-~¢ no...

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

‘As'LA‘iw

AUGfi 1 8 tech

 

 

 

 

 

 

 

MAR 1 51994
ID \18

onP 2 515%
. - “7—1::—

S
k - Jx’r If ‘1

 

 

   
    

/

 

 

 

. t‘r
g 3 ‘1.
AA" '
1*

MSU Is An Affirmative Action/Equal Opportunity Institution

 

 

 

 

 

 

 

WOMEN AND ANGER: THE RELATIONSHIP BETWEEN
SEX ROLE, SELF-ESTEEM AND AWARENESS, EXPRESSION,
AND CONDEMNATION OF ANGER

BY

Cynthia Anne Hockett

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Counseling,
Educational Psychology and Special Education

1988

{A
si}

'53. L:

L:
.J

(003

ABSTRACT
WOMEN AND ANGER:

THE RELATIONSHIP BETWEEN SEX ROLE, SELF-ESTEEM
AND AWARENESS, EXPRESSION, AND CONDEMNATION OF ANGER

BY
Cynthia Anne Hockett

This descriptive study investigated the relationship
between anger, sex role, and self-esteem in adult women
university employees. Three of the four primary hypotheses
were supported. The effects of demographic variables on the
main variables were also investigated. The sample showed
differences in.anger in four areas: sex role, marital status,
age, and educational level. Feminine women had the greatest
awareness of anger, but masculine, and to a lesser extent
androgynous, women were most likely to express their anger.
Contrary to predictions, no differences between the sex role
groups in physical expression of anger and condemnation of
anger were found. Women who had never been married were more
likely to express their anger than women who were or had
previously been married. The youngest women in the sample
were more aware of their anger than the oldest women. Women
without a college degree felt more condemning of their anger,
and were less likely to express it, than women with education
beyond the master's level.

The relationship found between anger and self-esteem is
complex. Self-esteem was negatively related to awareness of

anger, but positively related to verbal expression of anger.

Sex role differences in self-esteem were consistent with
earlier findings, with androgynous women having the highest
self-esteem, followed. by ‘masculine ‘womenn Feminine and
undifferentiated women reported the lowest self-esteem.

For masculine women, it was predicted that awareness and
expression of anger would be positively related to self-
esteem, and that condemnation of anger would be negatively
related to self-esteem. The inverse of these relationships
was predicted for feminine women. Self-esteem was not
expected to vary with anger for androgynous and
undifferentiated women. These predictions were not supported
since the effect of anger in predicting self-esteem was the
same for all sex role groups. Sex role was a more significant
predictor of self-esteem than anger.

The results.of this research.suggest.that.cultural values
prohibiting anger in women have a significant impact on self-
esteem regardless of whether a woman perceives herself in a

traditional or nontraditional role.

Copyright by
CYNTHIA ANNE HOCKETT

1988

This dissertation is dedicated to my husband,

David L. Froh

ACKNOWLEDGEMENTS

I offer my appreciation and gratitude to the following
individuals, each of whom has made important contributions to
this project.

To Dr. John Powell, chairperson of my dissertation
committee, who helped me begin the journey that has brought
me to this milestone in my life, and has extended his
unflagging support, encouragement, and trust in my judgment
and professional development.

To Dr. Linda Forrest, dissertation committee member, who
has served as a supportive role model and energized me with
her interest in and excitement about my research.

To Dr. Richard Johnson, dissertation committee member,
whose insights and probing questions helped me sharpen my
thinking.

To Dr. Max Raines, dissertation committee member, whose
fresh perspective and willingness to share himself personally
helped me to broaden my thinking and integrate theory with
humanistic applications.

To my dear friends and colleagues Linda Gold and Michael
Ranieri, whose generosity, support and flexibility have

provided a haven for me and enriched my life.

vi

To Dr. David Wenger, whose gentle mentorship has given
me reassurance and helped me develop resilience and strength.

To Tom Helma, coordinator of the Michigan State
University Employee Assistance Program, whose enthusiasm for
and commitment to this project were essential to its
completion.

To the women of Michigan State University who made this
project possible through their participation.

To David LaRoy Froh, my husband, who has given so much
of himself, whose encouragement, nurturing, and light-
heartedness have sustained me and helped make this dream a
reality, and whose nimble fingers and agile mind solved the
mysteries of word processing for the data collection and
innumerable drafts of this dissertation.

And to my parents, Rosemary and Howard Hockett, who
launched me on my quest for higher education and whose values

have always inspired me to reach beyond.

vii

TABLE OF CONTENTS

LIST OF TABIES O O O O O O O O O O O O O O O O O O O 0

Chapter

I.

II.

III.

INTRODUCTION 0 O O O O O I O O O O O O O O O O O 0

Need for Research on Anger in Women
Purpose and Procedures . . .
Hypotheses . . . . . . .
Assumptions . . .
Delimitations . .
Overview . . . . .

REVIEW OF THE THEORETICAL LITERATURE . . . . . . .

Relationship Between Anger and Aggression . . . .
Traditional Theory of Female Psychological
Development . . . . . . . . . . . . . . . . . .
Challenge to the Traditional View . . . . . . . .
Cultural Prohibitions on Women's Anger and
Aggression . . . . . . . . . . . . . . . . . .
Intrapsychic Inhibition of Anger in Women . .
Cultural Sources of the Intrapsychic Inhibition
of Anger . . . . . . . . . . . . . . . . . . .
Anger in the Mother-Daughter Relationship . . .
Loss of Self-Esteem Resulting from Inhibition of
Anger . . . . . . . . . . . . . . . . . . . . .
Sex Roles as a Context for Women's Anger and Self-
Esteem . . . . . . . . . . . . . . . . . . . . .
Summary . . . . . . . . . . . . . . . . . . . . .

REVIEW OF THE EMPIRICAL LITERATURE . . . . . . . .

Studies on Sex Differences in Aggression . . . . .
Studies on Anger in Women . . . . . . . . . . .
Studies on Sex Role, Self-Esteem and/or Related
Variables . . . . . . . . . . . . . . . . . .
Relationship Between Sex Role and Self-Esteem .
Relationship Between Sex Role, Self-Esteem and
Related Variables . . . . . . . . . . . . . .
Relationship Between Sex Role and Variables
Related to Psychological Adjustment . . . . .

viii

xi

H

O‘DWNIOtH

13

17
17

19
21

22
24

28

3O
33

37

38
54

73
74

78

81

IV.

VI.

Some Observations on Methodology . . . . . . . . .
summary 0 l O O O I O O I O 0 O O O O O O O O O O

DES IGN MD METHommGY O O O O O O O O O O O O O 0

Subjects . . . . . . . . . .
Method of Data Collection . . .
Procedures to Ensure Confidentiality
Instrumentation . . . . . . . . . .
Anger Self-Report . . . . . . . .
Personal Attributes Questionnaire
Texas Social Behavior Inventory .
Hypotheses . . . . . . . . . . . .
Statistical Procedures for Data Analysi
Assumptions of Analysis of Variance
Stepwise Multiple Regression . . . .
Summary . . . . . . . . . . . . . . .

ANALYSIS OF THE DATA . . . . . . . . . . . . .

Main Variables . . .
Demographic Characteristics
Normative Data . . . . . . .
Anger Self-Report . . . . . .
Personal Attributes Questionnaire
Texas Social Behavior Inventory
Hypotheses Tested . .
Hypothesis 1 .
Hypothesis 2 .
Hypothesis 3 .
Hypothesis 4 .
Summary . . . .

SUMMARY AND CONCLUSIONS . . . . . . . . . . . . .

Summary of the Study . . . . . . . . .
Conclusions . . . . . . . . . .
Sex Role Differences in Anger . . .
Marital Status Differences in Anger
Age Differences in Anger . . . . . . .
Differences in Anger by Educational Level
Relationship of Anger to Self-Esteem . . .
Sex Role Differences in Self-Esteem . . . . .
Interaction of Sex Role and Anger in Predicting
Self-Esteem . . . . . . . . . . . . . . . .
Discussion of Findings . . . . . . . .
Sex Role Differences in Anger . . .
Marital Status Differences in Anger
Age Differences in Anger . . . .
Differences in Anger by Educational Level
Relationship of Anger to Self-Esteem . . .
Sex Role Differences in Self-Esteem . . .

ix

93
96

99

99
100
102
103
103
108
113
116
121
122
123
127

129

129
131
131
131
134
135
135
135
149
151
154
157

160

160
163
163
165
165
165
165
166

166
167
167
170
171
172
173
174

Interaction of Sex Role and Anger in Predicting

Self-Esteem . . . . . . . . . . . . . . . . 176
Limitations . . . . . . . . . . . . . . . . . . . 180
External Validity . . . . . . . . . . . . . . . 181
Design and Methodology . . . . . . . . . . . . . 183
Instrumentation . . . . . . . . . . . . . . . . 183
Implications for Future Res sea arch . . . . . . . . . 185

APPENDICES O O I O O O O O O C O O O O I O O O O O O O O 18 9

A. Initial Letter Requesting Subjects' Participation 189
B. Interest Card Enclosed in First Mailing . . . . . 190
C. Follow-up Mailing Sent One Week After

Initial Letter . . . . . . . . . . . . . . . . 191
D. Questionnaire . . . . . . . . 192
E. Cover Letter Enclosed With Questionnaire . . . . 205
F. Follow-up Mailing Sent One Week After

Questionnaire . . . . . . . . . . . 206
G. Return Card Enclosed With Questionnaire . . . . 207
H. Expanded Abstract Sent to Subjects Requesting

Results Summary . . . . . . . . . . . . . . . 208
I. Cover Letter Enclosed With Results Summary . . . 211

RE FERENCES O O O O O O O O O O O O O O O O O O O O O O 2 1 2

LIST OF TABLES

Table Page
4.1 Pearson Correlations: Main Variables . . . . . . 126
5.1 Reliabilities and Descriptive Statistics . . . . . 132
5.2 Demographic Characteristics of Sample . . . . . . 133
5.3 Multivariate Tests of Significance . . . . . . . . 136
5.4 Univariate F-Tests . . . . . . . . . . . . . . . . 138
5.5 ASR Scores by Sex Role . . . . . . . . . . . . . . 139
5.6 ASR Scores by Marital Status . . . . . . . . . . . 145
5.7 ASR Scores by Age . . . . . . . . . . . . . . . . 146
5.8 ASR Scores by College Educational Level . . . . . 148

5.9 Regression of TSBI on ASR Scales
Analysis of Variance . . . . . . . . . . . . . . . 150

5.10 Regression of TSBI on ASR Awareness of Anger,
ASR Verbal Aggression & Household Income

Analysis of Variance . . . . . . . . . . . . . . . 152
5.11 Analysis of Variance for Self-Esteem on Sex Role . 153
5.12 TSBI Scores by Sex Role . . . . . . . . . . . . . 154
5.13 Regression of TSBI on ASR Awareness of Anger,

ASR Verbal Aggression & Sex Role

Analysis of Variance . . . . . . . . . . . . . . . 156

6.1 Overall Results of Hypothesis Testing . . . . . . 164

xi

CHAPTER I
INTRODUCTION

The Need for Research on Anger in Women

During the past ten years a body of theoretical
literature on the psychological development of women has been
growing (e.g., Chodorow, 1978: Cox, 1976; Gilligan, 1982;
Miller, 1976). An important component of this developmental
theory is the proposition that women have serious difficulties
with anger in the form of intrapsychic inhibitions of anger
as a result of two factors. First, it is proposed that anger ~
has been prohibited in the feminine sex role, which prescribes
that the primary role for women is caring for men and
children. Theorists believe that this prohibition has
resulted in powerful intrapsychic inhibitions of anger for
women (Bernardez-Bonesatti, 1986; Miller, 1976; Miller, 1983:
Symonds, 1976; Westkott, 1986) . Second, theorists assert that
developmental factors in the female child's relationship with
her mother have also led to strong intrapsychic inhibitions
of anger for women (Bernardez-Bonesatti, 1976; Chodorow, 1978;
Lerner, 1980).

A central tenet of the theoretical literature on women's
anger is that the inhibition of anger has negative
consequences for their mental health. It is argued that when
a woman does become aware of angry feelings, she experiences

a severe loss of self-esteem for two reasons. First, the

2

experience of her own anger threatens her identity as a woman
(Bernardez-Bonesatti, 1978; Kaplan, 1976; Miller, Nadelson,
Notman, & Zilbach, 1981; Nadelson, Notman, Miller, & Zilbach,
1982; Notman, 1982). Second, because the anger is experienced. \
as threatening, it is ultimately directed against the woman U
herself, resulting in.self-hatred (Bernardez-Bonesatti, 1978{//
Westkott, 1986).

Helping women become aware of and learn to express their
anger is assumed to be an essential aspect of psychotherapy
with women (Barrett, Berg, Eaton, & Pomeroy, 1974; Bernardez-
Bonesatti, 1978; Cline-Naffziger, 1974; Kaplan, Brooks,
McComb, Shapiro, & Sodano, 1983; Lerner, 1980). Approaches
to working with anger in therapy vary from beginning with
"screaming, kicking, yelling sessions to release the excess
and get the burden down to a manageable size" (Cline-
Naffziger, p. 55) to controlling and channelling anger to
bring about desired changes in circumstances (Lerner, 1985;
Tavris, 1982).

In spite of the importance placed on anger in therapy
with women, therapists appear to have considerable difficulty
in working with women's anger. Barrett et al. (1974) maintain
that the anger of women is often "diffused and rendered
powerless" (p. 12) in therapy. Kaplan et a1. (1983) argue
that the developmental paradigm for women, in which the
relational self is central, can prevent female therapists from

recognizing and exploring anger in therapeutic relationships

3
with women. Bernardez attributes the difficulties which
therapists have with women's anger to their unconscious fears
of female destructiveness. These fears may lead therapists
to discourage women from expressing their anger freely and
openly, and to induce them to feel guilty instead (Bernardez,
1985a; Bernardez-Bonesatti, 1976 & 1978).

The importance of anger in psychotherapy becomes obvious
when one considers the areas of emotional distress in which
anger is presumed to play an important role. Freud (1957)
maintained that melancholia was the result of reproaches
against a lost love object shifting away from the object to
the individual's own ego. This view is still espoused in thek
view that depression in‘women is the result of repressed anger‘
(Bernardez, 1985b; Cline-Naffziger, 1974; Eichenbaum & Orbach,
1983). Kaplan (1986) explains that although depressed women
are sometimes aware of their angry feelings, they experience
these feelings as destructive and as confirming their badness
and worthlessness. Given the hypothesized prohibition on
anger in the feminine sex role, it is not surprising that
depression is much more common among women than men and is
considered a woman's disorder. A recent study indicates that
twice as many women as men experience depressive episodes,
with ten percent of all women having a serious depression
during their lifetimes (Weissman & Klerman, 1979).

In addition to depression, anger is associated with

another serious problem for women. It is thought to be an

4

important factor in suicide attempts. For example, Weissman,
Fox and Klerman (1973) found that manifest hostility was an
important factor in distinguishing women who attempted suicide
from women who were acutely Irodépattyed. while
repression of anger seems to result in depression,
difficulties in its regulation and expression appear to be
associated with attempted suicide. These findings underscore
the importance both of addressing anger in therapy with women,
and of handling anger skillfully in therapy. Failure to
address anger issues may result in the anger remaining buried
and the woman being depressed. Helping a woman become aware
of her anger without also helping her use it constructively
may pose a risk of suicide or other less dramatic self-
destructive actions.

The importance of gaining greater understanding of anger
in women is also evident when the range of physical disorders
associated with anger is considered. Iewitan (1981) cites
studies which found that women with rheumatoid arthritis
handle anger in the following ways: becoming especially
concerned with controlling anger; glossing over frustrating
situations or turning their hostility inward: restricting
their expression of hostility; reacting to provocation by
feeling hurt and upset rather than angry; or becoming more
aggressive toward themselves and showing less overt
aggression. Zealley (1971) found that women attending a

hospital clinic for treatment of bronchial asthma had

5

psychological disturbances involving hostility. Greer and
Morris (1978) found that women with breast cancer showed a
persistent pattern throughout their adult lives of extreme
suppression of their anger. This pattern was not found in
women with benign breast disease. According to Pelletier
(1977), difficulties with anger are associated with stress
reactivity, cancer, rheumatoid arthritis, colitis, migraine
headache, and asthma. The indirect expression of anger has
shown significant correlations with psychosomatic symptoms in
research conducted by Heiser and Gannon (1984). Most
recently, several studies have reported significant
relationships between difficulties with anger and hostility
and the incidence of cardiovascular disease (Chesney &
Rosenman, 1985).

In spite of the important role attributed to anger in
mental and physical health, little empirical research has been
conducted in this area. A growing body of theoretical
literature proposes an important relationship between anger
and two areas important to women's functioning and well-being:
sex role and self-esteem. In spite of the important and
central role which anger is assigned in developmental theory
for women, virtually no empirical research investigating the
relationship of anger to sex role and self-esteem has been
found. In light of the theoretical emphasis placed upon

anger, the need for such a study is obvious.

6

This study is significant because it provides empirical
data in two important areas in which there are presently gaps
in knowledge. First, this study explores the empirical
foundations of an important aspect of female developmental
theory regarding anger as a determinant of self—esteem. Given
the crucial role which self-esteem plays in mental health, it
is imperative that research be conducted to determine the
accuracy of theory in this area. The results of this study
point the way to areas in which further research is essential
to deepen current understanding of factors which affect how
women regard their own anger, and the resulting impact on
their self-esteem.

This study also makes a significant contribution in
broadening knowledge which can be applied by practitioners
working with women in psychotherapy. The empirical data
regarding the relationship between anger, sex role and self-
esteem provide mental health professionals with valuable
insights and information regarding how, and to what extent,
women are able to integrate anger with their sex roles and
how anger affects self-esteem. Such insights and information
are valuable to therapists working with women in becoming more
skilled in working with anger and in becoming more sensitive
to their own countertransference reactions to anger issues.

Purpose and Procedures
The purpose of this descriptive study was to obtain and

analyze empirical data on the relationship between anger, sex

7

role and self-esteem in a large sample of women. The study
was conducted under the auspices of the Employee Assistance
Program at Michigan State University. The sample was composed
of voluntary female adult subjects employed at the University.
Data were obtained by means of anonymous mail surveys.
Relationships between the main variables were tested by
analyses of variance, stepwise multiple regression analyses,
and planned and post hoc contrasts.

The study addressed the following questions:

1. What is the relationship between sex role and these

three anger variables: awareness of anger, expression

of anger, and condemnation of anger?

2. What is the relationship between each of the three

anger variables (awareness of anger, expression of anger,

condemnation of anger) and self-esteem?

3. What are the differences in self-esteem between the

sex roles groups in this sample of adult women?

4. Does the relationship between the three anger

variables (awareness of anger, expression of anger,

condemnation of anger) and self-esteem vary according to

sex role?

Hypotheses
The hypotheses which were tested in this study are:
1. Masculine and androgynous sex roles will be

associated with higher levels of awareness and expression

8
of anger and lower levels of condemnation of anger than
feminine and undifferentiated sex roles.
2. Awareness and expression of anger will negatively
predict self-esteem, while condemnation of anger will
positively predict self-esteem.
3. The women with androgynous sex roles will have the
highest self-esteem, followed in descending order by the
women with masculine, feminine and undifferentiated sex
roles.
4. The relationship between anger and self-esteem will
vary according to sex role in the following manner:
a. For women with masculine sex roles, awareness
and expression of anger will be positively
associated with self-esteem, and condemnation of
anger will be negatively associated with self-
esteem.
b. For women with feminine sex roles, awareness
and expression of anger will be negatively
associated with self-esteem, and condemnation of
anger will be postively associated with self-esteem.
c. For women with androgynous or undifferentiated
sex roles, self-esteem will not vary according to

awareness, expression and condemnation of anger.

9
Assumptions

This research is based on the following assumptions:
1. The role which is considered culturally appropriate for
women is that of nurturing and caring for others. The active
expression of aggression, including anger, by women is
considered threatening to this role and is, therefore,
culturally prohibited.
2. The cultural prohibition of aggression and anger in women
has been a primary factor in the development of intrapsychic
inhibitions of anger for women. Because of these intrapsychic
inhibitions, women are prone to repress their awareness and
experience of their own anger.
3. For ‘women, self-esteem. is based on, their' personal
evaluation of their abilities to establish and maintain caring
and nurturing relationships with others. The intrapsychic
inhibition of anger in women results in their viewing anger
as threatening to their relational abilities and thus to their
identity and self-esteem.
4. Both the repression, and the awareness and expression of
their own anger, affect self-esteem in women. The impact
varies, however, according to the extent to which a woman
perceives herself as adopting cultural prescriptions for her
role and behavior.

Delimitations
The conclusions which can be drawn on the basis of this

research are applicable only to the main variables of anger,

10
sex role and self-esteem when conceptualized in a manner
corresponding to that found in this study. Dorgan, Goebel,

and House (1983) have pointed out that considerable confusion

and difficulty surround the conceptualization and assessment ’

of both sex role and self-esteem. The assessment of anger is
a relatively new endeavor in the social sciences, and its
operational definition varies considerably across different
studies (Siegel, 1986). Each measure of a main variable used
in this research was selected on the basis of the
appropriateness of its conceptualization for the variable
being assessed. The conceptualization of each of the main
variables in this study is explained below.

Anger: An emotion experienced when an individual
perceives that a wrong or injustice has occurred or that an
expected or desired outcome has not been fulfilled. This
emotion is usually accompanied by a desire to correct or
avenge the situation. An individual can, however, be aware
of feeling angry without expressing the anger verbally or
behaviorally. Thus anger may or may not be accompanied or
followed by action which has the goal of righting the wrong
or injustice, obtaining the expected or desired outcome, or
taking revenge. Individuals typically have an attitude
expressing some degree of condemnation or acceptance of their
own anger and that of others.

Sex role: Culturally sanctioned personality traits,

attitudes, and behaviors for males and females which differ

11
between the sexes and are determined and reinforced by
differential cultural values for masculinity and femininity.
An individual's sex role is her subjective evaluation of the
degree to which her personality reflects the traits,
attitudes, and.behaviors considered appropriate for males and
females.

Self-esteem: An individual's subjective and positive
evaluation of her/his own personal worth and value. This
evaluation is based on the individual's perception of her
competence in establishing and. maintaining interpersonal
relationships. Self-esteem is evidenced by positive thoughts

and feelings toward the self.

Overview

The theoretical literature which has led to the
development of the questions which will be addressed in this
study is reviewed in Chapter II. The empirical literature on
anger in women is reviewed in Chapter III. The literature
reviewed in Chapter III also includes research in these three
areas related to the study of anger, self-esteem and sex role:
sex differences in aggression: the relationship between sex
role and self-esteem, with some studies including related
variables: and the relationship between sex role and variables
related to psychological adjustment. The methodology which
was used in conducting the research is explained in Chapter
IV. The analysis of the data is reported in Chapter V. The

study concludes with Chapter VI, in which the research is

12
summarized, some conclusions and possible explanations
regarding ‘the .results are. suggested, limitations of the
findings are stated, and the need for future research is
discussed. The theoretical propositions upon which this study

is based are examined first.

CHAPTER II
REVIEW OF THE THEORETICAL LITERATURE
The Relationship Between
Anger and Aggression

There is no clear definition as to what constitutes the
experience of anger. The research literature is character-
ized by ambiguous and often interchangeable use of the terms
anger, hostility, and aggression, a phenomenon which has been
labelled the AHA! Syndrome (Spielberger, Johnson, Russel,
Crane, Jacobs, & Worden, 1985) . Spielberger, Jacobs, Russell,
and Crane (cited in Spielberger et al.) have proposed working
definitions of these three constructs.

Anger is generally considered to be a simpler
concept than hostility or aggression. The concept

of anger usually refers to an emotional state that

consists of feelings that vary in intensity, from

mild irritation or annoyance to fury and rage.

Although hostility usually involves angry feelings,

this concept has the connotation of a complex set

of attitudes that motivate aggressive behavior

directed toward destroying objects or injuring other

people.
While anger and hostility refer to feelings

and attitudes, the concept of aggression generally

implies destructive or punitive behaviors directed

towards other persons or objects. (p. 7)

It is generally recognized that anger is an important
a$13361: of aggression. Berkowitz (1962) describes anger as an
emotional state, resulting from frustration, which heightens
the probability of aggression and is often considered the
driVe for aggression. Whether aggression occurs depends on

the presence of a stimulus associated with the event which

armused anger .

13

l4

Buss (1961) defines anger as an emotional response with
facial-skeletal and autonomic components. Buss considers
anger one of the drives which lead to aggression and views
the energizing aspect of anger, which usually intensifies
aggression, as one of its salient characteristics. Novaco
(1975) defines anger as a strong emotional response to
provocation which has autonomic and central nervous system
components and cognitive determinants. Whether anger leads
to aggression depends, in Novaco's view, on the nature of the
provocation, situational constraints, and the individual's
coping style.

The concept of aggression originated in psychoanalytic
theory. Freud (1965) maintained that there are two primary
classes of instincts underlying human behavior, the aggressive
and the sexual. Dollard, Miller, Doob, Mowrer, and Sears
(1939) hypothesized that aggression always occurs as a result
of frustration in reaching a goal. They stated further that
the existence of frustration will always lead to some form of
aggression. Dollard et a1. include anger among their list of
the ways in which aggression may be manifested. According to
Feni(Ihel (1945) , an important aspect of aggression is its
rea“TE—rive character in respondingto and attempting to overcome
frustrations. To the object relations theorists, aggression
is not a primary motivating force for behavior, but occurs as
a reaction to the thwarting of, or failure in, the search for

relationship to an object (Greenberg & Mitchell, 1983) .

15

It is necessary to have a view of how aggression is
manifested to be able to study an aspect of aggression such
as anger. While there is no agreement among different
theories as to the source of aggression, it is possible to
define and classify aggressive behaviors. The definition of
aggression by Buss (1971) is generally accepted as describing
aggressive behaviors. According to Buss (1971), aggression
involves three factors: the delivery of noxious stimuli, an
interpersonal context, and an intention to act aggressively.

Aggressive behaviors may be classified in several
different ways. Buss ( 1971) divides aggression into two
classes: angry aggression and instrumental aggression. Buss
(1961) also distinguishes between verbal and physical
aggression, direct and indirect aggression, and active and
passive aggression. Edmunds (1980) also indicates that it is
important to classify the motivation involved (hostile or
instrumental) and to distinguish between reactive and
initiatory aggression. On the basis of these distinctions,
Edmunds has defined four classes of aggressive behavior:
initiatory instrumental, reactive instrumental, reactive
hostile, and initiatory hostile. An attempt to hurt the
victim is involved in all four classes. With the two
instrumental types, however, injury to the victim is secondary
to reaching an extrinsic goal. We extrinsic rewards are
involved in hostile aggression.

Anger and aggression are, then, intimately related to

each other. Kaplan (1976) defines anger as the emotional

16

component of aggression. .Although anger does not always lead
to aggression, anger may be considered a necessary
precondition for aggression motivated by anger and as such
may be considered to be one aspect of aggression. The common
definition of anger found in W
(1983) is "a strong feeling excited by a real or supposed
injury: often accompanied by a desire to take vengeance, or
to obtain satisfaction from the offending party" (p. 69).
Miller (1983) elaborates on the function of appropriate anger
when she says that

It tells us that something is wrong-~something

hurts-~and needs changing. Thus, anger provides a

powerful (and useful) recognition of discomfort and

motivation for action to bring about a change in
immediate conditions. It is a statement to oneself

and. to others. If it can Ibe recognized and

expressed, it has done its work. And, most

importantly, others can respond. (p. 5)

The desire to take action on one's own behalf appears to
be an important aspect of anger. Anger thus defined may be
viewed as an aspect of aggression, as can be seen in this
definition of aggression:

Those actions and impulses toward action and

assertion that give expression to the individual's

own aims and/or have an effect on others . . . .

Individual striving toward autonomous action and

self-assertion is included in this definition.

(Nadelson, Notman, Miller & Zilbach, 1982, p. 19)
Because anger and aggression are so intimately linked, much
of the theoretical and empirical work on aggression in women

speaks to the questions about women and anger which will be

addressed in this study. This theoretical work on aggression

17
provides the base for the study of anger in women and is,
therefore, summarized here.
Traditional Theory of Female
Psychological Development

During the past decade there have been major developments
in the theory regarding the psychological development of
women. Until that time, the traditional psychoanalytic view
of women formed the basis of most psychological theory about
female development. According to this view, women were
considered to be inherently passive, narcissistic and
masochistic - - the psychoanalytic triad (Miller, 1985).

According to this traditional view, a key task of female
development involved the conversion of aggression into
masochism and passivity (Deutsch, 1930). Deutsch considered
masochism."the most elementary force in feminine mental life"
(p. 60) and saw motherhood as the strongest form of
gratification of masochism and thus as the main goal of
existence for women. It was assumed that men and women
possessed the same quantity of aggression at birth, and that
the lifelong developmental task of women was to rid themselves
of their direct aggression in order to achieve femininity

(Nadelson, et al., 1982).

A Challenge to the Traditional View
Recent work has challenged this traditional view.
The traditional view of women as passive and masochistic
results not from their inherent nature, but from their

subordinate position in society, according to Miller (1976).

18

Miller has pointed out that what has been considered masochism
in women actually results from their objectification in our
culture. Being considered an object, according to Miller,
leaves women with an inner sense of being bad and evil. This
sense of self is confirmed when others who occupy more
important positions in society appear to believe that women
deserve to be treated as objects. The result of this process
is that women have come to believe there are valid reasons why
they are regarded as evil, and thus accept the subordinate
roles and abuse assigned to them.

Miller (1976) has also clarified that women are, in
contrast to the traditional psychoanalytic view, very active.
Their work is not recognized as activity, however, because it
is concerned to a large extent with helping others develop
rather than being directed toward their own goals. Miller
contends that this type of activity is not recognized as such
because "it is not activity in the male definition of it" (p.
53).

Caplan (1985) explains that the behavior of women that
has been considered to result from innate masochism actually
results from the limited options which society has made
available to women. Caplan believes that women's apparently
masochistic behavior can be understood as the consequence of
attempting to avoid punishment or guilt, putting the needs of
others first, delaying gratification, or having (or, as a
result of poor self-esteem, perceiving that one has) few or

no alternatives.

19

Cultural Prohibitions on Women's
Anger and Aggression

Developmental theorists are now asserting that the
inhibition of aggression in women has resulted from powerful
cultural prohibitions on its experience and expression rather
than from the inherent psychology and physiology of women.
This cultural prohibition on aggression in women originated
because of their mothering role (Bernadez-Bonesatti, 1978).
The prohibition arose to protect the human species by assuring
that in their capacity to bear children and in their assigned
role of socializing the young, women were assigned tender and
caring functions rather than aggressive ones. The inhibition
of women's anger, then, results from a very old and powerful
injunction against aggression in women.

Westkott (1986) traces this injunction against aggression
in women to the development of the "nurturing imperative" (p.
215) which arose with the shift from a mercantile to a
capitalist economy at the end of the eighteenth and beginning
of the nineteenth centuries. As a result of this shift, the
role of women changed from producing household commodities,
such as soap and cloth, to consuming them. Changes for men
moved their work from the home to offices, shops and
factories. In the midst of these major social changes, the
role of women became the preservation of traditional moral
‘values and the creation of a haven and refuge in the home from
the corruption of the marketplace. Westkott contends that the
"nurturing imperative” has now evolved into a "male

entitlement to female care" (p. 215). Women are expected to

20
nurture and care for men, their husbands and fathers, as well
as for their children. Thus the prohibition of women's anger
and aggression continues.

Miller (1976) analyzes the expectation that women will
be the caretakers of men and children in terms of the
subordinate position women hold relative to the dominant
position of men in society. Miller believes that women are
taught that their primary life goal is to serve others - men
and, later, children. As a result, women believe that they
can use all their attributes for others, but not for
themselves.

Miller (1983) has pointed out that the cultural
difficulty with anger extends to men as well as women.
Constraints, which are different for both sexes, are placed
on the experience and expression of anger by both men and
women. The hierarchical structure of society places men in
subordinate roles to other' men in ‘which anger; must. be
suppressed. Miller believes this hierarchical structure is
replicated within the family so that male children are not
permitted to express anger to their fathers. At the same
time, however, boys are encouraged to act aggressively. Early
in life males are discouraged from experiencing the emotion
of anger and are encouraged instead to act aggressively. Men,
as well as women, are not able to experience their anger in
a healthy way that allows them to communicate their hurt in

a nonaggressive manner.

21

Our conception of anger is based, according to Miller
(1983) , on the manner in which its expression has been allowed
in men. Constraints placed on its direct expression by men
has resulted in a distorted conception of anger. Miller
argues further that since men have not been assigned
responsibilities for the care and development of others, anger
is seen.as incompatible*with.the.ability to attend to and.care
for others that is central to the role of women.

