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.

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 B r i t i s h 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 B r i t i s h 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, i t was apparent that absolute levels of masculinity were significantly related  ii ±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. factors were obtained.  Four  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  iii 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 i t s significance to women's career aspirations.  iv  TABLE OF CONTENTS PAGE i  ABSTRACT LIST OF TABLES  vi  LIST OF FIGURES  viii  ACKNOWLEDGEMENTS  ix  CHAPTER I - RELATED LITERATURE Introduction Limitations in Vocational Theories Predictions Based on Abilities Predictions Based on Interests Predictions Based on Personality Abilities Motivation Socialization of Women Self Concept and Vocational Choice Self-Esteem Self-Esteem and Achievement Self-Esteem and Femininity  1 1 2 4 5 5 7 8 12 15 17 19 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 The Tennessee Self Concept Scale  45 47  The Self Concept of Ability Scale  49  Data Analysis  52  Summary  56  V  PAGE CHAPTER IV - RESULTS Response to Hypotheses 1, 2, and 3  68  Response to Hypothesis 4  61  Response to Hypothesis 5  65  CHAPTHER V - POST ANALYSES Methodology The Androgyny Score Dimensions of the BSRI Rescoring of the BSRI .. Second Phase of Hypothesis Testing Results The Androgyny Score Dimensions of the BSRI Phase II of Hypothesis Testing CHAPTER VI - DISCUSSION Congruence between Self Concept and Occupational Stereotype Sex Role Stereotype and Career Choice Sampling Factors Instrumentation Variations Within Career Groups Self Concept of Ability and Masculinity/Femininity of Career Choice Inter/Personal Self-Esteem and Masculinity/Femininity of Career Choice Relationships Among the Variables Within-Group Differences Relationships Between Sex-Role Stereotyping and Self-Esteem Androgyny and Self-Esteem Masculinity & Femininity Scales Feminine Factors of the BSRI & Self-Esteem Masculine Factors of the BSRI & Self-Esteem Recommendations for Future Research Additional Recommendations Summary REFERENCES  58  66 67 67 68 69 .... 70 72 72 76 89 98 99 100 100 101 103 105 107 107 107 110 110 113 115 118 120 122 126 128  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 I II  III IV V VI VII VIII IX X  TITLE MEANS AND STANDARD DEVIATIONS FOR CAREER GROUPS ON SELF CONCEPT VARIABLES  PAGE 59  UNIVARIATE AND MULTIVARIATE ANALYSIS OF VARIANCE ON SELF CONCEPT VARIABLES FOR CAREER GROUPS  60  INTERCORRELATION MATRIX FOR COMBINED GROUPS  62  STEPWISE MULTIPLE REGRESSION ANALYSES UPON SELF-ESTEEM PERCENTAGE OF ANDROGYNOUS SUBJECTS IN CATEGORIES FORMED ACCORDING TO "MEDIAN SPLIT" PROCEDURE  64 74  MEANS, STANDARD DEVIATIONS, AND F VALUES FOR ANDROGYNOUS CATEGORIES ON SELF CONCEPT VARIABLES  75  IDENTIFICATION OF BSRI ITEMS SUBJECTED TO FACTOR ANALYSIS  77  PRINCIPAL COMPONENTS SOLUTION: TRANSFORMED FACTORS  79  FACTOR PATTERN FOR FACTOR ANALYSIS OF BSRI ITEMS: PRINCIPAL COMPONENTS - 4 & 5 FACTORS  82  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 XIV XV  XVI XVII  INTERNAL CONSISTENCY STATISTICS FOR FACTORS OF THE BSRI MEANS AND STANDARD DEVIATIONS ON BSRI FACTORS AND SELF ESTEEM VARIABLES FOR SIX CAREER GROUPS  88 90  UNIVARIATE AND MULTIVARIATE ANALYSIS OF VARIANCE FOR SIX CAREER GROUPS ON BSRI FACTORS AND SELF-ESTEEM VARIABLES  91  INTERCORRELATION MATRIX FOR COMBINED GROUPS ON FACTORED BSRI AND SELF-ESTEEM VARIABLES  93  STEPWISE MULTIPLE REGRESSION ANALYSIS UPON SELF CONCEPT OF ABILITY - BSRI AS FOUR FACTORS  95  VI 1  LIST OF TABLES Cont'd TABLE  TITLE  XVIII XIX  XX XXI XXII  XXIII  XXIV  XXV  XXVI  XXVII  XXVIII  :  PAGE  ANALYSIS OF GROUP DIFFERENCES IN MEAN SELF CONCEPT OF ABILITY  150  MEANS AND STANDARD DEVIATIONS FOR LOW SELF-ESTEEM THIRD OF EACH OF THE SIX CAREER GROUPS ON REMAINING VARIABLES  152  INTERCORRELATIONS AMONG THREE COMPONENTS OF TOTAL POSITIVE SCORE AND ORIGINAL BSRI SCORES  153  INTERCORRELATIONS AMONG THREE COMPONENTS OF TOTAL POSITIVE SCORE AND FOUR FACTORS OF BSRI  154  MEANS & STANDARD DEVIATIONS OF CAREER GROUPS ON "IDENTITY" COMPONENT OF SELF-ESTEEM SCORE AND BSRI SCALES  155  UNIVARIATE & MULTIVARIATE ANALYSIS OF VARIANCE OF CAREER GROUPS ON "IDENTITY" COMPONENT OF SELF-ESTEEM SCORE AND BSRI SCALES  156  MEANS & STANDARD DEVIATIONS OF CAREER GROUPS ON "BEHAVIOR" COMPONENT OF SELF-ESTEEM SCORE AND BSRI SCALES  157  UNIVARIATE & MULTIVARIATE ANALYSIS OF VARIANCE OF CAREER GROUPS ON "BEHAVIOR" COMPONENT OF SELF-ESTEEM SCORE AND BSRI SCALES  158  MEANS & STANDARD DEVIATIONS OF CAREER GROUPS ON "SELF ACCEPTANCE" COMPONENT OF SELF-ESTEEM SCORE AND TWO BSRI FACTORS  159  UNIVARIATE & MULTIVARIATE ANALYSIS OF VARIANCE OF CAREER GROUPS ON "SELF ACCEPTANCE" COMPONENT OF SELF-ESTEEM SCORE AND TWO BSRI FACTORS  160  CORRELATIONS AMONG ITEMS OF THE BSRI  162  vi i i  LIST OF FIGURES  FIGURE  TITLE  1  SUMMARY OF MEASURES EMPLOYED  2  DISTRIBUTION OF SCORES ON MASCULINITY AND  PAGE 51  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 L i l l y 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 a l l 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. experience.  