5.1.... t:!..-| v v .vv (‘1! MMMMMMMMMMMMMMMMMMMMMMMMMMMOMMMM MMMM MMMMMMMMMMMMMMM ‘ LIBRARY Michigan State University L 4 ——f ———— This is to certify that the dissertation entitled AN INVESTIGATION OF THE RELATIOIxSHIPS BETWEEN HOSTILITY, SELF-ESTEEM, AND SEX-ROLE IDENTIFICATION IN ADULT WOMEN presentedby Diane E. Trebilcock has been accepted towards fulfillment of the requirements for Ph.D degree in Educam LTJ/LCLQ ! Major professor Date 4/§€)(/fi© MS U is an Affirmative Action/Equal Opportunity Institution 0-12771 PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before one due. DATE DUE DATE DUE DATE DUE 34:31}, i} M995 II I WM _ ;é; 0 [ 2031 H I 1 2 0 4 (M 1 ll ll ' MSU Is An Affirmative Action/Equal Opportunity Institution , ammo-m AN INVESTIGATION OF THE RELATIONSHIPS BETWEEN HOSTILITY, SELF-ESTEEM, AND SEX-ROLE IDENTIFICATION IN ADULT WOMEN By Diane B. Trebilcock A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Counseling Psychology Program Department of Counseling, Educational Psychology, and Special Education T990 ABSTRACT AN INVESTIGATION OF THE RELATIONSHIPS BETWEEN HOSTILITY, SELF-ESTEEM, AND SEX-ROLE IDENTIFICATION IN ADULT WOMEN By Diane B. Trebilcock This study investigated the relationships among self-esteem, sex-role identification, and hostility in adult women. The theoretical base for this research was developmental and psychoanalytic. The Bem Sex-Role Inventory, Buss-Durkey Hostility Scale, and Rosenberg Self-Esteem Scale were administered to l05 adult working women. Subjects also provided demographic information about themselves and responded to questions designed by the researcher. Pearson correlations were used to measure the relationships between the main variables of self-esteem, sex-role identification, overt hostility, and covert hostility. Qualitative information about the subjects was measured in relationship to the main variables by use of one-way ANOVAs and Scheffe post-hoc comparisons. It was found that global hostility was negatively related to self-esteem, self-esteem was positively related to masculine sex- role identification, androgyny was positively related to self- esteem, masculine sex-role identification was negatively related to Diane B. Trebilcock covert hostility, androgyny was negatively related to covert hostility, feminine sex-role identification was negatively related to overt hostility, and masculine sex-role identification was positively related to overt hostility. Findings also indicated that self-esteem was related to income, feminine sex-role identification was related to employment, masculine sex-role identification was related to employment, and masculine sex-role identification was related to education. In researcher-designed items, which were not statistically tested, subjects indicated a desire not to have to choose between relationships and personal goals, an equal valuing of self and other, and the experience of feeling bad about themselves when angry. These findings support theoretical notions that women feel negative about themselves when angry and that masculine and androgynous sex-role identifications are related to higher self- esteem and provide women with a broader range of behaviors. I dedicate this work with love to my mother, Corinne A. Trebilcock, and my sisters, Cheryl H. Trebilcock Greer and Kathy A. Trebilcock Sromalski. iv ACKNOWLEDGMENTS At the completion of this project I would like to thank my committee for their support: my chairman, John Powell, for his tireless effort, encouragement, and kind spirit; Imogen Bowers for her assistance and the inspiration she provided by her example; Richard Johnson for his clarity of thinking and gentle guidance; and Gersh Kaufman for being supportive and staying with me to the end. I wish also to thank Sue Cooley, whose fine mastery of style and presentation made this manuscript possible. I would also ‘like to acknowledge Bill Mueller, who helped provide vital building blocks to my development as a therapist-- trust in myself and faith in my unconscious. My friends who 'were endlessly loving, supportive, and encouraging of me throughout, I thank for their time, sustenance, and comic relief: Jim Brady, Kathy Long, Shelly Smithson, and Carolyn Paulter. My gratitude also goes to Karyn Boatwright and Polly Kent, my dearest friends, who held on to my hope for me in my darkest hours so that I could go on, celebrated with me in moments of joy, and literally loved me to success. Thanks also to Mary Ann Stehr, whose steadfast belief in me allowed for the unfolding of my possibilities. And finally, thank you to my family, who have stuck with me through it all. vi TABLE OF CONTENTS Page LIST OF TABLES ....................... ix Chapter 1. STATEMENT OF THE PROBLEM .............. 1 II. REVIEW OF THE LITERATURE .............. 8 Theoretical Literature .............. 8 Analytic Theorists ............... 8 Relational Theorists .............. 11 Social Theorists ................ 15 Integrative Theorists .............. 18 The Research Literature .............. 31 Sex-Role Identification ............. 31 Anger, Hostility, and Aggression ........ 35 Self-Esteem ................... 40 Summary ...................... 45 III. DESIGN OF THE STUDY ................ 48 The Sample Population ............... 49 Method and Procedure ............... 49 Data Analysis ................... 52 Variables ..................... 52 Definitions of Variables ............ 52 Hypotheses .................... 54 Measure of Self-Esteem .............. 55 Measure of Hostility ............... 56 Measure of Sex-Role Identification ........ 58 IV. ANALYSIS OF RESULTS ................. 6l Results on the Main Variables ........... 62 Sex-Role Identification ............. 62 Self-Esteem ................... 64 Hostility .................... 65 Hypothesis Testing ............... 67 Results of Supplementary Analyses ......... 70 vii Hostility Subscales ............... 70 Analysis of Variance .............. 72 Rank-Ordered Items ............... 74 Summary of Results ................ 79 V. CONCLUSIONS ..................... 82 Limitations of the Study ............. 82 Review and Discussion of the Results ....... 85 Main Variables ................. 85 Hypothesis Testing ............... 86 Analysis of Variance .............. 90 Rank-Ordered Items ............... 91 Summary ..................... 94 Recommendations for Future Research ........ 98 APPENDICES A. PERMISSION LETTER FROM THE UNIVERSITY COMMITTEE ON RESEARCH INVOLVING HUMAN SUBJECTS .......... 99 B. INFORMED CONSENT FORM AND DEMOGRAPHIC INFORMATION SHEET ........................ 100 C. ANOVA TABLES .................... 103 REFERENCES ......................... 107 viii Table A ##-h &#-h-h .IO .11 LIST OF TABLES Demographics ..................... Frequency Distributions on Main Variables ...... Frequency Distribution: Rosenberg Self-Esteem Scale . Frequency Distribution: Buss-Durkee Hostility Inventory ..................... Frequency Distribution: Hostility Subscales ..... Summary of the Pearson Corre1ation Coefficients: Self-Esteem .................... Summary of the Pearson Correlation Coefficients: Hostility ..................... Pearson Correlation Coefficients: Hostility * Self-Esteem and Hostility * SRI .......... Rank-Ordered Items: Defend Against Criticism Rank-Ordered Items: Relationship Versus Goals . . . . Rank-Ordered Items: Statements of Values ...... Behavior of Subjects, by Income and Educational Level ....................... Results of the Analyses of Variance ......... ix Page 50 63 64 65 66 67 68 71 75 76 77 CHAPTER I STATEMENT OF THE PROBLEM This study investigated the relationships among self-esteem, sex-role identification, and hostility in women. 'HJ date, little attention has been given to empirical studies on how the indirect expression of anger, aggression, and/or hostility affects the self- esteem of women. Yet low self-esteem, lack of assertiveness, and fear of their own aggression and anger have been shown to be problems for women in our culture. Various attempts have been made (e.g., assertiveness training, workshops on anger) to address these problems. However, only recently have theoretical concepts been developed to help understand relationships between variables important both to sound theoretical development and to practical approaches for assistance. Further, almost no significant attention has been given to the cultural context and its effect on the expression of anger and aggression and its relationship to self- esteem in women. In a culture that maintains prohibitions against the expression of anger or assertion in women, stereotypes of feminine behavior bind women with strong feminine identification, giving them no acceptable manner of expressing anger. These feelings are often expressed in nonconstructive, hostile behavior and can damage self-esteem. It is behavior, however, that fits readily with negative feminine sex-role stereotypes. In many respects, psychological literature has portrayed women in a manner that reflects a restricted position in society, without acknowledging the effect of cultural and societal pressures. Women have been described in the literature as masochistic, less advanced developmentally, passive, dependent, hysterical, and neurotic (Deutsch, 1945; Freud, 1965). They have been identified as victims and held responsible for their victimization. Women have been seen as victims of their own interpsychic weakness (Freud, 1965; Kohlberg, 1981). To the extent that these views accurately reflect the position of women in society, it is little wonder that women have suffered from problems of self-esteem. While women have been described as chronically struggling with low self-esteem, little serious attention has been paid to how the negative stereotypes and the societal restriction of women have contributed to this. Freud suggested that in order to be healthy and productive, one had to experience satisfaction in love and work. Yet society relegated to women only the realm of love. There is overwhelming evidence of wOmen’s comparative lack of advancement in the realm of work and their frequent status as victims. Until recently, little effort was made to integrate an understanding of the effect of the societal context into an understanding of these dilemmas for women. Recent theorists (Bernardez, l978; Flax, 1978; Westkott, 1986) have addressed the difficulties women have with aggression, anger, hostility, competition, and achievement as these relate to their struggle in developing high self-esteem. They have integrated into their theories the variable of the caretaker role women have been assigned in this culture. These theorists have described how women’s fear of their own aggression and the societal pressures to be nonaggressive have, in combination, held women in inferior positions, often creating depression and low self-esteem. Other theorists, in a broad sense, have offered important views. Psychoanalytic, developmental, and object-relations theorists have pointed out the importance of having access to both the libidinal and aggressive drives for healthy development (Freud, 1955; Horner, 1979; Kohut, 1971). Horner (1979) and Kohut (1971) described the importance of the individual’s need to express the aggressive drive in separation, while at the same time maintaining the internal and external relationship to the object as basic to the regulation of self-esteem. "When healthy aggression is not available for competence and achievement, the development of self- esteem is also blocked" (Horner, 1979, p. 211). Recent literature on women’s development has focused on the investment women have in relationships and how their identity and self-esteem are often tied to or even embedded in those relationships (Chodorow, 1978; Gilligan, 1983; Surrey, 1984). This refocusing in developmental theory has suggested differences in men’s and women’s development and helped to remove the pathological overtones to descriptions of women’s development and functioning. This literature has aided in identifying the competence, strength, and responsibility that women have developed in the caretaker role. It has been suggested women’s identity is embedded in relationships (Kaplan, 1984), that women have a need for affiliation (Miller, 1976), and that they possess a disposition to nurture (Chodorow, 1978) and a sense of responsibility for others (Gilligan, 1982). The historical mode of basing developmental norms on men’s development exclusively and thus suggesting that women’s development is abnormal or less advanced (Freud, 1965; Kohlberg, 1981) has been responded to with force and clarity. However, in the process of re- evaluating previously accepted notions, it is not helpful to women to focus on the positive role relationships have played in women’s lives to the exclusion of examination of the problems women have with separation and aggression in the context of those relationships. Other recent writers have identified that women’s concern with attachments and relationships is not "simply connective" but "constricting and binding" (Eichenbaum & Orbach, 1988). It is suggested that many women lose track of their own identities early in development. They thus live out their lives in submissive, accommodating roles because of the over-determined focus on relationships, and the internal and external prohibitions against women’s use of aggression on their own behalf (Bernardez, 1978; Flax, 1978; Horney, 1967; Westkott, 1986). One of the powerful forces maintaining the prohibitions against women’s aggression and anger is sex-role stereotypes. The feminine stereotype in this culture prohibits aggression and anger except in the service and defense of a child or mate. Aggressive strivings and nurturant feelings are seen as mutually exclusive (Bernardez, 1978; Chodorow, 1978). The nurturer role is female, and the aggressive, achieving role is male, in a culture where aggressiveness and achievement are valued (Broverman, 1972). It is believed that this nonaggressive, compliant nurturing role is passed on from mother to daughter from the very beginning of the female infant’s development (Bernardez, 1978; Chodorow, 1978; Horney, 1967; Westkott, 1986). A second set of powerful forces contributing to women’s difficulty with natural self-assertion, aggression, and anger is the prohibitions learned early in childhood. The strides of female children toward mastery, competence, and achievement are frequently redirected toward the care and consideration of others. Frequent punitive responses or no response to natural self-assertion leads to a fear of self-assertion and a primitive unmodulated rage in the child. This frightful rage, in conjunction with the unconscious memories of mother’s anger, sets up strong internal prohibitions against both the expression of anger and assertive striving on one’s behalf (Bernardez, 1978; Westkott, 1986). These early learnings are often sealed in place by the reactions of fathers who devalue their daughters’ assertive strides and sexualize their relationships (Horney, 1967; Westkott, 1986). As a result, assertion, aggression, and rage are repressed. This powerful unconscious material expresses itself in depression, seductiveness, manipulation, passive aggressive behavior, submission, dependence, and contempt, behaviors that are not often acknowledged as derivatives of anger. 'They are the expressions of powerless, frightened, and rageful women trying to have an effect on others. If her own aggression and assertion are forbidden, she will enlist the aggression of others. Any self- assertive strivings become potentially contaminated with hostility, a reflection of unconscious anger and fear. This unfortunately results in a self-fulfilling prophecy in that others then react negatively to hostility-laden behavior. Assertive strides may be followed by retreats and depression because of the anticipated punishment, guilt for the anger, and fear or loss of her relationships (Bernardez, 1978; Westkott, 1986). Women’s self-esteem and identity could thus be embedded in relationships not because of a predisposition to nurture and feelings of empathy or connectedness, but because self-assertive strivings must be kept unconscious and can be expressed only through others or on behalf of others. They are not available for independent action. Relationships become the vehicle for expression of aggression. The repression and unconscious anger are intensified over time by the continuation of patterns in adult life that prohibit independent striving, but instead support and encourage investing in relationships as a means to get what one wants and needs. The problem for women, then, is how to regain access to natural self-assertion and aggression, to use it. hi one’s behalf, and thus to build and maintain self-esteem. This must be done in the face of societal pressures against it. Authors who have addressed issues of helping women develop self-esteem frequently have focused on encouraging and supporting identification and development of positive attributes. Assertiveness training, as an example, is included as a part of what is needed for women to feel better about themselves. Seldom are other issues of aggression, such as competitiveness, rage, jealousy, envy, and hostility, addressed directly. Yet for women to feel better about themselves on an ongoing basis, these feelings and issues must also be addressed, normalized, and integrated in a more constructive manner. To empower women, what is needed is more than acknowledging strengths or changing nonassertive behavior. The beginning steps in empowering women include moving out of the victim role by reowning anger and aggression, redirecting hostility, and giving legitimacy to feelings. This is a difficult step for women because these feelings are thought to be in conflict with the role of caretaker. This study investigated the relationship of hostility and self- esteem in women’s experience. It sought to identify how women feel about themselves, how they see themselves relative to sex-role stereotypes, and how they experience their anger and hostility. The results may provide insight into ways in which clinicians can help women move to more powerful and less victimized positions in their own perceptions and in the world. CHAPTER I I REVIEW OF THE LITERATURE The basis for women’s difficulty with development of self- esteem appears to be related to poor access to healthy, natural aggression, self-assertion, and anger. This difficulty has its roots in long-standing social-cultural norms which profoundly influence early development and adult behavior. Earlier psychological theorists and researchers studying women’s development often highlighted different aspects of the developmental process without acknowledging the social/cultural context. In more recent years, however, theorists have tried to integrate into their explanations influences of social contexts on the development of self-esteem and access to aggression. This chapter contains a review of the theoretical literature on these aspects of women’s development, followed by a summary of research literature. Theoretical Literature Analytic Theorists Freud’s psychoanalytic theory has greatly affected the general view and understanding of human behavior. Yet Freud ignored the larger social-cultural environment in which he developed his theory. He theorized that women developed a sense of inferiority based on the discovery that they, in contrast to men, do not have a penis. Thus, Freud explained that women did not esteem themselves as they would a man because they lacked and envied possession of this all- important biological part. The envy was believed to contaminate their natural aggressive strivings and undermine development of healthy self-esteem. If women exercised their contaminated aggressive strivings, they would be viewed as castrating. If they did not exercise aggression, they were seen as passive and dependent. If they turned their aggression against themselves in the form of somatic problems and complaints, they were seen as hypochondriacal, neurotic, and hysterical. In a view that reflects the status and roles of’ women in the historical and cultural environment of his times, Freud purported that, for most women, self-esteem was to be gained by being a good helpmate to one’s husband and a good mother to one’s children. Freud’s attitude toward women is also revealed in his decision to withdraw his original stance regarding the sexualization of women and incest. His original suggestion was that fathers’ sexualization of their daughters was a significant contributor to the development of neurosis in women. When this was met with outrage from the patriarchal establishment, Freud withdrew this notion and replaced it with his theory of little girls’ fantasy of sexually seducing their fathers. Freud submitted to societal pressure and, instead of challenging the stereotype, sealed it; hi place. Women’s inferior status was a given, which she was to accept in order to be healthy. But again, Freud saw this as based in biological inferiority; the 10 lack of a penis and women’s resulting envy were the source of her problems (Freud, 1965). Adler (1956), a student of Freud, recognized in his theory that it is the woman’s situation in the family and culture that results in her feelings of inferiority, not her lack of a penis. According to Adler, women see and experience the universal dominance of men and the privileges of men denied to women. Since power and privilege are tied to gender, women come to view their femininity as without value. Women’s inferiority complex would come from the shaming rejection of their femininity as a valued asset. Adler saw the power structure being such that women were limited primarily to the options of pursuing power through the manipulation of men or by living through their children. Neither option would result in high esteem of one’s self. Horney (1967), also a student of Freud’s, objected to Freud’s view and also viewed women’s behavior and development in the context of the culture. Central to Horney’s theory was that women develop in the context of a culture that devalues them. It was Horney’s notion that women simultaneously integrate this devaluation as an accurate statement about 'themselves and fight against it. The devaluation makes it difficult for women to develop high self- esteem. The conflict about the devaluation, according to Horney, leaves women with anger that contaminates natural aggressive drives and ambition, seeps into her relationships, and is turned against the self and thus lowers self-esteem. 