Symonds (1976) explains that women are forced to suppress
their anger and to accept the "dependent solution to conflict"
(p. 195) . Rather than experiencing feelings of anger, or
acting aggressively; women report.that.they feel hurt, and.are
afraid to hurt others' feelings. Symonds labels this behavior
”horizontal aggression” which he defines as "emotionally
lying-down behavior where the person on the ground
dramatically indicates how and where the other has injured
him. He intends this behavior to move his listener to comply
through guilt" (p. 197). Symonds emphasizes that horizontal
aggression is not a sex-linked trait, but results from the
dependent position of those using it (primarily women).
Because the dependent person is fearful of being on her own,
she represses direct expressions of hostility that might
threaten relationships with those upon whom she depends
(Symonds, 1974).

The Intrapsychic Inhibition
of Anger in Women

The cultural prohibition of anger in the sex-role

assigned to women has resulted in profound intrapsychic

22
inhibitions of anger in women. These inhibitions have three
primary sources: the subordinate status of women in society,
the importance of relationships to women's identity and self-
esteem, and external perceptions of women's agency. In
addition to these cultural factors, several aspects of the
mother-daughter relationship play an important role in the
intrapsychic inhibition of female anger. The cultural sources
are discussed first.
q_ ,_ 1_ _.‘ ;; . t,. , f..; ,, .,_. ., . ,L.-

Miller (1983) maintains that although anger in women is
considered pathological, the subordinate status of women
constantly generates anger. Force, in the form of physical
violence or economic and social deprivation, is available to
keep women in their subordinate positions (Mueller & Leidig,
1976). In addition, because men are dominant, it can usually
be made to appear that there is no legitimate cause for a
woman's anger. As a consequence, when a woman feels angry,
she is likely to feel that there is something wrong with her.
The result, according to Miller, is:

that women generally have been led to believe that

their identity, as women, is that of persons who

should be almost totally without anger and without

the need for anger. Therefore, anger feels like a
threat to women's central sense of identity, which

has been called femininity. (p. 3).

Intrapsychic inhibitions of women's anger also arise from
the importance of the ability to form and maintain
relationships to women's identity and self-esteem. Miller
(1983) explains that to protect themselves in the subordinate

role assigned to them, women have developed the ability to be

23

sensitive and attuned to the needs of others and to maintain
and enhance their relationships with others. Because of this
necessity to develop relational abilities, and because women
are primarily responsible for the care and development of men
and children, the ”relational self” (Kaplan, Brooks, McComb,
Shapiro & Sodano, 1983 p. 30) is a central aspect of women's
identity (Bernardez-Bonesatti, 1978: Gilligan, 1982; Kaplan,
1986: Kaplan et al., 1983: Miller, 1976: Surrey, 1984).

The intrapsychic inhibition of women's anger develops
when women inhibit their own anger because they fear that it
will disrupt relationships that are central to their identity
as women (Kaplan, 1986). Bernardez-Bonesatti (1978) also
argues that because the experience and expression of anger is
inimical to the way in which women's role has been defined,
feeling angry constitutes a serious threat to a woman's sense
of identity, to her "relational self."

The way in which others perceive a woman's ability to act
on her own behalf, her agency, becomes another source of the
intrapsychic inhibition of women's anger when these
perceptions are internalized by women. Kaplan et al. (1983)
maintain that others see a woman's capacity for agency as
threatening to relationships. Women then internalize this
threatening response, experience their own agency as
threatening to their relational self, and lose their ability
to experience and express anger.

Miller (1976) has pointed out that there are two

additional mechanisms of the intrapsychic inhibition of

24
women's anger which result from their focus on the emotional
life of others and efforts to avoid threatening important
relationships. First, women are diverted and cut off from
their own emotional experiences and needs. Second, rather
than recognizing their own needs, women "transform" them and
come to see their own needs as identical to those of others
(p. 19). By meeting the needs they perceive others to have,
women then believe that they will also feel fulfilled.
e t -

Developmental factors in the female child's relation-
ship with her mother also result in strong intrapsychic
inhibitions of anger for women. According to Chodorow (1978) ,
there are important differences in pre-oedipal and oedipal
experiences of boys and girls due to the fact that girls are
mothered by an individual of their own sex. The preoedipal
period is longer for girls and their preoccupation with
preoedipal issues of primary identification and lack of
separateness or differentiation is carried into their
subsequent development and relationships. Girls continue
their attachment to their mothers during the oedipal period
and, according to Chodorow, never give up their mothers as a
love object. Chodorow believes that these developmental
factors result in women having a sense of self that is
continuous with, rather than separate from, others. As a
result, women define and experience themselves in terms of
their relationships. Men, in contrast, have a more

differentiated sense of self as a result of being mothered by

25
an individual of the opposite sex.

Chodorow (1978) also points out that women who are
mothers form intense identifications with their daughters
through which the mother may try to recreate herself, and/or
provide for her daughter the type of mothering she, herself,
would like to have had. Over time, a sense of mutuality and
responsiveness to each other develops between the mother and
daughter. The daughter has her first experiences with anger
within this intense bond with her mother (Kaplan et a1. ,
1983) , as she begins to differentiate from her mother and
experience her own competence. Kaplan et a1. maintain that
these moves away from the mother by the daughter are met with
anger by the mother, who fears abandonment. Since a sense of
connectedness to her mother is central to the daughter's sense
of herself, she is especially vulnerable to her own
abandonment fears that are provoked by her mother's anger.
The daughter's response is to correct her behavior to prevent
upsetting her mother. According to Kaplan and her associates,
it is likely that any experience of anger or rage by the
daughter will be curtailed in this process to avoid
displeasing the mother since her sense of attachment to her
mother is so crucial to the daughter's sense of herself.

Lerner (1980) maintains that there are two intrapsychic
determinants of women's fear of their own anger. First, she
believes that ‘women.lhave irrational fears of 'their’ own
destructiveness. During the process of attempting to separate

and individuate from her mother, the female child projects

26

onto her mother her own rage and aggression at her mother, her
object of dependency. These projections, along with the
punishments and narcissistic injuries caused by the mother,
result in the girl's internalized maternal imago containing
elements of the "bad, omnipotent, destructive mother" (p.
140). The girl's fear of identifying with this
internalization causes a defensive shift to experiencing
herself as powerless and helpless, which masks the girl's
experience of being destructive. Lerner points out that
women's fears of their own destructiveness are reinforced by
cultural stereotypes that encourage women to act as if they
are weaker than men and thus convey a message about how
destructive women might be if they were to be themselves.

Difficulties that daughters experience in the separation-
individuation process from their mothers account for the
second intrapsychic determinant of women's fear of their own
anger, according to Lerner (1980). These difficulties result
from the daughter's assimilation of sameness to her mother,
in contrast to a son's assimilation of difference from his
mother. The daughter must identify with her mother at the
same time she separates from her. This process may be
complicated by the mother's difficulties in encouraging her
daughter to become independent when the mother may be
attempting to live vicariously through her daughter. For this
.reason, the mother ‘may subtly undermine her daughter's
attempts to become more autonomous. The daughter, in turn,

may adopt feminine qualities, including the inhibition of her

27
anger, to protect her mother, whom she unconsciously perceives
as too fragile to tolerate her separation.

The result of these difficulties in separation-
individuation is that women have difficulty tolerating the
sense of being separate from others (Lerner, 1980) . The
experience of anger, therefore, is quite problematic since it
"involves the feeling of being separate, different, and alone
. . . . and apart from a relational context" (p. 140). Women
are likely to experience hurt, either instead of or in the
midst of, feeling anger. Whereas anger leaves a woman feeling
separate and alone, the expression of hurt emphasizes the
significance of the relationship rather than the woman's
autonomy. Lerner thus summarizes the dilemma that anger poses
for women:

The expression of legitimate anger and protest

is more than a statement of dignity and self-

respect. It is alsoa statement that one will risk

standing alone, even in the face of disapproval or

the potential loss of love from others....Not only

have women been taught that their value, if not

their very identity, rests largely on their loving

and being loved, but also, even more to the point,

many women have not achieved the degree of

autonomous functioning that would permit them to

stand separate and alone in the experience of their

anger. (pp. 145-146)

Bernardez-Bonesatti (1976) argues that the fear of
omnipotent female destructiveness is shared by members of both
sexes as a result of their experience of female power in their
early relationships with their mothers. She also maintains
that these fears are perpetuated in our society by the

inequalities that keep women in subordinate positions and

roles (Bernardez, 1985) . Because women see themselves as

28

powerfully destructive and men see themselves as totally
vulnerable to this destruction, men and women work in concert
to prevent the expression of anger by women (Bernardez-
Bonesatti, 1978). According to Miller et a1. (1981) our
cultural tendency to ignore and deny women's aggression is
directly related to this deeply-rooted fear of "the powerful
life-sustaining and potentially life-destroying woman" (p.
165).

Loss of Self-Esteem Resulting
from Inhibition of Anger

The inhibition of anger in women has serious consequences
for their mental health. Theorists contend that a drastic
loss of self-esteem results when a woman does become aware of
angry feelings. There are two factors involved in a woman's
self-esteem being threatened and destroyed by her experience
of her own anger.

Central to the first factor is the contention that a key
aspect of self-esteem for women is the ability to engage in
and.maintain mutually empathic, caring relationships (Surrey,
1985) . Because of the cultural prohibition against anger
associated with their caretaking role, the expresion of anger
by women is equated with the loss of the relationships that
are central to their self-esteem (Kaplan et al., 1983).
Kaplan and her associates state:

The expression of anger can come to be equated with

precipitating loss, not just of another person, but

more importantly of a key component of one's self-
esteem - one's capacity to maintain relational ties.

(P- 31)

29

The experience of anger directly contradicts the woman's
ego ideal, which has developed throughout her life as an
individual who is not aggressive (Miller et al., 1981.,
Nadelson et al., 1982). Notman (1982) maintains that direct
expressions of aggression or assertion by women are
accompanied by their feeling guilty or unfeminine. The
awareness of being aggressive reduces the woman's self-esteem,
leaving her with a sense of worthlessness or wrongdoing.

Bernardez-Bonesatti (1978) argues that women who express
anger openly risk the loss of ”one of their most valued
characteristics—-their loving regard for mankind" (p. 216) ,
as well as their sexual identity and attractiveness. In a
later work, Bernardez (1985) states that the definition of
femininity, in which active or angry behavior by women is
considered inappropriate, seriously impacts the confidence of
women, who believe that there is something wrong with them
when they are angry. When a woman does feel angry, she ”sees
herself as a failure, inadequate and inferior, and her already
low self-esteem is diminished further" (Nadelson et al. ,
1982) .

Kaplan (1976) describes the dilemma a woman faces as a
"a conflict between anger clearly expressed and her sel f-image
as a woman” (p. 357). The result of this dilemma has been
labelled "spiraling phenomena" by Miller (1983, p. 3) . Small
degrees of anger are repeatedly suppressed, leading to the
woman repeatedly experiencing frustration and inaction and

resulting feelings of weakness and lack of self-esteem. These

30

feelings generate more anger and the cycle continues, with the
woman developing a distorted sense of herself and her anger,
which seems to fill her completely and which she then judges
as irrational and unwarranted. Miller et al. ( 1981) point out
that spiraling phenomena result in women experiencing their
own aggression as unorganized, destructive, and overwhelming
to their personality. Thus, women's fears and self-
condemnation regarding their own anger are increased and their
difficulty with aggression is intensified.

The second factor in the loss of self-esteem associated
with the inhibition of anger in women is the turning inward
of the anger, where it is directed against the self. Westkott
(1986) asserts that the difficulties and fears which surround
anger for women ultimately lead, as a result of reaction
formation, to the anger being deflected and turned back upon
the woman herself, resulting in self-hatred. Because anger
sabotages the ”nurturing imperative" and the idealized self,
which sustains relationships by pleasing others, it is
directed against the woman herself and thus maintains self-
contempt. Bernardez-Bonesatti (1978) maintains that women
redirect their anger against themselves rather than risk
losing their own self-esteem and the support and approval of
men.

Sex Roles as a Context
for Women's Anger and Self-Esteem
This examination of the theoretical literature on anger

in women suggests the existence of a relationship between a

31
woman's acceptance and expression of her anger and her sex
role. The construct of sex role is an important aspect of
the theoretical literature on women and anger. An examination
of recent advances in theory and research on sex roles is
necessary to understand the relationships between anger, sex
role and self—esteem hypothesized in this study.

Sex roles are generally conceptualized as ”behaviors
individuals exhibit and feel are appropriate for them by
virtue of their being male or female" (Orlofsky, 1980, p.
656). Those traits and behaviors that a culture values for
males make up the masculine sex role, while traits and
behaviors valued for females make up the feminine sex role.
Masculinity has been described as an instrumental orientation,
giving priority to getting the job done, in contrast to the
expressive orientation of femininity that gives priority to
feelings, relationships, the welfare of others, and the
harmony of the group (Parsons & Bales, 1955). Bakan (1966)
has characterized masculinity as an agentic orientation, in
which one is concerned with oneself as an individual, and
femininity as a communal orientation, in which one is
concerned with relationships between oneself and others.

Historically masculinity and femininity have been
considered to constitute a single, bipolar dimension that
differentiated males. and females (Orlofsky, 1980). This
assumption began to be challenged in the early 1970s by
researchers who proposed that masculinity and femininity

comprised two separate dimensions and that individuals could

32
possess both masculine and feminine traits (e.g. , Block, 1973:
Constantinople, 1973). This challenge was quickly followed
by the development of new measures of sex role, predominantly
the Bem Sex-Role Inventory (BSRI) (Bem, 1974) and the Personal
Attributes Questionnaire (PAQ) (Spence, Helmreich & Stapp,
1974). Empirical use of these new measures demonstrated that
masculinity and femininity are two separate, independent
dimensions and that approximately one-third of college
students described themselves as androgynous, i. e. having
equal levels of both masculine and feminine traits (Kaplan &
Bean, 1976: Orlofsky, 1980). The term androgyny began to be
used to describe the sex role characterized as an integration
of both masculinity and femininity (Bem, 1974: Block, 1973).
Before this research was undertaken, Bakan (1966) had
conceptualized the moral imperative as "to try to mitigate
agency with communion" (p. 14).

The theoretical notion that anger has been prohibited in
the feminine sex role has been supported by sex role
researchers who claim that androgyny provides a model for 1
mental health. Bem (1978) argues that an androgynous
personality represents the best of masculinity and femininity
and is necessary for fully' effective and. healthy' human
functioning. In discussing the ways in which both men and
women have been constrained by their respective sex roles, Bem
points out that women are becoming aware of their fear of

expressing their own anger.

33

Kaplan (1976) believes that androgyny provides a model
of mental health that should be the goal of psychotherapy for
women. One of the most prominent implications of the
androgynous model of mental health for women, according to
Kaplan, is that women must become comfortable with and in
control of their own anger. Kaplan argues that becoming aware
and accepting of their anger is important for women moving
toward androgyny because female children learn to dismiss or
suppress their anger in accordance with a nonaggressive
feminine sex role.

Kaplan states:

And not every woman one sees in a clinical practice

will have trouble dealing with her anger. But. . .the

literature on sex-role socialization suggests that

with women one should be especially alert to

difficulties in this area. Within an androgynous

model of mental health, problems with anger are a

major potential mental health hazard. (p. 358)

Since anger is considered to be an important element of
the androgynous model of mental health for women, it is
surprising that there is a dearth of empirical literature on
the relationship between anger and sex role in women. The
few existing studies on anger in women, and related research

exploring the relationship between sex role, self-esteem

and/or related variables, are reviewed next.

Summary
The theoretical literature on women and anger which forms
the foundation of this study is reviewed in this chapter.
Anger can be understood best when it is viewed within the

larger context of work on aggression. Anger is considered the

34
emotional component of aggression. While anger may not always
lead to aggressive behavior, it does provide a motive for
changing negative conditions.

That anger has not been considered an appropriate form
of emotional expression for women is cogently expressed in
traditional psychoanalytic theory, which. views women as
inherently passive, narcissistic and masochistic. According
to this view, the lifelong developmental task of women is
ridding themselves of aggression to attain femininity. Recent
theoretical work on female development challenges the
traditional psychoanalytic view and argues that women's
apparent passivity and lack of anger and aggression are not
inherent in their makeup but result from their subordinate
status in society.

Current theorists assert that strong cultural
prohibitions on women's anger and aggression exist because
women have traditionally filled the mothering role, have
functioned as caretakers for children and men, and have been
taught that their primary goal in life is to serve others.
The conception of anger in this culture is based on its
expression by men, who have been socialized to act
aggressively rather than express their anger directly.

The cultural prohibition on anger in women has resulted
in intrapsychic inhibitions of anger that have three sources.
First, the subordinate status of women is reinforced by
«economic, social and physical forces that make anger in women

appear pathological and threatening to their feminine

35

identity. Second, women fear that their anger will disrupt
relationships which are central to the identity they have
developed in the caretaking role assigned to them. Third,
women have internalized the perceptions of others, who see the
ability of women to act on their own behalf as threatening.

Intrapsychic inhibitions of anger in women also arise
from developmental factors in the mother-daughter
relationship. Because girls are mothered by an individual of
their own sex, they experience greater difficulty than boys
in the separation-individuation process from their mothers.
As a result, girls learn to inhibit their anger to avoid
threatening their relationships with their mothers. The
problematic separation-individuation jprocess leaves ‘women
unable to tolerate the sense of separateness and
differentiation from others that accompanies the expression
of anger. In addition, women have irrational fears of their
own destructiveness arising from fears of identifying with
their own projected rage and aggression which has been
incorporated in their internalized image of their mothers.

There are serious consequences for women's mental health
which result from the inhibition of their anger. Women
experience a severe loss of self-esteem when they become angry
because the anger conflicts with the cultural ideal for women
and thus threatens their identityu The anger is turned inward
and directed against the self, which results in self-hatred

and loss of self-esteem.

36

Recent developments in sex role theory suggest that the
prohibition on, and intrapsychic inhibition of , anger in women
is an element of the feminine sex role which prescribes
culturally-valued female characteristics. The
conceptualization of masculinity and femininity has shifted
from the historical view of a single, bipolar dimension
differentiating males and females to two separate dimensions,
both of which may be present in both males and females. The
androgynous sex role, in which masculinity and femininity are
integrated within individuals of either sex, is considered by
sex role researchers to provide a model for mental health.
It is argued that for women, the androgynous model
necessitates becoming aware and accepting of, and able to

express, their own anger.

CHAPTER III

REVIEW OF THE EMPIRICAL LITERATURE

The developmental theory reviewed in Chapter III posits
a relationship between the prescribed sex role for women, the
awareness and expression of anger by women, and their level
of self-esteem.

In reviewing the existing literature, no studies were
found of the relationship between anger, sex role, and self-
esteem, the three variables discussed in the theoretical
literature. Only' a small number of studies addressing
questions regarding women's anger were found. The findings
and conclusions of researchers who have studied this question
are reviewed below.

Several investigations have been conducted of the
relationship between sex role and self-esteem, and some have
included other related variables. A small number of studies
have explored the relationship between sex role and variables
related to psychological adjustment which may have some
bearing on one's awareness of and attitude toward their own
anger. The existing literature in these areas is reviewed
following the examination of studies of anger in women.

Because aggression and anger are intimately linked, it
is important to provide a context for a study of anger in
women by examining the literature on aggression. To provide

the background necessary for understanding the literature on

37

38
anger, several major studies of sex differences in aggression
are summarized below. Also included in this section are
studies ‘which. demonstrate that. both aggressive and
nonaggressive responses can be conditioned and can provide
cathartic effects. Two studies which address the relationship
between aggression and sex role are included. The literature
on aggression is summarized prior to reviewing the literature
on anger in women and the relationship between sex role and
self-esteem, as ‘well as studies including' other' related

variables.

Studies on Sex Differences in Aggression

There appears to be widespread agreement that females
display significantly less aggression than males (e.g.,
Bardwick, 1971; Maccoby & Jacklin, 1974: Terman & Tyler,
1954). Several investigators, however, have questioned
whether such. differences are ‘the result of’ differential
socialization of males and females or are the result of innate
biological differences between the sexes.

~There is considerable debate as to whether males have a
greater biological predisposition to aggression than females.
For example, on the basis of an extensive review of the
literature regarding psychological differences between the
sexes, Maccoby and.Jacklin (1974) concluded that, in general,
there is little difference between parental treatment of sons

and daughters. In the area of aggression, they claim there

39
is no support for the belief that aggressive behavior is
reinforced for boys but not for girls. It may be, instead,
that boys are punished ‘more than girls for aggressive
behavior. Maccoby and. Jacklin argue that the greater
propensity for aggressive behavior among males is due to a
biological predisposition rather than resulting from a learned
fear of aggression among girls, or from a tendency for girls
to reinforce aggression in boys.

In refuting the argument that girls have a learned fear
of aggression, Maccoby and Jacklin (1974) note that the
studies they reviewed did indicate that boys do more
spontaneous copying of modeled aggressive behaviors than girls
do. Previous researchers have attributed this difference to
aggressive responses in girls being inhibited due to fear
resulting from negative socialization experiences. Maccoby
and Jacklin reject this explanation and argue that there are
differences between the sexes in the acquisition of aggressive
behaviors. To support this argument, they cite findings that
girls do not notice and retain the details of modeled
aggression to the same extent that boys do.

Maccoby and.Jacklin (1974) also reject the argument that
boys are more aggressive because their aggressive behaviors
are reinforced by passivity in girls. They state that the
studies they reviewed found that girls provided the same
contingencies for the aggressor's behavior as boys and "were

no more and no less likely than boys to cry or yield in

40
response to an aggressive attack" (p. 240). Interestingly,
these studies also showed that.women and girls were the object
of aggression less often than men and boys.

Maccoby and Jacklin (1974) state that the findings of
the studies they reviewed were conflicting regarding the
question of sex differences in preferences for verbal or
physical forms of aggression. They conclude that "if there
is a sex difference in the forms aggression takes, the verbal-
physical distinction does not accurately describe the
difference" (p. 235).

Maccoby and Jacklin (1974) argue that males have a
greater’biological predisposition to aggression than females,
which results in higher levels of preparedness to learn
aggressive behaviors. They base their argument in favor of
the biological foundation of sex differences in aggression on
these four factors:

(a) In all societies for which evidence is available,

males are more aggressive than females.

(b) Sex differences in aggression appear early in

life, when differential socialization pressures
do not appear to exist. Maccoby and Jacklin
reviewed many studies of parent-child
interactions which did not show that parents
are more permissive toward aggression in boys

than in girls.

41

(c) Similar sex differences are found in subhuman
primates.

(d) Aggression is related to levels of sex hormones and
can be changed by experimental administration of
hormones.

Block (1978) studied the literature reviewed by Maccoby
and Jacklin (1974) and concluded that their conclusions were
unwarranted. Block argues that existing theories of parent-
child socialization do not permit the development of
hypotheses which are sufficiently differentiated to detect
important sex differences in socialization. She also claims
that methodological problems in the studies reviewed by
Maccoby and Jacklin prevent the finding of significant
differences which may actually exist. Block maintains that
the questions posed by Maccoby and Jacklin could not be
answered on the basis of existing research.

To address the question of sex differences in parent-
child socialization practices, Block (1978) administered a
standardized instrument on child-rearing practices to 696
mothers, 548 fathers, and 1,227 young adults in 17
independent samples. This study found major differences in
parental treatment of male and female children. In the area
of aggression, Block found that there was a greater
discouragement of aggression in girls than in boys by their
parents. Both mothers and fathers expected ladylike behavior

from their daughters and discouraged them more than their

42
sons from rough-and-tumble games and fighting. For male
children, in contrast, aggression was encouraged through
competition, greater acceptance of participation in rough
games and fighting, and tolerance of teasing and expressions
of sibling rivalry.

Chodorow ( 1978) has also taken issue with the
conclusions of Maccoby and Jacklin (1974). She argues that
the findings of the studies which they reviewed are
inconsistent and could support almost any hypothesis
regarding differences in treatment of male and female
children. Chodorow also states that the studies which did
find differences between the treatment of boys and girls
consistently reported differences in the same direction.
Chodorow cites the proceedings of a panel on the psychology
of women of the American Psychoanalytic Association, which
reported growing evidence that mothers show differences in
attitude and handling toward male and female children from
early in the child's life (Galenson, 1976). Chodorow
contends that there are subtle differences, consisting of
nuance, tone, and quality, in maternal treatment of boys and
girls, with different developmental results for the sexes.

Whether there are innate biological differences that
result in greater aggression in males than in females has not
been definitively answered. Several researchers have
concluded, however, that regardless of whether biological

differences are present, different socialization practices

43
for male and female children appear to play a major role in
the lower levels of aggressive behavior in females than in
males. Three of these studies are summarized here.

Kagan and. Moss (1962) concluded that socialization
appears to be the major factor in determining differences in
aggression between males and females. In a major
longitudinal study of the continuity between childhood and
adult behavior, and the influences on the goals and behaviors
of children, they found that female children displayed
greater conflict over aggression than male children. In
adulthood, aggressive behavior was more stable for’men, while
dependency was more stable for women. Kagan and Moss also
state that the female child's conflict over aggression stays
with her into adulthood in the form of repression of anger
and inhibition of aggression.

Kagan and Moss (1962) maintain that girls are subject to

more severe socialization of aggression than boys as the

T result of two forces: rewards and punishments from parents,

teachers and friends: and the influence of the cultural
definition of the ideal female. They emphasize the role of
socialization in the difference between males and females in
aggression and dependency, concluding that sex role
identification plays a central role in the selection and
maintenance of aggression, passivity and dependency. Kagan

and Moss also suggest, however, that constitutional

44
variables, which are supported by different cultural
expectations for the sexes, may play a role.

On the basis of their review of the experimental
literature on aggression in adult males and females, Frodi,
Macaulay and Theme (1977) assert that aggressive behavior is
shaped primarily by social forces, which may either
strengthen or weaken any existing biological differences.
They state that there was not sufficient evidence in existing
studies to support any hypothesis regarding sex differences
in preferences for either verbal or physical modes of
aggression. The sex differences in aggression that can be
predicted by sex role stereotypes were found only in self-
reports of general hostility or aggressiveness.

Frodi et al. (1977) found that in studies of a single
physical or verbal aggressive response, women did not show
lower tendencies than men toward physical or verbal
aggression. They also did not find that women demonstrated
more indirect, prosocial, or displaced aggression than men.
They did find, however, that women have a tendency to avoid
physical and face-to-face verbal aggression.

Frodi et al. (1977) did find some differences in
aggression between men and women which they believe are
accounted for by three factors consistent with sex role
expectations. First, they found that the sex of the
instigator or the victim appears to be a factor in behavior

which conforms to sex role guidelines. For example, women

45

who received feedback on the pain caused by their aggressive
actions, displayed less aggression toward female victims than
either of the following two groups:

(a) women who did not receive feedback

(b) women or men with male victims, regardless of

whether they received feedback.
Second, women were found to be more anxious and guilty about
aggression and, as a result, to be more defensive against
perceptions of aggression or to avoid or inhibit their
expression of aggression. Less evidence was found to support
the third possible explanation for women being less
aggressive than. men--that women are more likely to be
empathic with victims of aggression, and thus less likely to

act aggressively.

Based on existing data which they reviewed, Frodi and a

her associates (1977) concluded that aggression is a trait
possessed by both men and women. Sex role expectations
influence behavior and are likely to account for some of the
sex differences found in adulthood, but can be negated by
either situational factors or relearning.

A subsequent investigation by Frodi (1977) suggests that
differences between men and women in perceptions of
experimental manipulations may confound findings of sex
differences in aggressive behavior. Following the review of
the literature on sex differences in aggression with her

colleagues (Frodi et a1. 1977), Frodi conducted a survey to

46
investigate whether what constitutes provocation differs for
men and women. She hypothesized that exposure to an
aggressive event may have different consequences for men and
women, and that a provocation to aggression may be perceived
differently by the two sexes.

In this investigation (Frodi, 1977), 60 male and 70
female psychology students were asked seven open-ended
questions. The results indicated that there are differences
in what men and women perceive as provocation. Male subjects
reported that they were most angered by physical or verbal
aggression by another male, and by a condescending attitude
on the part of a woman. For women, the most anger-provoking
behavior was being treated in a condescending manner,
regardless of the provoker's sex. On the basis of these
results, Frodi points out that the use of identical
provocations for men and women may be a problem in
experimental studies of aggression. Observed sex differences
in aggression may result not from differences in aggressive
behavior, but from different emotional states mediating the
overt behavior.

Following these findings of sex differences in what
constitutes provocation to aggression, Frodi (1978)
investigated sex differences in responses to sex—appropriate
provocations devised on the basis of the above results. In
a study of 48 male and female college students, subjects

given sex-appropriate provocations were compared to non-

47
provoked subjects on heart rate, blood pressure, and
electrodermal responses as physiological measures of
provocation. A stream of consciousness technique was also
given to half the subjects to examine emotional and cognitive
processes associated with response to provocation.

Subjects in this study were paired with a confederate,
who talked to the subject in either a negative (provocation
condition) or neutral (no provocation condition) manner.
Provocations consisted of directly hostile remarks for male
subjects and the expression of a condescending attitude for
female subjects. The dependent aggression measure was the
delivery by the subject of noxious sounds to the confederate
in a subsequent learning trial (Frodi, 1978).

Frodi (1978) found that when given sex-appropriate
provocations to aggression, there were no significant
differences between males and females in their responses,
"suggesting that provoked men and women alike were more
angry, disliked their partner more, and aggressed more
against their partner than did their non-provoked
counterparts" (p. 347). There were also no significant sex
differences in physiological measures among the provoked
subjects.

The stream of consciousness data indicated that the
provoked men were more angry, were more negative toward their
partner, and became more behaviorally aggressive than the

'provoked. women, who 'were preoccupied. with nonaggressive

48

thoughts in coping with their anger. This difference between
the sexes was not found among provoked subjects who were not
given the stream of consciousness technique. Frodi concludes
that this sex difference was the result of the verbalization
of feelings and thoughts. She also states that this
difference between men and women suggests that women may
perceive aggression less readily, or may wish to give the
provoker the benefit of the doubt (Frodi, 1978).

Bardwick (1971) reviewed the literature which suggests
that males are more aggressive than females and concluded
that, while females may have lower levels of aggression,
males and females may have similar aggressive needs.
Bardwick argues that women possess greater hostility and
aggression lthan can be demonstrated by experimental
procedures. She supports her contention by citing a study
in which women displayed aggressive behavior when the
aggressor and victim could not see each other (Rapoport and
Chammah, cited in Bardwick, 1971) . Aggression can be
expressed more directly and without guilt by males, and
females feel more guilt and conflict and inhibit the direct
expression of aggression. Bardwick believes that true levels
of aggression should be measured by examining not only
fighting, hitting and biting, but also verbal aggression,
interpersonal rejection, academic competitiveness, gossip,

deviation from sexual standards, passive aggression, and

49
manipulation with power, withdrawal, tears and somatic
complaints.

A series of experiments which suggest that the lack of
aggressive behavior in women may be the result of
instrumental learning has been conducted by Hokanson and his
associates (Hokanson 8 Edelman, 1966; Hokanson, Willers, 8
Koropsak, 1968). The results of these studies support the
contention of many investigators regarding the centrality of
socialization in determining sex differences in aggression.
The studies were designed to test the catharsis hypothesis
"that the carrying out of aggressive behavior, either
directly towards a frustrator or in a displaced fashion
towards a substitute target, is physically tension-reducing:
and, that this overt aggression serves, in a sense, to ‘drain
the person's reservoir of aggressive motivation'” (Hokanson,
1970, p. 74).

The earlier study (Hokanson 8 Edelman, 1966) compared
the effects of aggressive and nonaggressive counter-responses
to aggression on systolic blood pressure, which was used as
a measure of tension, in 12 male and 16 female undergraduate
volunteers between the ages of 18 and 24. The subjects had
the option of responding to electric shock which they
received by giving, in response, a shock, a reward, or no
response. For the male subjects, Hokanson and Edelman found
that a shock counter-response was apparently cathartic since

the subject's blood pressure dropped significantly following

50

this counter-response, but not after giving either a reward
or no response. The results for females, however, were
dramatically different. The catharsis effect of lowered
blood pressure was observed when females gave a friendly
counter-response, i.e. reward, but not when they gave either
an aggressive counter-response or no response. The
investigators hypothesized that this difference could be due
to females not being trained in or rewarded by aggression,
with the result that aggression for females is not
instrumental in terminating noxious events.

An experiment was then designed to test whether
aggressive counter-responses would be cathartic for females
and friendly counter-responses would be cathartic for males
under conditions in which these responses were reinforced
(Hokanson et al., 1968). For the ten female volunteers,
reward and. punishment contingencies which favored the
development of aggressive counter-responses were used. For
the 11 male volunteers, the contingencies favored the
friendly counter-responses. Hokanson and his associates
found that the use of aggressive counter-responses by the
females increased significantly and demonstrated a cathartic
effect in lowered blood pressure. For the males, use of
friendly counter-responses increased only negligibly but did
show the cathartic effect of lowered blood pressure.

In evaluating the results of these studies, Hokanson

(1970) proposed an alternative to the catharsis hypothesis

51

based on a learning model, in which individuals learn ways of
behaving which reduce aggression from others. These learned
behaviors result in a sense of relief, accompanied by
physical relaxation, when the aggression ceases.
Significantly, Hokanson points out that women respond to
aggression with friendliness because friendly behavior has
been effective in terminating hostility from others.