In themselves, roles can be useful tools in interpreting  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 f l e x i b i l i t y 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 f a i l 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, a b i l i t i e s ,  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 a b i l i t i e s , 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 l i f e 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 f i r s t 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 l i 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 a l l 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 g i r l s , 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 i t is for boys.  It appears to have less influence over their choices of  activities and goals, and less effect on the development of specialized a b i l i t i e s ; 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 a b i l i t i e s , 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 s k i l l s .  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 s k i l l s 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 1966).  in girls.  (Maccoby,  Poor spatial a b i l i t i e s , on the other hand, are found in conjunction  with high levels of passivity and dependence—traits fostered by the socialization 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 g i r l s , 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 a b i l i t i e s , 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 a f f i l i a t i o n (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 i s , 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 f e l t .  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 situations.  in competitive  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 " a f f i l i a t i v e " 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 a b i l i t i e s . 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 i n i t i a l 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 l i f e 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, a b i l i t i e s , 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  and 1 976)  ( 1 972  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 a b i l i t i e s , - 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 i t 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 a b i l i t i e s , 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 i t 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, a b i l i t i e s , 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 a b i l i t i e s , but to remain in a situation where their personal needs are not being satisfied. (Korman, 1967(b)). social and to vocational situations.  The phenomenon applied both to  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. satisfaction for expressed needs.  They also experienced greater  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. 1967).  (Coopersmith,  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 l i f e 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 s k i l l s at managing and nurturing relationships, but her sense of selfworth 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. 1962).  (Brookover, Patterson & Thomas,  Depending on the subject area being considered, students' self  20 concept of their ability varied. Concept of Ability.  It differed also from their overall Self  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 i t 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 s t r i c t l y 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 i l l o g i c a l ; in short, lacking in traits denoting competency and confidence. Clarkson, F., Rosenkrantz, P., 1972).  (Broverman, I., Vogel, S., Broverman, D., 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 a b i l i t i e s ,  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 l i f e 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' masculinity . The fact remains, however that girls' 7  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). 