11 Thompson (1964) also rejected Freud’s explanation of women’s position being determined by biological inferiority. Thompson viewed the penis as merely symbolic of the privileges men had that women did not have in a patriarchal society. Jealousy and envy are characteristic of all individuals who live in a competitive culture and would be more problematic for any groups that are disadvantaged in terms of privileges. Thus, if women do display more of these emotions, it is likely to be reflective of their disadvantaged status rather than reflective of neurosis. Thompson drew specific attention to how the integration of these negative, inferior, and disadvantaged positions would negatively affect women’s self- perception and self-esteem (Williams, 1977). Re tional h ori s A more recent school of thought, the relational theorists, postulates that even with acknowledgment of the social/cultural bias, women’s behavior and development are defined in pathological terms based on male norms. These theorists take a more positive stance about the development of both women and men while pointing out significant and legitimate differences. Relational theorists believe the roles assigned according to gender result in men’s identity being defined in terms of separation and competition, while women’s identities are based in relationships. This theory generally purports to explain women’s inhibition of aggression and anger as based in the fear of losing relationships. Women feel caught between caring for self or other. 12 It is suggested that women’s difficulty with self-esteem is related to social devaluing of' women as well as the general lack of acknowledgment that their conflict between self and other is legitimate. One of the relational theorists, Surrey (1984), based her view of women’s development in the notion of "self in relation." She believed that the mutual empathy between mother and daughter is also the basis for mutual self-esteem. According to Surrey, mothers have more capacity for empathy with their daughters. Women’s basic experience of self is relational, organized in important relationships. She contrasted this with male development, which she proposed to be based in disconnection, separation, and assertion because of their need to disidentify with mother. Also core to the self-worth of women is the need to understand the other in order to further the development of self and other. Another relational theorist who addresses moral development in her formulations is Gilligan. A student of Kohlberg, Gilligan (1982) found unacceptable Kohlberg’s description of women’s moral development as less advanced and more immature than men’s. Instead, Gilligan suggested that women’s deference to others comes not only from social subordination but from their values and moral concerns, in which their sensitivity to the needs of others and assumption of responsibility to others are primary. Gilligan saw women as defining themselves in the context of relationships and judging themselves by their ability to take care. This basic assumption of caring for others and protecting human bonding has priority for 13 women over the personal power of the individual. Personal power is the primary value held by men, to which women are negatively compared. It is also a value to which they compare themselves. Gilligan stated that women discover and hold on to the importance of intimacy, understanding, and relationship from early life, while men must rediscover it in middle life. Women’s dilemma is valuing relationships over personal power in a society that devalues relationships and caring. Esteeming one’s self in this situation becomes difficult. Women are expected to behave in caring ways and yet are seen as inferior to men because of it. Yet another relational theorist, Kaplan (1984), stressed the importance to women of being attuned to the feelings of others, being understood by and understanding others, and participating in others’ development. In her discussion of women and depression, she proposed that the generally accepted core elements of depression-- (a) experience of loss, (b) inhibition of anger or aggression, (c) inhibition of assertiveness and action, and (d) low self-esteem--are descriptive of the female experience of women in this culture in general. Kaplan suggested that women’s vulnerability to depression is based in loss of or fear of loss of relationship. Anger and aggressiveness are repressed because of feared destructiveness, action and assertiveness are repressed because of fear of major disruption in relationships, and low self-esteem both contributes to the repression and is the end result of it. If self-esteem depends on the quality of her relationships, her own wishes and expression 14 can threaten ‘those relationships and are 'thus inhibited. Kaplan suggested that the importance of relationships and the basis of identity in relationships make up the core experience of all women, rendering them more vulnerable to depression and problems with self- esteem and aggression. Chodorow (1978) is an interesting theorist who earlier aligned strongly with the analytic school but more recently has become more identified with the relational theorists. In a revision of psychoanalytic theory, she provided the basis for other relational theorists in her detailed description of women’s early development. Chodorow proposed that women have a nurturing disposition in large part because of the continuous relationship and identification with the same-gendered parent, mother. She also suggested that mothers will identify more closely with their daughters than their sons because of gender. According to Chodorow, this identification holds the daughter in an extended pre-oedipal dependency with mother, which results in a more continuous and intense relationship between mothers and their daughters than between mothers and their sons. The outcome is a relational identity and women’s capacity to nurture others. This outcome is based not in biological gender differences but in the contemporary parenting structure, in which women are the exclusive caretakers of infants and children and the expressors of the need to affiliate with others. Men, in the meantime, become the expressors for separate identities and nonfamilial work in the field. This division, Chodorow suggested, sets both men and women up for difficulty in relationships. 15 Chodorow (1978) objected to the strict social-cultural view of women’s psychological development as expressed by Horney. Chodorow did not believe the patriarchal structure alone is enough to explain women’s position in society. She did not believe, for instance, that men’s power over women can explain women’s position because she did not believe that anyone can be coerced to be a "good enough mother," a task many women accomplish. Rather, Chodorow, while taking into account the economic organization based on women’s exclusive mothering, suggested that the asymmetry in parenting leads to the reproduction of the same system, generation after generation, and can interfere (with healthy development in both sexes. She described the difficulty that women have with aggression as being based in the incompatibility between a nurturing position and aggression, self-assertion, and angeru 'This difficulty, according to Chodorow, is expressed in women’s lack of differentiation and problems 'with establishing a separate identity. This lack of differentiation results in the limiting of internally acceptable choices for women, and thus the danger of ambivalent attachments and role reversal, when mothers look to their children to meet their needs. Soci The r' Westkott (1986), in contrast, used Horney’s theory as a base to propose a new social-psychological theory of the development of women. She saw the problems with anger and self-esteem as central issues beginning early in childhood. Westkott disagreed with 16 Chodorow (1978) in contending that the feminine identification with the nurturer comes not from a continuous relationship with the same- gendered parent but from a different source. Westkott suggested that a girl child is devalued by all in the culture, including the child’s mother. The development of this devalued child involves the direct and subtle experience that her needs, wants, and wishes are devalued. Out of a fear of loss of safety, well-being, and love, the child soon learns to devalue her own experience, to deny her real self, and to repress the anger at this requirement. In a defensive stance, she turns her anger against herself, admires the devaluing parents, and experiences her real self as contemptuous and worthless. This is a sharp contrast to the views of Chodorow (1978), Gilligan (1982), and Kaplan (1984), who proposed a positive, healthy identification with mother as compared to Westkott’s view of it as a largely defensive stance. Westkott went on to say, "The repression of anger therefore results in self-contempt and in the conviction that one’s anger is powerfully destructive, especially to those whose love is desired" (p. 137). Westkott suggested that while Gilligan acknowledged that assertion seems dangerous to women in part because of their dependence on others, she idealized their altruism and ignored the underlying rage. Westkott (1986) explained woman’s inability to gain lasting self-esteem from her nurturing despite the strong identification when she described female altruism as defensive: 17 Female altruism is the characterological need to care for others. It emerges, I argue, not out of an extended attachment between mother and daughter but out of the parents’ need fer nurturing. The mother turns to her daughter not because they are the same gender but because the culturally rooted nurturing imperative presumes that all females are to nurture others. This, I suggest, does not lead to the daughter’s extended stay in the so-called pre-oedipal nest, but to the mother’s devaluation of the daughter’s needs, expressed in her imposing her own needs to be nurtured. In this respect a mother’s identification with her daughter as a female is important only to the extent that she has internalized the cultural assumption that daughters are of less value than sons, that their needs are less important, and that they exist to serve the needs of others, including those of the mother. (pp. 133-34) Thus her internal experience is devalued, whether that be needs for nurturance or more expansive needs for creativity and exploration. Her understandable anger about this is held in check with self- contempt. She feels a rage at being under-nurtured and a fear of the rage. In addition, she is trapped in a dependent stance because of unmet dependency needs and her inability to act on her natural aggression and other expansive needs, which would provide the basis for more independence and self-esteem (Westkott, 1986). Thus, in contrast to Chodorow, Westkott proposed that girls are not held in a pre-oedipal dependency by mother because of continuous identification. From this perspective, dependency and the inability to separate adequately arises because the daughter is expected to care for mother and others and to prematurely abandon her own needs for dependency, care, and nurturance. What Westkott was suggesting is that instead of a strong positive identification with mother, which could provide the core of self-esteem and the base for aggressive action, women are more likely to develop with feelings of low self-esteem because of unmet needs, which they have learned to 18 devalue, as well as a fear of their own aggressiveness. Both Chodorow and Westkott suggested, however, that women are both victims of an oppressive system and participants in their ongoing victimization. n tiv r Historically, Karen Horney was the first to address the issues of women’s development and behavior in the social/cultural context. Other theorists from the analytic and object relations schools have more recently also offered an integrated view of how internal dynamics and developmental processes are influenced by social pressures and norms. Although not using the word "victim,“ Horney also described women of the "feminine type” as victims of a culture that devalues and sexualizes women. Horney described at great length women’s conflict between ambition and attachment. She believed the base of the conflict to be in women’s struggle with society’s definition of their role. According to Horney, two possibilities presented to women are working toward growing independence or attaching to and caring for others. These alternatives are frequently experienced as mutually exclusive. She saw ‘the social pressures on women as pushing toward an overvaluation of love and undervaluation of pursuits of work and competence. This was the role defined for women by the culture and then, in part, maintained by their own conflicts. Horney viewed this overvaluation of love as leading to a 19 stance of self-effacing, compliant women who are dependent on their charges for substance and a sense of worth. In the context of a society that has an investment in women’s remaining dependent, for women the dependent stance becomes the defensive solution for dealing with disappointment and loss of being valued and supported. It defends against the anger and revengeful fantasies by keeping them unconscious. Unconscious anger and wishes for revenge are reactions to the disappointment and the prohibition against going after what she wants, using her natural aggressive strivings. Thus women, according to Horney, are left angry at self and other. The anger at self is important because it mirrors and expands on the societal devaluing. It also serves to internalize the process that used to be external in the family and society. The anger at self is experienced in two ways. First, anger from ‘the real self“ is directed toward the idealized self that continues to push for happily conforming to external standards. This anger is in reaction to the lack of acceptance of the real self, the hated submission to external devaluing and sexualizing standards and denial of the longed-for ambition strivings. There is anger because the impossible standards of the idealized self lead to inevitable failure and self-hate. Second, because of the passed-on submissive role, there is anger against the part of the self that identifies with mother and all women. This anger can be experienced as anger at being female, further internal devaluation. In the midst of this, aggression remains blocked by external standards and 20 internal conflicts, and self-esteem is low. Horney is historically the first theorist to describe how the internal devaluation is only an adjunct to the powerful societal devaluation of women. To help us understand further the consequences of a female sex- role identification and the importance of a father figure, Horney also addresses society’s idealized role for women. According to her, the societally suggested romantic solution for women as the caretaker of others, the admired beauty who is sought because of her tenderness and nurturance, is a false solution and cannot assure happiness and self-esteem. Finding happiness through men as wives and mistresses to the exclusion of her own independent ambition is often disappointing. In addition to the devaluing of a young girl’s strivings toward competence and achievement, Horney saw sexualization as a major threat to healthy development of self-esteem. By sexualization, Horney was referring to experiences ranging from incest to sexually loaded, intrusive or controlling behavior on the part of fathers and brothers. An additional variable is having a mother whose identity and self- esteem are centered on male admiration of her beauty. This kind of treatment and experience, according to Horney, leads to self-hate, dependency, fear, and disgust with her body. If the important male figures in her early life value her as an object or plaything because of her sexuality instead of her competence, achievement, and brightness, the girl will retreat from the world of aggression and assertion. According to Horney, what is lost to her is the vital 21 admiration, support, and encouragement to integrate masculine aspects of identity that allow her movement toward independence and self-esteem (Horney, 1967). Further understanding of the unique consequences of women living within society’s restrictive roles is given by Margaret Mahler. Mahler (1979, 1981) directly addressed the personal, internal issues for women when she explored the possibility of female children having more unneutralized aggression, an aggression that contaminates areas of development and relationships in such a way as to interfere with healthy progress and satisfaction. Mahler suggested that this is because of mother’s over-identification with or idealization of her daughter as well as mother’s own proclivity toward dependence. These dynamics contribute to a pattern of mother being less accurately attuned to the needs of her daughter, leaving her daughter traversing the developmental process of separation and individuation with unmet dependency needs. This, in turn, leaves the female child more vulnerable to the regressive pull back toward mother (i.e., undifferentiated and dependent) as well as more conflicted about the natural drive to move away from mother to a separate identity. According to Mahler, father helps children of both sexes to separate from mother by offering the toddler an alternative to the disappointing and now separate mother. Father represents moving out into the ‘world and offers the opportunity for a selective identification to assist in the separation process. Mahler also proposed that a strong identification with father, even in the face 22 of a poor relationship with mother, can significantly contribute to the maintenance of higher self-esteem. Mahler differed from Chodorow in her description of this identification process, suggesting that rather than the seduction of the ”like nature," the pull back to the symbiotic orbit may be stronger for the girl child due to the lack of adequate and accurate empathy on the part of the mother. This leaves the child with unmet needs and primitive rage, clinging to mother in hopes of yet getting her needs met. In summary, Mahler saw aggression as essential to the process of development, while she saw unneutralized aggression as interfering and female children more vulnerable to problems because of mother’s tendency to project her own needs and feelings onto her female children. Mahler (1979, 1981) also spoke to the difficulties girls have in separating from mother if father is missing or uninvolved. In a similar but more politically based description, Flax (1978) saw these difficulties as based firmly in mother’s own unresolved conflicts about her childhood, gender, and unmet needs. She believed that female children are much less likely to have their genuine needs for nurturance met in the early years, and this is what she purported to be in large part responsible for the difficulty female children have with separation/ individuation and access to aggression. She also suggested that mothers communicate to their female children their own conflicts about being female as well as demonstrating their less powerful role in the family and 23 society. For the girl, blows to self—esteem will be more difficult to rebound from because of deficits carried from earlier developmental phases. Girls are more likely to cling to mother, to become more engrossed with her instead 'of moving away as boys do. They are enmeshed in the ambivalent aspects of the relationship with mother and likely to be more depressed. In the context of a culture where men are the holders of power, she is to turn to mother, instead of moving forward to identification with all-powerful father. Fear of the loss of mother’s love at any moves toward autonomy, as well as longings to return to a regressive, symbiotic union with mother, are strongly present. According to Flax; women are left with a conflict between nurturance and autonomy, which is most often unconscious. Because of early experience with mother and growing up in a patriarchal society, women feel a need to repress their needs for autonomy. They also feel a fear of losing mother’s love if they make assertive strides toward autonomy. The anger felt because of this is fused with the drives of autonomy. Thus the drive itself is experienced as more dangerous. The anger is also expressed in hostility toward any women who act on their own wishes and attempt to escape the traditional female role. The variables involved in women’s development can often contribute to an over-identification with mother, which is really a mask for deep rage at her and all other women. This goes on to contribute to undermining the possibilities for women to get support from each other for moves toward autonomy as well as interfering 24 with possibilities for women nurturing women in adulthood as part of a potentially healing process (Flax, 1978). Thus Flax helped to explain how women in a society where mothers are the exclusive caregivers are likely to suffer from low self-esteem and unmet dependency needs. This results in cycles of women growing up unable to use their natural aggression in their own behalf and thus having difficulty building and maintaining self- esteem. While clearly identifying the restrictions of feminine sex- roles, Lerner (1980, 1985, 1987) looked more directly at the relationship between developing autonomy and the ability to be self- assertive and express anger. She suggested that this was an important variable in the development of self-esteem. Lerner explained that early' developmental experiences of attempting to separate from the same-sexed parent with whom a continuing identification is maintained results in women often failing to achieve the autonomous functioning required for a tolerance of aloneness. Confidence about a certain level of autonomous functioning is required to tolerate the separation anxiety inherent in the experience of anger. Without this level of autonomy as well as trust that one has it, the experience of anger is too dangerous. Lerner (1980) stated, ”The expression of legitimate anger and protest is more than a statement of dignity and self-respect” (p. 145). She tied the ability to freely express anger and protest with self-esteem, 'which she believed to be based in the ability to tolerate separateness. 