationship was also
examined in the current research. A correlational analysis revealed a negative

correlation between H and the initial self-esteem measure (Time 1), [(96) = -.23,

32

p < .05. Similar to Woolfolk et al. (1995), this finding suggests that greater self-
complexity is related to lower self-esteem.

As previously mentioned, in an investigation of the affective-extremity
hypothesis Linville (1985) found that those lower in self-complexity reported
greater change in mood and self-evaluation following both positive and negative
feedback than did those greater in self-complexity. In order to test for
replication, a correlational analysis was conducted with H and self-esteem
change scores (i.e., subtracting self-esteem at time 2 from self-esteem at time
1). A negative correlation between H and self-esteem change scores was
found, [(96)= -.24, p < .05, revealing that those lower in self-complexity reported
a greater drop in self-esteem following the negative feedback than did those
greater in self-complexity. This finding is supportive of the affective-extremity
hypothesis. Correlational analyses were also conducted with _l-1 and both the
positive and negative mood change scores (i.e., subtracting mood at time 2
from mood at time 1) separately. A marginal negative correlation between [I
and positive mood change scores was found, [(96) = -.18, p < .07. This finding
suggests that those lower in self-complexity tended to report a greater drop in
their positive mood following the negative feedback than did those greater in
self-complexity, which is also consistent with the affective-extremity hypothesis.
No correlation between [1 and negative mood change scores was found,

[(96)= -.02, n_s.
Finally, the conceptual process of thought suppression and subsequent

rebound suggests that there should be an inverse relationship between initial

33

suppression and subsequent expression within individuals, such that the less
an individual mentions the unwanted thought during suppression the more that
individual should mention it during subsequent expression. Correlational
analyses were conducted on both the number of mentions and the duration of
mentions between the initial suppression and subsequent expression periods
for each of the suppression groups separately. Contrary to the conceptual
definition, a marginal positive correlation between the duration of mentions in
the suppression period and the duration of mentions in the expression period
for those suppressing student-related thoughts while focusing on an external
distracter (i.e., the white bear) was found, [(29)= .32, p < .08. This finding
suggests that the longer one tends to talk about the to-be-suppressed thoughts
during suppression, the longer that individual talks about the previous to-be-
suppressed thoughts during subsequent expression. Interestingly, this finding
is not in accordance with the conceptual definition of thought suppression and
subsequent rebound. No correlation between the number of mentions in the
suppression and expression periods for those focusing on a white bear was
found, [(29) = .14, pg. In addition, no correlations between the number of
mentions or the duration of mentions between the suppression and expression
periods were found for those suppressing student-related thoughts while
focusing on internal distracters, [5(31)= .16 and .03, g, respectively.
Exploratog Analyses

Because the exploratory hypotheses involving compartmentalization focused

on how compartmentalization, positive importance, and the type of distracter

used during suppression (internal versus external) might interact to produce
differences in the various dependent variables, the following results are based
on multiple regressions involving the 3-way interaction among m, with/p
importance, and a contrast vector (comparing internal suppressers to external
suppressers). In addition, the following lower-order factors were included in the
regression analyses: pm, msitive importance, the contrast vector (internal
versus external), and the 2-way interaction between ptli and W
importance. Thus, the analyses excluded participants in the expression group.
And because theory related to compartmentalization is based on the
significance of the interaction between p_h_i and positive importance, which treats
each of their linear contributions as relatively uninteresting individual
differences, the other 2-way interactions were not included in the analyses.
Significant 3-way interactions among M. positive importance, and the contrast
vector comparing internal suppressers to external suppressers for each of the
dependent variables will demonstrate support for the predictions.

The first exploratory hypothesis was that negative compartmentalization
would be related to greater rebound following suppression of negative self-
relevant information while focusing on internal distracters. Contrary to
expectations, the 3-way interaction was not significant for either the number of
mentions, b__e_t_a = -.34, [(87) = -1.08, pg, or for the duration of mentions,

% = -.20, [(87) = -.64, pp. No other effects were significant.
The second exploratory hypothesis predicted that following suppression

while focusing on internal distracters, negative-compartmentalized individuals

35

would reveal self-judgments greater in negativity. Contrary to the expectation,
the 3-way interaction was not significant, pe_ta_ = -.18, [(87) = -.58, g. However,
a marginal main effect for m was found, pe_t_a_ = .75, [(87) = 1.72, p < .09,
suggesting that compartmentalized participants tended to report greater
expectations regarding how well they would perform on a task similar to that for
which negative feedback was provided. This main effect was qualified by a
marginally significant 2-way interaction between m1 and positive importance,
M = -.78, [(87) = -1.79, p < .08. Figure 4 (Appendix H) shows how the
relation between self-concept organization and self-judgment scores varies as a
function of whether participants had a fewer number of self-aspects that were
positive and important or a greater number of self-aspects that were positive
and important. Each of the slopes was non-significant.

