UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Sex roles and career goals of university women Waterman, Diane C. 1977

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


831-UBC_1977_A8 W38.pdf [ 8.3MB ]
JSON: 831-1.0053803.json
JSON-LD: 831-1.0053803-ld.json
RDF/XML (Pretty): 831-1.0053803-rdf.xml
RDF/JSON: 831-1.0053803-rdf.json
Turtle: 831-1.0053803-turtle.txt
N-Triples: 831-1.0053803-rdf-ntriples.txt
Original Record: 831-1.0053803-source.json
Full Text

Full Text

SEX ROLES AND CAREER GOALS OF UNIVERSITY WOMEN by Diane C. Waterman B.A., University of British Columbia, 1970 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in the Department of COUNSELLING PSYCHOLOGY We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA AUGUST, 1977 © Diane C. Waterman, 1977 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia '2075 Wesbrook Place Vancouver, Canada V6T 1W5 i ABSTRACT Difficulties in predicting female occupational choices according to the theories developed for men have stimulated research into components of a model effective for women. The objective of the present study was to investigate the usefulness of self concept measures in determining the masculinity or femininity of occupations chosen by female undergraduates at a Canadian university. The ninety subjects who participated in the study were chosen for their enrollment in one of three fields defined by previous research as traditionally feminine or in one of three fields defined as traditionally masculine. The two aspects of self concept considered relevant to the sex stereotype of vocational field were sex role stereotyping and self-esteem. The Bern Sex Role Inventory used to assess the former variable, allowed subjects to endorse both masculine and feminine traits, thus providing an index of psychological androgyny as well as an index of stereotypic masculinity or femininity. Self-esteem was also considered from more than one perspective; that of personal and interpersonal functioning and of academic functioning. The results of the study revealed that the subjects' sex role orientation did not correspond to the sex stereotype associated with their fields of specialization. Similarly, statistically significant differences in self-esteem in either the personal or achievement areas emerged from the comparison among career groups. There was, however, a nonsignificant tendency for women in Science to have higher self-esteem in the achievement area than women in the remaining five groups. When self-esteem relationship in the achievement area between sex typing and self-esteem were examined, however, it was apparent that absolute levels of masculinity were significantly related i i ±o levels of self-esteem in both the areas being measured. The effects on self-esteem of femininity and androgyny were not statistically significant. In the case of the femininity scores, there were certain deviations from the normative data for the Bern Sex Role Inventory. Further consideration of these findings in terms of the instruments employed led to a factor analysis of the Bern Sex Role Inventory. Four factors were obtained. They are identified as scales measuring the constructs Dominance, Independence, Nurturance, and Passivity. Utilizing these newly designed scales, an analysis was made of the previous statistically non-significant relationships. While sex of field could s t i l l not be predicted according to scores on these factors, predictions regarding levels of self-esteem were refined by use of factor scores. By thus isolating the independent effects of the two feminine factors--Nurturance and Passivity--upon Inter/Personal self-esteem, the source of the low correlations for femininity was identified. The positive effects of the Nurturance factor were counteracted by the negative effects of the Passivity factor, resulting in a low correlation with self-esteem and with Social Desirability for the original Femininity scale. The two aspects of Masculinity which were defined by the factor analysis also bore differing, although not opposite relationships to the self-esteem criteria. Qualities denoting Independence were of most importance in explaining levels of self-esteem in the personal and interpersonal areas, while traits suggestive of Dominance were most relevant to self-esteem in the achievement area. The effectiveness of sex role stereotyping in predicting the sex stereotype of women's career choices was therefore not confirmed by the results of the present study. The relevance of self-esteem as a variable which moderates the predictive ability of other determinants of career choice was similarly unconfirmed. A number of reasons were postulated as to why the i i i results differ from previous studies on patterns of career choice among women. Recommendations were made for further research into the measurement of psychological androgyny and its significance to women's career aspirations. iv TABLE OF CONTENTS PAGE ABSTRACT i LIST OF TABLES vi LIST OF FIGURES v i i i ACKNOWLEDGEMENTS ix CHAPTER I - RELATED LITERATURE 1 Introduction 1 Limitations in Vocational Theories 2 Predictions Based on Abilities 4 Predictions Based on Interests 5 Predictions Based on Personality 5 Abilities 7 Motivation 8 Socialization of Women 12 Self Concept and Vocational Choice 15 Self-Esteem 17 Self-Esteem and Achievement 19 Self-Esteem and Femininity 20 Femininity and Achievement 23 Relevance of the Present Study 29 Definition of Terms 31 CHAPTER II - HYPOTHESES 34 Problem 34 Hypotheses 35 Research Assumptions 40 CHAPTER III - METHODOLOGY 41 Design 41 Sample 42 Measures 45 The Bern Sex Role Inventory 45 The Tennessee Self Concept Scale 47 The Self Concept of Ability Scale 49 Data Analysis 52 Summary 56 V PAGE CHAPTER IV - RESULTS 58 Response to Hypotheses 1, 2, and 3 68 Response to Hypothesis 4 61 Response to Hypothesis 5 65 CHAPTHER V - POST ANALYSES 66 Methodology 67 The Androgyny Score 67 Dimensions of the BSRI 68 Rescoring of the BSRI .. 69 Second Phase of Hypothesis Testing .... 70 Results 72 The Androgyny Score 72 Dimensions of the BSRI 76 Phase II of Hypothesis Testing 89 CHAPTER VI - DISCUSSION 98 Congruence between Self Concept and Occupational Stereotype 99 Sex Role Stereotype and Career Choice 100 Sampling Factors 100 Instrumentation 101 Variations Within Career Groups 103 Self Concept of Ability and Masculinity/Femininity of Career Choice 105 Inter/Personal Self-Esteem and Masculinity/Femininity of Career Choice 107 Relationships Among the Variables 107 Within-Group Differences 107 Relationships Between Sex-Role Stereotyping and Self-Esteem 110 Androgyny and Self-Esteem 110 Masculinity & Femininity Scales 113 Feminine Factors of the BSRI & Self-Esteem 115 Masculine Factors of the BSRI & Self-Esteem 118 Recommendations for Future Research 120 Additional Recommendations 122 Summary 126 REFERENCES 128 APPENDICES 134 A Measures Employed 135 B Analysis of Group Differences in Mean Self Concept of Ability ..... 149 C Tables Corresponding to "Additional Recommendations For Further Study 151 D Correlations Among Items of BSRI 161 LIST OF TABLES TABLE TITLE PAGE I MEANS AND STANDARD DEVIATIONS FOR CAREER GROUPS ON SELF CONCEPT VARIABLES 59 II UNIVARIATE AND MULTIVARIATE ANALYSIS OF VARIANCE ON SELF CONCEPT VARIABLES FOR CAREER GROUPS 60 III INTERCORRELATION MATRIX FOR COMBINED GROUPS 62 IV STEPWISE MULTIPLE REGRESSION ANALYSES UPON SELF-ESTEEM 64 V PERCENTAGE OF ANDROGYNOUS SUBJECTS IN CATEGORIES FORMED ACCORDING TO "MEDIAN SPLIT" PROCEDURE 74 VI MEANS, STANDARD DEVIATIONS, AND F VALUES FOR ANDROGYNOUS CATEGORIES ON SELF CONCEPT VARIABLES 75 VII IDENTIFICATION OF BSRI ITEMS SUBJECTED TO FACTOR ANALYSIS 77 VIII PRINCIPAL COMPONENTS SOLUTION: TRANSFORMED FACTORS 79 IX FACTOR PATTERN FOR FACTOR ANALYSIS OF BSRI ITEMS: PRINCIPAL COMPONENTS - 4 & 5 FACTORS 82 X FACTOR PATTERN FOR FACTOR ANALYSIS OF BSRI ITEMS: IMAGE ANALYSIS - 4 & 5 FACTORS 83 XI ITEM STATISTICS FOR FACTORS OF THE BSRI 86 XII INTERCORRELATIONS AMONG THE BSRI FACTORS 87 XIII INTERNAL CONSISTENCY STATISTICS FOR FACTORS OF THE BSRI 88 XIV MEANS AND STANDARD DEVIATIONS ON BSRI FACTORS AND SELF ESTEEM VARIABLES FOR SIX CAREER GROUPS 90 XV UNIVARIATE AND MULTIVARIATE ANALYSIS OF VARIANCE FOR SIX CAREER GROUPS ON BSRI FACTORS AND SELF-ESTEEM VARIABLES 91 XVI INTERCORRELATION MATRIX FOR COMBINED GROUPS ON FACTORED BSRI AND SELF-ESTEEM VARIABLES 93 XVII STEPWISE MULTIPLE REGRESSION ANALYSIS UPON SELF CONCEPT OF ABILITY - BSRI AS FOUR FACTORS 95 VI 1 LIST OF TABLES Cont'd TABLE TITLE PAGE XVIII ANALYSIS OF GROUP DIFFERENCES IN MEAN SELF CONCEPT OF ABILITY 150 XIX MEANS AND STANDARD DEVIATIONS FOR LOW SELF-ESTEEM THIRD OF EACH OF THE SIX CAREER GROUPS ON REMAINING VARIABLES 152 XX INTERCORRELATIONS AMONG THREE COMPONENTS OF TOTAL POSITIVE SCORE AND ORIGINAL BSRI SCORES 153 XXI: INTERCORRELATIONS AMONG THREE COMPONENTS OF TOTAL POSITIVE SCORE AND FOUR FACTORS OF BSRI 154 XXII MEANS & STANDARD DEVIATIONS OF CAREER GROUPS ON "IDENTITY" COMPONENT OF SELF-ESTEEM SCORE AND BSRI SCALES 155 XXIII UNIVARIATE & MULTIVARIATE ANALYSIS OF VARIANCE OF CAREER GROUPS ON "IDENTITY" COMPONENT OF SELF-ESTEEM SCORE AND BSRI SCALES 156 XXIV MEANS & STANDARD DEVIATIONS OF CAREER GROUPS ON "BEHAVIOR" COMPONENT OF SELF-ESTEEM SCORE AND BSRI SCALES 157 XXV UNIVARIATE & MULTIVARIATE ANALYSIS OF VARIANCE OF CAREER GROUPS ON "BEHAVIOR" COMPONENT OF SELF-ESTEEM SCORE AND BSRI SCALES 158 XXVI MEANS & STANDARD DEVIATIONS OF CAREER GROUPS ON "SELF ACCEPTANCE" COMPONENT OF SELF-ESTEEM SCORE AND TWO BSRI FACTORS 159 XXVII UNIVARIATE & MULTIVARIATE ANALYSIS OF VARIANCE OF CAREER GROUPS ON "SELF ACCEPTANCE" COMPONENT OF SELF-ESTEEM SCORE AND TWO BSRI FACTORS 160 XXVIII CORRELATIONS AMONG ITEMS OF THE BSRI 162 vi i i LIST OF FIGURES FIGURE TITLE PAGE 1 SUMMARY OF MEASURES EMPLOYED 51 2 DISTRIBUTION OF SCORES ON MASCULINITY AND FEMININITY SCALES 73 3 PRINCIPAL COMPONENTS ANALYSIS SCREE TEST 78 4 IMAGE ANALYSIS SCREE TEST 81 ACKNOWLEDGEMENTS I wish to thank those who have, in many ways, contributed to the process and to the completion of this work: To Dr. Robert Tolsma, for his enthusiasm over research which stimulated rne in my area of interest. His criticisms and suggestions in the preparation of this manuscript were also appreciated. To Dr. Todd Rogers, whose continued help, direction and advice during the data analysis encouraged me to j experiment. To Dr. Sharon Kahn for her special interest in the area of counselling women, and women's achievement. To Lilly Jaffe, whose help in resolving the conflicts between "femininity" and "success" which arose during the course of this work is warmly acknowledged. To the participating students and their professors who gave their time and energy, and who co-operated in the gathering of data. To my supportive friends. 1 CHAPTER I RELATED LITERATURE Introduction The impact of the Women's Liberation Movement can be seen in the degree of attention that is being focussed on women's roles, their status, and the consequences of their changing roles in society. Heightened awareness of the social and economic barriers to their growth has led feminists to seek changes in social and political structures. Increasingly dissatisfied with the low status, poorly paid jobs open to them, women have directed attention to discriminatory hiring practices and unequal wage structures, with noticeable results. There has been progress in establishing day care centers, legislating against overt discrimination and providing support for women in accommodating job and family commitments. In addition to the external constraints upon women's work - the overt discrimination in the work place and the problems that arise in combining work with the family obligations that have traditionally fallen to women -there are internal psychological constraints which cause women to limit their options. Recent empirical evidence regarding women's career choices indicates that social and political moves alone will not bring about equality of the sexes. For example: A study (Karman, 1973) involving over 1600 college women all over the United States showed that even before experiencing the discrimination of the job market, and prior to coping with adult family responsibilities, the majority of women reduce their own aspirations. A full 50% of the entire sample planned to become teachers, and of the remainder 2 only 109 women aspired to other than the traditionally feminine roles of homemaker, social worker, clerical or other service oriented occupations. It is not the intrinsic worth of those conventional roles which is in question, but the fact that the potential of so many women with a wide range of unique identities and talents should be confirmed to such a narrow range of roles. In themselves, roles can be useful tools in interpreting experience. People rely on generally accepted forms of behaviour and patterns of relating to one another to understand the environment. It is when those patterns lose their flexibility and their purpose that human growth becomes restricted. Another negative effect of these limitations for women is that with the present trend for more women to join the labour force, their continued aspirations toward careers in the traditionally feminine fields is likely to lead them to unemployment as those fields become overrepresented. Of the three types of restraints upon women's employment mentioned; the political and economic, the social and familial, and the internal psychological, the present study focusses on the third; based on the assumption that a clearer understanding of the dynamics involved in career choice at present will aid in altering outcomes. In this research, some of the variables believed to be related to women's occupational choice are explored. Limitations in Vocational Theories Various theories of vocational choice have been applied to women at different stages of educational and occupational growth. Results of the attempts to predict women's choices or to counsel them vocationally led Osipow (1968) to conclude that "special problems exist for them as opposed 3 to men and that most of the masculine based tests and theories fail to provide a useful vehicle; for the understanding of the career development of women" (p.247). The information considered relevant to an individual's career choice varies with the vocational theory employed, and with the personality theory used to interpret the dynamics of choice. Interests, abilities, and personality traits are generally treated, by theorists, as important determinants of career choice; but the process by which they affect career decisions is described differently by psychoanalytic than by need theorists, for instance; and differently by trait-factor than by self concept theorists. Models of vocational development differ, too, in their conception of the etiology of interests, the acquisition of abilities, and the formation of personality. For example, Ann Roe (1956) is a theorist who attributes major importance to early family background, while Super (1963) sees several life stages as influencing career choice and satisfaction. The process of career choice is explained, by the trait-factor theory, as an act of matching personality traits with job requirements; while in terms of self theory i t is seen as the implementation of one's concept of himself. Psychoanalytic theory emphasizes identification with role models, the organization of ego defenses, and the sublimation of needs, while need theory sees occupations primarily in terms of their potential for need satisfaction. Some evidence supporting Osipow's assertion that these theories deal inadequately with women's career choices will follow. It introduces a review of research studies selected for their value in exploring reasons for these theoretical and practical limitations. 4 Predictions Based on Abilities In a survey of college women about to graduate (Baird, 1973) i t was learned that fewer women than men planned to pursue graduate studies, despite their superior grades. The phenomenon of low aspirations compared with actual abilities was again observed in a sample of 1188 first year women at the University of Wisconsin. Harmon (1971) compared women's present aspirations with careers they had considered earlier in their lives, and found that later preferences were restricted largely to typical female jobs, thus excluding the business and science worlds. A self perceived lack of talen or ability was the most common reason given for reducing their expectations. Similar findings are reported by Helen Olive (1973) in studying the vocational preferences of adolescents. The girls did not aspire to the most prestigious occupations, such as medicine, upper levels of administration, college presidents, etc., even though their intellectual abilities were not significantly different from the boys'. They tended to choose social work, teaching, or secretarial work, even in fantasy. Similar trends were found among elementary school children (O'Hara, 1962) with girls choosing the prescribed feminine roles and boys choosing a wide variety of occupations, including many that were purely fantasy roles. That intelligence scores are not a reliable predictor of future success for women was shown in an interesting longitudinal study by Terman and Oden (1947). They followed gtft'ed children of both sexes over a period of ten years to assess the extent to which levels of occupational achievement related to the measured I.Q. of their school years. A substantial relationship between the two was found for boys, but for girls, no relationship could be discovered. A high I.Q. for girls was equally likely to lead to high as to low achievement in later li f e . 5 Predictions Based on Interests Another major tool supplied by the trait-factor theorists, and used by many others, is the interest inventory. Its purpose is to assess the individual's motivation toward various activities. Interests do tend to motivate boys' choices of where to direct their energies. In school, for example, boys perform well in the subjects that interest them and poorly in those that bore them. The same cannot be said of girls. Their performance showed a uniformity that applied to all subjects, regardless of the comparative interest (Coleman, 1961) subjects held for them. A difference such as this illuminates, to a degree, Ann Roe's (1956) statement that interests and abilities are not linked in the same way for the two sexes. Her research showed that for girls, interests related more closely to attitudes than to abilities. The fact that interests have less bearing on school achievement for girls may be a factor in determining the patterns of their subsequent occupational achievement. A survey of Canadian youth (Breton & McDonald, 1967) showed that "while only 18% of fourth year high school girls preferred sales and clerical careers, 32% seemed to assume that they would eventually work in these fields even though more challenging areas interested them." Interest alone does not seem to be the motivating force for girls that it is for boys. It appears to have less influence over their choices of activities and goals, and less effect on the development of specialized abilities; while boys' interests are closely aligned with their realistic expectations and with the talents they actually cultivate. Predictions Based on Personality The applicability of Holland's personality theory to women's career designs has also been tested. According to his model, an individual's 6 perception of his needs leads him to choose an occupational environment likely to offer him satisfaction. He defines six typical ways of organizing needs, according to the individual's preferred mode of expressing masculinity/or femininity, aggression, needs for achievement; social and emotional needs. Each of the six "personal orientations" has a corresponding "occupational environment" in which the individual with that pattern of needs is likely to find satisfaction. D. W. Harvey and R. Whinfield (1973) investigated the validity of Holland's six personal styles for describing the needs of mature women. Scales from other measures of interest, values and needs thought to correspond with those described by each of Holland's types were selected, and the logical correspondences tested statistically. Results showed that only three of the six "types" were meaningful descriptors of the women's personality styles and vocational interests given by the criterion measures. The needs, interests and attributes Holland associates with the Realistic, Theoretical and Social Orientations did not characterize women with those orientations as measured by Holland's Vocational Preference Inventory. The preceding evidence suggests there are some differences between men and women in the organization of variables used to predict career choice. These different patterns could be attributable to either one or to both of two factors: 1) important differences on certain of the determinants themselves; differences, for instance, in abilities or intelligence, that go undetected by global measures, or 2) the "motivation" thought to translate traits and abilities into career aspirations. The possibility of sex differences in these determinants has received attention in research reviewed in the following section. 7 Abi1i ties In a comprehensive review of the research on the abilities of school children, Maccoby (Chap.3, 1974) concluded that there are no consistent sex differences in overall abilities. The next important question, which involved possible differences in specific abilities, was also summarized by Maccoby. The highlights of her conclusions are presented below. The greater achievement by boys in math and science is not attributable to a greater "aptitude" for those subjects. It may be due, rather, to their greater interest in scientific subjects, or to their superiority in certain relevant skills. Analytic ability, one such s k i l l , thought to be better developed in boys, proved not to differentiate the sexes in the research reviewed by Maccoby. However, the spatial relations skills often involved in analytic tasks of a scientific nature, do seem to be lacking in girls. They tend to perform better in tasks requiring verbal abilities. Beginning in adolescence, they tend to perform better in tasks requiring verbal abilities than do boys. At this time also, their verbal abilities exceed their spatial competencies. Whether differences in spatial abilities are sufficiently great to account for the vast differences in scientific achievement.'between the sexes remains to be seen. There is evidence to suggest that the differences in spatial ability are a product of differential training: Performance in spatial tasks has been shown to be enhanced by training (Kato, 1965; cited in Maccoby, 1974). Spatial abilities have been related to traits such as assertiveness and autonomy, which are encouraged in boys and discouraged in girls. (Maccoby, 1966). Poor spatial abilities, on the other hand, are found in conjunction with high levels of passivity and dependence—traits fostered by the socializa-tion of girls. Overall intellectual abilities too, have been linked to personality 8 characteristics, but the patterns are somewhat different for the two sexes. Girls described as "bold and assertive" showed greater general intellectual abilities and interests than other girls, but boys who were less active and less "masculine" were superior intellectually to other boys (Maccoby, Chap.3, 1974). Thus a situation similar to that previously discussed with interest and ability patterns may be the case with personality trait and ability patterns. Specifically, the inaccuracy of vocational predictions for girls that are based on the personality/ability patterns of boys may be due to the fact that those patterns are different for girls. As a result, a prediction based on a girl's personality would not necessarily coincide with one based on her abilities, and vice versa. Motivation Another possible source of inaccuracy in predicting women's career choice according to the criteria used for men is in the motivation linking interests, abilities and traits to actual achievement. It has been said that males "have a greater need for achievement and are more oriented to achievement for its own sake" (Maccoby, p. 134, Chapter 4, 1974). Females, on the other hand, are said to be motivated primarily by a need for affiliation (Kimball, 1974). These general statements refer to two kinds of achievement: the level of performance on a given task, which is the sort normally used in research on "need for achievement", and cumulative achievement: the kind resulting in career success, a work of art or an invention. (Atkinson, 1974). Examining the evidence on women's motivation toward achieving an immediate goal may lead to explanations of their comparatively reduced "cumulative" achievement; that is, success as the world recognizes and rewards i t . The summary of studies on children's "need for achievement" reviewed by 9 Maccoby, Chapter 4, 1974) contradicts the widespread belief that boys have a stronger achievement drive. They appear only to be more stimulated by competitive situations than girls, but even this factor does not raise their level of motivation above that of the females on specific tasks. The second assertion, that boys are more oriented toward tasks and girls toward people, also found no consistent support. Girls did not perform better than boys on "person oriented tasks", nor were they more influenced by the presence of others. In fact the reverse was true -- boys' performance improved under observation by their peers, and girls' performance was unaffected. Thus the idea that girls' achievement derives primarily from a need for affiliation or to please others is of questionable validity. The lack of evidence for a sex difference in "need for achievement" leaves the reality of differences in later achievement unexplained in terms of motivation. Maccoby suggests that since the decline in girls' achievement begins as they reach maturity, their underachievement in adulthood may be linked to the adult female sex role (Maccoby, Chapter 4, 1974). Research on the "need for achievement" among adult women tends to support the effect of the feminine sex role on reduced achievement. Mat.ina Horner's (1968) approach to the study of achievement motivation revealed interesting results. . She studied the subjective responses of college students to success stories about someone of their own sex. While 90% of the men were encouraged and inspired by the success cues, 65% of the women attributed negative consequences,such as social rejection and loss of femininity,to the successful woman. That equally negative responses were not found by Kimball (1974) in girls of age 13 would imply that the "fear of success" as a threat to femininity does not emerge until adulthood, when the conflicts arising out of the role are more acutely felt. 10 Extending her research from fantasy cues to actual achievement situations, Horner found her subjects' behaviour consistent with previously mentioned sex differences; i.e., that women, more than men, underachieved in competitive situations. The reverse was true for men, whose performance improved in a competitive situation. An interesting result was that a minority of women performed more similarly to the men than to other women and their responses to the "success" stories evidenced significantly less "fear of success" than those of the other women. A further comparison between these two groups of women on their respective career aspirations confirmed their differentness. Regardless of their high grade point averages, the women high in "fear of success" tended towards the traditionally feminine fields of housewife, nurse and teacher. The women whose stories had revealed l i t t l e "fear of success" were planning careers in demanding scientific fields, even though their grades were not as high as their goals would require (Horner, 1968). A later study on the effects of different levels of achievement motivation among women was conducted by Sara Kriger in 1972. In a study of college women, their career aspirations and achievement motivation, a direct correlation was found between the "femininity" of aspirations and strength of achievement needs. A high "need for achievement" distinguished those women planning careers in male-dominated fields such as medicine, chemistry, and architecture from those entering teaching, nursing, social work and other fields where women form the majority of the work force. A paper by S. Thurber (1976) describing his research into women's achievement motivation develops further the pattern discovered by Kriger. He hypothesized that the relatively poor predictive ability of achievement motivation scales for women's choices was due to the use of academic performance as the sole criterion for achievement. He thought for women 11 with a traditional attitude toward femininity, to whom achievement in the academic area might seem too competitive, a motivation scale might more accurately predict achievement in social or "affiliative" areas. Two groups of women were thus identified: those with conventional attitudes toward sex roles according to the "Fand Inventory," and those with contemporary attitudes. For the "conventional" women, the "achievement motivation" scale successfully predicted achievement in social organizations, time spent in social interaction, etc., but for the women with "contemporary" attitudes toward sex roles, the motivation to achieve was expressed in academic success. The author concluded that the predictive validity of an achievement motivation scale depends upon congruence between the type of achievement in question and the subject's sex role orientation. Should results such as Thurber's be confirmed, i t would suggest that women's comparative lack of "success" in terms of paid employment is not due to a general lack of motivation to achieve, but in the directions their motivation has taken. It would suggest also, that the comparative inadequacy of theories of career choice based on male patterns in explaining dynamics of female career choice and achievement may be due to the vast differences among women -- differences that may equal in complexity those between men and women. For this reason, the remainder of the literature to be reviewed deals only with women; with comparisons between the sexes mentioned only to clarify the differences among women. The kinds of differences in question are those relevant to the basic differences in career paths which have been described as "traditionally feminine" or "traditionally masculine". The underlying question, then, is what characteristics differentiate the women who have felt free to choose unconventional career paths from the majority who have not. What are the factors leading to such differences in direction, 12 and what are the attitudes toward sex roles