25 According to Lerner, women have a special problem with tolerating separateness and expressing anger. She explained that women are socialized to believe that their value and identity are based on loving and being loved to the exclusion of their separate identities and autonomy. These influential early lessons leave women more dependent on the approval of others and make their anger more dangerous. In addition, women’s anger is feared not only by the woman herself, but also by other women as well as by men. This fear is based on early images carried in the unconscious of the omnipotent, vengeful, destructive mother on whom one was dependent for survival. Because women are assigned the role of primary, often exclusive, caregiver to infants and children, mother is the first person to nurture and gratify impulses as well as to punish, withhold, and inflict narcissistic injuries as a part of everyday life. This early experience, as well as the primitive projections onto mother of one’s own unmodulated rage, sets in place an irrational fear of women’s anger (Lerner, 1980, 1985). Thus the cultural prohibition against women’s anger is internalized by all. The images of angry women being ”castrating bitches," manipulative, passive-aggressive housewives who control their husbands and children or irrational, hysterical females are not just to be explained away as sexist stereotypes. These are the combination of others’ projections and the unfortunate, self- destructive, and devalued stances some women assume to defend against their own feared power and anger. These positions are 26 reinforced by cultural prohibitions against direct female anger. These stereotypes become associated with expression of anger or aggression for women. Women also avoid direct anger in order to avoid the negative stereotype. According to Lerner (1980, 1985), women face a special difficulty because a girl also identifies with her mother and thus is faced with internalizing the image of the great nurturer, along with this terrifying "bad mother." In defense against that possibility, women often split off this dreaded potential, and with it go most of their aggressive impulses. The result is the position of helpless, castrated female, a position preferred by the woman and the culture to the possibility of being a "castrating" female. This position for girls leaves them an image'with which to identify that is partly terrifying, and much less "permission" to separate and individuate. The girl is often given less permission from mother to assert herself. Mother, because of her own experience and conflicts, reacts to her daughter’s assertiveness and difference as a narcissistic injury and thus something to be discouraged. In an unfortunate recapitulation of the sex-role stereotype, mothers are likely to be less supportive of their daughters’ striving toward autonomy. This may leave girls with the disapproval from mother for a stance of separateness and any expression of anger or assertiveness. Mothers are more likely to respond in an admiring way to boys’ aggressive strivings. This gives the boy a sense of autonomy and connectedness with mother. If mother experiences her daughter’s 27 strivings as a betrayal, the girl will experience terrifying aloneness in aggressive strivings without the sense of connectedness to mother. Thus girls are likely to be less able to tolerate the experience of anger, which, according to Lerner, has inherent in it a sense of standing alone and separate, a statement of difference. Lerner also saw feelings of guilt, self-criticism, sorrow, and depression as retreats from feelings of anger to feeling hurt. Because of the intolerable separation anxiety when angry, women abandon their assertive position and refocus on caring for the relationship. The hope is that expressions of hurt will draw the loved one back. In this process, however, the woman loses the self- respecting sense of the legitimacy of her feelings and often suffers a loss of self-esteem (Lerner, 1980, 1985). Lerner outlined clearly how the interplay between external sex-role expectations and internal conflicts binds women in nonassertive, self-devaluing stances. Nadelson et a1. (1982) specifically addressed self-esteem from an analytic, developmental perspective. As these authors described it, children grow to accept their realistic limitations and inhibit acting on impulse if impulses can be replaced with ego-syntonic goals, which bring pleasure to the child and gain the approval of the parents. This developmental process is part of the growth of self-esteem as the child begins to feel effective and "good." If ego-syntonic goals are not approved of, problems arise. Self-concept is achieved through reacting to and internalizing the attitudes and behaviors of the family. So, if parents react to 28 the girl’s aggressive strivings with disapproval, this negative evaluation will be internalized. Mastery of physical processes and the environment, as well as experience with other people, modifies self-esteem as the child develops. Negative reactions based on sex- role stereotypes will affect the developing self-esteem of the child. The parents’ self-esteem and self-images are also vital as the child depends on the parents for love, support, and acceptance and strives to live up to parental standards, expectations, and values. According to Nadelson et a1. (1982), unconscious conflict can undermine self-esteem and the internalization and identification with a parent who has low self-esteem. This is particularly relevant for female children if the maternal ego ideal does not include elements of aggression. If the child pursues activity for her own pleasure and enhancement in response to her own natural aggressive drive, she will be in conflict with mother and with the internalized maternal ego ideal, which makes the nurturance of others the primary source of self-esteem. These authors also made the point that the restriction of physical activities and the difficulties in acceptance of the body and its functions also make the development of self-esteem in girls more limited. The cultural standards for early restriction on physical activity' can be internalized as evidence that her own impulses and desires are "bad." The use of natural aggression to go out into the world and separate from a mother who feels devalued herself is more difficult. 29 For the daughter to develop high self>esteem would mean to reject mother when, in fact, mother’s love and support are needed on an ongoing, if more separate, basis. The impact of the fear of aggression with the resulting lack of integration and inhibition of aggressive expression is that. women frequently experience their aggressive strivings as evidence of a lack of worth or defectiveness, instead of the basis of self-esteem (Nadelson et a1., 1982). For further exploration of the effect of sex-role standards on women’s expression of anger, we turn to Bernardez. Bernardez (1978) equated a tolerance for one’s own anger with self-esteem and capacity for separateness. She stated that to feel anger is to experience standing alone. The culturally determined over-valuation by women of the needs and feelings of others results in a fear of women’s anger as potentially destructive. The potentially life-giving experience of anger as an aid to defining "me" and "not-me" is blocked by the fear of being destructive to the very relationships on which she) depends. Bernardez additionally explained that this blockage of the expression of anger and aggression in conjunction with women’s extreme sense of responsibility for the comfort of others sets up a situation that leads the woman to perceive herself as hateful. Aware on one level of an injustice, yet unable to discriminate between legitimate and irrational anger, women suppress the experience and instead develop self-deprecatory attitudes. Women work at obtaining the approval of others to the point of being fraudulent, misrepresenting their 30 desires and values. This results in women’s experiencing themselves as inadequate and dishonest, with an inevitable loss of self-esteem. Bernardez also pointed out that it is not only the suppression of direct anger but the defensive, indirect release of that anger that results in loss of the esteem of others as well as self-esteem. Resentment, bitterness, temper tantrums, or maniacal domination are not responded to by others as legitimate. Complaints are seen as invalid, and behavior is dismissed as infantile. Bernardez went on to point out that an additional loss for women is that anger turned against the self also prohibits objective self—evaluation. Unique personal assets and positive aspects of one’s behavior are lost in a self-deprecating attitude. Growth- giving self-evaluation is not possible in the presence of self-hate, nor is self-esteem. Esteem building, creativity, and realistic self-love are necessary for growth and change. This requires a freedom with the experience and expression of anger. Assertiveness occurs in an atmosphere of care and respect for self as well as others, according to Bernardez. Clarity and a sense of legitimacy about one’s anger are required for women to stand separately and maintain their deeply valued concern for others. Theoretical literature repeatedly has addressed the importance of the access of aggression for women and how vital that is for the development of self-esteem. Many theorists also have pointed out how sex-role stereotypes in this culture affect this and how a purely feminine sex-role identification makes access to aggression difficult. There is disagreement among the theorists regarding the 31 relationships among anger, self-esteem, and sex-role identification. Some theorists have suggested a high positive relationship between a feminine sex-role identification and self-esteem, without reference to access to aggression or problems associated with aggression for women (Surrey, 1984). Other theorists have focused on access to anger and aggression relative to the feminine sex-role and have suggested that high self-esteem is unlikely for women within the confines of a traditional feminine sex-role (Bernardez, 1978; Lerner, 1985). These theorists have delineated how a seemingly simple sex-role restriction results in complex problems with far- reaching consequences in the development of women. A survey of research literature on the relationships among these variables suggests some variance in understanding how they relate to each other. Since few studies have dealt directly with the relationships among these three variables, related studies that contribute understanding are reviewed. The Research Literature Sex-Role Identifigatign The research literature on sex-role stereotypes is vast, and a review of all such literature was beyond the scope and intent of this study. Research literature that relates to the interaction with other variables in this study will be summarized. It is clear, as cited in a landmark research report, that in this culture masculine traits are more highly valued than feminine traits (Broverman et al., 1972). They found that both men and women 32 ascribed more positive values to characteristics viewed as masculine. The positively valued masculine traits involved competence, rationality, and assertion, whereas positive female traits were warmth and expressiveness. Negatively valued traits seen as feminine included such things as passivity, sneakiness, and dependency. Sex-role standards were defined as “the sum of socially designated behaviors that differentiate between men and women" (p. 60). Rosenkrantz et a1. (1968) also indicated agreement between the sexes about sex differences, differences between the self- concept of the sexes, and more valuing of stereotypically masculine traits by both sexes. Culp, Cook, and Hansley (1983) found support for previous studies that demonstrated that parents behave differentially toward male and female children. This study plus previous research (Frisch, 1977; Seavy, Katz, & Zalk, 1975; Smith 8 Lloyd, 1975; Will, Self, & Daton, 1976) all suggested that there is subtle sex typing that occurs in parental treatment of infants. A child perceived as male in the Culp et a1. study was smiled at more and received more direct eye contact from both male and female adults. Female adults had less physical contact and more verbal contact with children of both genders. Male adults smiled more at the children and had more physical contact. A child perceived as female by the adult females was more consistently given a feminine toy to play with. In a similar study, Rathbart and Maccoby (1966) found the sex of the parent to be a better predictor of differential responses to 33 boys and girls. In a study based on self-report from parents as well as observation, they found that fathers allowed more autonomy than mothers“ Mothers showed more permissiveness and positive attention toward sons, while fathers were more permissive and positive with their daughters. Mothers tolerated more aggressiveness from boys and were more accepting of comfort-seeking behavior in their sons. This was found to be the same for fathers with their daughters. Maccoby and Jacklin’s (1974) review of research literature on the differential socialization of girls and boys found few differences in parental treatment. The exceptions were that parents tended to give children sex-typed toys and to discourage them from activities not believed to be appropriate to the sex of the child. Their review also indicated that parents were made more anxious by boys behaving like girls than by girls behaving like boys. Williams (1977) reported in her review of research literature on sex-role identification in women that it is not the sex of the parent model that matters but the ability of the model to integrate a balance of masculine and feminine traits. Girls identified with these more integrated models make better adjustments in life and have higher levels of functioning. There have been a number of studies investigating gender differences and psychological well-being as well as sex-role identification and well-being. In her study of sex-role attitudes and psychological well-being using Barrons’s Ego Strength Scale of the MMPI and the Fand Inventory, Gump (1972) found one can be 34 feminine and possess high ego strength, but it is questionable whether one can do both in the narrowly defined feminine role of nurturer of others. Gump looked at women who were wives and mothers and had careers, but their careers were in traditionally feminine fields. There is evidence that women are more prone to depression than men (Weissman & Klerman, 1987). These authors reported that a review of 40 studies done before 1977 found few exceptions to the higher incidence of depression in women. In 1987, these authors reported that new studies have indicated an equal rate for bipolar depression in men and women but a continuing higher rate of other categories of depression in women. Elpern and Karp (1984) found that sex-role identification is a better predictor of depression than is gender. Women who described themselves as more identified with the female stereotype were significantly more likely to be depressed than women identified with the masculine sex-role stereotype. In their' major study of’ depression in women, Weissman and Paykel (1974) found that acutely depressed women were more overtly hostile to others, particularly close family members, than normals. They also found defects in the depressed women’s self-esteem. The authors speculated that for some acutely depressed women, depression allowed overt expression of hostility otherwise repressed. They also found these women unable to communicate their needs, wishes, and feelings to others. 35 In a replication of studies looking at emotionalism in males and females, Bernham (1983) looked at preschoolers in an observational study. She found that anger was more characteristic for boys, and fear, happiness, and sadness were more characteristic for girls when they were expressing emotion. Bernham challenged the global statements regarding females being more emotional and suggested that when looked at more closely, males were more emotional along one dimension (anger), while females were more expressive of fear, happiness, and sadness. This study highlighted that these patterns were established early in development (subjects were 3 to 5 years of age), males were aggressive, females were responsive. Heeser and Gannon (1984) found positive associations between the number of physical symptoms expressed and the indirect expression of anger. These two factors also correlated with certain stereotypically feminine traits, such as passivity, submissiveness, and insecurity. Measures of irritability, indirect hostility, resentment, and suspicion also correlated with the number of physical symptoms subjects complained about. n e H t' nd A i Stereotypically, women are seen as being less assertive than men (Broverman et al., 1972). Holandsworth and Wall (1977), in a sampling of 702 conmunity mental health clients, found that men viewed themselves as being more assertive in dealing with bosses and supervisors, being more outspoken, and taking more initiative in 36 social contacts. Women described themselves as being more assertive in expressing affection and compliments in addition to expressing anger at their parents. Muehlenhard (1980) concluded from her review of the literature that women appeared to be less assertive than men in negative assertion and initiation of social contacts. Sanchez and Lewensohn (1980), in monitoring 12 depressed outpatients, found that there was a high negative correlation between the rate of assertive behavior and level of depression. They also found that the rate of assertive behavior on one day reliably predicted the level of depression the next day. Research literature on aggression is quite extensive, particularly as it investigates gender differences. There is little research literature that focused only on women and aggression. One frequently cited study was by Maccoby and Jacklin (1974), who reported males to be consistently more aggressive than females. The authors concluded that differences appear by around age 2 and decline with age. Males are still more aggressive than females in late adolescence. Little research investigating adults was found by these authors. Hyde (cited in Hyde & Linn, 1986), in a meta-analysis of gender differences in aggression, also found males to be more aggressive than females. However, she reported a pattern of differences being greater in naturalistic studies than in experimental studies. Hyde, suggested the difference may have been a function of the difference 37 between spontaneously displayed aggression versus stimulated aggression. She suggested that there was a capacity for aggression in women and, when stimulated in a situation that gave approval to aggressive behavior, there was an increase in women’s aggressiveness. Similar increases were not evident in the men. Hyde’s review of 143 studies indicated that gender differences began to appear between ages 2 and 3, tended to be larger differences in preschoolers, and became smaller among college students. Few studies have used older subjects, but there have been indications that aggression in both sexes tends to decrease with age. Hoppe (1979) found masculine males to be more aggressive than other males and females. She noted an increase in aggression in both sexes following provocation. She also found that the sex of the opponent influenced the male’s aggressiveness but not the female’s. Interestingly, she found that the Bem Sex Role Inventory scores did not necessarily relate to the social behaviors adopted by the individuals in her study. Brodzinsky, Messer, and Tew (1979) found boys to be more physically and verbally aggressive than girls but not more indirectly aggressive. Their results indicated that girls were just as likely to be aggressive when the aggression was indirect. Feshbach (1969) found girls to be higher in the indirect expression of aggression. She also found no significant differences in aggressiveness in boys’ and girls’ responses to strangers. Frodi (1977) found that men and women became equally angry when provoked. However, her study suggested that. males and females 38 adopted different coping strategies when angered. In recorded stream of consciousness, men revealed thoughts and associations that stirred up angry feelings. Women’s thoughts were prone to understanding the other and attempting to give the benefit of the doubt. Perry, Perry, and Rasmaussen (1986) saw negligible differences in self-perceived efficacy for aggression between boys and girls but large differences in the anticipated consequences for aggression. Girls expected the aggression to cause more harm and injury to the target (victim), and they also expected to be punished more severely by peers and by themselves. Hardin and Jacob (1978), in a study investigating aggression of preschoolers and teachers’ sex, discovered that overt aggression was highest when preschoolers were dealing with a female-female pair of teachers. This was the most aggressive condition for both girls and boys. The lowest aggression for girls occurred in the female-male teacher pair. This was not true for boys, who demonstrated least aggression in the male-male condition. Eagly' and Steffen (1989) found in their study that a significant deterrent to physical expression of aggression in women is their' concern for‘ the consequences of such action and anticipation that their aggression will be viewed negativelyu Men view their aggression positively because of their superior status and because of their ascribed social roles, in which aggression is central. Examples of such roles are the military, sports, and competition. 