'hon-competitive"image  Thus the "incompetent",  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 l i f e 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 i t s e l f 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 f i r s t 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 overcome in order to be seen as competent and professional.  In addition to proving  28 their a b i l i t i e s , 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 selfesteem 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  masculine or 1  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 d i f f i c u l t ies encountered by researchers in predicting women's choices from the commonly used career determinants which do not take  self-esteem ' into 1  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." STEREOTYPE:  (David & Brannon, 1976).  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 i t 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 l i f e 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: H  Q  :  y  F  - y  M  =  0  where y p  mean Androgyny Score on the BSRI for women in the traditionally feminine fields  and  mean Androgyny Score on the BSRI for women in the traditionally masculine fields.  y  M  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  hypothesis, in each case, is stated in non-directional form, indicating that two-tailed tests will be performed. In equation form, the alternate hypothesis i s : H  a  : 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 i s : 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: H  Q  : U  where u m  and l i p  M  - u  F  =0  = mean Self Concept of Ability score for women in masculine fields. = 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, i s : H :U - lip / 0 A  where  M  and U p 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, i t 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 i s : H: Q  u  M  wu  where y  =0  F  = mean level of self-esteem for women in masculine fields,  M  n  and  y  = 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"  F  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 i s : H: A  y  where  M  - y  F  p  0  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, i t 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: H  : xy = 0 p  Q  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: H  A  :  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 f a l l s 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 i s : There i s . 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: H  :H  0  A  =  ^  where ,  y  F mean level of self-esteem in a l l areas for women classified as androgynous  u  mean level of self-esteem in a l l areas for women sex typed masculine and  y  r  mean level of self-esteem in a l l 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 i n leve)of self-esteem from those sex typed feminine. In equation form, the alternate hypothesis may be represented as: H  A  :  ^A  * Vfn  y  F  where y ^ , y ^ , and  y p are defined as above.  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 i t 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". descriptions.  Procedures for scoring and analysing the data follow those  Three phases of data analysis were performed; an i n i t i a l  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 f i r s t 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 f i e l d , 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. academic level i s preferred for two reasons:  This particular  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, a l l 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 r e l i a b i l i t y 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. scores.  A third score, for Androgyny, is computed from these two  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. levels of mental health have been provided.  Norms for groups at various  A wide body of research supports  the validity and r e l i a b i l i t y 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 a l l 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 a r t i f i c i a l l y 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, c l i n i c a l l y 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 i t 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 l i f e 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 r e l i a b i l i t y 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 VARIABLE NAME  SCORING METHOD  TEST NAME  SCALES  Bern Sex Role Inventory  Masculinity  Masculinity  mean score for 20 items  Femininity  Femininity  mean score for 20 items  Androgyny  Androgyny  Social Desirability  Social Desirability  sum of scores on 20 items  all scales except Self Criticism  Inter/Personal Self-Esteem  sum of scores on 90 items (the TOTAL POSITIVE Score)  Self