The third exploratory hypothesis predicted that following suppression while
focusing on internal distracters, negative-compartmentalized individuals would
be more likely to avoid performing a task similar to that for which negative
feedback was provided. Contrary to expectations, the 3-way interaction was not
significant for either the analytical task preference responses, pe_t§ = -.08,

[(87) = -.27, pp, or for the creative task preference responses, peg = -.41,
[(87) = -1.34, g. However, the main effect for p_h_i was significant, _b_et_a_ = .89,
[(87) = -2.01, p < .05, suggesting that compartmentalized participants reported
a stronger preference for the creative task following suppression than did

evaluatively-integrated participants. This main effect was qualified by a

significant interaction between phi and positive importance, beta = -.85,

36

[(87) = -2.01, p < .05. Figure 5 (Appendix I) shows how the relation between
self-concept organization and creative task preference varies as a function of
whether participants had a fewer number of self-aspects that were positive and
important or a greater number of self-aspects that were positive and important.
Each of the slopes was non-significant.

The final exploratory hypothesis predicted that following suppression while
focusing on internal distracters, negative-compartmentalized individuals would
manifest a greater perseverance of negative beliefs even after being told that
the negative self-information was not accurate. Contrary to the expectation, the
3-way interaction was not significant, b_etp = -.16, [(87) = -.50, pp. No other
effects were obtained.

DISCUSSION

The main purpose of the current study was to investigate the role that self-
concept organization plays when one is attempting to distract oneself from
unwanted self-relevant thoughts. Although previous research has shown that
self-concept organization may be related to self-relevant affect, its implications
for thought suppression have not been explored. Thus, the current research
attempted to shed some light on this relationship. Furthermore, this
investigation attempted to understand why some prior investigations examining
the suppression of self-relevant thoughts have not found the rebound effect that
has been demonstrated in research examining the suppression of novel stimuli
and of stereotypical thoughts. Kelly and Kahn (1994) suggested that the reason

for not obtaining rebound effects for self-relevant thoughts might be because

37

individuals have more experience suppressing self-relevant thoughts than they
do suppressing non-self-relevant thoughts. Therefore, they may have
developed a network of successful distracters to be used during suppression.
The current research examined the role of the self-concept as one pctential
network of distracters. The primary hypotheses focused on self-complexity, one
way in which self-concept organization varies among individuals.

The first hypothesis predicted that following suppression of negative self-
relevant thoughts while focusing on internal distracters, those lower in self-
complexity would reveal stronger rebound effects. The analysis involving the
number of mentions supported this prediction. Participants lower in self-
complexity focusing on internal distracters revealed a greater number of
mentions of the to-be-suppressed thoughts following suppression. This
outcome could result for at least two reasons. First, because those lower in
self-complexity have fewer self-aspects overall, they have fewer internal
distracters available during suppression than do those greater in self-
complexity. Second, because their self-aspects are more interconnected in
memory, there are more associative links back to the to-be-avoided self-aspect.
Thus, those lower in self-complexity should have more difficulty suppressing
negative self-relevant information while focusing on themselves. This difficulty
in avoiding unwanted thoughts could lead to greater attempts to avoid self-
focused attention altogether, which may lead to more destructive and self-
defeating behaviors, such as alcoholism (Baumeister, 1991 ). On the other

hand, because those greater in self-complexity have a greater number of

38

independent self-aspects, they should have less difficulty suppressing negative
self-relevant information while focusing on themselves and, therefore, may be
less likely to deal with self-focused attention in destructive ways. Although it will
take additional work to explore the possibility that those lower in self-complexity
will engage in more destructive behavior, this line of thinking is consonant with
the research findings of Dixon and Baumeister (1991). These researchers
found that people lower in self-complexity attempted to escape from self—
awareness more quickly than those greater in self-complexity following negative
self-relevant feedback.

Contrary to expectations, the analysis involving the duration of mentions did
not reveal evidence consistent with the hypothesis. This may be because the
duration of mentions is a less reliable measure of rebound than the number of
mentions. Although one investigation (i.e., Wegner 8. Gold, 1995) using
duration of mentions was partly successful in demonstrating the rebound effect,
all other investigations have only reported evidence using the number of
mentions as the measure of rebound.