39 In looking at anger expressors and suppressors, Pape (1986) found males and females did not differ in predicted ways in private settings. However, males appeared to be more sensitive to anger provocation in public than private settings. Navoco (1977) looked at stress inoculation as a method for developing self-regulation of anger. He related the ability to cope with and use anger effectively with the subject’s sense of personal competence. One of the popular theories that has been used to explain the male having more access to aggression is the notion that females express less aggression because they are more empathic, caring for the other (Chodorow, 1976; Gilligan, 1983). In a study in which the authors expected to support this theory, Feshbach and Feshbach (1969) looked at aggressiveness and empathy in children 4 to 7 years of age. What they found was surprising. Girls with high empathy scores were also more aggressive than girls with low empathy scores. Aggression did correlate negatively with empathy in older boys (6 to 7 years). Aggression correlated negatively with intelligence in boys and correlated positively with intelligence in girls. It has been asserted that fathers who sexualize their relationships with their' daughters negatively affect their daughters’ self-concept and self-esteem (Horney, 1967; Westkott, 1986). LaBarbera (1984) found that college-aged women who reported sexualized relationships with their fathers described themselves as arrogant, hostile, and lacking in warmth, nurturance, and interest 40 in others. These women (ages 17 to 24) attributed more danger to male sexuality and female competitiveness. Women who reported their fathers to be disinterested in or discouraging of their achievement appeared unassertive, passive-dependent, and lacking in "nontradi- tional" values. - te In reviewing the research literature on self-esteem, an effort was made to focus on research that attempted to address develOpmental variables. This was done in an effort to integrate an understanding of the developmental variables that may contribute to high or low self-esteem instead of just looking at, as an example, high self-esteem and achievement. These latter types of studies did not indicate why some subjects were able to attain high self-esteem and achieve whereas others could not. This issue of the developmental variables a1501 is at the crux of ‘the theoretical disagreement. Coopersmith (1967), in his extensive study of antecedents to self-esteem, defined self-esteem as “the evaluative attitudes toward the self." These attitudes are lasting over time, express approval or disapproval of the self, and suggest the degree to which the individual holds herself to be capable, significant, successful, and worthy. Coopersmith viewed self-esteem to carry affective loadings and to have motivational consequences. While keeping in mind that Coopersmith used preadolescent boys (ages 10 to 12) as subjects, his findings are interesting and relevant, and his is the most 41 comprehensive developmental study of self-esteem to date. He found that individuals with high self-esteem responded more consistently to both external and internal events. Persons with low self-esteem had higher levels of anxiety but otherwise lower levels of expressed affect. Subjects with low selfkesteem were more socially withdrawn and isolated, whereas those with high self-esteem were affiliative and had greater social interests. In groups, subjects with high self-esteem engaged in discussions and expressed opinions, whereas those with low self-esteem were more passive. In general, high self-esteem was demonstrated in trust in self, feelings of worthiness, greater social independence, creativity, and more assertive, vigorous social actions. Coopersmith viewed the results as indicating that subjects with low self-esteem were less able to have genuine interest in others because of marked self-consciousness and preoccupations with inner problems. When examining relationships to parents, Coopersmith found parental warmth and acceptance to be enhancing to self-esteem, while rejection, disinterest, and distance led to a diminished sense of worth. Subjects with low self-esteem had more unconscious hostility and antagonism toward mother and received less acceptance and affection from her. In relations to father, boys with high self- esteem had closer relationships with father, mothers reported more satisfaction with their husbands, and boys confided more in father. The relationship Coopersmith feund between independence training and self-esteem is more puzzling. Subjects with high self- esteem and those with low self-esteem had greater and fairly equal 42 levels of independence. Subjects with medium self-esteem were dependent. The distinction between subjects with high or low self- esteem was that those with high self-esteem experienced success and a sense of competence from their independence, thus further enhancing self-esteem. The subjects with low self-esteem experienced failure, only reinforcing feelings of inferiority. Subjects with medium levels of self-esteem were found to have been over-protected by parents and to have experienced considerable self- doubt about independent action, to rely on others for assurance, and to be attuned to the needs of their parents. Parental values also were found to relate to and to affect children’s self-esteem (Coopersmith, 1967). Many children could not state what values were important to their parents. Children with low self¢esteem were more apt to claim that their parents valued accommodation. This included obedience, helpfulness, adjustment to others, and kindness. Children with high self-esteem believed that their parents valued achievement. Coopersmith also found persons of low self-esteem were more anxious, internalized their reactions to their anxiety, were more destructive toward objects, and were more prone to psychosomatic symptoms than people with high self-esteem. They were not, however, more aggressive. Coopersmith presumed that this is because aggressive acts require more initiative and assertiveness, traits lacking in persons with low self-esteem. Rosenberg’s (1965) research with adolescents demonstrated that paternal attention and concern were significantly related to 43 self-esteem and that closer relationships with father resulted in higher self-esteem for both boys and girls. Persons with low self- esteem were found to appear more depressed to others and more likely to express feelings of gloom and despair. Low self-esteem was related to a greater likelihood to have psychosomatic symptoms; those with the highest self-esteem experienced the fewest symptoms. Students with low self-esteem were "outstanding'I in their invisibility in the classroom. High self-esteem seemed to reflect feelings of being "good enough" and self-acceptance, whereas low self-esteem was related to feelings of self-contempt. In her study of sex-role behaviors relative to self-esteem, Karam (1977) found that women with high masculine scores on the Bem Sex Role Inventory also had high scores in emotional stability and self-esteem. In contrast, Spence et a1. (1975) found high self- esteem correlated with a high degree of masculine and feminine characteristics and feelings of agency. Barauch (1973) found that the evaluation of self as competent was clearly related to self- esteem. She also found that girls whose mothers showed a preference for a career and encouraged their daughters to achieve, fostered high self-esteem. Long (1986) found masculinity to be a predictor of self-esteem and self-acceptance in college students, professionals, and mental health clients. Femininity, on the other hand, was not a reliable predictor of self-esteem, suggesting it is neither helpful nor harmful. An additional study, of early adolescents, by Connell and Johnson (1970) suggests that the male 44 role results in higher self-esteem whether it is adopted by a girl or a boy. Follingstad, Robinson, and Pugh (1977) found consciousness-raising groups that helped participants to process their conflicts about women’s roles and socialization resulted in higher self—esteem for the participants. Maccoby and Jacklin (1974), in their review of studies of development, found that while males and females rated themselves with equally positive (or negative) self-images on a self-rating scale, males judged their performance on a task significantly more positively than women’s judgment of their own performance. Males were also more optimistic about grades, even when evidence was that girls would do better academically. In a study looking at self-esteem and academic achievement relative to sex-role perceptions, Robinson-Awana, Kehle, and Jenson (1986) found that, for both boys and girls, self-esteem rose with academic achievement. It was also found that boys rated themselves higher in self-esteem and girls ascribed higher self-esteem to boys. However, the one exception to this pattern was a group of academically high-achieving girls who had high self-esteem and rated boys as having lower self-esteem. The results again suggest that feelings of competence and agency are important for girls as well as boys. Hasenbeld (1986) found a positive correlation between self- esteem and agency in her study. She interpreted agency as the ability to use one’s aggression for one’s own personal interests and goals. 45 Mathes, Adams, and Davies (1985), in their studies of jealousy, found women more vulnerable than men to depression, loss of self- esteem, and anger' with the loss of' a relationship. This was particularly true if the relationship was lost under conditions other than fate (i.e., loss to a rival, partner destiny, or rejection). The authors concluded from the results of their two studies that women may be particularly vulnerable to loss of self- esteem with the loss of a relationship. They did not attempt to evaluate level of functioning in other areas of the subjects’ lives relative to the loss of self-esteem, and the measure was self-report in response to hypothetical situations. §HEEQEX Research literature has supported the notion of a positive relationship between high self-esteem and a sense of competence and assertiveness, as well as a negative relationship between high self- esteem and psychosomatic symptoms, withdrawal, or depression. The results have not been consistent relative to sex-role identification and self-esteem. However, notably absent was any study suggesting that high self-esteem was correlated positively with a feminine sex- role identification. Previously cited research literature on sex-role identification evidenced positive correlations between feminine sex-role identification and depression, as well as a general valuing of masculine traits over feminine traits for both sexes. This appears to have been borne out in studies of behaviors of parents with their 46 children. Stereotypes seemed to be upheld in parental interactions with their children, starting when the children were very young, with mothers demonstrating less tolerance for separateness and aggressiveness from their daughters. These previous findings tend to support 'the idea that femininity and aggressiveness are not consistently accepted together and that conflict about it is expressed in parenting behavior as well as adult women’s behavior. In general, the research literature supports the idea of differences between men and women in terms of aggression and assertion and points out that female subjects display aggression but in different ways and under different circumstances. Female subjects may also have been aggressive or assertive in response to different stimuli. While not directly addressing the issues of this study, this research broadly supports the general thesis of aggression in females that is expressed or displayed differently. While the idea of a unique base of strength in the mother- daughter relationship is appealing, solid support for it in research evidence is lacking. Research evidence, in fact, seems to provide more support for theories that propose more conflict in the mother- daughter relationship. There appear to be few studies that have looked directly at the relationship between self-esteem, sex-role identification, and hostility in women. These variables have been studied separately but not directly, relative to one another. Given the central role these variables have been described to play in women’s development, 47 investigating relative levels in adult women and examining how they occur together seems a natural extension of the theory. The purpose of this study was to directly address issues raised in the extensive theoretical literature about. women’s development and self- perceptions by measuring sex-role identification, self-esteem, and hostility in adult women subjects. It was hoped that such an investigation would provide some insight into the relationship between these variables, which will facilitate clinical work with women as well as further research. The theoretical and research literature together point to the central importance of the social/cultural context for understanding human behavior. The external values, expectations, and restrictions must be illuminated and clarified in order to understand how women internalize these standards and then struggle with them intrapsychically. Without this understanding, women appear to avoid simple solutions. Measuring women’s sex-role identifications gives some access to how women have integrated into their self-concepts the societal stereotypes. By looking at how these identifications related to levels of hostility and self-esteem, it was hoped that a better understanding would be gained of the complexity of the relationships between these variables. CHAPTER III DESIGN OF THE STUDY This study investigated the relationships between levels of self-esteem, degrees of hostility, and sex-role identification in adult women. It was designed as a survey using standardized tests as the method of data collection. The survey design has been used in a broad range of scientific disciplines and been particularly popular in education and psychology. Survey research is seen as an acceptable method for examining the distribution of one variable, as well as to look at the relationships between variables (Borg & Gall, 1963). The subjects, adult women, were asked to respond to a self- esteem inventory, a hostility inventory, and a sex-role- identification inventory, as well as to fill out a demographic information form. The resulting scores are based on self-report. Self-report measures are appropriate in this study because the writer is interested in what the women consciously view to be acceptable aspects of their feelings and beliefs about themselves and their behaviors. The scores from these measures were analyzed with correlations. 48 49 m l P n The subjects in this study' were 105 adult women who were members of women’s professional organizations or labor unions. They were all employed, 92 full time and 13 part time. Ages ranged from 18 to 76 years. Mean age was 36.97; the mode was 41 years. Sixty- nine percent of the women were between the ages of 27 and 46. Thirty-four and three-tenths percent were single, 51.4% were married, and 12.4% were divorced. The subjects reported that 42.9% of them had no children, 33.3% had one or two children, 18.1% had three to five children, and 2.9% had six or more children. Forty- three and eight-tenths percent of the women were professionals, 40% were clerical, and 12.4% were laborers. The educational levels ranged from high school to Ph.D. and M.D.; 10.5% were high school graduates, 35.2% had some college, 29.5% were college graduates, and 24.8% had graduate educations. The women’s family income varied widely; 12.4% reported belonging in the 0-515,000 range, 15.2% were in the $15,500 to $30,000 range, 30.5% were in the $30,500 to $45,000 range, and 41% were in the $45,500 and up range. (See Table 3.1.) W The researcher sought permission to conduct the study from the University Committee on Research Involving Human Subjects (UCRIHS) at Michigan State University. That permission was granted (see letter in Appendix A). 50 Table 3.1: Demographics (N - 105). Age Mean - 36.97 years SD - 9.564 Range - 18-76 Marital Status: Single 34.3% 36 subjects Married 51.4% 54 subjects Divorced 12.4% 13 subjects Employment: Professional 43.8% 46 subjects Clerical 40.0% 42 subjects Laborer 12.4% 13 subjects (92 full time, 13 part time) Education: High school 10.5% 11 subjects Some college 35.2% 37 subjects College 29.5% 31 subjects Graduate education 24.8% 36 subjects (4 subjects did not report) Number of Children: None 42.9% 45 subjects 1 or 2 33.3% 35 subjects 3 to 5 18.1% 19 subjects 6 or more 2.9% 3 subjects (3 subjects did not report) Family Income: $ 0-$15,000 12.4% 13 subjects $15,500-$30,000 15.2% 16 subjects $30,500-$45,000 30.5% 32 subjects $45,500 or more 41.0% 43 subjects Number of Siblings: 0 4.8% 5 subjects 1 14.3% 15 subjects 2 21.0% 22 subjects 3 23.8% 25 subjects 4 11.4% 12 subjects 5 10.5% 5 subjects 6 9.5% 10 subjects 7 1.0% 1 subject 9 3.8% 4 subjects Sibling Rank Oldest 32.4% 34 subjects Middle 38.1% 40 subjects Youngest 26.7% 28 subjects (3 subjects did not report) 51 An announcement of the research study was made at meetings and/or in organizational newsletters. A request for volunteers to participate in a research survey of women’s self-perceptions and the psychology of“ women was made with the explanation as brief as possible so as to avoid bias of the subjects’ responses. The packet of materials was distributed to women who volunteered. The packet contained the 'three inventories, an answer' sheet, a demographic sheet, an informed consent form, and instructions (see Appendix B). The materials were returned either through the chairwoman of the organization or in direct mail to the researcher. Unused packets were returned. Subjects were assured that all results were confidential, any identifying data would be destroyed after the data were analyzed, and that results would be anonymous so that the subjects’ identity could not be connected to their results. Subjects were informed that they could withdraw at any time. They were encouraged to answer all items on the inventories and demographic form, and informed that an accumulation of nonresponses by a subject would result in her results being invalid and therefore dropped from the study. Subjects were asked to sign an informed consent form, which included a statement that the results could be used in any publication that might result from the study, given that results were always held anonymous and confidential. 