The obvious implication of the work of Hokanson and his
associates is that both aggression and nonaggression can be
learned by both males and females. These studies appear to
provide support for the view that aggression is learned and
rewarded as part of the masculine sex role, while
nonaggression is learned and rewarded as part of the feminine
sex role.

Two additional studies are noteworthy since their
findings are relevant to the relationship between aggression
and sex role. In a study involving only female subjects,
Richardson, Vinsel and Taylor (1980) investigated the
relationship between either liberal or traditional beliefs
regarding the role of women and aggressive behaviors. In a
laboratory setting, they studied the aggressive behavior of
40 female introductory psychology students, half of whom had
liberal attitudes toward the role of women and half of whom
had traditional attitudes. The female subjects were paired
with male opponents and were instructed that they could avoid

being shocked by the opponent if they were faster than the

52
opponent in delivering shock. They found that women with
traditional attitudes were more aggressive in shocking their
opponents than those with liberal attitudes.

The findings of Hoppe (1978) more directly address the
question of the relationship between aggression and sex role.
In a study of 96 college undergraduates, she investigated
interpersonal aggression as a function of subject's sex,
subject's sex role, opponent's sex, and degree of
provocation. Half the subjects of each sex were told their
opponent was male, and half were told their opponent was
female. In a purported test of reaction time, the degree of
shock subjects believed they were administering to their
opponents was compared in unprovoked and.provoked conditions.
In both conditions, masculine subjects (male and female) with
male opponents administered significantly higher levels of
shock than any of the other groups, which did not differ from
each other. Subjects in all groups increased the level of
shock in response to increasing provocation from their
opponents.

Hoppe (1978) points out that these results do not b
support predictions of traditional sex roles that males will
be more aggressive than females. Only those males
identifying with the masculine sex role were more aggressive
than other males or females, while a high level of aggression
was also found among masculine females. Hoppe concludes that

masculine females do not differ from masculine males when

53

aggressing against a male opponent. She also suggests that
the lack of aggressive tendencies in androgynous individuals
may result from balancing masculine characteristics with
feminine qualities. She hypothesizes that aggression anxiety
may mediate responses in all sex roles except the masculine,
whose instrumental and agentic qualities may be antithetical
to such anxiety. The expressive, communal domain of
femininity, however, may facilitate anxious reactions.

It is clear that no definitive conclusions can be drawn
regarding the debate over biology vs. socialization in
determining sex differences in aggression. Maccoby and
Jacklin's (1974) conclusions that the difference has a
biological basis has been challenged by Block (1978) and
Chodorow (1978). Kagan and Moss (1962) and Frodi et al.
(1977) argue that socialization plays a major role in the
development of aggressive behaviors. Studies by Frodi (1977,
1978) have demonstrated that the provocation to aggression
differs for males and females and that, when appropriately
provoked, men and women display similar levels of aggressive
behaviors. Findings by Hokanson and his colleagues (Hokanson
8 Edelman, 1966: Hokanson, Willers, 8 Koropsak, 1968) provide
evidence that socialization is a major factor in determining
sex differences in aggression. Finally, two investigations
suggest a strong link between sex roles and differences in

aggression (Richardson et al., 1980: Hoppe, 1978).

54

Regardless of whether biological factors contribute to
sex differences, in aggression, it is clear that most
investigators in this field believe that socialization
practices play a determining role in the higher levels of
aggression displayed by males. Their conclusions lend
crucial support to theoretical propositions that strong
cultural prohibitions, expressed in the feminine sex role,
exist and result in prohibitions against the experience and

expression of aggression and anger by women.

Studies on Anger in Women

In reviewing the literature on women and anger, studies
involving hostility were also reviewed. Although hostility
is not equivalent to anger, it is a closely related state
which occurs as a result of the arousal of anger. Buss
(1961) defines hostility as an enduring attitudinal response
involving negative feelings and evaluations of people and
events. According to Buss, hostility occurs when stimuli are
identified and categorized during the arousal of anger, with
the labeling and identification enduring after anger
subsides. Hostility may be considered a conditioned anger
response.

Hostility has also been defined as an attitudinal set,
or even a personality trait, developing from a lack of trust
in the goodness of others and centered on the belief that

others are "generally mean, selfish and undependable"

55

(Williams, Barefoot, 8 Shekelle, 1985, p.173). These
researchers clearly differentiate hostility from anger, which
they define as "an emotional state made up of feelings
ranging in intensity from minor irritation to fury and rage."
They point out that anger and hostility are clearly related
and that individuals with hostile attitudes are likely to
experience anger more frequently and intensively than these
low in hostility.

Only a small number of studies involving either anger or
hostility in women were found. These studies are considered
in three groups. First, those studies which suggest that
anger or hostility is incompatible with the traditional
feminine sex role are reviewed. Second, two studies which
raise questions about the relationship between anger or
hostility, age, and sex role are examined. Finally, a single
study of the psychological settings associated with anger is
summarized. This section concludes with a description of
findings that do not appear consistent with the theorized
prohibition against anger in the feminine sex role.

A total of eight studies were found which lend support
to the theoretical proposition that anger is prohibited by,
and incompatible with, the feminine sex role. Four of these
studies address this issue rather directLy. The remaining
four studies may be regarded as suggesting support for this
theoretical proposition through examining the following

relationships: anger arousal and irrational beliefs;

56
hostility and educational level: hostility and career vs.
family-role commitment: and dogmatism, hostility, aggression,
and social desirability. The four studies which are more
direct in approach. are reviewed first, followed. by the
remaining four studies.

The four studies which provide the most direct
examination.of the incompatibility of anger’with the feminine
sex role took very different approaches. Biaggio (1988) has
studied sex differences in anger provocation and experience
in a field study in which 72 college students recorded all
incidents which aroused anger and their behavioral reactions
over a two-week period. Men reported more anger-arousing
events and more reactions of physical and verbal antagonism
than women. Conversely, the women reported.more incidents of
criticism or rejection, and of feeling hurt, and responded
with more reactions of passive consent.

Biaggio (1988) also conducted two laboratory studies of
anger provocation and experience. In the first investiga-
tion, 51 college students reported the extent to which they
experienced anger in response to imagined anger-provoking
scenes. The only sex difference found involved males
reporting more hostility and hate than females in response to
a criticism incident, while females reported greater hurt.
In the second laboratory study, 101 subjects reported their

behavioral responses after being exposed to an insulting

57
letter. No sex differences were found in degree of anger
reported.

Biaggio (1988) has pointed out that the difference in
research designs may help explain the findings of sex
differences in anger in the field studies, and the lack of
such findings in the laboratory studies. She states that the
pressures of real life present in the field studies may
reflect a greater likelihood for men to respond with anger,
to be subjected to anger-provoking situations, and/or to have
a lower threshold for recognizing aggression. The field
studies also may demonstrate that women are more likely to
inhibit anger when provoked in real life. In the absence of
real life pressures in the laboratory, Biaggio believes that
women may have viewed anger as a reasonable or expected
response.

The relationship between sex role and hostility guilt
has been investigated by Evans (1984) in 101 males and 135
female college undergraduates. Females reported
significantly more guilt over hostility than males.
Masculine sex-typed.males reported feeling significantly less
guilt over hostility than feminine sex-typed males, and
androgynous males fell between the two sex-typed groups.. For
the females, there were no differences between the sex role
groups on guilt over hostility. Although the difference was
nonsignificant, feminine sex-typed females did score higher

than other females on hostility guilt. Evans' study provides

58
support for the theoretical predictions that socialization in
the feminine caretaking role has resulted in women perceiving
anger as incompatible with their identity as females.

Holahan and Spence (1980) compared the relationship
between socially desirable and undesirable aspects of
masculinity and femininity and anger in 60 male and 94 female
college counseling clients. Undesirable aspects of
masculinity were measured by the following items on the M
scale: tarrogant, boastful, egotistical, greedy, dictatorial,
cynical, looks out only for self, and hostile. Undesirable
aspects of femininity consisted of two components, one
indicating lack of sense of self (Fc-) and one indicating
verbal aggressive qualities (Fva-) . Significant positive
relationships between anger and undesirable aspects of both
masculinity and femininity were found among females. There
were no significant relationships between anger and desirable
aspects of either masculinity or femininity for females. For
the males in the study, no significant relationships were
found between anger and either desirable or undesirable
aspects of masculinity and femininity. These results also\\
lend support to theoretical notions that women have learned i
to suppress anger and regard it as incompatible with their /
role and identity as women.

Atkinson and Polivy (1976) studied the effects of delay,
attack, and retaliation on depression and hostility. Forty-

three male and 43 female undergraduates were subjected to a

59

waiting condition and then received either a verbal attack
(high anger condition) or an apology (low anger condition).
Half the subjects in each group were given an opportunity to
retaliate through a written evaluation of the experimenter.
They found that subjects in the high-anger condition became
more hostile and more depressed than those in the low anger
condition, and devalued the experimenter more when given an
opportunity to retaliate. This retaliation, however, did not
decrease either the anger or depression.

In terms of sex differences, Atkinson and Polivy (1976)
found that males and females felt the same degree of
hostility after being attacked but males were more likely to
express their hostility outwardly. Females, however, were
more likely to retaliate against the experimenter, but felt
less hostile than males when they received an apology.
Females were also more likely to become depressed in the high
anger condition. In commenting on these findings, the
researchers hypothesize that a more direct physical
expression of hostility may be required to alleviate the
discomfort. If so, females may not feel free to express
themselves in this manner. Males, however, are not subject
to the same social restrictions and thus need not internalize
their anger.

Biaggio's (1988) work illustrates the greater tendency
of women to inhibit their anger, while men are more likely to

respond with angry behavior. The study by Atkinson and

60
Polivy (1976) demonstrates that women are as likely as men to
feel angry, but less likely to express their angry feelings
outwardly. Both the study by Evans (1984) and the work of
Holahan and Spence (1980) suggest that anger may be
associated with negative aspects of the self-concept for
women, regardless of their sex roles. These findings
obviously have significant implications for the proposition
that anger is prohibited by, and incompatible with, the
feminine sex role. Although their findings are important,
they do not provide the direct examination of the
relationship between anger, sex role and self-esteem that is
posited in the theoretical literature. Four studies of anger
or hostility in women were found whose conclusions, when
examined in depth, suggest support for the prohibition of
anger in the traditional feminine sex role. Hazaleus and
Deffenbacher (1985) studied the relationship between specific
irrational beliefs and anger arousal in 113 male and 229
female introductory psychology students. No differences were
found between the sexes in anger arousal, which Hazaleus and
Deffenbacher suggest demonstrates the necessity of separating
concepts of anger and aggression since studies have
consistently found men to be more aggressive than women. The
researchers also state that this finding suggests that anger
arousal in women may be followed by behaviors other than
aggression since aggression is {discouraged in role

expectations for women.

61

The question of the incompatibility of anger with the
feminine sex role is raised in Hazaleus and Deffenbacher‘s
(1985) findings regarding the irrational beliefs that
discriminate between women in high and low anger groups. The
four beliefs that discriminated between the two groups were
anxious overconcern, emotional irresponsibility,
helplessness, and dependency, which was inversely related to
high anger. Hazaleus and Deffenbacher conclude that the
finding that women high in anger arousal endorse the need to
be dependent on others less often than those low in anger
arousal suggests that high anger may be inconsistent with
traditional sex role socialization.

Martin and Light (1984) examined the differences in
anxiety, depression, hostility, perceived control over one's
own life, and attitudes toward autonomy in women with varying
levels of education ranging in age from 21 to over 63. In
their sample of 416 midwestern women who were members of the
National Organization for Women, the Catholic Guild, home
economics teachers, or nurses, they found significant
differences between hostility scores of women with varying
educational levels. Hostility was highest among women with
a high school education. Hostility decreased as educational
level increased, up to postgraduate educational levels, when
women with advanced degrees were higher on hostility than
women with four year degrees. Martin and Light also found

that the women's attitudes toward autonomy became

62

increasingly liberal with higher levels of education. In
discussing this finding they suggest that ”college education,
and particularly advanced degrees, can be assumed to be a
non-traditional role for women" (p. 404). When considered
from this perspective, it appears that higher levels of
hostility may be associated with nontraditional roles, i.e.,
either masculine or androgynous sex roles, for women.

Martin and Light's (1984) findings regarding the
relationship between hostility and educational level were not
supported in another investigation of 760 midwestern farm
women with an average age of 44 (Hertsgaard 8 Light, 1984).
In this study, hostility was not related to the women's
educational level.

In a study of 284 midwestern women ranging in age from
21 to 63 who were employed full-time, Light (1984) examined
levels of anxiety, depression, and hostility according to
their perceived career and family-role commitment. She found
that women who stated that they placed their career over
their families scored highest on all three variables, while
those who reported that their families had priority over
their careers scored lowest on all three variables. Women
who gave their families and careers equal priority scored
between these two groups. Light argues that her findings
support the hypothesis that deviating from the traditional
feminine sex role results in emotional turmoil and stress for

women. It is also apparent from these findings that women in

63
less traditional roles report feeling greater hostility than
women in more traditional, feminine sex roles.

Heyman (1977) investigated the relationships between
dogmatism, hostility, aggression, and social desirability in
74 male and 109 female college students. He found that
increasing dogmatism was accompanied by increasing feelings
of hostility in both males and females, and by increasing
aggression only in males. He also found that, for females,
both hostility and aggression were inversely related to
social desirability, while only aggression showed this
relationship for males. Heyman believes that these findings
suggest that hostility and aggression are more readily
integrated with the male role than the female role since both
are alien to the prescribed female role.

Although the number of studies is small, these four
studies demonstrated consistent findings of higher levels of
hostility, anger, and/or aggression in women whose
circumstances suggest that they have moved out of a
traditional feminine sex role into a less traditional,
possibly masculine or androgynous, role. The findings are
consistent with both college students (Hazaleus 8
Deffenbacher, 1985; Heyman, 1977) and women ranging in age
from 21 to over 63, thus representing many levels of
development in adulthood (Martin 8 Light, 1984: Light, 1984).

This conclusion, however, is based on an extrapolation

of life circumstances such as priority given to career vs.

64
family (Light, 1984) or educational level (Martin 8 Light,
1984) rather than on an actual measurement of sex role
identification. The conclusion is, therefore, tentative and
needs to be tested in further research on the relationship
between anger and sex role in women.

Contradictory findings of the relationship between anger
or hostility and age have been reported in two studies.
These findings are instructive, however, when examined fer
their implications for the relationship between anger or
hostility and the feminine sex role. In a study of 760
midwestern farm women with an average age of 44, Hertsgaard
and light (1984) found that younger women scored higher on
hostility than older women. In a ten year longitudinal study
of 60 middle-class divorcing families, however, Wallerstein
(1986) found that anger increased incrementally with age for
the divorced women. Specifically, of the women who were at
least 34 years old. at the time of divorce, over* half
continued to be angry ten years later. In contrast, of women
who were in their twenties at the time of divorce, less than
15 percent continued to be angry ten years later.

In both of these studies, anger was also associated with
factors other than age in the women's lives. An examination
of these additional factors suggests a tentative hypothesis
regarding the relationship between the women's level of anger
or hostility, their age, and their sex role. For the

midwestern farm women, higher levels of hostility were also

65

associated with having more than two children under the age
of 14, lack of involvement in decision-making regarding the
farming operation, and visiting friends less than once a
month (Hertsgaard 8 Light, 1984). These factors suggest that
the women were involved in a more traditional feminine sex
role. The younger women in this study had higher levels of
hostility, and it can be conjectured that they had both
greater exposure to ideas about more flexible roles for
women, and greater openness to these ideas, than older women.
It may be that their hostility expresses their
dissatisfaction with their traditional roles and their
preference for less traditional and.more flexible androgynous
roles.

Among the divorced women, anger also occurred in
conjunction with loneliness, being a single parent, anxiety
about living alone, and economic worries (Wallerstein, 1986).
These factors can be viewed as indicating discomfort with
displacement from the traditional feminine sex role. It is
interesting to consider this possibility in the light of the
findings that more older women than younger women reported
feeling angry, and that anger was more prevalent among women
who did not remarry. It can be hypothesized that the older
women were more angry as a result of being in less
traditional sex roles for which they may not have been

prepared adequately as a result of less opportunity for

66
exposure to ideas about more flexible roles for women, and/or
less openness to such ideas.

A single study which explored the psychological settings
associated with a variety of emotions, including anger, was
found (Gormly 8 Gormly, 1984). The subjects, 32 college
women between the ages of 18 and 25, described their
emotional experiences on a daily basis for three weeks. At
the end of that time, they were interviewed by the
investigators and asked to describe events, feelings, and
behavior which preceded and followed their experience of
particular emotions. The situations which were reported to
be associated with anger, in order of decreasing frequency,
were as follows:

1. The perception that her expectations for success
were not fulfilled.

2. The perception that she was being treated
disrespectfully by someone who was significant to her.

3. The perception of social rejections.

4. The perception that someone was disagreeing with
her.

Gormly and Gormly (1984) state that their results
suggest that a psychological situation in which one perceives
that an expected and desired outcome is not attained is "the
sufficient condition for experiencing anger" (p. 76). The
subjects in this study also reported that the behavior in

which they engaged.most frequently while angry was "screaming

67
and hollering" (p. 77). In addition, they reported three
acts of physical aggression while they were angry. Gormly
and Gormly also found that while anger was the focus of the
emotional state labeled anger, other emotions occurred along
with it, most prominently frustration, guilt, disappointment,
and anxiety.

This small number of studies of anger in women appears
to provide some support for the hypothesized relationship
between sex role and anger: that anger has been prohibited
in the traditional feminine sex role. The most direct
support is evidenced in findings that females feel
significantly more guilty over hostility than males (Evans,
1984), anger is significantly related to negative aspects of
both. masculinity and femininity for females (Holahan 8
Spence, 1980), that women are less likely than men to react
with anger when provoked (Biaggio, 1988), or to express anger
they feel at the same intensity as men (Atkinson 8 Polivy
1976) . Some support for the hypothesized relationship is
found in the inverse relationship between high anger arousal
and the need to be dependent (Hazaleus 8 Deffenbacher, 1985);
the higher levels of hostility in women with postgraduate
levels of education (Martin 8 Light, 1984) and in women with
greater commitment to their careers than to their family role
(Light, 1984): and in Heyman's (1977) findings that hostility
and aggression are not readily integrated into the female

role.

68

Less direct support for the theoretical notion that
anger is prohibited in the feminine sex role is demonstrated
in findings that younger farm women report more hostility
than older farm women (Hertsgaard 8 light, 1984), and that
more older divorced women report being angry than younger
divorced women (Wallerstein, 1986).

The conclusion that these findings provide support for
the prohibition against anger in the feminine sex role must
be tempered, however, by findings which do not appear to be
consistent with those above. Three findings directly related
to studies reviewed above will be discussed first. This
section then concludes with a review of three additional
studies whose findings seem not to support the hypothesized
prohibition of anger in the feminine sin their
study of farm.women, Martin and Light (1984) did not find the
relationship between hostility and educational level which
lends support to the hypothesis regarding sex role and anger.
More significantly, the finding that the most common behavior
of women while they are angry is "screaming and hollering"
indicates that many women do feel and express anger (Gormly
8 Gormly, 1984).

On the basis of the finding of Atkinson and Polivy
(1976) that females become more depressed than males after
being provoked to anger, Biaggio (1980a) hypothesized that
females who were low in anger arousal would score lower on

self-acceptance and sense of well-being than low-arousal

69

males. She studied anger arousal and personality
characteristics in 72 male and 78 female introductory
psychology students. The results of this study did not
support the hypothesis above regarding low anger females. In
addition, the hypothesis that low anger subjects of both
sexes would score lower on dominance and sense of well-being,
and higher on femininity, was not confirmed. Biaggio found
instead that the low anger subjects scored higher on well-
being. She suggests that a low anger arousal score may be
evidence not of repression of anger, but of a high threshold
for anger arousal.

The review of studies on anger in women concludes with
three studies of sex differences in anger whose findings do
not appear to support the hypothesis regarding anger and sex
role. Sex differences in emotionality have been investigated
in two studies. .Allen and. Haccoun. (1976) studied. sex
differences in anger, fear, joy, and sadness in the areas of
covert responsiveness, interpersonal expressiveness,
orientation (attitudes toward responses and expressions), and
situational determinants. In a sample of 61 male and 61
female undergraduate psychology students, they found that in
covert responsiveness, females reported a greater overall
experience of emotion but not in the experience of anger.
Females also scored higher than males on the interpersonal
expressiveness of all four emotions: the smallest sex

difference, however, was for anger. There was no significant

70
sex difference in orientation toward anger. Females gave
more interpersonal responses in ‘the area of situational
determinants of anger.

Allen and Haccoun (1976) conclude that of the four
emotions studied, the sexes are least different in anger,
which is likely to be associated with the agentic qualities
of masculinity. They point out that although it has been
documented that males are more aggressive than females,
expressing anger is not the same as intentionally harmful
aggression. Furthermore, some researchers have found no
difference between the sexes in verbal aggression.

A finding of Balswick and Avertt (1977) in a study of
523 social science undergraduates is relevant to the findings
of Allen and Haccoun (1976) regarding sex differences in
anger. Balswick and Avertt studied gender, interpersonal
orientation, and perceived parental expressiveness as
contributing factors to sex differences in emotional
expressiveness. They measured emotional expressiveness on
four scales: hate, love, sadness, and happiness. Anger was
a component of the hate scale, along with hate, resentment,
and rage. The rationale upon which their study was based did
not result in an expectation that there would be sex
differences in the expression of hate. The hate scale was,
therefore, considered irrelevant. They did report finding no

difference between males and females in this area.

71

In a series of four studies of the everyday experience
of anger, Averill (1982) found only' a small number' of
differences between the sexes. His findings are based on
written surveys in which 288 students or community residents
described either experiences of their own anger or as the
target of another person's anger, or kept records of their
experiences of anger and annoyance. The sample contained
equal numbers of men and women.

In Study I, which examined the subject's own experience
of anger, significant differences between the sexes were
found in nine of the 87 tests performed. These differences
occurred in four areas. First, women described their anger
as more intense and more out of proportion to the situation
than the men did. Second, women reported more often than men
that they would like to talk the incident over with the
instigator or a neutral party, and that they would deny the
instigator some customary benefit. Third, women were more
likely to cry (38% vs. 8%), have a shaky, cracking voice and
experience greater tension than men. Finally, Averill found
that men were more likely to be the targets of anger than
women.

These differences were largely confirmed in Study III,
which found differences between the sexes in the same areas
and the same direction as in Study 1. The only exception was
that men reported that their anger was more intense than the

intensity reported by the women (Averill, 1982).

72

In Study II significant differences were found between
the sexes in the experience of another person's anger. More
women than men (80% vs. 50%) reacted with hurt feelings,
while more men than women (45% vs. 28%) denied responsibility
for instigating the incident. Women reported believing that
violation of an expectation on the part of the angry person
was a greater factor in the instigation of anger than men
did. Women were also more likely than men to believe that
the intensity of the other person's anger was greater than
called for. Averill acknowledges that these findings are
consistent with the feminine stereotype, but points out that
the number of significant sex differences was quite small
compared to the total number of variables tested.

In Study IV, concerning the temporal dimension of anger,
Averill (1982) again found that women were four times more
likely than men to report crying while angry. Action which
violated the expectation of an interpersonal relationship was
an important factor in arousing their anger for 91% of the
women, compared to 79% of the men. The correlation between
intensity and duration of anger was strong for women (r ==
.54), but 'was significantly’ weaker for' men (r’ = .20).
Averill's interpretation of this finding is that men may
place :more emphasis on their’ behavior' when judging' the
intensity of an episode of anger, while women may emphasize

the duration.

73

In a later report of the same study, Averill (1983)
concludes that women's experience of anger is quite similar
to that of men. Averill states that the women in his studies
became angry as often, as intensely, and for the same reasons
as men, and expressed their anger as openly as men. He
believes that the only major difference between the sexes
found in his studies was that women reported crying when
angry about four times more often than men.

The majority of the small number of studies on anger in
women appears to provide some support for the prohibition on
women's anger proposed in the theoretical literature. This
support is, however, of an indirect nature and must be
extrapolated by integrating diverse findings. In addition,
findings in six studies do not appear to support the
theoretical construct. When both these factors are
considered, it becomes quite clear that there is a compelling
need to test, directly and empirically, the hypothesized
relationship between anger and sex role. The present
research carried out a test of this nature.

Studies on Sex Role, Self Esteem,
and/or Related Variables

Several investigations have been conducted on the
relationship between sex role and self-esteem, and some have
included other related variables. Although none of these
studies addresses issues regarding anger, this literature

provides the empirical foundation for a study of the

74
relationship between sex role, self-esteem, and anger. One
of the issues raised by the proposed study is whether anger
has a significant effect on the relationship between sex role
and self-esteem. For this reason, the literature on sex role
and self-esteem, including those studies which incorporated
related variables, is reviewed here.

Most of the studies in this section explored the
relationship between sex role and self-esteem and found
consistently that androgyny and masculinity show the
strongest relationships with self-esteem. Other variables
whose relationship to sex role and self-esteem has been
studied are satisfaction with one's own body and sexual
satisfaction, sex type of occupational field, and achievement
motivation. 'The review' of these studies follows those
examining sex role and self-esteem. This section concludes
with a review of studies of the relationship between sex role
and variables related to psychological adjustment.

MW

Several researchers have examined the relationship
between sex role and self—esteem. Spence and Helmreich
(1978) found that in a sample of 715 college students, the
highest self-esteem was found in those with androgynous sex
roles, followed in descending order by masculine, feminine
and undifferentiated individuals. These relationships were
the same for males and females. The same relationships

between self-esteem and sex role were also reported for a

75
sample of 752 high school students. For both of these
samples, Spence and Helmreich also found that within each of
the sex role categories, individuals who scored above the
median on the M-F scale, indicating the existence of
masculine characteristics, scored higher on self-esteem.

In a study of 248 male and 282 female college
undergraduates, Spence, Helmreich and Stapp (197 5) found
significant positive correlations between both masculinity
and femininity and self—esteem in both men and women. The
correlation between masculinity and self-esteem was stronger
in both sexes (r - .77 and .83 for males and females
respectively, compared to .42 and .30 for femininity). When
these subjects were classified into the four sex role
categories, those classified as androgynous were highest in
self-esteem, followed in descending order by those classified
as masculine, feminine, and undifferentiated. Spence et al.
hypothesize that the strong positive individual correlations
between both. masculinity and femininity and self-esteem
suggest that the two dimensions may contribute additively to
determine self-esteem.

Following these initial investigations, Spence,
Helmreich and Holahan (1979) expanded their sex role measure
to include scales for negative aspects of masculinity and
femininity. Use of this measure (EPAQ) in a study of the
relationship between sex role and self-esteem replicated

their earlier findings of positive relationships between

76
self-esteem and positive aspects of both masculinity and
femininity. They found that for both males and females, the
correlations between self-esteem and negative masculinity
approached zero. They also reported significant negative
relationships between self-esteem and the negative femininity
scales.

Orlofsky (1977) also studied the relationship between
sex role and self-esteem. The results of his study of 111
college men and women showed that high levels of self-esteem
were associated equally with androgyny and masculinity.
Students with a feminine sex role had low to moderate scores
on the self-esteem scales, with the exception of the
attractiveness scale. Those with an undifferentiated sex
role had the lowest self-esteem score. Masculinity was
positively related to most dimensions of self-esteem for both
men and women. Taken together, these findings indicate that
high masculinity, with or without femininity, is associated
with high levels of self-esteem.

The relationship between sex role and self-esteem in
graduate nursing students has also been explored. In a study
of 96 female graduate nursing students between the ages of 20
and 50, Gautheir and Kjervik (1982) found that women with
masculine or androgynous sex roles had higher self-esteem
than those with feminine or undifferentiated sex roles.

There were no significant differences in self-esteem between

77
either the masculine and androgynous women, or the feminine
and undifferentiated women.

Two studies have investigated the relationship between
sex role and self-esteem in samples of adults with more
diverse ages and lifestyles than samples limited to college
students. Spence et al.'s 1975 study of sex role and self-
esteem was later replicated using a sample of middle-aged,
upper middle class Caucasion adults (O'Connor, Mann 8
Bardwick, 1978). The 43 men and 48 women in this sample were
between the ages of 40 and 50 and had annual incomes ranging
from $50,000 to over $100,000. The results obtained by
O'Connor et al. were identical to those of Spence et al.
Self-esteem was highest among androgynous men and women,
followed by the masculine, feminine, and undifferentiated
groups for both sexes. O'Connor et al. concluded that their
data extends the generalizability of the work of Spence et
al. (1975).

Puglisi and Jackson (1980-81) studied the relationship
between sex role and self-esteem in 1029 males and 1040
females between the age of 17 and 89. Like O'Connor et al.
(1978) , they found that for both sexes, self-esteem was
highest among those with androgynous sex roles, followed by
the masculine, feminine and undifferentiated sex role groups.
Further analyses conducted by Puglisi and Jackson, however,
showed that masculinity is a far better predictor of self-

esteem than femininity.

78

The results of these seven studies indicate that self-
esteem is strongly associated with androgynous and masculine
sex roles for both men and women. Although femininity is
also positively related to self-esteem, the relationship is
weaker than that shown by androgyny and masculinity. These
findings suggest that androgyny and masculinity may be more
important contributors to emotional well-being than
femininity. Since no studies have been conducted of the
relationship between sex role, self-esteem, and anger, no
conclusions can be drawn regarding the relationship of anger

to emotional well-being viewed from this perspective.

 

Other researchers have explored the relationship between
sex role, self-esteem, and a variety of related variables.
Three studies which explored the relationship between sex
role, self-esteem, and the following variables are reviewed
next: satisfaction with one's own body and sexual
satisfaction, sex type of occupational field, and achievement
motivation.

In a study of 204 unmarried college undergraduate women,
Kimlicka, Cross and Tarnai (1983) found that women with
androgynous or masculine sex roles scored higher on self-
esteem, satisfaction with one's own body, and sexual
satisfaction than.women with feminine or undifferentiated sex
roles. Kimlicka et al. based their analysis on two, rather

than four, sex role categories, combining androgynous and

79

masculinity scores into a single group, and doing the same
with femininity and undifferentiated scores. Results of
their initial analysis before their data was grouped in this
manner, however, showed that a positive self-concept was
strongly associated with masculinity and was not greatly
increased by the addition of femininity in the androgynous
role.

Jones and Lamke (1985) studied the relationship between
sex role and self-esteem in 144 college women in a
feminine-typed occupational field of study and 143 college
women [in a masculine-typed occupational field of study. In
contrast to the above research, their results indicate that
femininity has an important adaptive function. They found
that women in the feminine occupational group had higher
self-esteem than women in the masculine occupational group.
When they examined the relationship between sex role and
self-esteem in each occupational group, they found that
androgynous and masculine women in both occupational groups
had higher self-esteem than the feminine and undifferentiated
women in both groups. There were no differences in self-
esteem between androgynous and masculine women in the
feminine occupational group. In the masculine occupational
group, however, masculine women had lower self-esteem than
androgynous women in both occupational groups, and than
masculine women in the feminine occupational group. Jones

and Lamke conclude that social prejudices that masculine

80
women in masculine occupations experience from their male
peers may account for their lower self-esteem. They point
out that although androgynous women in masculine occupations
experience the same prejudices, they have greater flexibility
in responding and adapting since they have feminine as well
as masculine characteristics.

The relationship between sex role, achievement
motivation, and self-esteem was explored in a study by
Stericker and Johnson (1977). In a sample of 312 male and
female college students, they found that achievement
motivation and self-esteem were associated with a masculine
sex role in both males and females. On the basis of these
findings, they suggest that high self-esteem may give
individuals, particularly women, the psychological strength
and freedom to deviate from the traditional feminine role.
They conclude that the relationship between optimal self-
esteem and appropriate sex role identification is not direct
and may, in fact, be inverse.

The findings of these three studies are inconsistent
regarding the role which femininity may play in emotional
well-being. The results of the study by Kimlicka et al.
(1983) are consistent with studies reviewed above which
showed that androgyny and masculinity are more strongly
related to self-esteem than femininity. The findings of
Stericker and Johnson (1977) support the importance of

masculinity to self-esteem. The findings of Jones and Lamke

81
(1985), however, suggest that femininity may make a
significant contribution.to the greater'well-being attributed
to androgyny.

’0 ‘ 21°. V- .-_.- z: i: -_ ‘0. 0 -; 2° 0.: ‘-. .0. __: u‘l

Studies of the relationship between sex role and these
variables related to psychological adjustment are reviewed
below: identity formation, self-image, and adjustment.

In his study of 111 college men and women, Orlofsky
(1977) found that an androgynous sex role orientation was
associated with higher levels of identity formation for both
men and women. Masculinity was an important aspect of this
achievement, as evidenced in Orlofsky's findings that
cross-sex-typed women and sex-typed men (both having a
masculine sex role) had high levels of identity formation.
For women, a feminine sex role was associated with
intermediate levels of identity formation. The lowest level
of identity formation was associated.with an undifferentiated
sex role for both men and women. Orlofsky states that:

Thus, rather than leading to difficulties in

integration and identity confusion, high levels of

both masculinity and femininity are highly conducive

to identity achievement. Although both types of

characteristics are important, however, it appears

that masculine characteristics such as autonomy,

independence, and assertiveness are more crucial to

identity formation than feminine characteristics

such as understanding, warmth, and tenderness. (p.