The second hypothesis predicted that following suppression of negative self-
relevant thoughts while focusing on internal distracters, those lower in self-
complexity would exhibit self-judgments greater in negativity. Contrary to the
prediction, this hypothesis did not receive support. It was suggested that those
lower in self-complexity would encounter the to-be-avoided self-aspect more
often, resulting in greater accessibility of negative self-relevant information for

them. It may be that most students had relatively positive beliefs about

39

themselves as students, and thus, recurrent thinking about their student self-
aspect was not completely negative. In fact, the findings for accessibility in the
current study do show that those lower in self-complexity did reveal greater
accessibility of student-related concepts. However, the student self-aspect
probably contained a sizable amount of positive information in addition to the
negative feedback information provided to them during the experiment. This
preexisting positive information in combination with the negative feedback
information probably resulted in an accessible student self-aspect that was not
overly negative. Therefore, the greater accessibility of the student self-aspect
would not increase negative self-judgments for those lower in self-complexity.
The third hypothesis predicted that following suppression of negative self-
relevant thoughts while focusing on internal distracters, those lower in self-
complexity would report a stronger resistance to participating in a task similar to
that which previously provided negative feedback. Analyses were conducted
on two task preferences: analytical and creative. Contrary to the predictions,
neither analysis revealed evidence consistent with the hypothesis. Similar to
the second hypothesis, this prediction rested on the expectation that those
lower in self-complexity would experience greater accessibility of negative self-
relevant information. As discussed above, the increased accessibility of the
student self-aspect as a whole may not have been experienced as negative as
was initially expected. If the student self-aspect contained a sizable amount of
positive information, raising its accessibility would not likely lead participants

lower in self-complexity to have a greater need to avoid tasks similar to that

40

which resulted in negative feedback. In fact, raising its accessibility may have
actually helped to alleviate the effects of the negative feedback.

Interestingly, a marginal interaction between self-complexity and
suppressers versus expressers was found for creative task preference.
Suppressers greater in self-complexity reported a stronger preference for the
creative task. As mentioned previously, it is possible that the self-aspects of
individuals greater in self-complexity reflect more distinct and varied interests.
Following suppression, the accessibility of student-related thoughts may have
led those greater in self-complexity to desire a different type of task from that
which was previously given. Thus, they were more likely to prefer the creative
task. Those lower in self-complexity, on the other hand, may have been more
comfortable performing a task similar to that which they completed previously.

The final hypothesis predicted that those lower in self-complexity would
demonstrate a greater perseverance of negative beliefs following suppression
while focusing on internal distracters. Contrary to the prediction, the analysis
did not reveal any evidence consistent with the hypothesis. Again, this may be
because the increased accessibility of the student self-aspect as a whole may
not have been experienced as negative as was initially expected.

In addition to the primary hypotheses discussed above, a number of ancillary
hypotheses were tested. The first of these was conducted to assess the
accessibility of student-related thoughts following the suppression. Previous
research by Macrae et al. (1996) found that participants who previously

suppressed stereotypical thoughts revealed greater accessibility of such

41

thoughts than did participants who previously expressed such thoughts.
Therefore, it was first expected that suppressers would reveal greater
accessibility of student-related thoughts than would expressers. It was also
expected, based on the current research, that this relationship would be
qualified by self-complexity and distracter type. As expected, participants who
previously suppressed thoughts about the feedback and about their student
self-aspect revealed greater accessibility of student-related thoughts than did
participants who were not told to suppress such thoughts. There was also
marginal evidence that this relationship varied as a function of self-complexity,
but not as a function of distracter type. In particular, suppressers lower in self-
complexity tended to reveal greater accessibility of student-related thoughts
than did suppressers greater in self-complexity. This finding provides some
support for the proposition that those lower in self-complexity may be more
likely to inadvertently raise the accessibility of the to-be-avoided self-aspect
during suppression. Interestingly, this finding along with the finding of the first
hypothesis for self-complexity provides some support for the Wagner’s (1994)
model of mental control and also the extension of this model posited by Macrae
et al. (1994). In particular, the first hypothesis for self-complexity showed that
the use of different types of distracters by the controlled operating process
during suppression leads to differences in the relative amount of rebound
between the two suppression groups. This finding is in accordance with
Wagner’s view that failures of the operating process (due to less suitable

distracters) produce greater accessibility of unwanted thoughts, which

42

subsequently leads to greater rebound. But the finding that suppressers lower
in self-complexity tended to reveal greater accessibility of student-related
thoughts, along with the fact that this relationship did not vary as a function of
distracter type, provides some support for Macrae et al.’s extension of Wagner’s
model. This extension proposed that the process of thought suppression itself
increases the accessibility of the unwanted thought, even if the operating
process is successful in its job. The current findings are in accordance with this
proposition.