52 Data Analy§i§ The writer hypothesized that there are relationships that exist between sex-role identification, self-esteem, and hostility. The data collected were scores from the Bem Sex Role Inventory, the Buss-Durkee Hostility Inventory, and the Rosenberg Self-Esteem Scale, as well as data from the demographic form. Measures of masculine sex-role identification, feminine sex-role identification, androgyny, and hostility were correlated with measures of self- esteem. These data were analyzed to determine whether a linear relationship existed, the level of that relationship (whether or not the relationship is statistically significant), the direction of the relationship, and whether or not the relationship had predictive value. Variable; The variables in this study were: Self-esteem--Score on the Rosenberg Self-Esteem Scale, which reflects global self-esteem. Hostility--Scores on the Buss-Durkee Hostility Inventory: global hostility, covert hostility, overt hostility. Sex-role identification--Scores on the Bem Sex Role Inventory: femininity score, masculinity score, androgyny score. Definitions of Variables The 'variables investigated 'hi this study, self-esteem, hostility, and sex-role identification, have been defined in various ways. Coopersmith (1967) views self-esteem as an attitude toward 53 the self that is enduring and tends to be maintained, reflects approval or disapproval, and indicates if the individual believes herself to be worthy, capable, and significant. Rosenberg (1965) defines self—esteem as a positive or negative attitude toward self. He thinks of high self-esteem as reflective of a person who respects herself and views herself as worthy. Low self-esteem to Rosenberg suggests self-rejection, self-dissatisfaction, and even self- contempt. Given that Rosenberg’s instrument was used in this study, his definition was used. This definition does not exclude the more expanded definition of Coopersmith. The second variable, hostility, is defined by the Ameritan Haritage Dictignary as feelings of antagonism or enmity. Roget’s International iha5agrg§ suggests synonyms of opposition, contrariety, and dislike. Anger is defined as a feeling state of extreme displeasure, with synonyms of ill humor and wrath. Buss and Durkee suggest that hostility includes behaviors such as malicious gossip and practical jokes, undirected aggression such as slamming doors cu“ temper tantrums, irritability, negativism, oppositional behavior, resentment, jealousy, and suspicion, as well as negative affect expressed in the style and content of speech. Sex-role identification is defined as an internalized sex-role standard that directs behavior to be masculine, feminine, or androgynous. The sex-role standard is an internalization of society’s sex-type standards. of' desirable behavior for men and women. Thus, a feminine sex-role identification reflects 54 personality characteristics that are socially desirable for females, a masculine identification reflects personality characteristics that are socially desirable for males, and androgyny reflects a mix of the two. The androgynous person has a high level of both feminine and masculine characteristics. Hypothaaas The hypotheses were: Hypothesis 1: There is a negative relationship between global hostility and self-esteem. )iypotheaia 2: There is a negative relationship between feminine sex-role identification and self-esteem. Hypgtheais 3: There is a positive relationship between mascu- line sex-role identification and self-esteem. Hypothasia i: There is a positive relationship between androg- yny and self-esteem. Hypothesia 5: There is a positive relationship between femi- nine sex-role identification and covert hostility. Hypothesis 6: There is a negative relationship between mascu- line sex-role identification and covert hostility. Hypothesi; 7: There is a negative relationship between androg- yny and covert hostility. Hypotheait 8: There is a negative relationship between femi- nine sex-role identification and overt hostility. Hypotheais 9: There is a positive relationship between mascu- line sex-role identification and overt hostility. Hypothesia IO: There is a negative relationship between androgyny and overt hostility. In addition to the relationships between these variables as stated in the hypotheses, other relationships were examined as the data were analyzed. The subclasses of hostility as delineated on 55 the Buss-Durkee were examined for any apparent patterns of relation- ship to self-esteem or sex-role identification. Items on the demographic information sheet were examined relative to variables of hostility, self-esteem, and sex-role identification. Items in 15 II, and III on the demographic sheet will also be examined, in particular to see if there are indicators of the response to the needs of self or other, defense of self or other, and relationship needs versus achievement needs. Because of the number of variables involved, formal hypotheses were not stated for all of these rela- tionships. Mea re 0 elf- te m For the purposes of this study, self-esteem refers to the positive (N‘ negative attitudes or feelings a person has about herself (Rosenberg, 1965). Low self-esteem indicates self- dissatisfaction and rejection. High self-esteem reflects a sense of worthiness as a person and self-respect. Global self-esteem is the sum of positive and negative feelings about the self. The measure for this variable, the Rosenberg Self-Esteem Scale, is comprised of a lO-item Guttman scale. The subjects are asked to respond to each item with ”strongly agree,” ”agree," "disagree," or "strongly disagree." Research on the inventory has demonstrated its face validity (Rosenberg, 1965). The scale is meant to represent the subject’s subjective estimate of the stability of self-image (Tippet & Silber, 1965). The scale measures global self-esteem, 56 i.e., an estimate of the subject’s overall sense of satisfaction or dissatisfaction with self. Global self-esteem is vfiewed as comprising a person’s self-evaluation in terms of feelings of worth or adequacy, being a "good” or "bad“ person, personal skill, sexuality, and appearance. This is in contrast to measures of social self-esteem or social competence in other scales. Social self-esteem and global self-esteem measures are moderately correlated, which demonstrates some convergence but also shows different factor loadings, demonstrating that they are measuring somewhat different variables (Whitely, 1983). For this study, global self-esteem is relevant because the point of interest is women’s evaluation of themselves in many areas of life, not just in social or relational aspects of self- evaluation. The Rosenberg also provides a global picture of self- esteem in terms of a general estimate of feelings about the self, as opposed to other measures that are identified as reflecting a more cognitive picture of self. The Rosenberg is also simple and quick to administer, thus not overloading the subjects with the demands of test taking and increasing the likelihood of completion (Whitely, 1983). e 5 re 0 il’ For the purposes of this study, hostility is defined as feeling or showing enmity (n: antagonism. The Buss-Durkee Hostility Inventory (BDHI) delineates seven subclasses of hostility: assault (A), indirect. aggression (IA), irritability (I), negativisnl (N), 57 verbal aggression (V), resentment (R), and suspicion (S). The inventory was designed to give both a quantitative measure of the intensity of the hostility and a descriptive measure of an individual’s preferred mode of expression of hostility. The inventory consists of 75 true-false items. Factor analysis of the items found two primary factors, covert hostility (CH) and overt hostility (OH) (Bendig, 1962). The covert hostility factor was drawn from the items of the Irritability and Guilt subscales. The overt hostility factor was drawn from items of the Assault and Verbal Hostility subscales. The inventory was normed using male and female college students. Social desirability was controlled for through the process of item construction, focusing on the wording of each item. Construct validity for the BDHI has been demonstrated (Geen & George, 1969; Simpson & Craig, 1967). Tests for predictive validity resulted in correlations from .31 to .51, demonstrating that this instrument does not have good validity for making predictions of hostility expression. There are few measures of anger, aggression, or hostility. The Buss-Durkee has been shown to be a reliable and valid measure of hostility. Although it does not demonstrate good predictive validity in a clinical population, it has demonstrated validity in measurement of feelings of hostility as it is experienced internally. Since many theorists have suggested that women’s hostility is likely to contaminate other aspects of feeling or behavior, the lack of predictive validity for expression of hostility does not eliminate it for this study. 58 Riley and Treibe (1989) in comparing the Buss-Durkee with other measures of hostility, using 120 adult subjects, found the measure to be a valid measure of hostility. asur - 01 n i i For the purposes of this study, sex-role identification is defined as the belief about the association between aspects of one’s self and behavior with that of a stereotypical sex role. The Bem Sex Role Inventory (BSRI) treats masculinity and femininity as two independent dimensions. Thus it becomes possible for one individual to describe herself as masculine, feminine, or androgynous. Previous inventories viewed masculinity and femininity as bipolar opposites. The inventory comprises a Masculinity scale and a Femininity scale, each with 20 personality characteristics. There are also 20 items that have been identified as neutral as to sex role, which make up the Social Desirability scale. The 60 items are scored on a Likert scale of "never or almost never true" or "always or almost always true." 0n the basis of these scorings, each individual receives a score in masculinity, femininity, and androgyny, as well as a social desirability score. Scores can also indicate cross-sex identification or undifferentiated identifica- tion. The scores indicate to what extent the individual describes herself by endorsing masculine or feminine traits (Bem, 1974). The scale has demonstrated high reliability (Bernard, 1984). It has also been demonstrated that the Masculinity and Femininity scales are logically independent. Masculinity and femininity scores 59 correlate with social desirability as the scale is designed, since the score reflects endorsement of socially defined sex-appropriate characteristics. However, androgyny has been shown not to correlate with social desirability and has been shown to measure a tendency to describe oneself with sex-typed standards for both men and women. In their study of the BSRI, Ramanacah and Martin (1984) were able to demonstrate the validity of the Masculine scale for measuring dominance and the Feminine scale for measuring nurturance. The Masculine scale on the BSRI demonstrated that it measures the dominant-instrumental dimension made up of themes of intellectual- social ascending, autonomy, and risk taking. The Femininity scale measures the nurturant, expressive dimension with themes of nurturance, affiliative, expressive concerns and self-subordination. Taylor (1984) also found in her study of the concurrent validity of the BSRI that the trait of instrumentality (Masculinity scale) was significantly related to instrumental behavior and that the trait of emotional expressiveness (Femininity scale) was significantly related to emotionally expressive behavior. Taylor suggested that the results indicated that the BSRI trait estimates were a good base for predicting instrumental and expressive behavior, a better predictor than the subject’s gender. The BSRI was chosen for this study particularly because of this because it allows for extrapolation beyond just the Femininity, Masculinity, and Androgyny scales to the themes that are more meaningful to this study. It is also apparent from the items in the Masculinity and Femininity 60 scales on the BSRI that it gives an indication of the subject’s ability to acknowledge personality characteristics associated with aggression, such as assertiveness, competitiveness, and forceful- 11655. CHAPTER IV ANALYSIS OF RESULTS This chapter' contains the analysis of' the results of this study. The results on the main variables, the hypothesis testing, analysis of variance, and post-hoe analysis are reported. The chapter also includes some limited discussion of the results as related to the qualitative measures. The first section reports the descriptive and statistical data that pertain to the main variables. The frequency distributions for each variable are reported and presented in tables. The established norms for each of the instruments used to measure the main variables are also reported when they are available. Also included is the report of the subscales of the instruments with relevant descriptive statistics. The second section restates the hypotheses and reports the Pearson correlations for each of the hypotheses. A brief description of the results of each tested hypothesis and tables are also included. The third section presents the results of the one-way analyses of variance (ANOVAs) on the main variables and descriptive data on the subjects. Twenty-four ANOVAs were performed, but only those 61 62 with statistical significance are discussed. The results of all 24 ANOVAs are presented in tables in Appendix C. The fourth section reports the results of the post-hoc analysis. The Scheffe method of multiple comparisons was used and is reported in discussion of significant results as well as in tables. The fifth section presents data collected on items designed by this researcher that were included on the demographic sheet. Frequency distributions are reported for some of these items; others are reported in simple counts and percentages. This section is followed by a summary of the results. esults n the Mai l s The main variables were sex-role identification (SRI), self- esteem, and hostility. Each is reported separately with frequency distributions of the scores of the instruments used for measurement. Sex-Role Identifitation The Bem Sex-Role Inventory (BSRI) was used to measure SRI of the subjects. 'This scale yielded three scores for each subject: (a) Femininity, (b) Masculinity, and (c) Androgyny. The Androgyny score is the sum of ‘the Femininity and Masculinity scores and identifies the subject as androgynous if the score is above the sum of the median norm for Femininity and Masculinity (F [4.90] + M [4.95] - A [9.85]). A subject is determined feminine in SR1 if her Femininity score is above a median split (4.90) and masculine in SRI if her Masculinity score is above the median split (4.95). She is 63 assessed as androgynous if both the Masculinity and Femininity scores are above the median split and undifferentiated if both the Masculinity and Femininity scores are below the median split. The mean of the Femininity scores for the total sample was 4.955 with a standard deviation (SD) of .53 (see Table 4.1). The range of scores was from 3.60 to 6.15. The mean Masculinity score for the total sample was 4.850 with a SQ of .758. The range of Masculinity scores was from 2.80 to 6.35. Of the total sample, 29 subjects scored as feminine in SRI, 23 as masculine in SRI, 30 as androgynous, and 23 as undifferentiated. Table 4.1: Frequency Distributions on Main Variables BSRI Sex-Role Identification Scores Mean Standard Deviation This Study This This Range Study Norma Adultb Study Norm Adult Femininity 4.955 5.05 5.11 .530 .53 .68 3.60-6.15 Masculinity 4.850 4.79 4.57 .758 .46 .77 2.80-6.35 SRI Number of Subjects % of Sample Feminine 29 27.6 Masculine 23 21.9 Androgynous 30 28.6 Undifferentiated 23 21.9 aNorms established with college-aged females (N - 340). bStudy of adult females (N = 59). 64 The norms used for the original form of the BSRI were established on 340 female Stanford University students tested in 1978. The mean was 5.05 with a _$_Q of .53 for Femininity. The Masculinity mean was 4.79 with a §Q of .66. A study done at Stanford on adult females ages 31 to 65 (N - 59) yielded norms as follows: Femininity mean 5.11 with SD of .52; Masculinity mean 4.57 with SD of .77. The subjects in the present study appear to have lower average Femininity scores and higher Masculinity scores. The Sfls did not differ greatly. Self-Eataem The Rosenberg Self-Esteem Scale is constructed to yield self- esteem scores ranging from a low of 5 to a high of 35. Each of the 10 items is responded to as (1) strongly agree, (2) agree, (3) disagree, or (4) strongly disagree. The results from this study are reported in Table 4.2. Table 4.2: Frequency Distribution: Rosenberg Self-Esteem Scale Mean SD Range Possible Range 27.39 5.18 12-35 5-35 The subjects in this study rated themselves with a mean score of 27.29 (SD = 5.18). The median was 28.0. These scores suggest that subjects tended to rate themselves in a more positive direction 65 on self-esteem. For subjects to describe themselves in the most negative would result in a score of 5. To choose all moderately negative responses, subjects would receive a score of 15. Subjects in this study rated themselves toward the positive end of the scale on self-esteem. H til'! The Buss-Durkee Hostility Scale was used to measure hostility in this study. The mean global hostility score was 24.143, with a §Q of 9.167 (see Table 4.3). The range of scores was from 6.00 to 50.00. This compares with the normative scores of a mean of 27.74 with a SQ of 8.75. It is important to note that these norms were established on college-aged females. Table 4.3: Frequency Distribution: Buss-Durkee Hostility Inventory Mean SD Range Possible Range Hostility Global 24.143 9.167 6.00-50.00 0-66 Covert 7.914 3.873 1.00-17.00 0-20 Overt 5.505 2.889 .00-14.00 0-14 Norms-~Global (college-aged 27.74 8.750 females, N=88) There are 10 subscale scores on the hostility measure. The first division was Covert and Overt Hostility. The mean Covert Hostility (CH) score was 7.914 with a §Q of 3.873 and a range of 66 1.00 to 17.00 out of a possible range of 0 to 20. The mean Overt Hostility (OH) score was 5.505 with a §Q of 2.889 and a range of .00 to 14.00 out of a possible range of .00 to 14.00. The seven subscales that. make up ‘the Global Hostility score are Assault, Indirect Hostility, Irritability, Negativism, Resentment, Suspiciousness, and Verbal Hostility. These scores are listed in Table 4.4. Table 4.4: Frequency Distribution: Hostility Subscales Mean SQ Range Possible Range Covert 7.914 3.873 1.00-17.00 0-20 Overt 5.505 2.889 .00-14.00 0-14 Assault 2.524 2.122 .00- 9.00 0-10 Indirect 4.952 1.918 1.00- 9.00 O- 9 Irritability 4.829 2.447 .00-11.00 0-11 Negativism 2.029 1.289 .00- 5.00 0- 5 Resentment 1.743 1.754 .00- 6.00 0- 8 Suspicion 2.381 2.077 .00- 8.00 0-10 Verbal 5.686 2.383 .00-12.00 0-13 Global 24.143 9.167 6.00-50.00 0-66 Guilt 3.410 2.069 .00- 9.00 O- 9 The distribution of Hostility scores in this sample reflects an overall lower Hostility score when compared to the norms. It is noteworthy, however, that the norms were established on a sample of college-age females in contrast to the adult subjects in this sample. It is generally believed that levels of aggression/ hostility peak in late adolescence and decline throughout adulthood. 67 Hypothesis Testing The alpha level of .05 was set for the Pearson correlations used in testing the hypotheses. Null hypotheses were tested. Hypotheses are stated here as research hypotheses to reveal the predicted outcome. Table 4.5 shows a summary of the Pearson correlation coefficients for self-esteem. Table 4.5: Summary of the Pearson Correlation Coefficients: Self- Esteem (Alpha - .05) r E Self-Esteem and Hostility -.42 5 .05 Self-Esteem and Femininity .04 = .34 Self-Esteem and Masculinity .46 5 .05 Self-Esteem and Androgyny .43 5 .05 Hypothesis !: There is a negative relationship between global hostility and self-esteem. The results of the Pearson correlation indicated a significant negative relationship between global hostility scores and scores on the self-esteem scale. These results supported the hypothesis that global hostility is negatively related to self-esteem. Nyppthaaia_2: There is a negative relationship between femi- nine sex-role identification and self-esteem. The results of the Pearson correlation indicated no significant relationship between the two variables (3; a .04, p = .34). This result does not meet expectations and is puzzling in light of relevant theory. It will be discussed later in the chapter in contrast to other findings. 68 Nypothesia 3: There is a positive relationship between mascu- line sex-role identification and self-esteem. The results of the Pearson correlation for this relationship were significant (1: a .46, p 5 .05). These results supported the hypothesis that self-esteem is positively related to masculine SRI. Nypothaai; 5: There is a positive relationship between androg- yny and self-esteem. The results indicated a significant positive relationship between the two variables (L - .43, p 5 .05). This supported the hypothesis that self-esteem and androgyny are positively related. A summary of the Pearson correlation coefficients for hostility is shown in Table 4.6. Table 4.6: Summary of the Pearson Correlation Coefficients: Hostility (Global, Covert, Overt) r 2 Global Hostility and Self-Esteem -.42 5 .05 Covert Hostility and Femininity -.O7 . .253 Masculinity -.23 a .01 Androgyny -.23 5 .05 Overt Hostility and Femininity -.26 5 .05 Masculinity .25 5 .05 Androgyny .06 3 .05 69 Hypothesis 5: There is a positive relationship between femi- nine sex-role identification and covert hostility. This hypothesis was not supported. The Pearson correlation was not significant (p = -.07, p . .25). The variable of feminine SRI was not found to be related to covert hostility. Hypothesis 6: There is a negative relationship between mascu- line sex-role identification and covert hostility. The result of the Pearson correlation indicated a significant negative relationship between masculinity and covert hostility (r = -.23, p = .01). This hypothesis was supported. Hypothesis 7: There is a negative relationship between androg- yny and covert hostility. This result also indicated a significant negative relationship between androgyny and covert hostility (i: - -.23, p 5 .05). This hypothesis was also supported. Hypothesis 8: There is a negative relationship between femi- nine sex-role identification and overt hostility. There was a significant negative relationship between overt hostility and femininity (p = -.26, p 5 .05). The hypothesis was supported. Hypothesia 9: There is a positive relationship between mascu- line sex-role identification and overt hostility. This hypothesis was supported. There was a significant positive relationship between masculinity and overt hostility (p = .25, p 5 .05). Hypothesis IO: There is a negative relationship between androgyny and overt hostility. This hypothesis was not supported (1 = .06, p 3 .05). 