571)

Because men in the foreclosure stage of identity formation

also had high levels of masculinity, however, Orlofsky

82
concluded that a strong masculine orientation is not always
sufficient for the achievement of identity formation.

The findings in a study of the relationship between
androgyny and self-image parallel Orlofsky's (1977) findings
regarding identity formation. Lee and Scheurer (1983) tested
their hypothesis that the superior adaptability believed to
characterize androgynous individuals would be manifested in
a self-image characterized by high self-monitoring, internal
locus of control, and positive expectations of achievement and
affiliation success. In their study of 243 college students
in introductory psychology, they found that masculinity was
more strongly associated with greater adaptiveness in self-
image than the combination of masculinity and femininity in
the androgynous sex role. Androgyny had a stronger
relationship to adaptiveness in self-image only for
expectations for affiliation success and only among female
subjects. Lee and Scheurer conclude that their findings lend
support to the view that it is advantageous for men and women
to have masculine characteristics because of the value placed
on instrumentality in this culture.

In a number of studies of the relationship between sex
role and psychological adjustment, investigators have found
that although masculinity and androgyny are both associated
with adjustment, masculinity is more important than femininity

in determining favorable adjustment.

83

Silvern and Ryan (1979) studied the relationship between
adjustment and sex role in two groups of undergraduate men and
women. They found that superior adjustment was associated
with androgyny in women, but that a traditional masculine sex
role was sufficient for men to achieve superior adjustment.
The least favorable levels of adjustment for both men and
women were associated with an undifferentiated sex role.
While their results did not indicate any negative
relationships between positive adjustment and femininity, they
do emphasize masculinity as a predictor of favorable
adjustment. Silvern and Ryan indicate, however, that

This should not be interpreted. . .as an indication

that traditionally masculine traits are inherently

more valuable than feminine traits. While masculine
traits may be more associated with personal comfort

or adjustment, self-reported feminine traits such

as ‘compassionate, ' may be highly valuable for

different reasons. (p. 761)

Hoffman and Fidell ( 1979) studied the relationships
between sex role and several characteristics, including self-
esteem and adjustment, of 369 middle-class women. They found
that women with masculine or androgynous sex roles had higher
self-esteem than feminine women, and that undifferentiated
women scored lowest on self-esteem. In terms of overall
adjustment, however, androgynous women did not show a markedly
stronger pattern of adjustment than the feminine women. Their
results did show that women with strong traits, either

masculine or feminine or both, indicate better adjustment than

undifferentiated women without strong traits of either kind.

84

Rendely, Holmstrom and Harp (1984) studied the
relationship between sex role, life style and mental health
in 97 white, suburban mothers. They found that masculine and
androgynous women reported symptoms much less frequently than
feminine and undifferentiated women on seven of the nine
dimensions of psychological distress. Their results suggest,
moreover, that it is the absence of masculine traits which is
associated with psychological distress.

The relationship between sex role and the following
measures of psychological adjustment has been studied in 45
male and 56 female introductory psychology students: The
Minnesota Multiphasic Personality Inventory (MMPI) , the Rathus
Assertiveness Schedule, and the Self-Efficacy Scale (Adams and
Sherer, 1985) . Adams and Sherer found that masculine subjects
(both males and females) scored significantly higher on self-
efficacy and assertiveness, and significantly lower on
depression and social introversion as measured by the MP1,
than androgynous, feminine or undifferentiated subjects.

Whitley (1984) conducted a meta-analysis of 32 studies
of the relationship between sex role and either psychological
depression or a more general measure of adjustment. The
overall results of his study provide support for the
masculinity model, which suggests that the relationship
between androgyny and psychological well-being is primarily
attributable to the masculinity component of androgyny, with
the influence of femininity being negligible. Whitley

85
suggests that the relationships between masculinity and both
low depression and high general adjustment reflect a strong
belief in self-efficacy attributable to masculinity.

In contrast to these findings regarding masculinity and
psychological adjustment, a smaller number of studies have
found that femininity also plays an important role in the
higher levels of adjustment associated with androgyny. Glazer
and Dusek (1985) investigated the relationship between sex
role and resolution of Eriksonian developmental crises in 139
female and 133 male undergraduates. Their results showed that
androgynous subjects had resolved the crises of trust versus
mistrust, initiative versus guilt, identity versus identity
diffusion, and intimacy versus isolation more successfully
than masculine, feminine, and undifferentiated subjects .
Glazer and Dusek argue that these results indicate that the
androgynous sex role is associated with better adjustment than
the masculine, feminine or undifferentiated roles. Both
masculinity and femininity were significant predictors of
adjustment in this study. The lowest levels of adjustment
were generally associated with an undifferentiated sex role.
The researchers concluded that their results do not support
the argument that adjustment results from higher masculinity
scores.

Burchardt and Serbin (1982) found that an androgynous
sex role was related to better personality adjustment for

women in normal and psychiatric groups. They investigated

86

the relationship between sex role and personality adjustment
as measured by an abbreviated version of the Minnesota
Multiphasic Personality Inventory (MMPI) in two studies of
106 female and 84 male undergraduate introductory psychology
students and 48 female and 48 male psychiatric inpatients.
In the college student sample, they found that androgynous
females showed better adjustment on the Depression and Social
Introversion Scales when compared to feminine females, and on
the Schizophrenia and.Mania Scales when compared to masculine
females. The findings were similar for female psychiatric
inpatients. In general, individuals with an undifferentiated
sex role had the poorest adjustment. Among females in both
the college student sample and the psychiatric sample,
androgynous individuals were the most symptom-free. Burchardt
and Serbin conclude that their findings provide support for
the notion that role flexibility is related to mental health
for women in normal and psychiatric groups.

Orlofsky and Windle (1978) have found that femininity is
strongly related to personal adjustment for females. They
studied the relationship between sex role orientation,
behavioral adaptability and personal adjustment in 58 male
and 53 female introductory psychology students. They found
the highest levels of personal adjustment among subjects whose
sex roles were consistent with their gender, i.e., masculine
and androgynous males and feminine and androgynous females.

For males, personal adjustment, as measured by the personal

87

integration scale of the Omnibus Personality Inventory, was
positively related to masculinity but unrelated to femininity.
For females, personal adjustment was positively related to
femininity but unrelated to masculinity. Orlofsky and Windle
also found that androgynous individuals displayed greater
behavioral adaptability, conceptualized as emotional
expressivity and interpersonal assertiveness, than sex-typed
or cross sex-typed individuals. It is clear that for females
in this study, femininity was the primary contributor to high
levels of adjustment.

Recent studies of the relationship between women's
involvement in multiple roles and psychological well-being
also address the question of the relationship between sex role
and psychological adjustment.

As part of a longitudinal study of life situations, Gore
and Mangione (1983) examined the relationships between social
roles, sex roles and psychological distress in a sample of 464
adult males and 647 adult females. Psychological distress was
measured by psychophysiologic symptoms and depressed mood.
They found that there were no differences in depression
between men and women who occupied the same social roles.

Among married persons who have children, however, women
who were not employed showed the highest level of depression,
which was significantly higher than working married men and
working women (Gore 8 Mangione, 1983) . On the measure of

psychophysiologic symptoms, women who had children liv ing 3‘

88

home reported significantly more symptoms than men in the same
role. This discrepancy existed for women who were employed
as well as for those who were not. Gore and Mangione conclude
that working appears to play a protective role against
depression for women who are also mothers. They also point
out that being employed does not provide this protection
against psychophysiologic symptoms, suggesting that having
children increases the psychological distress of married
working women and homemakers.

A later study of the stressfulness of daily social roles
for women was conducted by Kandel, Davies, and Raveis (1985) ,
who interviewed 197 women between the ages of 18 and 54.
Psychological distress was measured by the Center for
Epidemiological Studies Depression Scale, a psychosomaticism
scale, and a measure of role-specific stress. The lowest
level of distress was found among women who occupy the most
complex set of roles (those who are married, employed, and
parents), while higher levels of distress were reported by
single women and housewives. In addition, the occupancy of
multiple roles decreased stress within the marital role but
increased stress within the occupational role. Kandel et al.
conclude that while multiple roles clearly are associated with
higher levels of well-being for women, the net benefit results
from two opposing forces. These forces are evident in the

positive, buffering effect work has on marital stress, and the

89
negative, exacerbating effect parenting has on work-related
stress.

On the basis of previous research which suggested that
the role of paid worker was less stressful for women than more
traditional female roles, Barnett and Baruch (1985)
hypothesized that the particular roles a woman occupies, and
her quality of experience within each role, would be stronger
predictors of role strain and distress than the number of
roles occupied.

In a study of 238 Caucasian women between 35 and 55 years
of age, Barnett and Baruch (1985) examined the relationship
between role overload, role conflict, and anxiety, which were
considered stress indices, and. the following’ aspects of
multiple role involvement: the number of roles occupied, the
particular roles occupied, and the quality of experience
within each role. They found that the major source of stress
for women in this age group was the role of parent, not the
role of paid worker. The quality of experience within the
roles a woman occupied significantly predicted psychological
distress, while the role of paid worker did not. Barnett and
Baruch also found that role conflict and role overload were
strongly related.totanxiety for'nonemployed women, but not for
those who were employed. They point out that qualitative data
suggest this finding may be due to the lack of structure and
legitimacy which nonemployed women experience in their

commitments.

90

In a later analysis of this data set, Baruch and Barnett
(1986) explored the relationship between women's occupancy and
quality of experience in the roles of paid worker, wife, and
mother, and psychological well-being as measured by self-
esteem, depression, and pleasure. Contrary to popular beliefs
that entering the work force is impairing well-being for
women, Baruch and Barnett found no evidence that involvement
in multiple roles had a negative impact on well-being. They
found instead that the number of roles occupied was
significantly correlated with all three measures of well-
being. Occupancy of a particular role was associated with
well-being in only one case: the role of paid worker
predicted self-esteem. The quality of experience in all three
roles, however, was significantly related to well-being. They
also point out that there was one significant exception to
this finding: the role of mother did not predict pleasure.
Their findings in this analysis appear to confirm that it is
the quality of women's experience in their roles, not the
number of roles they occupy, which determines psychological
well-being.

Although sex role was not formally included in these
investigations of involvement in multiple roles and
psychological well-being, the findings of the four studies
reviewed above are quite relevant to the relationship between
sex role and psychological adjustment. Their overall results

demonstrate that women experience less psychological distress,

91

and greater psychological well-being, when their commitments
include the nontraditional role of paid worker. Of equal
significance is the finding that being a mother is a major
source of stress for women, while being employed is not. The
work of these researchers strongly suggests that a
nontraditional sex role, i.e., androgynous or masculine, is
associated with greater levels of psychological adjustment in
women than the traditional feminine sex role. These findings
may indicate that femininity alone may not be a sufficient
condition for well-being.

The results of the above studies indicate that higher
levels of identity formation and a more positive, and thus
adaptive, self-image are both positively related to
masculinity (Lee 8 Scheurer, 1983; Orlofsky, 1977). Findings
regarding favorable adjustment are equivocal; it is associated
with masculinity in five studies (Adams 8 Sherer, 1985;
Hoffman 8 Fidell, 1979: Rendely et al., 1984: Silvern 8 Ryan,
1979: Whitley, 1984) and with androgyny in three studies
(Burchardt 8 Serbin, 1982: Glazer 8 Dusek, 1985; Orlofsky 8
Windle, 1978) . Recent investigations of the relationship
between involvement in multiple roles and women's
psychological well-being suggest that femininity alone is not
strongly associated with favorable adjustment (Barnett 8
Baruch, 1985; Baruch 8 Barnett, 1986; Gore 8 Mangione, 1983:
Kandel et al. , 1985) . It appears that masculinity alone and,

to a lesser extent, femininity in conjunction with

92
masculinity, are much more strongly associated with adjustment
than femininity alone. None of these investigators, however,
included any anger variable in their measures of adjustment.
It is, therefore, difficult to conjecture what the
relationship between anger, sex role and self-esteem is.

The studies reviewed in this section suggest that
masculinity is more strongly associated with measures of
emotional well-being than femininity. Masculinity has been
found to be more strongly associated than femininity with
self-esteem (Hoffman 8 Fidell, 1974: Puglisi 8 Jackson, 1980-
81: Spence et al., 1975: Spence et al., 1979), sexual
satisfaction and satisfaction with one's own body (Kimlicka
et al., 1983), higher levels of identity formation (Orlofsy,
1977), more positive self-image (Lee and Scheurer 1983), and
adjustment (Adams 8 Sherer, 1985: Hoffman 8 Fidell, 1979:
Rendely et al., 1984: Silvern 8 Ryan, 1979: Whitley, 1984).
Other findings, however, demonstrate that femininity may be
an important component of emotional well-being for women
(Hoffman 8 Fidell, 1979: Jones 8 Lamke, 1985: Orlofsy 8
Windle, 1978). Androgyny is also associated with.measures of
self-esteem (Hoffman 8 Fidell, 1979: Spence 8 Helmreich, 1978)
and personality adjustment (Burchardt 8 Serbin, 1982: Glazer
8 Dusek, 1985). Those studies which show masculinity to be
more strongly related to self-esteem and adjustment than
femininity, and recent findings that employment enhances well-

being for women, while parenting is a major source of stress,

93

suggest that masculinity plays a more important role than
femininity in determining a global sense of well-being.

While these findings are not directly related to the
relationship between anger, sex role, and self-esteem, they
have important implications for the study of this
relationship. If the theoretical proposition that anger has
been prohibited in the feminine sex role is accurate, it may
help to explain why femininity appears to be less important
than masculinity in determining emotional well-being.
Theorists argue, however, that anger has also been distorted
in the masculine role. No empirical work has been found that
investigates the relationship between anger, sex role and
self-esteem in women. In light of the importance placed on
the suppression of anger in the psychological development of
women, and the crucial relationship which the theoretical
literature hypothesizes exists between anger and self-esteem
in women as a result of developmental factors, an empirical

study of these relationships was clearly needed.

Some Observations On Methodology
Most of the empirical literature reviewed above reported
the findings of descriptive studies. A variety of measures
was used to assess the variables of interest in this study:
anger, sex role and self-esteem. In this section some
observations are made regarding the instrumentation used to

measure these variables in the empirical studies reviewed.

94
Anger

A variety of measures was used in the 14 studies on anger
in women that were reviewed. Six of these studies used
measures of hostility or hostility guilt (Atkinson 8 Polivy,
1976: Evans, 1984: Hertsgaard 8 Light, 1984: Heyman, 1977:
light, 1984: Martin 8 Light, 1984). In only two studies did
the researchers specify that they used an inventory designed
particularly to assess anger (Biaggio, 1980a: Hazaleus 8
Deffenbacher, 1985). Biaggio used the Anger Self-Report and
Hazaleus and Deffenbacher used the Anger Inventory. The
remaining studies used questionnaires or instruments developed
by the researchers (Allen 8 Haccoun, .1976: Averill, 1982:
Holahan 8 Spence, 1980) , in-depth interviews (Wallerstein,
1986) , self-descriptions (Gormly 8 Gormly, 1984) , and a diary
recorded over a two-week period (Biaggio, 1988) . The Anger
Self-Report used by Biaggio (1980a) was selected for use in

the present study.

maple

Sixteen individual studies on the relationship between
sex role, self-esteem, and/or related variables have been
reviewed. All except one (Stericker 8 Johnson, 1977) used
either the Personal Attributes Questionnaire (PAQ: Spence,
Helmreich 8 Stapp, 1974) or the Ben Sex Role Inventory (BSRI:

Bem, 1974) as a measure of sex role. Of the 41 studies

4"

95
included in the meta-analysis by Whitley (1984) , only five did
not use either the BSRI or the PAQ to measure sex role.

The method of scoring most commonly used with both the
BSRI and the PAQ is based on the median split for the
masculinity and femininity scales (Spence, Helmreich, 8 Stapp,
1975). The median-split.method.was used in all but one of the
14 individual studies reviewed. Although scores on the BSRI
were originally calculated on the basis of the difference
between Femininity and Masculinity scores (Bem, 1974), Bem
subsequently adopted the median-split method of scoring
advocated by Spence, et al. The median-split method is
preferable because it differentiates between individuals who
score high on both masculinity and femininity, and those who
score low on both dimensions. The BSRI does not differentiate
between these two groups.

A hybrid scoring method which combines the two scoring
methods has been advocated (Orlofsky, 1977: Orlofsky, Aslin,
8 Ginsburg, 1977: Tesch, 1984). Of the 14 individual studies
reviewed here, the hybrid method was used in only one. Spence
and Helmreich (1979) have demonstrated that the median-split
method is a more useful and accurate method of scoring than
the hybrid method. The median-split method was used in the
present study.

W
Self-esteem was included as a variable and measured

directly in 10 of the 16 studies noted above. The Texas

96
Social Behavior Inventory (TSBI) was used as a measure of
self-esteem in the majority of these studies and was selected

for use in the present research.

Summary

No studies were found of the relationship between anger,
sex role and self-esteem in women. The literature on sex
differences in aggression is summarized to provide the
necessary context for understanding existing research on
women's anger. Although there is considerable controversy
regarding whether sex differences in aggression are innate or
learned, there is widespread agreement that males exhibit
higher levels of aggression than females and that
socialization plays a major role in determining this
difference. Studies which demonstrate that both aggression
and nonaggression can be learned by, and can be cathartic for,
both males and females support the role socialization plays
in determining sex differences in aggression. These
conclusions support theoretical propositions that the feminine
sex role contains strong prohibitions on anger in women.

Of the small number of studies on anger in women, the
majority provide support for the hypothesized prohibition on
anger in the feminine sex role. For example, women have been
found to express more guilt over hostility than men. Positive
relationships have been found between anger and negative

aspects of masculinity and femininity among women. It appears

97

to be more difficult for women than for men to integrate
hostility and aggression into their roles. Several studies
found that women in less traditional roles report higher
levels of hostility, anger, and/or aggression. The
nontraditional roles of these women were suggested by their
higher levels of education, lower need for dependency, more
liberal beliefs about autonomy, or greater commitment to
career than family.

The support provided by the empirical literature for the
hypothesized prohibition of women's anger is indirect and
extrapolated from diverse findings. In addition, six studies
were found whose findings are not consistent with the
supportive findings. These two factors underscored the need
for an empirical test of the hypothesized prohibition of anger
in the feminine sex role.

Studies of the relationship between sex role, self-esteem
and related. variables show' that :masculinity' is strongly
associated with self-esteem and other factors important to
psychological well-being, such as identity formation, self-
image, and adjustment. Androgyny has also been associated,
to a lesser extent, with self-esteem and adjustment. There
is limited.empirical support for femininity'being’an.important
aspect of psychological well-being for women. This lack of
support for femininity, in conjunction with recent findings

that the role of mother is the major source of stress for

98
women in mid-life, suggests that femininity alone is not
strongly associated with a global sense of well-being.

The hypothesized prohibition of anger in the feminine
sex role may help explain the strength of the relationship
between.masculinity and.well-being. Since no anger variables
were included in studies of adjustment or well-being, an
empirical exploration of the relationship between anger, sex
role and self-esteem was needed to help clarify questions

raised by existing research.

CHAPTER IV
DESIGN AND METHODOLOGY

In this chapter the procedures used in conducting the
research are described. Methods used to select the sample,
collect the data, and ensure confidentiality are set forth.
The instrumentation used for the data collection is then
described, followed by a statement of the research hypotheses.
The chapter concludes with an explanation of the statistical

procedures used in the data analysis.

Subjects

The subjects for this study were obtained from a
systematic sample of women employees listed in the 1987-88
Michigan State University Faculty-Staff Directory. Systematic
samples are generally considered.to be representative samples
and provide less opportunity than random samples for sampling
error to occur (Glass 8 Hopkins, 1984). The findings of both
systematic and random samples are generalizable (Glass 8
Hopkins). With random populations, systematic sampling is
equivalent to random sampling and either design can be used
(Scheaffer, Mendenhall, 8 Ott, 1979). Since the listing of
employees in the Faculty-Staff Directory was random with
respect to the variables considered in this study, a
systematic sample was acceptable and results of the study may

be generalized to the population'tof ‘women employees of

99

100
Michigan State University from which it is drawn.

The initial sample of 689 women who were asked to
participate in the study was drawn from the approximately
4,300 women employed at Michigan State University in the
following' categories: faculty, academic staff,
administrative-professional, supervisory administrative-
professional, police, and clerical-technical. Women.who were
employed off the main campus of Michigan State University were

not included in the sample.

Method of Data Collection

The data for this study were collected as a survey of
women employees conducted by the Employee Assistance Program
at Michigan State University using procedures recommended by
Dillman (1978). The 689 women identified in the systematic
sample received a letter through interdepartmental mail at
their work place describing the planned research and
requesting their participation (Appendix A). This mailing
included a post card which women who were willing to
participate in the survey'were asked.to return to the Employee
Assistance Program ‘with their names and home addresses
(Appendix B) . An informational brochure explaining the
services of the Employee Assistance Program was also enclosed.
A follow-up mailing was sent one week later, thanking those
who had responded and urging those who had not to do so

immediately. (Appendix C).

101

Of the 689 women who were identified in the initial
sample and asked to participate in the study, 396 indicated
that they were willing to do so. These subjects were sent a
questionnaire through. United States mail to their home
addresses (Appendix D). A cover letter thanking them for
their interest and emphasizing the importance of their
participation was included in this mailing with the
questionnaire (Appendix E). A stamped, addressed return
envelope was also included.

The researcher had originally planned to send a follow-
up mailing to each of the 396 women one week after the
questionnaire was mailed. The purpose of this mailing was to
have been to thank.those who had returned their questionnaire
and urge those who had not to do so. This procedure was
modified as a result of the response to a similar follow—up
mailing' to the initial letter' explaining the study and
requesting the subj ects' participation. Many of the women who
had responded to this initial mailing were confused as to why
they received the follow-up mailing, and were concerned that
their response had not been received. As a result of the
disruption caused by this misunderstanding, it was decided
that only those subjects who had not returned their
questionnaires within ten days of the date it was mailed would
receive a follow-up post card.

Of the 396 potential subjects who expressed willingness

to participate in the study, 52 did not return their

102

questionnaires. A total of 344 questionnaires were returned.
Of these, two were blank. Of the remaining 342 questionnaire,
70 subjects omitted one or more responses. Data from the 342
questionnaires which were partially or fully complete were

included in the data analysis.

Procedures to Ensure Confidentiality

All questionnaires were returned to the researcher
anonymously with the exception of those from three subjects
who requested information from the Employee Assistance
Program. When the questionnaire was mailed, subjects received
a post card which they were asked to return separately,
indicating that they had returned their questionnaire
(Appendix G). Use of the return post card made it possible
to remove the names of subjects who had returned the
questionnaire from the mailing list for follow-up mailings.
This assurance of anonymity was necessary since the subjects
were asked to provide personal information about their
feelings and self-perceptions to an agency which was part of
the organization employing them. This return card also
allowed respondents to indicate whether they wished to receive
a summary of the results of the study. Those who wanted this
information were mailed a expanded copy of the Abstract
(Appendix H) with a cover letter thanking them for their

participation in the research (Appendix I).

103
Instrumentation

The questionnaire had three parts. First, subjects were
asked questions regarding their use and perception of services
offered by the Employee Assistance Program. This data was
compiled and forwarded to the Employee Assistance Program for
use in planning future programs. The second part of the
questionnaire consisted of the following psychological
inventories described below which measured the variables of
anger, sex role and self-esteem: the Anger Self-Report, the
Personal Attributes Questionnaire, and the Texas Social
Behavior Inventory. The third part of the questionnaire
consisted of questions designed to gather the following
demographic information: marital status, age, years of
education completed by the respondent and her parents,
educational level, employee group, type of work, most
important work activities, and before-tax household income.

n e - e 0 AS

Subjects were administered five scales of the ASR (Zelin,
Adler, 8 Myerson, 1972) to measure the three anger variables
(awareness of anger, expression of anger, condemnation of
anger). The ASR was developed to discriminate between the
awareness of anger and its expression, and the five scales
selected for this research reflected the areas of concern in
this study. The following scales were administered: Awareness
of Anger, General Expression of Aggression, Verbal Aggression,

Physical Aggression, and Condemnation of Anger. The

104

researcher had originally planned to use the Total Expression
of Anger score, which is a composite of the scores on the
general expression of aggression, verbal aggression, and
physical aggression scales. During the analysis of the data,
however, it became obvious that.more useful information.would
be obtained by analyzing and reporting scores on the three
scales separately. For this reason it was decided that the
three separate expression scales, rather than the composite
Total Expression of Anger score, would be used to measure the
expression of anger variable.

The ASR gives the respondent a series of self-descriptive
statements such as ”I get mad easily," "It's right for people
to express themselves when they are mad," and "If I don't like
somebody, I will tell him so." The respondent is asked to
indicate how each statement applies to her on a six point
scale ranging from "strong' disagreement" (1) to "strong
agreement” (6).

The authors conducted validation studies of the ASR on
samples of 82 psychiatric patients and 67 college students
(Zelin et al., 1972). ASR scores of the psychiatric patients
were correlated with psychiatrists' ratings on 16 of the
Problem Appraisal Scales (PAS) (Endicott 8 Spitzer, 1972)
considered most relevant to the assessment of anger. On the
basis of their analysis of these correlations, Zelin and his
associates found that the ASR scales show substantial

convergent and discriminant validity.

105

Regarding convergent validity, awareness of anger
correlates positively with PAS anti-social attitudes and acts
(r - .24, p < .05) and negatively with PAS obsessive-
compulsiveness (r = -.37, p < .01). The physical aggression
scale correlates most highly with Assaultive Acts on the PAS
(r - .41, p < .01). The highest correlation for ASR general
expression.is with.PAS antisocial attitudes and acts (r'= .25,
p < .05). ASR verbal aggression has a significant negative
correlation of -.36 (p < .01) with PAS dependency, and
positive correlations of .31 (p < .01) with PAS anger-
belligerent negativism and .28 (p < .01) with anti-social
attitudes and acts. ASR condemnation of anger has significant
correlations with PAS suicidal thoughts, anxiety-phobias, and
obsessive—compulsiveness (r’ = -.22, .23, .22, p < .05,
respectively).

Discriminant validity of the ASR has also been
demonstrated by its correlations with the 16 PAS scales. For
example, ASR physical aggression is the only ASR scale which
correlates significantly with PAS assaultive acts.
Significant negative correlations were found between the
dependency scales of the PAS and ASR general expression and
verbal aggression, but not with ASR awareness of anger. The
authors believe that this indicates that highly dependent
people can be aware of anger but feel compelled to inhibit its
expression. This finding provides support for the validity

of the distinction between awareness of anger and its

106

expression. Discriminant validity is further demonstrated in
the significant positive correlations found between PAS
antisocial attitudes-acts and both ASR awareness (r = .24, p
< .05) and all the ASR expression scales, with general
expression showing'a1correlation.of .25 (p < .05) and physical
aggression and.verbal aggression both showing correlations of
.28 (p < .01) (Zelin, et al., 1972).

Peer ratings were used to assess the validity of the ASR
scores for the college student sample. The 67 students
completing the ASR.were rated on six scales by three students
who lived nearby in the students' dormitories (Zelin et al.,
1972) . For example, ASR verbal aggression correlates
positively only with ”To what extent does this person provoke
arguments?" on the peer rating scale (r = .29, p < .05). The
highest positive correlation for ASR awareness of anger is
with "To*what extent.does this person feel anger?” on.the peer
rating scale (r = .29, p < .01).

Split-half reliabilities for the ASR reported by Zelin
et al. (1972) range from .64 to .83. Reliabilities for the
scales used in this study are as follows:

Awareness of Anger .82
General Expression of Aggression .66
Physical Expression of Aggression .64
Verbal Expression of Aggression .78

Condemnation of Anger .81

107

Biaggio (1980b) administered four anger scales, including
the ASR, to 150 college undergraduates. The results of her
study indicate that the ASR shows concurrent validity with the
Buss-Durkee Hostility Inventory (BDHI) . In a later study,
Biaggio, Supplee and Curtis ( 1981) administered the four anger
inventories to 60 college undergraduates, who also filled out
a personal incidents record and participated in imaginal and
role-playing laboratory situations designed to provoke anger.
Significant correlations between ASR scales and the laboratory
self-report data and the subjects' personal incidents records
established good predictive validity for ASR awareness of
anger , general expression , physical aggression , and
condemnation of anger.

Means and standard deviations for the ASR have been
reported for psychiatric inpatients and college students
(Zelin et al. 1972: Biaggio, 1980b). Neither set of data
appears to provide appropriate norms for this study. There
are large differences between the scores for psychiatric
inpatients and college students in Zelin et al.'s study. In
addition, Biaggio used a different method of scoring than
Zelin and his associates, making it impossible to compare the
scores for college students in the two studies. The means for
the research sample were used to provide the norms for the
current study.

Each of the items on the ASR is scored in either the

agree or disagree direction. All items scored in the disagree

108

direction are given a negative value in scoring. .As a result,
the raw scores on the scales can have negative values.
Biaggio recommends adding a constant of 40 to the raw score
for each scale to eliminate the negative values (M. K.
Biaggio, personal communication, June 8, 1988). Raw scores
were modified in this manner in the present research in
accordance with Biaggio's recommendations.

e son 5 nna'r

The sex role attitudes of the respondents in this study
were measured by the short form of the PAQ (Spence, Helmreich
8 Stapp, 1974: Spence 8 Helmreich, 1978). The short form of
the PAQ consists of 24 bipolar trait characteristics such as
"Not at all aggressive--Very aggressive” and "Not at all
independent--Very independent." For eaCh characteristic the
respondent rated herself on a five point scale to indicate
where she fell between the two extremes.

The PAQ contains three scales, masculinity (M),
femininity (F), and masculinity-femininity (M-F). The M and
F scales each contain eight items which are considered
socially desirable characteristics for both sexes. Those
items on the M scale are believed to he possessed by males to
a greater degree than by females. The F scale contains items
which are believed to be possessed by females to a greater
extent than by males. The content of the scales reflects the
instrumental, agentic orientation of masculinity and the

expressive, communal orientation of femininity as

109
conceptualized by Parsons and Bales (1955) and Bakan (1966).
The M-F Scale contains eight characteristics whose social
desirability differs for the two sexes (Spence 8 Helmreich,
1978).

The full version of the PAQ contains 55 bipolar items
derived from a pool of 122 items developed by Rosenkrantz,
Vogel, Bee, Broverman and Broverman (1968) in studies of
characteristics believed to differentiate men and women. The
existence of sex role stereotypes was confirmed by Rosenkrantz
et al., who asked their subjects to rate the typical adult
male, the typical adult female, and the ideal individual (sex
unspecified) on each item. In developing the PAQ,
introductory psychology students were asked to rate the
typical adult of each sex, the typical college student of each
sex, or the ideal individual of each sex for each of the items
in the pool used by Rosenkrantz et al. Each student was also
asked to rate him/herself (Spence 8 Helmreich, 1978).

The 55 items for the original version of the PAQ were
selected from those items on which male and female ratings
differed significantly for the typical adult, typical college
student, and ideal individual. These significant differences
indicate the existence of consistent stereotypes regarding
differences between the sexes (Spence 8 Helmreich, 1978).

The short form of the PAQ which was used in this study
is highly correlated with the original 55 item PAQ.

Correlations between the scales of the original and short

110
versions of the PAQ in a sample of the college students are
.93, .93, and .91 for the M, F, and M-F scales respectively.
Scales for the short form have Cronbach alpha reliabilities
of .85, .82, and .78 for the M, F, and M-F scales respectively
(Spence 8 Helmreich, 1978).

Each item on the PAQ is scored from 0 to 4, with high
scores indicating either an extreme masculine or extreme
feminine response. Respondents were classified into one of
four sex role groups on the basis of whether their scores fell
above or below the sample median on the M and F scales. The
following fourfold classification was used: androgynous (high
masculine, high feminine): masculine (high masculine, low
feminine): feminine (high feminine, low masculine): and
undifferentiated (low masculine, low feminine) (Spence 8
Helmreich, 1978) . The sample medians were used in classifying
the subjects in accordance with the suggestion of Spence and
Helmreich that this procedure allows relationships within the
population to be revealed more clearly.

In a sample of 248 male and 282 female introductory
psychology students, the internal consistency of the PAQ has
been demonstrated in alpha coefficients of .92 and .90 for men
and women respectively on the stereotype ratings, and .73 and
.91 on self ratings (Spence et al., 1974). Later analyses
using data from several thousand high school and college
students and parents resulted in alpha coefficients in the mid

.708 for the M, F, and M-F scales (Helmreich, Spence 8

111
Wilhelm, 1981). Test-retest reliabilities for 31 subjects in
Spence et al.'s (1974) original study who retook.the PAQ after
13 weeks were .92 and .98 for men and women respectively on
stereotype ratings, and .80 and .91 on self ratings.