that accompany them? This involves a look at the socialization processes women in these two categories may undergo. Socialization of Women A number of investigators have recently directed their attention to the differences in backgrounds between two groups of women: those who have either expressed interest in, or achieved substantial success, in careers traditionally chosen by men, and those who have preferred the traditionally feminine roles. The relevance of family dynamics is stressed in a study by Leland and Lozoff (1969). Throughout their four college years, a number of women who differed in the types of success they sought were interviewed regarding their family background and their present conflicts. Those choosing careers were termed "autonomous", and those choosing homemaking above a career were termed "least autonomous". The "autonomous" group showed distinctly different levels of conflict among its members, and was accordingly divided. Those described as most autonomous and least conflicted were involved in the arts, psychology and teaching. Their family backgrounds tended to involve close relationships with fathers who were intelligent, ambitious, and who both encouraged and respected their daughters' unique abilities. The "autonomous conflicted" group were similarly preoccupied with their fathers rather than their mothers, but their fathers' personalities were cold, aloof and perfectionistic. They did stimulate competency in their daughters, but at the expense of not having satisfying personal relationships. These women had planned careers in science and mathematics, but most became discouraged with those fields before graduation. The "least autonomous" focussed their energies on the roles of wife and mother, and chose areas of study that would allow them to develop skills related to domestic l i f e . Their fathers were 13 successful, usually in business; were dominant in the family, and supported rigid definitions of sex roles. The dichotomized attitude toward appropriate activities prevalent in the home allowed women only limited development as persons. Interested in antecedents to female success in the business world, Margaret Hennig (1971, cited in Hennig, 1974) studied the lives of 25 women executives. Her findings attributed much of their positive achievement orientation to particularly strong, secure, healthy family backgrounds. The women saw both parents as complete persons, who supported and reinforced their exploration and mastery of skills related to both sex roles. Though they perceived relationships with both parents as close, they considered their rapport with their fathers as warmer and more sharing than most father-daughter relationships. The importance of parental role models is emphasized in a number of studies on female career commitment. Baruch (.1973) and Lozoff (1974) each found that autonomy in young women was associated with father identification. Investigating vocational trends among both sexes, Super (1968) saw that women who identified with their fathers typically had strong motivation to pursue a career. However, in families where "male" and "career" are not so inextricably linked, and the mother works outside the home, she may have a similar influence on her daughter. Using the term "career salience" to describe long term commitment to a career, Almquist & Angrist (1971) found that career salient women more often had working mothers than did non-career salient women. In light of the information that the most influential role models are those who are liked, who have status, influence, and who appear satisfied, and considering that daughters were more likely to prefer the mother as a model 14 If she were working (Baruch, 1974), Almquist & Angrist's finding supports the hypothesis that maternal modeling was the vehicle through which "career salience" was transmitted. Working mothers, reasoned Almquist, expand their daughters' view of appropriate female roles by demonstrating the compatibility of marriage and career. In predicting what seems to be a primary decision for women: whether to include work as a significant part of their lives, parental modeling seems to be a critical factor. The secondary decision, of which particular occupation to choose, owes its nature more directly to the influence of college professors and occupational role models, according to Almquist & Angrist's information. Thus the absence of female models in high positions in business and public l i f e , and in scientific fields, serves to depress young women's expectations where they might be enhanced. (Royal Commission on Status of Women, p. 180, 1967). The work of Sara Kriger (1972) placed similar importance on parental relationships to predict the initial career vs. homemaker' decision. Use of the model developed by Ann Roe led her to postulate that women from restrictive, over-protective families would have learned to depend on their relationships to other people for maintaining their own sense of self, to a significantly greater degree than would women from more permissive families. Her hypothesis was that homemaking would express the "person" orientation. She further postulated, in line with Roe's theory, that preference for a career over homemaking would represent, for women, a level of risk-taking and an orientation toward "non-persons" that would have required a permissive, accepting home environment in order to develop. Empirical results confirmed those hypotheses; with aspiring homemakers having perceived their parents as significantly more restrictive than the career oriented women had perceived theirs to be during their childhood. 15 In a later study by Klemmack and Edwards (1973), family status, rather than family relationships, provided additional information concerning women's aspirations. Where the father's occupational status was high, daughters' , career choices tended to be more stereotypically masculine in character; and where his status was low, more feminine. Occupations had previously been rated by a panel of judges on the degree of feminine sex-typing they represented. Of more importance to future aspirations than even these background variables were the present life situations and concerns of the young women, which mediated the relationships between background variables and career plans. Inclusion of these variables revealed that desires for a large family, current serious involvement in a heterosexual relationship, and younger ideal age for marriage characterized women aspiring to careers that were "primarily social-emotional, nurturant, and person-centered, and congruent with traditional sex-role learning." (Klemmack & Edwards, 1973, p.511). In noting the distinctly greater predictive power of these factors over family background, the authors emphasize how unique to women are these dynamics. The applicability of the models of career choice developed for men, suggest Klemmack & Edwards, may be severely limited by this factor. Self Concept and Vocational Choice The issue of traditional vs. non-traditional sex role learning has been invoked in a number of the preceding studies which sought to identify the determinants of female career choices. A degree of congruence has emerged between interests, abilities, achievement needs and career choice, on the one hand, and widely recognized "sex role stereotypes" on the other. The effect of role stereotyping has, however, been inferred rather than directly measured in the research cited to this point. The assessment of attitudes toward women's roles made by Thurber ( 1 972 and 1 976) and the influence 16 of present plans for relationships detected by Klemmack and Edwards (1973, p.15) suggest that a shift in focus from background variables to the woman's concept of herself in terms of sex-appropriate behaviour, may clarify the reasons underlying women's decisions. The shift in focus from antecedent, environmental variables to the present, internal influences upon career decisions this discussion will now take, has, as its rationale, the Self Concept Theory of vocational choice. Major proponent of the theory, Donald Super, conceives of career choice as a process of matching one's self concept - his self-perceived interests, needs, preferences, and general estimate of his abilities, - with an occupational role concept. The most satisfying choice is that which is most congruent with the self concept, allowing the individual's "self" the maximum expression in his occupation. Background factors, parent identification etc., receive attention, but only in terms of their formative influence on the self concept. Social and environmental factors, too, are important only insofar as they have become internalized. Super stresses the developmental process of acquiring a self concept, translating it into occupational terms, testing i t in realistic situations; implementing i t in one's chosen occupation. It is his contention, and that of vocational guidance in general, that one's development in this area "can be aided and guided by the provision of adequate opportunities for the utilization of aptitudes and for the development of interests and personality traits." (Super, 1963, cited in Zytowski, 1968, p.126). The implications for socialization into femaleness are that i f the opportunities for development are limited by role constraints; i f skills are not validated because they are seen as inappropriate; important abilities, and the concept of self that would accompany their acquisition, will either 17 not be developed, or, i f developed, will not be integrated into the self concept. Choices made in accordance with a restricted self image are not an accurate representation of the person's potential, and tend to maintain restrictions based on past development. Self-Esteem Self-esteem, according to Carl Rogers, is a critical factor in the development of an adequate self image. It is his contention that "low self-esteem leads to distortion and constriction of perceptions of which the self is a part," and "that it influences his perception of his ability to perform different tasks and roles." (Healy, C, Bailey, M., and Anderson, E., p.69, 1973). The implications of Rogers' observations for vocational theory have been explored by Abraham Korman in a series of studies. Reasoning that choice of an occupational role congruent with one's self concept, i.e., his self-perceived traits, abilities, and needs, is contingent upon feelings of adequacy and self worth, he compared the choices made by individuals high and low in those feelings (Korman, 1967(a)). The career choices made by men high in self-esteem were more likely to require the same abilities the men attributed to themselves than were the choices of men low in self-esteem. Low self esteem subjects were as likely to choose a work situation for which they perceived themselves lacking in ability as one in which they felt competent. People low in self-esteem are likely not only to make choices unsuited to their abilities, but to remain in a situation where their personal needs are not being satisfied. (Korman, 1967(b)). The phenomenon applied both to social and to vocational situations. The discrepancies between the needs for which subjects sought satisfaction and the satisfaction they actually obtained 18 within their chosen environments were significantly greater for the low self-esteem group. For high self-esteem subjects, greater congruence existed between the needs they expressed, and those they saw an opportunity to satisfy in the area of their choice. They also experienced greater satisfaction for expressed needs. In this way, suggests Korman, the person low in self-esteem perpetuates his self image as an inadequate, non-need-satisfying person. The explanation offered by Korman is that low self-esteem leads one to see situations more in terms of their general social desirability than in terms of his own desires. He allows others to influence his decisions because of his lack of confidence in his own judgments. The work of R. C. Smith (1972) follows directly from that of Korman and supports his thesis. Selecting an environment with the potential to satisfy a specific need, that for "nurturance". His study compared the satisfaction experienced by volunteer nurses differing in levels of self-esteem. Results showed that nurses with a high "need for nurturance" and high self-esteem found a significantly greater degree of satisfaction for that need than did nurses with low self-esteem and the same expressed need. Smith concluded that vocational predictions according to the "need reduction" theory are reliable only in the presence of high self-esteem. That is to say, individuals will choose occupational environments likely to satisfy salient needs, and experience satisfaction for those needs, i f their perceptions of themselves are not distorted due to low self-esteem. Self-esteem seems to mediate the relationships between determinants such as abilities and needs, and the occupational choices they seek to predict. 19 Self-Esteem and Achievement Self-esteem has an important role to play in achievement behaviour as well as in the choice of an occupation. Persons with low self-esteem have been found to lack persistence in completing tasks and to avoid taking risks with new tasks by comparison with high self-esteem persons. (Coopersmith, 1967). They tend to set lower ideals for themselves and to underestimate their performance on completed tasks more than people who believe themselves to be capable. High self-esteem has been associated with greater creativity, independence, and activity on both behavioural and written tests. People with high self-esteem have greater conviction in their own judgments and are able to ignore pressures to conform to others' expectations to a greater degree than those low in self-esteem. Individuals whose measured level of self-esteem is low are quieter, more passive, and comply more readily with others' opinions. In addition to an overall evaluation of oneself as worthy or unworthy, there is a tendency for self-esteem to vary depending on the area of one's life in question (Coopersmith, 1967). This "differential" nature attributed to self-esteem means that one's self-esteem will tend to be highest in those areas most important to him, and that he will be most affected by threats to his self-esteem in those areas. Thus a woman with a traditionally feminine identity may esteem herself highly in relationships to other people, and for her skills at managing and nurturing relationships, but her sense of self-worth in other areas may be considerably lower. Academic or professional achievement may represent a greater source of self-esteem to less traditional women, as Hennig's (1974) article on female executives indicates. Within achievement areas, self-evaluations may also vary, according to a study involving intermediate school children. (Brookover, Patterson & Thomas, 1962). Depending on the subject area being considered, students' self 20 concept of their ability varied. It differed also from their overall Self Concept of Ability. Self concepts in terms of overall school achievement, were positively related to actual achievement when measured intelligence was statistically controlled. High-achieving students were found to have a significantly higher mean self concept of ability than low-achieving students of comparable intelligence. Confidence in his "achievement self" corresponded, interestingly, to the confidence a student perceived parents, teachers and peers to have in his ability to achieve. Future aspirations were another variable directly affected by the ability self concept; with students planning a college education showing greater confidence in their ability than those who were not. Thus low self-esteem has implications for achievement, and by virtue of the "circular relationship between confidence and learning," (Loeb, p.88, 1973) i.e., reluctance to take the risks necessary for new learning, low self-esteem contributes to the development of actual deficiencies as well as the imagined ones resulting from devaluation of oneself. Self-Esteem and Femininity Virginia Satir (1964) observes that abilities which are not validated and traits that are not accepted do not contribute to self-esteem. This being the case, one would expect the denial of aspects of the self associated with femininity, (to be specified in the following section, "Femininity and Achievement") particularly in relation to achievement, to have a depressing effect on self-esteem. Evidence for a general sex difference in self-esteem is not apparent, although research using subjects beyond college age is limited (Maccoby, 1974). A study by Lyell, (1973) does suggest that adulthood seems to bring about a decline in self-esteem for adolescent girls but an 21 .increase in self-esteem for boys. The author considers the effect a result of girls' anticipation of roles which are less prestigious, less socially rewarding than those boys foresee. It seems equally plausible that to the extent the young woman identifies with the feminine stereotype, she denies important strengths and assumes characteristics which are socially devalued; both of which processes undermine her self-esteem. Comparisons made on this variable among women who differ in degree of identification with the stereotype have lent support to the above hypothesis. Sandra Bern (1973) cites a number of studies linking high levels of femininity with "high anxiety, low self-esteem, and low social acceptance" (e.g. Cosentina & Heilbrun, 1964; Gall, 1969; Gray, 1957; Sears, 1970; Webb, 1963). Research involving school girls and their mothers (Baruch, 1974) showed that low self-esteem accompanies a typically feminine self-perception, while seeing oneself as competent and endorsing "masculine" traits is coupled with high self-esteem. Taking orientation toward a career as an index of unconventional behaviour for women, Birnbaum"s(l971, cited in Baruch, 1974) evidence that career oriented women in their thirties have higher self-esteem than do domestically oriented women of the same age support Baruch's findings. Choice of the traditional feminine mode, homemaking, is coupled with reduced self-esteem Birnbaum (1971, cited in Baruch, 1974). The relevance of parental modeling is again emphasized by Satir (1964), who attributes considerable importance to identification with the same sex parent and to the demonstration of sex roles by the parental relationship in the development of one's self-esteem as a sexual person. In that context she notes that "a girl's esteem about herself will suffer i f it is her mother who looks the most hurt" (p.53). Investigating this effect, Grace Baruch (1974) reports the research of Connell & Johnson (1970) and Gray (1959), both of whom found that seeing oneself as feminine and identifying primarily with the mother 22 were negatively related to adjustment and self-esteem. If, on the other hand, the mother represents a different model of femininity, and shows personal strengths with which her daughter may identify, she may have a positive effect on the girl's self-esteem. Baruch (1973) found that mothers with a career commitment had daughters who saw themselves as competent and had high self-esteem; and conversely, daughters of non-working mothers, though seeing themselves as competent, did not show high levels of self-esteem. Coopersmith (1967) notes that overall self-esteem among children of both sexes is enhanced by their mother's desire to work and by her long term employment. The literature on the development of self-esteem tends to validate the negative associations with femininity. Coopersmith (1967) stresses the importance of internalizing and accepting personal successes in order to build self-esteem. His research shows that individuals who are successful by objective standards may s t i l l consider themselves "inadequate". Evidently successes must be perceived as such by the individual, not just by others, in order to enhance his self-evaluation. The term "success" has four aspects relevant to self esteem. They are "power", "competency", "significance", and "virtue". (Coopersmith, p.38, 1967). By definition, the feminine stereotype limits the areas open to women to only two of the four: significance; the acceptance and affection of others, and virtue; the adherence to moral and ethical standards. In excluding from her repertoire those activities involving mastery over problems, decision-making, reasoning and planning ahead, the woman who adheres strictly to the orthodox view of femininity seriously restricts opportunities to build self-esteem in herself. The relationships among self-esteem, achievement, and career choice make a discussion of "femininity", and achievement worthy of attention. 23 Femininity and Achievement Some recent research has addressed the issue of how women's sex role self concepts affects their achievement and their career choice. First to be discussed are those studies involving overall feminine stereotyping and subsequently, some which focus on particular aspects of the stereotype. Just why "femininity" of self concept should militate against significant achievement is evident from the discrepancy between society's definition of that term and the traits associated with high achievers. Atkinson's work on achievement motivation has resulted in a profile of the type of person who tends to be strongly achievement oriented (Atkinson & Raynor, 1974). He is typically self reliant, willing to take risks in competitive situations, works independently, and is able to delay gratification in favour of future rewards. Mastery of new skills is worth the anxiety of entering unfamiliar situations, due to the anticipation of success. The typical feminine profile on the other hand, describes a person who is dependent, passive, non-competitive and illogical; in short, lacking in traits denoting competency and confidence. (Broverman, I., Vogel, S., Broverman, D., Clarkson, F., Rosenkrantz, P., 1972). Competency and confidence-are features of the masculine stereotype, which bears a much closer resemblance to that of the achieving personality than does the feminine stereotype. According to the investigation of sex role stereotypes prevalent in American society, the desirable aspects of femininity were gentleness, sensitivity to the feelings of others and tactfulness, along with a tendency to be religious, neat, quiet, interested in art and literature, and express tender feelings. The attributes were notably fewer in number than the socially desirable traits making up the masculine stereotype; with the result that the majority of typically feminine traits denoted the absence of positive masculine qualities. 24 A number of studies have attributed women's suppressed ambitions to the fear of losing their femininity, for example Horner's "fear of success." Horner (1968). Meredith Kimball (1974) proposes a reversal in that logic; that "in the minds of many women, being feminine causes, or creates the impossibility of success" (p.12). Positions requiring leadership abilities, decisiveness and personal risk taking, might make a woman with a typically feminine self concept feel inadequate to the stresses of such a role. Because of the disparity between the role requirements and her own identity, however, theory and research evidence suggest she would be unlikely to aspire to such a position. With respect to aspirations and identity, Cowan and Moore (1971) designed a study to test the hypothesis that women choose vocations congruent with their sex role self concepts. Contrasting female college students aspiring to traditionally male fields with those planning traditionally feminine lifestyles, they found that women planning to enter masculine career areas endorsed the masculine traits on the Broverman Sex Role Stereotype Questionnaire to a significantly greater degree than did the traditional group. Similar outcomes were obtained subsequently by Hawley (1972). Her approach to the same task was rather unique. Believing that a woman's perception of the behaviours significant men in her life deem appropriate for women plays an important role in her choice of career, the author designed an instrument to assess those perceptions. The degree of traditional femininity indicated was then compared with femininity of career choice. The women selected to represent a typically feminine career choice were teachers in training. Their sex role definitions were indeed significantly more dichotomized than those of women in Mathematics-Science 25 fields. That is to say, they were more prone to define activities in a sex linked way; considering men better qualified to make decisions in financial matters, for instance, and women to take responsibility in home and family decisions. Women in the "male" fields did not tend to make gender the basis of those decisions. They thought men approved of a wider range of activities for women, including competing with men in business and professional areas, and at the same time placed more importance on supportiveness in their relationships with men than did the teacher group. The conclusion drawn by Hawley was that the choice of a career in a male dominated field such as Science was the result of a broader and less typically feminine' concept of sex roles, which allowed women to choose activities generally termed masculine, without identity conflict. Further to the evidence that adherence to stereotypic notions of femininity limits the scope of feminine achievement, there is reason to believe that certain dynamics of female role behaviour serve to simultaneously maintain feminine identity and thwart long term achievement. Traits such as competency, competitiveness, independence, and rationality are associated with' masculinity7. The fact remains, however that girls' school achievement generally equals or exceeds that of boys. Success in academics and the qualities required to attain i t are incompatible with femininity, as defined by social norms. How, then is a 'feminine - self concept maintained in the face of such seemingly contradictory behaviour? Meredith Kimball originated a "fluke" theory of success to account for the discrepancy: responsibility for the successful achievement is denied, and attributed instead to chance (Kimball, 1974). Thus the "incompetent", 'hon-competitive"image remains intact. Empirical support for this theory may be found in the literature 26 concerning the "locus of control". The extent to which one sees events resulting from his own actions as opposed to luck or chance is a measure of how much he "internalizes" control over those events. If he attributes the majority of events in his life to chance rather than to himself, he is said to have an "external" locus of control. Although sex differences on this variable in childhood have not been confirmed, there is a tendency for women in college to "externalize" the locus of control,particularly over their academic achievements, more frequently than men (Maccoby, Chapter 4, 1974). Successes are not internalized; and therefore cannot contribute to the self concept. This mechanism itself is a dimension of femininity: 1A women are seen as dependent. Their dependence is more than economic: they are seen as dependent upon others for their self evaluation. The effect of this characteristic dependence is to make the 'feminine woman most vulnerable to the pervasive social influences which reinforce her stereotypic identity. It may be such an "external" orientation which allows female high achievers to cope with the success vs. femininity conflict in school. Although achievement interest and effort is generally associated with an internal locus of control (Joe, p. 628, 1971) two separate studies, one by Thurber, (1972) and another by Duke and Nowicki (1974) discovered correlations between "externality" and sustained high achievement in female college students; the inverse of the relationship for males. Thurber reasons that women adopt an external orientation as a defense against the stresses of competition. With less personal accountability for their success, they may reduce performance anxiety and anxiety over the conflict between femininity and success. The external locus of control phenomenon offers some rationale for 27 female academic success. Why then, should women's achievement decline after their academic careers are over? The locus of control device is apparently less effective in dealing with later conflicts involving worldly success. A rationale, proposes Maccoby, (1974) is that while school is a relatively structured situation, the post-school world requires that one take initiative in planning and directing achievements. The external orientation is no longer required to deny success, since i t has already served to decrease the likelihood success will occur. The link between locus of control and the tendency to plan for the future is again discussed in the work of Piatt and Eisenmann (1968). Subjects with an external locus of control were less able to conceive of future events than those classified as "internal". The internals had both a longer perspective into the future and saw a greater number of potential events, presumably because of their stronger belief in personal control over their own lives. Thus, in addition to defending themselves against conflicts that might result from success, women with an external orientation may lack the ability to plan or to fantasize future events leading to such success. There remains a third possible way of managing the femininity-success conflict. If her sex role conceptions have not been sufficiently strong to either keep her from aspiring to types of achievement that would threaten a "feminine" self concept, or to cause her to deny her achievements, a woman may succeed in a male-dominated field - the kind of success the first two mechanisms sought to avoid -- and find i t necessary to deny "femininity". Arlie Hochschild (1974) thus discusses the "defeminization" (p.196) of successful women. Women who have entered occupations in which there are very few females encounter a negative stereotype of their sex which they must over-come in order to be seen as competent and professional. In addition to proving 28 their abilities, they face the social pressures from male colleagues to prove they are unlike other women, who are seen as undependable, unable to think analytically or, on the other hand, unlike those other professional women who are seen as lacking in the expressive, nurturant feminine traits. In this way they often cut their ties with other women; thus engaging in what Hochschild terms "defeminization". Alternatively, the conflict itself may be the object of denial, as the biographical statements of a number of successful women in the sciences reveal. After relating obstacles in their career paths which relate directly to discrimination, phrases such as "they are the same kind