.,. 3
it u
.

7
; 31:93..

.
. A $4
1.31,.)

 

 

Y‘fif .
t‘,
.L’-‘

~ «:02

This is to certify that the

dissertation entitled

Psych0pathic and Narcissistic Personality
Traits: Impact of Ego Threat on the

Experience of Anger
presented by

Angela Mae McBride

has been accepted towards fulfillment
of the requirements for

Pb - D _ degree in W

MW M

Major professor

 

 

Date 6/10/02

 

MS U is an Affirmative Action/Equal Opportunity Institution 0- 12771

 

 

 

LIBRARY
Michigan State
University

 

 

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c;/CIRC/DateDue.p65.p. 15

PSYCHOPATHIC AND NARCISSISTIC PERSONALITY TRAIT S:
IMPACT OF EGO THREAT ON THE EXPERIENCE OF ANGER

By

Angela Mae McBride

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Psychology

2002

ABSTRACT

PSYCHOPATHIC AND NARCISSISTIC PERSONALITY TRAITS: IMPACT OF EGO
THREAT ON THE EXPERIENCE OF ANGER

By

Angela Mae McBride

This study was designed to further our understanding of links between personality,
anger, and aggression, by building on previous research related to self-esteem and
narcissism. In this study, psychopathic and narcissistic traits in a noncriminal sample
were assessed, along with their relations to anger experience in the face of ego threats.
Participants (N = 162) completed scales of psychopathy (Psychopathic Personality
Inventory), narcissism (Narcissistic Personality Inventory) and self-esteem (Rosenberg’s
Self-Esteem Scale). The sample of participants was randomly divided into an
experimental manipulation group and a control group. Experimental group participants
read a series of vignettes in which they imagined themselves in a number of situations in
which their self-esteem was threatened (ego threat). Control group participants read
vignettes describing similar situations, but with neutral or unresolved outcomes. All
participants then completed a mood questionnaire and a measure of state and trait anger
(State-Trait Anger Expression Inventory).

Results indicated positive correlations between trait anger and both psychopathic
and narcissistic personality traits. Contrary to predictions, there was no interaction
between narcissistic traits and self-esteem on the outcome variable of state anger

following the ego threat, nor was there a main effect of self-esteem on state anger.

Levels of psychopathic and narcissistic traits were higher among males than
females, but no gender differences were found in level of self-esteem. No gender
differences were present in the relationship between psychopathic and narcissistic traits
and the effect of the ego threat on state anger. However, the ego threat affected males and
females differently; males who received the ego threat were less angry than males who

did not, while the reverse pattern was found for females.

ACKNOWLEDGMENTS

Writing this dissertation involved a lot of time working alone. However, the
finished product is the result of the efforts of many people. I would like to thank the
members of my guidance committee: Dr. Robert Caldwell, Dr. Raymond Frankmann, and
Dr. Dozier Thornton, for helping to clarify conceptual issues and dispensing much needed
statistical advice. Dr. Norman Abeles, my chairperson, deserves special thanks for
encouraging me to pursue my interests and supporting me at every step of the dissertation
process.

One of the most positive aspects of graduate school has been the friends I have
made along the way. Thanks to Alissa, Dave, David, Edwin, and Ruth Ellen -- without
your support this would have been a lonely enterprise. And Jodi — even though you were
living elsewhere through much of this process, you were always with me in spirit. I
couldn’t ask for a better friend.

My family has truly been the foundation that has allowed me to build my dream of
getting a Ph.D. They have helped in ways too numerous to list, but I haven’t forgotten any
of the things, great and small, that have made it easier to get where I am today. This

would not have been possible without you. Many thanks, and much love!

iv

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................ vi
LIST OF FIGURES .......................................................................................................... vii
INTRODUCTION ........................................................................................................... . l
The Psychopathic Syndrome ................................................................................ 2
Assessment of Psychopathy .................................................................................. 5
Gender Differences in Psychopathy and Psychopathic Traits ............................ .11
The Narcissistic Syndrome... ...13
Self-Esteem: Definitions and Associated Behaviors .................................. 16
Definition and Assessment ofAnger22
Hypotheses .................................................................................. 23
METHOD ....................................................................................................................... .27
Participants .......................................................................................................... 27
Measures .............................................................................................................. 27
Procedure .................................................................................... 31
RESULTS ........................................................................................................................ 35
Hypothesis I ................................................................................. 35
Hypothesis II. ................................................................................................ 36
Hypothesis III ...................................................................................................... 37
Hypothesis IV ...................................................................................................... 40
DISCUSSION ................................................................................................................ .42
Relationship of Trait Anger to Psychopathy and Narcissism. . . . . . . . . ....43
Narcissistic Traits, Self-esteem, and Anger ...................................................... 48
Psychopathic Traits, Self-esteem, and Anger ........................................... 50
Gender Differences in Psychopathic Traits.. ..................................................... .51
Gender Differences in Narcissistic Traits. ........................................................ 53
Gender Differences in Self-Esteem. ............................................................... 54
Anger as an Outcome of Reading Vignettes ........................................... 57
Limitations of the Study ........................................................................................ 60
Implications. . . . . . .. ................................................................................................ 61
APPENDIX A: Vignettes .......................................................................... 63
APPENDIX B: Mood Questionnaire ............................................................ 67
LIST OF REFERENCES ............................................................................................... 68

LIST OF TABLES

Table l — pg. 16

Subclinical personality trajfi vs. disorder

Table 2 - pg. 35

Correlations between personality traits and trait anger

Table 3 - pg. 36

State anger scores for low vs. high NPI and RSE (means and standard deviations)

Table 4 - pg. 37

State anger scores for low vs. highPPI and RSE (means and standard deviations)

Table 5 - pg. 39

Comparison of males and femles on PPI subscale tota_l§

Table 6 — pg. 40

Comparison of males and females on NPI subscgle totals

LIST OF FIGURES

Figure 1 - pg. 41

State anger scale scores

vii

INTRODUCTION

The study of factors that contribute to anger and aggression is an area that has
long been investigated by researchers in clinical and social psychology. For a number of
years, the predominant belief was that low self-esteem led individuals to behave
aggressively, and it was assumed that raising self-esteem would help to combat this
problem. However, in recent years, this belief has been challenged by the presence of
conflicting research results, some of which have linked high, rather than low, self-esteem
to violence. This study seeks to further our understanding of links between personality,
anger, and aggression, by building on findings related to self-esteem and narcissism. In
this study, psychopathic and narcissistic traits in a noncriminal sample are assessed, along
with their relations to an imagined anger experience in the face of ego threats.

This literature review begins with a discussion of the construct of psychopathy,
including issues regarding its definition, assessment, and links to theories of personality,
as well as gender differences in psychopathic traits. Then the construct of narcissism is
described from a range of theoretical orientations. The often fuzzy concept of “self-
esteem” is defined for the purposes of the proposed study, and previous research
regarding ego threats and their impact on individuals with high and low self-esteem is
reviewed. Anger is analyzed in terms of its varied components, including state and trait

anger. Finally, advantages of and predictions for the proposed study are described.

The Psychopathic Syndrome

In his seminal work on psychopathy, The Mas_k of Sam, Cleckley (1941) derived
a set of criteria for identifying individuals with this disorder. This consisted of a list of
sixteen personality traits and behaviors, including superficial charm and intelligence, lack
of remorse, poor judgment and failure to learn by experience, as well as pathological
egocentricity. Cleckley believed that both specific observable actions (such as lying and
other antisocial behavior) as well as inferred personality traits (such as absence of
nervousness and loss of insight) were part of the psychopathic syndrome, and he provided
the framework for later research into this subject.

Hare’s program of research has built upon, and provided an empirical basis for,
the Cleckley criteria. Most of his work has involved the study of incarcerated individuals,
and he developed the Psychopathy Check List - Revised (PCL—R; Hare, 1991) as a
diagnostic tool for use in prisons. His model of psychopathy (Hare, 1993) includes two
dimensions: an emotional/interpersonal factor, including symptoms such as egocentric
and grandiose, and deceitful and manipulative, as well as a social deviance factor,
comprised of symptoms such as impulsivity and poor behavioral controls. Hare has
found that a major component of the psychopathic syndrome is extreme self-centeredness
and grandiosity: “Psychopaths have a narcissistic and grossly inflated view of their self-
worth and importance, a truly astounding egocentricity and sense of entitlement, and see
themselves as the center of the universe, as superior beings who are justified in living
according to their own rules” (Hare, 1993, p. 38). Part of this egocentricity includes the

setting of lofty goals, without a realistic conception of the skills and effort required to

reach those goals. Psychopaths assume that they are highly deserving of praise and
respect from others, regardless of whether their actions justify this adulation.

Widiger and Lynam (1998) provide a link between Hare’s 20 Items of the PCL-R
and their corresponding domains in the Five Factor Model (FF M) of personality
functioning (McCrae & Costa, 1990). They state that the egocentricity aspect of
psychopathy is represented in the arrogance subcategory of antagonism (vs.
agreeableness). In addition, psychopaths are characterized by low or absent self-
consciousness (an aspect of neuroticism), and high excitement-seeking (a component of
extraversion), among other traits. The main subcategory relevant to the
emotional/interpersonal component of psychopathy appears to be the antagonism
construct (low agreeableness). The social deviance aspect is best captured by low
conscientiousness, such as actions done impulsively and with little regard to their impact
on other people. Poor behavioral controls contribute to this as well, including
psychopaths’ poor modulation of anger, hostility, and aggression (Hare, 1991).

Other attempts have been made to further subdivide the construct of psychopathy.
Millon and Davis (1998) delineated ten subcategories based on clinical observation and a
review of research findings. They argued that psychopathy has as its core a “marked self-
centeredness and disdain for the needs of others,” (p. 161), yet there are notably different
patterns of personality organization within this overarching construct. For example, they
identify the covetous psychopath as “completely self-centered and self-
indulgent...unfulfilled, empty...dissatisfied” (p. 164-65), while the malevolent psychopath
shows “minimal guilt or remorse for their violent acts, but may instead display an

arrogant contempt for the rights of the others.” No attempt has yet been made to develop

a diagnostic tool to capture these proposed types. However, if empirically validated, they
would further add to our fimd of knowledge about psychopathy.

There is disagreement as to whether psychopathy is a categorical variable or a
continuous construct. For diagnostic purposes, the Diagnostic and Statistical Manual of
Mental Disorders, Fourth Edition (DSM-IV) contains the diagnosis of Antisocial
Personality Disorder, which captures many of the symptoms of psychopathy (American
Psychiatric Association, 1994). To make this determination, specific cutoffs regarding
the necessary type and number of symptoms are used. However, there is a growing
number of researchers studying subclinical levels of psychopathy fiom a variety of
theoretical orientations, who argue that psychopathy is best conceptualized as a
continuum. Meloy (198 8), working from a psychodynamic perspective, argued that
psychopathy may be defined for purposes of diagnosis as a categorical list of traits and
behaviors, but that for treatment, it is most useful to view psychopathy as a continuous
variable, with differing treatment implications fi'om one person to another. Similarly,
writing from an interpersonal orientation, Widiger and Lynam (1998) stated that
psychopaths do not have a unique personality organization, but rather that they exist on a
continuum with normal personality functioning. Research by Levenson, Kiehl, and
Fitzpatrick (1995) concluded that psychopathy is most accurately viewed along a
continuum of traits and behaviors, adding, “It seems plausible that a psychopathic
interpersonal style, even in a muted form, could be situationally amplified with
destructive consequences” (p. 157). Thus, the argument can be made that understanding
psychopathy along a continuum of psychopathic traits rather than an either-or construct

may be beneficial for understanding both clinical and everyday interpersonal interactions.

Identifying “true” psychopaths may be useful for tasks such as the prediction of
dangerousness in criminals (Hare, 1993), but it seems unlikely that they represent a
qualitatively different type of person than those found in a normal population.

The cluster of symptoms comprising the psychopathic syndrome are varied, but
marked by a grandiose sense of self, as well as impulsivity, poor behavioral controls, and
a lack of concern about the feelings of other people. We now turn to a discussion of the
ways psychopathy has been measured, in both criminal and noncriminal populations.
This information on assessment of psychopathy is included in order to demonstrate how
research in the area has progressed since Cleckley (1941) first developed his sixteen
criteria, and to show how researchers approach the study of psychopathy from a variety of
perspectives and techniques, including self-report, observation, and interview methods.
In particular, information relevant to the study of psychopathic traits in non-incarcerated
populations is presented. This serves to help define psychopathy as a personality
characteristic which exists along a continuum, and to highlight the need for further

research in areas addressed by the proposed study.

Assessment of Psychopathy

The study and assessment of psychopathy is marked by disagreement as to how
best to measure this construct. Lilienfeld (1994) highlights several major debates in the
literature. One issue centers around whether psychopathy should be measured through a
personality-based or behavior-based approach, with the former focusing on identifying
traits, and the latter emphasizing observable actions. The trait approach more closely

matches Cleckley’s (1976) concept of psychopathy, and is thus considered by some to be

a more valid representation of the construct. In addition, the trait approach would include
those “successful psychopa ” who are high-functioning and do not have the extensive
contact with law enforcement that more overtly antisocial psychopaths do. However, the
behavior-based approach, which includes such criteria as criminal behavior, results in
greater interrater reliability among researchers and clinicians.

The Psychopathy Checklist (PCL; Hare, 1980) and its current, revised version
(PCL-R; Hare, 1985) are the most widely researched and clinically used instruments for
studying psychopathy. These measures were devised based on the Cleckley criteria for
psychopathy (Cleckley, 1976), and were designed to be used with an incarcerated
population. The PCL—R contains 20 items, scored on a scale from 0 to 2, with 0
indicating that the item does not apply, 1 indicating that it applies somewhat, and 2
indicating that it definitely does accurately describe the individual. Items are rated based
upon a semi-structured clinical interview, as well as a review of file information. A score
of 30 is used as a cutoff to determine psychopathy, however, this is intended to be a rough
estimate rather than an absolute determination (Hart, Hare, & Harpur, 1992). When this
cutoff is used, 15 to 25 percent of incarcerated offenders in North America are estimated
to meet criteria for psychopathy (Hare, 1998). The source for this figure is unclear as to
whether this includes both violent and nonviolent offenders, or what particular types of
crimes are represented in the offender population being studied.