The second ancillary test was conducted to examine the potential
relationship between self-complexity and self-esteem. Prior research exploring
this relationship has produced mixed results. In particular, Campbell et al.
(1991) found a positive relationship between them, Woolfolk et al. (1995) found
a negative relationship between them, and Niedenthal et al. (1992) found no
relationship between them. The current investigation revealed a negative
relationship between the two, showing that greater self-complexity was related
to lower self-esteem. This finding is accordance with that of Woolfolk et al.
(1995). Together, these mixed findings suggest that there may be a moderator
variable in the relationship between self-complexity and self-esteem that has
not yet been identified. One possibility may be the types of self-aspects that
make up one’s self-concept. In particular, it is widely accepted that human
beings have an almost innate need to develop and maintain social
relationships. Research (Gottlieb & Green, 1984; Lu 8. Chen, 1996; Shaver,

F urrnan, & Buhrmester, 1985) has shown that those who develop strong social

43

networks tend to cope better with life’s challenges than those who do not
develop social networks. It is possible that those with a self-concept made up
of self-aspects that reflect social relationships have higher self-esteem than
those with a self-concept made up of self-aspects that reflect nonsocial
interests. That is, self-aspects may vary qualitatively in ways that facilitate their
ability to ameliorate stress during difficult times.

Another possible moderator may be the breadth of the self-aspects that
make up one’s self-concept. More specifically, it may be that those who have
fewer, yet broader, self-aspects cope better than those who have many, yet
narrower, self-aspects. For instance, an individual who has a student self-
aspect that contains information about her general role as a student may be
more stable and self-assured than an individual who describes each of the
components of his student role (e.g., “when I am taking tests,” “when I get an
A,” “when I get a C,” “when lam studying”) separately. It would seem that the
latter individual would experience much more variance in his reactions to events
within his student role than would the former individual.

The third ancillary test examined the affective-extremity hypothesis
associated with self-complexity theory. This hypothesis suggests that lower
self-complexity is related to more extreme affective responses to self-relevant
feedback, whereas greater self-complexity is related to more moderate affective
responses given similar feedback. The methods used in the current study
allowed for a replication of Linville’s (1985) finding that following negative

feedback those lower in self-complexity revealed more negative self-evaluation

and mood than did those greater in self-complexity. The current research found
a relationship between self-complexity and self-esteem change, suggesting that
those lower in self-complexity reported a greater drop in self—esteem following
the negative feedback than did those greater in self-complexity. This finding
replicates that of Linville (1985) for those given negative feedback and is in
accordance with the affective-extremity hypothesis. A marginal relationship
between self-complexity and positive mood change was also found, suggesting
that after receiving negative feedback, those lower in self-complexity reported a
larger drop in positive mood than did those greater in self-complexity. Again,
this is consistent with Linville’s (1985) affective-extremity hypothesis. No
relationship between self-complexity and negative mood change was found. It
is possible that positive moods may be more variable in response to life events
than negative moods. Furthermore, because there was no relationship found
between the positive mood and negative mood scores in the currentresearch, it
could be argued that positive feelings and negative feelings fall along separate
dimensions. This reasoning is consistent with research (e.g., Diener &
Emmons, 1984; Watson & Tellegen, 1985) suggesting that positive and
negative affect are relatively independent dimensions. Similar to circumplex
models of affect, independent dimensions of affect could explain greater
variability across moods. For example, a truly negative mood could consist of
high ratings on negative mood adjectives and low ratings on positive mood

adjectives, whereas a truly positive mood would consist of the opposite ratings.

45

A modest negative mood could consist of moderate ratings on negative mood
adjectives and low ratings on positive mood adjectives, and so forth.

The final ancillary test involved an examination of the conceptual process of
thought suppression and subsequent rebound. This process suggests that the
more an individual tries to suppress a thought, the more that same individual
subsequently ends up thinking about it when suppression is no longer a goal.
Inverse relationships between the initial suppression and subsequent
expression periods for both the number of mentions and the duration of
mentions would be in accordance with this process. Contrary to expectations, a
marginal positive relationship between the duration of mentions in the
suppression period and the duration of mentions in the expression period was
found for those suppressers focusing on a white bear. This suggests that the
longer an individual talks about the to-be-suppressed thoughts in the
suppression period, the longer that individual tends to talk about the previous
to-be-suppressed thoughts in the subsequent expression period. This finding is
not in accordance with the conceptual process suggested by Wegner et al.
(1987). In addition, no relationship for the number of mentions between the two
periods was found for suppressers focusing on a white bear. Furthermore, no
relationships between the two periods were found for either the number of
mentions or the duration of mentions for those suppressers focusing on other
self-aspects. Wagner’s (1994) model of mental control suggests that each time
the operating process fails to find a suitable distracter, the unwanted thought