of difficulties a man would have run into" or, "I don't feel I have made any sacrifices at a l l " refute earlier statements in which the same women acknowledge their difficulties. (Hochschild, 1974, p.195). Conversely, other women often partially reject the professional woman; being less willing to work for a woman than for a man, finding faults with '• their lives, and emphasizing their differentness. A generalized form this type of denial takes is the devaluation of women's products and of the work women do. Women in both male and female fields judged articles as less well written, less persuasive and important, when they were ascribed to a female rather than to a male author (Goldberg, 1968). Women tend also to underestimate the quality of their own achievements by comparison with men. To a greater extent, women will underestimate their expected performance on a future task. By comparison with men, they also predict lower grades for themselves than their past performance would indicate, (Baruch, 1974). This same phenomenon has been observed also in individuals of either sex who have low self-esteem (Coopersmith 1967). But the restrictions both of aspirations and of self concept that come about as a result of low self-esteem may be heightened when accompanied by the restrictive influence of a 29 stereotypically "feminine" self concept. People vary in the degree to which they internalize the sex role stereotypes prevalent in North American society. Similarly, occupations vary in the extent to which they are seen as typically masculine1 or feminine . Degree of sex-typing may be seen as another aspect of self concept, important to the process of matching self and occupational role concept, and apparently of particular importance to women. The present study will focus on the importance of sex role stereotypes in the choices of future occupations among a sample of Canadian women. Relevance of the Present Study The present study questions, in the light of recent research both on role stereotyping and self concept, whether women who choose traditionally feminine roles out of conformity to sex role self concept might also be doing so out of the low self-esteem characteristic of highly feminine women. In so doing they limit their aspirations and may, i f Korman's theory is valid, be denying important personal needs. The special relevance of an intervening variable such as self-esteem for women, would serve to explain the difficult-ies encountered by researchers in predicting women's choices from the commonly used career determinants which do not take self-esteem1' into consideration. The link between femininity and self esteem suggests that modifications in that stereotype would tend to enhance self esteem. Conversely, an increase in self-esteem would tend to influence identification with the feminine stereotype. Both imply expanded vocational options, as self-perceptions change and individuals' strengths emerge. How the stereotype should be modified is an issue that has been effectively addressed by 30 Sandra Bern. She advocates a balance between the masculine and feminine roles; a balance involving characteristics of both roles, since both include desirable, useful attributes. In her view, which she has substantiated with data (Bern 1974), "strongly sex-typed individuals, masculine or feminine, might be seriously limited in the range of behaviours open to them as they move from situation to situation, whereas a person with an androgynous or mixed self concept can freely engage in both masculine and feminine behaviours." (p.155). In a subsequent study, Bern verified that both masculine and androgynous subjects (Bern, 1975) were better able to resist social pressures to conform than were feminine subjects. With respect to vocational behaviours, this implies that a masculine or androgynous person has the widest range of choices as he is not limited to those prescribed by sex role definitions. Comparisons thus far have centered on the differences in self-esteem between those who see themselves as either masculine or feminine. The results of these studies, which have found the levels of self-esteem and mental health to be consistently higher among masculine than feminine subjects, tend to support the current masculine bias and the further devaluation of feminine qualities. Evidence that persons with androgynous self concepts have higher self-esteem than either masculine or feminine subjects would suggest that the acceptance of both kinds of qualities is more satisfying than the abandonment of feminine in favour of masculine traits. Such findings would lend empirical support to Sandra Bern's hypothesis that "the androgynous person will come to define a more human standard of psychological health" (Bern, p. 162, 1974). The relationship between androgyny and self-esteem will be explored in the present study. The bearing of each of those two personality dimensions upon women's career choices will become the major focus of this investigation. 31 DEFINITION OF TERMS Certain terms which will be used extensively in this study are subject to varying definitions depending on the context in which they are used. Because they have a specific meaning.in the present research study, the following definitions have been adopted: ROLE: "Any pattern of behaviours which a given individual in a specified set of situations is both (1) expected and (2) encouraged to perform." (David & Brannon, 1976). STEREOTYPE: A social expectation placed on people for the particular situation or category to which they belong. SEX ROLE: The role of "male" or "female" defined by one's culture. It is a pattern of behaviours defined as appropriate to one's gender, which includes such specific actions as gestures, style of dress and speech, and such complex behaviours such as the career striving typically assigned to men and the homemaking or social-emotional orientation traditionally assigned to women, in western culture. SEX ROLE STEREOTYPE: The cultural image of "maleness" or "femaleness". The masculine and feminine stereotypes refer to the two sets of personal characteristics males and females are expected to display in role appropriate behaviour. MASCULINITY: For the purposes of this study, i t is defined in terms of socialization. It is not a single characteristic, but describes the endorsement of all those qualities which make up the masculine sex role stereotype. A "masculine" person is one who tends to be ambitious, independent, and who actively pursues goals of success, 32 leadership, and physical and psychological invulnerability. He is rational, competent, and aggressive, and is willing to make decisions and take responsibility for them. FEMININITY: In the stereotypic sense in which i t is used in the present study, "femininity" refers to the tendency to be compassionate, understanding, warm and affectionate. A "feminine" person is gentle, cheerful, and responsive to the needs of others, and has a tendency to depend on others for her evaluation of the environment,and upon her relationships with others,for her self-esteem. ANDROGYNY: Derived from the Latin "andro" meaning "masculine" and "gyno" meaning "feminine", the term "androgyny" is currently used to mean the psychological integration of both masculine and feminine qualities into the self concept. An androgynous person can be both instrumental and expressive, both assertive and yielding, and chooses behaviours from either sex role according to the requirements of a given situation. SEX TYPED: One is sex-typed in the extent to which he has internalized society's standards of behaviour for one sex-role stereotype and divorced himself from the other. Hence one may be sex-typed "masculine" or sex-typed "feminine", depending on the direction his standards of behaviour have taken. An androgynous person endorses both stereotypes equally and is seen as the least sex-typed person. SOCIAL DESIRABILITY: The tendency to endorse traits which are positively valued in American society and to reject those which are negatively valued. Widespread agreement as to which characteristics are seen 33 as desirable and which are not has made it possible to measure differences among individuals in the tendency to say socially desirable things about oneself. In the present study, subjects who describe themselves as, for example, always or almost always "friendly", "truthful", "helpful" and "adaptable"; but never or rarely "jealous" "inefficient", "moody" or "conceited" are said to be high on social desirability. SELF-ESTEEM: The estimate one makes of himself and which he tends to maintain over extended periods of time. The degree to which one esteems himself is the degree to which he sees himself to be capable and effective in various aspects of his l i f e ; from the personal, and social, to the mastery of skills,and achievements. It is a personal judgment of one's own worth. INTER-PERSONAL SELF-ESTEEM: Is an estimate of how confident one feels of his worth in his personal l i f e , his family and social relationships, and his physical and moral self. As i t is measured by the Tennessee Self Concept Scale, the instrument employed in the present study,self-esteem includes a description of one's self-perceived identity and behaviour, as well as the evaluation of the self one perceives. SELF CONCEPT OF ABILITY: Is an estimate of how confident one feels of himself as an achieving person, a dimension not included in the TSCS, but particularly relevant to this study. Achievement is defined as academic achievement, and a high self concept of ability is seen to be evidence of high self esteem, high feelings of competence in the academic sphere. 34 CHAPTER II HYPOTHESES Problem Chapter I presented a review of the available literature on those personal and situational factors thought to influence women's vocational plans. The comparisons which have been made between women who have made important contributions through their careers—particularly in fields which are not traditionally feminine--and women who have either preferred homemaking or typically feminine vocations, has shown that the factors involved in women's career decisions differ from those involved for men. The problem identified by previous researchers, that the vocational predictions for women have been inaccurate when based on theories developed for men, has been clarified to a certain extent by the discussion of the differences between men's and women's career paths, and of the differences between career paths among various groups of women. The question to be investigated in the present research deals with an area of life experience which has emerged from those comparisons as a potentially relevant variable; that of sex role stereotyping. The specific objectives of the present study are to obtain information regarding the relevance to female vocational preferences of adherence to sex roles. The type of vocational preference in question is the preference for a career in the traditionally feminine category or for a career in the traditionally masculine category. In conjunction with the measurement of sex typing, levels of self-esteem characteristic of women with differing sexual identities will be assessed. Thus the relationships between self-esteem and sex typing, and the 35 importance of self-esteem in choice of a masculine or feminine occupation will also be explored. Specifically, the hypotheses being investigated in this study are: Masculinity/Femininity of Career Choice Relative to Self Concept  Hypothesis I Scientific: Based on the self concept model of career choice, i t is expected that women's sex role self concept will be congruent with the sex stereotype associated with the career they choose. Hence women with traditionally feminine self concepts will tend to choose careers in the traditionally feminine domans of teaching, nursing and arts, and women with more masculine self concepts will tend toward more typically masculine careers such as those in scientific, mathematical or agricultural fields. Stated in null form, the statistical hypothesis to be tested is as follows: There will be no statistically significant differences between women in traditionally masculine fields and those in traditionally feminine fields in terms of the level and direction of sex role stereotyping as expressed by Androgyny Scores on the Bern Sex Role Inventory (BSRI). In equation form this can be represented as: 0 mean Androgyny Score on the BSRI for women in the traditionally feminine fields mean Androgyny Score on the BSRI for women in the traditionally masculine fields. The Alternate Hypothesis is that the mean Androgyny Score on the BSRI for women in traditionally masculine fields will differ significantly from that for women in the traditionally feminine fields. Although the literature suggests the alternative hypothesis be directional, the evidence is insufficient to justify use of a one^tailed test. Hence the alternate HQ : y F - y M = where y p and y M hypothesis, in each case, is stated in non-directional form, indicating that two-tailed tests will be performed. In equation form, the alternate hypothesis is: Ha : vip - UjYj / 0 where Up and are defined as above. Hypothesis 2 Scientific: Since a self-perceived lack of ability and lower achievement motivation have been associated with traditionally feminine careeer choices, i t is predicted that women pursuing a non-traditional course will have a higher concept of their academic abilities than will women who, in their junior or graduating year, are pursuing traditionally feminine goals. Stated in null form, the statistical hypothesis to be tested is: There will be no statistically significant differences between women in traditionally masculine fields and those in traditionally feminine fields in terms of self-esteem as regards achievement, as measured by the Self Concept of Ability Scale (SCA). In equation form this may be represented as: HQ : U M - u F = 0 where u = mean Self Concept of Ability score for women m in masculine fields. and l i p = mean Self Concept of Ability score for women in feminine fields. The Alternate Hypothesis predicts a different mean Self Concept of Ability score for women in the atypical or masculine fields than for women in the traditionally feminine fields. The alternate hypothesis as an equation, is: HA : U M - l i p / 0 where and Up are defined as above. 37 Hypothesis 3 Scientific: Because of the associations between femininity and low self-esteem, and between femininity and traditional career choice, it is predicted that women choosing careers in the female dominated areas will evidence lower levels of self-esteem than those choosing careers in the male dominated areas. Stated in null form, the statistical hypothesis is the following: There is no statistically significant difference in mean level of self-esteem in personal and social areas as measured by the Total Positive score of the Tennessee Self Concept Scale (TSCS), between women in traditionally feminine and those in non-traditional, or masculine career groups. In equation form, the null hypothesis is: HQ: u M w u F =0 where y M = mean level of self-esteem for women in masculine n fields, and y F = mean level of self-esteem for women in feminine fields, as measured by the Total Positive score of the TSCS; to be labeled "Inter/Personal Self-Esteem" The Alternate Hypothesis predicts that the mean level of Inter/Personal Self-Esteem for women in the masculine fields will differ, to a statistically significant degree, that of women in the feminine fields. As an equation, the alternate hypothesis is: HA: y M - y F p 0 where and pp are defined as above Sex Role Stereotyping and Self-Esteem Hypothesis 4 Scientific: To test the theory that persons who have integrated 38 characteristics of both male and female stereotypes are psychologically healthier than those who adhere rigidly to one stereotype or the other, it is proposed that the more androgynous one becomes, the higher is her self-esteem in the personal/interpersonal area, and in the achievement area. Stated in null form, the statistical hypothesis to be tested is as follows: There will be no statistically significant linear relationship between degree of sex typing, as measured by the Androgyny score of the BSRI, and level of self-esteem, as measured by both the Tennessee Self Concept Scale and the Self Concept of Ability scale. In equation form, this can be represented in the following way: HQ : pxy = 0 where x = self-esteem score (a) Inter/Personal (b) SCA and y = Androgyny Score (absolute value). The alternate hypothesis states that there will be a linear relationship between self-esteem and androgyny. As an equation, this prediction becomes: HA : P xy * 0 Hypothesis 5 Scientific: While the previous hypothesis suggests that as one becomes less sex typed, her self-esteem increases, there is evidence in the literature to suggest that the direction, as well as the degree, of sex typing bears a relationship to self-esteem. It is expected that women whose self concept falls within the category of feminine sex typing will have lower levels of self-esteem than women categorized 39 as typically masculine. It is further predicted that women who score androgynously will have higher self-esteem than the masculine,and consequently, higher also than the feminine sex typed groups of women. Stated in null form, the statistical hypothesis is: There is. no statistically significant difference in mean level of self-esteem as measured by the Self Concept of Ability scale and the TSCS, between women classified masculine, feminine, or androgynous by their Androgyny scores on the BSRI. In equation form, the null hypothesis may be represented as follows: yF mean level of self-esteem in all areas for women classified as androgynous mean level of self-esteem in all areas for women sex typed masculine mean level of self-esteem in all areas for women sex typed feminine. The alternate hypothesis is that women classified androgynous by their Androgyny scores on the BSRI will d i f f er i n level of self-esteem from those sex typed masculine; who will in turn, di f f er in leve)-of self-esteem from those sex typed feminine. In equation form, the alternate hypothesis may be represented as: HA : A^ * Vfn y F where y^, y^, and y p are defined as above. H0 : H A = ^ where u, and y r 40 Research Assumptions The preceding hypothesis are based on certain assumptions. To the extent that they are accurate, these assumptions permit the acceptance of the alternate hypotheses where results indicate the null hypotheses should be rejected. To the extent that these assumptions are inaccurate, they limit the generalizations which may be drawn from the results of the study, in terms of either the null or the alternate hypotheses. Assumptions 1) That the sex stereotypes assumed to be associated with the fields of study sampled are an accurate reflection of stereotypes commonly held by women at U.B.C. 2) That the sample is representative of women in the U.B.C. departments from which it is drawn. 3) That women at U.B.C. are representative of university women in general in North America. 4) That the psychological measuring instruments used are adequately sensitive to reflect actual differences in the constructs they are purported to measure: sex role stereotyping, self-esteem, social desirability. CHAPTER III METHODOLOGY This chapter describes the methods used to investigate the research problems identified in the previous chapter. The human and material resources employed will be detailed under the headings "Design", "Sample", and "Measures". Procedures for scoring and analysing the data follow those descriptions. Three phases of data analysis were performed; an initial hypothesis testing is followed by analysis of one of the instruments used i the testing. Information gained in this secondary analysis is used to further clarify results from the original hypothesis testing phase. This information is presented in Chapter IV, "Post Analyses". Design The first area of investigation, addressed by Hypotheses 1, 2 and 3, i concerned with understanding the sex-type dimension of career choice. Because self theory is the model proposed for explaining preference for either a masculine or a feminine field, variables related to self concept are examined for their relevance to this aspect of vocational preference. A multivariate model is employed to test these relationships. The independent variable, masculinity/femininity of aspirations is defined by membership in one of the specific fields associated with one or other of the sex role stereotypes. Choice of these fields was based on the classifications used by previous researchers, either according to the dominant sex employed in various fields (e.g. Karman, 1973, Almquist, 1971) 42 or according to the stereotype a majority of people sampled, connected with various occupations named. (Hawley, 1972; Kriger, 1972). For the present study, three fields within each of the traditionally masculine and traditionally feminine areas are selected to represent the diversity of career choices within those categories. They are, for the masculine area; Science, Mathematics and Agriculture; and for the feminine area; Education, Nursing, and Arts. The variables chosen as potential discriminators of masculine or feminine careeer choices represent two aspects of self concept: sex role stereotype endorsed and level of self-esteem. The former has been shown to affect the level and type of aspirations of women, and the latter variable to affect the degree of congruence between self concept and occupational role concept. An attempt is made in the present study to assess the effect of sex typing and self-esteem, both independently and in combination, on the stereotypy of aspirations. The remaining hypotheses deal with the relationships among the self concept variables. Correlational and multiple regression analyses are used to test their effects upon one another. Sample Ninety third and fourth year female students at the University of British Columbia make up the sample. The Nursing and Arts groups are fourth year students, while the other groups are third year students. The deviation in the case of the former two groups is due to the following factors: The third year Arts students were not chosen because their classes contained insufficient numbers for the sample. The third year Nursing students were determined to be inappropriate because they were following an experimental, rather than a traditional training approach. This particular academic level is preferred for two reasons: 43 1) that they, as opposed to high school graduates or freshmen, would have selected an area of specialization, but 2) neither their self image nor their career choices would have had to undergo modification by outside influences such as marriage, adult family commitments, or pressure to conform to the occupational stereotype once employed. Subjects were selected on the basis of their enrollment in the fall of 1975, in one of the fields listed preceding section entitled "Design". Those representing the stereotypically feminine :(F.) career choices are: Fl. - Fifteen Primary School Teachers in training, all enrolled in Education 310, "Growth and Development". F2. - Fifteen students from the graduating class in Nursing responded. F3. - Although the social science field in general has been associated in the literature with typically feminine choices, to be consistent with the level of female participation exemplified by the two preceding groups, a class composed mainly of females was preferred. For this reason a French class was chosen; the French 412 class in particular/it contained the largest number of women specializing in • i French. Subjects enrolled in the traditionally masculine (M) fields were: Ml. - Chemistry 315 students, numbering 12, and three 3rd year Physics majors, comprised the Science group. M2. - Fifteen women majoring in Mathematics. Because no single course grouped the women together, subjects whose names appeared on the l i s t of female students provided by the Mathematics Department were contacted individually and asked to participate. Twenty women chosen randomly from the l i s t were contacted, and of those, 15 made up the final group. 44 Procedures for selecting the final group are described under the following section, "Response Experience". M3. A third year Genetics class was the source from which the Agriculture group was taken. Class time was used for the testing in this case, so the entire class was tested and selection made from this pool according to ID numbers, as detailed in the following paragraph. Response Experience. During the.data collection phase, the sample size decreased from the intended 120 to 90 subjects. Due to class time constraints, testing for all groups except those in Agriculture, could not be conducted during a class meeting. Two separate procedures were required to accommoate the schedules of the remaining groups: For the Teaching, Nursing, Arts and Chemistry students, the project was introduced and questionnaires distributed at the beginning of a class session. Subjects were asked to complete them and to return the materials the following period when they would be collected. Some subjects in each group either failed to return the questionnaires or returned them with portions unanswered, causing the sample size to be reduced. The group with the lowest return rate were the Teachers; hence the others, including the Mathematics group, were reduced to match that number (15). This was accomplished by eliminating from the processing, those questionnaires in each of the remaining groups whose ID numbers exceeded "15". No bias would be expected to result from this procedure. In all cases, subjects were supplied with instructions for completing each of the instruments. They were given a brief verbal and written explanation of the purpose of the study prior to administration of the test packages, and assured that though group results would be shared, individuals' rights to confidentiality would be respected. 45 Measures The Bern Sex Role Inventory Few of the studies linking stereotypic occupational aspirations to women's beliefs about sex appropriate behaviour have employed self inventories. They have tended to measure attitudes and motives thought to tape the dimension of sex role stereotypes. Since the present study is concerned with implementation of the self concept in occupational choice, a measure of how the individual sees herself with respect to stereotypic traits was required. 1) Advantages The Bern Sex Role Inventory was chosen for this purpose because of the following advantages: (1) Normative data for a college age population. (2) Control for social desirability. (3) Adequate reliability and validity. (4) Its major advantages over other instruments are two innovations; the operationalization of the concept "androgyny", and the treatment of Masculinity and Femininity as two independent dimensions rather than as polar opposites on a single continuum. 2) Scale Descriptions The BSRI consists of three scales, two of which assess sex role stereotyping and a third which measures social desirability. The Masculinity and Femininity Scales each contain 20 characteristics, each of which is rated on a scale ranging from 1 (never or almost never true of myself) to 7 (always or almost always true of myself.) Classification of items into the appropriate scale was based on the judgements of 100 male and female college students. The Masculinity Scale is composed of those items judged to be significantly more desirable in a man than in a woman; traits 46 indicative of an instrumental orientation and a cognitive, rather than an emotional focus. The Femininity Scale consists of those items judged to be significantly more desirable in a woman than in a man in American society. The feminine items reflect an "expressive orientation, an affective concern for the welfare of others" (Bern, p. 156, 1974). Subjects receive a score for Masculinity and a score for Femininity, which equal the mean self-rating of the 20 masculine and 20 feminine traits respectively. A third score, for Androgyny, is computed from these two scores. By measuring the difference between an individual's endorsement of the feminine over the masculine traits or vice versa, and testing that difference for its statistical significance relative to the individual's own standard deviation among item scores, the Androgyny score evaluates both the degree of sex typing and whether one favours the masculine or the feminine stereotype. One whose Androgyny score shows no statistically significant difference between her endorsement of traits belonging to the two scales is said to be "androgynous". The larger the discrepancy between scores on the two scales, the more sex typed (and less androgynous) the person is said to be. The 20 item Social Desirability scale is "completely neutral with respect to sex". (Bern, p. 156, 1974). The original purpose of this scale was to show whether or not ascribing to the socially desirable sex role attributes implied a general social desirability response set. The conclusion drawn by the scale's author on correlating this scale with the other two and with Androgyny was that the Androgyny Score taps, not a general tendency to say positive things about oneself, but a more specific tendency to identify with sex-appropriate standards that did not extend to traits unrelated to sex roles. The Masculinity and Femininity scales, on the other hand, do correlate with this scale, confirming the decisions of those who 47 judged their sex-appropriateness. 3) Psychometric Data for the BSRI The instrument is normed on 900 male and female university and college students. Internal consistency and stability estimates range between .70 and .86, and between n. = .89 and ^ = .93 respectively. Concurrent validity for the scales has been investigated both through conventional methods of correlation with comparable tests and through a factor analytic approach. Construct validity has been demonstrated through behavioural studies involving direct observation of subjects assessed by the inventory. A more detailed report on the psychometric characteristics of the BSRI is presented in Appendix A. The Tennessee Self Concept Scale The Total Positive Score from the Tennessee Self Concept Scale was chosen as the measure of self-esteem in the personal and social area. The Self Criticism score, also from the TSCS is included for its value as a validity check on the total positive score. 