In addition to a total score, factor scores are also obtained. Two factors comprise
the PCL-R (Templeman & Wong, 1994); the first contains items regarding interpersonal
behavior and affect. According to Harpur, Hare, and Hakstian, (1989), this factor has

been shown to be most closely linked to measures of narcissism, negatively linked to

anxiety, as well as matching the prototypical psychopath as described by Cleckley.
Examples of items include “lack of remorse or guilt” and “egocentricity/grandiose sense
of self-worth”. Factor 2 consists of items descriptive of antisocial behavior,
irresponsibility, and impulsivity. This factor correlates more with DSM diagnoses of
antisocial personality disorder, scales 4 and 9 of the MMPI-2, and diagnoses of substance
abuse. Examples of items include “parasitic lifestyle” and “lack of realistic long-term
plans.” These factors correlate in the range of 0.5 to 0.6 in both forensic patients and
inmates (Hart, Hare, & Harpur, 1992). Although these factors are found to be robust
across studies, it is important to note that statistical analyses such as inter-item
correlations and principal components analyses support the unidirnensionality of the scale
as well (Harpur, Hare, & Hakstian, 1989).

A shorter, 12-item version of the PCL—R was recently devised to screen for
psychopathy in forensic populations (Hart, Cox, & Hare, 1995). The Psychopathy
Checklist — Screening Version (PCL:SV) contains the same factor structure as the PCL-R,
and provides adequate validity and reliability, thus, it can be used as a parallel form of the
longer scale (Hare, 1998). A cutoff score of 18 is used to diagnose psychopathy. An
item-response theory analysis of this measure indicated that as with the PCL-R, items in
Factor 1 have a higher threshold for scoring than those in Factor 2. That is, Factor 1
characteristics must be more prominent and noticeable to be scored (Cooke, Michie, Hart,
& Hare, 1999). Possibly this is due to the greater level of inference required to determine
whether the individual possesses traits as opposed to documented behaviors. Both forms
of the Checklist are effective in predicting violence and recidivism among inmates and

forensic patients (Hare, 1998). Salekin, Rogers, and Sewell (1996) wrote of the PCL—R

as being “unparalleled” as a predictor of dangerousness. In their meta-analysis of studies
which employed the PCL/PCL-R, using violence as an outcome measure, they found
effect sizes ranging fi'om 0.42 to 1.92; the mean effect size was 0.79. This robust finding
has led to use of the PCL/PCL-R by parole boards and others who undertake the serious
and difficult task of predicting dangerousness.

Prompted to a large extent by Hare’s development of psychopathy scales, other
researchers have developed instruments to assess psychopathy, ranging from observer
ratings to self-report measures. These measures are usefirl in that they can provide
additional information to the researcher or clinician, and are particularly important for use
among nonforensic, nonincarcerated populations, thus overcoming one potential
limitation of the Hare scales.

Reise and Oliver (1994) designed the Psychopathy Q-Sort (PQS) in order to allow
for observer ratings of psychopathy based on interpersonal interaction. Seven judges used
the 100 items of the California Q-Set to develop a Psychopathy Prototype, a measure of
primary psychopathy. Items were sorted into nine categories based upon the judges’
conceptions of prototypical psychopathic traits. Most characteristic items included “self-
indulgent,” “guileful, deceitful, manipulative, and opportunistic,” and “personally
charming.” Least characteristic items included “behaves in a giving way toward others,”
“ has a readiness to feel guilt,” and “is basically anxious.” When complete, the person’s
prototype receives a score from -1 .0 to 1.0, from complete dissimilarity from to total
agreement with the prototype.

The authors compared the psychopathy prototype to those for other constructs

which had been developed independently by other researchers. Prototype Q-Sorts for

hysteria and narcissism were compared to that of psychopathy, with notable similarities
and differences being found. The Psychopathy and Narcissism prototypes correlated
moderately highly, g = 0.51, indicating a great deal of overlap between the types.
However, the Narcissism prototype received higher values for items including,
“concerned with own adequacy as a person,” “is basically anxious,” “ and “is thin-
skinned; sensitive to anything that can be construed as an interpersonal slight.” Items
which received higher values in the Psychopathy prototype included, “is calm, relaxed in
manner,” and “is personally charming.” Reise and Oliver conclude, “Typically, primary
psychopaths show a marked absence of self-concern and do not react defensively when
criticized. . .In point of fact, they appear not to be able to fully appreciate insults or
criticism and do not often report experiencing such complex affective concepts as
humiliation or shame” (pp. 142-143).

In a follow-up study, Reise and Wink (1995) further explored the construct
validity of the Psychopathy Q-Sort by analyzing its correlations with DSM-1H personality
disorder clusters. According to the DSM-1H (American Psychiatric Association, 1980),
Cluster A personality disorders include Paranoid, Schizoid, and Schizotypal, and
generally refer to individuals who are viewed by others as odd or eccentric. Cluster B
encompasses borderline, narcissistic, histrionic, and antisocial personality disorders,
which entail behaviors which are emotional, dramatic, or highly variable. Finally, Cluster
C refers to the Avoidant, Dependent, and Obsessive-Compulsive personality disorders,
which refer to a high degree of anxiety or fearfulness in the individual’s personality.
Reise and Wink (1995) found high correlations between the PQS and Cluster B

personality disorders in DSM-III, and low correlations with Clusters A and C. However,

particularly among females, the PQS did not discriminate adequately among the Cluster B
disorders, suggesting that the construct assessed by the PQS has some degree of overlap
with all four of the disorders in this Cluster (defined above). The PQS also correlated
positively with measures of self-assurance and social poise and negatively with measures
of impulsivity. The participants in this study were normal individuals recruited fi'om the
community.

Levenson’s Self-Report Psychopathy Scale (LSRP) was developed by Levenson,
Kiehl, and Fitzpatrick (1995) in order to produce a two-factor psychopathy measure
appropriate for use in noninstitutionalized populations. They were interested in
differentiating primary from secondary psychopaths, using trait anxiety scales to
differentiate these types. This measure assesses primary psychopathy, similar to Hare’s
Factor 1, secondary psychopathy, aligned with Hare’s Factor 2, and antisocial action,
which load onto three scales. Thirty items are each endorsed on a 4-point scale, ranging
fi'om “disagree strongly” to “agree strongly.”

The most comprehensive self-report measure of psychopathy was developed by
Lilienfeld and Andrews (1996). The Psychopathic Personality Inventory (PPI) was
designed to assess the presence of psychopathic traits in noncriminal populations. As
discussed above, the Psychopathy Check List - Revised (PCL-R; Hare, 1991) is the most
widely used measure of psychopathy among incarcerated criminals; however, it has not
proven to be generalizable outside of the prison population (Lilienfeld & Andrews, 1996).
In addition, the PCL-R requires extensive interviewing, institutional record review, and
case history review, making it less applicable to “successful psychopa ”-- individuals

who do not have a legal record (Lykken, 1995). The PPI, a relatively new instrument, has

10

demonstrated high positive correlations with the Cleckley criteria for psychopathy, as
determined by interview and peer ratings. The 160-Item, true/false questionnaire also has

the advantage of easy administration due to the self-report format.

M Differences in Psychopathy and Psychopathic Tra_it§

A number of studies of noncriminal populations suggest that higher levels of
psychopathic traits are found in males than in females. Forth, Brown, Hart, and Hare
(1996) found higher scores for males than females on the majority of items of the
PCL:SV. However, they note the possibility of rater bias in this study; since ratings were
made by females only, they may have systematically ascribed different traits to males than
females. Lilienfeld and Andrews (1996) found that the factor structure of the PPI was the
same for males and females but that males scored significantly higher on the PPI total
score than females, in addition to obtaining higher scores on subscales such as
Machiavellian Egocentricity and Fearlessness. Support for a similar factor structure for
men and women who completed the LSRP was found by Lynam, Whiteside, and Jones
(1999); however, these authors did not report data on overall levels of psychopathy for
both genders.

Hamburger, Lilienfeld, and Ogben (1996) found that males with psychopathic
traits exhibited more Antisocial Personality Disorder characteristics, while females with
psychopathic traits exhibited more Histrionic Personality Disorder characteristics. They
suggest that, if future research supports this finding, psychopathic traits may be a
common basis for two syndromes which are moderated by biological gender. Notably,

they investigated the possibility that gender roles, rather than biological gender, were

11

responsible for this difference. However, this hypothesis was not supported. On a
measure containing subscales for primary and secondary psychopathy, a large gender
difference was found between males and females for primary psychopathy, with males
exhibiting much higher subscale means (Levenson, Kiehl, & Fitzpatrick, 1995).
However, the gender difference for secondary psychopathy, while significant, was much
smaller.

Among a sample of female inmates with a wide variety of offenses, Salekin,
Rogers, and Sewell (1997) found a lower prevalence ofpsychopathy (16%) as determined
by the PCL-R, compared to previous studies of prevalence among male inmates. Among
males, prevalence in correctional samples ranges from 25 to 30% (Hare, 1991). In
contrast to the findings from a noncriminal sample in which items loaded onto similar
factors for males and females (Lynam et al., 1999), Salekin et al. found a different factor
structure for incarcerated females than males. While a generally similar 2-factor structure
(i.e., traits and behaviors) did emerge for the PCL-R in their female sample, the types of
items comprising these factors varied from previously reported studies of males.

In terms of severe psychopathy, the estimated prevalence among men (1%) in the
United States is roughly three times that for women (Meloy, 1992). A case study of a
female psychopath which employed objective and projective personality tests (N esca,
Dalby, & Baskerville, 1999) indicated that the subject had high levels of anger, was very
aggressive, displayed a number of antisocial behaviors, and had difficulty modulating her
affect. These characteristics are in line with those described by researchers studying male
psychopaths. One notable finding of this study was that, while the subject scored high on

a measure of psychopathy (PCL-R = 34), she did not evidence an abnormal egocentricity

12

ratio on the Rorschach (0.39). The authors interpret this to mean that she did not have
elevated levels of narcissism, and they suggest that the theoretical link between
psychopathy and narcissism merits further investigation.

Varying approaches to the assessment of psychopathy highlight the diversity of
this construct, the consistent finding of gender differences in levels of psychopathic traits,
as well as the complexity of measuring psychopathy among people who, by definition,
have little insight into their own inner functioning. It is worthwhile at this point to
discuss another personality construct that is similarly diverse, and can also be destructive

when it exists to an extreme degree.

The Narcissistic Sm_dlome

The criteria for a DSM-IV (American Psychiatric Association, 1994) diagnosis of
Narcissistic Personality Disorder include five or more of the following traits: has a
grandiose sense of self-importance, has fantasies of power or success, feels special or
unique, requires excessive admiration, has a sense of entitlement, is interpersonally
exploitative, lacks empathy, often is envious of others, and shows arrogant behaviors or
attitudes. In addition, the DSM-IV states that individuals with NPD are highly sensitive
to perceived criticism fi'orn others, and that they may react with rage to what they view as
an attack. In terms of gender differences, men are more likely to exhibit the narcissistic
syndrome, with males representing 50-75% of those diagnosed with narcissistic
personality disorder (American Psychiatric Association, 1994).

The above criteria effectively capture the narcissistic syndrome described by

theorists flour a range of theoretical orientations. However, while the symptoms are

13

generally agreed upon, the purported causes and treatment for the syndrome are varied. In
addition, theorists do not generally draw a clear distinction, as does the DSM-IV, between
narcissistic disorder and normal levels of narcissism. Rather, narcissistic traits are
viewed as existing along a dimension rather than being a categorical distinction.
Kemberg (1984) has conceptualized narcissism along a continutun of normal to
pathological; the normal adult is able to regulate self-esteem and integrate object
representations. Normal infantile narcissism represents a stage of development in which
self-esteem regulation and goal attainment occur through processes which are appropriate
for this age. Finally, pathological narcissism involves either regression to this level of
development, which is a relatively mild type of pathology; narcissistic object choice, in
which the person projects their infantile self onto another object; or most severely,
narcissistic personality disorder, which involve self-love, object love, and superego which
are all of a pathological nature (Kemberg, 1998). This most severe level of pathology
involves “infantile values,” such as an emphasis on external status symbols such as
money or physical attractiveness. It includes a feeling of excessive self-centeredness and
superiority, as well as grandiosity and an “inordinate dependency on external admiration”
(p. 36). Kernberg (1998) states that individuals with narcissistic traits fall along the
above continuum from nearly normal to borderline functioning.

A somewhat different view of narcissism was taken by Kohut (1966), working
from the orientation of self-psychology. Kohut saw pathological narcissism as an arrest
in the normal development of the libido, in which the individual does not integrate the
idealized parent and grandiose self personality structures. These structures eventually

become split off, and the narcissistic person seeks to firlfill, unconsciously, the need for

14

grandiosity and a perfect other. Therefore, these individuals are prone to disappointment
when they discover that their view of the other has been shattered by realistic interactions.

Millon (1969) developed a social learning approach to understanding the
development of narcissism. In this model, one or both parents provide the child with a
grandiose sense of self-worth through both modeling behavior and direct feedback. This
level of self-worth becomes internalized but is not confirmed by external reality, beyond
the reactions of the parent(s). Drawing upon some of these ideas, Beck, Freeman, and
Associates (1990) describe the cognitive theory of narcissism, which involves a distorted
view of self, world, and future. This view is developed, as in Millon’s model, through
interactions with significant others who provide skewed feedback about the person being
more special, important, and unique than others. Notably, this can also involve negative
treatment, such as, being the target of abuse by a parent who does not abuse the other
siblings. Over time, the individual comes to believe they are different from others, and
they do not get disconfirming feedback. Beck et a1. (1990) explain that this may
“contribute to the hypersensitivity to evaluation so common among narcissists” (pp. 23 8-
239).