70 Overall, results from this sample and analyses indicated hostility to be lower in subjects with higher self-esteem and higher in subjects with lower self-esteem. The relationship between SRI and hostility was found to be negative between masculinity and hostility, and androgyny and hostility. There was no support for the hypothesized positive relationship between femininity and hostility. Self-esteem was positively related to masculinity and androgyny. Again, there was no support for the hypothesized negative relationship between femininity and self-esteem. ll significant negative relationship was found between the variable of feminine SRI and overt hostility. Other hypotheses predicting relationships with feminine SRI were not supported. Results of Supplementary Apalysas Hostility Subscales The subscale scores on the Buss-Durkee are relevant though not addressed in the formal hypothesis testing. A summary of results can be seen in Table 4.7. Significant negative relationships were found between self-esteem and the following subscales: irritability (p = -.47, p 5 .05), negativism (p - -.28, p 5 .05), resentment (y -.54, p 5 .05), suspicion (y - -.48, p 5 .05), guilt (p - -.52, p IA .05), and covert hostility (L = ~.66, p 5 .05). Feminine SRI was found to be significantly, positively related to guilt (p - .18, p 5 .05). Feminine SRI was found to have significant negative relation- ships with global hostility (L - -.18, p - .03), irritability (E = -.22, p - .01), verbal (y; - -.27, p 5 .05), and overt hostility 71 Table 4.7: Pearson Correlation Coefficients: Hostility * Self- Esteem and Hostility * SRI Self-Esteem Femininity Masculinity Androgyny Assault -.03 -.ll .09 .02 p-.4l p-.13 p-.l7 p-.44 Indirect -.21 .12 .08 .14 Aggression p=.02 p-.12 p-.21 p-.08 Irritability -.47 -.22 -.08 -.20 p5.05 p-.01 p-.20 p5.05 Negativism -.28 -.05 -.ll -.12 p5.05 p=.32 p=.14 p=.ll Verbal .04 -.27 .35 .14 Aggression p=.35 p5.05 p5.05 p=.08 Resentment -.54 -.15 -.13 -.20 p5.05 p-.06 p=.lO p=.02 Suspicion -.48 -.O7 -.11 -.14 p5.05 p=.25 p=.13 p=.08 Guilt -.52 .18 -.26 .12 p5.05 p5.05 p5.05 p=.12 Global -.42 -.18 .04 .07 Hostility p5.05 p=.03 p-.33 p-.23 Covert -.66 -.O7 -.23 -.23 Hostility p5.05 p=.25 p5.05 p5.05 Overt -.02 -.26 .25 .06 Hostility p-.44 p5.05 p5.05 p-.28 72 (E = -.26, p,5 .05). Masculine SRI was found to have a significant positive relationship with verbal (j; = .35, p 5 .05) and overt hostility (; = .25, p - .05). Masculine SRI had a significant negative relationship with guilt (3; - -.26, p 5 .05) and covert hostility (p - —.23, p 5 .05). Androgyny was found to have significant negative relationships with irritability (L s -.20, p 5 .05), resentment (y = -.20, p - .02), and covert hostility (L = - .23, p 5 .05). The results lend support to the notion that unexpressed anger which is experienced as hostility negatively affects self-esteem. The findings of significant relationships between feminine SRI and hostility subscales, which were in the opposite direction of the significant relationships with masculine and androgynous SRIs, also lend support to SRI as being an important correlate in the subjects’ experience of hostility. Analysia pf Variance To understand possible relationships between the main variables of this investigation and the demographic, socioeconomic variables of this sample, ANOVA was employed. The alpha level of .05 was set for this analysis. Data were analyzed by one-way ANOVA based on the qualitative (demographic) data about the subjects and the main dependent variables of self-esteem scores and SRI scores. Significant results are presented here. A summary of all ANOVAs is found in Appendix C. 73 Self-esteem was related to income (E - 4.30; g = 3,99; p 5 .05). Femininity was related to employment (E = 4.02; fl = 2,98; p - .02). Masculinity was related to employment (E = 4.43; at = 2,98; p . .01) and education (,5 . 3.14; at - 3,101; p - .03). Results between femininity and education (5 - 2.61; at = 3,101; p = .056) and self-esteem and employment (5 - 2.69; at - 2,97; p - .07) missed significant levels. The Scheffe method of multiple comparisons was used to test significance of difference between group means. The comparison of group means of self-esteem by income indicated that the highest income group ($45,500 and up) reported the highest self-esteem, and this group mean of self-esteem (28.98) was significantly different from the self-esteem group means for other lower income groups (24.80, 24.57, 27.58). In the feminine SRI by employment comparison, the clerical group had a significantly higher mean femininity score than the professional group (4.81) or the labor group (5.05). In the masculine SRI by employment comparison, the professional group had a significantly higher mean masculinity score (5.05) than the clerical (4.61) or the labor group (4.65). In the Scheffe post-hoc comparison, no groups were significantly different at the .05 level of significance for the masculine SRI by education comparison. However, the ANOVA reached significance, and group means on the masculine SRI scale were informative (high school = 4.64, some college = 4.62, college = 4.96, graduate school = 5.14). The feminine SRI by education test 74 missed significance, and the group means on the SRI were equally informative (high school - 5.29, some college a 4.95, college = 4.99, graduate school = 4.78). In summary, the comparison of socioeconomic data and the dependent variables lends some support to the notion that higher income is related to higher self-esteem and that SRI is related to education and employment level. Rank-Ordered Items The items included and completed on the Demographic Sheet designed for this research are also of' relevance. 'These were designed in an effort to understand the subjects’ sense of the importance of personal/professional needs/goals, and that of relationships and significant others. Three series of forced-choice items to be rank ordered 1 through 5 were scored with frequency distributions for each item. The results are shown in Tables 4.8 through 4.10. These items, which subjects reported anecdotally as "difficult," suggest support for the theoretical literature which views women as more invested in their relationships and the needs of others than in themselves and their own goals. These results were not tested for statistical significance. Subjects were almost equal in the percentage who would defend child or self as a first priority (see Table 4.8). It should be noted that this was a population of women, 42.5% of whom had no children. "Defend friend or mate" was most frequently chosen as second choice (52.4%). "Defend self" was a third choice for more than 33.3% of the sample. Looking at these choices suggests that 75 these subjects valued and would come to the defense of important others, sometimes before defending self. It is also worthy of note that some subjects would rank defending more distant entities such as country or leader before self, child, or mate. This is a small group of subjects when one might have hypothesized that no one would make these first choices. Table 4.8: Rank-Ordered Items: Defend Against Criticism Sifltr Item 166 Item 167 Item 168 Item 169 Item 170 1 41.9% 42.9% 21.9% 3.8% 2.9% 2 13.1% 30.5% 52.4% 4.8% 2.9% 3 35.2% 19.0% 21.9% 13.3% 8.6% 4 3.8% 1.0% 2.9% 52.4% 29 5% 5 1.0% 5.7% 1.0% 25.7% 56.2% Key: Item l66--Defend self. Item l67--Defend child. Item 168--Defend mate or friend. Item 169--Defend country. Item l70--Defend religious or political leader. The forced choices that placed relationship at odds with goals suggest that the women did not want to have to make the choice (see Table 14.9). Fifty-four' and three-tenths of “the subjects chose leaving a relationship to pursue a goal as their last choice. Subjects were close in first choice of altering goals to preserve a relationship (35.2%) and tolerating criticism to pursue goals (29.5%). Sacrificing goals to preserve a relationship was last 76 choice for about one-third of the subjects (34.3%). The ranking of the first and second choices seems to reflect women’s wanting to preserve some of both. Table 4.9: Rank—Ordered Items: Relationship Versus Goals Rank Order Item 174 Item 175 Item 173 Item 172 Item 171 1 14.3% 35.2% 29.5% 19.0% 5.7% 2 17.1% 21.0% 27.6% 28.6% 4.8% 3 10.5% 23.8% 28.6% 22.9% 15.2% 4 22.9% 16.2% 10.5% 25.7% 19.0% 5 34.3% 2.9% 2.9% 2.9% 54.3% Key: Item l74--Sacrifice goals to preserve relationship. Item l75--Alter goals to preserve relationship. Item l73--Tolerate criticism from loved one to pursue goal. Item l72--Alter relationship to preserve goal. Item l7l--Leave relationship to pursue goal. In the statements of values, subjects responded again in the direction of placing others first (see Table 4.10). Protecting other was first or second choice for 48% of the subjects. "Helping someone I care about reach a goal" was first or second for 75.2% of the subjects, whereas "Achieving a goal" was first or second choice for 54.3% of the subjects. "Winning a competition" was last choice for 71.4% of these subjects. Competitiveness, a trait stereotypically associated with masculinity, was clearly not an important part of how these subjects viewed themselves. This is in a group of subjects of whom 21.9% reported a masculine SRI and 28.6% reported an androgynous SRI. 77 Table 4.10: Rank-Ordered Items: Statements of Values Rank Order Item 176 Item 180 Item 179 Item 171 Item 178 1 19.0% 33.3% 9.5% 38.1% 1.0% 2 29.5% 41.9% 19.0% 16.2% 3.8% 3 23.8% 15.2% 24.8% 24.8% 8.6% 4 21.0% 6.7% 34.3% 17.1% 13.3% 5 4.8% 1.0% 10.5% 1.9% 71.4% Key: Item 176--Protecting others. Item 180--Helping someone I care about reach a goal. Item l79--Asserting my rights. Item 17l--Achieving a goal. Item l78--Winning a competition. In response to incomplete sentences "The way I usually behave when I’m angry is . . ." and "I usually feel this way about myself when I am angry . . . ," subjects described themselves as more likely, as a group, to be withdrawn and feel bad or to be hostile and feel bad about themselves (61.9%). Nine and five-tenths percent of the subjects viewed themselves as expressing TM” handling their anger constructively and feeling good about themselves. This group appeared to have a global hostility mean (20.9) lower than the mean of ‘the total sample (24.14). The sample population described themselves most predominantly as not assertive and not feeling good about themselves when they were angry. This information provides an interesting contrast to the inventory scores. In looking at these subgroups of subjects and their income and education, there do not appear to be any meaningful trends (see Table 4.11). 78 Table 4.11: Behavior of Subjects, by Income and Educational Level Behave Behave Appropriately Withdraw Hostilely % of % of % of n Subgroup p Subgroup p Subgroup Educational Level Graduate school 4 40.0 5 15.0 8 25.0 College 1 10.0 10 31.0 12 37.5 Some college 4 40.0 14 43.0 10 31.0 High school 1 10.0 4 12.5 2 6.0 Income Level $45,000 & up 6 10 15 $30,500-$45,000 3 6 9 315,500-330,000 0 12 5 $0-$lS,OOO l 4 3 In response to a Likert-scale item concerning whether subjects saw themselves as like mother or father, 36.2% saw themselves as more like mother, 39.0% as more like father, 20% as like both, and 3.8% as like neither. Results indicate that the subjects in the "more like mother" group had higher feminine SRI scores, whereas subjects in the "more like father" group had higher masculine SRI scores. The subjects in the “like both" group had the highest self- esteem, and the subjects in the "like neither" group had the lowest self-esteem. 79 In response to items requesting three adjectives to describe mother, father, and self, most subjects described themselves with all positive adjectives; only 23 subjects used any negative adjectives to describe themselves. Twenty-eight subjects responded with all positive adjectives for mother, father, and self. Twenty- two subjects described only mother negatively, 23 subjects described only father negatively, and 26 subjects described both parents with negative adjectives. So while only 23 subjects described self as negative, 69% of them described at least one parent as negative. Four subjects described parents as positive and self as negative, and two subjects did not respond to this item at all. Summary pf Results The following summarizes the results presented in this chapter: 1. The total sample mean for the SRI of feminine indicated a sample 'that, (”1 average, viewed themselves as feminine or androgynous in identification. The sample included representative subgroups of feminine, masculine, androgynous, and undifferentiated identifications. . 2. The total sample overwhelmingly reported viewing themselves as having high self-esteem. 3. The scores on the hostility measure indicated that the total sample had a normal distribution on the hostility scales. Taking into account the ages of these subjects, the distribution appears a close match for the college-age distribution on which the 80 instrument was normed. On the whole, these females did not report high levels of hostility. 4. A significant relationship existed between hostility and self-esteem in this sample; the lower the self-esteem, the higher the hostility. 5. No statistically significant relationship between feminine SRI and self-esteem was found in this group of subjects. 6. A statistically significant relationship was found between a masculine SRI and self-esteem. In this sample, the more masculine the SRI, the higher the self-esteem. 7. A statistically significant relationship was found between androgyny and self-esteem. In this sample, the more androgynous the SRI, the higher the self-esteem. 8. No relationship was found between feminine SRI and covert hostility. A negative relationship was found between feminine SRI and overt hostility. ‘This relationship was statistically significant. 9. A negative relationship of statistical significance was found between masculine SRI and covert hostility. A positive relationship was found between overt hostility and masculinity. 10. A negative relationship of statistical significance was found between androgyny SRI and covert hostility. There appeared to be no relationship between androgyny SRI and overt hostility. 11. Self-esteem was found to be related to income; the lower the income, the lower the self-esteem. 81 12. Feminine SRI was related to type of employment. The more unskilled the employment, the more feminine the SRI. l3. Masculinity was related to level of education. The higher the level of education completed, the more masculine the SRI. 14. A relationship between masculine SRI and employment was found. The higher the masculine SRI, the more likely the subject to be professional. 15. Subjects reported a trend toward giving equal priority to self and other but greater priority to the importance of relation- ships over goals. 16. Subjects reported predominantly not feeling good about themselves when they were angry and evaluated their behavior when angry as not constructive. Subjects who reported positive feelings about themselves and their behavior in angry situations also had a lower average hostility score. 17. Subjects reported more positive descriptors for themselves than for their parents. The majority of subjects had something negative to say about at least one parent, but only a small group described themselves in negative terms. 18. The subjects’ scores on the Likert scale "like mother or father" supported the findings on the BSRI. The percentage of subjects in categories ”like mother," ”like father," or "like both" closely matched the percentages of feminine, masculine, and androgynous scores on the BSRI. CHAPTER V CONCLUSIONS This chapter is devoted to the discussion of the results of this research study designed to investigate the relationships among self-esteem, sex-role identification (SRI), and hostility; First, the limitations of the study are discussed. Next, the results on the main variables, the hypothesized relationships, and the relationship between main variables and demographic data are reviewed and discussed. Finally, explanations for the findings are explored and overall implications of the study examined. In general, this study found masculinity and androgyny to be related to hostility and self-esteem. Femininity did not show strong relationships to the other main variables but did appear related to lower' education and employment status. The results revealed interesting relationships between some demographic data and the main variables. These findings suggest strong relationships between external, socioeconomic variables and internal self- descriptors of the subjects. In an effort to provide parameters for this study, the limitations are discussed first. Limitatipus pf the §tugy As with most research, this study has limitations that restrict its generalizability and usefulness. The sample population of this 82 83 study was drawn from groups of working women. The age range was wide, from 18 years to 76 years, with the mean age of 36 years. The majority of the subjects (51%) were married, and 42% had no children. A.large percentage of subjects (71%) reported a family income above $30,000. The sample also included many professional women (43.8%), and 54% of the subjects had at least a college degree. The sample was largely educated, upper- or middle-class women who predominantly held professional jobs. The instruments used were the Bem Sex-Role Inventory (BSRI), the Buss-Durkee Hostility Inventory, and the Rosenberg Self-Esteem Scale. All of these instruments are self-administered and rely on self-report. Research on the BSRI has demonstrated it to have high reliability. It also has been demonstrated to have validity in measuring stereotyped SRIs. It is a heavily researched instrument that avoids polarizing femininity and masculinity and has established androgyny as a measurable characteristic. It has been shown to have good predictive validity for instrumental and expressive behaviors. The major limitations of the BSRI may be that it assumes that SRIs are aligned with socially desirable traits and stereotypical views of sex-roles. The Rosenberg Self-Esteem Scale is brief and easy to take. It has face validity, although no research was found to substantiate validity or reliability. 'It was chosen because of its brevity since it was being used with other lengthy inventories. Asking yet more 84 time from the subjects could have resulted in a lower rate of return. The Buss-Durkee inventory is limited in its predictive validity, although it has shown reliability and face, convergent, and discriminant validity. Social desirability of items was not eliminated but reduced significantly by item construction. The Buss-Durkee has been established as an accepted measure of hostility in research. The methodology and design limitations include having to rely on a nonrandom sample of the subjects asked to complete the survey. Subjects volunteered and thus self-selected. It is likely that individuals who would volunteer for and complete this survey have higher self-esteem and are more highly motivated toward self- understanding and achievement than individuals who heard about the study but chose not to participate. Women who belong to organizations in general may represent a select population of women who are more active, involved, and feel more empowered and supported than the average working woman. The analysis is not a complete inventory of all possible ways these data could have been examined. Limited choices were made based on relevance, economy, and clarity. Other correlations or ANOVAs might have provided different or additional information. The analysis of the results was also limited to the use of descriptive statistics. A sample of adult women working outside of the home was drawn, and measures were used to determine whether the hypothesized notions had a basis in reality for a sample of women. 85 The results provided a description of the sample and a summary of how these women described themselves. The survey method of research is also limiting in that many individuals do not complete the survey. In this case, more than 300 surveys were distributed and 110 were returned completed. The kind of information obtained is limited by the survey' method. 