C r i t i cism  Self-Criticism  sum of scores on 10 items  test contains a single scale  Self Concept of Ability  sum of scores on 16 items  Tennessee Self Concept Scale  Self Concept of Ability Scale  t-ratio of difference between Masculine and Feminine scale means  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 f i r s t 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 + b X + Q  +  ]  b  5 5 X  bX  ]  +  b  2  6 6 X  2  + b  + bX 3  7 7 X  +e where Y = Masculinity/Femininity of Field a n d  X = Social Desirability ]  X  2  - Self Criticism  X^ = Masculinity  3  + b X 4  4  54 = Femininity Xg  = Androgyny  Xg  = Inter/Personal  Xy  = Self Concept of Ability  bg  = constant (y-intercept)  Self Esteem  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 a l l 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 i s :  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 a l l 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 i t 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 i n i t i a l 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: Y  l,2  =  b  0  +  b  l l  +  +  b  X  b  2 2  5 5  + e  X  X  +  +  b  b  3 3  6 6 X  X  +  b  44 X  56 -where  a n d  Y  ]  = Self Concept of Ability  Y  2  = Interpersonal Self-Esteem (TSCS Total P)  ]  = Social Desirability  X  X  2  - Self Criticism  X^ = Masculinity X^ = Femininity Xg  = Androgyny  Xg  = Inter/Personal Self-Esteem (for Y-j) = Self Concept of Ability  bg  (for Y ) 2  = 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: f i r 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, i f 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 f i r s t stated. Because the same statistical procedures were used to test simultaneously the f i r s t three hypotheses, a brief summary of them will introduce the f i r s t 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 i t 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  Inter/Personal Self-Esteem  Self Concept of A b i l i t y  Self Criticism  99.67 9.45  348.07 27.70  58.33 6.26  36.00 5.55  0.25 2.34  104.73 8.50  355.67 38.52  59.47 7.74  35.87 6.99  0.80 2.36  96.33 8.61  339.60 24.65  55.13 8.30  36.07 3.90  0.40 1.87  101.60 10.25  355.27 41.60  63.33 6.97  35.07 6.06  0.06  100.93  350.40  52.80  34.20  1.51  10.53  44.05  7.01  5.48  0.30 2.51  102.20 7.28  357.40 20.41  60.13 6.63  33.00 4.55  Femininity  T.OO 1.65  Masculinity  Social Desirability  Means and Standard Deviations of Career Groups on Self Concept Variables  Teaching n=15  x SD  4.66 0.60  5.12 0.51  +  Nursing n=15  x SD  4.86 0.75  4.91 0.58  +  Arts n=15  x SD  4.47 0.77  4.81 0.56  +  Science n=15  x SD  4.85 0.61  5.07 0.56  +  Agriculture n=15  x  4.61  4.66  +  SD  0.65  0.38  Math n=15  x SD  4.68 0.81  4.75 0.70  XI  Career Groups  'Z >  1  1  * >> E  >>  CD O S-' T3 • E • • • • =£  +  *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 i s , 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 selfesteem.  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. Self Concept of Ability 2. Social Desirability  1  2  3  1.00  .26*  .24*  -.02  .43**  .63**  -.23*  .26*  .13  -.12  -.37**  .35**  .17  -.17  -.03  -.21  -.10  1.00  .05  -.75**  1.00  .59**  1.00  1.00  3. I/P SelfEsteem  4  1.00  4. Self C r i t i cism 5. Masculinity 6. Femininity  5  6 29**  ** |*|  -.12  1.00  7. Androgyny  * Ul  7  >  .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  Increment in R  2  Variable Symbol & Name  R  X-, Social Desirability  .3990  .3990  2.  X£  Self Criticism  .4553  .0563  3.  X^  Masculinity  .4941  .0388  1.  2  1  Variables listed are significant to thea = .05 level  ACHIEVEMENT AREA: SELF CONCEPT OF ABILITY Order of Predictor Variables  Variable Symbol & Name  1.  Xg  2.  X  5  Masculinity Androgyny  R  2  Increment 2 in R  .1815  .1815  .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 i s : 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 f e l l 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 a l . , 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 f i r s t 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 a l . , 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. Principal Components Analysis  Procedures used were:  69 Image Factor Analysis from the AGFA Program (Hakstian, 1972) Orthogonal rotation with varimax transformation was used with both factor analytic models. also used.  With the Image model, oblique transformation was  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 reduced as the smaller factors were collapsed.  