Factor analyses of the PAQ M and F scales result in two
factors which consist of masculinity and femininity and show
a.mean intercorrelation of -.04 (Helmreich, Spence 8 Wilhelm,
1981). Studies conducted during the development of the PAQ
show that the male-valued and female-valued items correlated
negatively for both sexes (Spence, Helmreich 8 Stapp, 1974).

The validity of the PAQ has been established in several
different contexts. Spence et al. (1974) performed a cross
validational study of the use of the PAQ to measure sex role
stereotypes by giving the PAQ to a second sample of
introductory psychology students of 56 males and 108 females.
They found significant stereotypes for both sexes on all
items. In a study of two independent samples of college
students, significant differences were found between men and
women on every item, with men scoring higher on the M and M-
F items scored in a masculine direction, and lower on the F
items, scored in a feminine direction (Spence, Helmreich 8
Stapp, 1975).

Construct validity for the PAQ has been demonstrated in
a variety of studies. In a study of 248 male and 282 female
introductory psychology students, self and peer ratings on the

PAQ of the ideal male and female fell toward the same pole,

112

supporting the conceptualization of masculinity and femininity
as a dualism. The content of the male-valued and female-
valued items was also consistent with the agency-communion
(Bakan, 1966) and instrumental-expressive (Parsons 8 Bales,
1955) distinctions drawn between masculinity and femininity
(Spence, Helmreich 8 Stapp 1975). In a study of 123 female
introductory psychology students, Klein and Willerman (1979)
found that undifferentiated and feminine women were
significantly less dominant than masculine and androgynous
women in laboratory settings and.cn the Dominance scale of the
California Psychological Inventory.

The findings of two additional studies are relevant to
the construct validity of the PAQ. PAQ scores which indicate
the perception of large differences between the sexes have
been found to correlate with traditional sex role attitudes
on the Attitudes Toward Women Scale (Spence 8 Helmreich,
1978). The power strategies which 50 male and 50 female
heterosexual college. students reported. that. they' use in
romantic/sexual relationships were found to be consistent with
sex roles as measured by the PAQ. For example, women and
males and females with feminine sex roles reported using
indirect and unilateral power strategies. In contrast, men
and masculine males and females used direct and bilateral

strategies (Falbo, 1982).

113
as c' v' v S

The short form of the TSBI was used to measure the self-
esteem of respondents in this study (Helmreich, Stapp 8 Ervin,
1974: Helmreich 8 Stapp, 1974). The short form of the TSBI
consists of 16 statements which.were designed to assess self-
confidence and competence in social situations (Spence 8
Helmreich, 1978). For each item the respondent rated herself
on a five point scale ranging from "Not at all characteristic
of me" to "Very much characteristic of me" (Helmreich et al.) .

The TSBI was developed from a pool of 60 items dealing
with aspects of personal worth and social interaction which
were administered to over 1000 introductory psychology
students. The 32 items on the original form of the TSBI were
selected from this pool on the basis of factor analysis. The
correlation between the long and short forms of the TSBI is
.97 (Helmreich 8 Stapp 1974). Use of the TSBI has
demonstrated test-retest reliabilities of .94 for males and
.93 for females. All correlations between individual items
and the whole TSBI are significant, with.mean correlations of
.53 for males and .55 for females (Helmreich et al., 1974).

The validity of the TSBI has been demonstrated in a
number of contexts. In studies of variables thought to have
an important relationship to self-esteem, experimental
hypotheses have been confirmed using the TSBI. For example,
on the basis of previous research, Helmreich et al. (1970)

predicted that individuals with moderate self-esteem would

114
perceive a competent person who blundered as attractive, while
individuals with high or low self-esteem would not. The
findings of their study, using the TSBI, supported their
hypothesis. Kimble and Helmreich (1972) conducted another
study based on previous findings that high and low self-esteem
individuals have a greater tendency to evaluate themselves on
the basis of how others view them than do individuals with
moderate self-esteem. Kimble and Helmreich hypothesized that
individuals with high and low self-esteem would demonstrate
a greater need for social approval than individuals with
moderate self-esteem. ‘The findings of their study, using the
TSBI, supported this hypothesis.

The validity of the TSBI has also been demonstrated
through its correlations with other psychological measures.
Correlations of .50 for males and .52 for females with the
self-esteem scale of the California Personality Inventory
support the construct validity of the TSBI (Helmreich et al.,
1974). The TSBI is strongly correlated with the masculinity
and femininity scales of the Personal Attributes Questionnaire
(PAQ). Correlations with the masculinity scale are .81 and
.83 for males and females, respectively: correlations with the
femininity scales are .42 and .44 respectively (Helmreich 8
Stapp, 1974) . Discriminant validity is illustrated in the
small and nonsignificant correlations between the TSBI and
intelligence, as measured by the Scholastic Aptitude Test

(Helmreich et al.).

115

In selecting a measure of self-esteem for use in this
study, it was important to consider whether the
conceptualization of self-esteem in the inventory was
consistent with the theoretical base of the study which argues
that, for women, self-esteem is derived from relationships.
It was important, therefore, that the inventory chosen to
measure self-esteem reflect this conceptualization of self—
esteem as being rooted in one's interactions with others.
Stake and. Orlofsky (1981) have addressed. this issue in
pointing out that measures of self-esteem used with female
populations must have the capacity to measure communal as well
as agentic aspects of self-esteem. They contend that since
a communal orientation is central to the feminine sex role,
measures of self-esteem which tap only the agentic qualities
more characteristic of the masculine sex role lack predictive
validity when used with female subjects.

Since the TSBI is a measure of social competence, its
conceptualization of self-esteem is consistent with the
communal, relationship-based nature of self-esteem defined in
the theoretical base of this study. Evidence of this
consistency is found. in research reported. by’ Stake and
Orlofsky (1981). They examined the relationship between the
TSBI and specific measures of agentic and communal self-esteem
and found that the TSBI has three factors, one of which
measures both agentic and communal aspects of self-esteem.

Seven items loaded on this factor, while only four loaded on

116
each of the other two factors. Thus, the philosophical and
theoretical appropriateness of the TSBI for this study is

evident.

Hypotheses

The following hypotheses were tested in this study:

1. There will be differences between women who have masculine
or androgynous sex roles as measured.by the PAQ, and women who
have feminine or undifferentiated sex roles as measured by the
PAQ, with respect to scores on the ASR awareness of anger,
general expression of aggression, verbal aggression, physical
aggression, and condemnation of anger scales. More
specifically:

a. Women with masculine or androgynous sex roles as
measured by the PAQ will score higher on the ASR awareness of
anger scale than women who have feminine or undifferentiated
sex roles as measured by the PAQ.

b. Women with masculine or androgynous sex roles as
measured by the PAQ will score higher on the ASR general
expression of aggression scale than women who have feminine
or undifferentiated sex roles as measured by the PAQ.

c. Women with masculine or androgynous sex roles as
measured by the PAQ will score higher on the ASR verbal
aggression scale than ‘women ‘who have feminine or

undifferentiated sex roles as measured by the PAQ.

117

d. Women with masculine or androgynous sex roles as
measured by the PAQ will score higher on the ASR physical
aggression scale than ‘women ‘who have feminine or
undifferentiated sex roles as measured by the PAQ.

e. Women with masculine or androgynous sex roles as
measured by the PAQ will score lower on the ASR condemnation
of anger scale than 'women ‘who have feminine or
undifferentiated sex roles as measured by the PAQ.

2. Scores on the ASR scales will play a role in predicting
the self-esteem of individual women as measured by the TSBI.
More specifically:

a. ASR awareness of anger scores will negatively predict
TSBI self-esteem scores.

b. ASR general expression of aggression scores will
negatively predict TSBI self-esteem scores.

c. .ASR‘verbal aggression scores will negatively predict
TSBI self-esteem scores.

d. ASR physical aggression scores will negatively
predict TSBI self-esteem scores.

e. ASR condemnation of anger scores will positively
predict TSBI self-esteem scores.

3. There will be differences in TSBI self-esteem scores
between the four groups of women having masculine,
androgynous, feminine, or undifferentiated sex roles as

measured by the PAQ. More specifically:

118

a. Women having androgynous sex roles as measured by the
PAQ will score higher on the TSBI than women having masculine,
feminine, or undifferentiated sex roles as measured by the
PAQ.

b. Women having masculine sex roles as measured by the
PAQ will score higher on the TSBI than women having feminine
or undifferentiated sex roles on the PAQ.

c. Women having feminine sex roles as measured by the
PAQ will score higher on the TSBI than women having
undifferentiated sex roles as measured by the PAQ.
4. There will be an interaction between sex roles as measured
by the PAQ and ASR awareness of anger, general expression of
aggression, verbal aggression, physical aggression and
condemnation of anger scores in predicting TSBI self—esteem
scores. More specifically:

a. ASR awareness of anger scores will:

(1) positively predict TSBI self-esteem scores for
women having masculine sex roles as
measured by the PAQ.

(2) negatively predict TSBI self-esteem scores for
women having feminine sex roles as measured by
the PAQ.

(3) not predict TSBI self-esteem scores for women
having androgynous or undifferentiated
sex roles as measured by the PAQ.

b. ASR general expression of aggression scores will:

(1)

(2)

(3)

119
positively predict TSBI self-esteem scores for
women having masculine sex roles as
measured by the PAQ.
negatively predict TSBI self-esteem scores for
women having feminine sex roles as measured by
the PAQ.
not predict TSBI self-esteem scores for women
having androgynous or undifferentiated

sex roles as measured by the PAQ.

c. ASR verbal aggression scores will:

(1)

(2)

(3)

positively predict TSBI self-esteem scores for
women having masculine sex roles as

measured by the PAQ.

negatively predict TSBI self-esteem scores for
women having feminine sex roles as measured by
the PAQ.

not predict TSBI self-esteem scores for women
having androgynous or undifferentiated

sex roles as measured by the PAQ.

d. ASR physical aggression scores will:

(1)

(2)

positively predict TSBI self-esteem scores for
women having masculine sex roles as

measured by the PAQ.

negatively predict TSBI self-esteem scores for
women having feminine sex roles as measured by

the PAQ.

120

(3) not predict TSBI self-esteem
scores for women having androgynous or
undifferentiated sex roles as measured by
the PAQ.

e. ASR condemnation of anger scores will:

(1) negatively predict TSBI self-esteem scores for
women having masculine sex roles as
measured by the PAQ.

(2) positively predict TSBI self-esteem scores for
women having feminine sex roles as measured by
the PAQ.

(3) not predict TSBI self-esteem scores for women
having androgynous or undifferentiated
sex roles as measured by the PAQ.

In reviewing the literature, very little empirical data
were found that suggest the existence of relationships between
sex role, anger, self-esteem, and demographic variables such
as age, educational level, and occupational classification.
It was not feasible, therefore, to formulate hypotheses which
addressed the effects of demographic variables on the
variables of interest in this study. 1 There was reason to
believe, however, that such demographic variables might affect
the other variables being investigated in this study. For
this reason, information on several demographic variables was
gathered and used to explore relationships between the primary

variables of sex role, anger and self-esteem. These variables

121
were martial status, age, educational level, employee
classification, and household income. Data were also gathered
for use in a future study of the impact of socioeconomic

status on anger in women.

Statistical Procedures for Data Analysis

Descriptive statistics were computed for each of the main
variables: awareness of anger, general expression of
aggression, verbal aggression, physical aggression,
condemnation of anger, self-esteem, and sex role.

The following statistical procedures were used to analyze
the data for this study:
1” .A multi-variate analysis of variance was used to test
hypothesis 1, which stated the expected differences in
awareness of anger, general expression of aggression, verbal
aggression, physical aggression, and condemnation of anger,
between women with masculine or androgynous sex roles, and
women with feminine or undifferentiated sex roles. Planned
and post hoc contrasts were used to test the specific parts
of the hypotheses. The effects of the demographic variables
were explored for each of the anger variables using a multi-
variate analysis of variance and post hoc contrasts.
2. A stepwise multiple regression analysis was used to test
hypothesis 2, which described the manner in which awareness
of anger, general expression of aggression, verbal aggression,

physical aggression, and condemnation of anger were expected

122
to predict self-esteem. Stepwise multiple regression analyses
were used to test the effects of the demographic variables in
predicting self-esteem.
3. A one-way analysis of variance was used.to test hypothesis
3, which. stated. the: expected. differences in self-esteem
between the four sex role groups. The specific parts of the
hypothesis were tested by pairwise contrasts using the Tukey
procedure, and by pairwise contrasts using a less conservative
contrast-based alpha level.
4. A stepwise multiple regression analysis was used to test
hypothesis 4, which described the expected variability,
according to sex role, of the manner in which awareness of
anger, general expression of aggression, verbal aggression,
physical aggression, and condemnation of anger were
anticipated to predict self-esteem.
s ' V 'a

Analysis of variance was used to test hypotheses 1 and
3. The assumptions of analysis of variance are:
1. The observations are sampled from normally distributed
populations.
2. The variances of the populations from which the
observations are drawn are homogeneous.
3. The observations are independent.
An examination of the frequency distributions and.descriptive
statistics for the data confirmed that these assumptions were

met in this study. The homogeneity of variance assumption was

123
also tested.by the Bartlett test (Glass 8 Hopkins, 1984). The
results confirmed that the assumption had been met by the
data.
'5 e s 'o

Stepwise multiple regression analyses were used to test
hypotheses 2 and 4. In this type of analysis the best
predictor of the dependent variable is entered into the
regression equation in step one. The variable which accounts
for the greatest amount of variance, in conjunction with
variables already in the equation, is selected and added in
each successive step. Stepwise multiple regression was
employed because it shows the change in the variance of the
outcome variable which is accounted for by the inclusion of
each independent variable in the regression equation. This
quality was desirable in the present research because there
was no empirical base for predicting the relative importance

of the ASR variables in predicting self-esteem.

WW

Regression analyses are based on the following three
assumptions: normality, linearity, and homoscedasticity of
residuals. The normality assumption was tested by histogram
and normal probability plots of the residuals, which proved
to be normally distributed. The linearity assumption was
tested by plotting the actual TSBI scores against the

predicted TSBI scores. The resulting plot demonstrated that

124

the actual and predicted scores were linearly related.
According to Berenson, Levine, 8 Goldstein (1983) , the
homoscedasticity assumption may be tested by plotting the
residuals against the scores on the independent variables.
When no pattern emerges in the scatterplot, the
homoscedasticity assumption has been met. Accordingly, the
residuals for the TSBI scores were plotted against the scores
on each of the five ASR scales. Random scatterplots indicated
that the assumption was met for those ASR scales which proved
to be significant predictors of TSBI self-esteem scores.

In addition to testing the three assumptions of
regression analysis described above, the residuals were also
examined to determine whether the observations were
independent. Independence may be tested by plotting the
residuals against each case (Berenson et al., 1983) . The
casewise plots were random, indicating that the observations

were independent.

NEW

In performing regression analyses it is important to
inspect relationships between the independent variables for
the existence of multicollinearity or singularity.
Multicollinearity occurs when independent variables are highly
correlated with each other. Singularity refers to linear
dependency among the independent variables, i.e. , one variable

is a linear combination of other variables. Pedhazur (1982)

125

has demonstrated that these conditions can result in a lack
of precision in the results of the regression analysis.

Correlation coefficients were computed between all the
main variables to detect the possibility of multicollinearity.
Some significant correlations were found between the ASR
scales, most notably a correlation of .65 between verbal
aggression and general expression of anger, and a correlation
of .51 between general expression of anger and awareness of
anger. Other correlations between ASR scales ranged from -
.16 to .42. Zelin et al. (1972) reported the existence of
correlations between the scales and addressed the question of
whether the scales were sufficiently independent to result in
differential diagnostic information on. the Ibasis of ‘the
difference between scores on the scales. Zelin at al.
concluded that there was more than adequate reliable,
independent variance between the scales.

Correlations between the PAQ scales and the ASR scales
did not pose problems of multicollinearity. The highest
correlation was found between PAQ masculinity and ASR verbal
aggression (.31). Table 4.1 contains the complete
correlations matrix for all main variables in this study.

The problem of singularity was not found in the present
research. ane of the independent variables were linearly

dependent.

126

*NH.

sevo.u

444mm.

«*mH.I

seed.)

444mm.

seams.

Hume

*NH.

***oo.

***mm.

***mm.l

***Hm.

AH.I

enma.

mmoz< ho
onwmmmmxm
Admmzmo

***oo.

8*mH.|

*«umo.

oo.

mo.)

onmmmmoO< onmmmmoc<

A<0Hw>=m

*summ.

***mw.

*muoo.

usemm.u

4*woe.

Ho.l

assem.

A<mmm>

«mascara» can:

44mm.

444mm.1

*nma.l

esmm.l

sesvm.l

mo.)

sesom.n

mmoz< ho
zo~a<zzmnzoo

" mac w #0 Hflhhoo GOOEGOQ

4*ss.- 4.4mm.
8*«Hm. H..-
***Nv. oo.
***oe.u so.-
44*e~.- mo.n
In so.:
3.- ..
use..- me.
mmcz< so mszHzme
mmmzme<z<

“H.o canoe

Hoe. v a «4*

do. v s *4

mo. v m 8
seems. Home
same. mmoz< no
onmmmmmxm.zmo
mo.n onmmmmoo<
A<0Hw>mm
easam. onmmmmoo¢
A<mmm>
«ssom.u mmoz< no
one<zzmozoo
us®~.l mmoz< mo
mmmzmm<z<
mo. mszHzmm
II m2~q=om<£

mzHADom<z

127
Summary

Subjects for this study were obtained from a systematic
sample of women employed by Michigan State University and
listed in the 1987-88 Faculty-Staff Directory. Data were
collected by a questionnaire mailed to the subjects as a
research survey conducted by the Employee Assistance Program
at Michigan State University. The questionnaire included the
Anger Self-Report, the Personal Attributes Questionnaire, and
the Texas Social Behavior Inventory. Demographic data and
information regarding the respondents' perceptions and use of
the services of the Employee Assistance Program were also
gathered. Confidentiality was assured since all
questionnaires were returned anonymously.

Four hypotheses were tested in this study. The first
hypothesis stated the expected differences in awareness of
anger, general expression of aggression, verbal aggression,
physical aggression, and condemnation of anger between women
with masculine or androgynous sex roles, and women with
feminine or undifferentiated sex roles. The second hypothesis
described the manner in which awareness of anger, general
expression of aggression, verbal aggression, physical
aggression, and condemnation of anger were expected to predict
self-esteem. Hypothesis 3 stated the expected differences in
self-esteem between the four sex role groups. The fourth
hypothesis described the expected variability, according to

sex role, of the manner in which awareness of anger, general

128
expression of aggression, verbal aggression, physical
aggression, and condemnation of anger were expected to predict
self-esteem. Multi-variate analyses of variance, stepwise
multiple regression analyses, one-way analyses of variance,
and planned and post hoc contrasts were used to test these

hypotheses.

CHAPTER V

ANALYSIS OF THE DATA

The analysis of the data collected for this research is
presented in Chapter V. The main variables in the study are
discussed first, followed by a report of the reliabilities
and descriptive statistics for each of the scales used in
measuring the variables. The demographic characteristics of
the sample are then described, followed by normative data on
the sample for each instrument. The main body of the chapter
is devoted to formal tests of the hypotheses. The chapter

concludes with an overall summary of the data analysis.

Main Variables

The main variables of interest in this study were sex
role, measured by the Personal Attributes Questionnaire (PAQ) ,
self-esteem, measured by the Texas Social Behavior Inventory
(TSBI), and awareness of anger, expression of anger, and
condemnation of anger, measured by the Anger Self-Report
(ASR). The ASR total expression of anger score is computed
by adding the scores of these three subscales: general
expression of anger, verbal aggression, and physical
aggression. Preliminary analyses indicated that more useful
information was available by analyzing the three subscales
separately, rather than as a composite score. The hypotheses

as they are stated and tested in this chapter reflect this

129

130
more precise measurement of expression of anger.

The primary dependent variable in this study was self-
esteem. The effects of the primary independent variables,
sex role and anger, in predicting self-esteem were explored
both separately and in interaction with each other. Anger
was also used as a dependent variable in the first analysis,
in which the differences in anger between women in the four
sex role groups were explored. Five demographic variables
were of interest in this research: marital status, age,
college degree, employee group classification, and household
income. These variables were used as independent variables
in two analyses. Their effects on anger were explored both
separately and in interaction with sex role in the first
analysis. The effects of the demographic variables on self-
esteem were explored in conjunction with anger in the second
analysis.

Reliabilities and descriptive statistics were computed
for each of the scales used in this study. Reliability was
tested by Cronbach's alpha, a measure of internal consist-
ency. Due to the manner in which the ASR scales are
constructed, two separate alpha statistics are reported for
each of the ASR scales. Each.ASR item is answered with either
an agree or a disagree response. All items scored in the
disagree direction are given a negative value in scoring.
Because of these negative values, it was necessary to compute
separate alphas for the agree and disagree items. Alphas

ranged from a high of .89 for the TSBI and .77 for the PAQ M

131
scale to lows of .32 for the agree items on the ASR physical
aggression scale and .37 for the disagree items on the ASR
general expression of anger scale. Reliabilities and
descriptive statistics are reported in Table 5.1. The alphas
for the agree items are designated with a +, those for the
disagree items with a -.
Demographic Characteristics

The sample for this study consisted.of 342‘women.employed
at a large midwestern university and ranging in age from 20
to 68, with a mean age of 39.6 years. The majority of the
women (60.5%) reported that they were currently married, 24%
indicated that they were formerly married (separated,
divorced, or widowed), and 15.2% stated that they had never
been married. The sample consisted primarily of well-
educated, middle class women, with an average educational
level of 15.4 years and a median household income of $30,000
to $34,999., Approximately half (51.1%) of the women were
employed in clerical or technical positions. Slightly over
one-quarter (29.1%) held administrative and/or supervisory
positions. Women in faculty or academic staff positions made
up 19.8% of the sample. Demographic characteristics of the

sample are summarized in Table 5.2.

Normative Data
S - e t
The women in this sample scored considerably lower on

the ASR awareness of anger scale than the introductory

Table 5.1:

132

Reliabilities and Descriptive Statistics

 

 

 

Measure/Scale Alpha Mean Std.Dev. Range

PAQ M Scale .77 22.17 4.8 0-32

PAQ F Scale .72 24.59 3.86 13-32

PAQ M-F Scale .53 14.13 4.29 0-27

TSBI .89 42.68 9.37 15-63

ASR Awareness .58+ * 28.86 9.92 -1-57
0 68- **

ASR Gen.Expression .43+ * 39.65 6.47 23-57
.37- **

ASR Ver.Aggression .55+ * 41.47 7.23 23-61
.46- **

ASR Phy.Aggression .32+ * 36.23 4.35 31-51
.53- **

ASR Condemnation .75+ * 41.60 6.01 33-62
.61- **

 

* Cronbach's alpha for
** Cronbach's alpha for

agree items
disagree items

 

133

Table 5.2: Demographic Characteristics of Sample

 

N %*

 

Marital Status

 

Never Married 52 15.2
Married 202 59.1
Formerly Married 80 23.4
Age
20-29 57 16.8
30-39 130 37.9
40-49 92 27.0
50 8 above 58 17.3

 

Educational Level

No degree 165 48.2
Bachelor's 81 23.7
Master's 53 15.5
Above Master's 38 11.2

 

Employee Classification

Clerical-Technical 170 49.7
Administrative/Supervisory 97 28.4
Faculty/Academic 66 19.3

 

Household Income

 

Below $20,000 52 15.2
20,000-29,000 84 24.5
30,000-39,000 51 15.0
40,000-49,000 42 12.3
50,000 8 above 104 30.4

 

 

* May not total 100% due to some subjects omitting item.

134

psychology students studied by Biaggio (1980b), with a mean
score of 28.86 (s.d. 9.92) compared to 39.12 (s.d. 13.11) for
Biaggio's subjects. Consistent with this result, the subjects
in this study scored considerably higher on the ASR
condemnation of anger scale than the college student sample,
with.a:mean score of 41.601 (s.d. 6.0) compared.to 32.92 (s.d.
6.93).

The scores for the university employees in this study
were similar to those of the college students (Biaggio, 1980b)
on the ASR general expression of anger, verbal aggression, and
physical aggression scales. The subjects in this study had
a mean score of 39.65 (s.d. 6.47) on general expression of
anger, compared to 36.61 (s.d. 9.48) for Biaggio's sample.
On verbal aggression the university employees in this study
had a mean score of 41.47 (s.d. 7.23) compared to 37.25 (s.d.
9.44) for the college student sample. The subjects in this
study had a mean score of 36.23 (s.d. 4.35) on the physical
aggression scale, while Biaggio's subjects had a mean of 34.27
(s.d. 6.11)

u s onnaire

The medians for the PAQ masculinity (M), femininity (F),
and M-F scales for this sample were slightly different than
those reported by Spence and Helmreich (1978) for a sample of
college students. The M and F scale medians were 22 and 24
respectively, compared to the college student sample medians
of 21 and 23. On the M-F scale the median was 14, compared

to 15 in Spence and Helmreich's sample. These discrepancies

135
may be accounted for by the fact that the college student
sample was composed of both males and females, while the
current sample was females only. Due to this fact, and.to the
restricted age range of the college student sample, the
medians for the university employee sample were used in
scoring the PAQ for this study.
ve o
The TSBI self-esteem scores of the women in this study
were somewhat higher than the scores for college students
reported by Spence and Helmreich (1978). The mean TSBI score
for the female university employees was 42.68 (s.d. 9.37)
compared to 38.74 (s.d. 7.86) for the female college students

and 38.88 (s.d. 7.27) for the male college students.

Hypotheses Tested
Multivariate analyses of variance, bivariate regression
analyses, analysis of variance, multiple regression analyses,
and planned and post hoc contrasts were used to test the main
hypotheses. The particular type of analysis used for each
hypothesis is given with the description of the results for

each hypothesis.

W

There will be differences between women who have
masculine or androgynous sex roles as measured by
the PAQ, and women who have feminine or undiffer-
entiated sex roles as measured by the PAQ, with
respect to scores on the ASR awareness of anger,
general expression of. aggression, verbal aggression,
physical aggression, and condemnation of anger
scales.

136

Null hypothesis:

There will be no difference between women who have

androgynous, masculine, feminine, or

undifferentiated sex roles as measured by the PAQ
with respect to scores on the ASR awareness of
anger, general expression of anger, verbal
aggression, physical aggression, and condemnation

of anger scales.

A multivariate analysis of variance was used to test this
hypothesis. The Wilks multivariate test of significance
indicated that differences existed between the sex role groups
on the 5 ASR scales (p = .000). The univariate F-tests
revealed that the sex role groups differed significantly on
the ASR awareness of aggression (p = .017) , general expression
of anger (p = .001), and verbal aggression (p = .000) scales.
The null hypothesis was, therefore, rejected. Table 5.3
contains the multivariate test of significance and Table 5.4

contains the univariate F-tests for sex role on the ASR

scales.

Table 5.3: Multivariate Tests of Significance

 

Test F-Value Sig. of F

 

Sex Role on ASR Scales
Wilks 4.33 .000
Marital Status on ASR Scales
Wilks 2.15 .019
Age on ASR Scales
Wilks 1.76 .036
Educational Level on ASR Scales

Wilks 1.91 .020

 

 

 

137

To test the specific parts of hypothesis one which are
stated below, the mean of the masculine and androgynous groups
was compared to the mean of the feminine and undifferentiated
groups on each of the five ASR scales. These planned
contrasts were tested at the .05 level. Post hoc pairwise
contrasts were then.performed to determine whether additional
differences existed.between individual sex role groups on ASR
awareness of anger, general expression of aggression, and
verbal aggression. Because post hoc contrasts require the use
of a family-based alpha level, the contrasts of individual
groups were tested at the .008 level of significance. The
results of both the jplanned and. post hoc analyses are
explained for each specific portion of the hypothesis. Table
5.5 contains the means and standard.deviations of the four sex
role groups on the five ASR scales.

WM:

Women with masculine or androgynous sex roles as

measured by the PAQ will score higher on the ASR

awareness of anger scale than women who have
feminine or undifferentiated sex roles as measured

by the PAQ.

Null hypothesis:

There will be no difference between women who have

androgynous, masculine, feminine, or undifferen-

tiated sex roles as measured by the PAQ on the ASR
awareness of anger scale.

The post hoc pairwise contrasts revealed that women
having a feminine sex role on the PAQ scored significantly

The planned contrast indicated that no significant difference

was found between the mean of the masculine and

Table 5.4:

138

Univariate F-Tests

 

 

 

Group Hypoth MS Error MS F Sig.
Sex Role on ASR Scales (DF 3, 278)
Awareness 330.45 95.16 3.47 .017
Gen Expression 235.92 40.71 5.80 .001
Verb Aggression 398.36 49.93 7.98 .000
Phys Aggression 5.57 18.93 .29 .829
Condemnation 63.66 35.63 1.79 .150
Marital Status on ASR Scales (DF 2, 279)
Awareness 157.26 96.58 1.63 .198
‘ Gen Expression 131.95 41.81 3.16 .044
Verb Aggression 188.55 51.57 3.66 .027
Phys Aggression 13.73 18.95 .73 .485
Condemnation 37.92 35.37 1.07 .344
Age on ASR Scales (DF 3, 281)
Awareness 646.88 91.50 7.07 .000
Gen Expression 131.71 45.20 3.12 .026
Verb Aggression 74.57 52.86 1.41 .240
Phys Aggression 13.37 18.92 .71 .549
Condemnation 53.31 35.75 1.49 .217
Educational Level on ASR Scales
Awareness 137.60 97.28 1.41 .239
Gen Expression 178.72 41.60 4.30 .006
Verb Aggression 93.81 52.65 1.78 .151
Phys Aggression 8.47 19.02 .45 .721
Condemnation 138.48 34.96 3.96 .009

 

 

139

 

 

Table 5.5: ASR Scores by Sex Role
Sex Role Mean Std.Dev. N
Awareness of Anger*
Undifferentiated 28.40 9.57 50
Feminine 32.00 10.35 60
Masculine 29.77 9.50 56
Androgynous 27.14 9.63 116
Total 28.92 9.88 282
General Expression of Anger*
Undifferentiated 38.52 5.79 50
Feminine 38.02 6.35 60
Masculine 42.59 6.63 56
Androgynous 39.61 6.51 116
Total Sample 39.67 6.54 282
Verbal Aggression*
Undifferentiated 38.76 6.68 50
Feminine 40.12 7.65 60
Masculine 40.93 7.76 56
Androgynous 42.20 6.55 116
Total Sample 41.69 7.33 282
Physical Aggression
Undifferentiated 36.46 3.51 50
Feminine 36.30 4.28 60
Masculine 36.75 4.83 56
Androgynous 36.10 4.47 116
Total Sample 36.34 4.33 282
Condemnation of Anger

Undifferentiated 43.32 5.73 50
Feminine 41.65 6.64 60
Masculine 41.71 5.04 56
Androgynous 40.98 6.11 116
Total Sample 41.68 5.99 282

 

 

 

* Significant differences between the groups were
this scale.

found on

140
androgynous groups and the mean of the feminine and
undifferentiated groups (t = -1.43, p = .155).

The planned contrast indicated that no significant
difference was found between the mean of the masculine and
androgynous groups and the mean of the feminine and
undifferentiated groups (t = -1.43, p = .155).

The post hoc pairwise contrasts revealed that women
having a feminine sex role on the PAQ scored significantly
higher'on‘the.ASR.awareness of anger scale (mean.= 32.00) than
women having an androgynous sex role (mean = 27.14, p = .002) .
It should be noted that this difference was in the opposite
direction of that predicted in hypothesis 1(a) . No other
significant differences between the sex role groups were found
on the ASR awareness of anger scale.

WM:

Women with masculine or androgynous sex roles as

measured by the FAQ will score higher on the ASR

general expression of aggression scale than women

who have feminine or undifferentiated sex roles as

measured by the PAQ.

Null hypothesis:

There will be no difference between women who have

androgynous, masculine, feminine, or

undifferentiated sex roles as measured by the PAQ

on the ASR general expression of aggression scale.

The planned contrast indicated that the masculine and
androgynous groups (mean.= 41.10) scored significantly higher
than the feminine and undifferentiated groups (38.27, t =
3.53, 1):: .000). The post hoc pairwise contrasts between

individual sex role groups indicated that women with a

masculine sex role scored higher (mean = 42.59) on the ASR

141

general expression of aggression scale than women with
undifferentiated (mean = 38.52), feminine (mean = 38.02), and
androgynous (mean = 39.61) sex roles (p = .001, .000, .004
respectively) . These results were consistent with
predictions, with the exception that no difference was
expected between the masculine and androgynous groups. The
null hypothesis was rejected.

W91:

Women with masculine or androgynous sex roles as

measured by the PAQ will score higher on the ASR

verbal aggression scale than women who have feminine

or undifferentiated sex roles as measured by the

PAQ.

Null Hypothesis:

There will be no difference between women who have

androgynous, masculine, feminine, or

undifferentiated sex roles as measured by the PAQ

on the ASR verbal aggression scale.