1) Advantages: The TSCS was chosen for its scope and for its statistical advantages. It measures attitudes toward the self in 5 main areas, or aspects of life-functioning; from the "Physical", "Moral", and "Personal", to the "Family" and "Social" areas. Norms for groups at various levels of mental health have been provided. A wide body of research supports the validity and reliability of the total test. 2) Description of Scales: The Tennessee Self Concept Scale (TSCS) is described by its author, William H. Fitts as focussing on "the self as observed, experienced and judged by the individual himself" (Fitts, p.14, 1971). The self is seen as composed of three "subselves": the self-as-object (Identity Self); the self-as-doer (Behavioural Self); and the self-as-observer and evaluator of the other two selves (Judging Self). The pool of 90 items 48 are statements about these three subselves which the subject rates on a five-point scale ranging from "completely false" to "completely true". The sum of the three subscores, which is equivalent to the sum of scores for the five areas listed above, is the Total Positive Score, which Fitts defines as follows: the TOTAL P score is the most important single score on the Counselling form of the TSCS. It reflects the overall level of SELF-ESTEEM. Persons with high scores tend to like themselves, feel that they are persons of value and worth, have confidence in themselves, and act accordingly. People with low scores are doubtful about their own worth, see themselves as undesirable, often feel anxious, depressed and unhappy, .and have l i t t l e faith or confidence in themselves. (Fitts, p. 2, 1965). The Self Criticism Score is included among the variables in the present study both for its function as a check on the accuracy of Total Positive scores, and out of an interest in correlations between tendency to be self critical and the tendency to follow rigid sex role definitions. It is defined in the manual in this way: Self Criticism Score: This scale is composed of 10 items. These are all mildly derogatory statements that most people admit as being true for them. Individuals who deny most of these statements most often are being defensive and making a deliberate effort to present a favourable picture of themselves. High scores generally indicate a normal, healthy openness and capacity for self-criticism. Low scores indicate  defensiveness, and suggest that the Positive Scores are probably  artificially elevated by this defensiveness. (Fitts, p.2, 1965). 3) Psychometric Data for the Tennessee Self Concept Scale Normative data for the TSCS is available for a broad sample of the U.S. population, including groups differing in psychological health, clinically determined. Internal consistency is estimated at .91, while stability estimates range from .^= .88 to .^ = .92. Both behavioural and self report measures have been used to establish the concurrent validity of the TSCS, with particular emphasis on the validity of the Total Positive score 49 as an estimate of self-esteem. That the scale does measure the construct "self-esteem" is evident from correlations between Total Positive scores and clinical descriptions of individuals' functioning. Additional information regarding these aspects of the TSCS are found in Appendix A. The Self Concept of Ability Scale 1) Advantages The test to be used as a measure of self-esteem in the achievement area is the Self Concept of Ability Scale developed by Brookover, Paterson, & Thomas (1962). The instrument was designed to measure the effect on school achievement of differing self concepts of the abilities required in school. Among the children for whom it was intended, Self Concept Ability score was independent of measured intelligence in predicting school achievement. A high level of confidence in ability was related both to high present performance and higher future educational aspirations. The test reflects how self-esteem in this area, as in others, is affected by one's perceptions of the esteem in which others hold him. Correlations were found between a student's own self concept of his ability, and the one he thought significant others in his life had of his ability. Since self concept was shown to be related to actual performance, i t was suggested that those significant others could, through communicating different expectations of the child to him, influence his performance and aspirations in a positive way. This type of sensitivity to external influences demonstrated by the scale was seen as an advantage in the present study, given the links that have been found between women's achievement and the external locus of control. Because the scale was written for children, certain adaptations were required for use with women. The adaptation involved changes in wording, such as to substitute "University" for "school" in questions involving the present, and "postgraduate" or "doctoral" for "university", in questions regarding future aspirations. 2) Description of Scale The test consists of a set of 16 questions to which the subject is asked to respond on a 5-point scale as to her opinion of her ability, her work, and her comparative academic standing. A high total score is indicative of a high self concept of ability, while a low score reflects a low self concept of ability. 3) Psychometric Data for the Self Concept of Ability Scale No information as to the reliability of the scale were available. Validity for the construct being measured is determined by the relationships observed between scores on this scale and actual achievement, and attitudes toward achievement. Further details are present in Appendix A. Summary of Measures: In Table II are listed the instruments employed in the present study. The relevant scales from each instrument are specified, along with the variable names assigned to them for purposes of this study. Scoring procedures for each scale are summarized. Copies of each of the instruments, with the exception of the TSCS, are included in Appendix A . Copies of the TSCS can be obtained from the publisher listed in Appendix A . The instructions given to participants prior to responding to these instruments are also reproduced in the same Appendix. FIGURE 1 SUMMARY OF MEASURES EMPLOYED IN THE PRESENT STUDY TEST NAME SCALES VARIABLE NAME SCORING METHOD Bern Sex Role Inventory Masculinity Masculinity mean score for 20 items Femininity Femininity mean score for 20 items Androgyny Androgyny t-ratio of difference between Masculine and Feminine scale means Social Desirability Social Desirability sum of scores on 20 items Tennessee Self Concept Scale all scales except Self Criticism Inter/Personal Self-Esteem sum of scores on 90 items (the TOTAL POSITIVE Score) Self Criti-cism Self-Criti-cism sum of scores on 10 items Self Concept of Ability Scale test con-tains a single scale Self Concept of Ability sum of scores on 16 items 52 Processing of Raw Data Responses Responses to the Tennessee Self Concept Scale were made on specially designed optical scan sheets which were returned to Tennessee for scoring. This provided computer output with profiles for each person, along with scale scores, group means and standard deviations, and the punched IBM output for further analysis. Scoring for the Self Concept of Ability Scale was done by calculator, which generated totals for each person. The sum of scores on the 16 items constitutes the score used in data analysis. The Bern Sex Role Inventory was scored by computer, using the UBC MULTIVAR program (Finn, 1974). After raw scores attached to individual ID numbers were keypunched onto IBM cards, totals were generated for the Masculinity, Femininity, and Social Desirability scales. The Social Desirability score remained in this form, with a high score indicating the greatest tendency to describe oneself in socially favourable terms, and a low score indicating the least tendency to do so. Computation of the Androgyny score required mean scores for each of the two 20-item Masculinity and Femininity scales, rather than sums of scores. The ^-ratios which were subsequently derived from the difference between mean Femininity and Masculinity (F-M) became the Androgyny scores. High scores indicate feminine sex typing; near zero scores, androgyny; and low (negative) scores, masculinity. Data Analysis Summarized below are the first three hypotheses, which deal with sex stereotype of field and the self concepts of women in those fields: Hypothesis 1. - There will be no statistically significant differences between women in traditionally masculine fields and those in traditionally feminine fields in terms of level and direction of 53 sex role stereotyping. Hypothesis 2. - There will be no significant difference in mean level of self-esteem in the achievement area, as measured by the Self Concept of Ability Scale, between women in traditionally feminine fields and women in fields considered atypical for women, but typical for men. Hypothesis 3. - There will be no significant difference in mean level of self-esteem in personal and social areas as measured by the Total Positive score of the TSCS, between women in traditionally feminine and those in non-traditional or masculine career groups. The following analyses were performed to test, simultaneously, the preceding three hypotheses: A multivariate analysis of variance using UBC MULTIVAR (Finn, 1974) was employed to discriminate between women in the three feminine fields from those in the three masculine fields on the basis of their scores on the self concept measures. The discriminant analysis may be represented by the following general equation: Y = b Q + b ]X ] + b2X2 + b 3X 3 + b 4 X4 + b5 X5 + b6 X6 + b 7 X 7 + e where Y = Masculinity/Femininity of Field a n d X] = Social Desirability X2 - Self Criticism X^  = Masculinity 54 = Femininity Xg = Androgyny Xg = Inter/Personal Self Esteem Xy = Self Concept of Ability bg = constant (y-intercept) bX (1-7) = slope of X(l-7) ^ e = error The analysis was performed according to the following procedure: A multivariate test for significance (Finn, 1974) of the differences among means of all groups of subjects was computed, followed by a univariate test corresponding to each of the dependent variables. In addition, a discriminant function analysis was performed to identify any linear combinations of those variables which might successfully distinguish either the six subgroups, or the two categories, masculine and feminine fields, containing three subgroups each. Hypothesis Four was tested by observing the magnitude of the correlation between Androgyny and self-esteem. Hypothesis Four is: There is no linear relationship between degree of androgyny as measured by the BSRI and level of self-esteem in either the Inter/Personal or the achievement dimensions. Disregarding the sex stereotype of fields to which subjects belonged, the relationships between their scores on variables related to sex typing and to self-esteem were measured. The UBC CORN Program (Halm, 1975) was used to 55 compute Pearson product-moment correlation coefficients among all variables. These correlation coefficients were tested for statistical significance, also using the CORN program. To detect, in addition to the linear relationships defined using the CORN procedures, any non-linear relationships that could exist, a further analysis was performed. The BMD 02V Program (Halm, 1970) was employed to produce scatter plots of the correlations described in the preceding analysis, and those plots examined visually for patterns of a curvilinear nature. In a subsequent step undertaken to clarify the relationships derived from the correlational procedures described above, regression procedures were employed. Stepwise Multiple Regression analysis using the UBC TRIP program (Bjerring & Seagraves, 1974) was performed on self-esteem variables Inter/Personal Self-Esteem and Self Concept of Ability when it was found that variables correlating with each of these self-esteem scores were themselves significantly intercorrelated. Through multiple regression, the independent effects of each of the variables contributing to the variance in the two self-esteem variables were measured. The UBC TRIP program automatically selects for steps beyond the initial one, only those variables which contribute a significant (p<.05) proportion to the variance of the dependent variable, and ranks them in order of greatest to least predictive ability. The regression equation took the following general form: Yl,2 = b0 + b l X l + b2 X2 + b3 X3 + b4 X4 + b5 X5 + b6 X6 + e 56 -where Y] = Self Concept of Ability Y2 = Interpersonal Self-Esteem (TSCS Total P) a n d X] = Social Desirability X2 - Self Criticism X^  = Masculinity X^  = Femininity Xg = Androgyny Xg = Inter/Personal Self-Esteem (for Y-j) = Self Concept of Ability (for Y2) bg = constant (y-intercept) bX(l-6) = slope of x(l-6) e = error Hypothesis Five - that there is no significant difference in mean level of self-esteem between women classified masculine, feminine or androgynous by their Androgyny Scores on the BSRI will be tested according to a multivariate procedure. The differences among means for the three groups were tested for significance, with both multivariate and univariate tests, using the MULTIVAR program. Summary A study of the impact of sex role stereotyping upon career choice and upon self-esteem was carried out at the University of B.C. in the fall of 1975. Subjects were 90 female undergraduates, chosen for their enrollment in either one of three traditionally feminine fields, or alternatively, one of three traditionally masculine fields of study. The responses of these women to measures of sex role stereotyping and self-esteem (in both personal and academic areas) were examined in several ways: fir s t , to assess the ability of these variables of self concept to predict membership in one of the masculine or feminine fields, and secondly, to identify which, if any, of the criterion measures are meaningfully related and could be used to predict one another. A third approach to the data was to explore dimensions of the overall Masculine and Feminine stereotypes relevant to this sample. (This will be described in the Section "Post Analyses"). The strategies employed to analyse the data in each of the ways described above was outlined. 58 CHAPTER IV RESULTS This chapter details the findings based on the analyses outlined in the preceding chapter. They will be presented in the order in which the hypotheses were first stated. Because the same statistical procedures were used to test simultaneously the first three hypotheses, a brief summary of them will introduce the first set of results: Women in traditionally masculine fields do not differ significantly from women in traditionally feminine fields of study on any of three measures of self concept: Hypothesis 1. - on Sex Role Stereotyping, measured by the BSRI Hypothesis 2. - on self-esteem in the achievement area, measured by the SCA scale Hypothesis 3. - on self-esteem in the personal and interpersonal area, measured by the TSCS. Response to Hypotheses one, two, and three Table I presents the means and standard deviations for each of the six career groups on the self concept variables. The results of the multivariate analysis, which appear in Table II, explain why the null hypotheses, one, two, and three were not rejected. In Table II it can be seen that the differences among group means are not statistically significant. The statistically nonsignificant multivariate F indicates that groups did not differ significantly on the set of variables. It indicates further that no linear combination 59 TABLE I Means and Standard Deviations of Career Groups on Self Concept Variables X I Career 'Z Groups > Masculinity Femininity * >> E >> CD O S-' T3 • E • • • • =£ Social Desirability Inter/Personal Self-Esteem Self Concept of Ability Self Criticism Teaching n=15 x SD 4.66 0.60 5.12 0.51 +T.OO 1.65 99.67 9.45 348.07 27.70 58.33 6.26 36.00 5.55 Nursing n=15 x SD 4.86 0.75 4.91 0.58 +0.25 2.34 104.73 8.50 355.67 38.52 59.47 7.74 35.87 6.99 Arts n=15 x SD 4.47 0.77 4.81 0.56 +0.80 2.36 96.33 8.61 339.60 24.65 55.13 8.30 36.07 3.90 Science n=15 x SD 4.85 0.61 5.07 0.56 +0.40 1.87 101.60 10.25 355.27 41.60 63.33 6.97 35.07 6.06 Agric-ulture n=15 x SD 4.61 0.65 4.66 0.38 +0.06 1.51 100.93 10.53 350.40 44.05 52.80 7.01 34.20 5.48 Math n=15 x SD 1 4.68 0.81 1 4.75 0.70 +0.30 2.51 102.20 7.28 357.40 20.41 60.13 6.63 33.00 4.55 *Signs are included to indicate source of standard deviations which exceed mean scores: Androgyny scores may be either positive or negative values. TABLE II Univariate & Multivariate Analysis of Variance on Self Concept Variables for 6 Career Groups Source Univariate F (5,84) Masculinity 0.6737 p<.6448 Femininity 1.5538 p<.1822 Androgyny 0.4435 p <.8169 Social Desirability 1.4005 p<.2325 Inter/Personal Self-Esteem 0.5697 p<.7230 Self Concept of Ability 4.0840 p<.0023 Self Criticism p<.5942 Multivariate Test F(35,330.55) = 1.1593 P < .2525 61 of the self concept variables could successfully discriminate among them, either as six independent groups, or as two broad groups corresponding to the categories "masculine fields" or "feminine fields." It must be assumed, from these results, that variations in self concept within each of these groups exceed the variations among them. The second series of analyses were performed to test Hypothesis Four. It is, in summary: Hypothesis 4. - There is no statistically significant linear relationship between degree of sex typing, as measured by the Androgyny Score of the BSRI, and level of self-esteem, as measured by both the Tennessee  Self Concept Scale and the Self Concept of Ability scale. In Table III is presented the matrix of Pearson product-moment correlations among variables of self-esteem (Interpersonal Self-Esteem, Self Concept of Ability, Self Criticism,) and Sex Role Stereotyping (Masculinity, Femininity, and Androgyny). The Androgyny Score, representing the level of feminine relative to masculine, sex typing, bears no linear relationship to any of the variables referring to self-esteem. In view of this result, the presence of curvilinear relationships was explored. The possibility of a curvilinear relationship accounting for the zero order correlation was discarded on inspection of the scatter plots for the above correlations. Masculinity and Femininity, however, do show relationships with self-esteem. These correlations vary, depending on whether i t is self-esteem in the interpersonal and personal area, or in the achievement area, which is being considered. First with respect to: 1) Inter/Personal Self-Esteem (TSCS) The correlation between the Femininity Scale and Score and the Inter/Personal Self-Esteem Score is statistically nonsignificant: *(89) = .172. 62 TABLE III INTERCORRELATION MATRIX FOR COMBINED GROUPS Variable 1 2 3 4 5 6 7 1. Self Concept 1.00 of Ability .26* .24* -.02 .43** 29** -.12 2. Social Desir-ability 1.00 .63** -.23* .26* .13 -.12 3. I/P Self-Esteem 1.00 -.37** .35** .17 -.17 4. Self Criti-cism 1.00 -.03 -.21 -.10 5. Masculinity 1.00 .05 -.75** 6. Femininity 1.00 .59** 7. Androgyny 1.00 * Ul > .22 is significant at the <* .05 level, df = 89. ** |*| > .26 is significant at the ^  .01 level, df = 89. Leading decimal points are omitted. 63 The Masculinity Scale Score, on the other hand, does bear a statistically significant relationship to scores for Inter/Personal Self Esteem, *(89) = .353, p < .01. In order to verify the conclusion that the two sex role scales differ in their relationships with the Inter/personal self-esteem scores, a multiple regression analysis was undertaken. From this analysis i t was learned that the Masculinity Score-s^H+T accounts for a significant proportion of the variance in Inter/Personal Self-Esteem scores when the contributions of the remaining variables are statistically controlled. The results of the stepwise regression analysis presented in Table IV indicate that although Social Desirability is clearly the most powerful predictor of the level of self-esteem, accounting for 39.9% of the variance in Inter/Personal Self-Esteem Scores, the Masculinity Scale does contribute a statistically significant increment (3.88%) to the overall variance in self-esteem. In relation to: 2) the Self Concept of Ability, both Masculinity and Femininity scores are significant. In the case of Masculinity, however, the relationship is more pronounced; *(89) = .29, p < .01; for Femininity. These two variables, along with Social Desirability and Inter/Personal Self-Esteem, produce a significant effect on the ability self concept score. Results of the multiple regression analysis clarify which of the variables mentioned above have significant independent effects on Self Concept of Ability. From Table IV i t can be seen that Masculinity is the strongest 2 predictor of scores on the Self Concept of Ability; n = .1815, p < .05. Second in importance is the Androgyny score, which had not emerged as significant in the correlational analysis. Androgyny adds an increment of 64 TABLE IV Stepwise Multiple Regression Analyses For Self-Esteem A. PERSONAL/INTERPERSONAL AREA: INTER/PERSONAL SELF-ESTEEM (TSCS) Order of Predictor Variables Variable Symbol & Name R2 Increment in R2 1. X-, Social Desir-1 ability .3990 .3990 2. X£ Self Criticism .4553 .0563 3. X^  Masculinity .4941 .0388 Variables listed are significant to the a = .05 level ACHIEVEMENT AREA: SELF CONCEPT OF ABILITY Order of Predictor Variables Variable Symbol & Name R2 Increment 2 in R 1. Xg Masculinity .1815 .1815 2. X5 Androgyny .2719 .0904 65 9.04% to the explained variance in the Self Concept of Ability scores. The unique variance of Femininity, on the other hand, is too small to reach significance. Summarizing the effects of sex typing on the achievement self concept presented in Tables IV and V, i t appears that 1) the more strongly the women endorse masculine characteristics, the higher their academic self concept tends to be, 2) differences in androgyny, i t , the degree of discrepancy between endorsement of masculine and feminine qualities, imply differences in academic self concept, and 3) how much the feminine qualities are endorsed has no relationship to academic self concept. A final set of analyses were to be undertaken in response to Hypothesis 5, which is: There are no statistically significant differences in mean level of self-esteem, measured by the Self Concept of Ability Scale and the Tennessee Self Concept Scale, between women classified androgynous by their Androgyny Scores on the BSRI, and those classified either masculine or feminine by the same score. The multivariate analysis which would have been used to test this hypothesis was not carried out, due to the results obtained in the testing of Hypothesis Four. The groups described in Hypothesis Five were to be formed on the basis of differences in Androgyny scores; with the Androgynous group being composed of subjects whose scores fell between plus and minus one, the Masculine group being composed of subjects scoring below -.2.025, and the Feminine group being composed of those scoring above 2.025. These groups were to be compared on the basis of their mean self-esteem scores. However, because the results found in testing the relationship between Androgyny and self-esteem indicated that the two variables do not covary to a statistically significant degree, no significant results would be expected from comparing groups differing in Androgyny Scores. It may be assumed, therefore, that the null hypothesis in this case should not be rejected. 66 CHAPTER V.". POST ANALYSES On completion of the analyses outlined in Chapter III, the results reported in Chapter IV were examined in terms of the solutions they offered to the research questions posed in the present study. A second phase of analyses designed to gain insight into these results was carried out for the following reasons: 1) The results pertaining to both hypotheses contradict those of the body of previous research. 2) Previous research cited in the rationale for the present study used measures other than the BSRI to evaluate sex role stereotyping. 3) Correlations among the Bern scales, particularly the relationships of the Femininity scale with Social Desirability, are at variance with those reported for the BSRI Norm group. Given that the Androgyny score is derived from scores on Masculinity and Femininity, irregularities in distributions of those scores or irregular response patterns within them, of which the lack of correlation between Femininity and Social Desirability could be a symptom, would have an effect on the observed Androgyny levels. 4) Recently, the computation of Androgyny by use of a t ratio was questioned. (Spence, et al., 1975). A similar questioning procedure was undertaken with data from the present sample in an effort to interpret the concept of "androgyny" in relation to the self-esteem variables. 67 Methodology The procedures used to clarify the results reported to this point involve an analysis of the BSRI and a second phase of testing for the original hypotheses. The aspects of the BSRI to be analysed are: 1) the computation and the definition of the Androgyny Score 2) dimensions of the Masculinity and Femininity Scales, to be explored through factor analysis. The information gained by investigation of these two aspects of the BSRI will be used to perform the second phase of hypothesis testing which is described immediately following the description of procedures involved in the first two questions. The Androgyny Score J: : Computation of the Androgyny score - The use of a t ratio to compute Androgyny presupposes that distributions of score on the two component scales are normal. Should these distributions not be normal, use of a t ratio, and the cut-off points differentiating androgynous from non-androgynous subjects, could not be justified. Histograms representing the frequency distributions of Masculinity and Femininity scores were drawn, and examined visually for conformity to the normal curve. Operational definition of Androgyny The use of a ratio score for androgyny, which indicates only the differences between the levels of masculine and feminine traits endorsed and ignores the absolute level of endorsement each stereotype could obscure important differences among women classified as "androgynous". (Spence et al., 1975). A significant difference in self-esteem between groups of androgynous women could conceivably affect the correlations observed between Androgyny and self-esteem. This was the specific focus of the second hypothesis of the 68 present study, and therefore deserving of attention. For this purpose, the women scoring "androgynous" according to the t ratio; i.e., having Androgyny Score between (+1 and -1) were re-classified according to the "median split" method introduced by Spence. On the basis of their Masculinity and Femininity scores, subjects formed three groups: Low-Low scorers: those scoring below the sample median on both Masculinity and Femininity. High-Low scorers: those scoring above the median on Masculinity and below the median on Femininity High-High scorers: those scoring above the sample median on both Masculinity and Femininity These three groups of Androgynous women were compared on the basis of their mean Self Concept of Ability scores, Inter/Personal Self-Esteem, Social Desirability and Self-Criticism scores. The multivariate procedure used for this was OWMAR (Hakstian, 1977), a program designed to more efficiently identify just which of the groups among a set of groups show significant differences. Like MULTIVAR, i t also performs discriminant function analysis, to see whether, i f there is not one variable that clearly discriminates between groups, a linear combination of variables might succeed in doing so. Response to the second question regarding dimensions of the Bern Sex Role  Inventory took the following form: The most meaningful factor structure of the instrument was identified by studying the results of different factor analytic models, and by comparing several solutions within each model. Procedures used were: Principal Components Analysis 69 Image Factor Analysis from the AGFA Program (Hakstian, 1972) Orthogonal rotation with varimax transformation was used with both factor analytic models. With the Image model, oblique transformation was also used. The criterion for deciding the number of factors to be rotated was the number of eigenvalues greater than one. After rotation, those factors with eigenvalues greater than one were interpreted in terms of the items contributing significantly to them. Since the proportion of the factor variance attributable to an item equals the square of its loading on that factor, items with loadings greater than .40 were considered to be significant. Those items were identified for each factor and studied subjectively for a common construct. By assigning specific numbers of factors to both Principal Components and Image models at this stage, the number of uninterpretable factors was reduced as the smaller factors were collapsed. Choice of the numbers of factors to be assigned was facilitated by inspection of the Scree Plot for the Principal Components model, i.e., the graphic representation of the ratio of eigenvalues to the number of factors, to see at what point the amount of variance attributable to the factors defined drops appreciably. The factors generated by each of these solutions were compared and the most common factor structure with also the most meaningful factor (both statistically and subjectively) identified. The results of this comparison became the final "composite" solution which was then utilized in the remaining analyses. Rescoring of the BSRI according to the Factor Solution To rescore the BSRI in terms of the factors identified by the factor analysis, raw scores for each person on the 33 items involved were subjected to analysis by the LERTAP program (Nelson, 1974). Scores for these factors 70 were sums of the individual item scores and not weighted "factor scores". To further check the internal consistency of the factors, the resulting scores were subjected to LERTAP (Nelson 1974) procedures" for computing item/scale correlations, means and standard deviations for each factor, and correlations among factors. Second Phase of Hypothesis Testing In order to determine whether the reorganization of the Bern Sex Role Inventory into these factors would change the answers to the initial hypotheses, a separate set of analyses identical to the first were undertaken; with the exception that the variables "Masculinity" and "Femininity" were replaced by the scores on the new factors. re Hypotheses 1, 2 & 3 - Accordingly, the same multivariate analysis using the MULTIVAR program, was performed on the six career groups, focussing particularly on the Bern Factors as potential discriminators between fields designated "male dominated" and those called "female dominated". Some additional analyses relating to Hypothesis 3 , Self-Esteem and Sex of Field, using the factored BSRI were done to explore some possible interactions between: 1) Self-Esteem and endorsement of the various BSRI Factors 2) Components of the TOTAL POSITIVE SCORE. Details of these analyses are reported in Chapter V. re Hypothesis 4 - The factors of the BSRI were used to further study the relationships betwen Self-Esteem (Inter/Personal and Self Concept of Ability) and sex role orientation. Pearson product-moment correlations had been produced by the LERTAP program^used to score the test, and the resulting matrix for the intercorrelations among all variables including the scores for 71 the new factors, was inspected for significance. Following examination of the correlation matrix, a multiple regression analysis was performed, again using the UBC TRIP, to identify the most important factors relating to the Self Concept of Ability Score. The regression equation took the following general form: Y = b Q + + b 2x 2 + b 3x 3 + b 4x 4 = b 5x 5 + b 6x 6 + b 7x 7 = b 8x 8 + e where Y = Self Concept of Ability x l — Social Desirability h = Self Criticism h = Androgyny h Inter/Personal Self-Esteem h = BSRI Factor I \ = BSRI Factor II X7 = BSRI Factor III X 8 = BSRI Factor IV b0 = constant e error 72 Results of Post Analyses 1. The Androgyny Score a) Computation: The frequency distributions for Masculinity and Femininity scores respectively are shown in Figure 2. On visual inspection the distributions appear to be normal, and thus to qualify as components of a ;£-ratio. Therefore the method of computing Androgyny described by the author of the BSRI is also valid for this populations insofar as it meets the criterion of normality. b) Definition: i) Alternate definition - The re-classification of the 34 subjects designated "Androgynous" by £ ratio into the categories appropriate to their absolute levels of masculinity and femininity is represented in Table V • It will be seen that in addition to the two groups identified by Spence (1975) - the High Masculine, High Feminine,and the Low Masculine, Low Feminine scorers - a small group of androgynous women scored High Masculine, Low Feminine, according to the median split method. The presence of this third group may be attributed to the relatively narrow range of masculinity scores in this, a totally female group. In comparison with males, these women would likely become "Low-low" scorers and the re-classification would more closely resemble that of the mixed group on which the system was first applied. For the present group, the median Masculinity and Femininity scores were 4.89 and 4.71, respectively. i i ) Analysis of Androgynous subjects according to classification by Median Split. The results or the multivariate analysis of variance set forth in FIGURE 8 Frequency Distributions of Masculinity and Femininity Scores 12 3. 