becomes accessible, leading to an overt mention of the thought. The current

46

lack of relationship between overt mentions in the suppression period and overt
mentions in the expression period suggests that the conceptual process of
thought suppression cannot be explained entirely by failures of the controlled
process. As mentioned previously, Macrae et al. (1994) suggest that the
monitoring process by itself may be sufficient to produce the basic rebound
effect. With this in mind, the monitoring process may be better able to account
for the conceptual process involved in thought suppression and subsequent
rebound than can the operating process. It should be noted that because
participants were provided with distracters on which to focus during
suppression, the current study might not be a reliable examination of this
relationship. As previously mentioned, Wegner, et al. (1987) found that the
rebound effect was attenuated when participants were provided with a specific
distracter on which to focus during suppression. The same attenuation of
rebound may have occurred in the current study. Thus, further research is
needed to address this issue.

In addition to the above analyses, compartmentalization was also examined
in a series of tests that paralleled the primary hypotheses.
Compartmentalization refers to another way in which self-concept organization
can vary among individuals; and accordingly, its role in thought suppression
was examined. Contrary to the predictions, the analyses did not reveal
evidence consistent with any of the hypotheses.

However, there were two findings involving compartmentalization that are of

interest. First, contrary to what would be expected based on previous research

47

involving compartmentalization, compartmentalized participants with a fewer
number of positive and important self-aspects (i.e., negatively-
compartmentalized participants) tended to report greater expectations regarding
how well they would perform on a task similar to that for which negative
feedback was provided. One possible explanation of this finding is that
negatively-compartmentalized individuals may tend to have greater
expectations of their future performance than is justified based on their past
performance. Such expectations may lead to greater disappointment following
future failures, which would add to the negative self-evaluation that is
characteristic of negatively-compartmentalized individuals.

Second, consistent with what would be expected based on theory involving
compartmentalization, negatively-compartmentalized participants reported a
stronger preference for the creative task following negative feedback This
finding along with the finding above suggests that negatively-
compartmentalized individuals may believe that they will perform a previously
failed task better in the future, but when given the opportunity they prefer
experiences that will not provide such objective feedback.

For the present, it can be assumed that compartmentalization does not play
an important role in the process of thought suppression and subsequent
rebound. Unlike self-complexity, compartmentalization does not take the
relationship among self-aspects into account, which may be one limitation of its

role in thought suppression.

48

In conclusion, it seems that self-concept organization can play an important
role in the process of thought suppression. In particular, the self-concept may
serve as an aid in distracting oneself from unwanted thoughts. In terms of self-
complexity, the self-concept allows those with a greater number of independent
self-aspects to avoid a currently unpleasant self-aspect without avoiding self-
focused attention altogether. However, those with fewer and more interrelated
self-aspects do not have the luxury of focusing on other self-aspects as a way
of successfully avoiding an unpleasant self-aspect. The relatively greater
number of interconnections among fewer self-aspects has the undesirable
consequence of raising the accessibility of the to-be-avoided self-aspect,
leading them to think about the unpleasant self-aspect more often following
attempts at suppression. Accordingly, those lower in self-complexity should be
more likely to attempt to reduce self-focused attention (Dixon & Baumeister,
1991), which may lead them to engage in more self-destructive behaviors, such
as substance abuse.

The role of distracters in the process of thought suppression and subsequent
rebound has received surprisingly little attention. However, the current study
shows that the type of distracter used during suppression plays a very important
role in the relative success of thought suppression. In particular, this research
demonstrated that different types of distracters can impact the success of
thought suppression, and the organization of the self-concept itself may

influence the relative efficacy of a distracter that is associated with the self.

49

Although previous research has found a relationship between self-concept
organization and affective reactions to life events, the process underlying such
a relationship is not well understood. The current research illustrates one way
in which self-concept organization can influence the experience of affect by
revealing its impact on suppressing unwanted self-relevant thoughts.

Future research should focus on the role of other distracters involved in
successful versus unsuccessful thought suppression. Attention should also
focus on trying to discover the precise process underlying thought suppression
and subsequent rebound, and it should further develop a model of thought
suppression that integrates both past and present findings. In terms of self-
complexity, future research should continue to focus on trying to discover a
potential moderator variable involved in the relationship between self-
complexity and self-esteem.