The planned contrast for verbal aggression showed that
the masculine and androgynous groups (mean = 43.56) scored
significantly higher than the feminine and undifferentiated
groups (mean = 39.44, t = 4.64, p = .000).

The post hoc contrasts revealed that women in both the
masculine (mean = 44.93) and androgynous (mean = 42.20) groups
scored significantly higher on the ASR verbal aggression scale
than women in the undifferentiated group (mean = 38.76) at
the p = .001 and .004 levels of significance, respectively.
Women in the masculine sex role group also scored higher than

women in the feminine sex role group (mean = 40.12, p = .000).

The null hypothesis was, therefore, rejected.

142
The results of the contrasts will not be reported for
the last two parts of hypothesis 1 pertaining to the ASR
physical aggression and condemnation of anger scales since
the univariate F-tests indicated no significant differences
between the sex role groups on these scales. These two parts
of hypothesis 1 are stated below, with the significance level
of the F-test.
W:
Women with masculine or androgynous sex roles as
measured by the PAQ will score higher on the ASR
physical aggression scale than women who have
feminine or undifferentiated sex roles as measured
by the PAQ.
Null Hypothesis:
There will be no difference between women who have
androgynous, masculine, feminine, or undifferen-
tiated sex roles as measured by the PAQ on the ASR
physical aggression scale.
The univariate F-tests showed no significant differences
between the sex role groups on ASR physical aggression (F =
.29, p== .829). The null hypothesis was, therefore, retained.
W:
Women with masculine or androgynous sex roles as
measured by the PAQ will score lower on the ASR
condemnation of anger scale than women who have
feminine or undifferentiated sex roles as measured
by the PAQ.
Null Hypothesis:
There will be no difference between women who have
androgynous, masculine, feminine or undifferentiated
sex roles as measured. by the PAQ on the .ASR
condemnation of anger scale.
The univariate F-tests showed no significant differences

between the sex role groups on ASR condemnation of anger (F

143
= 1.79, p = .150) . The null hypothesis was, therefore,
retained.
W

The effects of the five demographic variables of interest
in this study were explored for each of the ASR scales. A
multivariate analysis of variance was performed on the five
ASR.scales for the interaction.of each.of these variables with
sex role: marital status, age, educational level, employee
classification, and household income. There were no
significant interactions between sex role and any of these
variables on the five ASR scales. Multivariate analyses of
variance were performed for each of the demographic variables
alone on the ASR scales. Significant effects were found for
marital status, age and college degree. These results are
reported below.

W. Subjects were classified in three groups
according to marital status: never married, married, and
formerly married, including women who were separated, divorced
or widowed. The Wilks multivariate test of significance
showed significant differences between these groups on the ASR
scales (p = .019). The univariate F-tests showed significant
differences between the groups on the ASR general expression
of aggression and verbal aggression scales (p = .044, .027
respectively). Table 5.3 contains the multivariate test of
significance and Table 5.4 contains the univariate F-tests for

marital status on the ASR scales. Pairwise comparisons were

144
performed with a family-based alpha level of .05, with each
comparison tested at the .016 level of significance.

On the ASR.general expression of aggression scale, women
who had never been married scored higher (mean = 41.67) than
those who were formerly married (mean a 38.61, p = .013).
On the ASR verbal aggression scale, women who had never been
married scored higher (mean = 44.17) than married women (mean
= 40.95, p‘a .007). Table 5.6 contains the means and standard
deviations of the three marital status groups on each of the
ASR scales.

Age. Subjects were classified in four groups according
to age: 20-29 years, 30-39 years, 40-49 years, and 50 years
or above. The Wilks multivariate test of significance showed
differences between the age groups on the five ASR scales (p
= .036). The univariate F-tests indicated that there were
significant differences according to age on the ASR awareness
of anger (p - .000) and general expression of aggression
scales (p = .026). Table 5.3 contains the multivariate test
of significance and Table 5.4 contains the univariate F-tests
for age on the ASR scales.

Pairwise comparisons were performed with a family-based
alpha level of .05, with each comparison tested at the .008
level of significance. On the ASR awareness of anger scale,
women in the youngest age group (20-29 years) scored higher
(mean = 31.29) than women in the oldest age group (mean =
23.79, p = .000). Because the contrasts were tested at the
.008 level of significance, no significant differences were

found on the ASR general expression of aggression scale.Table

145

5.6: ASR Scores by Marital Status

 

Mean Std.Dev. N

 

Awareness of Anger

Never Married 31.26 9.23 46
Married 28.79 9.71 166
Formerly Married 27.99 10.47 70
Total Sample 28.99 9.85 282

General Expression of Anger*

Never Married 41.67 6.34 46
Married 39.58 6.63 166
Formerly Married 38.61 6.14 70
Total Sample 39.68 6.52 282

Verbal Aggression*

Never Married 44.17 7.79 46
Married 40.95 7.01 166
Formerly Married 41.83 7.17 70
Total Sample 41.69 7.25 282

Physical Aggression

Never Married 35.70 4.01 46
Married 36.56 4.49 166
Formerly Married 36.27 4.24 70
Total Sample 36.35 4.35 282

Condemnation of Anger

Never Married 40.63 5.51 46
Married 41.49 6.02 166
Formerly Married 42.27 6.04 70
Total Sample 41.55 5.95 282

 

 

 

* Significant differences between the groups were found on
this scale.

Table

 

 

 

5.7: ASR Scores by Age
Age Mean Std.Dev. N
Awareness of Anger*
20 - 29 31.29 10.10 51
30 - 39 30.74 9.23 110
40 - 49 28.09 10.03 77
50 or Above 23.79 8.94 47
Total Sample 28.98 9.87 285
General Expression of Anger
20 - 29 40.33 6.47 51
30 - 39 40.35 6.46 110
40 - 49 40.01 6.70 77
50 or Above 37.09 6.28 47
Total Sample 39.72 6.57 285
Verbal Aggression
20 - 29 41.84 7.74 51
30 - 39 42.25 7.39 110
40 - 49 41.99 7.57 77
50 or Above 39.72 5.80 47
Total Sample 41.69 7.29 285
Physical Aggression
20 - 29 36.94 4.57 51
30 - 39 36.20 4.20 110
40 - 49 36.43 4.93 77
50 or Above 35.70 3.30 47
Total Sample 36.31 4.34 285
Condemnation of Anger
20 - 29 41.35 5.79 51
30 - 39 40.97 5.60 110
40 - 49 41.70 6.16 77
50 or Above 43.15 6.70 47
Total Sample 41.60 5.99 285

 

* Significant differences between the groups were
found on this scale.

 

147
Table 5.7 contains the means and standard deviations of each
of the five age groups on the ASR scales.
W-

Subjects were classified in four groups according to
educational level: no college degree, bachelor's degree,
master's degree, and above master's degree. The "no college
degree" group included those with two-year associate's
degrees. The "above master's" group consisted of womengwho
had completed all requirements for a Ph.D except the
dissertation, or“whoihad.earned.a Ph.D., M.D., D.O., or D.V.M.
degree. The Wilks multivariate test of significance indicated
the existence of significant differences between the
educational level groups on the ASR scales (p = .020). The
univariate F-tests revealed significant differences on the ASR
general expression and condemnation.of.anger scales (p== .009,
.006 respectively). Table 5.3 contains the multivariate test
of significance and Table 5.4 contains the univariate F-tests
for educational level on the ASR scales.

Pairwise contrasts were performed with a family-based
alpha level of .05, with individual contrasts tested at the
.008 level of significance. On the general expression of
aggression scale, women in the "no college degree" group
scored lower (mean = 38.36) than women in the "above master's"
group (mean = 42.19, p = .000). On the condemnation of anger
scale, women in the "no college degree" group scored higher
(mean = 42.82) than those in the "above master's" group (mean

= 40.67, p = .001). Table 5.8 contains the means and standard

148

Table 5.8: ASR Scores by Educational Level

 

 

 

Mean Std.Dev. N
Awareness of Anger
No degree 27.84 10.51 140
Bachelor's 30.52 9.36 69
Master's 30.15 8.81 48
Above Master's 28.89 9.34 27
Total Sample 28.98 9.89 284
General Expression of Anger*
No degree 38.36 6.73 140
Bachelor's 40.86 5.91 69
Master's 40.46 6.44 48
Above Master's 42.19 6.31 27
Total Sample 39.69 6.56 284
Verbal Aggression
No degree 40.99 7.53 140
Bachelor's 42.20 7.37 69
Master's 41.35 7.14 48
Above Master's 44.33 5.41 27
Total Sample 41.66 7.29 284
Physical Aggression
No degree 36.27 4.48 140
Bachelor's 36.77 4.09 69
Master's 35.96 4.61 48
Above Master's 35.89 3.91 27
Total Sample 36.30 4.35 284
Condemnation of Anger*
No degree 42.82 5.90 140
Bachelor's 40.30 6.15 69
Master's 40.42 5.58 48
Above Master's 40.67 5.94 27
Total Sample 41.60 6.01 284

 

* Significant differences between the groups were found on
this scale.

 

149
deviations of each of the educational level groups on the ASR
scales.

W

Scores on the ASR scales will play a role in
predicting the self-esteem of individual women as
measured by the TSBI. More specifically:

(a). ASR awareness of anger scores will negatively
predict TSBI self-esteem scores.

(b). ASR general expression of anger scores will
negatively predict TSBI self-esteem scores.

(c). ASR verbal aggression scores will negatively
predict TSBI self-esteem scores.

(d) . ASR physical aggression scores will negatively
predict TSBI self-esteem scores.

(e). ASR condemnation of anger scores will posi-
tively predict TSBI self-esteem scores.

Null Hypothesis:
There will be no relationship between scores on the

ASR scales and the self-esteem of individual women
as measured by the TSBI.

W

A stepwise multiple regression analysis was used to test
Hypothesis 2. When the analysis was run with all five ASR
scales as independent variables, only ASR awareness of anger
and verbal aggression proved to be significant predictors of
self-esteem as measured by the TSBI. Together, these two
variables accounted for 17.15% of the variance in self-esteem
(R2 = .1715), and had a multiple R value of .4141. The
correlation coefficients, t-values and significance levels ,
and the analysis of variance for the regression equation are

reported in Table 5.9.

150

Table 5.9: Regression of TSBI on ASR Scales

 

 

Variable B t-value Sig. t
Verbal Aggression .56 7.02 .000
Awareness of Anger -.30 -5.12 .000
(Constant) 28.47 9.14 .000

Analysis of Variance

 

 

 

 

Source of Variance DF MS F Sig. F
Regression 2 2092.49 27.95 .000
Residual 270 74.87

 

Because ASR awareness of anger and verbal aggression were
significant in the regression equation, the overall null
hypothesis was rejected. The results were consistent with the
prediction of part (a) of hypothesis 2, that awareness of
anger would negatively predict self-esteem. The finding that
ASR verbal aggression predicted self-esteem in a positive
direction was the opposite of the negative relationship
predicted in part (b) of the hypothesis.

e ' es

The effects of the five demographic variables of interest
in this study were explored to determine whether they played
a significant role, along with the ASR scales, in predicting
self-esteem. Multiple regression analyses were performed for
each of these demographic variables with ASR awareness of
anger and verbal aggression: marital status, age, educational
level, employee classification, and household income. The
remaining.ASR.scales were not included.in these analyses since

it had previously been determined that only awareness of anger

151
and verbal aggression were significant predictors of self-
esteem.

Of the five demographic variables studied, only household
income was a significant predictor of self—esteem.when tested
with ASR awareness of anger and verbal aggression. Subjects
were classified according to five categories of household
income as follows: below $20,000: $20,000 - $29,999: $30,000
- $39,999: $40,000 - $49,999: $50,000 and above. When
household income was added to the regression equation with
awareness of anger and verbal aggression, R? increased from
.1813 to .2217, a significant increment of .0405 (p = .005).
Multiple R increased from .4257 to .4709. Because a family-
based alpha level was used, the significance of individual
groups was tested at the .005 level. An examination of the
individual groups revealed that membership in group 1 (below
$20,000) was a negative predictor of self-esteem. Household
income did not interact significantly with either awareness
of anger or verbal aggression. The correlation coefficients,
t-values and.significance levels, and.thetanalysis of variance
for the regression equation are reported in Table 5.10.
Mitosis—.3.

There will be differences in TSBI self-esteem scores

between the four groups of women having masculine,

androgynous, feminine, or undifferentiated sex roles

as measured by the PAQ. More specifically:

(a) . Women having androgynous sex roles as measured

by the PAQ will score higher on the TSBI than women

having masculine, feminine, or undifferentiated sex
roles as measured by the PAQ.

152

Table 5.10: Regression of TSBI on ASR Awareness of Anger, ASR

Verbal Aggression, and Household Income

 

 

 

 

Variable B t-value Sig.t
Awareness of Anger -.31 -5.55 .000
Verbal Aggression .55 7.24 .000
X3 -1.33 -1.21 .227
X2 1.97 2.14 .033
X1 -3.36 -2.97 .003
X4 1.27 1.07 .284
(Constant) 28.52 9.54 .000
Analysis of Variance
Source of Variance DF MS F Sig.F
Regression 6 967.07 13.58 .000
Residual 286 71.29

 

 

 

(b). Women having masculine sex roles as measured
by the PAQ will score higher on the TSBI than women
having feminine or undifferen-tiated sex roles on
the PAQ.

(c). Women having feminine sex roles as measured
by the PAQ will score significantly higher on the
TSBI than women having undifferentiated sex roles
as measured by the PAQ.

Null Hypothesis:
There will be no difference in TSBI self-esteem
scores between women having masculine, androgynous,
feminine, or undifferentiated sex roles as measured

by the PAQ.

A one-way analysis of variance was used to test
hypothesis 3. The analysis of variance indicated that
significant differences in self-esteem existed between the sex
role groups (F = 71.688, p = .000). The null hypothesis was,

therefore, rejected. Table 5.11 contains the analysis of

variance.

153

Table 5.11: Analysis of Variance for Self-Esteem on

 

 

Sex Role
Source of Variance DF MS F Sig. F
Between Groups 3 3676.21 71.69 .000
Within Groups 312 51.28
Total 315

 

 

 

The specific parts of hypothesis 3 pertaining to the
actual differences in self-esteem between the sex role groups
were tested at the .05 level of significance by pairwise
contrasts using the Tukey procedure. These contrasts
indicated that women in the masculine and androgynous sex role
groups scored higher on the TSBI than women in the feminine
and undifferentiated sex role groups. In addition, women in
the androgynous group scored higher than women in the
masculine group. These results were consistent with
predictions made in parts (a) and (b) of Hypothesis 3.
Contrary to the prediction made in part (c) of Hypothesis 3,
no significant difference was found between women in the
feminine sex role group and women in the undifferentiated sex
role group.

The above contrasts were performed using the family-based
alpha level employed. in ‘the 'Tukey’ procedure. Pairwise
contrasts using a less conservative contrast-based alpha level
were subsequently performed to determine whether a significant
difference in self-esteem existed between the feminine and

undifferentiated sex role groups. No significant difference

154
was found (t = -1.79, p = .075). Table 5.12 contains the
means and standard deviations for each of the sex role groups

on the TSBI.

Table 5.12: TSBI Scores by Sex Role

 

 

 

 

 

Sex Role N Mean Std.Dev.

Undifferentiated 57 34.54 7.02

Feminine 71 36.82 7.80

Masculine 61 45.33 6.86

Androgynous 127 48.56 6.99

Total Sample 316 42.77 9.26
es's

There will be an interaction between sex roles as
measured by the PAQ and ASR awareness of anger,
expression of anger, and condemnation of anger
scores in predicting TSBI self-esteem scores. More
specifically:

(a) ASR awareness of anger scores will:

(1) positively predict TSBI self-esteem scores
for women having masculine sex roles as
measured by the PAQ.

(2) negatively predict TSBI self-esteem scores
for women having feminine sex roles as measured
by the PAQ.

(3) not predict self-esteem scores for women
having' androgynous or 'undifferentiated sex
roles as measured by the PAQ.

Cb) ASR general expression of aggression verbal
aggression, and physical aggression scores will:

(1) positively predict TSBI self-esteem scores
for women having masculine sex roles as
measured by the PAQ.

(2) negatively predict TSBI self-esteem scores
for women having feminine sex roles as measured
by the PAQ.

155
(3) not predict TSBI self-esteem scores for
women having androgynous or undifferentiated
sex roles as measured by the PAQ.

(c) ASR condemnation of anger scores will:

(1) negatively predict TSBI self-esteem scores
for women having masculine sex roles as
measured by the PAQ.

(2) positively predict TSBI self-esteem scores
for women having feminine sex roles as measured
by the PAQ.

(3) not predict TSBI self-esteem scores for
women having androgynous sex roles as measured
by the PAQ.

Null Hypothesis:

There will be no interaction between sex role as

measured by the PAQ and ASR awareness of anger,

general expression of aggression, verbal aggression,
physical aggression, and condemnation of anger
scores in predicting TSBI self-esteem scores.

A stepwise multiple regression analysis was used to test
hypothesis 4. The analysis was performed twice. In the first
analysis, all five ASR scales were entered into the regression
equation along with the four sex role categories and
interactions of sex role with all the ASR scales. This
analysis was performed to determine whether the addition of
sex role to the equation would change relationships between
the ASR scales and the TSBI scores found in testing hypothesis
2. In that test, ASR.awareness of anger and.verbal aggression
were found to be significant predictors of self-esteem.

The results of this first analysis confirmed the findings
of the test of hypothesis 2 regarding the prediction of TSBI

scores by ASR scores. Awareness of anger and verbal

aggression again appeared to be the only ASR scales which

156
predicted self-esteem. The first analysis also showed that
sex role is a stronger predictor of self-esteem than any of
the ASR scales and does not interact with any of the ASR
scales.

Table 5.13: Regression of TSBI on ASR Awareness of Anger,
ASR Verbal Aggression, and Sex Role

 

 

Variable B t-value Sig. t
X3 3.2739 3.76 .000
X2 -3.4518 -4.15 .000
X1 -6.2124 -7.04 .000
Verbal Aggression .3271 4.68 .000
Awareness of Anger -.1728 -3.43 .001
(Constant) 33.0442 12.34 .000

 

 

 

Analysis of Variance

 

 

Source of Variance DF MS F Sig. F
Regression 5 2159.09 42.37 .000
Residual 267 50.96

 

 

 

A second analysis was then performed in which TSBI scores
were regressed on ASR awareness of anger and verbal
aggression, sex role, and the interaction of sex role with
the two ASR scales. This regression analysis showed that sex
role was the most important predictor of self-esteem,
accounting for 39.14% of the variance (R2 = .3914). ASR
verbal aggression and awareness of anger increased R2 by .03
and .02, respectively. Sex role did not interact with either

ASR awareness of anger or verbal aggression in predicting

157
self-esteem. It was, therefore, not possible to reject the
null hypothesis.

The final regression equation contained sex role,
awareness of anger, and verbal aggression, with a multiple R
of .67 and R?== .44. The correlation coefficients, t-values
and significance levels, and the analysis of variance are

contained in Table 5.13.

Summary

The results of the analysis of the data gathered in this
study are presented in Chapter 5. Reliabilities, computed by
Cronbach's alpha, and descriptive statistics were computed for
each of the scales used to measure the variables in the study.
Reliabilities ranged from highs of .89 for the TSBI and .77
for the PAQ M scale, to lows of .32 for the agree items on the
ASR physical aggression scale and .37 for the disagree items
on the ASR general expression of anger scale. Demographic
characteristics and normative data were reported for the
sample.

The first hypothesis was tested using a multivariate
analysis of variance. The main null hypothesis was rejected
since differences between the sex role groups on the ASR
scales were indicated. Significant differences between the
groups were found on the ASR awareness of anger, general
expression of aggression, and verbal aggression scales. Sub-
hypotheses were tested by planned contrasts, which showed that
the masculine and androgynous groups differed from the

feminine and undifferentiated groups on the general expression

158
of aggression and verbal aggression scales. Post hoc
contrasts were also performed to identity differences between
individual sex role groups on these two ASR scales and on ASR
awareness of anger.

The effects of five demographic variables were also
explored in the analysis of differences between the sex role
groups on ASR scores. These demographic variables were not
found to interact with sex role. Significant effects on the
ASR scores were found, however, for marital status, age, and
college degree.

The second hypothesis was tested by a stepwise multiple
regression analysis. ASR awareness of anger and verbal
aggression were found to be significant predictors of self-
esteem, allowing the null hypothesis to be rejected. The
effects of the five demographic variables, in conjunction with
ASR awareness of anger and verbal aggression, were explored
using multiple regression analyses. Only household income was
found to be a significant predictor of self-esteem. Income
did not interact with either ASR scale in predicting TSBI
scores.

Hypothesis 3 was tested using a one-way analysis of
variance. The null hypothesis was rejected since significant
differences between the sex role groups were found on self-
esteem. The sub-hypotheses pertaining to the specific
differences between the groups were tested by pairwise
contrasts using the Tukey procedure. As predicted, self-

esteem was found to be highest among the androgynous group,

159

followed in descending order by the masculine and feminine
groups. Contrary to predictions, there was no difference in
self-esteem between.the feminine and undifferentiated groups.

A stepwise multiple regression analysis was used to test
Hypothesis 4. The null hypothesis could not be rejected since
sex role did not evidence a significant interaction with any
of the ASR scales in predicting self-esteem. Sex role was the
most important predictor of self-esteem, with ASR awareness
of anger and verbal aggression accounting for only small
increments in the proportion of variance in self-esteem
accounted for by the regression equation.

The findings of the study are summarized and discussed,

and some conclusions suggested in Chapter VI.

CHAPTER VI

SUMMARY AND CONCLUSIONS

Summary of the Study

This study investigated the relationship between anger,
sex role and self-esteem in a large sample of women. The
study is rooted in a growing body of theoretical literature
on the psychological development of women which challenges
the traditional psychoanalytic view of women as inherently
passive, narcissistic, and.masochistic. Theorists argue that
powerful prohibitions against anger in the culturally
prescribed feminine sex role have resulted in intrapsychic
inhibitions of anger in women. Difficulties in the
separation-individuation process also leave women unable to
tolerate the sense of separateness accompanying anger and
fearful of their own destructiveness. This inhibition of
anger is thought to have serious negative consequences for
the mental health of women in drastic losses of self-esteem
when women become aware of angry feelings, threatening their
female identity, or“when anger'is repressed, becoming a source
of self-hatred.

Difficulty in handling anger is thought to be an
important factor in psychological difficulties such as
depression and attempted suicide, and in a wide range of
physical disorders. Although anger is considered an important

issue in psychotherapy with women, some researchers argue that

160

161

many therapists have difficulty in working with women's anger.
The purpose of this study was to obtain and analyze empirical
data on the relationship between anger, sex role and self-
esteem in order to explore the empirical foundations of one
aspect of female developmental theory, and provide insight
into how women integrate anger with sex role, and how anger
affects self-esteem.

In spite of the importance placed on anger by
theoreticians and practitioners, there have been few empirical
investigations of anger in women. Existing literature
provides limited support for the theoretical notion that anger
has been prohibited in. the feminine sex role. Empirical
investigations of the relationship between sex role, self-
esteem and psychological well-being show that masculinity and,
to a lesser extent, androgyny, are more strongly associated
with self-esteem and adjustment than femininity. This study
addressed the question of whether the hypothesized prohibition
of anger in the feminine sex role helps explain the strength
of the relationship between masculinity and self-esteem.

The primary dependent variable in this study was self-
esteem. The main independent variables were sex role and
anger. Anger was also treated as a dependent variable. In
addition, the effects of these five demographic variables were
explored: marital status, age, educational level, employee

classification and household income.

162

Self-esteem was measured by the Texas Social Behavior
Inventory (TSBI) . Sex role was measured by the Personal
Attributes Questionnaire (PAQ) . Awareness of anger,
expression. of anger’ and condemnation. of anger' were the
dimensions of anger of interest in this research, and were
measured by five scales of the Anger Self-Report (ASR) .
Cronbach's alpha was used to compute the reliability of the
TSBI, the PAQ, and the ASR. Correlations computed for all
the main variables showed substantial correlations between
the five ASR scales.

The sample for the study consisted of 342 women with a
mean age of 39.6 years employed at Michigan State University.
Subjects were generally well-educated and predominantly
middle-class. Half were employed in clerical-technical
positions, 30% held administrative positions, and 20% were
classified as faculty or academic staff. Participation was
voluntary and data were collected by an anonymous mail survey.

There were four main hypotheses in this study. The null
forms of'hypotheses 1, 2, and.3 were rejected. For hypothesis
1, which examined the differences in anger between the sex
role groups, significant differences were found between the
groups on awareness of anger, general expression of
aggression, and verbal aggression. Significant effects on
anger were also found for marital status, age, and educational
level. For hypothesis 2, which tested the ability of anger

to predict self-esteem, awareness of anger and verbal

163

aggression were found to be significant predictors of self-
esteem. Household income also predicted self-esteem.

Self-esteem was found to vary according to sex role, as
predicted by hypothesis 3. Self-esteem was highest among
androgynous women, followed by masculine women, and was lowest
among feminine and undifferentiated women. Hypothesis 4 could
not be rejected since sex role did not interact significantly
with any of the anger variables in predicting self-esteem.
Sex role was found to be the most important predictor of self-
esteem. The overall results of the hypothesis testing are

summarized in Table 6.1.

Conclusions
er c

It was hypothesized that the masculine and androgynous
groups would be higher than the feminine and undifferentiated
groups on awareness of anger, general expression anger, verbal
aggression, and physical aggression. Women in the masculine
and androgynous groups were expected to be lower than feminine
and undifferentiated women on condemnation of anger.

The masculine and androgynous groups together did not
differ from the feminine and undifferentiated groups on
awareness of anger. Feminine women had greater awareness of
anger than androgynous women.

The masculine and androgynous groups together were higher

on general expression of anger than the feminine and

164

Table 6.1: Overall Results of Hypothesis Testing

 

f j

Hypothesis 1

There will be differences between masculine or and androgynous
women and feminine or undifferentiated women, with respect to
awareness of anger, general expression of aggression, verbal
aggression, physical aggression, and condemnation of anger.

Result: Significant differences on awareness of anger,
general expression of anger, and.verbal aggression.

 

Hypothesis 2
Anger will play a role in predicting self-esteem.

Result: Significant for awareness of anger and verbal
aggression.

 

Hypothesis 3

There will be differences in self-esteem between the sex role
groups.

Result: Significant, with androgynous group scoring highest
and feminine and undifferentiated groups scoring lowest.

 

Hypothesis 4

There will be an interaction between sex role and anger in
predicting self-esteem.

Result: Not significant.

 

165
undifferentiated groups. Women with masculine sex roles had
greater general expression of anger than women in all other
sex role groups.

The masculine and androgynous women together were more
verbally aggressive than the feminine and undifferentiated
women. Both masculine and androgynous women were higher on
verbal aggression than undifferentiated women. Only masculine
women, however, perceived themselves as more verbally
aggressive than feminine women.

The women in this study reported equal levels of physical
aggression and condemnation of anger, regardless of their sex
roles.

a s ' nces n e

Women who had never been married showed greater general
expression of aggression than formerly married women, and
greater verbal aggression than married women.

D er s ' er

Women in the youngest age group, 20-29, were more aware
of anger than women in the oldest age group, 50 and above.

er d t ve

Women who did not have a college degree were less
expressive, and more condemning, of anger than women who had
attained education beyond a master's degree.

ti sh' of An e t Sel - st em
It was hypothesized that awareness and expression of

anger would be inversely related to self—esteem, and that

166
condemnation of anger would be positively associated with
self-esteem. —

Being aware of anger was negatively related to self-
esteemi for the women in this study. Verbal aggression,
however, was positively associated with self-esteem.‘

i In conjunction with the above relationships of awareness
of anger and verbal aggression with self-esteem, the lowest
level of household income (below $20,000) was associated
negatively with self-esteem.

Sex Role f e enc s in elf-Esteem

It was hypothesized. that women with androgynous sex roles
would have the highest self-esteem, followed in descending
order by masculine, feminine, and undifferentiated women.

Androgynous women had higher self-esteem than women in
the other sex role groups. Masculine women had higher self-
esteem than feminine and undifferentiated women. The self-
esteem of feminine women, however, was equivalent to that of

undifferentiated women.

Interaction of Sex Role and Anger
in Pnedicting Self-Esteem

For masculine women it was hypothesized that awareness
and expression of anger would be positively related to self-
esteem, and that condemnation of anger would be negatively
related to self-esteem. For feminine women, the predictions
were the inverse of those for masculine women. Self-esteem
was not expected to vary with differences in anger for

androgynous or undifferentiated women.

167
The effect of anger in predicting self—esteem was the
same for all sex role groups. Sex role was found to be a more

important factor than anger in predicting self-esteem.

Discussion of Findings

In. this section. the findings of this research are
discussed and observations are made regarding their
implications. The discussion is organized by headings which
correspond to those describing the major areas of conclusions
in the preceding section.

Se 0 e e es e

The finding that masculine and androgynous women were
more expressive of anger generally, and more verbally
aggressive, is supportive of the theoretical framework on
which this study is based. ‘Theorists contend that expres-
sion of anger and other forms of aggression, are more
acceptable as a masculine quality than as a feminine quality.
Empirical data (Atkinson 8 Polivy, 1976: Biaggio, 1988: Evans,
1984: Maccoby 8 Jacklin, 1974) support this notion, which is
also confirmed in the present findings. The fact that
androgynous women were not found to be more verbally
aggressive than feminine women, while masculine women were,
illustrates a further aspect of this differentiation between
the sex roles. Apparently, the feminine elements of androgyny
are sufficiently strong to balance the aggressive quality of

the masculine role.

168

The finding that feminine women are more aware of anger
than androgynous women is, at first glance, contrary to the
theoretical proposition that there are powerful intrapsychic
inhibitions on anger for the traditional female role. There
are two interesting possibilities which suggest themselves,
however, when this finding is contemplated. First, it may be
conjectured that it is the expression of anger which is more
powerfully inhibited intrapsychically than the awareness of
anger. This possibility is certainly consistent with the
powerful cultural prohibitions against the expression of anger
by ‘women in caretaking roles (Bernadez-Bonesatti, 1978:
Westkott, 1986). It.may be that feminine women are more aware
of their anger than women in a less traditional androgynous
role, but feel compelled to keep its expression under wraps
in order to conform to the cultural ideal for the feminine
woman.

Second, the finding of greater awareness of anger among
feminine than androgynous women may reflect the societal
conditions which some theorists argue are responsible for
women feeling they must conform to the traditional feminine
sex role. Miller (1983) has pointed out that the subordi-
nate status of women constantly generates anger. The
subordination is maintained, according to Miller and others
(e.g., Mueller' 8 Leidig, 1976), by economic and social
deprivation and/or physical force, and by the appearance that

there is no legitimate cause for a*woman's anger. The overall

169
result, according to the theorists, is that when a woman feels
angry, she concludes that there is something wrong with her
and feels that her identity as a woman is threatened. The
finding of low self-esteem for feminine women, coupled with
their higher awareness of anger, is consistent with this
theoretical position.

The lack of difference between the sex role groups on
physical aggression and condemnation of anger appears to be
consistent with the theoretical notion that there are strong
sanctions against anger for women. Although masculine and
androgynous women are more expressive of their anger, and more
verbally aggressive, they are as unlikely to be physically
aggressive as feminine and undifferentiated women, and are
equally condemning of anger. These findings suggest that
there are strong prohibitions against physical aggression in
women across all sex role groups.

The lack of difference between the sex role groups on
condemnation of anger also clearly implies that although
masculine and androgynous women may be more expressive of
anger, they may not feel positively toward and accepting of
their anger. One potential explanation is that the ability
to engage in and maintain empathic relationships may be
central to the identity of all women (Surrey, 1985),
regardless of sex role, and the experience and expression of
anger is threatening to this aspect of self and is thus

condemned. Thus, even masculine and androgynous women who

170
express their anger as part of the agency characteristic of
masculinity, may feel conflict between their more instrumental
orientation and that aspect of their identity as women in
‘which relationships are central. It is important.to note that
the mean score of 36.23 on the ASR Physical Aggression scale
is less than six points above the lowest.possible score on the
scale of 31, and that 31 was the modal response in this
sample. Thus, not only were there no sex role differences on
physical aggression, but the scores on the scale were
remarkably low.
t a s D' enc s ' n e

When considering the centrality of the relational self
to women's identity, it is also instructive to examine the
finding that women who had never been married showed greater
expressiveness of anger than both married and formerly married
women. Several possible explanations suggest themselves.
Perhaps the "never-married" women were less traditional than
women in the other two groups, and correspondingly more likely
to express their anger. A second hypothesis is that women who
are or have been married are less likely to express anger as
a result of experiencing the negative impact of anger in an
intimate relationship. 'This conjecture seems consistent with
the idea that anger may threaten the relational self for
women.