Although theoretical differences exist between these models, all seem to have as a
major component an emphasis on dysregulation of self-esteem; that is, failure to take a
realistic view of the self and one’s abilities, leading to a skewed perception of one’s self-
worth. This in turn leads to maladaptive interactions with other people.

It may be clarifying at this point to illustrate the concepts of psychopathy and
narcissism as personality traits which fall along a continuum, resulting in a disorder in the

extreme. While there is no clear point at which these traits obviously become a disorder,

15

the chart below was devised to show the criteria that are often used when making this
determination. The proposed study aims to assess traits among people who would fall
under the heading of “normal sample,” and can be measured using instruments such as
the Psychopathic Personality Inventory and Narcissistic Personality Inventory.
Individuals who meet criteria for a disorder are best assessed through use of the DSM-IV,

PCL-R or PCL:SV.

Table l: Subclinical personality Laits vs. disorder

 

Normal Sample Disorder
Traits including Score above 30 on PCUPCL-R
egocentricity, superficial Score above 18 on PCL:SV

Psychopathy charm, fearlessness, lack of
empathy or remorse, poor No DSM-IV diagnostic

 

 

 

 

 

igmlse control category

Traits including vanity, Narcissistic Personality
Narcissism grandiosity, attention-seeking Disorder

behavior, feelings 0f (must meet at least 5 of 9

supenorrty DSM-IV criteria)

 

It is appropriate at this point to further explore the construct of self-esteem, a key

aspect of both narcissistic and psychopathic personality traits.

Selffleem: Definitions_ and Assofited Behaviors
Self-esteem is a concept that is as difficult to define as it is integral to the study of
individuals’ psychological functioning. Mruk (1995) eloquently speaks to the heart of the

dilemma in studying self esteem:

l6

“. . .it makes achieving a respectable degree of scientific validity very

difficult. But another way of looking at this situation is to see the

complexity as a necessary evil in self-esteem research: it simply reflects

the fact the self-related phenomena coexist in an intricate,

multidimensional, interlocking network of structures that cannot be

completely unraveled. Although this condition is not an excuse for poor

research or analysis, it does set the parameters under which we must work

and, therefore, limits the kind of certainty we can realistically expect” (p.

26).

Given this situation in current research on self-esteem, Mruk argues that rather than
searching for an all-encompassing definition for which a universal measure could be
created, a more reasonable goal is to simply make clear what definition of self-esteem is
to be used. This necessitates a choice about what aspects of self-esteem, fiom its myriad
potential components, will be measured. A major contribution to the area of self-esteem
research was made by Rosenberg (1965), in his description of self-esteem as making
cognitive, attitudinal judgments about the self, as well having an evaluative dimension
which involves approval or disapproval of the self. He developed a brief inventory which
assesses self-esteem according to this definition, and is widely used today.

A point that further clarifies this definition was made by Steffenhagen (1987),
who argued that self-esteem and ego strength are often treated as equivalent concepts in
psychological literature, while they are more accurately viewed as separate constructs.
What makes this distinction important is his notion of ego strength as an individual’s
functional orientation to reality, including feelings of adequacy, as well as the ability to
be flexible and spontaneous. As mentioned earlier, self-esteem refers to one’s overall

estimation of the self. According to this view, then, one may have poor ego strength, but

relatively high self esteem as long as reality orientation is poor but the self-view is

17

positive. This would seem to be the case among individuals with psychopathic and
narcissistic traits.

In recent years, Baurneister has studied self-esteem extensively, concurring with
Mruk’s (1995) view that people tend to have a consistent baseline level of self-esteem,
which may vary somewhat at different points in time, as well as along different domains
(Baurneister, 1998). Measures of global self-esteem have been used most often in
research and have produced more notable results than measures of self-esteem in various
domains of functioning. In general, studies have found that people with high self-esteem
are happier and in better health than those with low self-esteem, and tend to persist longer
when faced with failure. However, Baurneister (1998) points to some potential pitfalls of
high self-esteem. People with high self-esteem may be more prone to interpersonal
violence, particularly in response to an ego threat. They may also respond in self-
destructive ways when faced with an ego threat, a possibility which was investigated in
the laboratory study described below, which measured the personal financial costs of
overinvestrnent in one’s self-esteem.

Baurneister, Heatherton, and Tice (1993) investigated the tendency of people with
high self-esteem to overstate their abilities and to predict success for themselves. They
predicted that this tendency, while often adaptive, may lead to failure when individuals
are presented with an ego threat, reasoning, “...people with high self-esteem should
outperform people with low self-esteem and self-management under normal,
nonthreatening conditions, but an ego threat (or other esteem challenge) will tend to make
self-esteem become an influential factor in the decision process” (p. 143). If it is true that

people with high self-esteem become more invested in maintaining self-esteem when

18

faced with an ego threat, they may be more likely to exhibit poor judgment in decision-
making tasks than people with low self-esteem, or when not presented with an ego threat.

In the Baurneister et a1. (1993) study, participants practiced playing a video game
and were then allowed to wager money on the last trial of the series, with the amount of
the bet being determined by the participant. Results from this study showed that in cases
where participants with high self-esteem were faced with an ego threat by the
experimenter (suggesting that the participants might be the type to “choke under
pressure”), they were likely to lose more money than when not faced with the threat. This
was because they bet larger amounts of money and then were unsuccessful on the critical
trial.

Baurneister (1996) discussed two main strategies that people use to minimize the
effects of an ego threat. One method is to avoid attending to the threat, so that it is not
processed, or is only minimally processed. The second technique, a refutational defense,
involves interpreting information that cannot be avoided in order to place it in a positive
light for the individual. The extent to which these strategies are effective, then, may be a
determining factor in the extent to which ego threats impact people with high self-esteem.

In addition to difficulties with self-regulation that impact only themselves, people
with high self-esteem may respond to ego threats by experiencing negative affect and
attacking others. Bushman and Baurneister (1998) designed a study to investigate whether
high or low self-esteem would be correlated with aggression toward others. They
predicted that, “In particular, inflated, grandiose, or unjustified favorable views of self
should be most prone to causing aggression, because they will encounter the most threats

and be chronically most intolerant of them” (p. 220). They propose that narcissism,

l9

which would entail the sort of self-view described above, may be the strong need to
believe that one is superior, whereas true high self-esteem would mean that a person truly
thinks well of himself or herself. Since narcissism could be a trait of people with either
high or low self-esteem, this is a distinction that would make it possible to explain
differences in aggressive behavior theoretically, among both high and low self-esteem
people.

Bushman and Baurneister (1998) predicted that people who scored high on the
Narcissistic Personality Inventory would produce the highest rates of aggression when
presented with an ego threat in the form of a negative evaluation of an essay they wrote.
Aggression in this study was assessed by the duration and intensity of a blast of noise the
participant would deliver to the person who they believed had negatively evaluated them.
Their prediction was confirmed in two related studies, and they further discovered that
self-esteem ratings were not related to aggression, through either main effects or an
interaction.

Kernis, Grannernann, and Barclay (1989) investigated how participants’ overall
levels and stability of self-esteem were predictive of their self-reported experiences of
hostility and anger. They defined self-esteem as an overall self-view, and stability of self-
esteern as the amount and intensity of change in this evaluation within a period of time.
This study assessed changes in level of self-esteem by sampling individuals’ feelings 40
times over a one-week time span using a pager system called the Experience Sampling
Method. Participants were given a list of 20 positive and 20 negative self-feelings, such
as “confiden ” or “unsure,” and asked to circle whatever words described their current

feelings at the time they were paged. The score for each event was computed by

20

subtracting the number of negative words circled fi‘om positive words circled, and
dividing by the total number of circled words. The level of self-feelings was defined as
the average of these 40 individual scores, and the standard deviation was computed and
used as the primary measure of stability. The authors predicted that individuals with
unstable self-esteem would have more negative reactions to ego threats, in terms of angry
feelings, than people with more stable self-esteem. As in the Bushman and Baurneister
(1998) study, these researchers sought to elucidate the mechanism which may account for
why individuals with both high and low self esteem may become aggressive.

The Kernis et a1. (1989) study hypothesized that individuals with high, unstable
self-esteem would be most prone to experience and express anger, and those with stable
high self-esteem would be least likely to respond in this way. The authors hypothesized
that individuals with low self-esteem would respond at a level between these two groups.
They argued that stability of self-esteem would be less of an issue for people with low
self-esteem, due to higher levels of depression and increased likelihood of inhibiting
angry responses among this group. One week after the last self-esteem score was
collected, participants filled out measures of anger and hostility, including Novaco’s
Anger Inventory (N ovaco, 1975), the Trait Anger Scale (Spielberger, Jacobs, Russell, &
Crane, 1983) and the Buss-Durkee Hostility Inventory (Buss & Durkee, 1957).

Results from this study were in line with predictions. Individuals with high,
unstable self-esteem endorsed the highest levels of anger and one component of hostility;
those with high, stable self-esteem reported the lowest levels. Participants with low self-
esteem had scores between these two groups. Kemis et a1. (1989) explain these results in

terms of high sensitivity to evaluative feedback among the unstable high self-esteem

21

group. Among those with low self-esteem, they suggest that negative feedback may
confirm an already negative self-view and will thus have less impact upon feelings of
anger. The authors add that they would expect to see even more significant correlations
between self-esteem stability and angry feelings if they assessed anger aroused
specifically in reaction to threats to self-esteem. In addition, they call for further
investigation of the role of personality factors in anger-proneness. To understand the

usefulness of measuring anger-proneness, it is necessary first to define what anger is.

Definition and Assessment of Ange_r

Spielberger, Reheiser, and Sydernan (1995) discuss definitional differences
between anger, hostility, and aggression. According to these researchers, anger describes
an emotional state which may contain a range of feelings, fi'om low levels such as
irritation, to high levels such as rage. It includes physiological factors, such as increased
heart rate or body temperature, an affective component (the subjective feeling of being
angry), behavioral aspects, such as clenched fists, and a cognitive component, which
involves thoughts related to intending harm towards another person. Hostility may
include the feeling of anger, but is also associated with attitudes that can lead someone to
want to hurt objects or people. Aggression, then, is the actual behavior involved in acting
on these feelings and attitudes. In constructing the State-Trait Anger Scale (STAS)
Spielberger (1980) developed two subscales relevant to the intensity of the subjective
experience of anger (State subscale), as well as the frequency of the experience of anger
(Trait subscale). Anger itself may be subdivided further, as Spielberger (1988) has done

in constructing the State-Trait Anger Expression Inventory (STAXI). This measure

22

contains subscales which tap factors such as Anger-In, Anger-Out, and Anger-Control, in

addition to State (immediate) and Trait (chronic) anger.

Hymtheses

For each of the four hypotheses presented below, a brief summary of the relevant
research findings that guided each prediction is provided.

Research on personality correlates suggest that trait anger may be associated with
psychopathy. Widiger and Lynam (1998) found that psychopathy was correlated with
low agreeableness, or antagonism, as measured by the Five Factor Model (F FM) of
personality functioning. Hart and Hare (1994) found strong negative correlations between
the warmth-agreeableness factor, and scores on the PCL-SV for both students and
inmates. Harpur, Hart, and Hare (1994) indicate that high scorers on the PCL are those
with high levels of hostility, antagonism, and aggressiveness. Previous research has also
shown correlations between narcissism and anger. McCann and Biaggio (1989) found
that, in an undergraduate sample, participants with high levels of narcissism were more
verbally expressive of anger than individuals low in narcissism. Rhodewalt and Morf
(1995) found that higher levels of narcissism as assessed by the NPI were associated with
the traits of hostility and antagonism.

Hyppthesis l: Trait anger will be positively correlated with both psychopathic
and narcissistic traits, for participants in both the experimental and control groups. The

possibility of an unintended effect of group assignment will be tested.

23

A body of research literature exists linking narcissism and high self-esteem to
anger and aggression (Baumeister, Heatherton, & Tice; Bushman & Baurneister, 1998;
Kernis, Grannernann, & Barclay, 1989; Rhodewalt & Morf, 1998), in contrast to the
commonly held belief that high self-esteem is a protective factor against negative affect.
Specifically, these studies found that individuals with high narcissism and high
(particularly unstable) self-esteem, evidenced a tendency to respond with anger and/or
aggression to a perceived failure or negative interpersonal feedback.

Hymthesis 2a: High narcissism on the NPI and high self-esteem on the RSE are
expected to be associated with high levels of state anger on the STAXI-Z following the
ego-threat manipulation. Low narcissism and high self-esteem will be associated with low

levels of state anger.

The psychopathy literature indicates a general shallowness of emotional
experience and interpersonal response on the part of psychopaths (Hare, 1993; Reise &
Oliver, 1994) as well as low levels of neuroticism (W idiger & Lynam, 1998). These
findings indicate that, unlike narcissists, psychopaths are not easily “rattled” by criticism
from others and would thus be less susceptible to the type of ego threat present in the
current study.

Hypgthesis 2b: High psychopathy on the PPI and high self-esteem on the RSE
are expected to be associated with low levels of state anger on the STAXI-2 following the
manipulation. Low psychopathy on the PP] and high self-esteem on the RSE are
expected to be associated with low levels of anger. Low self-esteem is expected to be

associated with high levels of anger for individuals at all levels of psychopathic traits.

24

These effects are predicted only for the experimental group, as the vignettes are not
intended to produce an angry reaction among control subjects. Pilot testing will be done
to provide support for this idea.

This hypothesis predicts a negative correlation between self-esteem and state
anger. For theoretical reasons outlined above, psychopathy is not expected to interact with
self-esteem to affect state anger, in contrast to the predicted interaction between
narcissism and self-esteem. This expected difference is a key component of the study.
The prediction of no interaction is somewhat controversial, because a lack of interaction
may be due to a number of reasons, while statistically significant findings support the
presence of a true difference. In developing this hypothesis, it was necessary to determine
whether adequate statistical power was present to detect a difference if one did exist.
Group size was determined based upon the statistical tables created by Cohen (1992) in
his article regarding power analysis. This article states that an ANOVA will be able to
detect a large effect at the alpha = .05 level if there are at least 18 participants per group.
The design of the current study meets this criterion; therefore, the prediction of no

interaction is supported by the study methodology.