factors was  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 i n i t i a l hypotheses, a separate set of analyses identical to the f i r s t 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 a l l 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 x 2  2  +  b x 3  3 +  b x 4  4  + e where Y = Self Concept of Ability x  l  —  Social Desirability  h = Self Criticism h = Androgyny h Inter/Personal h = BSRI Factor I \ X  X  b  7 8  0  e  = BSRI Factor II  = BSRI Factor III = BSRI Factor IV = constant  error  Self-Esteem  = b x 5  5  +  b x 6  6  +  b x 7  7  = b x 8  8  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 population  s  insofar as i t 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 f i r s t 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 Masculinity Scores  12  T  11 10  1  1  r  5  T  1  1  n = 90  9  o  6 5  be  CO +-> o (J o <L> •<—> -Q 7 zs CO  4  i~ E  3  I. ~3~  4 Femininity Scores  r  TABLE V Percentage of Androgynous Subjects in Categories Formed According to "Median Split" Procedure  Androgynous S's N - 34  Low Masculinity Low Femininity  41.18% (n=14)  High Masculinity  11.76%  Low Femininity  (n=4)  High Masculinity High Femininity  47.06% (n=16)  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  1. Self Concept of Ability  54.64  7.56  60.75  1.26  61.56  6.49  2. Social Desirability  97.86  7.89  106.50  9.33  104.63  8.94  394.44  40.42  3. Inter/Personal 344.00 21.69 Self-Esteem 4. Self Criticism  35.71  5.64  367.25 15.50 30.25  2.63  36.19  Significance Test F (2,31)  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 i s , having a magnitude greater than .40 are listed in Table VIII.  From this table i t can be seen that while the f i r s t 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 f i r s t 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 Feminity Scale  Original BSRI Masculinity Scale Variable #  Variable # 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  78  FIGURE 3 10.  PRINCIPAL COMPONENTS ANALYSIS SCREE TEST (Graph of Eigenvalues against Number of Eigen values)  10  3  5  i i i i i i i i i i l t I—i i I—n—n—i 1 5 10 15 ZO  i—i i i 25  Number of Eigenvalues  t  I  I i l 30  i i i i i i i 35 40  TABLE VIII Principal Components Solution Transformed Factors Variables Masc. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20  .058 -.088 .126 -.006 .134 -.053 .060 -.069 -.293 -.123 -.132 -.365 -.072 .043 -.400 -.171 -.262 -.198 -.126 -.389  .642 .080 .674 -.180 .662 .252 .167 .063 .205. .641 .198 .742 .073 .840 .129 .112 .150 .673 .357 .304 -.002 .343 .789 .074 .136 .725 .289 -.052 .108 .555 .742 .059 .567 .619 .728 .130 .099 .141 .168 .190  Fern. 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40  -.327 -.553 -.071 -.560 -.269 -.515 -.417 -.711 -.746 -.762 -.791 -.764 -.324 -.744 -.771 -.074 -.122 -.381 -.509 -.794  -.299 .272 -.329 .176 .285 .105 .012 -. 088 .096 .065' .254 -.036 -.476 .108 .091 .006 .183 -.080 .105 -.104  :  I  II  III  IV -.094 .067 -.091 -.011 -.358 -.136 .076 .067 -.205 -.022 -.255 .005 .238 -.034 -.053 -.026 -.128 -.079 .148 .081  .072 -.012 .249 -.012 .777 -.037 -.022 -.140 -.060 -.020 .151 .063 .315 .337 .381 .090 .190 .100 .172 .117 .150 .053 .131 -.017 .229 .037 .200 -.052 -.081 -.201 .057 -.015 -.248 .701 .481 .047 .019 -.001 .033 -.063  V -.037 .105 .286 .106 -.105 .153 -.034 -.021 -.011 -.1.26 .195 .038 -.108 .026 .040 -.065 -.095 -.121 -.131 .067  VI -.164 .033 -.208 .127 -.067 -.104 -.011 .027 .082 .132 .037 -.022 .293 .786 .093 .247 " .180 .243 .066 .314  VII  VIII  IX  -.430 -.209 -.065 .167 -.164 -.041 .043 -.761 -.089 -.221 -.215 -.008 .166 -.005 -.171 .274 -.121 .028 -.144 .073  -.158 -.044 -.227 -.770 -.052 .033 -.007 .082 -.270 -.435 -.147 -.013 .002 -.177 .183 -.022 -.289 .166 -.024 -.182  -.020 .095 .135 -.192 .086 -.161 .123 -.136 .036 -.02 -.610 -.129 -.069 -.037 -.327 -.153 -.104 -.127 -.165 .247  -.503 -.194 .011 -.136 -.370 .057 -.007 .129 .014 -.240 , .131 -.151 .011 -.187 -.074 .044 -.118 .152 -.207 .087  .149 -.085 -.281 -.187 .164 -.070 -.233 -.308 .154 -.204 -.054 .005 -.196 -.010 -.008 .038 -.216 .056 -.044 .138 .017 -.077 -.081 -.031 .286 -.337 .179 -.067 -.129 -.286 -.013 -.058 -.029 -.109 .047 -.086 -.494 .022 .134 -.250  -.391 -.002 -.109 .240 -.050 -.101 -.094 .043 -.016 .280 -.041 -.156 .255 -.416 -.210 .023 .500 -.161 .111 -.093 -.452 -.060 .135 -.159 .020 -.095 .079 .139 -.154 -.009 .044 -. 