Yet another potential explanation is that women who have

never been married do not just express more anger, but are

171
actually more angry than currently or previously married
women. Possible reasons they could feel more anger include
the strong societal expectations for marriage, the loneliness
and social isolation of not having an intimate partner, and/or
the hardships of being the sole source of economic support for
oneself. The group of "never-married" women may also include
those who do not seek fulfillment of their need for intimacy
through marriage. Some of these women may be lesbians who
establish.primary relationships with other women. Regardless
of their sexual orientation, however, there may be women in
the "never-married" group who are angry at men and not
interested in primary relationships with them. The greater
expression of anger by these women may be due to their lack
of conformity with the traditional female role and/or their
anger at men.
W

The possible explanations for the youngest women having
greater awareness of anger than the oldest women parallel the
areas of conjecture regarding marital status. The youngest
women could be more aware of their anger as a result of less
experience in relationships than the oldest women, who may
have learned to repress awareness of their anger because they
found it threatened their identity as caretakers in
relationships. A more likely explanation, however, is that

the impact of the women's movement has been felt by the

172
youngest women, who may be less likely to feel constrained by
traditional role expectations than the oldest women.

A third alternative to consider is that perhaps the
youngest women feel cheated by the loss of the traditional
role and the accompanying ambiguity of current role
expectations for women in professional and family life.
Research focused on the areas in which anger is experienced
by women in various age groups would help to clarify these
possibilities. The finding that the youngest women had
greater awareness of anger than the oldest women is consistent
with the findings of Hertsgaard and Light (1984) in a study
of midwestern farm women.

e es ' E uc 'on e

The impact of the women's movement may also help to
account for the finding that women without college degrees
were less expressive and more condemning of anger than women
with education beyond a master's degree. Women with higher
levels of education would be more likely to have been exposed
to feminist ideas and more androgynous role models. This
finding in the present research is consistent, in part, with
findings of Martin and Light (1984), who found a curvilinear
relationship between educational level and hostility. In
their study, hostility was highest among women with a high
school education, and.then decreased.with educational levels,

up to postgraduate levels of education. Women with advanced

173
degrees were higher on hostility than those with four year
degrees.

It is also possible that the differences found in anger
by educational level may be more strongly related to
socioeconomic status than to education. Women with the lowest
level of education and no college degree are also likely to
occupy a lower socioeconomic position than the women with
education beyond the master's level. The women in the lower
socioeconomic levels may be less critical of their status, and
correspondingly less expressive of anger, than those at higher

socioeconomic levels.

e ' i A e t -Est

The finding that awareness of anger negatively predicted
self-esteem appears to be consistent with theoretical
arguments that anger is threatening to those relationships
which are a central factor in self-esteem for women (Kaplan
et al., 1983). Several other theorists have argued that the
experience of anger contradicts the woman's ego ideal
(Bernardez-Bonesatti, 1978: Miller, et al., 1981: Nadelson et
al. 1982: Norman, 1982), creating obvious difficulties in
self-esteem. The negative relationship between awareness of
anger and self-esteem found in this study appears to support
that theoretical position.

The finding in the present research that verbal

aggression was positively related to self-esteem does not

174
appear to be consistent with the theoretical argument
described above. It is important to remember, however, that
theorists also argue that anger which comes to awareness and
is then suppressed becomes a source of self-hatred and results
in a distorted sense of self (Kaplan, 1976: Miller, 1983:
Westkott, 1986) . It may be that anger which comes to
awareness but is not expressed has a negative impact on self-
esteem, but that once anger comes to awareness its expression
enhances self-esteem because it is externalized rather than
directed inward against the self. The present findings, that
awareness of anger negatively predicted self-esteem while
verbal aggression positively predicted self-esteem, appear to
support a formulation of this nature. When viewed as a whole,
then, the present findings regarding the relationship between
awareness of anger, verbal aggression and self-esteem, appear
to be consistent with the theoretical foundations of this
research.
c - e

The findings regarding sex role differences in self-
esteem were, for the most part, consistent with previous
research. In the majority of studies reviewed, self-esteem
was highest among androgynous subjects, followed by masculine
and feminine subjects (O'Connor, Mann 8 Bardwick, 1978:
Puglisi 8 Jackson, 1980-81: Spence 8 Helmreich, 1978: Spence,
Helmreich 8 Holahan, 1978: Spence, Helmreich 8 Stapp, 1975).

Unlike the present study, however, these researchers also

175
found that feminine subjects were higher on self-esteem than
undifferentiated subjects.

Two possible explanations may be suggested for the
finding that feminine and undifferentiated women did not
differ on self-esteem. First, the women in the present
research scored higher on masculinity than subjects in
previous research. The mean M scale score for the subjects
in this study was about 2.5 points higher than the female
subjects in Spence and Helmreich's (1978) research. The
median M scale score in the present study was 22, one point
higher than the median score reported by Spence and Helmreich.
Several studies have found a strong correlation between self-
esteem and masculinity (Spence 8 Helmreich, 1978: Spence,
Helmreich 8 Holahan, 1979: Spence, Helmreich 8 Stapp, 1975).
The higher degree of masculinity found in the women in the
present study may have accounted for the lack of difference
in self-esteem between feminine and undifferentiated women.
Even though women in both groups scored below the median on
the M scale, the masculinity scores may have been sufficiently
strong to correlate with self-esteem at equivalent levels.

The lack of difference in self-esteem scores between
feminine and undifferentiated subjects may also be explained
by the fact that the present research utilized only female
subjects, while the studies noted above used both males and
females. The inclusion of male subjects in the feminine and

undifferentiated groups may result in a greater differenti-

176
ation between the two groups. The lack of difference in self-
esteem between feminine and undifferentiated women was also
found by Gautheir and Kjervik (1982) in their study of female

graduate nursing students.

122W
W

Contrary to predictions, the effect of anger in
predicting self-esteem was not found to vary according to sex
role. Sex role, in effect, was found to be more important
than anger in predicting self-esteem as it was conceptualized
and measured in this study. An obvious explanation for this
finding is that the degree to which one identifies with
socially accepted sex roles is a primary determinant of self-
esteem and as such is more important than anger in determining
self-esteem. Since self-esteem was measured by the TSBI in
this study and is, therefore, conceptualized as a measure of
social competence, this explanation appears to have some
validity. One's social competence is obviously strongly
related to one's conformity to cultural standards for sex
role.

When the lack of interaction between sex role and anger
in predicting self-esteem is considered in depth, however, the
above explanation may be simplistic. This finding clearly
suggests that the socialization of women regarding anger is
sufficiently strong so that its impact on self-esteem as

measured in this study is equivalent whether a woman sees

177

herself as more traditional, i.e., feminine, or less
traditional, i.e. masculine or androgynous. The link.between
anger and self-esteem seems to be deeply rooted and not easily
changed by recent movements away from traditional roles for
women, and toward less traditional and more androgynous roles.
This explanation is clearly consistent with the centrality of
the relational self to women's self-esteem, and the threat
which anger is thought to pose to the nurturing and caretaking
roles assigned to women.

Another possible explanation exists for both the weak
relationship found between anger as an independent variable
and self-esteem as a dependent variable, and for the lack of
interaction between sex role and anger in predicting self-
esteem. It may be that a woman's self-esteem determines how
she deals with anger, rather than anger being a determinant
of self-esteem as hypothesized in this study.

In considering this explanation, it is important to
remember two of the core theoretical propositions upon which
this study is based. Theory suggests that the relational self
is a crucial factor in self-esteem for women, and that anger
has been prohibited in the nurturing and caretaking roles
which form the core of the relational self. If relational
qualities are a major component of self-esteem for women, the
possibility that self-esteem as currently measured by
psychological inventories may be a predictor of anger for

women.appears quite likelyu Thus the theoretical base of this

178
study supports the possibility that the relationship between
anger and self-esteem is the inverse of that explored in this
research.

Finally, consideration must be given to the possibility
that the measurement of anger in this study did not address
its relationship with self-esteem. It may be that there are
aspects of anger other than awareness, expression and
condemnation which have not been quantified for measurement
and which are strong predictors of self-esteem. Since theory
predicts that how anger is handled, and whether it is allowed
into awareness, are crucial factors in determining self-
esteem, consideration must be given to the possibility that
the measurement of anger in the current research was not
valid. The assessment of anger is relatively recent in
psychology, and it is true that existing instruments may be
considered somewhat crude compared to those utilized in
measuring more easily quantified attributes such as sex role
and self-esteem. In considering the validity of the
assessment of anger in this and other research involving
women, it must be remembered that the experience of anger for
women is exceedingly complex. Several aspects of anger which
may confound its accurate measurement are delineated in
theory.

First, consideration must be given to the theoretical
proposition that the cultural conception of anger is based on

the manner in which its expression is allowed in men (Miller,

179

1983). Cultural constraints have Zbeen ‘placed. upon the
expression of anger by men, resulting in a distorted
conception of anger which is linked to aggressive acting out
rather than direct expressions of anger. The attempt to
measure anger in this and other research may be contaminated
to an unknown extent by an overriding cultural distortion in
the understanding of the experience of anger. If subjects
consider anger equivalent to aggressive behavior, their
perceptions of their own anger may be similarly distorted.

A second distortion of anger, in which it is expressed
indirectly'as Thorizontal.aggression" (Symonds, 1976, p. 197),
may also confound attempts to measure anger. Symonds explains
that women accept this indirect solution to conflict out of
fear of threatening relationships with those upon whom they
depend. The goal of this indirect means of expressing anger,
according to Symonds, is to effect compliance through guilt.
Attempts to measure anger directly, such as that made in this
research, would not tap this potentially important means of
expressing anger.

Finally, Lerner (1980) has argued that because women have
difficulty tolerating the sense of separateness which
accompanies anger, they may feel hurt instead of, or in
conjunction with, anger. Feeling hurt rather than angry
serves a function similar to horizontal aggression in
attempting to protect the woman's relationships with important

others. Existing measures of anger, including that used in

180
this research, do not include the experience of hurt as a
possible expression of anger.

In summary, the findings of this research appear to be
generally supportive of recent theoretical work on the role
of anger in the psychological development of women. The study
confirms the existence of a relationship between anger, sex
role and self-esteem which is important in achieving a full
understanding of factors which are crucial in female
psychological development. The empirical data also provide
information valuable to mental health practitioners, such as
the different relationships awareness of anger and verbal
aggression have to self-esteem. Such insights could be
important to therapists working with women toward the
establishment and maintenance of self-esteem. The findings
also raise compelling questions about the nature of anger, and
the relationship between anger, sex role and self-esteem which
call for further exploration to clarify these important

issues.

Limitations
In this section the limitations of the findings and the
problems encountered in this research are discussed. Threats
to and limitations on external validity are presented first,
followed by difficulties with design and methodology, and

instrumentation.

181
W

The external validity of this study is limited by the
characteristics of the sample which was studied. The sample
consisted of well-educated, middle-class women employed at a
large university. Because of the method by which the sample
was obtained, no women employed below the clerical-technical
level were included. Groups which were excluded include
skilled trades, grounds and maintenance workers, housekeeping
staff, and food service workers. As a result, the findings
of this research can be generalized only to women who are
employed in clerical-technical, administrative, or faculty
positions in a university setting.

Generalization of the findings of this research is also
limited to women who would be likely to participate in a
voluntary research study conducted by a mail questionnaire.
Such women might be more committed to research or academic
freedom, more altruistic, more organized, or might have more
time for such pursuits because of favorable life circum-
stances. Women choosing not to participate, on the other
hand, may have difficult life circumstances which.would cause
them to feel more anger. If this were the case, the sample
would reflect a lower level of anger than that found in the
general population.

Additional limitations on the generalizability of the
findings and threats to this study's external validity may

exist in possible interactions between self-selection of

182

subjects to participate in the study, and the outcome of the
research. Because the study was conducted by an agency of
the university by which all the subjects were employed, the
women's feelings toward their employer may have influenced
their willingness to participate. It may be conjectured that
women who felt angry toward the university or their work
situation would be less likely to agree to participate in the
study.

In the same manner, any previous contact which the women
had with the university's Employee Assistance Program under
whose aegis the study was conducted may also have influenced
their willingness to complete the questionnaire. It seems
likely that women who had an unsatisfactory experience with
that agency would be less likely to agree to participate in
the study than those who had positive feelings toward the
agency.

Conversely, it is also possible that women who felt angry
about either their employment or the Employee Assistance
Program may have viewed participation in the study as an
opportunity to ventilate or take revenge for their feelings.
If this were the case, these women would have probably been
more likely to complete the questionnaire. Any strong
feelings which the subjects had toward either their employment
or the agency involved in conducting the research could

potentially have affected the responses.

183
MW

Two potential problems with the study's design and
methodology may have affected the outcome of the research.
First, subjects were initially contacted through a letter sent
through the university's interdepartmental mail to their work
place. Being solicited while in their work setting to
participate in a research study may have affected the response
of the subjects to the initial letter and their subsequent
willingness to cooperate in completing the questionnaire.

The collection of the data in this study through a mail
questionnaire posed an additional difficulty with the design
and methodology. Subjects who agreed to participate in the
study received the questionnaire at their home address. A few
women, however, requested that the questionnaire be sent to
their work place. The conditions under which subjects
completed the questionnaire were obviously outside the
researcher's control. Whether questionnaires were completed
at home, or in some other setting, the conditions probably
varied widely and had an undetermined effect on the results
obtained.

Mentoring

Difficulties with the instrumentation used in this
research exist in four areas: distortions due to the use of
self-report measures, questions regarding what is measured.by
the PAQ, the low reliabilities of some of the ASR scales for

this sample, and the use of male pronouns in some ASR items.

184

All three psychological inventories used in the study
are based on the use of self-report in responding. Distor-
tions and inaccuracies may result from subjects deliberately
falsifying their answers, engaging in unconscious self-
deception, lacking self-awareness, or being influenced by
social desirability. In addition, the ASR contains items
which refer to the subject's childhood, responses to which may
be inaccurate due to faulty memory.

The results obtained in the present research may have
been influenced by the use of the PAQ to measure the sex role
of the subj ects. Questions have been raised in the literature
regarding’ whether the 'masculine and feminine dimensions
assessed by the PAQ and other sex role inventories are related
only to personality attributes, or to sex-related behaviors
as well as personality traits. On the basis of their findings
in a study of psychological androgyny and sex role
flexibility, Helmreich, Spence, and Holahan ( 1979) have argued
that the traits measured by the PAQ are not significantly
associated with sex-related behaviors. If their contention
is correct, the correlations between sex role and awareness,
expression, and condemnation of anger obtained in the present
study may be weaker than the actual relationships between the
variables.

The findings in this research may also have been affected
by the low reliabilities obtained for three of the ASR.scales.

For ASR verbal aggression and physical aggression, reliabil-

185

ities for either the agree or disagree items were below .50.
For ASR general expression, reliabilities were below .50 for
both disagree and agree items. These reliabilities were
computed by Cronbach's alpha and reflect the internal
consistency of the subjects' responses. These results suggest
that the likelihood of obtaining equivalent results in a later
study may be questionable. The accuracy of the present
results may, therefore, be subject to some degree of question.

Finally, responses to the ASR may have been skewed due
to the use of male pronouns in some items. Some subjects
commented that they found the use of male pronouns in a
questionnaire for'women offensive. Others questioned whether
the items were intended to apply only to their interactions
with and feelings about men. It seems clear that male
pronouns did have some impact on the subjects' responses to
the.ASR items. The nature and extent of this impact, however.

is unknown.

Implications for Future Research
The present research has not demonstrated the strong
relationship between anger and self-esteem which is expected
on the basis of the theoretical literature. The exploration
of possible explanations for these results leads to the
consideration of implications for future research involving
the relationships between anger, sex role, and self-esteem.

Three aspects of“ the research. which could. be addressed

186
profitably in future research are the choice of
instrumentation to measure the main variables, the direction
of the relationship between anger and self-esteem, and the
study of a more diverse population of women.

For a number of reasons, consideration needs to be given
to the use of different measures for all three main variables.
In the area of self-esteem, Dorgan, Goebel and House (1983)
have pointed out that self-esteem is a complex concept, making
it problematic to conceptualize and measure appropriately.
These researchers obtained very different results in a study
of the relationship between sex role and self-esteem using two
different measures of self-esteem. More could be learned
about the relationship between anger and self-esteem if
various measures of self-esteem which conceptualize it
differently than the TSBI were used in future studies.

Problems with the reliabilities of the ASR scales
discussed above may also play a role in the lack of strength
in the relationship between anger and self-esteem found in
this study. The attempt to measure several dimensions of
anger in this study may have resulted in less precise and more
diffuse results. More could be learned about the relationship
between anger, sex role, and self-esteem in future studies by
using a more narrowly focused trait measure of anger.

The contention of Helmreich et a1. ( 1979) and Spence
(1983) that the PAQ and other sex role inventories measure

personality traits which are not strongly related to sex-

187

related behaviors also has implications for the instrumenta-
tion in future studies of the relationship between sex role
and anger. In their study of psychological androgyny and sex
role flexibility, Helmreich et al. found that scores on the
Attitudes Toward Women Scale, a measure of general sex role
attitudes, were more strongly related to behavioral
preferences than scores on the PAQ. Consideration could be
given to use of this inventory in future research.

In addition to the need for research using different
measures of anger, sex role and self-esteem, attention could
also be given to further exploration of the direction of the
relationship between anger and self-esteem. As noted
previously, it may be that the manner in which an individual
deals with anger is dependent on self-esteem, rather than the
opposite relationship which was tested in this study.
Establishing the nature and direction of the relationship
between anger and self-esteem has important implications for
the treatment of difficulties with anger in psychotherapy as
well as the understanding of female psychological development,
and is, therefore, an important avenue for future research.
If the handling of anger is dependent on self-esteem, working
toward awareness and expression of anger before healthy self-
esteem has been developed could result in clients being
inadequately prepared for the risks inherent in acknowledging,

experiencing, and expressing anger.

188

The findings of the present research are based on a
sample of well-educated, middle-class women holding clerical
and. professional positions in a large 'university. The
relationship between anger, sex role and self-esteem may be
very different among women having lower levels of education
and income, and either employed in service, maintenance, or
unskilled positions, or not employed outside the home. More
can be learned in future studies of women in lower
educational, socioeconomic, and occupational levels.

In conclusion, much remains to be learned about the
relationship between anger, sex role and self-esteem. The
findings of the current study suggest that additional
exploration of the nature of the relationship between anger
and self-esteem may have important implications for the
treatment of difficulties with anger in psychotherapy with
women, and.the understanding of the psychological development
of women. Avenues for future research include the use of
different measures of anger, sex role and self-esteem, a
different conceptualization of the direction of the
relationship between anger and self-esteem, and the study of
subjects from lower socioeconomic, educational and

occupational levels.

APPENDICES

APPENDIX A

INITIAL LETTER REQUESTING SUBJECTS' PARTICIPATION

APPENDIX A
INITIAL LETTER REQUESTING SUBJECTS' PARTICIPATION
Dear (name of respondent):

We need your help! You are part of a fascinating and
diverse group of women employed by Michigan State University.
You are probably very busy juggling the demands of your work
and your personal life. We at the EAP would like to know how
you feel about. yourself and the things that upset and
frustrate you. As a woman working at MSU, you may face a
variety of complex challenges. We would like to provide
services that will help MSU women deal with these challenges
more effectively and enjoy their jobs and lives more.

You are one of a small number of women working at the
University who are being asked to fill out a questionnaire
about your feelings and attitudes toward yourself. It will
take only about 20 minutes to complete. We need to find out
what problems and challenges are shared by most women and how
these issues change with differences in such areas as age,
marital status, and job classification. Your participation
is important in helping us understand these differences. We
can improve our services for women only if we have this
information.

Your participation is voluntary. You may choose not to
participate or not to answer certain questions on the
questionnaire without recrimination. Should you participate,
your response to the questionnaire will be completely
anonymous. Please let us know that you will help by filling
out the enclosed card and returning it by campus mail. We
will send a questionnaire and a return envelope to your home
address. There is no way we can identify you from the
questionnaire you return. Instead, we will send you a post
card which you can return to us separately, letting us know
that you have returned your questionnaire. We can then remove
your name from our mailing list for this study.

We will be happy to provide you with a summary of the
results of this research at your request. If you have any
questions, please call us at 355-4506.

We're looking forward to learning what you have to teach
us! Please complete the enclosed card and return it to us in
campus mail as soon as possible. Thank you so much for your
time and assistance.

Sincerely,
Tom Helma Cynthia Hockett
Program Coordinator Staff Counselor

189

APPENDIX B

INTEREST CARD ENCLOSED IN FIRST MAILING

APPENDIX B

INTEREST CARD ENCLOSED IN FIRST MAILING

 

 

Yes! I will help the EAP learn more about MSU's
women employees. Please send the questionnaire
to me at this address.

Name 3

 

Home Address:

 

 

 

 

190

 

APPENDIX C

FOLLOW-UP MAILING SENT ONE WEEK AFTER INITIAL LETTER

APPENDIX C

FOLLOW-UP MAILING SENT ONE WEEK AFTER INITIAL LETTER

Last week a letter asking you to help us learn more about
the needs of MSU's women employees was sent to you. ‘Your name
was selected in a random sample of women working at MSU.

If you have already returned the card telling us you are
willing to help, please accept our sincere thanks. You will
receive your questionnaire very soon.

If you have not yet returned the card, please do so
today. Because we have contacted only a small, but
representative, sample of MSU's women employees, it is
essential that we be able to include your views in our study.

If by some chance you did not receive the first letter,
or have misplaced the return card, please call us immediately
at 355-4506.

Tom Helma Cynthia Hockett
Coordinator Staff Counselor
Employee Assistance Program Employee Assistance Program

191

APPENDIX D

QUESTIONNAIRE

APPENDIX D

QUESTIONNAIRE

MSU WOMEN - AN ASSESSMENT OF YOUR

FEELINGS, CONCERNS AND NEEDS

Your answers to this questionnaire will help us
to better understand the needs of women working
at Michigan State University. Please take the
few minutes needed to answer the questions.
Your answers will enable us to provide more
effective services to assist you in meeting the
complex challenges you face.

Employee Assistance Program
205 Olds Hall
Michigan State University
East Lansing, MI 48824

Telephone (517) 355-4506

192

PLEASE NOTE:

Copyrighted materials in this document have
not been filmed at the request of the author.
They are available for consultation, however,
in the author's university library.

These consist of pages:

193-204

U-M-I

193

Please answer the following questions about your knowledge and use of
the Employee Assistance Program (EAP). Select the answer for each
question which is most accurate for you and circle the number to the
left of that answer.

1.

Did you know about the Employee Assistance Program (BAP) before you
received the letter about this needs assessment research?

1 NO
2 YES
Have you ever been to the EAP for help?
1 NO
2 YES
If yes, how helpful was your visit to the EAP? .
1 NOT AT ALL HELPFUL
2 SOMEWHAT HELPFUL
3 VERY HELPFUL
Have you ever attended a program or workshop presented by the EAP?
1 NO
2 YES
If yes, how helpful was the program or workshop?
1 NOT AT ALL HELPFUL
2 SOMEWHAT HELPFUL
3 VERY HELPFUL
Would you consider going to the EAP for help now?
1 NO
2 YES
If no, why not?
1 CONCERN ABOUT CONFIDENTIALITY

2 SAP HAS NOT BEEN HELPFUL IN THE
PAST

194
3 CANNOT COME BETWEEN 8 AM & 5 PM
4 LACK OF PARKING

5 OTHER

 

 

8. What type of additional services would you like the EA? to provide?
1 LONG TERM COUNSELING

2 MORE SUPPORT AND COUNSELING
GROUPS

3 MORE WORKSHOPS WITH OO-WORKERS
4 MORE EDUCATIONAL PRESENTATIONS

5 OTHER

 

 

The following items ask you to describe your reactions and
feelings when you are around other people. Each item has a scale,
marked with the letters A, B, C, D, and E, with (A) indicating "not at
all characteristic of me" and (E) "very characteristic of me," and the
other letters, points in between.

For each item, please circle the letter which best describes how
characteristic the item is of you.

1. I am not likely to speak to people until they speak to me.

 

A B C D E
NOT AT ALL NOT VERY SLIGHTLY FAIRLY VERY MUCH
CHARACTER- CHARACTER-
ISTIC OF ME ISTIC OF ME

2. I would describe myself as self confident.

 

A B C D E
NOT AT ALL NOT VERY SLIGHTLY FAIRLY VERY MUCH
CHARACTER- CHARACTER-

ISTIC OF ME ISTIC OF ME

195

3. I feel confident of my appearance.

 

A B C D E
NOT AT ALL NOT VERY SLIGHTLY FAIRLY VERY MUCH
CHARACTER- CHARACTER-
ISTIC OF ME ISTIC OF ME

4. I am a good mixer.

 

A B C D E
NOT AT ALL NOT VERY SLIGHTLY FAIRLY VERY MUCH
CHARACTER- CHARACTER-
ISTIC OF ME ISTIC OF ME

5. When in a group of peOple I have trouble thinking of the right
thing to say.

 

A B C D E
NOT AT ALL NOT VERY . SLIGHTLY FAIRLY VERY MUCH
CHARACTER- CHARACTER-
ISTIC OF ME ISTIC OF ME

6. When in a group of people, I usually do what the others want rather
than make suggestions.

 

 

A B C D E
NOT AT ALL NOT VERY SLIGHTLY FAIRLY VERY MUCH
CHARACTER- CHARACTER-
ISTIC OF ME ISTIC OF’ME
7. When I am in disagreement with other people, my opinion usually

prevails.

A B C D E
NOT AT ALL NOT VERY SLIGHTLY FAIRLY VERY MUCH
CHARACTER- CHARACTER-
ISTIC OF ME ISTIC OF ME

8. I would describe myself as one who attempts to master situations.

 

A B C D E
NOT AT ALL NOT VERY SLIGHTLY FAIRLY VERY MUCH
CHARACTER- CHARACTER-

ISTIC OF ME ISTIC OF ME

196

9. Other people look up to me.

 

A B C D E
NOT AT ALL NOT VERY SLIGHTLY FAIRLY VERY MUCH
CHARACTER- CHARACTER-
ISTIC OF ME ISTIC OF ME

10. I enjoy social gatherings Just to be with people.

 

A B C D E
NOT AT ALL NOT VERY SLIGHTLY FAIRLY VERY MUCH
CHARACTER- CHARACTER-
ISTIC OF ME ISTIC OF ME

11. I make a point of looking other people in the eye.

 

A B C D E
NOT AT ALL NOT VERY SLIGHTLY FAIRLY VERY MUCH
CHARACTER- CHARACTER-
ISTIC OF ME ISTIC OF ME

12. I cannot seem to get others to notice me.

 

A B C D E
NOT AT ALL NOT VERY SLIGHTLY FAIRLY VERY MUCH
CHARACTER- CHARACTER-
ISTIC OF ME ISTIC OF ME

13. I would rather not have very much responsibility for other people.

 

 

A B C D E
NOT AT ALL NOT VERY SLIGHTLY FAIRLY VERY MUCH
CHARACTER— CHARACTER-
ISTIC OF’ME ISTIC OF ME
14. I feel comfortable being approached by someone in a position of

authority.

A B C D E
NOT AT ALL NOT VERY SLIGHTLY FAIRLY VERY MUCH
CHARACTER- CHARACTER-

ISTIC OF‘ME ISTIC OF ME

197

15. I would describe myself as indecisive.

 

A B C D E
NOT AT ALL NOT VERY SLIGHTLY FAIRLY VERY MUCH
CHARACTER- CHARACTER-
ISTIC OF ME ISTIC OF ME

16. I have no doubts about my social competence.

 

A B C D E
NOT AT ALL NOT VERY SLIGHTLY FAIRLY VERY MUCH
CHARACTER- CHARACTER-
ISTIC OF ME ISTIC OF ME

The items below inquire about what kind of a person you think you
are. Each item consists of a pair of characteristics with the letters
A-E in between. For example:

Not at all Artistic A....B....C....D....E Very Artistic

Each pair describes contradictory characteristics-that is, you
cannot be both at the same time, such as very artistic and not at all
artistic.

The letters form a scale between the two extremes. Please circle
the letter which describes where you fall on the scale. For example, if
you think you have no artistic ability, you would circle A. If you
think you are pretty good, you might circle D. If you are only medium,
you.might circle C, and so forth.

1. Not at all aggressive A....B....C....D....E Very aggressive

2. Not at all independent A....B....C....D....E Very independent

3. Not at all emotional A....B....C....D....E Very emotional

4. Very submissive A....B....C....D....E Very dominant

5. Not at all excitable Very excitable in
in a major crisis A....B....C....D....E a major crisis

6. Very passive A....B....C....D.. E
7. Not at all able to de-

vote self completely

to others A... B . C....D....E
8. Very rough A....B....C....D....E
9. Not at all helpful to

others A....B....C....D....E
10. Not at all competitive A....B....C....D....E
11. Very home oriented A....B....C....D....B
12. Not at all kind A....B. .C....D....E
13. Indifferent to others’

approval A ...B....C....D....E
14. Feelings not easily

hurt A....B....C....D....E
15. Not at all aware of

feelings of others A....B....C....D....E
16. Can make decisions

easily A....B....C....D....E
17. Gives up very easily A....B.. C. ..D.. E
18. Never cries A ...B....C....D....E
19. Not at all self-

confident A....B....C....D.. E
20. Feels very inferior A.. .B.. C....D....E

Very active

Able to devote self
completely to others

Very gentle

Very helpful to
others

Very competitive

Very worldly

Very kind

Highly needful of
others’ approval

Feelings easily hurt

Very aware of feel-
ings of others

Has difficulty making
decisions

Never gives up easily

Cries very easily

Very self—confident

Feels very superior

199

21. Not at all understand- Very understanding of
ing of others A....B....C....D....E of others

22. Very cold in relations Very warm in rela-
with others A....B....C....D....E tions with others

23. Very little need for Very strong need for
security A....B....C....D....E security

24. Goes to pieces under Stands up well under
pressure A....B....C....D....E pressure

We would like you to consider carefully the following statements and
indicate as accurately as you can how it applies to you. There are no
right or wrong answers, we just want to know how you feel.

Please mark next to each statement according to the amount of your
agreement or disagreement by using the following scale:

1 STRONG DISAGREEMBNT 4 SLIGHT AGREEMENT
2 MODERATE DISAGREEMENT 5 MODERATE AGREEMENT

3 SLIGHT DISAGREEMENT 6 STRONG AGREEMENT

Please mark all statements!

If a statement is unclear to you place an "X" next to it in the
margin but mark it anyggy. If a statement somehow does not apply to
you, place a "?" next to it in the margin but mark it anygay.

l. I get mad easily.

2. I am often inclined to go out of my way to win a point with
someone who has opposed me.

3. I never feel hate towards members of my family.
4. People should never get angry.

5. It’s right for people to express themselves when they are
mad.

6. Some of my family have habits that bother and annoy me very
much.

10.

ll.
12.
13.
14.

15.
16.

17.
18.
19.
20.
21.
22.
23.
24.
25.

26.
27.

200

l STRONG DISAGREEMENT 4 SLIGHT AGREEMENT
2 MODERATE DISAGREEMENT 5 MODERATE AGREEMENT
3 SLIGHT DISAGREEMENT 6 STRONG AGREEMENT
When I get mad, I say nasty things.
Even when my anger is aroused, I don’t use strong language.
If I am mad, I really let people know it.
Sometimes I feel that I could injure someone.
I will criticize someone to his face if he deserves it.
I get into fist fights about as often as the next person.
People should never get irritated.

I find that I cannot express anger at someone until they have
really hurt me'badly.

Even when people yell at me, I don’t yell back.

At times I have a strong urge to do something harmful or
shocking.

I have many quarrels with members of’my family.
Feeling angry is terrible.

I wouldn’t feel ashamed if people knew I was angry.
I never do anything right.

It doesn’t make me angry to have people hurry me.
If I don’t like somebody, I will tell him so.

I have physically hurt someone in a fight.

At times I feel like smashing things.

Whatever else may be my faults, I never knowingly hurt
another person’s feelings.

I find it easy to express anger at people.

My parents never made me angry.

45.

201

l STRONG DISAGREEMENT 4 SLIGHT AGREEMENT
2 MODERATE DISAGREEMENT 5 MODERATE AGREEMENT

3 SLIGHT DISAGREEMENT 6 STRONG AGREEMENT

Even when someone does something mean to me, I don’t let him
know I’m upset.

At times I hurt a person I love.

I hardly ever feel like swearing.

I couldn’t hit anyone even if I were extremely angry.
I hardly ever get angry.

Even though I disapprove of my friend’s behavior, I just
can’t let them know.

I find it hard to think badly of anyone.

I can think of no good reason for ever hitting anyone.
When people are angry, they should let it out.

I am rarely cross and grouchy.

In spite of how my parents treated me, I didn’t get angry.
I could not put someone in his place even if he needed it.

When I really lose my temper, I am capable of slapping
someone.

It’s easy for me not to fight with those I love.

If someone annoys me, I an apt to tell him what I think of
him.

It’s useless to get angry.
If someone crosses me, I tend to get back at him.

I think little of people who get angry.

202

Finally - please answer the following questions.

What is your present marital status? Please circle the apppropriate
number.
1 NEVER MARRIED
MARRIED
DIVORCED

SEPARATED

macaw

WIDOWED

What is your present age?

How many years of education have you completed?
(For example - completed elementary education equals 8 years - high
school equals 12 years - 2 years of college equals 14 years, etc.)

 

 

If you have a college degree, please indicate level.