Gender differences have been found which consistently show males to have
higher levels of psychopathy (Forth, Brown, Hart, & Hare, 1996; Lilienfeld & Andrews,
1996; Meloy, 1992; Salekin, Rogers, & Sewell, 1997), narcissism (Watson, Taylor, and
Morris, 1987), and self-esteem (Kendler, Gardner, & Prescott; Kling, Hyde, Showers, &
Buswell). However, it was unclear to what extent this would hold true in a nonclinical,

college population.

25

Hypothesis 3: Overall, females are expected to have lower levels of self-esteem
(RSE score), narcissistic traits (NPI score), and psychopathic traits (PPI score) than

males.

Previous investigations into ego-threat manipulations have used a variety of
methods, including the suggestion of upcoming failure on a task (Baurneister, Heatherton,
& Tice, 1993), and negative evaluation of an essay written by the participant (Bushman &
Baurneister, 1998). The current study was designed to test and implement the use of
vignettes as a way of administering an imaginary ego threat. This was done in order to
minimize any long-term negative affect resulting fiom the manipulation, as well as to tap
into various domains of self-esteem. This would not be possible given the other methods
reported in the literature, which tend to focus on one area of performance (e.g., score on a
video game, writing ability).

Hymthesis 4: Participants in the control group will endorse lower levels of state
anger as measured by the STAXI-Z after receiving the neutral manipulation than will

participants in the experimental group, who will receive the ego-threat manipulation.

26

METHOD

Participants

One hundred sixty-two undergraduate participants were self-selected fi'om an
introductory psychology class at a large Midwestern university. All students read the title
of the experiment, “Personality Traits and Reactions to Life Events” among the on-line
list of experiments, and signed up to participate in a group administration session. They
received course credit for their psychology class. The sample included 106 females and
56 males, ranging fiom 18 to 24 years of age (M = 19.31; SD = 1.37). All participants
were single. Participants self-reported their racial background as follows: Caucasian,
80.9%; Afiican-American, 7.4%, Asian, 4.9%, Hispanic, 4.3%; Native American, 1.2%;

multiracial, 1.2%.

Measures

Demoggphic Information

A brief questionnaire was given to obtain demographic information for
participants, including age, race, marital status, native language, and occupation.

Psychopathic Personality Inventog (PPI; Lilienfeld & Andrews, 1996). The PPI
was developed as a means of assessing the presence of psychopathic traits in noncriminal
populations, and was validated using a sample of undergraduates at colleges in
Minnesota. Validation was conducted through three rounds of item writing, distribution
and data collection, and item analysis. The first two rounds of data collection were done
exclusively on males in order to maximize the expected variance on psychopathic traits

endorsed. After eliminating items which had high social desirability and those which

27

assessed overt antisocial behaviors, the items were subjected to firrther factor analysis and
finally distributed to both males and females. At this point scores on the PPI were
correlated with existing measures of psychopathy (see below).

The PPI is a relatively new instrument and has demonstrated high positive
correlations with the Cleckley criteria for psychopathy, as determined by interview and
peer ratings. The l60-Item, true/false questionnaire also has the advantage of easy
administration due to the self-report format.

The measure contains two validity scales, the Deviant Responding (DR) scale,
designed to detect malingering and random response style, and the Variable Response
Inconsistency (VRIN) scale, which detects inconsistent responses to items with similar
content. The PPI consists of 8 subscales, including Machiavellian Egocentricity, Social
Potency, Coldheartedness, Carefree Nonplanfulness, Fearlessness, Blame Extemalization,
Impulsive Nonconforrnity, and Stress Immunity. The internal consistency of these
subscales computed by Cronbach’s alpha ranged from 0.70 to 0.90, with the PPI total
score having an internal consistency of 0.90 to 0.93. Test-retest reliability ranged fiom
0.82 to 0.94 for the subscales; PPI total score was very high, at 0.95. Finally, the PPI was
found to correlate significantly with existing self-report measures of psychopathy,
including the Levenson Psychopathy Scale (r = 0.37), the Self-Report Psychopathy Scale-
Revised (r = 0.91) and the MMPI ASP scale (r = 0.58).

Narcissistic Personalng Inventog (N PI; Raskin & Hall, 1979). This 40-ltem scale
is presented in a true-false format, and contains seven subscales: Superiority, Authority,
Exhibitionism, Self-Sufficiency, Vanity, Exploitativeness, and Entitlement. The NPI was

designed to assess the extent to which a person reports having narcissistic personality

28

traits, such as believing they are superior to other people and possess extraordinary
capabilities. Thus, it was not designed to assess narcissistic personality disorder per se,
and is appropriate for use in nonclinical populations.

Raskin and Hall (1979) found the split-half reliability of the original 80—item scale
to be 0.80. In 1981, two alternate forms were developed, each containing 40 items.
When forrrrs A and B were administered eight weeks apart to a sample of college
undergraduates, the reliability coefficient was 0.72, indicating a stable construct
underlying this measure (Raskin & Hall, 1981). However, the authors note that this
general component is not strong: “. . .it is worth bearing in mind that both theoretically
and clinically, narcissism represents a syndrome of relatively diverse behaviors that
would not predict for a particularly strong or overriding general factor” (p. 899). The
internal consistency of the NPI was found to be 0.83, and the measure was essentially
uncorrelated with age or gender (Raskin & Terry, 1988).

Rosenberg’s Self-Esteem Seals (RSES; Rosenberg, 1965). This instrument was
deve10ped to provide a measure of individuals’ global self-evaluations. Rosenberg’s self-
esteern scale is widely used in research on self-esteem and has been well-validated in this
domain. Fleming and Courtney (1984) and Gray-Little, Williams, and Hancock (1997)
both reported an internal consistency of alpha = 0.88 using different samples of college
students. Test-retest reliability has been found to range from r = 0.50 on a 1-year retest
(McCarthy & Hoge, 1984) to r = 0.82 over a one-week period Fleming and Courtney,
1984). A factor analysis by Gray-Little et a1. (1997) determined that a single-factor
solution to the RSE was the best fit, supporting Rosenberg’s original contention (1965)

that his scale was a unidirnensional measure of a person’s perceived self-worth.

29

State-Trait Anger Expression Inventory. Revised (STAXI-Z; Spielberger, 1998).
The STAXI-Z is a 57-Item scale consisting of six subscales relevant to the intensity of the
subjective experience of anger (State subscale), frequency of the experience of anger
(Trait subscale), and subscales assessing the ability to control one’s anger. The main
subscale of interest in the present study is State Anger, which is defined by the test
developer as subjective feelings that vary in intensity from irritation to intense firry, and
are accompanied by autonomic nervous system arousal. In addition, the author states, “It
was assumed that S-Anger would fluctuate over time as a fimction of frustration,
perceived affi‘onts, injustice, or being verbally or physically attacked” (Spielberger,
Reheiser, & Sydernan, 1995; p. 213). This subscale makes it particularly appropriate as
an operationalization of the construct of interest in the present study.

Vigr_1ettes

Participants read a series of four vignettes, written by the author of the study, in
which they were asked to picture themselves in a variety of situations (see Appendix A
for full texts of the vignettes). For the experimental group, these scenarios involved an
imagined ego threat, including receiving a poor grade in an important class, being
rejected by a potential dating partner, being denied a coveted job offer, and being told that
their family disapproves of them. Participants in the control group read similar scenarios,
but those in which the outcome was neutral or unresolved (e.g., waiting to hear about the
final course grade or the results of a job application).

Mood Scales

This is a questionnaire in the form of Likert scales, developed by the author (see

Appendix B). The individual is asked to think about how they are feeling at that moment,

30

and to rate a series of mood words (e.g., happy, sad, embarrassed) on a scale from 1 to 5,
with “1” indicating “not at all” and “5” indicating “extremely.” These scales were
included in order to reduce the salience of experimental constructs assessed by the anger
inventory, but were not used in any of the analyses.

Procedure

Before the full data collection began, pilot testing was done on 16 participants
from an introductory psychology class (13 females and 3 males); half in the control and
half in the experimental group. They were asked to read the vignettes, write their
responses, and then fill out the mood questionnaires. This was done to determine
whether the vignettes were effective at eliciting different types of responses between the
experimental and control scenarios.

The state anger subscale on the STAXI-2 was used as one dependent measure of
the effectiveness of the vignettes. As expected, the score for the experimental group (M =
22.25, SD = 7.87) was higher than that of the control group (M = 17.38, S_I2 = 4.53),
although this difference was not significant (p < .15). Inspection of the score distribution
revealed one participant in the control group with a score of 28, falling more than 2
standard deviations above the group mean. When this outlier was eliminated from the
analysis, group differences closely approached significance E(1, 13) = 4.424, p_< .06.
Significant differences (p < .05) between groups were observed for several items on the
mood questionnaires. Experimental participants reported feeling more annoyed,
depressed, and tense than controls, and controls reported feeling more proud, happy,
confident, relaxed, and lighthearted than those in the experimental group. These results

suggested that reading the vignettes did tend to induce a more angry and less positive

31

feeling among the experimental group. Therefore, it was deterrrrined that the vignettes
were effective and could be used for their intended purpose in the study.

After pilot testing was completed, the full data collection began. Participants came
to the administration room, which was a large classroom with desks, in groups ranging in
size from 1 to 27. Over the three-month data collection period, 162 participants were
assigned to receive either the experimental manipulation or to be a member of the control
group, resulting in 81 participants in each group. A packet was given to each participant,
with either control or experimental group vignettes inside. Every second person received
a packet with control vignettes. Control vignettes consisted of stories with neutral
outcomes; experimental vignettes contained negative outcomes. Included in the packet
was a consent form, demographic questionnaire, PPI, NPI, RSE, 4 vignettes, STAXI-2,
and mood scales. Each packet contained the same order of questionnaires and vignettes.

Participants were given an overview of the study procedures, and read and signed
an informed consent form. They were told that the experimenter was interested in the
ways that people describe themselves, their ability to imagine themselves in a variety of
interpersonal situations, and the ways that people respond in these situations. They
received the following instructions:

“In this experiment you will be asked to complete a set of three questionnaires
about your personality. The instructions for each questionnaire are on the front page.
Please take your time and answer each question honestly. You may skip any questions
that you do not wish to answer. After this, turn to the first story in your packet. Read the
story and then write a paragraph describing how you would feel and what you would do

in that situation. You will have as long as you want to write your response. Then go on

32

and do the same for the other three stories. After you are done, please complete the two
questionnaires at the back of the packet. The instructions are on the top of each page. If
you have any questions while you are working, please don ’t hesitate to ask the
experimenter. Be sure that you answer all of the questions truthfidly—-there are validity
scales built into the questionnaires. Please keep the pages in your packet in the same
order they are now. You may begin whenever you are ready. ”

All participants first completed three questionnaires: the Psychopathic Personality
Inventory (PPI), Narcissistic Personality Inventory (N PI), and Rosenberg’s Self-Esteem
Scale (RSE). The PPI contains 160 items that are in the form of statements about one’s
personality traits, habits, and preferences. For each item; the individual must indicate
whether the statement is false, mostly false, mostly true, or true, recording their answers
on a scale fiorn one to four. The NPI has 40 items, and is similar to the PM in that each
item is a statement about a personality trait to which the individual must respond with
“true” or “false” as it applies to him or her. The RSE contains 10 items assessing the
domain of self-worth, which are endorsed on a scale of l to 5, with 1 corresponding to
“strongly agree” and 5 corresponding to “strongly disagree.” In order to increase
participants’ investment in responding honestly to the personality measures, they were
told that the questionnaires have built-in validity scales, and that they should answer
truthfully.

Following completion of the three questionnaires, participants read the four
vignettes. After reading each vignette, participants wrote a short paragraph on the page
beneath the story, describing what they would do and how they would feel in each

situation. This was intended to increase participants’ focus on the situations they read

33

about. At this point, participants completed a brief set of Likert scales asking about their
mood, as well as the State-Trait Anger Expression Inventory - Revised. Individuals
completing this self-report measure are asked to read each item (e.g., I feel furious) and
rate their immediate feelings on a scale of 1 to 4, with 1 corresponding to “not at all” and
4 corresponding to “very much so.” The additional mood questions/Likert scales were
given to participants in order to reduce the salience of experimental constructs and
variables of interest to the experimenter.

After the measures were completed, participants filled out a final questionnaire in
which they were asked whether they were able to guess the purpose of the study, whether
they attempted to imagine themselves in the situations and answered the questionnaires
honestly, and if reading the stories had any effect on their mood. It was planned that data
from individuals who successfully guessed the study’s purpose or did not complete the
measures truthfully would not be included in the analyses. A check of participants’ post-
experiment questionnaires as well as PPI validity scales revealed that no data had to be

eliminated due to knowledge of the hypotheses or improper responding to the measures.

34

RESULTS

Hyppthesis 1: Trait anger was expected to be positively correlated with both
psychopathic and narcissistic traits. This was expected to be true for participants in both
the experimental and control groups.

Although not part of the original hypothesis, post-hoe analyses for Hypothesis 4
revealed a significant gender by group interaction; therefore, it was deemed necessary to
assess any effects of gender on trait anger for Hypothesis 1. A 2 (group) X 2 (gender)
AN OVA was conducted to assess for equality of the outcome measure, trait anger.
Results showed no main effects of group or gender on trait anger. The interaction was
not significant. Because there were no significant differences between the experimental
and control groups, or males and females, on the outcome variable of trait anger, further

analyses collapsed data across these groups.

Table 2. Correlations between persorprlity traits and trait anger

PPITOTAL NPITOTAL STAXI-T

PPI Total Score 1.000
NPI Total Score .533" 1.000
STAXI-T .222" . 130* l .000

** p < 0.01, one-tailed
* p < 0.05, one-tailed

Results of Pearson correlations are presented in Table 2. The correlation between

PPI and trait anger was significant, indicating a positive relationship between these

35

variables. The correlation between NPI and trait anger was also statistically significant,
but the correlation was not as strong. Thus, the prediction of a positive relationship with

trait anger was true for both psychopathic and narcissistic personality traits.