050 -.077 -.345 -.078 -.038 .036 -.086 -.165 .117 .045 -.006 .072 -.078 -.086 -.022 .042 -.912 .101 -.296 -.024 -.319 .133 -.070 -.237 -.003 -.145 .060 -.043 .034  X .171 .090 -.099 .011 -.084 -.081 -.040 -.148 -.075 -.317 -.192 -.059 -.095 -.119 -.234 -.101 -.076 -.178 -.793 -.507  * -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 f i r s t 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 f i r s t 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 r e l i a b i l i t y 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** Variable 1 2 3 4 5 6 7 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  1 .014 -.005 .076 -.162 .127 -.029 .047 -.078 -.303 -.310 -.201 -.434 -.153 .001 -.386 -.211 -.313 -.285 -.344 -.479 -.493 -.651 -.150 -.623 -.432 -.521 -.507 -.708 -.700 -.746 -.717 -.831 -.360 -.729 -.775 -.274 -.290 -.355 -.596 -.731  4-Factor 2 3 .007 .644 .170 .187 .575 .675 .796 .155 .650 .356 .031 .062 .795 .515 .595 .781 .639 .205 .253 .312 -.279 .240 -.265 .140 .214 .251 -.104 -.029 .102 .091 .236 -.049 -.436 .098 .076 .034 .236 -.055 .111 -.071  .689 -.049 .696 .076 .337 .366 .044 .351 .246 .349 .613 .669 -.040 -.126 .268 -.009 .136 .549 .117 .047 -.153 .221 -.284 .041 -.202 -.128 .226 -.112 .119 .071 .061 -.113 -.032 .110 .023 -.309 -.521 .026 .078 .045  4 .048 .089 -.085 -.486 -.262 -.093 -.045 .416 -.239 -.155 -.127 .669 .035 -.051 .126 -.196 -.217 .074 .111 .038 -.188 .221 .604 -.300 -.429 .185 .117 .409 .246 .145 .273 .005 .334. -.077 -.175 -.130 .251 .609 -.099 .151  1 -.005 .652 -.172 -.198 -.574 -.694 -.799 -.141 -.658 -.321 -.007 -.007 -.788 -.504 -.627 -.797 -.647 -.172 -.230 -.301 .323 -.230 .339 -.137 -.197 -.261 .166 .079 -.100 -.063 -.229 .093 -.493 -.080 -.075 .021 -.182 .104 -.115 .101  2  5-Factor 3  -.004 -.083 -.021 -.072 .099 -.207 -.051 -.058 -.321 -.003 -.053 -.223 .031 .250 -.261 -.060 -.171 .051 -.102 -.235 -.210 -.511 .107 .535 -.212 -.179 -.353 -.567 -.689 -.637 -.708 -.742 .161 -.629 -.752 .055 -.050 -.274 -.534 -.697  .742 -. 054 .772 .150 .341 .335 .103 .385 .269 .462 .512 .751 .009 -.118 .196 -.041 .156 .616 .229 .128 -.035 .216 -.182 -.023 -.101 -.245 .251 -.130 .064 .066 -.014 -.093 .010 .106 -.049 -.201 -.393 .057 .069 .019  4 .079 .131 .025 -.348 -.253 -.112 .022 .375 -.174 -.003 -.259 -.039 .044 -.030 -.031 -.245 -.194 -.052 .171 .121 -.001 -.323 .679 -.412 -.277 .190 .084 .441 .175 .168 .266 .083 .376 -.132 -.260 .056 .433 .619 -.055 .123  5 -.051 .051 -.084 -.414 -.063 .112 -.124 .054 -.181 -.409 .238 .078 -.125 -.091 .290 -.024 -.194 .126 -.306 -.311 -.625 -.121 -.133 -.088 -.575 .080 -.019 -.003 .155 -.089 .145 -.326 -.110 -.153 -.066 -.478 -.396 .094 -.147 .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** Variable 1 2 3 4 5 6 7 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  1 .077 -.684 -.093 .024 -.424 -.679 -.875 -.199 -.468 -.169 .162 .077 -.849 -.390 -.583 -.731 -.453 -.080 -.184 -.247 .440 -.089 -.041 .024 -.043 -.320 .079 -.113 -.171 -.069 -.365 .121 .372 .015 .104 .057 -.419 -.126 -.024 .106  4-Factor 2 3 .010 .087 .023 -.359 -.004 -.005 .100 .105 -.408 -.149 -.076 -.112 .127 .108 -.038 -.058 -.330 .099 -.122 -.229 -.538 -.524 .393 -.586 -.504 -.004 -.176 -.223-.345 -.452 -.288 -.724 -.066 -.577 -.789 -.176 .073 .090 -.509 -.505  .651 -.114 .676 .013 .216 .268 -.050 .270 .125 .260 .504 .756 -.135 -.190 .165 -.141 -.014 .580 .055 -.004 -.196 .179 -.200 -.027 -.261 -.213 .229 -.115 .094 .025 -.001 -.165 .039 .085 -.068 -.302 -.507 .056 .002 .016  4 -.051 -.116 -.033 .374 .280 -.068 -.068 -.283 .194 .138 .137 -.148 -.001 .241 -.105 .187 .294 .004 -.043 -.064 .159 .064 -.589 .144 .315 -.192 -.264 -.591 -.440 -.324 -.546 -.253 -.330 -.093 .019 .131 -.311 -.594 -.078 -.325  1 .102 -.652 -.059 -.113 -.465 -.600 -.880 -.122 -.524 -.238 .224 .166 -.882 -.461 -.470 -.766 -.544 -.008 -.196 -.273 .270 -.091 .007 .012 -.211 -.278 .138 -.017 -.046 -.027 -.220 .106 .417 .024 .119 -.067 -.471 .001 -.022 .201  2  5-Factor 3  .100 .033 .086 -.155 .029 -.135 .146 .079 -.325 .035 -.214 -.138 .202 .161 .276 -.076 -.220 .020 -.060 -.134 -.191 -.529 .504 -.648 -.249 -.077 -.179 -.251 -.487 -.432 -.454 -.625 -.033 -.563 -.889 .047 .266 .024 -.498 -.640  .752 -.110 .757 .131 .223 .192 .052 .312 .181 .420 . 343 .758 -.017 -.160 -.016 -.158 .051 .518 .128 .102 .066 .152 .022 -.136 -.090 -.238 .265 -.044 .030 .087 -.059 -.062 .107 .094 -.203 -.123 -.251 .106 .008 -.082  4 -.065 -.138 -.039 .247 .241 -.015 -.168 -.265 .126 .021 .280 -.069 -.128 -. 160 .031 .140 .196 .089 -.091 -.139 -.044 .092 -.695 .209 .135 -.178 -.234 -.577 -.331 -.324 -.445 -.296 -.321 -.075 .120 -.034 -.511 -.556 -.068 -.203  *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  5 .091 -.114 .042 .335 .021 -.236 .013 -.079 .134 .255 -.206 -.071 .061 .063 -.373 -.041 .175 -.161 .087 .143 .583 .033 .113 -.203 .433 -.112 -.003 -.032 -.192 .061 -.238 .220 .060 .072 -.053 .358 .268 .126 .063 -.145  84 Description of Factors Factor 1. - This factor is characterized by variables such as V7 — forceful (M) VT3 - dominant (M) VI3 - aggressive (M)  and  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 i t s 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 which do not share the "nurturing" sense. V  23 -- shy  (F)  V  37 -- childlike  (F)  V  36 —  (F)  gullible  are typical of this group of six items.  