How many years of education have your parents completed?
(For example - completed elementary education equals 8 years - high
school equals 12 years - 2 years of college equals 14 years, etc.)
Father years
Mother years
Please circle the number next to your employee group.

1 CLERICAL-TECHNICAL UNION (CTU)

2 ADMINISTRATIVE-PROFESSIONAL
ASSOCIATION (APA)

3 ADMINISTRATIVE~PROFESSIONAL
SUPERVISORS ASSOCIATION (APSA)

4 FACULTY
5 ACADEMIC STAFF

6 FRATERNAL ORDER OF POLICE (FOP)

203

What kind of work do you do?

 

Example - secretary, office manger, advisor, professor, manager,
administrator.

What are your most important activities or duties?

 

Example - typing and filing, supervision, research, teaching.

What is your before-tax household income? Please circle the appropriate
number.

1 Below $10,000
$10,000 - 14,999
$15,000

19,999
$20,000 - 24,999
$25,000

29,999
$30,000 - 34,999
$35,000 - 39,999
$40,000 - 44,999

coooqmoaawm

$45,000 - 49,999

10 $50,000 or above

THANK YOU VERY MUCH FOR COMPLETING THIS QUESTIONNAIRE2!

204

Is there anything else you would like to tell us about
yourself or the services of the Employee Assistance Program?
If so please use this space for that purpose.

Any comments you wish to make that you think might help us
in future efforts to understand the feelings and needs of
women working at MSU would be appreciated either here or in
a separate letter.

Your contribution to this study is greatly appreciated. If
you would like to receive a summary of the results please
check the box on the enclosed return post card.

APPENDIX E

COVER LETTER ENCLOSED WITH QUESTIONNAIRE

APPENDIX E

COVER LETTER ENCLOSED WITH QUESTIONNAIRE
Dear (name of respondent):

Thank you for agreeing to help us learn more about MSU's
women employees! We appreciate your taking time in your busy
life to help us with this important project.

You are one of a small number of women we have
contacted. It is very important that we receive input from
women employed in many types of positions. Your response is
essential to the success of our efforts to learn about the
feelings and attitudes of MSU's‘women.employees so that.we can
tailor our programs to the challenges facing this unique
group.

It will only take about 20 minutes of your time to fill
out the questionnaire. Please return your completed
questionnaire to us in the postage-paid envelope we have
provided. Your responses will be completely anonymous. When
we receive the questionnaire, we will have no way of
identifying who completed it.

We have also enclosed a post card which we are asking you
to return to us WW. Please
print your name on this card to let us know you have returned
your questionnaire. Receiving this card will allow us to
remove your name from the mailing list for this research so
that you won't receive follow-up mailings. You can also check
the box on this card if you would like to receive a summary
of the research results.

Your participation is voluntary. If for any reason you
are unable to complete the questionnaire, please return it in
the enclosed envelope and send the post card with your name
on it separately.

If you have any questions, please call us at 355-4506.

Thank you very much for your assistance.

Sincerely,
Tom Helma Cynthia Hockett
Program Coordinator Staff Counselor

205

APPENDIX F

FOLLOW-UP MAILING SENT ONE WEEK AFTER QUESTIONNAIRE

APPENDIX F

FOLLOW-UP MAILING SENT ONE WEEK AFTER QUESTIONNAIRE

Last week a questionnaire was sent to you asking you to help
us learn more about MSU's women employees in order to improve
our services. Your name was selected in a random sample of
women working at MSU.

If you have already completed and returned it to us, please
accept our sincere thanks. If not, please do so today.
Because it has been sent to only a small, but representative,
sample of MSU's women employees, it is exremely important that
we have your input. Your participation will help us improve
our services for all women employees at MSU.

If by some chance you did not receive the questionnaire, or
it has been misplaced, please call us immediately at 355-4506.
We will send you another one today.

Tom Helma Cynthia Hockett

Coordinator Staff Counselor
Employee Assistance Program Employee Assistance Program

206

APPENDIX G

RETURN POST CARD ENCLOSED WITH QUESTIONNAIRE

APPENDIX G

RETURN POST CARD ENCLOSED WITH QUESTIONNAIRE

 

 

I have returned my questionnaire separately.

 

Your Name
(Please Print)

 

I would like to receive a summary of
the result of this survey.

 

 

 

Thanks again for your help.

 

207

 

 

APPENDIX H

EXPANDED ABSTRACT SET TO SUBJECTS
REQUESTING RESULTS SUMMARY

APPENDIX H
EXPANDED ABSTRACT SENT TO SUBJECTS
REQUESTING RESULTS SUMMARY

This descriptive study investigated the relationship
between anger, sex role, and self-esteem in adult women
university employees. Three of the four primary hypotheses
were supported. The effects of demographic variables on the
main variables were also investigated. The sample showed
differences in anger in four areas: sex role, marital status,
age, and educational level. Feminine women had the greatest
awareness of anger, but masculine, and to a lesser extent
androgynous, women were most likely to express their anger.
Contrary to predictions, no differences between the sex role
groups in physical expression of anger and condemnation of
anger were found. Women who had never been married were more
likely to express their anger than women who were or had
previously been married. The youngest women in the sample
were more aware of their anger than the oldest women. Women
without a college degree felt more condemning of their anger,
and were less likely to express it, than women with education
beyond the master's level.

The relationship found between anger and self-esteem is
complex. Self-esteem was negatively related to awareness of
anger, but positively related to verbal expression of anger.
Sex role differences in self-esteem were consistent with

earlier findings, with androgynous women having the highest

208

209
self-esteem, followed. by' masculine ‘womenm Feminine and
undifferentiated women reported the lowest self-esteem.

For masculine women, it was predicted that awareness and
expression of anger would be positively related to self-
esteem, and that condemnation of anger would be negatively
related to self-esteem. The inverse of these relationships
was predicted for feminine women. Self-esteem was not
expected to vary with anger for androgynous and
undifferentiated women. These predictions were not supported
since the effect of anger in predicting self-esteem was the
same for all sex role groups. Sex role was a more significant
predictor of self-esteem than anger.

The results.of this research suggest.that cultural values
prohibiting anger in women have a significant impact on self-
esteem regardless of whether a woman perceives herself in a
traditional or nontraditional role.

The findings of this research have important implications
for women who are employed in a university setting. Self-
esteem is an important factor contributing to one's sense of
fulfillment resulting from work endeavors. As such, it is
important that self-esteem.be developed and emphasized.within
the work setting. This research indicates that an important
element in the promotion of self-esteem may be for supervisors
and managers to allow and encourage the expression of
frustration and anger'by'employees. It is important that this

expression be viewed as a means of enhancing self-esteem and

210
the sense of empowerment and autonomy in one's work experience
rather than as a formalized system of complaining. Such
opportunities should be available to women (and men) in all
employment classifications, and are particularly crucial in
the university setting in which the values of equity and

personal growth are espoused.

APPENDIX I

COVER LETTER ENCLOSED WITH RESULTS SUMMARY

APPENDIX I
COVER LETTER ENCLOSED WITH RESULTS SUMMARY
Dear Survey Respondent:

About six months ago you were asked to participate in a survey
of MSU women employees conducted by the Employee Assistance
Program. You were a volunteer for this research and indicated
that you would like to receive a summary of the study's
results.

We greatly appreciate the time and effort you invested in this
effort. You played an important part in helping us learn more
about women who work at MSU. The results of this research
will be helpful to the Employee Assistance Program in
developing future programs for women.

The study has now been completed. We are enclosing a summary
of the results for your information. In addition, data
regarding the use of the Employee Assistance Program have been
compiled and will be used by the staff in planning future
programs. If you have any questions, please contact one of
us at the phone numbers below.

Again, thank you for your assistance with this important
project.

Sincerely,

Tom Helma Cynthia Hockett
Program Coordinator Consultant
355-4506 (616) 454-0112

211

LI ST OF REFERENCES

References

Adams, C. H., & Sherer, H. (1985). Sex-role orientation and
psychological adjustment: Implications for the
masculinity model. Sex_Bglea..12. 1211-1218.

Allen, J. G., & Haccoun, D. H. (1976). Sex differences in
emotionality: A multidimensional approach. nomon
BSlQEiQDfi: 22: 711-722.

Atkinson, C., 8 Polivy, J. (1976). Effects of delay,
attack, and retaliation on state depression and

hostility- 12urnal_9f_Abn2rmal_£§!thlggx. 35. 570-576.

Averill. J. R- (1982). Anger_and_aggressigni__An_e§sax_2n
emotion. New York: Springer-Verlag.

Averill, J. R. (1983). Studies on anger and aggression:
Implications for theories of emotion.

Esxghglggisf. 15. 1145-1160.
Bakan. D- (1966)- Ibe_dualitx_9f_buman_exi§ten§ei_lsglatign
anQ_92mm_nign_in_se§§ern_man

11 . . Boston 3 Beacon Press .

Balswick, J., & Avertt, C. P. (1977). Differences in
expressiveness: Gender, interpersonal orientation, and
perceived parental expressiveness as contributing

factors. l9urnal_9f.aarriage_and_tbe_£amilx 12. 121-
127.

Bardwick, J. M. (1971). Esyohology of women; A stody of
bigzgultural_22nfligts- New York: Harper 8 Row.

Barnett, R. C., & Baruch, G. K. (1985). Women's involvement
in multiple roles and psychological distress. loorno1_of
Rera9nalifx_and_§221al_2§12b21292. A2. 135-145-

Barrett, C. J., Berg, P. I., Eaton, E. M., & Pomeroy, E. L.
(1974). Implications of women's liberation and the

future of psychotherapy- Esxsbgfhera22i__1begrxl
Beseargb_and_2rastise. II. 11-15-

212

213

Baruch, G. K., & Barnett, R. (1986). Role quality, multiple
role involvement, and psychological well-being in midlife

women. 12urnal_2f_2era9nalif1__nd_§_gial_£§xs__i_gL 51.
578-585.

Bem, S. L. (1974). The measurement of psychological
androgyny. a sult c

BEYQDQIQQY: A2, 155-162-

Bem, S. L. (1978). Beyond androgyny: Some presumptuous
prescriptions for a liberated sexual identity. In J. A.
Sherman & F. L Denmark (Ede-). The_nsxeholggx_2f_xgmeni
Euture_direstigns_in_re§earsh (pp- 1-23)- New York:

Psychological Dimensions.

Berenson, M. L., Levine, D. M., & Goldstein, M. (1983).
ed at ' t al ods a 1 cat .
Englewood Cliffs, NJ: Prentice-Hall.

Berkowitz, L. (1962). 0 °
onolysis. New York: McGraw-Hill.

Bernardez, T. (1985, June). A psychosocial approach to
psychotherapy: Understanding society' 8 bias against

women. Badeliff__QuarterlL pp- 3-4-

Bernardez, T. (1985). Prevalent disorders of women:
Attempts toward a different understanding and treatment.
In C. T. Mowbrey, S. Lanir, & M. Hulce (Eds.), Womog g
menfal_healtbi__neg_diregfigns_f2r_gbange (Special double
issue of flomon_ono_1norooy, 3(3/4)). New York:

Harrington Park Press.

Bernardez-Bonesatti, T. (1976). Unconscious beliefs about

women affecting psychotherapy- ngrtb_£arglina_lgurnal_2f
Mental_flealth. 1(5). 63-66. .

Bernardez-Bonesatti, T. (1978). Women and anger: Conflicts
with aggression in contemporary women. 1oornol_of
W ' O a , 1;, 215-219.

Biaggio, M. K. (1980a). Anger arousal and personality
characteristics. 0 d c'

EEYEDQlQQYr 22: 352-356-

Biaggio, M. K. (1980b). Assessment of anger arousal.
J2urnal_9f_2er§2nalit2_A§§e§§menf. is. 289-298-

Biaggio, M- K- (1988 August) Sex_differenses_in_angeri_
Aro_th_y_rool? Paper presented at American Psychological
Association Convention, Atlanta, GA.

214

Biaggio, H. K., Supplee, K., 8 Curtis, N. (1981).
Reliability and validity of four anger scales. Qoorool

9f_2er§9nalif2_5§§essment. $5. 639-648-

Block, J. H. (1973). Conceptions of sex role: Some cross-
cultural and longitudinal perspectives. Amorioon

Rexsbglggist. 25(6). 512-526-

Block, J. H. (1978). Another look at sex differentiation in
the socialization behaviors of mothers and fathers. In
J. A. Sherman & F. L. Denmark (Eds.), Tho_p§yohology_ofi
° ° ' ' (pp. 29-87). New
York: Psychological Dimensions.

Burchardt, C. J. (1982). Psychological androgyny and
personality adjustment in college and psychiatric

populations. Sox_3olo§, a, 835-851.

Buss, A. H. (1971). Aggression pays. In J. L. Singer

(Ed-). The_22ntr2l_Qf_aggressien_and_xiglengei segnitixe
and_n§xsnglggieal_fagtgrs. (pp 7-18)- New York:

Academic Press.

Caplan, P. J. (1985). m ' s . New
York: E. P. Dutton.

Chesney, H. A. & Rosenman, R. A. (Eds.). (1985). Ango;_oog

O O ‘ O O O O ‘
v - . I - - . . . A

Washington, DC: ‘Hemisphere Publishing.

Chodorow, N. The_renredugfign_2f_m2thering- Berkley:
University of California Press, 1978.

Cline-Naffziger, C. (1974). Women's lives and frustration,
oppression, and anger: Some alternatives. Jon;no1_of
counseling_£§xehelgsx. 21. 51-56.

Constantinople, A. (1973). Masculinity-femininity: An
exception to a famous dictum? Esyohologiool_fiollotin,
89, 389-407.

Cox, S. (1976). s 0 ° ' e f.
Chicago: Science Research Associates, Inc.

Deutsch, H. (1930). The significance of masochism in the

mental life of women. International_lgurnal_2f
Baxehganalxsia..ll. 48-60.
Dillman, D. A. (1978). a' d s s-

toto1_oosign_mothod. New York: John Wiley & Sons.

215

Dollard, J., Miller, N. E., Doob, L. W., Mowrer, O. H., &
Sears, R- R. (1939)- Irustrat12n_and_aggre§§i_n. New

Haven: Yale University Press.

Dorgan, M., Goebel, B. L., & House, A. E. (1983). General-
izing about sex role and self-esteem: Results or effects?

Sex_321es. 2. 719-724-

Edmunds, G. (1980). Aggression (psychological). In R. H.
Wood (Ed-). EnsY212pedia_2f_slini§al_asse§§ment (Vol. 1.
pp. 551-558). San Francisco: Jossey-Bass Publishers.

Eichenbaum. L-. & Orbach. 8. (1983). understanding_wpmeni
e o . New York: Basic
Books.

Evans, R. G. (1984). Hostility and sex guilt: Perceptions
of self and others as a function of gender and sex-role

orientation. Sox_3o1o§, 19, 207-215.

Falbo, T. (1982). PAQ types and power strategies used in

intimate relationships. 2sx2h91992.9f.fl9men_nuarterlx.
g, 399-405.

Fenichel, O. (1945). c c o .
New York: W. W. Norton.

Freud, S. (1957). Mourning and melancholia. In J. Strachey

(Ed. and Trans. ), The standard editioh of the oomploto

psY2h_l9gi2_1_wgrk§_2f_§igmund_fireud (Vol 14. pp- 239-
238). London: Hogarth Press. (Original work published

1917).

Freud, S. (1961a). The economic problem of masochism. In
J. Strachey (Ed. and Trans.), The standatd edition of tho
22mp1ete_psx2h9l99i2a1.29rks_9f_§igmund_£reud (Vol 19
pp. 157-170). London: Hogarth Press. (Original work
published 1924).

Freud, S. (1961b). Some psychical consequences of the
anatomical distinction between the sexes. In J.

Strachey (Ed. and Trans. ), The standard edition of tho
g9mplete_p_22b_12sieal_29rk__2f_§igmu_d_£_eud (Vol 19.

pp. 243-258L London: Hogarth Press. (Original work
published 1925).

Freud. 8- (1965)- New_intr9du2t2r1_le2ture§_2n_psxgh2:
analyois. In James Strachey (Ed. and Trans.). New York:
W. W. Norton Company.

216

Frodi, A., (1977). Sex differences in perception of a

provocation. a survey- Eersept251_and_ugter_§kills. 41.
113-114.

Frodi, A., (1978). Experiential and psychological responses
associated with anger and aggression in women and men.
, T2, 335- 349.

Frodi, A., Macaulay, J., E Thome, P. R. (1977). Are women
always less aggressive than men? A review of the

experimental literature- Rsxsbglggigal_pulletin. 51.
634-660.

Galenson, E. (1976L Scientific proceedings - panel
reports: Psychology of women.
, 35, 141- 160.

Gauthier, J., & Kjervik, D. (1982). Sex-role identity and
self-esteem in female graduate nursing students. Soy

MI 9.: 45-55-

Gilligan. C- (1982). In_a_d1fIerent_xgigei__£sxehelpsisal
the9r2.and_wgmenls_dexelepmento Cambridge: Harvard

University Press.

Glass, G. V., 8 Hopkins, K. D. (1984). Stotiotioo1_mothoo§

in_edusatign_and_psxsb212gx (2nd ed-)- Englewood Cliffs,
NJ: Prentice-Hall.

Glazer, C. A., & Dusek, J. B. (1985). The relationship
between sex-role orientation and resolution of Eriksonian
developmental crises. §ox_3olo§, 1;, 653-661.

Gore, 8., & Mangione, T. W. (1983). Social roles, sex roles
and psychological distress: Additive and interactive

models of sex differences. J2urnal_9f_nealth_and_§221al
Bebaxigr. 21. 300-312.

Gormly, A. V., & Gormly, J. B. (1984). Research studies: A

psychological study of emotions. Ttahsoctiohal AhaTysTs
IQEIDQII 11: 74-79.

Greenberg, J. R., 8 Mitchell, S. A. (1983). ghioot
s o t' e . Cambridge, MA:
Harvard Unversity Press.

Greer, S, & Morris, T. (1978L The study of psychological
factors in breast cancer: Problems of method. Sooiol
Selense_§_nedisine. 12. 129-134-

217

Hazaleus, S. L., 8 Deffenbacher, J. L. (1985). Irrational

beliefs and anger arousal- J9urnal_9f_§211ege_§tudent
Personnel. 2Q. 47-52-

Heiser, P., 8 Gannon, L. (1984). The relationship of sex-
role sterotypy to anger expression and the report of

psychosomatic symptoms. §o3_RoTo§, 1Q. 601-611.

Helmreich, R., Aronson, E., 8 LeFan, J. (1970). To err is
humanizing--sometimes: Effects of self-esteem,
competence, and a pratfall on interpersonal attraction.

I .‘Lé. 259'
264.

Helmreich, R. L., Spence, J. T, 8 Holahan, C. K. (1979).
Psychological androgyny and sex role flexibility: A test
of two hypothesee- I9urnal_9f_£er§2nalitx_and_§osial

E§¥§h21291.11. 1531-1544-

Helmreich, R. L., Spence, J. T., 8 Wilhelm, J. A. (1981). A
psychometric analysis of the personal attributes
questionnaire. Sox_3o1oo, 1, 1097-1108.

Helmreich, R., 8 Stapp, J. (1974). Short forms of the Texas
social behavior inventory (TSBI), an objective measure of
self-esteem. Bullet1n_of.the_Esxgh2n9mis_aogietx. 1.
473-475.

Helmreich, R., Stapp, J. 8 Ervin, C. (1974). The Texas
social behavior inventory (TSBI): An objective measure
of self-esteem or social competence. JSAS gotolog_ot

Selested_noouments.1n_£sxsbolog¥. 1. 79-

Hertsgaard, D., 8 Light, H. (1984L Anxiety, depression,

and hostility in rural women. Rsxsbologioa1_aeports. 5..
573-574.

Heyman, S. R. (1977L Dogmatism, hostility, aggression, and

gggder roles. J2urna1_of_glinigal_£_xsh919gL 11. 594-

Hoffman, D. M., 8 Linda S. Fidell (1979). Characteristics
of androgynous, undifferentiated, masculine, and feminine
middle-class women. fiox_go1o§, 5, 765-781.

Hokanson, J. E., 8 Edelman, R. (1966L Effects of three
social responses on vascular processes. Journal of
o , o, 442- 447.

Hokanson, J. E., Willers, K. R., 8 Koropsak, E. (1968). The
modification of autonomic responses during aggressive

interchanqe- Journal_2f_£ersonalitx. 19. 386-404-

218

Holahan, C. K., 8 Spence, J. T. (1980). Desirable and
undesirable masculine and feminine traits in counseling

clients and unselected students. Journol_of_§oh§olt1hg
and_slinical_£s¥chologx. 11. 300-302.

Hoppe, C. M. (1979). Interpersonal aggression as a function
of subject's sex, subject's sex role identification,
opponent's sex, and degree of provocation. Journal_2f

Personality. 11. 317-329.

Jones, S. L., 8 Lamke, L. K. (1985). The relationship
between sex role orientation, self esteem, and sex-typed
occupational choice of college women. Poyohology_ot

E2men.Quarterlx. 2. 145-152.

Kagan, J. 8 Moss, H. A. (1962). °
. New York: Wiley.

Kandel, D. 8., Davies, M., 8 Raveis, V. H. (1985). The
stressfulness of daily social roles for women: Marital,

occupational and household roles. Jootho1_ot_3oolth_ohd
222121.2eh5212r. 11. 64-?8-

Kaplan, A. (1986). The "self-in-relation": Implications
for depression in women. Egyohothotooy, 21, 234-242.

Kaplan, A. G. (1976). Androgyny as a model of mental health
for women: From theory to therapy. In A. G. Kaplan 8 J.
P. Bean (Eds.). Be29nd_sex:role_stere2122es. (pp. 353-
362). Boston: Little, Brown 8 Co.

Kaplan, A. 6., Brooks, 8., McComb, A. L., Shapiro, E. R., 8
Sodano, A. (1983). Women and anger in psychotherapy.

Homen_and_1herapx. 2. 29-40.

Kimble, C., 8 Helmreich, R. (1972). Self-esteem and the

need for social approval. 2522n9n9m12_2§1en2e. 19(6).
339-342.

Kimlicka, T., Cross, H., 8 Tarnai, J. (1983). A comparison
of androgynous, feminine, masculine, and undifferentiated
women on self-esteem, body satisfaction, and sexual

satisfaction. 2s22h91291.2f_flomen_guarterlx. 1. 291-294.

Klein, H. M., 8 Willerman, L. (1979). Psychological
masculinity and femininity and typical and maximal

dominance expression in women. Journal_of_2ersonalitx
and_§ocial_£§29hologx. 11. 1059-2070.

Lee, A. G., 8 Scheurer, V. L. (1983). Psychological
androgyny and aspects of self-image in women and men.

Sex_Bole§. 2. 289-306.

219

Lenney, E. (1979). Concluding comments on androgyny: Some
intimations of its mature development. Sox_3o1oo, 5,

Lerner, H. (1977). Taboos against female anger. The
C .V ' g, 4.11.

Lerner, H. E. (1980). Internal prohibitions against female
anger. The1Amer1can_1surnal_of_2522hoanalxsis. 12. 137-
148.

Lerner. H. G. (1985). The_dance_2f_anger1__a_womanls_guide
t9_changins_the_patterns_2f_1ntimate_re1ationships

. New
York: Harper 8 Row.

Levitan, H. L. (1981). Patterns of hostility revealed in
the fantasies and dreams of women with rheumatoid

arthritis. 2s2ch2therapx_1_2§19bosomatics. 11. 34-43.

Light, H. K. (1984). Differences in employed women's
anxiety, depression, and hostility levels according to
their career and family role commitment. Eoyohologiool

BEEQILfi. 25. 290-

Maccoby, E. E., 8 Jacklin, C. N. (1974).
oox_oittotohoo§. Stanford, CA: Stanford University
Press.

Martin, R. E., 8 Light, H. K. (1984). Education: Its

positive impact on women. We]. 11.
401-405.

Megargee, E. I., 8 Hokanson, J. E. (1970). T na f
° d' u on n 'o
oholyooo. New York: Harper 8 Row.

Miller. J. B. (1976). WWW.

Boston: Beacon Press.

Miller, J. B. (1983). The construction of anger in women

and men. HQIE_IE_EIQQI§§§ 83(01). Wellesley, MA: Stone

Center Working Paper Series.

Miller, J. B. (1985). Women's mental health issues: moving
forward with awareness and program alternatives. In C.
T. Mowbray, S. Lanir, 8 M. Hulce (Eds. ), Womoh and month;
hea1tn1_New_directions_for_cbange. (pp. 29- 35) New

York: Harrington Park Press.

220

Miller, J. B., Nadelson, C. C., Notman, M. T., 8 Zilbach, J.
(1981). Aggression in women: A reexamination. In S.
Klebanow, (Ed.), h c s o '

(pp. 157-167). New York: Gardner Press.

Mueller, K., 8 Leidig, M. W. (1976). Women's anger and

feminist therapy. Ir9nt1ers1__Amlournal_9f_Eomenls
Studies. 1(3). 23-30.

Nadelson, C. C., Notman, M. T., Miller, J. B., 8 Zilbach, J.
(1982). Aggression in women: Conceptual issues and
clinical implications. In M. T. Notman, 8 C. C. Nadelson

(Eds.). The_22man_patient (Vol 3). (pp. 17-28). New

York: Plenum Press.

Notman, M. T. (1982). Feminine development: Changes in
psychoanalytic theory. In C. C. Nadelson, 8 M. T. Notman

(Eds.). Tbe_woman_patient (Vol. 1. pp 3-29)- New York:

Plenum Press.

O'Conner, H., Mann, D. W., 8 Bardwick, J. M. (1978).
Androgyny and self-esteem in the upper-middle class: A
replication of Spence. n

£l1nical_£§ychology. 49. 1168-1169.

;{ Orlofsky, J. L. (1977). Sex-role orientation, identity
formation, and self-esteem in college men and women. Soy
MI 3.! 561-575-

Orlofsky, J. L. (1980). Sex-role orientation. In R. H.

Wood (Ed.). En2y2l2ped1a_of_211nical_asses§ment (Vol. 2.
pp. 656-672). San Francisco: Jossey-Bass.

Orlofsky, J. L., 8 Windle, M. T. (1978). Sex-role
orientation, behavioral adaptability and personal

adjustment. §o3_3o1o§, 1, 801-811.

Parsons. T. & Balsa. R. F. (1955). Easilyl_§oeialization
ohd_ihtotoot1oh_ptooo_§. Glencoe, IL: The Free Press.

Pedhazur, E. J. (1973). holtiolo togtossioh in hehaviotol
zooootoh (2nd ed.). New York: Holt, Rinehart 8 Winston.

Pelletier, K. R. (1981). s e a er° A
n911st12_approach_to_prerenting_stress_disorder§. New

York: Dell.

Puglisi, J. T., & Jackson, D. w. (1931). Sex-role identity

and self esteem in adulthood. International_lournal_of
Ag1ng_and_fluman_neyelopment. 11(2). 129-138.

221

Rendely, J. G., Holmstrom, R. M., 8 Karp, S. A. (1984). The
relationship of sex-role identity, life style, and mental
health in suburban American homemakers: 1. Sex role,
employment and adjustment. Sox_3o1o§, 11, 839-848.

Richardson, D., Visel, A., 8 Taylor, S. P. (1980). Female
aggression as a function of attitudes toward women. Soy

89122. 2. 255-271-

Rosenkrantz, P., Bee, H., Vogel, S., 8 Broverman, I. (1968).
Sex-role stereotypes and self-concepts in college
students.

12, 287- 295.

Rubin, J. A., Provenzano, F. J., 8 Luria, z. (1974). The
eye of the beholder: Parents' views on sex of newborns.

Amer1ean_1ournal_of_9rthop§ych1atry. 11. 179-186.

Scheaffer, R. L., Mendenhall, W., 8 Ott, L. (1979).

E1ementary_suryey_sampling (2nd ed.). North Scituate.
MA: Duxbury Press.

Siegel, J. M. (1986). The multidimensional anger inventory.
12urnal_2f_2ers2na11ty_and_fiocial_2222bology

I .11.: 191'-
200.

Silvern, L. E., 8 Ryan, V. L. (1979). Self-rated adjustment
and sex-typing on the Bem sex-role inventory: Is
masculinity the primary predictor of adjustment? Soy
82122. 2. 739-763-

Spence, J. T. (1983). Comment on Lubinski, Tellegen, and
Butcher's ”Masculinity, femininity, and androgyny viewed
and assessed as distinct concepts".1ootho1_of

pers2nal1ty_and_§221al_EsychologL 11. 440-446.

Spence, J. T., 8 Helmreich, R. L. (1978). H1§EQIIDILY_E
tominihity. Austin: University of Texas Press.

Spence, J. T., 8 Helmreich, R. L. (1979). On assessing
androgyny. 2ex_Bole§. 1. 721-738-

Spence, J. T., Helmreich, R. L., 8 Holahan, C. K. (1979).
Negative and positive components of psychological
masculinity and femininity and their relationships to
self-reports of neurotic and acting out behaviors.

12urna1_of_2ers2nal1ty_and_SQQial_Psycbolggy. 11. 1673-
1532.

222

Spence, J. T., Helmreich, R., 8 Stapp, J. (1974). The
personal attributes questionnaire: A measure of sex role
stereotypes and masculinity-femininity. JSAS gotoTog_ot
1ele1ted_Document§_1n_£sxcbology. 1. 127.

Spence, J. T., Helmreich, R., 8 Stapp, J. (1975). Ratings
of self and peers on sex role attributes and their
relation to self-esteem and conceptions of masculinity
and femininity. 19urna1_9f.21rsonality_and_1221al
£§¥Eh919§!. 12. 29'39-

Spielberger, C. D., Johnson, E. H., Russell, S. F., Crane,
R. J., Jacobs, G. A., 8 Worden, T. J. (1985). The
experience of anger: Construction and validation of an
anger expression scale. In M. A. Chesney, 8 R. H.
Rosenman (Eds. ), An e ' 'o u

and_behay1oral_d1sorders (pp. 5- 30). Washington. DC:
Hemisphere Publishing.

Stake, J. E., 8 Orlofsky, J. L. (1981). On the use of
global and specific measures in assessing the.self-esteem
of males and females. Sox_3o1o§, 1, 653-662.

Stericker, A. B., 8 Johnson, J. E. (1977). Sex-role
identification and self-esteem in college students: Do
men and women differ? Sox_Bo1o§, 1, 19-26.

Surrey, J. L. (1984). Self-in-relation: A theory of

women's development. floth_1h_ptogto§§ 84(02).
Wellesley, MA: Stone Center Working Paper Series.

Symonds, M. (1974L The liberated woman: Healthy and
neurotic. Tbe_Amer12_n_1ournal_of_zsy_hoanalxs1§. 11.
177-183.

Symonds, M. (1976). Psychodynamics of aggression in women.

0 , go, 195-203.
Tavris. C- (1982). Anger1__Ibe_nisunderstood_em211on. New

York: Simon 8 Schuster.

Terman, L. M., 8 Tyler, L. E. (1954L Psychological sex

differences. In H. S. Langfeld (Ed), nohool_ot_ohilg
noxohology (pp. 1064- -1114L New York: John Wiley 8

$0118 .

Tesch, S. A. (1984). Sex-role orientation and intimacy
status in men and women. §o3_BoTo§, 11, 451-465.

Wallerstein, J. S. (1986). Women after divorce:
Preliminary report from a ten-year follow-up. hmotiooh
Journal_of_grtbopsychiatry. 11. 65-77.

223

W ' w ce tu 'ct'on . (1979). New
York: Simon and Schuster.

Weissman, M., Fox, K., 8 Klerman, G. L. (1973). Hostility
and depression associated with suicide attempts.

Amer12an_lournal_of_fsycb1atry. 112. 450-455.

Weissman, M. M., 8 Klerman, G. L. (1977). Sex differences

and the epidemiology of depression. Azohiyo§_ot_§ohozol
BEYthALIY. 11. 98-111-

Westkott, M. (1986). Historical and developmental roots of
female dependency. Poyohothotopy, 21, 213-220.

Whitley, B. E. (1984). Sex-role orientation and
psychological well-being: Two meta-analyses. §o3_3o1o§,
12, 207-225.

Williams, R. B., Barefoot, J. C., 8 Shekelle, R. B. (1985).
The health consequences of hostility. In M. A. Chesney 8
R. H. Rosenman (Eds.), Ange; and hostility in
card19yas2ular_and_behayioral_disorders (pp. 173-185).
Washington, D.C.: Hemisphere Publishing.

Yoder, J. D., Rice, R. W., Adams, J., Priest, R. F., 8
Prince, H. T. (1982). Reliability of the attitudes
toward women scale (AWS) and the personal attributes
questionairre (PAQ). fioy_3oio§, 8, 651-657.

Zealley, A. K. (1971). Bronchial asthma: A problem
attributable to sampling when establishing its

psychopathology. 2sych9tberapy_and_25ycbosomat1cs. 12.
37-46.

Zelin, M. L., Adler, G., 8 Myerson, P. G. (1971). Anger
self-report: An objective questionnaire for the
measurement of aggression. J o and

£11n1251_£sychology. 12. 340.