Hymthesis 2a. High narcissism on the NPI and high self-esteem on the RSE were
expected to be associated with high levels of state anger on the STAXI-2 following the
ego-threat manipulation. Low narcissism and high self-esteem were expected to be
associated with low levels of state anger. This hypothesis was not supported.

To test this hypothesis, a median split was performed for scores on both the NPI
(median score = 62) and the RSE (median score = 41). State anger scores for the four
groups based on the median split are presented in Table 3. A 2 (low vs. high NPI) X 2
(low vs. high RSE) ANOVA was conducted to assess for equality of the outcome
measure (state anger). Results showed no main effects of NPI score or RSE score on
state anger. The interaction was not significant.

Table 3. State anger scores for low vs. highNPI and RSE (means and standard

 

 

 

deviations)
Low NPI High NPl
Low RSE 16.482 (2.63 7) 20.938 (12.162)
(n=27) (n=16)
High RSE 16.824 (4.035) 17.450 (4.662)
(n=l7) (n=20)

 

 

 

 

 

Hypothesis 2b. High psychopathy on the PPI and high self-esteem on the RSE was
expected to be associated with low levels of state anger on the STAXI-2 following the

manipulation. Low psychopathy on the PPI and high self-esteem on the RSE were

36

expected to be associated with low levels of anger. Low self-esteem was expected to be
associated with high levels of anger for individuals at all levels of psychopathic traits.
These effects were predicted only for the experimental group, as the vignettes were not
intended to produce an angry reaction among control subjects.

To test this hypothesis, a median split was performed for scores on the PPI
(median score = 367) and the RSE (median score = 41). State anger scores for the four
groups based on the median split are presented in Table 4. A 2 (low vs. high PPI) X 2
(low vs. high RSE) ANOVA was conducted to assess for equality of the outcome
measure (state anger). Results showed no main effects of PPI score or RSE score on

state anger. As expected, the interaction was not significant.

Table 4. State anger scores for low vs. high PPI aad RSE (means and standard deviations)

 

 

 

 

 

 

 

Low PPI High PPI
Low RSE 17.958 (6.410) 18.368 (9.581)
(n=24) (n=l9)
High RSE 17.000 (4.648) 17.286 (4.200)
(n=16) (n=21)

 

Hymthesis 3: Overall, females were expected to have lower levels of self-esteem,
narcissism, and psychopathic personality traits than males.

An independent samples t-test was conducted to compare the mean scores for
males and females on the Rosenberg Self Esteem Inventory (RSE). Scores on this

measure range from 0 to 50, with higher numbers indicating higher self-esteem. The

37

mean scores for females on the RSE (M = 40.08, S_D = 7.35) were not significantly
different than mean scores for males on this measure (M = 40.69, SD = 8.33). This
indicates similar levels of self-esteem among males and females, and does not support the
prediction that female levels of self-esteem would be lower.

On the Psychopathic Personality Inventory (PPI), an independent samples t-test
showed that the mean total score for females (M = 352.40, SD = 36.75) was significantly
lower than the mean total score for males (M = 393.32, S_D = 30.00), t(160) = -7.16, p <
0.0005, in line with the prediction that females would show lower overall levels of
psychopathy than males. Further analysis of gender differences was then performed for
all subscales of the inventory (see Table 5). A correction was made for multiple
comparisons; an alpha of .05 was divided by 8, so that the level of significance was set at
0.00625. Levene’s test was performed for equality of variance between scores for males
and females, and for the subscales on which variance for males and females differed
(Social Potency, Coldheartedness), analyses are reported with equal variance not
assumed. Significant differences in subscale scores were found on the subscales of
Machiavellian Egocentricity, F earlessness, Coldheartedness, Impulsive Nonconforrnity,
and Stress Immunity. In all of these cases, the total scores for males were higher than

those of females.

38

Table 5. Comparison of males and females on PPI subscale totala

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Subscale Mean Mean t df Mean Difference
(female) (male)

Machiavellian 64.132 73.179 -4.721 160 -9.047

Egocentricity“I

Social Potency 64.462 67.143 -1.681 139 -2.681

Fearlessness“ 44.330 54.375 -6.176 160 -10.045

Coldheartedness“ 38.094 42.339 -2.812 83 -4.245

lrnpulsive 35.453 40.929 -4.232 160 -5.476

Nonconforrnity"

Blame Extemalization 35.925 39.339 -2.413 160 -3.415

Carefiee Normflanfulness 35.038 35.696 -0.626 160 -0.659

Stress Immunity" 27.076 31.054 -3.932 160 -3.978

" p < .0005

* p < .005

On the Narcissistic Personality Inventory (NPI), an independent samples t-test

revealed that the mean total score for females (M = 60.66, E = 6.01) was significantly

lower than the mean total score for males (M = 65.00, SD = 6.01), t(160) = -4.37, p <

0.0005, in line with the prediction that females would show lower overall levels of

narcissism than males. Further analysis was conducted to assess gender differences on

subscales of the measure (see Table 6). A correction was made for multiple comparisons;

an alpha of .05 was divided by 7, so that the level of significance was set at 0.00714.

Levene’s test for equality of variance was performed between males and females, and for

the subscale on which variance for males and females differed (Authority), analyses are

reported with equal variance not assumed. Significant differences were found on the

subscales of Authority, Exploitativeness, and Entitlement, with males showing higher

scores than females.

39

 

Table 6. Comparison of males and females on NPI subscale tota_ls

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Subscale Mean Mean t (if Mean Difference
(female) Qnale)

Authority" 12.528 14.500 -6.619 133 -1.972
Self-Sufficiency 9.859 10.375 -2.585 160 -0.517
Superiority 8.481 8.661 -1.033 160 -0.180
Exhibitionism 9.359 9.500 -0.496 160 -0. 142
Exploitativeness" 7.104 7.964 -3.982 160 -0.861

Vanity 4.236 4.054 1.013 160 0.182
Entitlement" 9.094 9.946 -3.659 160 -0.852

** p < .0005

Hyppthesis 4: Participants in the control group were expected to endorse lower levels of
state anger after reading the vignettes than participants in the experimental group. A t-
test was conducted to compare the State-Anger subscale mean scores for control subjects
to mean scores for experimental subjects. Results revealed no significant difference
between subscale scores for control subjects (M = 17.64, S_D = 6.04) and those for
experimental subjects (M = 17.68, S_D = 6.42). A post-hoe analysis of clusters within the
subscale also did not reveal any significant differences between the groups in the clusters
of Feelings, Verbal, or Physical State Anger.

Further analysis was performed to determine whether the combined effects of
group assignment and gender were related to the outcome variable of state anger. An
ANOVA showed that a gender X group interaction did achieve statistical significance,
E(1, 158) = 9.931, p< .005). Follow-up contrasts were performed to test for simple
effects. Results (see Figure 2) showed that females in the control group (M = 16.33, S_D =

2.20) were significantly less angry as measured by the STAXI-2 state anger subscale than

40

 

males in the control group (M = 20.26, S_D = 9.57), E(1, 79) = 8.305, p< 0.005. Females
in the experimental group (M = 18.54, SQ = 7.73) did not have significantly different
state anger score after the manipulation than males in the experimental group (M = 16.14,
S_D = 2.23). Females in the control group had significantly lower state anger scores than
females in the experimental group, E(l , 104) = 4.054, p < .05. Males in the control group
had significantly higher state anger scores than males in the experimental group, E(1, 54)
= 5.084, p<.05. Thus, the manipulation affected each gender differently. Specifically,
undergoing the manipulation was associated with higher levels of state anger among
females, while males who received the manipulation showed lower levels of state anger

than males who did not.

Figufi 1. State anger scale scores

25

2°: ><j .

15 f w , , *
l. -_ q p ‘1 +Males
10 ’ 1 LI—Females

 

 

 

 

 

 

 

 

 

o '- - -7 - II. “L; I. .w.’. Imam“. M.
Control Experimental

 

41

DISCUSSION

This study was designed to further our understanding of links between personality,
anger, and aggression, by building on previous research related to self-esteem and
narcissism. In this study, psychopathic and narcissistic traits in a noncriminal sample
were assessed, along with their relations to anger experience in the face of ego threats.
Some theoretical overlap exists between these personality constructs (Reise & Oliver,
1994): both involve a grandiose sense of self as well as low empathy, and it appeared that
studying anger might further our understanding of the conceptualization of these clusters
of traits.

One aim of this study was to investigate the association between chronic (trait)
anger and psychopathic and narcissistic personality traits. In addition, while a body of
literature exists linking narcissism and high self-esteem to anger and aggression
(Baurneister, Heatherton, & Tice; Bushman & Baurneister, 1998; Kernis, Grannemann, &
Barclay, 1989; Rhodewalt & Morf, 1998), very little study has been made of the
interaction between high self-esteem, psychopathic traits, and anger. This study was
designed to investigate further how self-esteem might interact with personality traits to
produce anger in response to an attack on one’s self-esteem. The way in which
individuals respond to an attack was also seen as a potentially useful way to distinguish
the psychopathic and narcissistic syndromes.

Gender differences have been found that consistently show males to have higher
levels of psychopathy (Forth, Brown, Hart, & Hare, 1996; Lilienfeld & Andrews, 1996;

Meloy, 1992; Salekin, Rogers, & Sewell, 1997), narcissism (Watson, Taylor, and Morris,

42

1987), and self-esteem (Kendler, Gardner, & Prescott; Kling, Hyde, Showers, &
Buswell). This study aimed to determine whether this pattern held true with a
nonclinical, college undergraduate sample. Finally, the methodology of writing responses
to vignettes was used in conjunction with an anger inventory to determine the effects of

reading neutral or threatening vignettes.

Relationship of Trait Anger to Psychopagr and Nicissism

For all participants, a positive correlation was predicted between psychopathic
traits and trait anger. Specifically, higher scores on the PPI were expected to be
associated with higher levels of trait anger as measured by the STAXI-2 trait anger
subscale. Trait anger is not one of the core components of psychopathy as defined by ~
Hare (1993) and is not part of the operational definition of psychopathy as it is assessed
by measures in current use. However, research on personality correlates suggest that trait
anger may be associated with psychopathy. Widiger and Lynam (1998) found that
psychopathy was correlated with low agreeableness, or antagonism, as measured by the
Five Factor Model (FFM) of personality functioning. Hart and Hare (1994) found strong
negative correlations between the warmth-agreeableness factor, and scores on the PCL-
SV for both students and inmates. Harpur, Hart, and Hare (1994) indicate that high
scorers on the PCL are those with high levels of hostility, antagonism, and
aggressiveness. In addition, Hare (1991) found that psychopaths evidence poor
modulation of anger, hostility, and aggression.

This does not imply a causative relationship between anger and psychopathy such

that all expressions of anger are indicative of psychopathy. Anger may occur in response

43

to a wide variety of situations, and is experienced to some degree by all individuals
regardless of their personality type. However, given the above findings from the
personality literature, there is empirical evidence to suggest that the personality variables
associated with psychopathy are associated with hostility, aggressiveness, and
antagonism, and negatively related to warmth and agreeableness. While anger has not
generally been studied as separate from the variables mentioned above, there is reason to
expect that it would show similar correlations with psychopathic traits as do hostility,
disagreeableness, and antagonism.

It is important to note, however, that the construct of psychopathy involves
experiencing shallow emotions, and that psychopaths often have difficulty describing
their emotions verbally (Hare, 1993). Therefore, it was important to investigate whether
individuals with high levels of psychopathic traits would endorse higher levels of chronic
anger than those with lower levels of the trait. While the literature suggests an
association between psychopathy and anger, it was unclear whether this would hold true
for individuals with subclinical levels of psychopathy, and whether their feelings would
adequately be captured by a self-report measure.

Results indicate that this hypothesis was supported. Higher levels of psychopathic
traits were associated with higher levels of trait anger. This indicates that, at subclinical
levels of psychopathy, trait anger may be consistently positively related to psychopathic
traits. Trait anger was measured in the current study as the frequency of experiencing the
emotional, physiological, affective, and behavioral aspects of anger as defined by
Spielberger, Reheiser, and Sydernan (1995). As discussed above, characteristics such as

hostility, antagonism, and irritability have been linked to psychopathy in previous

research; the current findings suggest that anger may be a component deserving of firture
study.

It is also worthwhile to consider the issue of psychopathy on a continuum. The
DSM-IV (American Psychiatric Association, 1994) does not include a psychopathic
personality disorder; however, Hare (1998) conceptualizes psychopathy as a categorical
phenomenon. The current study has investigated psychopathy as a continuous construct,
which suggests the question of how much psychopathy is necessary for it to become
clinically meaningful. For example, when does a person feeling very good about himself
become grandiose; when does risk-taking behavior become impulsive? In addition, what
are the implications of having a small amount of psychopathy? There is no way to
definitively answer these questions; however, it may be useful to consider this in terms of
the negative consequences, both personal and interpersonal, of having these traits. That
is, even a low level of psychopathy may be harmful if it results in distress. Lower levels
of psychopathy are unlikely to result in criminal behavior, and it is among the
incarcerated population that psychopathy has been most studied. However, psychopathic
traits may still lead to aversive outcomes for oneself. For example, sensation-seeking and
poor behavioral controls may result in engaging in unsafe sex with many partners,
resulting in contracting a sexually transmitted disease. Low empathy toward others
combined with a feeling of grandiosity about the self make intimate relationships
frustrating and unbalanced. These traits may be present in varying degrees, with different
types of behavioral outcomes.