'The instruments must be simple, straightforward, nonintrusive, under- standable by a range of individuals, and thus amenable to self- administration. The conditions under which subjects filled out the survey were not controlled. vi w ' ' 1 Naip Variablaa The subjects identified themselves to be fairly evenly divided in SRI on the BSRI. In this sample, 27.6% of the subjects scored as feminine SRI, 28.6% as androgynous, 21.9% as masculine, and 21.9% as undifferentiated. As compared to norms on adult females, this sample was less feminine, more masculine and androgynous. The majority of subjects in this study reported high self- esteem (mean - 27.39, 50 - 5.18). In the total sample, 84% scored above 22, which is a moderately positive self-esteem score. This is a surprising finding. However, as noted earlier, subjects with higher self-esteem may have been more likely to fill out the survey. It is also possible that the self-esteem instrument is inadequate to get a true picture of self-esteem in adult females. 86 The subjects reported an average level of hostility that reflected a frequency distribution lower than reported on college- aged females. It is believed that this result is reflective of adult subjects rather than adolescent subjects. The subscale scores demonstrated that subjects acknowledged high levels of indirect hostility, irritability, and verbal hostility with lower levels of resentment, suspiciousness, and assault. This is in keeping with the stereotypes regarding women and hostility. It is in contrast, however, with the negative correlations found in this study between the Feminine SRI and Verbal Hostility, Resentment, and Irritability subscales. Nyppthasis Tagtipg oth 's . The findings of a low to moderate relationship between global hostility and self-esteem were in the expected direction and gave solid support to the hypothesis. Self-esteem was higher when hostility was lower. These findings had further support in the items designed by the researcher, which indicated that subjects who reported being able to be directly assertive and felt good about themselves when they were assertive had a lower average hostility score than subjects who repressed their anger and reported feeling bad about themselves. These findings indicate that for women maintaining high self-esteem is related to being able to use anger: constructively and directly instead of' developing hostile feelings. 87 Nyppthaaia_z. The results showed no significant relationship between femininity and self-esteem. This was an unexpected finding. While this result may reflect the inadequacies of either scale, it suggests also that femininity is complex and does not lend itself to direct linear relationships. It is possible that the group of feminine women was composed of at least two subgroups. The subgroups could be divided between those women who had a strong feminine SRI and felt good about themselves and one group with a strong feminine SRI who felt bad about themselves. Hypothasis 3. The data strongly supported Hypothesis 3. There was a positive relationship between masculine SRI and self-esteem. The more masculine the SRI, the higher the self-esteem scores. Women subjects reported themselves feeling more positive about themselves when they identified with more stereotypically masculine traits. This contrasts with the lack of relationship with feminine SRI and self-esteem. This suggests that women feel better about themselves if they identify with the societally more valued masculine traits such as competitiveness and achievement. The internalization of the value associated with the traits was supported by these results. Hypothagig g. The results on Hypothesis 4 also were in the predicted direction. Androgyny was related positively to higher self-esteem. The mixture of high feminine and high masculine traits was correlated positively to higher self-esteem. This strongly supports the theoretical hypothesis that individuals with the ability to identify with a broad range of characteristics--feminine 88 and expressive as well as masculine and instrumental--feel better about themselves. These individuals also have more adaptations to choose from and thus may feel more generally competent, which has also been shown to be related to higher self-esteem. Nypotheaia 5. The test of this hypothesis indicated no relationship between feminine SRI and covert hostility. It was believed that there would be a positive relationship between these two variables. This again raises the question about the group of individuals in the high feminine SRI group. There is also a question about the Femininity SRI scale since none of the findings provided strong significant relationships between feminine SRI and any other main variable. flypothesis 6. This hypothesis test revealed a negative relationship between covert hostility and masculinity. The masculine-identified subjects were less likely to use covert methods of expressing hostility. It was expected that this would be a stronger relationship. It is not known whether these results would be different if the subjects were men instead of women. H ot esi 7. The correlation of covert hostility and androgyny revealed a negative relationship. Subjects with a mix of high feminine SRI and high masculine SRI reported less covert hostility, as was expected. The degree of the relationship was less than expected but suggested at least a trend for androgynous subjects to be less likely to use covert hostility. This finding also pointed out a difference between high feminine SRI subjects and androgynous who had both high feminine SRI and masculine SRI. 89 Hypothesia 8. The testing of Hypothesis 8 resulted in findings in the expected direction. There was a negative significant relationship between femininity and overt hostility. The more femininely identified the subject, the less likely she was to express hostility in an overt manner. This result, in conjunction with the results on Hypothesis 5 (feminine SRI and covert hostility), raises a number of questions. It is possible that the Bem SRI scale is composed of items that fail to acknowledge much hostility of any kind. The more feminine the SRI, the less likely she is to be able to consciously acknowledge any form of hostility. The self-report situation also may result in subjects presenting themselves in a more positive light, and for feminine-identified subjects that could mean a greater need to deny hostility. This fits with the stereotype of women not having these feelings or motives. Hypothesis 9. The results of this hypothesis test were significant and were in the predicted direction. Women who identified with more masculine characteristics also reported a trend toward more overt expression of hostility. Identification with a stereotypical masculine role would allow for the endorsement of more direct expression of anger and hostility. These results, in combination with the negative relationship between feminine SRI and overt hostility, match stereotypes of how men and women express anger. It would be interesting to see with male subjects whether 90 the SRI is related to these variables in the same direction, to determine whether the important variable is SR1 or gender. H 0th 1 0. Results (”1 this hypothesis revealed no relationship between overt hostility and androgyny. Subjects with combined high feminine SRI and masculine SRI reported higher self- esteem and less hostility. The more individuals are able to identify with a wide range of characteristics, the better they feel about themselves and the less likely they are to experience hostility. Analysis of Variance The feur combinations that were statistically significant at the .05 level were negative self-esteem by income, feminine SRI by employment, masculinity by education, and masculinity by employment. These results strongly support the relevance of the social/cultural context when considering state/trait variables of individuals or groups of individuals. Lower income women had self-evaluations that were less positive than those of higher income women. Women in professional jobs had lower femininity scores than women in clerical or labor jobs and higher masculinity scores. The women with graduate educations also had higher masculinity scores. These results suggest that women’s status in this society, which usually involves being in lower paid jobs with less access to higher education and professional positions, is related to the women’s self-evaluation. The results also support the notion that women internalize the social values associated with status, income, and 91 education. Having a low income means one is a less valued person and thus values herself less. Since the vast majority of adult individuals living under the poverty line in this country are women, this has serious implications for these women. Masculinity was also associated with higher income and status as well as higher self- esteem in this sample of women. This again supports that the higher value given to masculine traits in this culture affects personal, individual self-evaluations. The subjects who identified themselves as single had more negative self-esteem than the married or divorced group. This can be thought of in terms of societal expectations that women are supposed to be married. It also may be related to lower income as a single person. Since subjects reported family income, it is likely that the single women had a lower income. The ANOVAs and post-hoe comparisons consistently indicated that the subjects felt. better about themselves when they had higher income and that masculinity was associated with higher levels of education and professional levels of employment, both of which were likely to be associated with higher income. Wm As subjects responded to items designed by the researcher, they confirmed their positive self-esteem in that only 23 of the 105 subjects used any negative adjectives in their self-descriptions. The subjects were less positive about their parents, with 71 subjects describing one or both parents with negative terms and only 92 28 describing both parents with all positives. This contrasts with 96% of the subjects describing themselves on Item 165 as like either mother, father, or both (36% like mother, 39% like father, 20% like both). It is difficult to interpret the meaning of subjects making negative statements about parents whom they also say they are like and ,yet describing themselves in entirely positive terms. An optimistic view might be that subjects acknowledged the best of what they were given in their relationships. It is also possible that it is more comfortable or acceptable to criticize others than to criticize self. This would not fit with the feminine stereotype of blaming self and protecting others. Subjects did not describe themselves as feeling good when angry or feeling good about how they behaved when angry. Only 9.5% of the subjects indicated that they believed they behaved appropriately, assertively, and felt good about themselves when angry. Thirty- three subjects (31.4%) described being withdrawn and feeling bad, 30% felt hostile toward others and bad about themselves, while 7% said they felt hostile toward others but okay about themselves. These results support the notion that women have internalized a societal standard that women are not supposed to get angry because if they get angry they will feel bad about themselves whether they express it or not. The responses also revealed that the negative statements about self were not objective evaluations of behavior but strong statements about being "stupid," "silly,” "out of control" for having the angry feelings or reaction. The vast majority did 93 not see themselves as reacting to their anger in a manner they felt good about. The subjects generally viewed themselves positively on other measures but when angry reported opposite evaluations. The SRI of the respondent did not relate to these negative self- evaluations. It was also true that for the subjects who did feel assertive and positive there was a mix of SRIs. It would appear that this is related to societal expectations based on gender rather than SRI. Items designed by the researcher that were rank ordered indicated that masculine values were rejected by subjects as first or: second choice. If“ items such as "asserting my rights" and "winning a competition" are viewed as indicative of the willingness to stand separate from others, the results showed that these subjects chose staying connected to others. In fact, 71.4% of the subjects ranked "winning a competition" as a last choice. "Asserting my rights" was ranked fourth as first choice (9.5%). This is in contrast to the nearly equal choice of "achieving a goal" (38.1%) and "helping someone else achieve a goal" (33.3%) for first choice. In the "defend against criticism” series of items, it was clear that "defending self" and someone close were priority choices. Given the results on other items, if this study were to be repeated it would be helpful to change the "defend country" item to "defend a parent." The ‘results obtained were not surprising except that "defending self" was given equal ranking with "defending child." 94 This is not what was predicted. The theoretical notion that women will feel more comfortable using aggression to defend a child or mate rather than defending themselves was not supported here. ”Defending self” and "defending child” were almost equal choices for the subjects. These items might be more revealing if redesigned to read "use my anger to defend.” This reported willingness to defend self on these items does not match with the other results in this study, in which subjects reported feeling bad about themselves when angry and not acting assertively or defending themselves but withdrawing. The items designed to create a conflict between relationships and goals produced results indicating subjects did not want to have to make a mutually exclusive choice. The subjects did not want to have to give up relationships for their goals. Seventy-three percent identified "leave relationships to pursue a goal" as their fourth or fifth choice. At the same time, subjects were also not eager to sacrifice their goals entirely for relationships since only 14.3% chose "sacrifice goals” as their first choice. Summar This study of adult working women yielded some results in the predicted directions and raised further questions. The subjects were fairly equally divided in SRI, and the vast majority reported high self-esteem. The hostility scores produced a normal distribution slightly lower than the distribution of the college- age norms. This is believed to be reflective of the fact that this 95 was an adult population and hostility is believed to decrease from adolescence through adulthood. The correlations revealed positive linear relationships between androgyny and self-esteem, masculinity and self-esteem, and masculinity and overt hostility. A negative linear relationship was found between hostility and self-esteem, masculinity and covert hostility, androgyny and covert hostility, and femininity and overt hostility. No relationship was found between femininity and self- esteem, femininity and covert hostility, and androgyny and overt hostility. The ANOVA results indicated that the more masculine the SRI, the higher the education and level of employment. The more feminine the SRI, the more likely the subject was to have a lower level of employment. The results also revealed that the lower the income, the more likely the subject was to have lower self-esteem. Single subjects were also more likely to have lower self-esteem. Subjects appeared to choose relationships over singular pursuit of goals but avoided making mutually exclusive choices. Subjects avoided choices that resulted in a ”standing alone or separate" status and rejected masculine-valued choices such as "winning competitions." Subjects reported difficulty with assertiveness and did not feel good about themselves when angry. They were more willing to make negative evaluations of parents than self but saw themselves as like one or both parents. 96 The results supported the hypothetical notion that hostility damages self-esteem and that women have a difficult time integrating the experience of anger and hostility into a positive picture of themselves. This is an especially relevant finding since this was a group of women who were more highly educated, with higher family incomes and more professional status than the general population. The findings also indicated that SRI was related to the subject’s level and expression of hostility. There were also findings, however, that suggested that difficulties with direct expression of anger may be both gender related and related to SRI as well. All the subjects were female, and the vast majority (91%) reported feeling bad about themselves and how they behaved when angry. This same group of subjects was fairly equally divided in SRIs. This indicates that no matter what the subject’s SRI, she was still very likely not to feel that she handled herself well when angry. While femininity was not negatively related to self-esteem or positively related to hostility, the lack of a directional relationship suggests that femininity is a complex variable. ‘The possibility exists that the scale includes two groups of women: one who feel good about themselves and are not hostile, and one group who feel bad about themselves and are hostile. It is also possible that the subjects were as ambivalent as the culture is about femininity, and about themselves as they embrace it. Certainly the negative relationships between femininity and socioeconomic data and positive relationships between self-esteem and the same data reflect the generally lower socioeconomic status of women in this society. 97 Overall, it would be informative to contrast these studies with male subjects. The question remains whether SRI is the important variable versus gender. It is also assumed that men have better access to their anger and aggression, but do they feel they use their anger in constructive and self-enhancing ways? 00 men, in fact, feel good about themselves when angry? Where are the differences based on gender and SRI, and what are the restrictions that men experience because of societal norms and expectations? It does not seem likely that one gender can exist in this society feeling totally free while the other is restricted. The findings of this study have raised many questions. The adult women subjects in the study informed us about themselves in many ways and left us curious in many others. Their self- descriptions revealed the effect of the societal context in which they lived and, to a certain extent, how that context affected their internal experience of themselves as women. It leads to the conclusion that one cannot assume that the problems and struggles of women are internal and self-created. Internal conflicts must be distinguished from legitimate reactions to a restrictive (society that continues to bind in many subtle ways. The findings support that development, behavior, and personality traits must be viewed in the context of societal expectations, pressures, and constraints. If societal norms restrict direct expression of anger and repressed anger leads to hostility and hostility damages self-esteem, then women are caught 98 in a circular bind with no route to feeling better about themselves and owning their own anger. If low income is a contributing factor to low self-esteem, then women’s restricted access to higher paying employment will also continue to contribute to the struggle for higher status and higher self-esteem. These restrictions are also likely' to result. in higher: levels of' hostility that fuel this negative, circular pattern for women in this culture. Racommendations for Futura Bagearch Future investigation into women’s ability to tolerate standing separately while maintaining self-esteem would be informative. A relevant issue not raised in this study is how much fear plays a role in the difficulty with anger and assertiveness versus the concern over loss of a relationship. One must also wonder how much poor access to direct anger played a role in the high self-esteem scores for these subjects. Would more objective, balanced self- appraisals be gained from subjects with more comfort when angry? In general, more research is needed to study how societal norms and stereotypes influence human behavior for both genders. APPENDICES APPENDIX A PERMISSION LETTER FROM THE UNIVERSITY COMMITTEE ON RESEARCH INVOLVING HUMAN SUBJECTS 99 MICHIGAN STATE UNIVERSITY UNIVERSITY COMMITTEE ON RESEARCH INVOIVING EAST LANSING 0 IICHIGAN 0 «524-1111 HUIIAN sonnets (UCRIHS) 206 ram HAIL (517) 353-973. November 23. 