4 5 Masculinity Scores 12 11 10 9 CO +-> o (J o <L> •<—> -Q 7 zs CO 6 o 5 i~ be 4 E 3 n = 90 T 1 1 r T 1 1 r ~3~ 4 Femininity Scores I. TABLE V Percentage of Androgynous Subjects in Categories Formed According to "Median Split" Procedure Androgynous S's N - 34 Low Masculinity 41.18% Low Femininity (n=14) High Masculinity 11.76% Low Femininity (n=4) High Masculinity 47.06% (n=16) High Femininity 75 TABLE VI Means, Standard Deviations and F values For Androgynous Groups on Self Concept Variables Androgynous Subjects N = 34 Variable High M-High F (n=14) • x SD High M-Low F In=4) x SD Low M-Low F £n=16) x SD Signifi-cance Test F (2,31) 1. Self Concept of Ability 54.64 7.56 60.75 1.26 61.56 6.49 2. Social Desir-ability 97.86 7.89 106.50 9.33 104.63 8.94 3. Inter/Personal Self-Esteem 344.00 21.69 367.25 15.50 3 9 4 . 4 4 40.42 4. Self Criticism 35.71 5.64 30.25 2.63 36.19 6.67 Multivariate Test F (8,56) = 1.672 p< .1258 76 Table VI reveal no statistically significant differences among the means of the three "androgynous" groups across the set of self-esteem variables. The nonsignificant overall F (8,56) = 1.67, p .126; shows that no discriminant function of those variables could be found to distinguish among the three groups. Since the three types of androgynous scorers exhibit comparable levels of self-esteem, the original definition of androgyny by i-ratios describes a homogeneous group in this respect. Therefore the lack of correlation between the variable "Androgyny" and the self-esteem set cannot be traced to wide within group variations in self-esteem among androgynous women, that might have been revealed by categorizing them according to a median split. Dimensions of the Masculine and Feminine Stereotype Scales  Factor Analysis by Principal Components Method The items introduced into the factor analysis are the 40 items from the BSRI masculinity and femininity scale and are listed in Table VII. J/m_ Intercorrelations among the 40 items appear in Table XXII. Application of a Principal Components procedure resulted in 27 factors, accounting for 39.96% of the total variance. Using the criterion of number of eigenvalues exceeding 1.00, ten factors, accounting for 27.88% of the variance were retained for the varimax rotation. The items loading significantly on these transformed factors; that is, having a magnitude greater than .40 are listed in Table VIII. From this table i t can be seen that while the first four factors are logical groups of attributes, having some common meaning, the remaining factors are both too brief and the items within them too contradictory to be useful. Since the majority of the variance is shared by the first four factors, leaving l i t t l e common variance to the remaining variables, this is an understandable outcome. 77 TABLE VII Identification of BSRI items subjected to Factor Analysis Original BSRI Masculinity Scale Original BSRI Feminity Scale Vari-able # Vari-able # 1. self-reliant 21. yielding 2. defends own beliefs 22. cheerful 3. independent 23. shy 4. athletic 24. affectionate 5. assertive 25. flatterable 6. strong personality 26. loyal 7. forceful 27. feminine 8. analytical 28. sympathetic 9. has leadership 29. sensitive to the needs of others 10. willing to take risks 30. understandi ng 11. makes decisions easily 31. compassionate 12. self-sufficient 32. eager to soothe hurt feelings 13. dominant 33. soft-spoken 14. masculine 34. warm 15. willing to take a stand 35. tender 16. aggressive 36. gulli ble 17. acts as a leader 37. childlike 18. i ndividualistic 38. does not use harsh language 19. competitive 39. loves children 20. ambitious 40. gentle 10 3 5 78 FIGURE 3 10. PRINCIPAL COMPONENTS ANALYSIS SCREE TEST (Graph of Eigenvalues against Number of Eigen values) i i i i i i i 35 40 i i i i i i i i i i l t I— i i I — n — n — i i—i i i 1 5 10 15 ZO 25 I I i l 30 Number of Eigenvalues t TABLE VIII Principal Components Solution Transformed Factors Vari-ables I II III IV V VI VII VIII IX X Masc. 1 .058 .080 .642 -.094 -.037 -.164 -.430 -.158 -.020 .171 2 -.088 .674 -.180 .067 .105 .033 -.209 -.044 .095 .090 3 .126 .252 .662 -.091 .286 -.208 -.065 -.227 .135 -.099 4 -.006 .167 .063 -.011 .106 .127 .167 -.770 -.192 .011 5 .134 .641 .205. -.358 -.105 -.067 -.164 -.052 .086 -.084 6 -.053 .742 .198 -.136 .153 -.104 -.041 .033 -.161 -.081 7 .060 .840 .073 .076 -.034 -.011 .043 -.007 .123 -.040 8 -.069 .112 .129 .067 -.021 .027 -.761 .082 -.136 -.148 9 -.293 .673 .150 -.205 -.011 .082 -.089 -.270 .036 -.075 10 -.123 .304 .357 -.022 -.1.26 .132 -.221 -.435 -.02 -.317 11 -.132 -.002 .343 -.255 .195 .037 -.215 -.147 -.610 -.192 12 -.365 .074 .789 .005 .038 -.022 -.008 -.013 -.129 -.059 13 -.072 .725 .136 .238 -.108 .293 .166 .002 -.069 -.095 14 .043 .289 -.052 -.034 .026 .786 -.005 -.177 -.037 -.119 15 -.400 .555 .108 -.053 .040 .093 -.171 .183 -.327 -.234 16 -.171 .742 .059 -.026 -.065 .247 .274 -.022 -.153 -.101 17 -.262 .619 .567 -.128 -.095 " .180 -.121 -.289 -.104 -.076 18 -.198 .130 .728 -.079 -.121 .243 .028 .166 -.127 -.178 19 -.126 .141 .099 .148 -.131 .066 -.144 -.024 -.165 -.793 20 -.389 .190 .168 .081 .067 .314 .073 -.182 .247 -.507 Fern. -.085 21 -.327 -.299 -.012 .072 -.391 -.002 -.109 -.503 .149 22 -.553 .272 .249 -.012 -.050 -.101 .240 -.194 -.281 -.187 23 -.071 -.329 -.037 .777 .043 -.016 -.094 .011 .164 -.070 24 -.560 .176 -.022 -.140 -.041 -.156 .280 -.136 -.233 -.308 25 -.269 .285 -.060 -.020 -.416 -.210 .255 -.370 .154 -.204 26 -.515 .105 .063 .151 -.161 .500 .023 .057 -.054 .005 27 -.417 .012 .337 .315 -.093 -.452 .111 -.007 -.196 -.010 28 -.711 -. 088 .090 .381 -.159 .135 -.060 .129 -.008 .038 29 -.746 .096 .100 .190 .079 .020 -.095 .014 -.216 .056 30 -.762 .065' .117 .172 -.009 .139 -.154 -.240 -.044 .138 31 -.791 .254 .053 .150 .044 -. 050 -.077 , .131 .017 -.077 32 -.764 -.036 -.017 .131 -.345 -.078 -.038 -.151 -.081 -.031 33 -.324 -.476 .037 .229 .036 -.086 -.165 .011 .286 -.337 : 34 -.744 .108 .200 -.052 -.006 .045 .117 -.187 .179 -.067 35 -.771 .091 -.081 -.201 -.086 -.078 .072 -.074 -.129 -.286 36 -.074 .006 -.015 .057 -.912 .042 -.022 .044 -.013 -.058 37 -.122 .183 -.248 .701 -.296 -.024 .101 -.118 -.029 -.109 38 -.381 -.080 .047 .481 .133 -.070 -.319 .152 .047 -.086 39 -.509 .105 -.001 .019 -.237 -.003 -.145 -.207 -.494 .022 40 -.794 -.104 .033 -.063 .034 .060 -.043 .087 .134 -.250 * -Those factor coefficients greater than .40 in absolute value are statistically significant and have been underlined. 80 Reduction of the number of uninterpretable factors by assigning specific numbers for rotation resulted in more meaningful patterns. Solutions of four, five, and six factors each had meaning, but since the last two, three, and four factors respectively were notably less consistent than the first two, further clarification seemed necessary. Factor Analysis by Image Model The Image model, said to be a more precise method (Hakstian, 1972) was employed to further clarify the factor structure explored above. Of the 40 factors extracted by this method, 27 were retained for varimax transformation, according to the criterion of eigenvalues greater than 1.00. To reduce the number of meaningless factors, specific numbers of factors, decided upon by inspection of the Scree plot (Fig. 4) were again assigned for the rotation. The results are comparable to those of the Principal Components model, with a reversal in the ordering of the first two factors. From a comparison of the solutions generated by both Principal Components and Image solutions, for which the details appear in Table IX and Table X, a single solution was derived. Statistical considerations, such as the consistent clustering of items, magnitude of item loadings on various factors, and omission of items consistently failing to reach significance; and subjective judgments, such as conformity to a common construct as suggested by the items of highest loadong on each scale, led to this composite solution. A total of 33 items from the original inventory were retained according to these criteria. Table XI presents the correlations of the items with their respective scales, along with the label "M" or "F" to indicate their original designation. Table XII contains the intercorrelations among the four factors, and Table XIII the means, standard deviations and reliability coefficients of the test rescored in terms of the new factors. 30 FIGURE 4 IMAGE ANALYSIS SCREE TEST (Graph of Eigenvalues against Number of Eigenvalues) Number of Eigenvalues 82 TABLE IX Factor Pattern* For Factor Analysis of BSRI Items (Principal Components, Varimax - 4-Factor Solution** - 5-Factor Solution** 4-Factor 5-Factor Variable 1 2 3 4 1 2 3 4 5 1 .014 .007 .689 .048 -.005 -.004 .742 .079 -.051 2 -.005 .644 -.049 .089 .652 -.083 -. 054 .131 .051 3 .076 .170 .696 -.085 -.172 -.021 .772 .025 -.084 4 -.162 .187 .076 -.486 -.198 -.072 .150 -.348 -.414 5 .127 .575 .337 -.262 -.574 .099 .341 -.253 -.063 6 -.029 .675 .366 -.093 -.694 -.207 .335 -.112 .112 7 .047 .796 .044 -.045 -.799 -.051 .103 .022 -.124 8 -.078 .155 .351 .416 -.141 -.058 .385 .375 .054 9 -.303 .650 .246 -.239 -.658 -.321 .269 -.174 -.181 10 -.310 .356 .349 -.155 -.321 -.003 .462 -.003 -.409 11 -.201 .031 .613 -.127 -.007 -.053 .512 -.259 .238 12 -.434 .062 .669 .669 -.007 -.223 .751 -.039 .078 13 -.153 .795 -.040 .035 -.788 .031 .009 .044 -.125 14 .001 .515 -.126 -.051 -.504 .250 -.118 -.030 -.091 15 -.386 .595 .268 .126 -.627 -.261 .196 -.031 .290 16 -.211 .781 -.009 -.196 -.797 -.060 -.041 -.245 -.024 17 -.313 .639 .136 -.217 -.647 -.171 .156 -.194 -.194 18 -.285 .205 .549 .074 -.172 .051 .616 -.052 .126 19 -.344 .253 .117 .111 -.230 -.102 .229 .171 -.306 20 -.479 .312 .047 .038 -.301 -.235 .128 .121 -.311 21 -.493 -.279 -.153 -.188 .323 -.210 -.035 -.001 -.625 22 -.651 .240 .221 .221 -.230 -.511 .216 -.323 -.121 23 -.150 -.265 -.284 .604 .339 .107 -.182 .679 -.133 24 -.623 .140 .041 -.300 -.137 .535 -.023 -.412 -.088 25 -.432 .214 -.202 -.429 -.197 -.212 -.101 -.277 -.575 26 -.521 .251 -.128 .185 -.261 -.179 -.245 .190 .080 27 -.507 -.104 .226 .117 .166 -.353 .251 .084 -.019 28 -.708 -.029 -.112 .409 .079 -.567 -.130 .441 -.003 29 -.700 .102 .119 .246 -.100 -.689 .064 .175 .155 30 -.746 .091 .071 .145 -.063 -.637 .066 .168 -.089 31 -.717 .236 .061 .273 -.229 -.708 -.014 .266 .145 32 -.831 -.049 -.113 .005 .093 -.742 -.093 .083 -.326 33 -.360 -.436 -.032 .334. -.493 .161 .010 .376 -.110 34 -.729 .098 .110 -.077 -.080 -.629 .106 -.132 -.153 35 -.775 .076 .023 -.175 -.075 -.752 -.049 -.260 -.066 36 -.274 .034 -.309 -.130 .021 .055 -.201 .056 -.478 37 -.290 .236 -.521 .251 -.182 -.050 -.393 .433 -.396 38 -.355 -.055 .026 .609 .104 -.274 .057 .619 .094 39 -.596 .111 .078 -.099 -.115 -.534 .069 -.055 -.147 40 -.731 -.071 .045 .151 .101 -.697 .019 .123 .048 Those factor coefficients greater than .40 in absolute value are statistically significant and have been underlined. ** Factors I, II, and IV have been reflected 180° for purposes of interpretation. 83 TABLE X Factor Pattern* For Factor Analysis of BSRI Items (Image Analysis Varimax Transformation - 4-Factor Solution** - 5-Factor Solution** 4-Factor 5-Factor Variable 1 2 3 4 1 2 3 4 5 1 .077 .010 .651 -.051 .102 .100 .752 -.065 .091 2 -.684 .087 -.114 -.116 -.652 .033 -.110 -.138 -.114 3 -.093 .023 .676 -.033 -.059 .086 .757 -.039 .042 4 .024 -.359 .013 .374 -.113 -.155 .131 .247 .335 5 -.424 -.004 .216 .280 -.465 .029 .223 .241 .021 6 -.679 -.005 .268 -.068 -.600 -.135 .192 -.015 -.236 7 -.875 .100 -.050 -.068 -.880 .146 .052 -.168 .013 8 -.199 .105 .270 -.283 -.122 .079 .312 -.265 -.079 9 -.468 -.408 .125 .194 -.524 -.325 .181 .126 .134 10 -.169 -.149 .260 .138 -.238 .035 .420 .021 .255 11 .162 -.076 .504 .137 .224 -.214 . 343 .280 -.206 12 .077 -.112 .756 -.148 .166 -.138 .758 -.069 -.071 13 -.849 .127 -.135 -.001 -.882 .202 -.017 -.128 .061 14 -.390 .108 -.190 .241 -.461 .161 -.160 -. 160 .063 15 -.583 -.038 .165 -.105 -.470 .276 -.016 .031 -.373 16 -.731 -.058 -.141 .187 -.766 -.076 -.158 .140 -.041 17 -.453 -.330 -.014 .294 -.544 -.220 .051 .196 .175 18 -.080 .099 .580 .004 -.008 .020 .518 .089 -.161 19 -.184 -.122 .055 -.043 -.196 -.060 .128 -.091 .087 20 -.247 -.229 -.004 -.064 -.273 -.134 .102 -.139 .143 21 .440 -.538 -.196 .159 .270 -.191 .066 -.044 .583 22 -.089 -.524 .179 .064 -.091 -.529 .152 .092 .033 23 -.041 .393 -.200 -.589 .007 .504 .022 -.695 .113 24 .024 -.586 -.027 .144 .012 -.648 -.136 .209 -.203 25 -.043 -.504 -.261 .315 -.211 -.249 -.090 .135 .433 26 -.320 -.004 -.213 -.192 -.278 -.077 -.238 -.178 -.112 27 .079 -.176 .229 -.264 .138 -.179 .265 -.234 -.003 28 -.113 -.223- -.115 -.591 -.017 -.251 -.044 -.577 -.032 29 -.171 -.345 .094 -.440 -.046 -.487 .030 -.331 -.192 30 -.069 -.452 .025 -.324 -.027 -.432 .087 -.324 .061 31 -.365 -.288 -.001 -.546 -.220 -.454 -.059 -.445 -.238 32 .121 -.724 -.165 -.253 .106 -.625 -.062 -.296 .220 33 .372 -.066 .039 -.330 .417 -.033 .107 -.321 .060 34 .015 -.577 .085 -.093 .024 -.563 .094 -.075 .072 35 .104 -.789 -.068 .019 .119 -.889 -.203 .120 -.053 36 .057 -.176 -.302 .131 -.067 .047 -.123 -.034 .358 37 -.419 .073 -.507 -.311 -.471 .266 -.251 -.511 .268 38 -.126 .090 .056 -.594 .001 .024 .106 -.556 .126 39 -.024 -.509 .002 -.078 -.022 -.498 .008 -.068 .063 40 .106 -.505 .016 -.325 .201 -.640 -.082 -.203 -.145 *Those factor coefficients greater than .40 in absolute value are statistically significant and have been underlined. r*Factors I, II, and IV have been reflected 180° for purposes of interpretation 84 Description of Factors Factor 1. - This factor is characterized by variables such as V7 — forceful (M) VT3 - dominant (M) and VI3 - aggressive (M) The items on this scale suggest a socially powerful, confident personality who leads rather than following others. It was named Dominance to reflect the interpersonal focus and controlling stance. It is made up of ten Masculine and one Feminine item which was negatively loaded. Factor 2. - Highest loadings on this factor were consistently given to the following items: V32 -- eager to soothe hurt feelings (F) V35 — tender (F) The second factor emphasizes feminine qualities of emotional sensitivity and responsiveness to the needs of others. It is named Nurturance for the warm, supportive orientation to others suggested by these ten items. It is correlated significantly with the Dominance Factor {n. = .286, p< .01). Factor 3. - The name for this factor was determined by the tone of these items, which gave i t the highest loadings: VI2 — self-sufficient (M) V 3 -- independent (M) A sense of self-confidence and personal strength, suggestive of a functioning adult is common to the items on this factor. It was named Independence, for its focus on the sense of self rather than on social confidence as suggested by the Dominance scale. However, there is a moderate correlation between the two. [n. = .326, p< .01). 85 Factor 4. - This factor is statistically less reliable than the others, and the items composing i t are those characteristics of the Femininity scale which do not share the "nurturing" sense. Terms such as : V 23 -- shy (F) V 37 -- childlike (F) V 36 — gullible (F) are typical of this group of six items. They suggest a tendency to be timid and to externalize the locus of control, for which reasons i t has been called Passivity. Its strongest correlation with another factor is negative: as Passivity increases, Independence decreases at the rate of [n. = -.320, p< .01). With the Nurturance scale i t covaries to a significant degree; [ft = .220,p <^ .05), but differs from that scale in the way it relates to the two masculine scales. Nurturance does not correlate with Independence, but does correlate with Dominance; while Passivity correlates significantly in a negative direction with Independence but does not correlate with Dominance. TABLE XI Item Statistics for Factors of the BSRI Factor I Factor II Factor III Factor IV Dominance Nurturance Independence Passivi ty correlation correlation correlation correlation item with factor item with factor item with factor item with factor 9. has leader- 35. tender (f) .676 3. independent (m) .657 37. childlike(f) .484 ship abilities (m) .714 32. eager to soothe 1. self-reliant(m) .627 28. sympathe-7. forceful (m) .699 hurt feelings (f) .674 tic (f) .333 12. self-5. assertive (m) .698 31. compassionate .597 sufficient (m) .632 23 shy (f) .321 16. aggressive (m) .683 29. sensitive to the 18. individual - 36. gullible (f) .247 needs of others (f) .596 istic (m) .474 6. strong 38. does not use personality (m) .678 30. understanding (f) .567 10. willing to take harsh langu-risks (m) .430 age (f) .164 13. dominant (m) .668 34. warm (f) .558 11. makes decisions 25. flatterable(f)- .01 17. acts as a 40. gentle (f) .541 easily (m) .385 leader (m) .648 39. loves children (f) .471 15. wi 11 i ng to ta ke a stand (m) .571 22. cheerful .470 2. defends own 24. affectionate (f) .460 bel ief s 33. soft-spoken (f) .470 14. masculine (m) .323 BSRI items omitted from solution: 4. athletic 8. analytical 19. competitive cR 20. ambitious 21. yielding 26. loyal 27. feminine TABLE XII Intercorrelations Among the BSRI Factors Factor 1. 2. 3. 4. Dominance 1.00 Nurturance _ 29** 1.00 Independence .33** .16 1.00 Passivity -.06 .22* -.32* 1.00 *fi~> .22 is significant at the^.OS level .26 is significant at the°=.01 level Leading decimal, points are omitted. 88 TABLE XIII Internal Consistency Statistics For Factors of the BSRI Dominance Nurturance Independence Passivi ty No. of Items 11 10 6 6 mean : 48.33 53.48 31.01 24.99 standard deviation 8.60 7.47 4.87 4.73 Reliability 0.82 0.85 0.77 0.47 SEM 3.49 2.77 2.13 3.13 Range 41 35 22 29 89 Phase II of Hypothesis Testing Utilizing the results of the preceding Analysis of the Bern Sex Role Inventory, the original hypotheses were again subjected to statistical tests. The previous responses to those research questions are clarified by the following additional information: re. Hypothesis I : Sex Of Field And Self Concept Table XIV shows the means and standard deviations for the six career groups on the self-esteem variables observed previously and the four factors into which the BSRI has been broken down. The multivariate test for significance of differences among those means, summarized in Table XV, shows that the reorganization of BSRI items into factors does not alter previous conclusions: No significant differences emerged between women in male and female dominated fields on any of the BSRI factors (Independence, Dominance, Nurturance or Passivity), as indicated by the nonsignificant overall F. (40,338.43) = .9999, p<.4956). Also as before, no discriminant function of the set of variables could be identified as a significant differentiating factor. The results for Hypothesis I may be summarized as follows: Hypothesis I: In terms of self concept, women in traditionally feminine fields do not differ significantly from those in traditionally masculine fields. No group is: a) more sex typed, in either masculine or feminine directions, as measured by the BSRI Androgyny Score; b) more highly sex typed in terms of the level of endorsement of either masculine or feminine traits, irrespective of the relationship between levels of those two types of traits for an individual. TABLE XIV Means and Standard Deviations on BSRI Factors and Self-Esteem Variables for 6 Career Groups Variables Dominance Nurturance Independ- Passivity Self Concept Social I/P Self Self ence Ability Desirability Esteem Criticism Group Teaching X 49.53 56.67 31.20 26.07 58.33 99.67 348.07 36.00 n = 15 S. D. 7.32 6.67 3.75 3.01 6.26 9.45 27.70 5.55 Nursing X 48.87 55.07 32.00 24.20 59.47 104.23 355.67 35.87 n = 15 S. D. 9.40 7.89 4.83 3.67 7.74 8.50 38.52 6.97 Arts X 46.33 51.20 29.67 25.80 55.13 96.33 339.60 36.07 n = 15 S. D. 10.72 5.75 5.58 5.73 8.30 8.61 24.65 3.90 Science X 50.47 55.00 32.00 27.00 63.33 101.60 355.27 35.07 n = 15 S. D. 6.91 6.69 4.63 4.16 6.96 10.25 41.60 6.06 Agric. X 46.87 51.33 30.33 23.07 52.80 100.93 350.40 34.20 n = 15 S. D. 7.88 7.42 4.59 4.38 7.01 10.53 44.05 5.48 Math. X 47.93 51.60 30.87 23.80 60.13 102.20 357.40 33.00 n = 15 S. D. 9.47 9.20 5.91 6.13 6.63 7.28 20.41 4.55 TABLE XV Univariate and Multivariate Analysis of Variance for Career Groups on BSRI Factors & Self-Esteem Variables Source UnjfV?£Jitfi Dominance 0.4964 P .7781 Nurturance 1.5716 P .1770 Independence 0.5279 P .7546 Passivity 1.6116 P .1659 Self Concept of Ability 4.0840 P .0023 Social Desirability 1.4005 P .2325 Inter/Personal Self-Esteem 0.5697 P .7230 Self Criticism 0.7421 P .5942 Multivariate Test: F (40,338.43) = 0.9999 p < 0.4756 a 92 c) higher in self-esteem, either in the personal and social areas measured by the Tennessee Self Concept Scale, or in the academic achievement area, measured by the Self Concept of Ability Scale. re. Hypothesis II: Relationships of Sex Typing to Self-Esteem The matrix of correlations between the Self-Esteem variables (Inter/Personal Self-Esteem, Self Concept of Ability, Self Criticism, and Social Desirability) and the four newly defined factors of the Bern Sex Role Inventory (Independence, Dominance, Nurturance and Passivity) appear in Table XVI . While the conclusions concerning the Androgyny variable were not altered by the results of additional analyses (see p.91), those concerning the original BSRI scales are modified by the intercorrelations discovered between their component factors. These correlations will be discussed separately in relation to the two areas of self-esteem tested; the Inter/Personal area, and the achievement area. i) Inter/Personal Self-Esteem (TSCS Total Positive Score) Significantly related to scores on the self-esteem variable are the scores on those masculine items denoting Independence (^ =.380, P <.01), and Dominance (^ =.250, p<.05). The two feminine factors now prove to be correlated with the criterion where the Femininity scale as a whole did not. The reason for this is shown by the opposite direction in which the two feminine factors relate to self-esteem: the negative influence of high levels of Passivity (^ = -.285,p<.01) is balanced by the positive effect on self-esteem of the Nurturance factor (Ji =.279, p<.01). This same factor has a negative effect also, upon the Social Desirability variable; the variable whose major importance in determining level of self-esteem was reported in Chapter IV. This correlation explains the zero order correlation between TABLE XVI Intercorrelation Matrix For Combined Groups On Factored BSRI And Self-Esteem Variables Variable 1 2 3 4 5 6 7 8 Dominance 1.00 Nurturance .29** 1.00 Independence .33** .16 1.00 • Passivity -.06 .22* -.32** 1.00 Self Concept of Ability .38** .35** .22* .13 1.00 Social Desir-ability .13 .37** .39** -.43** .26* 1.00 Inter/Personal ; Self Esteem .25* . 28** .38** > 29** .24* .63** 1.00 Self Criticism .01 -.19 -.03 .02 -.02 -.23* -.37 1.00 * k l > .22 is significant at the ^ .05 level **|*| > .26 is significant at the « .01 level Leading decimal points are omitted. 94 Social Desirability and the original Femininity scale which has been discussed in the preceding chapter. i i) Self Concept of Abi1i ty The subscales of the BSRI identified by the factor analysis are correlated with the Self Concept Ability, as the results appearing in Table XVI indicate. This pattern shows why previously the Masculinity scale had emerged more significant to variance in SCA than the Femininity scale. Both the scales composed of items originally designated as masculine correlate significantly with the Self Concept of Ability scores; (t. = .382, p <.01) a n c j (^.= .217,p<.05) respectively. Only one of the scales made up feminine items correlates significantly with SCA scores; namely the Nurturance scale (t=.348,p<.01). Using stepwise regression procedures to control the interdependence of these subscales, particularly for that of Nurturance with Dominance, and Independence with Dominance, the above influences on Self Concept of Ability were qualified in the following manner: Dominance is the strongest single predictor of Self Concept of Ability scores, accounting for 14.60% of the variance in the dependent variable. When the effect of Dominance upon the remaining factors is removed, Nurturance s t i l l adds a statistically significant increment to the total variance in SCA.( .0618, p<.05). Independence, though correlated with the Ability Self Concept, does not contribute enough unique variance to that variable to emerge as significant. Thus the conclusions reached in Phase I of these correlational analyses should be modified to state that: when the original BSRI scales are discussed, Masculinity is more important to the variance in the ability self concept than Femininity; however, when the four dimensions of the original BSRI are in question, i t becomes apparent that only one aspect of Masculinity, namely TABLE .XVII Stepwise Multiple Regression Analysis Self Concept of Ability BSRI as 4 Factors Self Concept of Ability Order of Predictor Variables Variable Symbol & Name R2 Increment in R2 1. X-| Dominance .1460 .1460 2. X2 Nurturance .2078 .0618 Variables listed are significant to the oc = .05 level. 96 Dominance, is important in that relationship. And, when the two dimensions of Femininity are free to vary independently of each other, Nurturance emerges as a significant contributor to the variance in Self Concept of Ability scores; more important, in fact, than the second masculine factor called Independence. The results pertaining to Hypothesis 4. may be summarized as follows: Hypothesis 4: No linear relationship exists between degree of androgyny measured by the BSRI Androgyny Score, and either level of Inter/Personal Self-Esteem^or Self Concept of Ability. Response to Hypothesis 4: a) No relationship was found between Androgyny and either measure of self-esteem, b) Relationships were found between its components, Masculinity and Femininity, and both measures of self-esteem. Relationships beteen subscales of Masculinity and Femininity as identified by a factor analysis also showed significant relationships with the self-esteem variables. With respect to: i) Inter/Personal Self-Esteem: Of the two original BSRI scales, only the Masculinity scale proved influential. As levels of Masculinity increase, so does Self-Esteem. Of the four BSRI subscales, three correlate positively with self-esteem, but one correlates significantly in the negative direction. The negative effect of this factor, made up of feminine items,offers an explanation for the zero order correlation 97 .observed for the femininity scale taken as a whole. ii ) Self Concept of Ability: Masculinity and Androgyny were identified as important influences on this aspect of self-esteem. However, when the four separate dimensions of the sex role stereotypes are considered, the Dominance traits within the Masculine stereotype appear most important, followed by that group of feminine items termed Nurturance. 98 CHAPTER VI DISCUSSION The purpose of this study was to explore two major issues: 1) The degree to which women's choices of fields of specialization leading to a career are related to two dimensions of self concept: sex role stereotyping and self-esteem. 2) The relationship between degree of conformity to conventional definitions of sex roles and level of self-esteem. For both these questions, self-esteem was evaluated in terms of personal and interpersonal functioning, and in terms of academic achievement attitudes involving the self. The following discussion centers on results obtained from an empirical investigation of these two issues. The results relating to the first of the two major issues, treated by testing Hypotheses one, two and three, will first be discussed in their entirety. It is followed by discussion of the issue addressed by Hypotheses four and five, and will include details of those discrepancies which stimulated the second phase of analysis described under "Post Analyses". Results of this second phase, which included an analysis of the Bern Sex Role Inventory, will then be used to clarify the discrepancies found in responding to Hypotheses four and five. Conclusions will be drawn from the present findings where those findings prove to be statistically significant, where they are not, explanations based on other relevant research will be proposed. 