For all participants, a positive correlation was predicted between narcissistic traits

and trait anger. Specifically, higher scores on the NPI were expected to be associated

45

with higher levels of trait anger as measured by the STAXI-2 Trait Anger subscale.
Previous research has shown correlations between narcissism and anger. McCann and
Biaggio (1989) found that, in an undergraduate sample, participants with high levels of
narcissism were more verbally expressive of anger than individuals low in narcissism.
Rhodewalt and Morf (1995) found that higher levels of narcissism as assessed by the NPI
were associated with the traits of hostility and antagonism. Other research has found
individuals who were higher in narcissism to react more angrily to ego threat (Bushman
& Baurneister, 1998), and failure (Rhodewa1t& Morf, 1998). Although not part of
diagnostic criteria for Narcissistic Personality Disorder, the DSM-IV (American
Psychiatric Association, 1994) states that individuals with NPD are highly sensitive to
perceived criticism from others and are likely to react with feelings of rage. Given their
heightened sensitivity to negative interpersonal evaluations, it seems likely that
individuals with narcissistic traits would frequently encounter this type of perceived
attack and thus, often feel angry. In addition, because narcissistic individuals have
unduly high opinions of themselves, they may be more likely to irritate other people, thus
eliciting the very sort of negative feedback fi'om others which they find so aversive
(Papps & O’Carroll, 1998).

This hypothesis was supported. Higher levels of narcissistic traits were associated
with higher levels of trait anger. This finding is consistent with previous studies which
have showed a link between narcissism and the tendency to experience anger. Research
strongly suggests that this tendency toward experiencing anger is related to narcissistic
individuals’ characteristic responses to criticism. Specifically, the fiagile sense of self-

worth that is typical of narcissists makes them extremely sensitive to negative evaluations

46

from others, resulting in feelings of anger, whereas individuals whose self-esteem is more
stable would not show this frequent angry response.

Although coping styles were not assessed in the current study, it is useful to
consider whether the sensitivity to interpersonal criticism seen in narcissistic individuals
would apply equally to the concepts of “repressors” and “sensitizers.” Weinberger,
Schwartz, and Davidson ( 1979) studied the coping responses of individuals with low
anxiety and high defensiveness (repressors) and their counterparts with high anxiety and
low defensiveness (high anxious, previously described in the literature as “sensitizers”).
They suggest that repressors cope with disturbing cognitions through avoidance, while
hi gh-anxious individuals make no effort to avoid aversive thoughts. Although
speculative, it seems likely that narcissists would comprise a third category, individuals
with high anxiety and high defensiveness. Narcissists seek to convey an image of
unflappability, yet underneath they are quite concerned with the opinions of others.
Therefore, they may use a coping strategy with elements of both avoidance of unpleasant
self-referent thoughts, as well as heightened sensitivity to negative interpersonal
feedback.

Finally, a moderate correlation emerged between scores on the PPI and NPI. This
is not surprising, given the high degree of conceptual overlap between these clusters of
traits (Reise & Oliver, 1994), including elements of grandiosity, exploitativeness of
others, and low empathy. However, there are differences between these constructs as
well. Narcissism involves a need for excessive admiration, envy towards others, and
sensitivity to criticism, which are not characteristic of psychopathy. Psychopathy

includes a tendency toward early behavior problems and shallow emotional experience,

47

both of which differentiate this syndrome fi'om that of narcissism. While individuals with
both types of traits do not truly empathize or identify with the needs of others, narcissists
appear dependent on the responses of others to validate their sense of self-worth, and may
have very strong emotional reactions to interpersonal feedback. Psychopaths, by contrast,
are concerned with other people only to the extent that others can help them to get what
they want. Therefore, we would expect to observe a moderate correlation between the

two traits, as was found in the current study.

Narcissistic Traits, Self-esteem. and Ange_r

For participants receiving the ego-threat, high narcissism and high self-esteem
were expected to be associated with high levels of state anger following the ego threat.
Low narcissism and high self-esteem was expected to be associated with low levels of
state anger. Low self-esteem was expected to be associated with moderate levels of
anger. Specifically, higher scores on the NPI and higher scores on the RSE were
expected to be associated with higher scores on the STAXI-2 State Anger subscale.
Lower scores on the NPI and higher scores on the RSE were expected to be associated
with lower scores on the STAXI-2 State Anger subscale. All levels of NPI scores
combined with low RSE scores were predicted to be associated with moderate scores on
the STAX1-2 State Anger subscale.

This pattern was expected based upon previous research which has consistently
found a link between narcissism, high self-esteem, and vulnerability to negative
interpersonal feedback (Baurneister & Boden, 1998; Baurneister, Heatherton, & Tice,

1993; Bushman & Baumeister, 1998; Kernis, Grannemann, & Barclay, 1989; Papps &

48

O’Carroll, 1998). The purpose of the present study was in part to determine whether this
finding could be replicated using an imagined ego threat (reading vignettes) as a form of
negative interpersonal feedback. It was proposed that if results were in line with
predictions, this would suggest that individuals with narcissistic traits have a form of self-
esteern which is grandiose, but is fragile in the face of an ego threat, and is a risk factor
for felt anger. This hypothesis was not supported. Results indicated no main effects of
level of narcissistic traits or level of self-esteem on state anger following the ego threat.
Apparently, responses to the ego-threat were not significantly affected by participants’
levels of self-esteem or narcissistic personality traits.

While these negative results were unexpected, some reasons are proposed as to
why they occurred, including features of the vignettes, participants, and statistical
analyses. It is possible that the vignettes did not contain the level of negative
interpersonal feedback that would be required to provoke an angry response among
individuals with narcissistic traits. Results from pilot testing indicated that the vignettes
were effective at inducing feelings of anger. However, it may have been that the vignettes
created a generally negative mood state, but because the negative information in them did
not come from a real, identifiable other person, they failed to produce an angry response
from participants. Alternatively, participants may not have been motivated to concentrate
fully on the task. In this case, the vignettes may have shown more of an effect with
individuals who were more invested in engaging in the task.

It is also possible that, in the statistical analyses, using a median split was not
sufficient to identify groups that were sufficiently high or low on the variables of interest.

A larger sample size would have allowed for a quartile split which would have alleviated

49

this problem somewhat. Results from a linear regression model were also not significant,
however, suggesting that no effect was present. Another potential explanation is that the
RSE, being a brief, unifactorial measure of self-esteem, did not elicit enough variance in
responses to allow for meaningful differences between the high and low self-esteem
groups. Bushman and Baurneister (1998) suggest that the use of more complex measures

of self-esteem may be more efficacious for research of this type.

Psychopathic Traits, Self-Egeem. and Ange_r

Among participants receiving the ego threat, high self-esteem was expected to be
associated with low levels of state anger following the ego threat. Low self-esteem was
expected to be associated with higher levels of anger. These predictions were expected
for individuals at all levels of psychopathy. Put simply, self-esteem was predicted to be
negatively correlated with state anger, regardless of level of psychopathy. To date, there
are no published studies that have specifically investigated the role of psychopathic traits
as they relate to self-esteem and ego threat. The above prediction was expected based
upon the psychopathy literature which indicates a general shallowness of emotional
experience and interpersonal response on the part of psychopaths (Hare, 1993; Reise &
Oliver, 1994) as well as low levels of neuroticism (W idiger & Lynam, 1998). These
studies indicate that, unlike narcissists, psychopaths are not easily “rattled” by criticism
from others and thus would not be susceptible to the type of ego threat presented in the
current study. Therefore, an interaction between psychopathy and self-esteem would not

be expected.

50

This hypothesis was partially supported. As predicted, there was no significant
interaction between self-esteem and psychopathic traits on the outcome variable of state
anger. However, contrary to predictions, results indicated no main effect of level of self-
esteern on state anger following the ego threat. Participants with low self-esteem did not
differ in their levels of state anger from those with high self-esteem. As discussed earlier,
sufficient statistical power was present to detect an interaction between psychopathy and
self-esteem (large effect) if one was present. The lack of interaction, while not as
compelling statistically as an interaction effect, does provide support for the notion that
psychopathy and self-esteem did not produce combined effects on state anger in the
current study.

These results may be interpreted to suggest that psychopathic traits do not interact
with self-esteem in determining an individual’s response to an ego threat/attack on self-
esteem. This supports the idea that it is not high self-esteem per se which leads to anger
and aggression, but that it must be combined with a grandiose (Bushman & Baurneister,
1998) or unstable (Kernis et al., 1989) sense of self. Psychopathic traits did not appear to

be a protective or a risk factor against anger.

Gender Differences in Paychopathic Tra_i_ts

It was predicted that females would have lower levels of psychopathic traits than
males. Specifically, females were expected to have lower total scores on the PPI than
males. These results were expected based on previous studies using both self-report and
observer ratings of psychopathic traits which showed higher levels of these traits in

males. These findings emerged in both criminal (Forth, Brown, Hart, & Hare, 1996;

51

Meloy, 1992; Salekin, Rogers, & Sewell, 1997) and noncriminal samples (Lilienfeld &
Andrews, 1996). Gender differences on several PPI subscales were found by Lilienfeld
and Andrews (1996). Thus, prior research suggested that males would show both higher
overall levels of psychopathic traits as well as higher levels of particular factors within
psychopathy.

This hypothesis was supported. Females had lower overall levels of psychopathic
traits than males. In addition, analysis of subscales within the overall total score revealed
that males showed higher levels of Machiavellian Egocentricity, Fearlessness, Stress
Immunity, Impulsive Nonconforrnity, and Coldheartedness. Results were in line with
predictions, and this appears to be a robust finding, as it has been shown consistently in
the psychopathy literature. Most studies to date have been done with incarcerated
populations, among individuals who have been classified as psychopaths using some
version of Hare’s Psychopathy Checklist. This study provides further support for the
gender difference among individuals who are not incarcerated and do not have clinical
levels of psychopathy. It is also interesting to note that the validation study of the PPI by
Lilienfeld and Andrews (1996) revealed gender differences on the same subscales in
which differences emerged in the present study.

One possibility for why this gender difference exists is that similar personality
traits are expressed differently in men and women, appearing to cluster into different
domains. Hamburger, Lilienfeld, and Ogben (1996) found that psychopathic traits were
more associated with Antisocial Personality Disorder in males, while these traits had a
stronger link to Histrionic Personality Disorder in females. There is some additional

support for these findings in the current study. The subscale traits on which males scored

52

higher than females tend to be those which support antisocial behavior rather than
behavior that is dramatic and attention-seeking. The authors’ proposition that the
expression of psychopathy may be moderated by gender appears to be partially supported
by data from the current study. In addition, studies have consistently found gender
differences in aggressiveness (e.g., Paris, 1998), and the traits of Fearlessness and
Impulsive Nonconforrnity, which involve risk-taking and a disregard for social mores, are

the two subscales most closely related to aggressiveness.

GendeQifferences in Narcissistic Traifi

It was predicted that females would have lower levels of narcissistic traits than
males. Specifically, females were expected to have lower total scores on the NPI than
males. Previous research suggests that different levels of narcissism may be present
between males and females. Watson, Taylor, and Morris (1987) found higher levels of
narcissism in males than in females. In addition, expectations for different gender roles
may explain why studies have found females to show lower levels of narcissism than
males. One study of gender and narcissism investigated responses to displays of
entitlement and self-absorption by a male or female target. Results indicated that both
males and females showed higher levels of negative affect, disliked the target more, and
attributed more negative qualities to the target when the target was female (Carroll,
Hoenigmann-Stovall, & Whitehead, 1996). Thus, there is some support for the notion
that these qualities are more socially accepted in males than in females, and females may

inhibit their expression due to societal pressure.

53

This hypothesis was strongly supported. Males showed higher overall levels of
narcissistic traits than did females. In addition, males had higher scores on the NPI
subscales of Authority, Exploitativeness, and Entitlement. Similar to the findings for
psychopathic traits, these results are generally consistent with previous findings showing
higher overall levels of narcissistic traits in males than in females, as well as higher levels
on particular factors within the overall construct of narcissism. The Authority subscale of
the NPI measures such traits as dominance, assertiveness, and leadership, while the
Exploitativeness subscale is characterized by rebelliousness, hostility, and lack of concern
for the feelings of others. Finally, the Entitlement subscale summarizes ambitiousness,
power-seeking, and lack of self-control. Tschanz, Morf, and Turner (1998) argue that due
to social role expectations, women are less likely to engage in displays of dominance and
leadership than are men. They found that the Entitlement and Exploitativeness
components of the narcissistic syndrome as measured by the NPI were less strongly
correlated with the other factors of the NPI in women than in men. Results from the
present study were consistent with the notion that women show lower levels of these trait
than men and, similar to the findings regarding psychopathic traits, the expression of the
syndrome may be different for males and females. Even narcissistic women may operate

within the bounds of socially constructed gender-stereotypic rules of behavior.
Gender Differences in Self-Eateem

It was predicted that females would have lower levels of self-esteem than males.

Specifically, females were expected to have lower total scores on the RSE than males.

54

Previous research has shown gender differences in self-esteem that favor males (Kendler,
Gardner, & Prescott, 1998; Kling, Hyde, Showers, & Buswell). Explanations that have
been put forth to explain this effect include gender roles, peer interactions, school
experiences, cultural expectations of appearance, and violence against women. While this
finding is generally robust, it is not entirely consistent across studies. A recent meta-
analysis (Kling, Hyde, Showers, & Buswell, 1999) was conducted to clarify disparate
findings. They examined 216 effect sizes across 185 studies, and found a small overall
difference (d = 0.21) favoring males. A strength of their study was that they also
investigated gender differences among different age groups. They found that gender
differences in self-esteem varied over time, but were greatest (d = 0.33) in late
adolescence (ages 15-18). Among college-aged participants (19-22), the difference
dropped to 0.18, and declined further with age, finally disappearing in adults aged 60 and
older. Thus, it was hypothesized that participants in the present study, the majority of
whom were college freshmen and sophomores, would show a gender difference in self-
esteem favoring males.