On the practical side, the current research suggests that developing distinct
and varied interests may be a healthy way to avoid unpleasant self-relevant
thoughts. Although dividing your eggs among various cognitive baskets may
prove challenging, keeping them numerous and separate can be beneficial

when one is left with egg on one’s face.

50

APPENDICES

51

APPENDIX A

52

APPENDIX A

Calculation of H using the Self-Aspects of a Participant in the Current Study

“At School” “th Friend_s_” “Wit_h Girlfriend” “With Family: “With Strangers”

 

Lazy Communicative Comfortable Tense Insecure
Successful Giving Self-centered Disagreeing Communicative
Independent Friendly Inferior Insecure Independent
Organized Self-centered Communicative Independent Tense
Happy Energetic Isolated

Happy Hardworking

Giving Incompetent

Insecure

Optimistic

Hardworking

Combination #1 (traits in Group 1 only): Lazy, Successful, Organized (n1 = 3)

Combination #2 (traits in Group 2 only): Friendly (n2 = 1)
Combination #3 (traits in Group 3 only): Comfortable, Inferior, Energetic,

Optimistic (n3 = 4)
Combination #4 (traits in Group 4 only): Disagreeing, Isolated,

Incompetent (n4 = 3)
Combination #5 (traits in Groups 1 and 4): Independent (ns = 1)
Combination #6 (traits in Groups 2 and 3): Giving, Happy, Self-centered (n5 = 3)
Combination #7 (traits in Groups 3 and 4): Insecure, Hardworking (n7 = 2)
Combination #8 (traits in Groups 4 and 5): Tense (n3 = 1)

Combination #9 (traits in Groups 2 and 3): Communicative (n9 = 1)

Remaining Group (traits in no group): Capable, Confident, Disorganized, Fun
and Entertaining, Hopeless, Immature, lndecisive, Intelligent, Interested,
Irresponsible, Irritable, Like a failure, Lovable, Mature, Needed, Outgoing, Sad

and Blue, Uncomfortable, Unloved, Weary, Worthless (n10 = 21)

fl = logz40 - 1/40 (3logz3 + “0921 + 4logz4 + 3l0923 + 1logz1 +3logz3 +
210922 +1Iogz1 +1logz1 +10logz10)

fl = 5.34 - 1l40 (4.77 + 0 + 8.00 + 4.77 + 0 + 4.77 + 2.00 + 0 + 0 + 92.55)
H = 5.34 — 2.92
[:1 = 2.42

53

APPENDIX B

APPENDIX B

§_t_r_e_§_m of Cfionsciousness lnstytgtions

During the next several periods, you will simply be asked to verbally report
your stream of consciousness. That is, to indicate what is going through your
mind. Measures have been taken to ensure your privacy and to guarantee
confidentiality concerning your participation (no identifying information will be
associated with the audio tapes and none of the tapes will be used until fie;
the semester is over). When asked to report on your thoughts, please convey
whatever information you can on your ongoing thoughts at that moment. Your
report might include, but is not limited to, images, ideas, memories, feelings,
fantasies, plans, sensations, observations, daydreams, objects that catch your
attention, and efforts to solve a problem. There are no restrictions,
qualifications, conventions, or expectations; simply report on whatever is going
through your mind (whatever you are conscious of or aware of). It is important
that you continuously verbalize your thoughts during the wh_ole 5-minute

period...try not to pause for long periods of time.

Period 1 (All Participants):
For the first 5-minute period, simply report on everything that comes to mind,
even if what comes to mind involves the feedback you were given on the

analogy task and how it may be related to your academic life.

55

Period 2 (Manipulation):
Internal Suppressers:

Another experiment not related to this study had the computer output the
groups that you described in the first session of the experiment and put them in
this envelope.

For the second 5-minute period, please verbalize your thoughts as you did
before, with one exception. This time, try not to think about the feedback you
were given on the analogy task or your academic life, but mention it if you do.
Instead, try to think about one or more of the other groups that you described in
the first session listed in this envelope. Go ahead and look over those for a few
minutes. ( In order to guarantee confidentiality, participants were given a new
envelope in which to put the list of their groups and were asked to sealed it.)
Again, remember, _d_p_r_1’_t think about the feedback or your academic life, but

mention it if you do.

External Suppressers:

For the second 5—minute period, please verbalize your thoughts as you did
before, with one exception. This time, try not to think about the feedback you
were given on the analogy task or your academic life, but mention it if you do.
Instead, try to think about a white bear. Again, remember, M think about the

feedback or your academic life, but mention it if you do.