99 Congruence Between Self Concept and Occupational Stereotype  Hypothesis 1 - Sex Typing and Masculinity/Femininity of Career Choice The portion of the study dealing with sex typing of occupations and how it relates to the self concepts of women who choose them yielded unexpected results. Contrary to expectations based on previous research into sex typed occupational choice, no consistently feminine sex role orientation characterized women in traditionally feminine fields, and no consistently masculine orientation characterized those in traditionally masculine fields. This applied to all measures of sex typing used in the study: the Androgyny score; measuring preference for the masculine over the feminine sex or the feminine over the masculine role; the Masculinity and Femininity scores measuring absolute levels of endorsement stereotypic qualities, and those scores obtained for the two dimensions of each stereotype (see Post Analyses). Hypothesis 2. - Self Concept of Ability and Masculinity/Femininity of  of Career Choice It was evident from the results of the multivariate analysis that a significant univariate F emerged for this variable. The source of the difference was identified by the Tukey procedure for Multiple Comparisons (Glass and Stanley, Chapter 16, 1970) as the higher mean Self Concept of Ability Score of the Science women relative to that of women in either -Arts or Agriculture. This table, XXII, is presented in Appendix B. Because of the results of the multivariate test were nonsignificant, this difference among means must be interpreted with caution. It is however, .consistent with the tendency for both higher achievement motivation scores and higher grades (Korman, 1973) to distinguish women in Science from their liberal arts peers (Sundheim, 1963, Kriger, 1972). 100 Hypothesis 3. - Inter/Personal Self-Esteem and Masculinity/Femininity  of Career Choice-Findings of the present study revealed no statistically significant differences in social and personal self-esteem between women in traditionally feminine fields and those in fields generally considered masculine. Hence the hypothesis that women who choose traditionally feminine occupations do so out of a need to raise their.self-esteem, rather than out of a need to achieve, was not supported. The needs motivating women to choose careers as Teachers, Nurses, or in the Arts fields; or to choose careers in Science, Mathematics or Agriculture have not been identified by these findings. It is clear, however, that neither level of self-esteem nor level of sex typing distinguishes.women choosing unconventional career paths from those following the traditionally feminine occupational roles. Several explanations for the failure, in this study, to find distinctive profiles for any of the six career groups are possible. Those explanations pertaining to sex role conceptions will receive attention fir s t ; to be followed by explanations related to self-esteem in, the areas of social/personal functioning and achievement. Sex Role Stereotype and Career Choice 1) Sampling Factors y The first potential source of deviance from expectations is in the sample itself. Conclusions from previous research, on which the predictions were based, would not be generalizeable to the present sample were it found not to represent the same population. Both cultural and historical factors were considered. The studies on which the predicted differences were founded all originated in the United States with women in colleges and universities. The possibility 101 that Canadian stereotypes might vary from those prevailing in the United States was therefore worthy of attention. However, the possibility of significant cultural differences seems unlikely in light of information gathered in the 1968 Royal Commission Report on the Status of Women in Canada. Surveys revealed that "traditional views on their status and role... are held even among highly educated women, including college and university students' (p.13). The report details the beliefs adhered to as those that suggest men prefer women with l i t t l e ambition for a career, that many traits are gender linked, and that mothers should stay at home with their young children. Beliefs such as these are reflected in career aspirations which exclude, for the vast majority of women, professional or highly responsible positions (p.14). Aspirations tend also to exclude fields such as medicine, the pure sciences, and commerce; where women are in the minority (p.9), as they are in the United States.. In general the report tends to confirm the similarity of Canadian and American women in conformity to the imposed stereotypes. The possibility also exists that a shift in attitudes since the time of this report and the time at which previous research was conducted may have occurred in response to the Women's Movement (Kravetz, 1976). The lack of conclusive information on this point make this possibility difficulty to assess. There is, in addition, the possibility that the observed discrepancies may result from instrumentation factors. 2) Instrumentation The predictions for differences in.sex role orientation according to sex of field were based on research which focussed with varying degress of directness, on feminine socialization. The work of Horner (1968), Leyland & Lozoff (1969), Hennig (1974), Almquist & Angrist (1971) and others cited in Chapter I, implicated the female role as an influence upon variables such as family 102 background, need for achievement, locus of control and parental modeling. The importance of women's own attitudes to their roles was emphasized by Hawley (1972), and Thurber (1976). The precedent for measuring women's self-perceptions in terms of stereotypic characteristics was set by Cowan and Moore (1971), and it is this method of direct self concept measurement as opposed to attitudes or antecedent conditions, which was adopted in the present study. The discrepancies between the present findings and those based on sex role attitudes can be explained with reference to the recent work by Spence, Helmreich, and Stapp (1975). To determine the extent to which self concepts are influenced by societal standard, the authors used a revised version of the Rosenkrantz Sex Role Stereotype Questionnaire to compare men and women's concepts of themselves with their concepts of the roles. They found women's self concepts generally to be less typically feminine than their images of the typical woman; that is, they tended not to distort their self perceptions in the direction of the stereotype to a significant degree. It is possible, therefore, that in the present study, the use of a measure of "self" concept rather than "stereotype" concept, could result in a reduction in the variance required for strong correlations. The "self" profile would tend to be less extreme than one based on opinions about the typical woman or about women in general. Results differed also from those of Cowan and Moore (1971) who did use self concept ratings. The instrument employed in their study was the Rosenkrantz scale, which is, like the BSRI, a self descriptive measure. In its original form, the form in which i t was used by Cowan and Moore, i t differs from the BSRI in an important way: i t measures masculinity and femininity in opposition while the BSRI allows for endorsement of both 103 stereotypes simultaneously. It can be seen from the description of the BSRI included in Appendix A, that the scores showed low correlations with those of bipolar scales such as the CPI M/F scale and the M/F scales of the Guilford Zimmerman Temperament Survey. To determine whether the bipolar nature of the Rosenkrantz scale places i t in a category equally as dissimilar to the BSRI as these scales and would require a study focussed on testing the concurrent validity of both instruments. Until this is determined, i t cannot be assumed that the present sample differs sufficiently from that tested in the Cowan and Moore study to make alternate interpretations of the discrepant findings conclusive; 3) Variations Within Career Groups Accepting for purposes of this discussion the limitations to generalizability that the introduction of a new instrument imposes, certain information is provided by the BSRI about the present sample, in terms of the concepts of masculinity and femininity measured by that scale. The statistically nonsignificant results of the multivariate analysis indicate that the variance within each of the career groups is sufficiently great to equal the variance among their means. This could be interpreted to mean that among the women in the masculine fields, who would be expected to have a more masculine profile, were a sufficient number with highly feminine scores to moderate this profile. A similarly heterogeneous group were found in the traditionally feminine fields. This unexpected diversity within groups can be tentatively explained by some factors whose influences was not taken into consideration in the present study. Creativity, for example, has been shown to affect the degree of femininity expressed by female mathematicians. By comparing women considered highly creative in the field of mathematics with a group working 1 •in the same field but not considered to be outstandingly creative, Ravenna Helson (1971) discovered that the more creative group described themselves, on the M/F scale of the MMPI, as more feminine than their less creative counterparts. A previous study (Helson, 1968) suggested that creativity may have an androgynizing effect on women in a variety of fields; such as Mathematics, English, Arts and Social Sciences. The creative women differed from other women in their fields by possessing not only an inner orientation and receptivity to emotional and environmental stimuli, but the ability to direct and order their cognitive processes toward autonomous self-expression. This is consistent with Torrance's (1962, cited by Howe, 1974) statement that the creativity personality is relatively androgynous. Direct measurement of the tendencies to be either masculine, feminine, or androgynous could help to clarify the extent to which this factor may neutralize the stereotypes associated with fields of all kinds, and whether it may have been relevant to the diversity observed within the present sampl Some subjective data gathered from subjects in the masculine fields at the time questionnaires were collected provided additional clues as to the wide range of scores, at least within this group. Masculine scores may be a product of what Hochschild (1974) called a subtle "de-feminization". (Hochschild, p. 196, 1974). Subjects expressed satisfaction with their minority status within the field. The "partial acceptance" they experienced did seem to be an advantage to them, and went along, as Hochschild suggests, with greater affiliation with male than with female peers; and with the cutting of social and psychological ties with other women. Some women in the present sample expressed a disapproving attitude toward feminine characteristics; an attitude which could, as Hochschild suggested, stem from the desire to reject the negative stereotype of women with which they are 105 faced in the masculine fields they have chosen. Alternatively, other women appeared to deal with their minority status by downplaying their intellectual abilities. By behaving in a shy, unobtrusive manner, they avoided attracting attention to themselves. They expressed concerns over appearing masculine; as successful women in unconventional fields are often seen to be. It may be these women whose profiles formed the feminine end of the spectrum. The individual responses that are possible among women in this minority position are varied, and could account for a wide range of sex role orientations. Another conceivable explanation for the variety of personal stereotypes within each group stems from Korman's (1967(a) and 1967(b) vocational theory involving self-esteem. As with any other self concept variable, mismatching according to sex role stereotype could result from the distorted perception noted to accompany low self-esteem. The wide variance within groups on the self-esteem variable, to be discussed in the following section, does allow for this possibility. The analyses described in "Additional Recommendations for further study" at the end of this chapter, were carried out in an effort to determine the existence of an interaction between sex role identity and self-esteem. Although no definitive answers could be obtained due to the small number of subjects in the high and low self-esteem categories of each group, the results suggest an area that warrants further study. Self Concept of Ability and Masculinity/Femininity of Career Choice Although a trend suggesting that women in Science do have greater confidence in their academic abilities than women in Arts and Agriculture did emerge, the question which arises from these findings is why more significant differences did not emerge involving all of the career groups. The research cited in the literature review which did find differences in attitudes toward 106 achievement, between women in masculine and feminine career groups, found those differences by using scales directly measuring either achievement motivation,(and the alternate motivation to avoid success.) or indices of actual achievement. That the Self Concept of Ability scale measures the self concept with regard to achievement, rather than the need to achieve, may be a sufficiently important difference to result in the deviation from predictions. Secondly, although the link between Self Concept of Ability and actual achievement has been assessed (Brookover et al., 1962); and school achievement has in turn been linked with sex of field (Korman, 1973); Self Concept of Ability and sex of field have not been previously correlated. Potentially useful in clarifying the present set of findings would be a study of the relationship between the 'achievement motivation" and the achievement self concept as measured by the Self Concept of Ability Scale. An understanding of this relationship could help to clarify the apparent superiority of the motivational measures in predicting masculine vs feminine career choices. The assumption on which predictions in the present study were based was that the two constructs were closely related. It is possible that, instead, a high achievement motivation may be characteristic of the person who is slightly dissatisfied with his achieving self and hence has an ability self concept that is low in relation to his motivation. In terms of need theory, dissatisfaction would tend to place the achievement need in a position of priority on the individual's hierarchy of needs, resulting in a strong achievement motivation. A replication of the present study using both these variables could clarify both the relationship between the achieving self concept and achievement motivation, as well as their relative strengths as determinants of career choice. 107 Inter/Personal Self-Esteem and Masculinity/Femininity  of Career Choice 1) Relationships Among The Variables Contrary to predictions, women in the traditionally feminine fields were not lower in self-esteem according to their Total Positive scores on the Tennessee Self Concept Scale than those in the masculine fields. Drawing on the information available from Coopersmith's (1967) work, i t was expected that the nonconformity evident in an unconventional career choice, would be an indicator of a relatively high level of self-esteem. Conformity to traditionally feminine occupational roles was expected to coincide with a relatively low self-esteem score. The basis for these predictions lay in: (1) studies correlating femininity with typically feminine career choice, (Klemmack & Edwards, 1973, Cowan and Moore, 1971, Hawley, 1972) and (2) those correlating femininity with low self-esteem (cited on p. 21 of this study). In retrospect, it appears that the predictions relied too heavily on the replicability of these previous findings. Because the congruence between personal sex typing and sex typing of field was not demonstrated, any subsequent link between self-esteem and career choice would in any case, make personal sex typing irrelevant to that relationship. In addition, since femininity was not correlated with low self-esteem, as i t had been predicted to be, the basis for a relationship between femininity of field and self-esteem was removed. In view of these findings, the nonsignificant results for "Self-Esteem and Sex of Field" are a possible consequence. 2) Within-Group Differences The fact that particular career groups did not manifest different mean levels of self-esteem indicates that differences in self-esteem within each group were highly variable. Although this within-group variation can not 108 be attributed to a corresponding variation in sex role self concept, since relationships between the sex typing and self-esteem proved to be more complex than expectations would have suggested, certain other explanations are possible: Status Striving. While feminine occupations are generally seen as lower in status than those dominated by men, and low aspirations seen as evidence of low self-esteem; the converse, that high status career striving implies high self-esteem is not necessarily true. It has been observed that "much status striving stems from a need to maintain a favourable self-evaluation" (Coopersmith, p.3, 1967). Despite high aspirations and actual attainments, the- person with a low self-esteem continues to strive for goals beyond his capabilities and, falling short of them, maintains his low self-esteem by keeping a discrepancy between his real and ideal self. (Coopersmith, 1967). By imposing a societal rather than a personal definition (i.e. specific to each subject) of high or low status aspirations,on the career choices women made, the present study obscured any individual trends such as this. It is possible that some subjects may have sought a high status field as an opportunity to raise a low self-esteem rather than to express a high level of confidence in themselves. Creativity. A second possible source of variation is suggested by additional findings from the previously mentioned study on creativity among women in mathematics (Helson, 1971). Her unexpected finding was that both creative and control groups had low scores on measures of social poise and assurance. The creative group exhibited, in addition, a greater number of signs of psychological difficulties than the comparison group. Helson attributes the latter result to the stress of working in a "cross-sex field" (Helson, p.213, 1971). To be creative in such a field, deduces the author 109 from the sum of her findings, involves the degreee of "rejection of outside influence" these women displayed, and "the need to resist conventional patterns (Helson, p.213, 1971). The author reasons that those two tendencies may be particularly strong in female mathematicians because of the strength of social pressures against sex-inappropriate behaviour of which creativity in the field of mathematics is seen as an example. The accompanying signs of psychological difficulty suggest that this conflict is detrimental to the women's self-esteem. If this is the case, and i f potentially creative women were included in the masculine groups, of which mathematics was one, the trends observed by Helson may have a bearing on the low self-esteem found among women in those groups. Comments from some.of the women in the Science and Mathematics groups suggest that for some, participation in these fields does create a conflict. For example: A third year physics student described how her self-confidence was affected by the doubts male peers frequently expressed about her competency. She began to seriously doubt her own competence, both in physics, and in dealing with male physicists. Several other women had become so discouraged by the negative assessments of themselves that they faced in social interactions }that they had transferred from Physics to Mathematics or to Arts. This information parallels that received by Leland' & Lozoff (1969) in their study autonomy and career choice. Discouragement with the rigourous course work, as well as with the negative social pressures,were cited as the reasons for moving out of the Science fields. This group, whose initial goal was a career in the Sciences, also showed the highest level of intrapsychic conflict, and had more difficulties with interpersonal relationships than women in more typically feminine fields. And, despite the fact that no statistically significant differences in no self concept characterized women in the scientific fields, they were, as a group significantly fewer in number at U.B.C. If the conflicts expressed by the women in the present study, and by those in the Lei and and Lozoff (1969) sample, are indeed a deterrent from continuing to pursue career goals in those areas, it would seem that identifying those conflicts and seeking resolutions for them is an important goal. By studying a larger sample, and focussing on differences in sex role self concept and self-esteem between women who do experience high degrees of personal conflict in the atypical, masculine environments and those who do not, could clarify some of these issues. Relationships Between Sex Role Stereotyping  and Self-Esteem Hypothesis 4 Androgyny and Self-Esteem (Inter/Personal and Academic) The second major objective of this study was to explore the relationship of sex role identity to self-evaluation. Previous research repeatedly demonstrated negative correlations between high levels of femininity and self-esteem. Greater mental health was associated with a more masculine self concept. On the basis of these studies, similar differences in self-esteem were predicted to appear for women whose Androgyny scores denoted masculine or feminine sex typing. It was further predicted that those with androgynous, or integrated sex role self concepts would evince higher self-esteem than subjects typed either masculine or feminine. The correlational analysis showed that women on the masculine or the feminine ends of the sex role continuum did not differ in a statistically significant way on the self-esteem variables, either from each other, or from those with a more balanced identity. n r Scoring Procedures The Androgyny Score, as a measure of preferential sex typing, was regarded as a possible source of error variance and examined in detail. The possibility that sampling differences between this and the BSRI norm group resulting in abnormal distributions of Masculinity and Femininity Scores which would have invalidated the use of the Androgyny Score, was tested and discarded. The alternate explanation, that wide variations in self-esteem among the androgynous scorers had reduced the correlations was also considered, and rejected for want of statistical evidence to that effect. Androgyny in a Social Context Speculating as to why self-esteem was not higher among androgynous women than among sex typed women, as Bern's theory suggests i t should be, it seems possible that since sex typing is the norm, and androgyny a relatively new concept, the nonsignificant results for Androgyny should be interpreted in the light of the present social context. In her original studies on sex typing, Bern (1974) noted that androgyny scores were not correlated with Social Desirability; while the masculinity and femininity scores from which Androgyny is computed,were significantly related to social desirability. Perhaps, in a society where the stereotypes masculinity and femininity s t i l l prevail, the merging of the two may not yet be seen as socially acceptable or desirable. Since the results of the correlational analyses in the present study emphasize the strength of the relationship between Social Desirability and self-esteem, a consequence of the low correlation betwen Androgyny and social desirability could be the similarly low correlation between Androgyny and self-esteem. A second possible explanation of the fact that androgynous women did not evidence higher levels of self-esteem than their sex typed peers has to do 112 with the stress involved in personal change, and/or, that which conceivably could result from the lack of social validation received by one with this unconventional identity. Both these processes could have a negative effect on self-esteem while they are being experienced; even though in the long run, the flexibility inherent in the androgynous condition seems to imply the optimum in psychological health, and hence, high self-esteem. The results of a recent study investigating the sources of psychological distress among college educated women (Powell & Reznikoff, 1976) suggest that these two hypotheses regarding the androgynous identity may find support. Contrary to expectations, women having what the researchers termed a "self", or contemporary attitude toward sex roles as measured by the revised Fand Inventory, exhibited significantly higher scores for symptoms of psychological problems than did women with "other", or traditional orientations. The authors attributed the problems associated with the non-traditional, or "self" orientation, to the stresses involved in facing cultural role expectations that conflict with the personal goals these women are now pursuing. There is, however, conflicting evidence on this issue. High.social self-esteem was associated with an androgynous identity, and lower self-esteem with a masculine or a feminine identity, in the recent work by Spence, Helmreich and Stapp (1975). Since the instruments used in that study, (the Texas Social Behaviour Inventory to assess self-esteem, and the adaptation of the Rosenkrantz et al., scale to measure sex roles) were not those employed in the present study, differences in instrumentation could partly account for the discrepancies. A comparison of the two sets of instruments would aid in clarifying this point. In the absence of such clarification, the interpretation that conflicts related to changing roles 113 contributed to the failure of androgynous subjects in the present study to exhibit higher levels of self-esteem than their sex typed peers remain tentative. Masculinity and Femininity Scale Scores and Inter/Personal Self-Esteem While the Androgyny score measured the discrepancy between endorsement of the two sets of sex role attributes, and this discrepancy bore no apparent relationship to self-esteem, scores on the separate stereotype scales, measuring level of each kind of stereotyping, produced a number of significant correlations. Because several of these correlations seemed contradictory to the expectations for this study, and to the normative data for the BSRI, the scales were analyzed to determine whether the existence of distinct subscales could account for these apparent incongruities. The correlations in question were the following: 1) Inter/Personal Self-Esteem and Masculinity/Femininity The Masculinity Scale alone contributes positively to Inter/Personal Self-Esteem. The lack of a negative effect associated with the Femininity Scale of the BSRI compared to other Masculinity/Femininity scales may be partly due to the fact that on the BSRI,subjects are not obliged to reject masculine qualities in order to score high on this scale. However, the fact that high levels of femininity according to the Androgyny Score, which would have implied a preference for feminine over masculine qualities, similarly failed to produce a correlation of any kind makes this an insufficient justification. 2) Social Desirability and Masculinity/Femininity According to the normative data for the BSRI,both the Masculinity and Femininity Scales should correlate positively with Social Desirability. As the results in Table III indicate, only Masculinity showed the predicted relationship in this study. When the significant relationship observed between Social Desirability and Inter/Personal Self-Esteem is taken into consideration, both the positive effect of Masculinity and the neutral effect offemininity on Inter/Personal Self-Esteem is understandable. What is not clear is the reason for the lack of relationship between Femininity and Social Desirability. 3) Self Concept of Ability and Masculinity/Femininity Both Masculinity and Femininity were found to correlate significantly and positively with the ability self concept. Because of the statistical relationship of both Masculinity and Self Concept of Ability scores with Social Desirability; a relationship which Femininity does not share, i t was unclear whether the unique effect of Masculinity would be statistically significant. The theoretical importance of clarifying this pattern lies in determining whether or not the link between "achievement" tendencies and a masculine "instrumental" orientation is supported by the evidence in the present study. By means of multiple regression analysis, some of these questions were answered. Masculinity was shown to be the strongest predictor of high academic self image, with the influence of Femininity being felt indirectly, through the Androgyny Score. The lack of correlation between the Femininity scale and either Social Desirability or Inter/Personal self-esteem was left unexplained by these analyses. At this point i t was hypothesized that the presence of distinct factors within the femininity scale, i f they were to bear equal and opposite relationships to the criteria (Inter/Personal Self-Esteem and Social Desirability), would offer an explanation for these ambiguities. 