This hypothesis was not supported. Results showed no significant differences
between males and females on the measure of self-esteem. While the lack of a difference
is a surprising finding, it is arguably a positive one; few would argue that lower self-
esteem in women is a good thing. An exhaustive discussion of gender and self-esteem is
beyond the scope of this paper; however, the literature provides some general themes
which may be operating to explain this finding. Although results did not support the
predicted difference in self-esteem, one explanation is suggested by the work of Kling,

Hyde, Showers, and Buswell, (1999). The meta-analysis by these authors provides strong

55

support for a trend in which gender differences peak at around age 16, but then diminish
over time. Their study suggests that by college, differences are decreasing. It seems
plausible that attending college works to increase females’ self-esteem, perhaps by
exposing them to alternative viewpoints and opening up a new set of potential areas of
mastery. The use of a global measure of self-esteem, while commonly used in studies of
this construct, necessarily masks any differences among domains of esteem, such as
physical appearance, academic achievement, athletic ability, or popularity among peers.
Perhaps as individuals pass from late adolescence into young adulthood, these domains
begin to change in importance in a way that favors females. A study by Burnett,
Anderson, and Heppner (1995) suggests that more stereotypically masculine traits, such
as independence, assertiveness, and decisiveness are associated with higher levels of self-
esteem. It may be that after high school, it becomes more socially accepted for women to
display these traits. ’

Gender differences in personality traits are in part a function of the sex roles
present in our society, and therefore, they may change over time. This seems particularly
relevant in the area of self-esteem. As society has become more accepting of women
taking on more stereotypically masculine roles (i.e., breadwinner, professional), perhaps
self-esteem has risen on average for women while it has not changed for men. In terms of
narcissistic and psychopathic traits, it would be interesting to administer these personality
measures to a cohort twenty years from now. This would allow us to assess whether any
gender differences were present, and, if so, to consider how they might reflect the

changing nature of sex roles.

56

Anger as an Outcome of Reading Vigpettes

Participants in the control group were expected to endorse lower levels of state
anger than participants in the experimental group. Specifically, individuals who read the
neutral vignettes were expected to have lower scores on the STAXI-2 State Anger
subscale than individuals who read the ego-threatening vignettes. These results were
expected based upon pilot testing which showed that participants who read the ego-
threatening vignettes reported more anger than those who read the neutral vignettes.
These vignettes were devised by the author and have not been used in previous studies.
However, a similar methodology was used by Lyubomirsky and Nolen-Hoeksema (1995)
in their study of the effects of self-focused rumination in college students. These authors
used a modified version of the Cognitive Bias Questionnaire (Krantz & Hammen, 1979)
to elicit open-ended written responses to six vignettes, including three with achievement
themes and three with interpersonal themes. The purpose of the CBQ is to assess
responses to situations that involve a depressive bias.

The vignettes used in the current study were designed to tap into several domains
of self-esteem that have been discussed in the literature, such as home and parents,
academic competence, and attractiveness (Quatrnan & Watson, 2001), as well as
employment. It was not possible to determine the full range of effects, if any, that reading
the vignettes would have. However, for the present study, pilot data were taken as
evidence that reading the ego-threatening vignettes resulted in higher levels of state anger
than reading the neutral vignettes. Based upon results of the Lyubomirksy and Nolen-
Hoeksema (1995) study, which found no gender differences in responses to their

vignettes, no interaction between gender and vignette type was expected.

57

This hypothesis was partially supported. Initial comparisons of the experimental
and control groups revealed no differences in level of state anger. However, further
analysis revealed that there was a significant interaction between gender and group
assignment. Females in the control group were significantly less angry than males in the
control group. Females in the control group were significantly less angry than females in
the experimental group. Males in the control group were significantly more angry than
males in the experimental group. Thus, it appears that the ego-threat affected males and
females differently, females who read the ego-threatening vignettes were more angry than
those who did not, while males who read these same vignettes were less angry than those
who did not.

It appears that the initial finding of no group differences was due to gender
differences in response to the ego-threat, which cancelled out overall group effects. Why
would gender differences occur? There are several possibilities, one being that females in
the study were more affected by vignettes in the domains selected for the study (work,
school, family, romantic relationships). An attempt was made to choose domains which
would not be gender biased, based on the notion that men are more oriented toward
independence and achievement, and women are more interpersonally oriented. In
addition, participants were asked to write responses to the vignettes, to insure that they
read closely enough to express a reaction. It is possible that women made more of an
attempt to follow the instructions by imagining themselves in the situation; however,
there is no clear reason why this would occur. One possibility is suggested by the work
of Nolen-Hoeksema regarding gender differences in coping. She has found that a

rurninative coping style, characterized by “passively focusing on one’s symptoms of

58

distress and the circumstances surrounding those symptoms” (N olen-Hoeksema,
McBride, & Larson, 1997, p. 855) is associated with an increased probability of
experiencing distress, and for longer periods of time, than not ruminating (N olen-
Hoeksema, 1991; Nolen-Hoeksema, Morrow, & Fredrickson, 1993). Her research has
shown that females engage in more ruminative coping than males, who tend to use
distractions (i.e., purposely thinking about or doing something else), and that this
difference in coping strategies may explain part of the gender difference in rates of
depression (Butler & Nolen-Hoeksema, 1994; Nolen-Hoeksema, 1991; Nolen-Hoeksema,
Morrow, & Fredrickson, 1993; Nolen-Hoeksema, Larson, & Grayson, 1999). I

Expanding on her response styles theory as it relates to depression, Rusting and
Nolen-Hoeksema (1998) investigated coping strategies in response to an angry mood.
This study found that a rurrrinative style increased anger, while distraction either did not
affect participants’ level of anger or decreased anger. However, this study also found that
women were more likely to distract themselves than to ruminate when angry, and men
chose the strategies of rumination and distraction equally across different mood
conditions. In the present study, it is possible that females tended to ruminate more after
reading the vignettes, thereby exacerbating any negative effects, while males actively
distracted themselves fi'om the negative information. This would explain why males in
the experimental group, paradoxically, were less angry than those in the control group,
who read the neutral stories.

Another explanation for the current findings is discussed by Kring and Gordon
(1998), and has to do with gender differences in the experience and expression of

emotion. Kring and Gordon’s review of a large number of studies concluded that women

59

are more expressive of anger, among many other emotions. This has been found
consistently despite the fact that anger is a less socially acceptable emotion for women to
display. Interestingly, Kring and Gordon found that this difference in emotional
expressiveness does not appear to be due to gender differences in the experience of
emotion. Thus, it is possible that males and females both experienced anger in reaction to
the vignettes, but women were more willing to endorse feeling angry on the
questionnaire. It is particularly interesting that men in the control group were
significantly more angry than men in the experimental group. This suggests that, for
males, the vignettes had the opposite effect than was intended. It would be interesting to
conduct further studies to determine whether this is a consistent finding using the vignette

methodology.

Limitations of the Stufl

A number of interesting findings emerged from the present study. However,
while these results may be representative of the college population, they do not easily
generalize to most adults. As Sears (1987) argued, college students have stronger
cognitive skills, are more willing to submit to authority, and have a less developed sense
of self than the general population of adults, most of whom are older. These variables
could potentially influence responses to the vignettes which deal with disapproval by an
authority figure; they may also result in a more intellectualized, less emotional response
to the vignettes than would be seen in the general population. It also seems likely that
personality traits which reflect higher levels of self-involvement, as psychopathy and

narcissism do, will be overrepresented in college-age students as compared to the general

60

population. In addition, this sample did not reflect the ethnic diversity of the population
in general. A larger sample size would also have been useful to allow for more statistical
power when comparing high and low groups on such factors as self-esteem and
narcissistic and psychopathic traits. The relative lack of significant findings related to the
ego threat may be due in part to the size of the sample. While enough power was present
to detect large effects, small or medium effects may have been present, but insufficient
power existed to detect these effects.

This study was entirely reliant on self-report information. While a validity check
on the PPI and a post-experiment questionnaire were employed to eliminate data from
participants who did not answer honestly, these methods are more likely to detect obvious
deception or lack of effort. It is not possible to know with certainty whether the traits and
feelings participants reported were entirely accurate, or were affected by social
desirability factors, or minimal investment in performing the tasks.

One final limitation, which was not evident until after data collection, was that
males and females were not affected in the same ways by reading the vignettes. While
this makes group analyses difficult to interpret because of the likelihood that effects were
washed out, it raises a number of interesting questions about gender differences in
response to situations that threaten self-esteem. Thus, a methodological shortcoming

produced an unexpected finding worthy of further exploration.

Implications
This study expanded on previous research by addressing questions which have not

yet been investigated, including furthering our understanding of how individuals with

61

 

psychopathic traits may respond in the face of an ego threat. Psychopathy was examined
as a cluster of traits relevant to ego threat, but somewhat different than narcissism. This
study was designed to contribute to our understanding of the distinction between the
theoretical constructs of psychopathy and narcissism, which show a moderate to high
degree of overlap in previous research (e.g., Reise & Wink, 1995). Studying individuals
with subclinical psychopathic traits rather than the full syndrome as defined by Hare
(1993) is important for obtaining results that may generalize to a nonincarcerated
population.

Use of the PPI permitted investigation into more specific domains of psychopathy,
such as blame externalization and stress immunity. Use of the NPI also permitted an
analysis of specific domains within narcissism, such as entitlement and exploitativeness.
Gender differences were apparent on several of these subscales, as well as overall trait
scores. Individuals’ experience of anger was investigated in relation to ego-threatening
scenarios, which may have different implications than studying aggressive behavior.
Investigating anger as a dependent variable, as discussed above, can tell us more about
the affective experience which may precede aggressive behavior. For example, by acting
aggressively, one may be able to discharge feelings of anger and regain previous levels of

self-esteem.

62

APPENDD( A

Vignettes for Experimental and Control Groups

Instructions: Please do your best to imagine yourself in the following
situations. Try to create a visual image of the scene unfolding, and imagine what
you would be thinking and feeling if these events were occurring. Then please write
a paragraph describing your thoughts and feelings in each situation, and (in the
experimental group only) explain what action you would take to resolve the

situation.

1) You are taking a challenging course that is required for your major. All semester, you
have done your best to keep up with the work, including going to see the professor and
TA’s during their office hours. This class has taken up more of your time than any other
one so far, but you don’t mind because the subject matter is so interesting and you know
it will be relevant to your future work after you’ve graduated. In addition, you need to get
at least a 3.0 in the class to keep your scholarship. You feel confident going into the final
exam that you will 4-point the class. However, while you are taking the test you realize
that there was a whole section of your notes that you forgot to study. Besides that, the
test questions seem really difficult and you can’t seem to come up with the right answers
for things that you did study. After the test, you ask your fiiends what they thought of it.

They say it wasn’t bad, and in fact was easier than the midterm.

63

Stacy endings:

Experimental: When you check your grades, your fears are confirmed: while you were
expecting a 4.0, your final grade is a 2.5.

Control: When you check your grades, you find that they haven’t been posted yet.

2) A few weeks ago, you met someone who was introduced by a mutual friend. As
you’ve talked with them more, you’ve discovered that you have a lot in common with this
person—a similar sense of humor, like the same movies and the same activities. In
addition, you feel very comfortable talking with this person—they accept you for who
you are and make you feel good about yourself. You find yourselves spending more and
more time together, you’ve become very attracted to them, and you learn that they aren’t
in a relationship with anyone at the moment. Eventually, it seems like a good time to let

them know how you feel.

Stacy Endings:

Experimental: When you tell them you’d like your relationship to move beyond
friendship, you get a look of surprise in return. The person tells you that while you’re a

nice person, they don’t think a relationship would work out because you’re not really their

type.

Control: You decide to tell them you’d like your relationship to move beyond

friendship.

3) You’ve just finished a second-round interview at a place that seems like your
dream job. The work sounds really interesting, the people there are fiiendly, and the
salary is very competitive. You haven’t had the best success interviewing at other places,
but this one seems different. You feel like you connected well with the people you
interviewed with, and they seemed to be impressed with your experience and
communication skills. At the end of the day, the person in charge of hiring decisions
smiles, shakes your hand, and tells you that you’re in their top group of applicants; they’ll

let you know within the week whether you’ve got the job.

Stan: Endings:

Experimental: The next day, your phone rings and your Caller ID shows that it’s from
this potential job. You excitedly pick up the phone, fairly sure at this point that the job is
yours. However, the person explains to you that it turns out there were other, better

qualified people for the job and an offer will not be extended.

Control: No additional ending

65

4) You haven’t seen your favorite sibling in quite a while, but this weekend you have
plans to visit with him or her. When you get there, you are happy to find that you’re

having a great time, relaxing and enjoying being together again.

Stogy Endings:

Experimental: This is fine for a while, but then you notice a change. Your sibling starts
acting different, and you can’t figure out why. Finally, they tell you that the rest of the
family is really disappointed in some of the choices you’ve made recently. Basically, the
family chose your closest sibling to let you know of their disapproval, thinking he or she
could help you get back on the right track. This is a real surprise to you, to find out that

you’ve failed in their eyes.

Control: The time seems to go by quickly, and eventually you head home, hoping you

can visit again soon.

66

APPENDD( B
Mood Questionnaire

Please think about how you are feeling right now, and write a number for each item
below using the following scale:

not at all somewhat moderately very much extremely

_ 1. Happy

2. Embarrassed

3. Annoyed

4. Sad

5. Hopeful

6. Serious

7. Bored

_ 8. Depressed
9. Anxious

10. Confident
11. Relaxed

12. Pessimistic
13. Lightlrearted
l4. Surprised
15. Irritated

16. Tense

17. Fearful

18. Proud

67

References

American Psychiatric Association (1980). Diagnostic and statistical manual of
mental disorders (3rd ed.). Washington, DC: American Psychiatric Association.

American Psychiatric Association (1994). Diagnostignd statistical manual of
mental disorders (4th ed.). Washington, DC: American Psychiatric Association.

Baurneister, R. F. (1996). Self-regulation and ego threat: Motivated cognition,
self-deception, and destructive goal setting. In P. M. GollWitzer & J. A. Bargh (Eds.),
The psychology of action: Linking cognition and motivation to behavior, (pp. 27 -47).
New York: Guilford.

Baurneister, R. F. (1998). The self. In D. T. Gilbert, S. T. Fiske, & G. Lindzey
(Eds.), Handbook of social psycholagy (4th ed., Vol 1). New York: McGraw-Hill.