56

Expressers:
For the second 5-minute period, again simply report everything that comes
to mind, even if what comes to mind involves the feedback you were given on

the analogy task and anything else related to your academic life.

Period 3 (All Participants):
For the last 5-minute period, simply report everything that comes to mind,
even if what comes to mind involves the feedback you were given on the

analogy task and how it may be related to your academic life.

57

APPENDIX C

58

[Jght
skHl
[est
smart
fooJ
grape
shame
prain
dpnce
report
Learn
sharp
guess

know

APPENDIX C

Word Completion Items

guH
plass
dpsk
Wu
bag
pen
fine
slow
book
[an
[each
poor

keen

59

APPENDIX D

60

APPENDIX D

Table 1. Descriptive Statistics for Primary Variables across All Participants

 

 

Variable M _S_[_)
Scott’s [1 2.49 .92
Total Number of Self-Aspects 4.58 2.37
Compartmentalization .67 .29
Positive Importance 30.68 7.76
Self-Esteem (Time 1) 31.35 4.91
Positive Mood (Time 1) 29.46 7.16
Negative Mood (Time 1) 17.67 5.71
Self-Esteem (Time 2) 30.69 5.09
Positive Mood (Time 2) 26.97 8.14
Negative Mood (Time 2) 17.31 6.20
Number of Mentions (Period 1) 2.73 1.29
Number of Mentions (Period 2) 1.15 1.11
Number of Mentions (Period 3) 1.70 1.22
Duration of Mentions (Period 1) 84.25 67.62
Duration of Mentions (Period 2) 43.40 63.71
Duration of Mentions (Period 3) 74.75 85.93
Self-Judgment 3.20 1 .41
Analytical Task Preference 2.72 1.58
Creative Task Preference 5.44 1.26
Belief Perseverance 1 1.65 4.95
Implicit Accessibility 9.91 2.17

 

Hie, All of the statistics, excluding compartmentalization, are based on a
sample size of 98. Compartmentalization scores can only be calculated for
participants who have at least two negative and two positive traits across their
self-aspects. In this study, nine participants had less than two negative traits
across their self-aspects. Therefore, they are not included in the

compartmentalization measure.

61

APPENDIX E

62

APPENDIX E

Figure 1. Interaction Between Self-Complexity and Distracter Type for Number
of Mentions

 

Internal Suppressers - - - External Suppressers

 

Number of Mentions
to
l

 

 

 

 

Low High
Self-Complexity

63

APPENDIX F

APPENDIX F

Figure 2. Marginal Interaction Between Self-Complexity and Suppressers
Versus Expressers for Creative Task Preference

 

Suppressers — — Expressers

 

01
I

 

Creative Task Preference
co 4s
1 l

 

 

 

Low High
Self-Complexity

65

APPENDIX G

66

APPENDIX G

Figure 3. Marginal Interaction Between Self-Complexity and Suppressers
Versus Expressers for Number of Student-Related Words Created

 

 

 

 

 

Suppressers - - - Expressers
12
1: 11 '-
.93.
g 10 -
o - - — - _ - - -
it) 9 " - " - '-
'2
E 8 -
8 -
E 7
a)
E 6 -
7%
.3 5 "
m D
'05 4
6
.0 h
E 3
3
z 2 -
1 I 1
Low High
Self-Complexity

67

APPENDIX H

68

APPENDIX H

Figure 4. Marginal Interaction Between Self-Concept Organization and Positive
Importance for Self-Judgment

 

Low Positive Importance - - - High Positive Importance

 

00
I

 

Self-Judgment

N
I

 

J l

 

 

Compartmentalized Integrated
Self-Concept Organization

69

APPENDIX I

70

APPENDIX I

Figure 5. Interaction Between Self-Concept Organization and Positive
Importance for Creative Task Preference

 

Low Positive Importance - - - High Positive Importance

 

 

Creative Task Preference

 

 

 

0 l l
Compartmentalized Integrated

Self-Concept Organization

71

REFERENCES

72

REFERENCES

Baumeister, R. F. (1991). Escaping the self: Alcoholism. spirituality,
masochism and other flights from the burden of selfhoog New York: Basic
Books.

 

Campbell, J. D., Chew, 8., & Scratchley, L. S. (1991). Cognitive and
emotional reactions to daily events: The effects of self-esteem and self-
complexity. Journal of Personality, 59, 473-505.

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

Dixon, T. M., & Baumeister, R. F. (1991). Escaping the self: The moderating

effect of self-complexity. Personality and Social Psychology Bulletin, 17,
363-368.