1988 W3 Diane B. Trebilcock 1200 Woodcrest East Lansing, MI 48823 Dear Ms. Trebilcock: Subject: ”AN INVESTIGATION OF THE RELATIONSHIPS BETWEEN HOSTILITY, SELF-ESTEEM AND SEX-ROLE IDENTIFICATION IN WOMEN W" The above project is exempt from full UCRII-IS review. The propgsed research protocol has been reviewed by another committee mem r. The rights and welfare of human subjects appear to be protected and you have approval to conduct the research. You are reminded that UCRIHS approval is valid for one calendar year. If you plan to continue this pro'ect be ond one year, please make provisions for obtaining agpropriate CRI S approval mum]; Anyjchan‘gles in procedures involving human subjects must be reviewed by CR1 8 prior to initiation of the change. UCRIHS must also be notified promptly of any roblems (unexpected side effects, complaints, etc.) involving human su jects during the course of the work. Thank on for bringing this project to my attention. [f] can be of any future elp, please do not hesitate to let me know. Sincerely, JKH/sar cc: .1. Powell APPENDIX B INFORMED CONSENT FORM, INVENTORIES, AND DEMOGRAPHIC INFORMATION SHEET TOO INFORMED CONSENT I agree to participate in a research investigation being conducted by Diane B. Trebilcock, a doctoral student at Michigan State University. I understand that this is a research project investigating the psychology of women. I understand that the extent of my participation involves filling out three standardized inventories (the Buss-Durkee Inventory, the Rosenberg Scale, and the Bem Inventory) as well as a demographic information sheet. My participatiOn will take approximately thirty minutes to an hour. I understand that my participation is voluntary and I may choose not to answer all questions and may withdraw at any time. All results will be treated with strict confidence, and mw results will be anonymous in any report of research findings. If I am interested in the results I may attend a meeting in which the results of the research will be reported and discussed. Other questions about the research may be directed to Diane Trebilcock at (517) 332-7160. Signed Date instructipns tp the Participant; This is a research investigation designed to provide information about the psychology of women. Your participation is entirely voluntary. Therefore, you may withdraw from this study at any time. You are encouraged, but not required, to answer all items and questions. Answering all the items will help me to obtain the most reliable results. It is expected that it will take between thirty minutes and one hour for you to complete these surveys. All results will be held in strict confidence, and your results will be anonymous to this investigator as well as in any report of research findings. Each packet contains 1 Informed Consent and Instruction Sheet, 1 Buss-Durkee Inventory, 1 Rosenberg Scale, 1 Bem Inventory, and 1 Demographic Information Sheet. Please complete the Demographic Sheet first. Responses should be written on the Demographic Sheet itself. The responses to the other three inventories should be recorded on the answer sheet provided. Instructions for the individual inventories appear with the questions. Please be sure to record your responses so that the number of the item is the same as the number on the answer sheet. Thank you for your participation!! lOl DEMOGRAPHIC INFORMATION Age___ Marital Status No. of Children Ages Employment (please indicate type of position held) Part time Full time Was there a period of time when you were out of the job market because of schooling or child-care responsibilities? No Yes If yes, dates Level of education: High school Some college College Graduate school Household income: $O-$15,000 $15,500-$30,000 $30,500-$45,000 $45,500 or above Have you used the services of a health professional in the last year? Please check: Physician Nurse or physician’s assistant Psychologist or psychotherapist Social worker Dentist Chiropractor Hobbies, recreational activities, sports: Family Histpry Number of siblings Your rank in these Three words you would use to describe your mother: Three words you would use to describe your father: Please place an X on the line indicating whom you are most like: Mother Father Three words you would use to describe yourself: 102 Please rank order the following statements, using numbers l,2,3,4,5. The number 1 indicates this is the first choice or most like you; 5 indicates last choice or least like you. I. I will defend myself against unfair criticism. I will defend my child against unfair criticism. I will defend my mate or friend against unfair criticism. I will defend my country against unfair criticism. I will defend a political or religious leader against unfair criticism. Please rank order the following statements, using numbers l,2,3,4,5. The number 1 indicates this is the first choice or most like you; 5 indicates last choice or least like you. 11. _I will leave a relationship to pursue my own interests. :I will alter my personal goals to preserve a relationship. :I will tolerate disapproval from someone I love to pursue _my own goals. Ihwill sacrifice my personal goals to preserve a relation- 5 ip. I will alter a relationship to pursue my own goals. Below are five statements of values. Please rank order the follow- ing 1 through 5, with l as more like you or more important to you and 5 as not like you or least important to you. III. Protecting others Achieving a goal _Winning a competition _Asserting my rights :Helping someone I care about reach a goal Please respond to the following by describing what is most like you. The way I usually behave when I’m angry is I usually feel this way about myself when I am angry APPENDIX C ANOVA TABLES 103 Table C.l: Results of the Analyses of Variance Mean E- Sign. Group Source at Square Value p - Means Count Group Self-Esteem x Education Between 3 55.23 2 13 10 26.73 11 High school Within 100 25.99 ' ' 26.16 37 Some college 27.42 31 College 29.44 25 Graduate school Self-Esteem x Income Between 3 107.09 4 30 005 24.80 13 50-515,000 Within 99 24.50 ' ' 24.57 16 515,500-530,000 27.58 31 530,500-545,000 28.98* 43 545,500+ Self-Esteem x Employment Between 2 71.00 2.69 .073 28.04 45 Professional Within 97 26.40 27.64 42 Clerical 24.38 13 Labor Self-Esteem x Marital Status Between 2 38.48 1.43 .244 26.11 35 Single Within 99 26.86 27.91 54 Married 28.08 13 Divorced Femininity x Income BGCWEEM 3 .4] 1 45 23 4.97 I3 50-315,000 Within 100 .28 ' ' 5.12 16 515,500-330,000 5.03 32 530,500-545,000 4.83 43 $45,500+ 104 Table C.l: Continued. Mean fi- Sign. Group Source at Square Value p - Means Count Group Masculinity x Income Between 3 1.10 1 97 12 4.86 13 $O-$15,000 Within 100 .56 ° ’ 4.53 16 515,500-530,000 4.75 32 $30,500-S45,000 5.02 43 $45,500+ Androgyny x Income Between 3 .18 23 .87 9.80 13 $0-$15,000 Within 100 .79 ' 9.64 16 $15,500-$30,000 9.77 32 $30,500-S45,000 9.85 43 $45,500+ Femininity x Like Mom or Dad Between 3 .23 79 50 5.02 38 Mom Within 100 .29 ° ' 4.96 21 Both 4.86 41 Dad 5.16 4 Neither Masculinity x Like Mom or Dad Between 3 .56 97 4] 4.79 38 Mom Within 100 .58 ' ' 5.08 21 Both 4.84 41 Dad 4.51 4 Neither Androgyny x Like Mom or Dad Between 3 .55 70 55 9.81 38 Mom Within 100 .79 ° ° 10.04 21 Both 9.70 41 Dad 9.68 4 Neither 105 Table C.l: Continued. Mean 5- Sign. Group Source at Square Value p . Means Count Group Self-Esteem x Like Mom or Dad Between 3 20.50 752 524 26.63 38 Mom Within 99 27.23 ' ' 28.67 21 Both 27.48 40 Dad 26.25 4 Neither Femininity x Employment Between 2 1.040 4 020 02] 4.812 46 Professional Within 98 .259 ' ' 5.112* 42 Clerical 5.054 13 Labor Masculinity x Employment Between 2 2.352 4 431 014 5.054* 46 Professional Within 98 .531 ° ' 4.612 42 Clerical 4.654 13 Labor Androgyny x Employment Between 2 .539 343 710 9.886 46 Professional Within 98 76.906 ’ ' 9.724 42 Clerical 9.708 13 Labor Femininity x Marital Status Between 2 .377 4.832 36 Single Within 100 .280 1°348 '265 5.014 54 Married 5.000 13 Divorced 106 Table C.l: Continued. Mean E- Sign. Group Source at Square Value p - Means Count Group Masculinity x Marital Status Between 2 .103 153 834 4.925 36 Single Within 100 .565 ° 4.838 54 Married 4.812 13 Divorced Androgyny x Marital Status BEtWEED 2 .097 .123 .885 9.757 35 Single Within 100 .793 9.852 54 Married 9.812 13 Divorced Femininity x Education Between 3 .701 2 608 056 5.291 11 High school Within 101 .269 ° ' 4.945 37 Some college 4.995 31 College 4.779 26 Graduate school Masculinity x Education Between 3 1.703 3 143 029 4.635 11 High school Within 101 .542 ° ' 4.615 37 Some college 4.958 31 College 5.142 26 Graduate school Androgyny x Education BETWEEN 3 1.134 1 484 224 9.927 I] High SChOOI Within 101 .764 ' ' 9.561 37 Some college 9.953 31 College 9.921 26 Graduate school REFERENCES REFERENCES Adler, A. (1956). individual psychplpgy at Alfred Aglar (H. L. Ansbacher & R. R. Ansbacher, Eds.). New York: Basic Books. Amaritah_haritaga_ditt1phary (New College ed.). (1976). Boston: Houghton Mifflin. Baruch, G. (1973). Feminine self-esteem self-ratings of competence and maternal career commitment. qurhal of taunaaling Psychol- pgy, ap(5), 487-488. Bem, S. L. (1974). The measurement of psychological androgyny. Journal pf Consulting and Clinipal Psycholpgy, fl2(2), 155-162. Bendig, A. W. (1962). Factor analytic scales of covert and overt hostility. Journal of tonsulting Payghplogy, 25(2), 200. Bernard, L. C. (1984). The multiple factors of sex role identifi- cation: Rapprochement of unidimensional and bidimensional assessment. Journal of Clinica1,E§ythplpgy, 1(4), 986-991. Bernardez, T. (1983). Women in authority: Psychodynamic and interactional aspects. r0 0 ' wom r work wi h hah. Haworth Press. Bernardez, T- (1935. June). A psychosocial approach to psychother- apy: Understanding society’s bias against women. Baggljffg Quarterly, pp. 3-4. Bernardez-Bonesatti, T. (1978). Women and anger: Conflicts with aggression in contemporary women. n f A ’ n Me i 1 Women’s Aaspgiatiph, 55(5), 215-219. Bernbaum, D. W. (1983). Preschoolers’ stereotypes about sex dif- ferences in emotionality: A reaffirmation. f en ic ESXSDQIQQI. 123. 139-140. Borg, W. R., & Gall, M. D. (1973). figuratipha1_ra§garth. New York: David McKay. Brodzinsky, D. M., Messer, S. B., & Tew, J. D. (1979). Sex differ- ences in children’s expression and control of fantasy and overt aggression. Child Devalopment, 59, 372-379. 107 108 Broverman, 1., Broverman, 0., Clarkson, F., Rosenkrantz, P., & Vogel, S. (1981). Sex-role stereotypes and clinical judgments of mental health. In E. Howe11 a M. Bayes (Eds.), meen and mental health. New York: Basic Books. Broverman, I., Vogel, S., Broverman, 0., Clarkson, F., & Rosen- krantz, P. (1972). Sex-role stereotypes: A current appraisal. qurpgl pf §Q§1§l Issues. Z§(2), 59-80. Buss, A., & Durkee, A. (1957). An inventory for assessing differ- ent kinds of hostility. MW. 21(4), 343-349. Chodorow, N. (1976). Oedipal asymmetries and heterosexual knots. §pcial Problems, 23(4), 454-467. Chodorow, N. (1978). The reproduction of mptheuing. Berkeley: University of California Press. Connell, D. M., & Johnson, J. E. (1970). Relationship between sex- role identification and self-esteem in early adolescents. Developmental Psycholugx. §(2), 260. Coopersmith, S. (1967). The nte ent f -es m. San Francisco: W. H. Freeman. Culp, R. E., Cook, A. S., & Housley, P. C. (1983). A comparison of observed and reported adult-infant interactions: Effects of perceived sex. Sex Rolss. 2(4), 475-479. Deutsch, H. (1945). he h 1 w ' tic interpretgtion (Vol. 1). New York: Grune & Stratton. Eagly, A., & Steffen, V. (1989, March). Violent reaction. §£L£. Edmunds, G. (1976). The predictive validity of the Buss-Durkee Inventory. qurnul of Clinical Psyspolpgy, 32(4), 818-820. Eichenbaum, L., & Orbach, S. (1988). Begween women. New York: Viking Press. Elpern, 5., & Karp, S. (1984). Sex-role identification and depres- sive symptomatology. Sex Roles. 19(11/12), 987-992. Feshbach, N. D. (1969). Sex differences in children’s modes of aggressive responses toward outsiders. Mgrrill-Ealmer Quurtgrly, 13, 240-258. Feshbach, N. 0., & Feshbach, S. (1969). The relationship between empathy and aggression in two age groups. Developmental Psychology, 1(2), 102-107. 109 Finman, R., & Berkowitz, L. (1989). Some factors influencing the effect of depressed mood on anger and overt hostility toward another. Journal_of_Besearsb_1n_£ersonalitx. 2;, 70-84. Flax, J. (1978). The conflict between nurturance and autonomy in mother-daughter relationships and with feminism. EQEIDISL 3tudies, s, 171-189. Follingstad, 0., Robinson, E., & Pugh, M. (1977). Effects of conscious-raising groups on measures of feminism, self-esteem and social desirability. Journal_of.§eunseling_£sxcholegx. 35(3), 223-230. Freud, S. (1965). New in h New York: W. W. Norton. Frisch, H. L. (1977). Sex stereotypes in adult-infant play. guild ngglppmens, 18, 1671-1675. Frodi, A., & Macaulay, J. (1977). Are women always less aggressive than men? A review of the experimental literature. Psychp- lpgjspl Bullngn, ss(4), 634-660. Geen, R. G., & George, R. (1969). Relationship of manifest aggres- siveness to aggressive word associations. Esychplogical Reports, 23, 711-714. Gilligan, C. (1962, June). Why should a woman be more like a man? Ps holo a . Gilligan, C. (1982). In i r v i . Cambridge, MA: Harvard University Press. Gump, J. P. (1972). Sex- role attitudes and psychological well- being. Jo urpsl pf Spgjgl Issues, 28(2), 79- 92. Harden, R. R., & Jacob, S. H. (1978). Relationship between aggression of preschoolers and teachers’ sex. Esyppplpgigul Eggpyss, 53, 1172-1174. Hasenbeld, R. H. (1986). Some factors related to self-esteem in women. Disseptusjpn Abstrasts, 16(11), 40538-40548. Heerboth, J. R. & Ramanaiah, N. (1985). Evaluation of the BSRI masculine and feminine items using desirability and stereotype ratings Jou_nal_o__£ensonalitx_Assessment 32(3). 264- 270 Heiser, P., & Gannon, L. (1984). The relationship of sex-role stereotype to anger expression and the report of psychosomatic symptoms. Sex Rules, 19(7/8), 601-611. 110 Heterington, E. M. (1972). Effects of father absence on personal- ity development in adolescent daughters. v 10 me P - gholggl. 1(3), 313-326. Hochschild, A., & Machany, A. (1989). lbs sgsond shift. New York: Viking Press. Hollandsworth, J. G., & Wall, K. E. (1977). Sex differences in assertive behavior: An empirical investigation. qupnsl pf tpunselipg Esychplpgy, 21(3), 271-272. Hoppe, C. (1979). Interpersonal aggression as a function of subject’s sex, subject’s sex-role identification, opponent’s sex and degree of provocation. Journul pf Egrspnsljty, 51, 317-329. Horner, A. (1984). b'e r 'ons an h v 1 i in therapy. New York: Jason Aronson. Horney, K. (1967). i ' e c . New York: W. W. Norton. Hyde, J. S., & Linn, M. C. (1986). lhs psytpplpgy pf gender advunces thruugh metu-anulysjs. Baltimore: Johns Hopkins University Press. Kaplan, A. G. (1984). The self-in-relation implications for depression in women. Work in Progrgss (Wellesley College). No. 84-03. Karam, M. L. (1977). The relationship between sex role behaviors and dimensions of personality adjustment and self-esteem. Dissertation Abstrusts, 38, 18868. Knot, P. 0., & Drost, B. A. (1970). Sex-role identification, interpersonal aggresSion and anger. Esyphplpgjgul Reports, 31, 154. Kohlberg, L. (1981). Th hilo o h o m r 1 e1 ment. New York: Harper & Row. Kohut, H. (1971). lbs anslysjs pf tug sslf. New York: Inter- national Universities Press. LaBarbera, J. D. (1984). Seductive father-daughter relationships and sex roles in women. §gx Rplgs, 11(9/10), 941-951. Lerner, H. (1974). The hysterical personality: A "women’s disease." Comprehensixe_fisxshiatnx. 15(2). 157-164. lll Lerner, H. (1980). Internal prohibitions against female anger. Amerigan qurnal pf Esyshpaualysls, 59(2), 137-148. Lerner, H. (1981). Early origins of envy and devaluation of women. In E. Howell & M. Bayes (Eds.), flumgp and manta! hgalth. New York: Basic Books. Lerner, H. (1985). T an . New York: Harper & Row. Lerner, H. (1987). Female depression: Self-sacrifice and self- betrayal in relationships. In R. Formanek & A. Gurion (Eds.), W. New York: Springer. Leventhal, D. B., & Shemberg, K. M. (1967). Sex-role adjustment and nonsanctioned aggression. Journal pf Experimsptal Rasgarsh in Esrspnaljty, 3, 283-286. Leventhal, D. B, & Shemberg, K. M. (1968). Effects of sex-role adjustment upon the expression of aggression. qurnal pf Per- WW. 8(4). 393 396 Long, V. 0. (1986). Relationship of masculinity to self-esteem and self- -acceptance in female professionals, college students, clients and victims of domestic violence. Jou 5 1t- jpg Q1 jpjtal Esyth plpgy, 55(3), 323- 327. Maccoby, E. E., & Jacklin, C. N. (1974). [he psytholpgy pf ssx differenses. Stanford, CA: Stanford University Press. Mahler, M. (1979). The selatteu papgrs pf Margaret 3. Mahler (Vol. 2). New York: Jason Aronson. Mahler, M. (1981). Aggression in the service of separation- individuation. P a ic u . Mathes, E. W., Adams, H. E., 3 Davies, R. M. (1985). Jealousy: Loss of relationship rewards, loss of self-esteem, depression, anxiety and anger. JoumaLoLEemaUMnLSoflaLfismhoj; pgy, 53(6), 1552-1561. Mikulincer, M. (1988). The relationship of probability of success and performance following unsolvable problems. Reactance and helplessness effects W 12(2) Miller, A. (1981). h h ' h d. New York: Basic Books. Miller. J. B. (1976). MW. Boston: Beacon Press. 112 Miller, J. B. (1982). Women and Power. or P r (Wellesley College). No. 8201. Miller, J. B. (1983). The construction of anger of women and men. W (Wellesley College)- Miller, J. B. (1984). The development of women’s sense of self. Work in Prograss (Wellesley College). Muehlenhard, C. L. Women’s assertion and the feminine sex-role stereotype. In Spegial prpplgms jp ljyjpg. Unpublished manuscript. Nadelson, C. C., Notman, M. T., Miller, J. B., & Zilbach, J. (1982). Aggression in women: Conceptual issues and clinical implications. In M. T. Notman & C. Nadelson (Eds.), Tha woman patiant (Vol. 3, pp. 11-28). New York: Plenum Press. Novaco, R. (1977). Stress inoculation: A cognitive therapy for anger and its application in a case of depression. Journal pf Consulting and Clinital Esychology, 55(4), 600-608. Drlofsky, J. L., & Stake, J. E. (1981, Winter). Psychological masculinity and femininity relationship to striving and self concept, in the achievement and interpersonal domains. Psyphol- ogy of Women Quarterly, 1(6), 218-233. Perry, 0. G., Perry, L. G., & Rasmussen, P. (1986). Cognitive social learning mediators of aggression. Child Development, 51, 700-711. Ramanaiah, N. V., & Martin, H. J. (1984). Convergent and discrimi- nant validity of selected masculinity and femininity scales. Sex Roles, 19(7/8), 493-504. Riley, W. T., & Treibe, F. A. (1989, May). The validity of multi- dimensional self-report of anger and hostility measures. qur- gal of Clinical Esycholpgy, 55(3), 397-404. Robinson-Awana, P., Kehle, T. J., & Jenson, W. R. (1986). But what about smart girls? Adolescent self-esteem and sex-role percep- tions as a function of academic achievement. qurnal pf Educa- tjonal Psychology, 15(3), 179-183. Roget’s jnterpatjpnal thgsaurus (4th ed.). (1977). New York: Thomas Y. Crowell. Rosenberg, M. (1965). So i d o n ' . New York: Princeton University Press. 113 Rosenkrantz, P., Bee, H., Vogel, S., & Broverman, D. M. (1968). Sex-role stereotypes and self-concepts in college students. mm of ClinicaLandLonsultjmmpgx. 3.2. 287-295. Rothbart, M. K., & Maccoby, E. E. (1966). Parents’ differential reactions to sons and daughters. qurpa1_pf_flgrsppa11ty_apg Sosial_£sxshologx. 1(3). 237-243- Rust, J. 0., & McCraw, A. (1984). Influence of masculinity- femininity on adolescent self-esteem and peer acceptance. W9 12(74): 359'366. Sanchez, V., & Lewensohn, P. N. (1980). Assertive behavior and depression. MW. 4.8. 119-120. Sears, R. R., Ran, L., & Alpert, R. (1965). lusptifisatipn and thilg_rgaripg. Stanford, CA: Stanford University Press. Seavey, C. A., Katz, P. A., & Zalk, S. P. (1975). Baby X: The effects of gender labels on adult responses to infants. 55; 39155, 1, 103-109. Silber, E., & Tippett, J. S. (1965). Self-esteem: Clinical assessment and measurement validation. Esytpplpgy_flgpprts, 15, 1011-1071. Simpson, H. M. & Craig, K. 0. (1967). Word associations to homogenic and neutral stimuli as a function of aggression. W 29. 351- 354 Sims, S. M. (1989, March). Violent reactions. 5ELE, p. 159. Smith, C., & Hoyd, B. (1978). Maternal behavior and perceived sex of infant: Revisited. thjld Dayalppmant, 52, 1263- 1265. Spence, J. T., Helmreich, R., & Strupp, J. (1975). Ratings of self and peers on sex-role attributes and their relation to self- esteem and conceptions of masculinity and femininity. qurpal pf Parspnaljty and Spsjal Psychology. 32(1), 20-39. Stiver, I. (1983). Work inhibitions in women. uprt_1p_£rpgrgss (Wellesley College). Surrey, J. L. (1984). Self-in-relation: A theory of women’s development. Work inJErpgrsss (Wellesley College). Symonds, A. (1974). The liberated woman: Healthy and neurotic. American qurpal pf Esysppanalysjs, 35, 177-183. 114 Symonds, A. (1985). Separation and loss: Significance for women. American qurpal pf Esyspoanalysjs, 55(1), 53-57. Symonds, M. (1976). Psychodynamics and aggression in women. rn o , 35, 195—203. Taylor, D. (1984). Concurrent validity of the Bem Sex-Role Inven- tory: A person-environment approach. 5gx_flplgs, 15(9/10), 713-723. Thompson. C. (1964). Wagon (ll- R- Green, Ed.). New York: Basic Books. Tippett, J., & Silber, E. (1965). Self-image stability: The problem of validation. Esytnplpgisa1_fiepprts, 11, 323-329. Weissman, M., & Klerman, G. L. (1987). Gender and depression. In R. Formanek & A. Gurian (Eds.), Woman apu depressiop; A lifsspan perspgptjys. New York: Springer. Weissman, M. K., & Paykel, E. (1974). ng_ggprgssgg_meap. Chicago: University of Chicago Press. Westkott, M. (1986). h mi is le a K ren Hor e . New Haven, CT: Yale University Press. Whitely, B. (1983). Sex- role identification and self- esteem: A critical meta- analysis review. r W 44(4) 755- 778 Will, J. A., Self, P. A., a Datan, N. (1976). Maternal behavior and perceived sex of infant. m r th - psyshiatry, 55, 135-139. Williams, J. H. (1977). o f ° vi pjpspcial cpntsxt. New York: W. W. Norton.