Baurneister, R. F., & Boden, J. M. (1998). Aggression and the self: High self-
esteem, low self-control, and ego threat. In R. G. Geen & E. Donnerstein (Eds.), Human
aggession: Theories, research. and implications (pp. 111-137). San Diego: Academic
Press.

Baurneister, R. F., Heatherton, T. F., & Tice, D. M. (1993). When ego threats lead
to self-regulation failure: Negative consequences of high self-esteem. Journal of
Personality and Social Psychology, M, 141-156.

Beck, A. T., Freeman, A., & Associates (1990). Cogm'tive therapy of mnality
disorders. New York: Guilford Press.

Burnett, J. W., Anderson, W. P., & Heppner, P. P. (1995). Gender roles and self-
esteem: A consideration of environmental factors. Journal of Counseling and
Developr_nent, 3, 323-326.

Bushman, B. J ., & Baurneister, R. F. (1998). Threatened egotism, narcissism, self-
esteem, and direct and displaced aggression: Does self-love or self-hate lead to violence?

Journal of Personalig and Social Psychology, M, 219-229.

Buss, A. H., & Durkee, A. (1957). An inventory for assessing different kinds of
hostility. Journal of Consulting Psychology, 2_l_, 343-349.

Butler, L. D., & Nolen-Hoeksema, S. (1994). Gender differences in responses to
depressed mood in a college sample. Sex Roles, 3_0, 331-346.

68

Carroll, L., Hoenigmann-Stovall, G. 1., & Whitehead, G. I. (1996). The
interpersonal impact of narcissism: A comparative study of entitlement and self-
absorption factors. Journal of Social Behavior and Personalig, _l_l_, 601.613.

Cleckley, H. (1976). The maik of sanity (5th ed.). St. Louis, MO: Mosby.
Cohen, J. (1992). A power primer. Psychological Bulletin. 112, 155-159.

Cooke, D. J ., Michie, C., Hart, S. D., & Hare, R. D. (1999). Evaluating the
screening version of the Hare Psychopathy Checklist-Revised (PCL:SV): An item
response theory analysis. Psychological Assessment. M, 3-13.

Fleming, J. S., & Courtney, B. E. (1984). The dimensionality of self-esteem: H.
Hierarchical facet model for revised measurement scales. Journal of Persorglity and
Social Psychology, M, 404-421.

Forth, A. E., Brown, S. L., Hart, S. D., & Hare, R. D. (1996). The assessment of
psychopathy in male and female noncriminals: Reliability and validity. Personalig and
Individaal Differences. _20, 531-543.

Gray-Little, B., Williams, V. S. L., & Hancock, T. D. (1997). An Item Response
Theory analysis of the Rosenberg Self-Esteem Scale. Personalig and Social Psychology
Bulletin, _2__3, 443-451.

Hamburger, M. E., Lilienfeld, S. O., & Hogben, M. (1996). Psychopathy, gender,
and gender roles: Implications for antisocial and histrionic personality disorders. Journal
of Personality Disorders. m, 41-55.

Hare, R. D. (1991). Manual for the Reviaed Psychopathy Checklist (1st ed.).
Toronto, Canada: Multi-Health Systems.

Hare, R. D. (1993). Without conscience: The disturbing world of the psychopaths
among us. New York: Guilford.

Hare, R. D. (1996). Psychopathy: A clinical construct whose time has come.
Criminal Justice and Behavior. _2_3_, 25-54.

Hare, R. D. (1998). The Hare PCL-R: Some issues concerning its use and misuse.
Legal and Criminological Psychology, 3, 99-119.

Harpur, T. J ., Hare, R. D., & Hakstian, A. R. (1989). Two-factor
conceptualization of psychopathy: Construct validity and assessment implications.

Psycholog'cal Assessment, 1, 6-17.

69

Harpur, T. J ., Hart, S. D., & Hare, R. D. (1994). Personality of the psychopath. In
P. T. Costa & T. A. Widiger (Eds.), Personality disorders and the Five-Fgor Model of
paramajdpp. 149-173). Washington, DC: American Psychological Association.

Hart, S. D., Cox, D. E., & Hare, R. D. (1995). The Hare Psychopaahy Checklist:
Screenig Version. Toronto, Canada: Multi-Health Systems.

Hart, S. D., Hare, R. D., & Harpur, T. J. (1992). The Psychopathy Checklist--
Revised (PCL-R): An overview for researchers and clinicians. In J. C. Rosen & P.
McReynolds, P. (Eds.), Advapces in psychological assessmenLVol. 8 (pp. 103-130).
New York, NY: Plenum Press.

Kendler, K. S., Gardner, C. O., & Prescott, C. A. (1998). A population-based twin
study of self-esteem and gender. Psychological Medicine, 28, 1403-1409.

Kemberg, O. F. (1984). Severe persoaality disorders: Psychotherapautic
strategies. New Haven, CT: Yale University Press.

Kemberg, O. F. (1998). Pathological narcissism and narcissistic personality
disorder: Theoretical background and diagnostic classification. In E. F. Ronningstam
(Ed.), isordera of n_arcissis_m: Diagnostic. clinica_l_and empirical implicaflaa (pp. 29-
51). Washington, DC: American Psychiatric Press.

Kernis, M. H., Grannemann, B. D., & Barclay, L. C. (1989). Stability and level of
self-esteem as predictors of anger arousal and hostility. J oumal of Personality and Socia_1
Psychology, M, 1013-1022.

Kling, K. C., Hyde, J. S., Showers, C. J ., & Buswell, B. N. (1999). Gender
differences in self-esteem: A meta-analysis. Psychological Bulletin, _1_2§, 470-500.

Kring, A. M., & Gordon, A. H. (1998). Sex differences in emotion: Expression,
experience, and physiology. Journal of Personality and Social Psychology, 3, 686-703.

Kohut, H. (1966). Forms and transformations of narcissism. Journal of the
American Psychomlytic Associatiop, M, 243-272.

Krantz, S., & Hammen, C. L. (1979). Assessment of cognitive bias in depression.
Journal of Abnormal Psychology, §_8_, 611-619.

Levenson, M. R., Kiehl, K. A., & Fitzpatrick, C. M. (1995). Assessing
psychopathic attributes in a noninstitutionalized population. Journal of Personalig and
Social Psychology, 68, 151-158.

Lilienfeld, S. O. (1994). Conceptual problems in the assessment of psychopathy.
Clinicflsychology Review. l_4, 17-38.

70

Lilienfeld, S. O., & Andrews, B. P. (1996). Development and preliminary
validation of a self-report measure of psychopathic personality traits in noncriminal
populations. Jourral of Personality Assessment, Q6, 488-524.

Lykken, D. T. (1995). ThQntisocial personalities. Hillsdale, NJ: Erlbaurn.

Lynam, D. R., Whiteside, S., & Jones, S. (1999). Self-reported psychopathy: A
validity study. Journal of Personality Assessmeni, 13, 110-132.

Lyubomirsky, S., & Nolen-Hoeksema, S. (1995). Effects of self-focused
rumination on negative thinking and interpersonal problem solving. Journal of
Persofinality Ed Social Psychology, Q, 176-190.

McCann, J. T., & Biaggio, M. K. (1989). Narcissistic personality features and
self-reported anger. Psychological Rgparts, M, 55-58.

McCarthy, J. D., & Hoge, D. R. (1984). The dynamics of self-esteem and
delinquency. American Journal of Sociology, 20, 396-412.

McCrae, R. R., & Costa, P. T. (1990). Personality in adulthood. New York:
Guilford.

Meloy, J. R. (1988). The psychopathic miad: Origins, dyr_1amics, and treatment.
Northvale, NJ: Aronson.

Meloy, J. R. (1992). Violent attachments. Northvale, NJ: Aronson.

Millon, T. (1969). Modern psychopathology: A biosocial approach to maladaptive
learning and functioning. Philadelphia: W.B. Saunders.

Millon, T., & Davis, R. D. (1998). Ten subtypes of psychopathy. In T. Millon, E.
Simonsen, M. Birket-Srrritlr, & R. D. Davis (Eds.), Psychopathy: Antisocial, criminal,
and violent behavior, (pp. 161-170). New York: Guilford.

Mruk, C. (1995). Self-esteem: Research, theog, and mactice. New York:
Springer.

Nesca, M., Dalby, J. T., & Baskerville, S. (1999). Psychosocial profile of a female
psychopath. American Journal of Forenaic Psychology, l_7, 63-7 7.

Nolen-Hoeksema, S. (1991). Responses to depression and their effects on the
duration of depressive episodes. Jourral of Abnormal Psychology, m, 569-582.

71

Nolen-Hoeksema, S., Larson, J ., & Grayson, C. (1999). Explaining the gender
difference in depressive symptoms. Joml of Personaligg and 80ml Pswholgy, fl,
1061-1072.

Nolen-Hoeksema, S., McBride, A., & Larson, J. (1997). Rurnination and
psychological distress among bereaved partners. J oumal:of Personality anASocM
Psychology, 12, 855-862.

Nolen-Hoeksema, S., Morrow, J ., & Fredrickson, B. L. (1993). Response styles
and the duration of episodes of depressed mood. Journal of Abnormal Psychology, 1_0_._2_,
20-28.

Novaco, R. W. (1975). Anger control: The deve10pment and evaluation of an
exm'mental treatment. Lexington, MA: Heath.

Papps, B. P., & O’Carroll, R. E. (1998). Extremes of self-esteem and narcissism
and the experience and expression of anger and aggression. Aggressive Behavior. _23,
421-438.

Quatrnan, T., & Watson, C. M. (2001). Gender differences in adolescent self-
esteem: An exploration of domains. J 011ml of Genetic Psychology, L6_2, 93-117.

Raskin, R. N, & Hall, C. S. (1979). A narcissistic personality inventory.
Psychological Ramrts, 4_5, 590.

Raskin, R. N., & Hall, C. S. (1981). The Narcissistic Personality Inventory:
Alternate form reliability and ftuther evidence of construct validity. Journal of

Personally Assessmeni, 4;, 159-162.

Raskin, R., & Terry, H. (1988). A principal-components analysis of the
Narcissistic Pesonality Inventory and further evidence of its construct validation. Journal
of Persoglity and Social Psychology, M, 890-902.

Reise, S. P., & Oliver, C. J. (1994). Development of a California Q-Set indicator
of primary psychopathy. Journal of Personality Assessment, Q, 130-144.

Reise, S. P., & Wink, P. (1995). Psychological implications of the Psychopathy Q-
Sort. Journal of Personalig Assessment, M, 300-312.

Rhodewalt, F., & Morf, C. C. (1995). Self and interpersonal correlates of the
Narcissistic Personality Inventory: A review and new findings. J oumal of Research in

Personally, 22, 1-23.

72

Rhodewalt, F., & Morf, C. C. (1998). On self-aggrandizement and anger: A
temporal analysis of narcissism and affective reactions to success and failure. Journal of
Personality m SociaM’sychology, 141, 672-685.

Rosenberg, M. (1965). Society and the adolescent self-imaga. Princeton, NJ:
Princeton University Press.

Rusting, C. L., & Nolen-Hoeksema, S. (1998). Regulating responses to anger:
Effects of rumination and distraction on angry mood. Joumal of Personality and Socifial
Psychology, 14, 790-803.

Salekin, R. T., Rogers, R., & Sewell, K. W. (1996). A review and meta-analysis of
the Psychopathy Checklist and the Psychopathy Checklist--Revised: Predictive validity of
dangerousness. Clinical Psychology: Science and Practice, _3, 203-215.

Salekin, R. T., Rogers, R., & Sewell, K. W. (1997). Construct validity of
psychopathy in a female offender sample: A multitrait-multimethod evaluation. Journal
of AbnormaiPsycholgy, 106, 576-585.

Sears, D. O. (1987). College sophomores in the laboratory. Influences of a narrow
data base on social psychology’s view of human nature. Journal of Personalig and Social
Psychology, _5_l, 515-530.

Spielberger, C. D. (1980). Preliminm manual for the State-Trait Anger Scale
(STAS). Tampa, FL: University of South Florida, Human Resources Institute.

Spielberger, C. D., Jacobs, G., Russell, 8., & Crane, R. (1983). Assessment of
anger: The State-Trait Anger Scale. In J. N. Butcher & C. D. Spielberger (Eds.),

Advances in monalig assessment (V 01. 2, pp. 159-187). Hillsdale, NJ: Erlbaum.

Spielberger, C. D., Reheiser, E. C., & Sydeman, S. J. (1995). Measuring the
experience, expression, and control of anger. Issues in Comprehen_sive Pediatric Nursing,
18, 207-232.

Spielberger, C. D., Sydeman, S. J ., Owen, A. E., & Marsh, B. J. (1999).
Measuring anxiety and anger with the State-Trait Anxiety Inventory (STAI) and the State-
Trait Anger Expression Inventory (STAXI). In In M. E. Maruish (Ed.), The use of

psychological testing for treatment planning and outcomes assessment (2nd ed. ), (pp.993-
1021). Mahwah, NJ: Lawrence Erlbaum Associates.

 

Steffenhagen, R. A., & Burns, J. D. (1987). The social dmamics of self-esteem:
Theory to therapy. New York: Praeger.

73

Templeman, R., & Wong, S. (1994). Determining the factor structure of the
Psychopathy Checklist: A converging approach. Multivarfie ExperimeMal Clinical
Research, 19, 157-166.

Watson, P. J ., Taylor, D., & Morris, R. (1987). Narcissism, sex roles, and self-
firnctioning. Sex Roles, i6, 335-350.

Weinberger, D. A., Schwartz, G. E., & Davidson, R. J. (1979). Low-anxious,
high-anxious, and repressive coping styles: Psychometric patterns and behavioral and
physiological responses to stress. Journai of Abnormal Psychology, 88, 369-380.

Widiger, T. A., & Lynam, D. R. (1998). Psychopathy and the Five-F actor Model
of personality. In T. Millon, E. Simonsen, M. Birket-Smith, & R. D. Davis (Eds.),

Psychopathy: antisocial, criminal, and violent behavior (pp. 171-187). New York:
Guilford.

74