Gottlieb, N. H., & Green, L. W. (1984). Life events, social network, life-style,
and health: An analysis of the 1979 National Survey of Personal Health
Practices and Consequences. Health Education Quarterly, 11, 91-105.

Kelly, A. E., & Kahn, J. H. (1994). Effects of suppression of personal
intrusive thoughts. Journal of Personality and Social Psychology, 66, 998-1006.

Linville, P. W. (1985). Self-complexity and affect extremity: Don’t put all your
eggs in one cognitive basket. Social Cpgnition, 3, 94-120.

Linville, P. W. (1987). Self-complexity as a cognitive buffer against stress-

related illness and depression. Journal of Personality and Social Psychology,
p; 663-676.

Lu, L., & Chen, C. S. (1996). Correlates of coping behaviors: lntemal and
external resources. Counseling Psychology Quarterly, 9, 297-307.

Macrae, C. N., Bodenhausen, G. V., Milne, A. 8., & Jetten, J. (1994). Out of
mind but back in sight: Stereotypes on the rebound. Journal of Personally and

Social Psychology, 67, 808-817.

Macrae, C. N., Bodenhausen, G. V., Milne, A. B., & Wheeler, V. (1996). On
resisting the temptation for simplification: Counterintentional effects of
stereotype suppression on social memory. Social Cognition, 14, 1-20.

Monteith, M. J., Shaman, J. W., & Devine, P. G. (1998). Suppression as a

stereotype control strategy. Personalig and Social Psychology Review, 2, 63-
82.

73

Niedenthal, P. M., Setterlund, M. 8., & Wherry, M. B. (1992). Possible self-
complexity and affective reactions to goal-relevant evaluation. Journal of

Personality and Social Psychology, 63, 5-16.

Pelham, B. W., & Swann, W. 8., Jr. (1989). From self-conceptions to self-
worth: On the sources and structure of global self-esteem. Journal of

Personality and Social Psychology, 57, 672-680.

Pope, K. S. (1978). How gender, solitude, and posture influence the stream
of consciousness. In K. S. Pope & J. L. Singer (Eds), The Stream of
Consciousness (pp. 259-299). New York: Plenum Press.

Rosenberg, M. (1965). Society and the adolescent self-image. New Jersey:

Princeton University Press. Journal of Personality & Social Psychology, 32,
880-892.

Ross, L., Lepper, M. R., & Hubbard, M. (1975). Perseverance in self-
perception and social perception: Biased attributional processes in the

debriefing paradigm. Journal of Personality & Social Psychology, 32, 880-892.

Scott, W. A. (1969). Structure of natural cognitions. Journal of Personality
and Social Psychology. 1; 261-278.

Shaver, P., Furman, W., & Buhrmester, D. (1985). Transition to college:
Network changes, social skills, and loneliness. In S. Duck & D. Perlman (Eds),

Understanding personal relationships: An interdisciplina[y approach (pp.
193-219). London, England: Sage Publications.

Showers, C. J. (1992a). Compartmentalization of positive and negative self-
knowledge: Keeping bad apples out of the bunch. Journal of Personality and

Social Psychology, 62, 1036-1049.

Showers, C. J. (1992b). Evaluatively integrated thinking about

characteristics of the self. Personality and Social Psychology Bulletin, 18, 719-
729. -

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

Personali[y and Social Psychology, 54, 1063-1070.

Watson, 0, & Tellegen, A. (1985). Toward a consensual structure of mood.
Psychological Bulletin, 98, 219-235.

Wegner, D. M. (1994). Ironic processes of mental control. Psychological
Review 101 34-52.

 

74

Wegner, D. M., & Gold, D. B. (1995). Fanning old flames: Emotional and
cognitive effects of suppressing thoughts of a past relationship. Journal of

Personality and Social Psychology, 68, 782-792.

Wegner, D. M., Schneider, D. J., Carter III, S. R., & White, T. L. (1987).
Paradoxical effects of thought suppression. Journal of Lersonalitv and Social

Psychology, 53, 5-13.

Wenzlaff, R. M., Wegner, D. M., & Klein, S. B. (1991). The role of thought
suppression in the bonding of thought and mood. Journal of Personality and

Social Psychology, 60, 500-508.

Woolfolk, R. L., Novalany, J., Gara, M. A., Allen, L. A., & Polino, M. (1995).
Self-complexity, self-evaluation, and depression: An examination of form and

content within the self-schema. Journal of Personality and Social Psychology,
pp, 1108-1120.

75

"IililllllllllllllllTill“