115 ^ Speculation as to the effect of such factors on the Androgyny score and its correlations with the same criteria arose from this hypothesis. The factor analysis of the Bern Sex Role Inventory's Masculinity and Femininity items resulted in a pattern of four logically and statistically coherent factors which split each of the two original scales into two components. The relationships found between these factors and the self-esteem variables did succeed in clarifying the previous set of results. Feminine Factors of the BSRI and Inter/Personal Self-Esteem An examination of the correlations between each of the two factors composed of feminine items revealed that these two aspects of the feminine stereotype are differentially related to self-esteem as measured by the Tennessee Self Concept Scale. A comparison of the characteristics included in these scales with those from the Rosenkrantz et al., scale provides a rationale for this pattern. The Nurturance Factor is composed of items comparable to those feminine items from the Rosenkrantz scale which made up the "warmth-expressiveness" cluster, and were seen as particularly desirable in women, but also desirable in men (Broverman et al., 1972). That these "female-valued" traits should have a positive impact upon self-esteem is logical, since the ability to communicate and to understand others are "important aspects of psychological well-being" (Woolfolk, p.123, 1976). The effect of the second group of feminine traits which was labelled Passivity is markedly different from that of the Nurturance factor. Seeing oneself as shy, gullible and flatterable coincides with the tendency to place a negative value on oneself for the women in this sample. Qualities similar to these composed a part of the Rosenkrantz scale which was negatively valued for both men and women. Internalizing this negative view would, understandably, have a depressing effect on self-esteem, in view of the close 116 link between self-esteem and social desirability observed in the case of the TSCS. Since Social Desirability appears to be inextricably linked, both practically and empirically, with self-esteem, further discussion of self-esteem and sex role stereotypes must include consideration of this variable. It was previously noted that the Femininity scale as a whole did not show the expected positive correlation with Social Desirability. However, when the two subscales, Nurturance and Passivity, are considered individually, a pattern similar to that with Self-Esteem emerges: The traits denoting Passivity have a significantly negative association with Social Desirability. The mean social desirability ratings for the original Femininity scale was achieved through a balance between extreme high and low ratings on certain characteristics, while the masculine items received relatively homogeneous ratings (Bern, Note 1). It appears that in the present instance, the negative ratings were sufficiently extreme for the items identified as coming from the Passivity factor, to depress the social desirability rating of the entire scale to a greater degree than that occurring for the norm group. This may be a result either of sampling factors or evidence of a change in social values since the test was normed. The results of the correlations discussed above indicate that each of the two feminine dimensions is sufficiently powerful to moderate the effect of the other, thus producing zero-order correlations with both Inter/Personal Self-Esteem and Social Desirability. It is possible that the neutralizing effect of these two factors on the overall femininity score used to compute androgyny may be a partial explanation for the lack of correlations observed in the case of the androgyny scores. Future research might compare the behaviour of an androgyny score based only on the difference between 117 scores on those two factors which do correlate significantly and positively with Social Desirability, as the original BSRI scales did; the Nurturance Factor and the Independence Factor. Although the use of a self-report measure which is as subject to the influence of social desirability as the TSCS has been shown to be, in this and in other research, (Robinson and Shaver, 1973) may be criticized inasmuch as this clouds the assessment of other variables; in the case of self-esteem, this clouding may be unavoidable - even desirable. The ratings by health professionals of the stereotypic traits comprising the Rosenkrantz et al., scales showed that "concepts of health correspond to concepts of social desirability" (Broverman et al., 1972). Items for the BSRI were chosen for their social desirability. Although the desirability of items was gender-linked, i t is the contention of the scale's author that the socially desirable attributes of both sexes, i f integrated, would represent optimal psychological health, thus supporting the interdependence of social desirability self-esteem. If the strong negative value associated by women in the U.B.C. sample with the Passivity traits is an indication of a changing evaluation of the qualities considered typical of women, then these women are presently caught in a stressful situation. Despite a negative evaluation of these qualities which is more severe than that of the norm group, they continue to integrate these qualities, along with the more positive aspects of femininity, into their self concepts. The positive correlation between the two factors which appeared is evidence of this tendency. And this aspect of femininity appears detrimental to self-esteem, according to the present findings. Isolating the effects of the two dimensions of the Femininity Scale could be useful in interpreting some of the results of other research in the 118 area. A recent study conducted by Sandra Bern (1975) on the behavioural flexibility of sex typed versus androgynous subjects yielded some unexpected results. The feminine subjects were expected to be more responsive in a situation calling for nurturing behaviour than either masculine or androgynous subjects, but instead, they demonstrated significantly less responsiveness than the other two groups. The explanation possible in light of the dimensionality of the femininity scale, which might be shown to generalize to other samples, involves the differential effects of the two factors. While subjects high in femininity would have endorsed the nurturing qualities expected to result in a willingness to be playful and nurturing, they would also be high in those feminine qualities signifying passivity. The latter traits could conceivably inhibit the spontaneous behaviour expected by the experimentors of the feminine subjects. Use of separate scores for the two dimensions would allow for the investigation of this possibility. Masculine Factors of the BSRI and Self-Esteem It was noted at the beginning of the discussion on relationships between Masculinity and Femininity scale scores (p. 81 ) that Masculinity scale scores were important to the variance of self-esteem in both the social and the achievement areas. The breakdown of this scale into two components which were labelled "independence" and "dominance", reveals that the relevance of the Masculinity scale score to the two indices of self-esteem is based on a different facet of masculinity in each case. Statistically, i t is understandable that the Independence scale should relate more significantly to Inter/Personal self-esteem than the Dominance scale does, because of the fact that both Independence and Self-Esteem correlate with Social Desirability, while Dominance does not. 119 The psychological explanation may have to do with the acceptability to women of these two types of characteristics. Self reliance, self sufficiency, and the ability to make decisions are not as opposed to stereotypic femininity as tendencies to be aggressive, dominating, and to show leadership, and may be easier to integrate into their self concepts. Clarification could be obtained for this theory by comparing social desirability ratings of men and women for both these groups of characteristic. A sex difference in ratings of the two factors would tend to support the proposed explanation. . The contrast between the two masculine factors in their influence on self-esteem is- evident also in their respective correlations with the ability self concept. The regression analysis showed that only the Dominance factor is significant to the variance of scores on the Self Concept of Ability Scale. The second most important contributor was the Nurturance factor, not the remaining masculine subscale. That these qualities should be relevant to self confidence in abilities suggests that they are either relevant to the achieving tendency itself, which has been associated with the "instrumental" orientation these qualities denote, or, that qualities such as dominance are required of women who deal with the social pressures that often deter women from pursuing intellectual achievement. The lack of correlation between the Dominance scale and Social Desirability implies an ambivalence toward social norms which would be consistent with the second suggestion. The emergence of Androgyny as a significant contributor of the variance in Self Concept of Ability scores is now seen to be a result of the positive effect of the Nurturance factor alone; for neither Androgyny itself nor the Passivity factor made significant contributions to its variance in the second phase of analysis. This finding points to the tendency reported-by Maccoby for female high achievers to be "not only dominant and striving, 120 (characteristics labeled "masculine") but also... more anxious to do things for other people...(behaviours normally classified as "feminine")" (Oetzel, 1961; cited in Maccoby, 1966). While studies with adult women have tended to limit the characteristics predicting achievement motivation to the masculine stereotype, the present findings emphasize the importance of the positive feminine qualities as well. More empirical support for this observation would seem obtainable given the subjective data which is consistent with the androgyny-achievement link. The finding that high self-esteem in social areas is determined by a different set of characteristics than is high self-esteem in the achievement area, suggests.that integration of a wide variety of qualities which are not typically feminine, is necessary for women to enjoy success in both those areas. Recommendations for Future Research Throughout this discussion, issues unresolved by the present study have presented questions for further research. A replication of the present analysis of masculine and feminine career choices employing measures of sex role attributes and attitudes used in the previous research cited would evaluate the effects of instrumentation and sampling on the nonsignificant results. The relationships between creativity and both sex role orientation and self-esteem, have been opened for exploration. With a sample sufficiently large to ensure adequate numbers of subjects in each category, the possibility of interactions between sex role orientation and self-esteem within career groups could be investigated. The effects of role conflict upon the self-esteem of women in male-dominated fields, and the effects upon self-esteem of learning to adapt to new roles would be a useful area of exploration. 121 A comparison between the "need for achievement" and the "self concept of ability" would aid in clarifying their relative importance in predicting career choices. Should the effects of instrumentation upon the discrepancies between the present findings and those of previous researchers into the area of career choice prove to be insignificant, the possibility that other variables, such as cognitive styles and spatial/verbal abilities may be more relevant, should be considered. A revision in the social desirability ratings of the BSRI items should be undertaken. Of interest also would be an item rating in terms of "psychological health", comparable to the rating made on the Rosenkrantz scale, with a view to clarifying the varying effects upon self-esteem of the factors defined in the present study. Identifying the behavioural correlates of the four factors could yield useful information concerning the behaviours associated with the overall stereotypes. A factor analytic study with a large sample should be undertaken to test the applicability of the factor structure proposed in the present study. 122 Additional Recommendations for Further Study  Differences in Sex Role Self Concept among Women in Similar Fields The literature on Self-Esteem and need satisfaction through appropriate vocational choice focusses on the individual with a low self concept. He is the one who tends to make non-need satisfying choices, due to this distorted self-perception, lower expectations of success, or both. For this reason it seems possible that while a woman with high self-esteem may have chosen the field of Nursing to actualize her "nurturing" potential, a similar choice could be made by a person low in self-esteem in the hope of expressing a side of her personality not generally nourished by the Nursing environment. That is, she may have misperceived the opportunities in that field. Or, feeling inadequate in the role of a "nurturing" person but valuing those traits, may hope to downplay parts of herself which are more evident, such as independence or aggressiveness, and through this choice, change her self-image. The result in terms of the profile of Nurses as a group would be a very diverse picture. Were those individuals low in self-esteem to be isolated, within their respective career groups, differences in sex role orientation between them and those reporting high self-esteem could be explored. An interaction between self-esteem and sex role attribution might serve to explain the apparent absence of differences in stereotypes across groups. In the present sample, the top and bottom thirds in Self-Esteem were separated in each group. Only the Low Self-Esteem women, five in each case, were compared in an effort to predict their group membership by scores on the BSRI Factors. Results of the multivariate analysis indicated that no significant effects were associated with those variables. However, the means of the various groups, reported in Table XIX (See Appendix C) 123 suggest some interesting trends. For instance, the Teacher group, whose overall mean on Nurturance was highest of all the groups, include women low in self-esteem who are lower in Nurturance and simultaneously higher in self reported Dominance than the Teacher group as a whole. Due to the small sample size the differences were not tested for significance. Future research using a greater number of subjects could more clearly assess the importance of this interaction. Measurement of Self-Esteem Another approach to the problem of discriminating Masculinity or Femininity of field, i.e., applying self concept theory to vocational choice, is in the instrument used to measure the variable Self-Esteem. Since previous studies involving Self-Esteem and either vocational choice or sex role stereotyping had not used the Tennessee Self Concept Scale, its appropriate-ness was questioned. Just as the dimensions "Masculinity" and "Femininity" appeared to be too global to discriminate levels of self-esteem, while their subscales were not, the Total Positive score might also be too general a measure of self-esteem to distinguish masculine from feminine stereotyping, or masculinity/femininity of field. The possibility, that one.or more of its components might be more sensitive than the global measure was approached in a manner similar to the breakdown of the stereotype scores: Sex of Field and Components of "Self-Esteem" A multivariate analysis of variance tested the effect of career choice on each of the three components of the Total Positive self-esteem score described by Fitts (1965). Results indicated no overall differences on either Identity, Behaviour or Self Acceptance for male vs female dominated career groups (or on any other scale). 124 An interest in the relative contributions of these components of Inter/Personal Self-Esteem to the variance noted in Masculinity (Table III) led to the analysis of correlations among BSRI variables and the three components. Pearson Product moment correlation coefficients were computed both for the original Masculinity and Femininity scales and for the reorganized factors. While the Total Positive score had been significantly related only to Masculinity and not to Femininity, the same pattern emerges now only in the case of "Self Satisfaction" or "Self Acceptance" - the estimate of how positively a person evaluates the self she describes. (For exact figures, see Table XX in Appendix C). With the variable Identity, however, both Masculinity and Femininity covary positively while neither "Masculinity" nor "Femininity" help determine Behavioural Self scores. The absolute value of Androgyny remains unrelated to any of the subareas of self-esteem as i t did with the Total Positive score. BSRI factors and components of self-esteem: Table XXI (also in Appendix C) reproduces the correlations between Factors of the Bern Inventory and the components of Self-Esteem. These three dimensions appear to be differentially related to the four aspects of the sex role stereotypes. Identity: High self ratings of "Independence" and of Nurturance predict a positive self image on the Identity scale. The degree of "Dominance" or of "Passivity" seems to have l i t t l e bearing on self concept in this area. Self Satisfaction or Acceptance: In terms of Self Satisfaction, the Independence element was again the most highly correlated (A.=.281, p <.01). The woman who is the most accepting of herself sees herself in terms of self reliance and individuality, but to what extent she endorses either passive or aggressive traits is not predictable. Behaviour: None of the correlations with Behaviour Self ratings reached significance, suggesting that sex role attitudes are not being tapped by the kinds of behaviour measured in this inventory. It means also that the variance in Inter/ Personal Self-Esteem (Total P) associated with aspects of sex roles, (as noted in previous sections of this study) can be due only to variations in two of its subscales, and that the scale Behavioural Self is irrelevant to that discussion. 126 Summary The objective of the present study in identifying role-related internal constraints upon women's career choices was not met in ways anticipated at the outset. It is clear from the findings of this research that whether or not women follow paths which have been traditionally assigned to women in western culture is a result of factors more complex than the degree to which they have internalized the feminine role. However, certain personality attributes drawn from the masculine and feminine sex role stereotypes as they are presently conceived did appear to have a bearing on how optimistic and how self confident women were of their ability to achieve academically. There was also some evidence that academic self-confidence and feelings of competence may help to distinguish those women who gravitate toward the scientific fields in preference to the educational, health and social science areas. The results of the inquiry into determinants of career choices led to further investigation into the measures used to assess one of the key determinants, sex role stereotyping. The Bern Sex Role Inventory was found to include four personality dimensions, rather .than the two dimensional Masculine-Feminine construct. These four aspects of the two stereotypes related differentially to levels of self-esteem in both the social and personal and the achievement areas, emphasizing the range of factors involved in developing high self-esteem simultaneously in both these aspects of l i f e . Of particular interest was the negative impact upon self-esteem produced by a group of feminine qualities. Endorsement of this group of traits correlated negatively with the tendency to say socially desirable things about oneself. The remaining feminine qualities, and the two masculine groups of traits had varying, but generally positive, effects upon 127 upon self-esteem and Social Desirability. The strong relationship found between Inter/Personal Self-Esteem and Social Desirability emphasizes the effect of the social environment on self attitudes. An environment which supports women's changing roles could therefore be expected to contribute to an increase in their self-esteem. Although the aspects of sex role stereotypes identified by factor analysis in the present study could be shown to effect self-esteem in either positive or negative ways, the effect of sex role integration, or "androgyny" was not demonstrated. Because "androgyny" is a construct which is being widely discussed by those involved in altering attitudes toward sex roles, its meaning; in terms of the qualities i t includes, the way i t relates to self-esteem, and the effect the androgynous self concept may have on life style choices, warrants further research. 128 REFERENCES Almquist, E.M., & Angrist, S.S. Role model influences on college women's career aspirations. Merrill - Palmer Quarterly of Behaviour and  Development, 1971, ]]_, 3, 263-279. Atkinson, J. W., & Raynor, J.O. (Eds.) Motivation and Achievement. New York, Halsted Press Division, 1974. Baird, L. The Graduates. Princeton, N.J.: Educational Testing Service, 1973. In G.K. Baruch, The traditional feminine role: Some negative effects. The School Counsellor, 1974, March. Baruch, G.K. Feminine self-esteem, self-ratings of competence, and maternal career commitment. Journal of Counselling Psychology, 1973, 20, 5, 487-488. Baruch, G.K. The traditional feminine role: Some negative effects. The School Counsellor, 1974, March. Bern, S.L. and Bern, D.B. Case study of a nonconscious ideology: Training the women to know her place. In M.H. Garskof (Ed.), Roles Women  Play. Belmont, Calif.: Brooks Cole, 1971. Bern, S.L. The measurement of psychological androgyny. Journal of Consulting  and CIinical Psychology, 1974, 42, 2, 155-162. Bern, S.L. Sex role adaptability: One consequence of psychological androgyny. Journal of Personality and Social Psychology, 1975, 31_ 4, 634-643. Bern, S.L. On the utility of alternate procedures for assessing psychological androgyny. Journal of Consulting and Clinical  Psychology, in press. Bjerring, J. & Seagraves, P. UBC Trip Triangular Regression Package. Through facilities of U.B.C. Computing Center, 1972 (revised 1974). Breton, R. & McDonald, J. Career decisions of Canadian youth Vol. 1. In Report of the Royal Commission on the Status of Women in Canada Ottawa: Information Canada, 1967. Brookover, W.B., Paterson, A., & Thomas, S. Self concept of ability and school achievement. Final Report of Co-operative Research Project  No. 845. East Lansing, Michigan: Michigan State University, Office of Research and Publication, 1962. Broverman, I.K., Vogel, S.R., Broverman, D.M., Clarkson, F., & Rosenkrantz, P. Sex Role stereotypes: A current appraisal. Journal of Social Issues, 1972, 28, 2. Coleman, J.S. The adolescent society. Glencoe, 111.: The Free Press, 1961. In E.E.' Maccoby (Ed.), The development of sex differences. Stanford, Calif.: Stanford University Press, 1966. 129 Coopersmith, S.E. The antecedents of self-esteem. San Francisco: W.H. Freeman, 1967. Cowan, G. & Moore, L. Female identity and occupational commitment. Michigan: Unpublished doctoral dissertation, April, 1971. David, D.S. & Brannon, R. (Eds.) The forty-nine percent majority: The male sex role. Reading, Mass.: Addison-Wesley Publishing Co., 1976. Duke, M., & Nowicki, S. Locus of control and achievement - The confirmation of a theoretical expectation. Journal of Psychology, 1974, 87, 263-267. Finn, J.D. UBC MULTIVAR Program. In Univariate and multivariate analysis  of variance, covariance and regression. National Educational Resources, Release 2, 1974 (available through U.B.C. Educational Statistics Laboratory). Fitts, W.H. Manual for the Tennessee Self Concept Scale. Nashville: Counsellor Recordings and Tests, 1965. Fitts, W.H. The Second concept and self-actualization: Research Monograph  No.3. Nashville, Tennessee: Dede Wallace Center Studies on the Self Concept and Rehabilitation Monograph Series, July, 1971. Frankel, P. Sex role attitudes and the development of achievement need in women. Journal of Col 1ege Student Personnel, Mar., 1974, ]_5, 2, 114-119. Glass, G., & Stanley, J. Statistical methods in education and psychology Englewood Cliffs, N.J.: Prentice-Hall, 1970, chap. 16. Goldberg, P. Are women prejudiced against women" Transaction, j>, April, 1968, 28-30. Hakstian, R. Manual for the Alberta general factor analysis program University of Alberta, Division of Educational Research, 1972. Hakstian, R. AGFAP: Alberta general factor analysis program. University of Alberta, Division of Educational Research, 1972. Available through facilities of the U.B.C. Educational Statistics Laboratory. Hakstian, R. OWMAR: Program for multivariate analysis of variance. University of Alberta, Division of Educational Research, 1977. Available through facilities of U.B.C. Educational Statistics Laboratory. Halas, CM. Sex-role stereotypes: Perceived childhood socialization experiences and the attitudes and behaviour of adult women. Journal of Psychology, 1974, 88, 261-275. Halm, J. UBC CORN. In Parametric and non-parametric correlations and tests of significance. U.B.C. Computing Center, June, 1973, (rev. 1975). 130 Harmon, L. The childhood and adolescent career plans of college women. Journal of Vocational Behaviour, 1971, Vol. 1. Harvey, D.W., & Whinfield, R.W. Extending Holland's theory to adult women. Journal of Vocational Behaviour, 1973, 3_, 115-127. Hawley, P. Perceptions of male models of femininity related to career choice. Journal of Counselling Psychology, 1972,• 1_9, 4, 308-313. Healey, C.C., Bailey, M.L., & Anderson, E.C. The relation of esteem and vocational counselling to range of incorporation scores. Journal of Vocational Behaviour, 1973, 3^» 69-74. Helson, R. Generality of sex differences in creative style. Journal of  Personality, 1968, 36, 1, 33-47. Helson, R. Women mathematicians and the creative personality. Journal of Consulting and Clinical Psychology, 1971, 36, 2, 210-220. Hennig, M.M. Family dynamics and the successful woman executive, In R.B. Kundsin (Ed.), Women & success. New York: Wm., Morrow & Co., Inc., 1974. Hjelli, L A., & Butterfield, R. Self actualization and women's attitudes toward their roles in contemporary society. Journal of Psychology, 1974, 87, 225-230. Hochschild, A. Making i t : Marginality and obstacles to minority conscious-ness. In R.B. Kundsin (Ed.), Women & success. New York: Wm. Morrow & Co., Inc., 1974. Holland, J.L. Need reduction as an occupational determinant. In D. Zytowski (Ed.), Vocational behaviour: Readings in theory and research. New York: Holt, Rinehart and Winston, Inc., 1968. Horner, M.S. Sex differences in achievement motivation & performance in competitive and non-competitive situations. Unpublished Doctoral Dissertation. University of Michigan, Ann Arbor, Mich., 1968. In M.S. Horner & M.R. Walsh. Psychological barriers to success in women. (In R.B. Kundsin Ed. , Women & success. New York: Wm. Morrow & Co., Inc., 1974. Howe, F. Sexual stereotypes and the public schools. In R.B. Kundsin (Ed.) Women & success. New York: Wm. Morrow & Co., Inc., 1974. Information Canada. Report of the Royal Commission on the status of women  in Canada. Ottawa: 1970. Janeway, E. Man's world, woman's place. New York: Dell Pub!ishinging Co:, 1971. Joe, V.C. Review of the internal-external control construct as a personality variable. Psychological Reports, 1971, 28, 619-640. 131 Kato, N. A fundamental study of rod-frame test. Japanese Psychological Research, 1965, 7, 61-68. In E. E. Maccoby & C.N. Jack!in (Eds.), The psychology of sex differences. Stanford, Calif.: Stanford University Press, 1974. Karman, F. Women: Personal and environmental factors in role identification and career choice. Based on Ed. D. Dissertation, University of California, Los Angeles, Aug., 1973. Kerlinger, F.N., & Pedhazur, E.J. Multiple regression in behavioural research. New York: Holt, Rinehart and Winston Inc., 1973. Kimball, M. Achievement: A second look at some assumptions. Paper presented at the Symposium: Women as portrayed by psychology & sociology at Canadian universities. University of Toronto, Oct. 5-6, 1974. Klemmack, D.L., & Edwards, J. M. Women's acquisition of stereotyped occupational aspirations. Journal of Sociology and Sociological  Research, 1973, 57, 4, 510-525. Korman, A.K. Self-esteem as a moderator of the relationship between self-perceived abilities and vocational choice. Journal of Applied  Psychology, 1967, 51_, 1 , 65-67 (a). Korman, A.K. Relevance of personal need satisfaction as a function of self-esteem. Journal of Applied Psychology, 1967, 51_, 6, 533-538 (b). Kriger, S.F. N-Achievement and perceived parental child-rearing attitudes of career women and homemakers. Journal of Vocational Behavior, 1972 4, 419-432. Leland, C.A., & Lozoff, M. College influences on the role development of female undergraduates. Bethesda, Md.: 1969, Eric Document Reproduction Service No. ED-026975. In M. Lozoff, Fathers and autonomy in women. (In R.B. Kundsin (Ed.), Women & Success. New York: Wm. Morrow & Co., Inc., 1974. Levy, B. Do teachers sell girls short? reprint from Today's Education, Dec, 1972. (by the B.C.T.F., Jan., 1974). Limpus, L. Liberation of women: Sexual repression and the family. This Magazine is About Schools, 1969, 3_, 1, 61-75. Loeb, D. Building sexual confidence. Medical Aspects of Human Sexuality. 1973, May_, 82-106. Lozoff, M. Fathers and autonomy in women. In R.B. Kundsin (Ed.), Women & success. New York: Wm. Morrow & Co., Inc., 1974. Lyell, R.G. Adolescent and adult self-esteem and cultural values. Adolescence, 1973, 8, 29, 85-88. Maccoby, E.E. (Ed.). The development of sex differences. Stanford, Calif.: Stanford University Press, 1966. Maccoby, E.E. & Jacklin, C.N. (Eds.). The psychology of sex differences. Stanford, Calif.: Stanford University Press, 1974. 132 Mansfield, R. Self-esteem, self-perceived abilities and vocational choice. Journal of Vocational Behavior, 1973, 3_, 433-441. Morrison, T.L., Duane, T.M., & Weaver, J.S. Self-esteem and self-estimates of academic performance. Journal of Consulting and Clinical Psychology. 1973, 41_, 3, 412-415. Nelson, L.R. LERTAP, Dunedin, N.W.: University of Otago, Education Department, 1974. Available through facilities of U.B.C. Educational Statistics Laboratory. Nunnaly, J. Psychometric theory. New York: McGraw-Hill, 1967. Oetzel, R. The relationship between sex role acceptance and cognitive abilities. Unpublished master's thesis, Stanford University, 1961. In E.E. Maccoby, (Ed.), The development of sex differences. Stanford, Calif.: Stanford University Press, 1966. O'Hara, R. The roots of Careers. Elementary School Journal, 1962, 5_, 277-280. Olive, H. Sex differences in adolescent vocational preferences. Vocational Guidance Quarterly, 1973, March, 199-201. Patterson, J.M.& Whitaker, R.A. UBC C-GROUP Cluster Analysis Program. U.B.C. Department of Geography, 1973. Available through facilities of U.B.C. Computing Center. Piatt, J. & Eisenmann, R. Internal-external control of reinforcement, time perspective, adjustment and anxiety. Journal of General  Psychology, 1968, 79, 121-128. Powell, B. & Reznikoff, M. Role conflict and symptoms of Psychological distress in college education women. Journal of Consulting and  Clinical Psychology, 1976, 44, 3, 473-479. Roe, A. The psychology of occupations. New York: Wiley Press, 1956. Roe, A. Childhood experiences as occupational determinants. In D. Zytowski (Ed.), Vocational Behavior: Readings in theory and research. New York: Holt, Rinehart and Winston, Inc., 1968. Robinson, J., & Shaver, P. Measures of social psychological attitudes. Ann Arbor, Mich.: University of Michigan, Survey Research Center, Institute for Social Research, 1973. Satir, V. Conjoint family therapy. Palo Alto, Calif.: Science & Behavior Books Inc., 1964. Smith, R.C. Self-esteem as moderator variable in the relationship between manifest need of nurturance and satisfaction. Journal of Clinical  Psychology, 1972, July, 347-348. 