scale  Terms such as :  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 i t 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  item  Factor I Dominance correlation with factor  item  Factor II Nurturance correlation with factor  35. tender (f)  .676  .699  32. eager to soothe hurt feelings (f)  .674  5. assertive (m)  .698  31. compassionate  .597  16. aggressive (m)  .683  29. sensitive to the needs of others (f) .596  .678  30. understanding (f)  .567  .668  34. warm (f)  .558  40. gentle (f)  .541  39. loves children (f)  .471  22. cheerful  .470  24. affectionate (f)  .460  9. has leadership abilities (m)  .714  7. forceful (m)  6. strong personality (m) 13. dominant (m) 17. acts as a leader (m)  .648  15. wi 11 i ng to ta ke a stand (m)  .571  2. defends own bel ief s 33. soft-spoken (f)  .470  14. masculine (m)  .323 BSRI items omitted from solution: 4. athletic 8. analytical 19. competitive 20. ambitious 21. yielding 26. loyal 27. feminine  Factor III Independence correlation item with factor  Factor IV Passivi ty correlation item with factor  3. independent (m) .657  37. childlike(f)  .484  1. self-reliant(m) .627  28. sympathetic (f)  .333  23  .321  12. selfsufficient (m)  .632  18. individual i s t i c (m)  .474  10. willing to take .430 risks (m) 11. makes decisions easily (m) .385  shy (f)  36. gullible (f) .247 38. does not use harsh language (f) .164 25. flatterable(f)- .01  cR  TABLE XII Intercorrelations Among the BSRI Factors  Factor Dominance  1.  _ 29**  Independence  .33**  *fi~>  3.  4.  1.00  Nurturance  Passivity  2.  -.06  1.00 .16  1.00  .22*  -.32*  1.00  .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 No. of Items  11  Nurturance 10  Independence 6  Passivi ty 6  : 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  mean  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 i s :  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 Dominance  Variables  Nurturance  Independence  Passivity  Self Concept Ability  Social Desirability  I/P Self Esteem  Self Criticism  Group Teaching X n = 15 S.D.  49.53 7.32  56.67 6.67  31.20 3.75  26.07 3.01  58.33 6.26  99.67 9.45  348.07 27.70  36.00 5.55  Nursing n = 15  X S.D.  48.87 9.40  55.07 7.89  32.00 4.83  24.20 3.67  59.47 7.74  104.23 8.50  355.67 38.52  35.87 6.97  Arts n = 15  X S.D.  46.33 10.72  51.20 5.75  29.67 5.58  25.80 5.73  55.13 8.30  96.33 8.61  339.60 24.65  36.07 3.90  Science n = 15  X S.D.  50.47 6.91  55.00 6.69  32.00 4.63  27.00 4.16  63.33 6.96  101.60 10.25  355.27 41.60  35.07 6.06  Agric. n = 15  X S.D.  46.87 7.88  51.33 7.42  30.33 4.59  23.07 4.38  52.80 7.01  100.93 10.53  350.40 44.05  34.20 5.48  Math. n = 15  X S.D.  47.93 9.47  51.60 9.20  30.87 5.91  23.80 6.13  60.13 6.63  102.20 7.28  357.40 20.41  33.00 4.55  TABLE XV Univariate and Multivariate Analysis of Variance for Career Groups on BSRI Factors & Self-Esteem Variables  jf?£Jitfi  Un  V  Source 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. separately in relation  These correlations will be discussed  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 Dominance  1 1.00  2  3  Nurturance  .29**  1.00  Independence  .33**  .16  1.00  .22*  -.32**  Passivity  -.06  4  5  8  1.00  .38**  .35**  .22*  .13  Social Desirability  .13  .37**  .39**  -.43**  .26*  Inter/Personal ; Self Esteem  .25*  . 28**  .38**  >  29**  .24*  Self Criticism  .01  -.03  .02  *kl **|*|  7  •  Self Concept of Ability  -.19  6  1.00  -.02  1.00 .63** -.23*  > .22 is significant at the ^ .05 level > .26 is significant at the « .01 level  Leading decimal points are omitted.  1.00 -.37  1.00  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)  j (^.= .217,p .05) respectively. <  a n c  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  Increment in R  2  Variable Symbol & Name  R  2  1.  X-|  Dominance  .1460  .1460  2.  X  Nurturance  .2078  .0618  2  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. i i ) 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 f i r s t of the two major issues, treated by testing Hypotheses one, two and three, will f i r s t 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 i t 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 a l l 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 ChoiceFindings 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 f i r 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 f i r s t potential source of deviance from expectations is in the sample i t s e l f .  Conclusions from previous research, on which the predictions  were based, would not be generalizeable to the present sample were i t found not to represent the same population.  