133 Spence, J., Helmreich, R., & Stapp, J. Ratings of self and peers on sex role attributes and their relation to self-esteem and conceptions of masculinity and femininity. Journal of Personality and  Social Psychology, 1975, 32, 1, 29-39. Super, D.E. Self Concepts in Vocational Development In D.G. Zytowski (Ed.), Vocational Behavior: Readings in theory and research. New York: Holt, Rinehart and Winston Inc., 1968. (Reprinted from Career Development: Self Concept Theory. New York: The College Entrance Examination Board, 1963, p.1-16). Tatsuoka, M.M. Discriminant analysis: the study of group differences. Champaign, 111.: Institute for Personality and Ability Testing, 1970. Terman, L.M. & Oden, M.H. The gifted child grows up. Stanford: Stanford University Press, 19477 In E.E. Maccoby (Ed.), The development of sex  differences. Stanford, Calif.: Stanford University Press, 1966. Thurber, S. Defensive externality and academic achievement by women. Psychological Reports, 1972, 30, 454. Thurber, S. The achievement motivation-performance relationship as moderated by sex role attitudes. Educational & Psychological Measurement, 1976 36, 1075-77. Torrance,P. Guiding creative talen. Englewood Cliffs, N.J.: Prentice Hall, Inc., 1962. Cited by F. Howe, Sexual stereotypes and the public schools, in R.B. Kundsin (Ed.), Women & success New York: Wm. Morrow & Co. Inc., 1974. Wakefield, J.A., Sasek, J., Friedman, A., & Bowden, J. Androgyny and other measures of masculinity-femininity. Journal of consulting & Clinical Psychology, 1976, 44, 5, 766-770. Woolfolk, R.K. Absurd communication as an adjunct to assertion training. Psychotherapy: Theory, Research and Practice, 1976, J_3, 2, 123-127. Zingberg, D. College: When the future becomes the present. In R.B. Kundsin (Ed.), Women & success. New York: Wm. Morrow & Co., Inc., 1974. Zytowski, D.V., (Ed.). Vocational behavior: Readings in theory and research. New York: Holt, Rinehart & Winston, 1968. REFERENCE NOTES 1. Bern, S.L. Personal communication, March 10, 1977. 134' APPENDICES 135 APPENDIX A MEASURES EMPLOYED PAGE 1. Explanation to Participants. 136 2. The Bern Sex Role Inventory - Psychometric Data 1;37 - Test 140 3. The General Self Concept of Ability Scale - Psychometric Data 142 - Test 143 4. Indication of Career Plans 145 5. Tennessee Self Concept Scale - Psychometric Data 146 - availability 148 136 (1) EXPLANATION TO PARTICIPANTS As a student in Counselling Psychology at U.B.C., I am interested in working to improve the effectiveness of Vocational Counselling, particularly for women. To contribute in this area I need to learn more about the ways in which women presently make choices about their future careers. I am requesting your help, along with that of women in five other departments of the university, because I feel the most valuable source of information about their needs and ideas is in women themselves, while they are engaged in this process. Due to the personal nature of the questionnaire, the confidentiality of your responses will be safeguarded by replacing names with numbers and destroying the personal identification, required only for the purpose of obtaining your consent. You are asked to return the questionnaires along with your signed consent form, requested by the U.B.C. Research Administration, in the envelope provided. It will remain sealed, to be opened only by the principal researcher. Please answer the questions as honestly as you can - there are no right or wrong answers, only personal choices. Arrangements will be made to inform you of the overall results of the study when it has been completed, and any questions you may have will be answered fully at that time, since at present I do not wish to influence your responses in any way. And know that your part in this study, an invaluable part, is greatly appreciated. CONSENT FORM I, agree to participate in this study, on (first and last name) the understanding that the information I provide will be confidential and used only for the purpose of research in Counselling. Date 137 (1) Psychometric Data For the Bern Sex Role Inventory A) Norms: Normative data was collected from 900 undergraduate students at Stanford University and Foothill Junior College in 1973. Results showing that males score higher on the Masculinity scale and females higher on the Femininity scale than the opposite sex in each case shows that the stereotypes which had been defined for men and women in general are shared by people individually. This trend is borne out by sex differences on the Androgyny Score, with males scoring on the masculine side of androgyny and females on the feminine side. B) Reliability: i) Internal Consistency - The coefficient alpha computed to estimate internal consistency showed the reliability estimate of the scales for the two norm groups to be: Stanford Foothill Masculinity .86 .86 Femininity .80 .82 Androgyny .85 .86 Social Desir- .75 .70 ability i i ) Test-retest reliability - Over a four week period, the reliability of scores for 28 males and 28 females from the Stanford normative sample, determined be product-moment correlations between the two scores are as follows: Masculinity J = .90 Femininity = .90 Androgyny = .93 Social Desir-ability = .89 138 C) Validity: i) Concurrent: Since the BSRI is unique in its method of measuring masculinity and femininity, i.e., as independent concepts, comparisons with other measures of sex role self concept which place the two at opposing extremes, are inadequate validation for the BSRI. Hence the low correlations between the BSRI scales and the Masculinity/Femininity scales of both the California Psychological Inventory and the Guilford-Zimmerman Temperament survey. (Bern, p.160, 1974) Bern concludes that "the BSRI is measuring an aspect of sex roles which is not directly tapped by either of these two scales." A factor analytic approach to validating the BSRI was taken by Wakefield, Sasek, Friedman and Bowden (1976). From the Masculinity/Femininity scales of four personality tests including: the BSRI, the "Need for Heterosexuality Scale" from the Adjective Check List, and the Masculinity/Femininity scales of the MMPI, the California Psychological Inventory and the Omnibus Personality Inventory, three principal components relating to sex roles were extracted. They were a Masculinity/Femininity factor, a Need for Hererosexual-ity factor and an Androgyny factor. The BSRI scales were then compared with the other measures in terms of their relative contributions to those common factors. In contrast with the other Masculinity/Femininity scales, the Masculinity and Femininity scales of the BSRI loaded not onto the Masculinity/ Femininity factor, but on the "Need for Heterosexuality". The third factor, Androgyny, on which the BSRI Androgyny score loaded highly, proved to be orthogonal to the first two factors. Again, the BSRI's unique definition of the two stereotypes is the logical source of this variation. The results emphasize the difficulty of establishing concurrent validity for a test which sets out to distintuish itself from existing instruments. 139 i i ) Construct - Validity for the construct "Androgyny" as the balance between the two sex roles has been demonstrated by Sandra Bern's behavioural studies. (Bern, 1975). In a series of studies, subjects who scored "androgynous" on the BSRI consistently displayed a wider range of behaviours related to both sex roles than did either masculine or feminine sex typed subjects. The flexibility the concept implied was borne out by the ease with which androgynous men and women displayed typically masculine or feminine behaviours, according to the situation. The Masculinity and Femininity scales, although they do not correlate with other measures, are capable of discriminating males from females, as are the other scales. (Bern, 1974). This is evidence to. suggest that commonly held stereotypes are indeed descriptive of men and women to different degrees. 140 On the following page, you will be shown a large number of personality characteristics. I would like you to use those characteristics in order to describe yourself. That is, I would like you to indicate, on a scale from 1 to 7, how true of you these various characteristics are. Please do not leave any characteristic unmarked. Example: sly Mark a 1 i f it is NEVER OR ALMOST NEVER TRUE that you are sly. Mark a 2 i f i t is USUALLY NOT TRUE that you are sly. Mark a 3 i f it is SOMETIMES BUT INFREQUENTLY TRUE that you are sly. Mark a 4 i f it is OCCASIONALLY TRUE that you are sly. Mark a 5_ i f i t is OFTEN TRUE that you are sly. Mark a 6 i f i t is USUALLY TRUE that you are sly. Mark a 7_ i f i t is ALWAYS OR ALMOST ALWAYS TRUE that you are sly. Thus, i f you feel it is sometimes but infrequently true that you are "sly" never or almost never true that you are "malicious", always or almost always  true that you are "irresponsible", and often true that you are "carefree", then you would rate these characteristics as follows: Irresponsible Carefree 1 4 V 1 NEVER OR ALMOST NEVER TRUE DESCRIBE YOURSELF 3 USUALLY NOT TRUE SOMETIMES BUT INFREQUENTLY TRUE OCCASION- OFTEN USUALLY ALWAYS OR ALLY TRUE TRUE ALMOST TRUE ALWAYS TRUE Self reliant Yielding Helpful Defends own beliefs Cheerful Moody Independent Shy Conscientious Athletic Affectionate Theatrical Assertive Flatterable Happy Strong personality Loyal Unpredictable Forceful Feminine Reliable Analytical Sympathetic Jealous Has leadership abilities Sensitive to the needs of others Truthful Willing to take risks Understanding Secretive Makes decisions easily Compassionate Sincere Self-sufficient Eager to soothe hurt feelings Conceited Dominant Soft-spoken Likeable Masculine Warm Solemn Willing to take a stand Tender Friendly Aggressive Gullible Inefficient Acts as a leader Childlike Adaptable Individual istic Does not use harsh language Unsystematic Competitive Loves children Tactful Ambitious Gentle Conventional 142 (3) Psychometric Data For the Self Concept of Ability Scale A) Reliability - No information on the scale's consistency or stability were available. Validity - construct Evidence for the validity of the self concept of ability as a measure of confidence in one's ability to achieve academically can be seen in the positive correlation between ability self concept and actual achievement, when intelligence is statistically controlled. Students with high self concepts of their abilities tend to see grades as more important than do those with low ability self concepts. Feelings of competence in particular subjects is reflected in preferences for those subjects and in higher performance in them, suggesting that the scale taps a personality dimension that is relevant to achievement. Limitation - since the results detailed above have not been replicated with a college-age population, the scale is included as an exploratory measure, and as an adjunct to the TSCS in measuring self-esteem. 143 GENERAL SELF-CONCEPT OF ABILITY Circle the letter in front of the statement which best answers each question. ,1. How do you rate yourself in academic ability compared with your close friends? a) best b) above average c) average d) below average e) poorest 2. How do you rate yourself in academic ability compared with those in your classes at university? a) among the best b) above average c) average d) below average e) among the poorest 3. Where do you think you would rank in your year in university? a) among the best b) above average c) average d) below average e) among the poorest 4. Do you think you have the ability to complete a graduate program in your present field? a) yes, definitely b) yes, probably c) not sure either way d) probably not e) no 5. Where do you think you would rank in your class in a graduate program? a) among the best b) above average c) average d) below average e) among the poorest 6. How likely do you think i t is that you could complete work up to the doctoral level? a) very likely b) somewhat likely c) not sure either way d) unlikely e) most unlikely •144= 7. Forget for a moment how others may grade your work. In your own opinion, how good is your work? a) excellent b) good c) average d) below average e) much below average 8. What kind of marks do you think you are capable of getting? a) mostly first class b) mostly high second class c) mostly low second class d) mostly pass e) mostly fail 9. How important to you are the marks you get? a) most important b) very important c) important d) not particularly important e) totally unimportant 10. How. important is i t to you to be high in your class academically? a) most important b) very important c) important d) not particularly important e) totally unimportant 11. How do you feel i f you don't do as well in your courses as you know you are able to? a) very badly b) quite badly c) a l i t t l e disappointed d) not particularly disappointed e) i t doesn't bother me at all 12. How important is i t to you to do well in university? a) highly important b) quite important c) somewhat important d) not particularly important e) it doesn't matter to me at all 13. Which statement best describes your most comfortable or typical position: a) getting better marks than everyone else b) getting better marks than almost everyone else c) getting about the same marks as everyone else d) getting slightly lower marks than everyone else e) I don't care about any particular marks. 145 14. Academically, do you try to do better than others a) all of the time b) most of the time c) about half the time d) occasionally e) never 15. How important are good marks to you, compared with other aspects of your life at present? a) they are the most important aspect b) good marks are among the important aspects c) some other things in my l i f e at present are more important d) many other things in my life are more important e) good marks are the least important area of my life 16. What level of academic achievement do you aim for? a) mostly first class b) mostly high second class c) mostly low second class d) mostly pass e) mostly fail 1. Please indicate by circling "yes" or "no", whether you have decided to pursue a career after graduation. "Yes" you intend to have a career or "No" you do not 2. If you d£ intend to enter a career after graduation, please indicate whether or not the career you have chosen is directly related to your present field of study. (4) CAREER PLANS Yes No Yes No 3. If the answer to question #2 is "No", please state your career choice below 146 (5) Psychometric Data For the Tennessee Self Concept Scale A) Norms: The TSCS was normed on people from many socio-economic levels, geographical areas, and in various age groups and was found to be relevant despite these demographic variations. Norms are reported in the manual for three groups along the mental health continuum, from psychiatric patients, to the normal population, to a group described as high in "personality integration". Sex differences on profiles and on individual scores are termed negligible. B) Reliability i) Internal Consistency - The split half reliability for the Total P score was reported by Nunnaly (1968) to be =.91. i i ) Test-retest reliability - A study involving psychiatric patients (Congdon, 1958) reported in the TSCS manual found a reliability coefficient =.88 for the Total Positive score. For another sample, Fitts reports a reliability of .92 for the Total P score, and adds that the overall profile patterns, i.e., scales contributing to the Total P score, tend to maintain their individual distinctions over long periods of time; a year or more. C) Validity i) Concurrent: The TSCS shows negative correlations with measures of psychological disturbance as measured by the MMPI High Total Positive scores imply a lack of such disturbances. Measures of physical fitness have been significantly and positively correlated with the positive score for Physical Self (Christian, 1969, reported in Monograph III). Support for the validity of the Total P score was found in the correlation with the Bills' Index of Adjustment and Values, another measure of self concept, (Gay, 1966; report in Monograph III). i i ) Construct - The scale has been found to differentiate between 147 patient and non-patient groups, and to relate positively to "clinical indices of psychological health" (Robinson and Shaver, 1973). Personality change in both patients and non-patients has been reflected in changes in their Total P scores. In addition, certain subscales serve to identify types of psychological problems as well as their severity. On the other extreme, people characterized as highly functional and integrated tend to have higher Total Positive scores than the norm group, indicating that high self-estem, as evidenced by their Total P scores, is translated into functional, integrated behaviour. 148 Tennessee Self Concept Scale This test, written by William H. Fitts, is available from the publishers, "Counsellor Recordings and Tests," at the following address: Box 6184 - Acklen Station Nashville, Tennessee 37212 U.S.A. APPENDIX B ANALYSIS OF GROUP DIFFERENCES IN MEAN SELF CONCEPT OF ABILITY TABLE XVIII Analysis of Group Differences in Mean Self Concept of Ability Gr. M2 Gr. F3 Gr. Fl Gr. F2 Gr. M3 Gr. Ml * Gr.M2 Agric. 2.3333 5.5333 6.6667 7.3333 10.53 Gr. F3 Arts - 3.200 4.334 5.0000 8.200* Gr.Fl Teach. - 1.1334 1.800 5.000 Gr. F2 Nurse. - 0.666 3.866 Gr.M3 Math - 3.200 Gr. Ml Science -Minimum difference required = 7.64 * difference is significant to a = .05 level. 151 APPENDIX C Tables Corresponding to "Additional Recommendations for Further Study." PAGE Table XIX - Means & Standard Deviations of Low Self-Esteem 152 Third of each Career Group on remaining variables. Table XX - Intercorrelations among three Components of Inter/Personal Self-Esteem and original BSRI 153 Table XXI - Intercorrelations among three Components of Inter/Personal Self-Esteem and Four Factors of BSRI. 154 Table XXII Means and Standard Deviations of Career Groups on "Identity" Component of Self-Esteem Score and BSRI Scales. 155 Table XXIII - Univariate & Multivariate Analysis of Variance of Career Groups on "Identity" Component of Self-Esteem Score and BSRI Scales. 156 Table XXIV Means and Standard Deviations of Career Groups on "Behavior" Component of Self-Esteem Score and BSRI Scales. 157 Table XXV Univariate & Multivariate Analysis of Variance for 6 Career Groups on "Behavior" Component of Self-Esteem Score and BSRI Scales. 158 Table XXVI Means and Standard Deviations of Career Groups on "Self Acceptance" Component of Self-Esteem Score and Two BSRI Factors. 159 Table XXVII Univariate & Multivariate Analysis of Variance for 6 Career Groups on "Self Acceptance" Component of Self-Esteem Score and Two BSRI Factors. 1 6 0 152 TABLE XIX Means & Standard Deviations of Low Self-Esteem third of each Career Group (on remaining variables) n = 30 Teachers Nurses Arts Science Agric. Math. Dominance X SD 54.20 9.03 44.60 8.82 48.40 5.90 46.20 4.32 44.60 7.70 37.80 7.95 Nurturance X SD 54.40 3.78 52.20 8.53 51.00 7.25 53.00 3.16 50.20 6.87 42.40 7.63 Independence X SD 28.80 5.17 30.00 4.80 26.00 6.00 28.80 5.97 30.80 2.68 28.00 7.65 Passivity X SD 25.20 1.92 24.20 4.32 26.80 7.79 27.20 3.34 25.20 2.59 24.40 6.19 Self Con-cept Abi 1. X SD 56.80 5.89 61.20 10.83 55.80 7.40 58.00 2.35 52.80 7.19 57.20 8. 01 Soc. Desir. X SD 93.20 7.82 100.20 7.40 92.20 5.07 92.20 10.47 94.20 7.83 99.00 4.64 Self.Crit. X SD 37.40 5.18 43.00 3.46 38.00 5.70 38.60 1.34 37.20 5.81 31.40 5.73 TABLE XX Intercorrelations Among 3 Components of Inter/Personal Self-Esteem and original BSRI Variable 1 2 3': 4 5 6 7 8 9 1 Self Concept of Ability 1.000 2 "Identity" 0.294** 1.000 3 "Self Acceptance" 0.162 0.636** 1.000 4 "Behaviour" 0.042 0.040 0.016 1.000 5 Social Desirability 0.264* 0.523** 0.555** -0.090 1.000 6 Self Criticism -.024 -0.321** -0.312** -0.071 -0.231* 1. 000 7 Masculinity 0.430** 0.324** 0.269* 0.052 0.261* -0. 030 1.000 8 Femininity 0.295* 0.289** 0.019 -0.130 0.132 -0. 206 0.053 1.000 9 Androgyny 0.075 0.011 0.036 -0.041 0.045 -0. 025 -0.354 0.229 1.000 * = significant to <* .05 level ** = significant to <= .01 level (leading decimal points are omitted) TABLE XXI Intercorrelations Among 3 Components of  Inter/Personal Self-Esteem 4 Factors of  Bern Sex Role Inventory Variable 1 2 3 4 5 6 7 1 Identity 1.000 2 Self Acceptance .636** 1.000 3 Behaviour .040 .016 1.000 4 Dominance .176 .198 -.058 1.000 5 Nurturance .257* .072 -.138 .286** 1.000 6 Independence .382** .281** .094 .325** .163 1.000 7 Passivity -0.199 -.323** -.061 -.062 .219* -.320** 1.000 * = significant to«:.05 level ** = significant to <* .01 level (leading decimal points are omitted) TABLE XXII Means and Standard Deviations of Career Groups on "Identity" Component of Self-Esteem Score and BSRI Scales Career Groups Masculinity Femininity Androgyny Social Desirability Identity Self Concept of ability Self Criticism Teaching n = 15 X SD 4.66 0.60 5.12 0.51 1.00 1.65 99.67 9.45 127.20 7.85 58.33 6.26 36.00 5.55 Nursi ng n = 15 X SD 4.86 0,75 4.91 0.58 0.25 2.34 104.73 8.50 126.73 9.44 59.47 7.74 35.87 7.00 Arts n = 15 X SD 4.47 0.77 4.81 0.56 0.80 2.36 96.33 8.61 122.13 7.96 55.13 8.30 36.07 3.90 Science n = 15 X SD 4.85 0.61 5.07 0.56 0.40 1.87 101.60 10.25 125.60 13.06 63.33 6.97 35.07 6.06 Agriculture n = 15 X SD 4.61 0.65 4.66 0.39 0.06 1.51 100.93 10.53 122.27 12.24 52.80 7.01 34.20 5.48 Mathematics n = 15 X SD 4.68 0.81 4.75 0.70 0.30 2.51 102.20 7.28 126.33 8.41 60.13 6.63 33.00 4.55 TABLE XXIII Univariate & Multivariate Analysis of Variance of Career Groups on "Identity" Component of Self-Esteem Score and BSRI Scales Source Univariate F (5,84) Masculinity 0.6734 p<.6448 Femininity 1.5538 p<.1822 Androgyny 0.4435 p<.8169 Soc i a1 Desirability 1.4005 p<.2325 Identity 0.7628 p<.5792 Self Concept of Ability 4.0840 p<.0023 Self Criticism 0.7421 p<.5942 Multivariate Test F (35,330.55) = 1.2257 p<0.1846 TABLE XXIV Means and Standard Deviations of Career Groups on "Behavior" Component of Self-Esteem Score and BSRI Scales Career Groups Masculinity Femininity Androgyny Social Desirability Behavior Self Concept of Ability Self Criticism Teaching n = 15 X SD 4.66 0.60 5.12 0.51 1.00 6.65 99.67 9.45 109.60 9.77 58.33 6.26 36.00 5.55 Nursing n = 15 X SD 4.86 0.75 4.91 0.58 0.25 2.34 104.73 8.50 111.67 13.89 59.47 7.74 35.87 7.00 Arts n = 15 X SD 4.47 0.77 4.81 0.56 0.80 2.36 96.33 8.61: 106.33 8.43 55.13 8.30 36.07 3.90 Science n - 15 X SD 4.85 0.61 5.07 0.56 0.40 1.87 101.60 10.25 114.73 14.15 63.33 6.97 35.07 6.06 Agriculture n = 15 X SD 4.61 0.65 4.66 0.39 0.06 1.51 100.93 10.53 168.13 211.19 52.80 7.01 34.20 5.48 Mathematics ;n = 15 X SD 4.68 0.81 4.75 0.70 0.30 2.51 102.20 7.28 112.60 5.80 60.13 6.63 33.00 4.55 TABLE XXV Univariate & Multivariate Analysis of Variance for 6 Career Groups on "Behavior" Component of Self-Esteem Score & BSRI Scales Source Univariate F (5,84) Masculinity 0.6734 p<.6448 Femininity 1.5538 p<.1822 Androgyny 0.4435 p<.8169 Social Desirabi1ity 1.4005 p< Behavior 1.1001 p<.3666 Self Concept of Ability 4.0840 p<.0023 Self Criticism 0.7421 p<.5942 Multivariate Test F (35,330.5459) = 1.3433 p<.0991 TABLE XXVI Means and Standard Deviations of Career Groups on "Self Acceptance" Component of Self-Esteem Score and Two BSRI Factors Career Groups Dominance Nurturance Social Desirability Self Acceptance Self Concept of ability Self Criticism Teaching n = 15 X SD 49.53 7.32 56.67 6.67 97.67 9.45 111.27 • 12.73 58.33 6.26 36.00 5.55 Nursing n = 15 X SD 48.87 9.40 . 55.07 7.90 104.73 8.50 117.27 19.53 59.47 7.74 35.87 7.00 Arts n = 15 X SD 46.33 10.72 51.20 5.75 96.33 8.61 111.13 12.24 55.13 8.30 36.07 3.90 Science n = 15 X SD 50.47 6.91 55.00 6.69 101.60 10.25 114.93 18.68 63.33 6.97 35.07 6.06 Agriculture n = 15 X SD 46.87 7.88 51.33 7.42 100.93 10.53 115.80 19.97 52.80 7.01 34.20 5.48 Mathematics n = 15 X SD 47.93 .9.46 51.60 9.20 102.20 7.28 118.47 9.36 60.13 6.63 33.00 4.55 160 TABLE XXVII Univariate & Multivariate Analysis of Variance for 6 Career Groups on "Self Acceptance" Component of Self-Esteem score and two BSRI Factors Source Univariate F (5,84) Dominance 0.4964 p< .7781 Nurturance 1.5716 p< .1770 Social Desirability 1.4005 p <.2325 Self Acceptance 0.5472 p <.7400 Self Concept of Ability 4.084:0 p < .0023 Self Criticism 0.7421 p < .5942 Multivariate Test F (30,318.00) = 1. p < 0.2291 1928 APPENDIX D TABLE XXVIII CORRELATIONS AMONG ITEMS OF THE BSRI 162 TABLE XXVIII Correlations* among items of the BSRI 1 2 3 4 5 6 7 8 X 5.7444 6.0000 5.7111 3.8222 4.5778 5.2667 4.1111 5.2889 S.D. 1.2958 . 0.9068 1.0354 1.7737 1.4831 1.3233 1.3699 1.2583 n 90 90 90 90 90 90 90 90 Var. 1 . .0851 .4253 .1011 .2618 .1888 .0786 .2838 2 , .0710 .1520 .3305 .4260 .4473 .0487 3 .1777 .2968 .3806 .2655 .1152 4 .1658 .1906 .1956 -.0566 5 .4900 .5919 .1725 6 .6211 .1806 7 .1554 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 163 TABLE XXVIII (cont'd) 9 10 11 12 13 14 15 16 X 4.6111 4.5222 4.0667 5.3889 4.0111 2.3889 5.3556 4.0778 S.D. 1.3638 1.4394 1.4667 1.0923 1.4719 1.1709 1.1285 1.3600 n 90 90 90 90 90 90 90 90 Var. 1 .2707 .2622 .3247 .5019 -.0160 -.1176 .0773 -.0644 2 .4133 .1532 -.0084 -.0337 .3080 .2302 .5212 .4235 3 .3060 .4144 .2980 .6102 .1917 .0102 .2115 .1579 4 .2700 .2844 .1241 .0873 .1582 .2419 .0094 .2222 5 .4352 .3531 .1866 . 1768 .3992 .1521 .3884 .4405 6 .5377 .3702 .2484 .2742 . 5233 .1700 .5615 .5565 7 .4989 .2692 -.0590 .1122 .6882 .2224 .4416 .6633 8 .1496 .2112 .2184 .1608 .0583 -.0008 .2720 -.0521 9 .3242 .1185 .2805 .5114 .2965 .4580 .5200 10 .2204 .2807 .2123 .1700 .2003 .2233 11 .2751 -.0673 .0043 .2609 .0200 12 .1010 -.1443 .1492 .0918 13 .3327 .3321 .6768 14 .1981 .3020 15 .4308 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 164 17 18 X- 3.7667 5.4444 S.D. 1.3170 1.0553 n 90 90 Var. 1 .1734 .4243 2 .3815 .0232 3 .2276 .4633 4 .3152 .0838 5 .4729 .2406 6 .4119 .3368 7 .4085 .2041 8 .1211 .1460 9 .7413 .2205 10 .3105 .1764 11 .1116 .2178 12 . 0940 .4670 13 .5344 .1184 14 .3254 .0849 15 .4296 .2312 16 .5250 .1540 17 .0906 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 LE XXVIII (cont'd) 19 20 21 .3444 5.2889 4.1778 .4465 1.3017 1.4266 90 90 90 .0825 .0174 .0536 .0593 .1506 -.1890 .1629 .2186 -.0555 .0542 .1474 .1662 .1040 .1265 -.1326 .1668 .2069 -.3076 .1769 .2499 -.1011 .2200 .0033 .0147 .1974 .1575 -.0159 .2765 .2457 .1225 .0939 -.0275 -.0588 .0629 .0929 -.1157 .1913 .2419 -.1597 .1177 .2398 -.1212 .2245 .1494 -.2877 .1671 .2133 -.1732 .1763 .2143 -.0844 .1472 .1330 -.2296 .3779 .0242 .1399 22 23 24 5.3889 3.7000 5.3889 0.9750 1.5737 1.2710 90 90 90 .1051 -.1193 .0064 .0755 -.1246 .0578 .2216 -.1282 -.0328 .3037 -.1783 .2278 .2675 -.3970 .1225 .4281 -.3244 .2819 .2841 -.1701 .1539 -.0282 .0325 -.1189 .4986 -.4530 .2860 .1325 -.2104 .1259 .2230 -.2513 .0934 .3591 -.2036 .1392 .2295 -.2096 .1105 -.0644 -.1658 -.0867 .2482 -.3403 .2367 ..2540 -.4044 .2525 .2353 -.3769 .2202 .0480 -.2676 -.0377 .1573 -.0522 .1749 .2006 -.0228 .1537 .0062 .1326 .0477 -.2936 .3984 -.2861 165 TABLE XXVIII (con 25 -26 -27 28 X 4.0000 6.1222 5 .0889 5 .7222 S.D. 1.4682 0.8797 1 .2261 .9193 n 90 90 90 90 Var. 1 -.0759 -.0798 .1961 - .0409 2 .2253 .2229 - .0300 .0267 3 .1096 -.1808 .1953 - .0960 4 .2517 .0780 .0226 -.1461 5 .1276 -.0115 -.0588 -.2653 6 .1430 .1247 .0744 - .0944 7 .2486 .1731 .0140 - .0373 8 -.1684 -.0319 - .0311 .1558 9 .3385 .1322 .0340 .0202 10 .2051 .0812 - .0452 - .1591 11 -.1651 -.0580 .0400 - .1428 12 -.1247 -.1420 .2314 -.0744 13 .1440 .1105 -.0929 .0105 14 -.0582 .1480 -.4652 -.1474 15 .0738 .1017 .1458 .0095 16 .2059 .1685 - .0908 -.1694 17 .3103 .0342 .0197 - .0251 18 -.0287 .0013 .0897 -.0904 19 .1465 -.0593 .0266 .0135 20 .2035 .0953 -.0788 .1599 21 .1963 .0712 .1307 .1054 22 .2331 -.0425 .1944 .1703 23 -.1491 -.0538 .1060 .2727 24 .2501 -.0346 .1489 .0164 25 -.1118 .1420 .0823 26 -.0719 .1794 27 .1994 28 29 30 31 32 33 34 35 36 37 38 39 40 d) 29 30 31 32 5.5222 5.5778 5.3222 5 .5778 1.0773 .8942 1.0732 1 .1448 90 90 90 90 .1274 .1562 .0352 _ .0353 .1024 .1782 .2512 - .0214 .0854 .0603 .0738 - .0935 -.0153 .1628 -.1217 .0779 -.1471 -.0423 .0576 -.1246 .2452 .0670 .2759 .0450 .0962 .0564 .2477 - .0055 .1346 .1084 .1039 .1155 .2744 .2480 .2526 .1154 .0104 .1713 .0997 .0191 . 0342 -.0124 .0428 - .1156 .1579 .1567 .1206 .1668 .0734 .0458 .1173 -.0368 -.1170 -.0130 -.1174 - .1593 .2220 .1488 .2907 .0216 .0557 .0270 .1884 .0496 .1015 .1994 .1239 .0968 .0206 -.0484 -.0382 - .1298 .0842 -.0078 .1003 .0811 .1380 .1430 .2198 .1490 .0336 .2069 -.0011 .3929 .3147 .1757 .2840 .3366 -.0321 .0363 .0046 .0407 .2655 .2227 .3317 .3190 .0773 .1016 .0705 .2446 .1788 .0797 .2643 .0402 .1920 .1356 .2485 .2167 .3821 .4251 .4737 .5115 .5634 .4791 .3950 .4428 .4880 .4634 166: TABLE XXVIII (cont'd) 33 34 35 36 37 38 39 40 X 4.0333 5.3111 4.9556 3.7667 3.2000 4.5556 5.2222 5.2111 S.D. 1.5808 .9735 1.0843 1.5709 1.4621 1.8566 1.6585 1.1105 n 90 90 90 90 90 90 90 90 Var. 1 -.1824 -.0402 -.2124 .1613 .0158 -.1291 -.0169 .0140 2 .1875 -.0476 .0763 -.0334 -.0913 -.0049 -.0741 -.0691 3 .-797 .2253 -.0365 -.2538 .0058 -.1095 -.0859 .0681 4 -.0452 -.0276 -.1347 -.3192 .1399 .1893 .0719 -.1604 5 .2271 .1295 .1128 .1383 -.2300 .0386 -.0865 .1023 6 .2141 .0897 -.0854 .1239 -.0432 -.0076 .0931 .0583 7 .0218 -.1766 .1023 -.2008 -.1383 -.2239 -.1773 .1847 8 -.1458 -.0009 .1661 .0852 -.0052 -.1863 .0684 -.0835 9 .1944 -.0234 -.0482 -.0428 .0417 .1052 -.1175 -.0496 10 -.0623 -.0695 -.0021 .0064 . 1076 .0360 .0082 -.0647 11 -.0053 -.3279 .0076 .0737 .3497 -.1190 .2868 -.1190 12 .0082 .1234 -.0982 .0634 .1112 .0088 -.1244 -.0399 13 -.0438 .0440 -.0558 -.1361 -.0522 .0685 -.0587 .0140 14 -.1581 .0681 -.0445 .1814 .0557 .1400 .1042 -.3042 15 .1405 -.1577 .3548 .1877 -.0440 .0545 .0841 -.1415 16 -.0596 -.0359 -.0341 -.1687 .0239 .0186 -.0695 .0779 17 .0337 -.1034 -.0284 -.0482 -.1482 .1987 .1261 .0780 18 -.0045 -.0218 .2533 .3079 .0167 -.0504 .0132 .1445 19 .0305 -.2467 .2044 .0507 .0124 .0422 .0544 .1019 20 -.0785 -.0489 -.1120 -.1908 -.0398 -.0719 -.1795 .1617 21 -.1160 -.0983 .1516 -.0925 -.2772 .0218 -.0139 -.2295 22 .1357 -.0232 -.1438 -.0961 .2649 .0199 -.0696 -.0803 23 .1241 .0798 .1160 -.4298 .0336 -.1932 -.0376 -.1208 24 .2248 -.0645 .0431 -.0474 .2060 .1552 -.2523 -.1075 25 .0944 -.0435 -.0525 -.1380 -.2471 -.1327 -.1546 .2214 26 .1084 -.0408 -.0871 .2105 .0240 -.1850 .0986 -.0033 27 -.0219 -.0248 .0723 -.2629 -.0835 .1012 -.1248 .0597 28 -.2861 -.0289 -.2153 .0088 -.1676 -.3648 -.0061 -.0402 29 -.1782 -.1618 -.2157 -.0507 .0254 -.0580 .1797 -.2778 30 -.2913 .0953 -.2485 -.1096 -.0841 -.0626 .0070 -.3125 31 .0488 .0366 -.1258 .0347 -.0237 -.1540 -.1868 .0168 32 -.2022 -.1205 -.0182 .2118 -.1249 -.0319 .1863 -.0843 33 .0574 .1818 .0251 .0330 .0562 -.1750 .0201 34 -.1313 -.1050 .1040 -.0751 -.3471 -.0584 35 .0951 .0380 .1191 -.1569 .0246 36 -.1258 .0137 .2261 .1540 37 -.1287 -.0126 -.0266 38 -.0165 -.0059 39 -.1702 40 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items