Both cultural and historical factors  were considered. The studies on which the predicted differences were founded a l l 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 i s , 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 i t 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 i s , 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 i s , 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 i t 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 a f f i l i a t i o n 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 a l . , 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, i t 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 f i e l d , 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 i n i t i a l 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, i t 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. self concept.  Greater mental health was associated with a more masculine 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, i t 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 f l e x i b i l i t y 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 i s , 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 a l . , 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. the results in Table III indicate, only Masculinity showed the predicted  As  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 f e l t indirectly, through the Androgyny Score.  The lack of correlation between the Femininity  scale and either Social Desirability or Inter/Personal self-esteem was l e f t 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 i t s 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 a l . , 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 a l . , 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 in the case of the androgyny scores.  for the lack of correlations observed 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 a l . , scales showed that "concepts of health correspond to concepts of social desirability" (Broverman et a l . , 1972). their social desirability.  Items for the BSRI were chosen for  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  f l e x i b i l i t y 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 selfesteem 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  i t 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 i s , 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 appropriateness 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 positively a person evaluates the self she describes. see Table XX in Appendix C).  how  (For exact figures,  With the variable Identity, however, both  Masculinity and Femininity covary positively while neither "Masculinity" "Femininity"  help determine Behavioural Self scores.  nor  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 l i f e 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. and CIinical Psychology, 1974, 42, 2, 155-162.  Journal of Consulting  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 u t i l i t y 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 f a c i l i t i e s 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. W.H. Freeman, 1967.  San Francisco:  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 C l i f f s , 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 f a c i l i t i e s 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 f a c i l i t i e s 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. Personality, 1968, 36, 1, 33-47.  Journal of  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. H j e l l i , 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 consciousness. 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. 1971.  Man's world, woman's place.  New York:  Dell Pub!ishinging Co:,  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. 1973, May_, 82-106.  Medical Aspects of Human Sexuality.  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 f a c i l i t i e s 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. 277-280.  Elementary School Journal, 1962, 5_,  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 f a c i l i t i e s 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 C l i f f s , 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 - Test  137 140 ;  3. The General Self Concept of Ability Scale - Psychometric Data 142 - Test 4.  Indication of Career Plans  5. Tennessee Self Concept Scale - Psychometric Data - availability  143 145 146 148  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 i t 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 r e l i a b i l i t y estimate of the scales for the two norm groups to be: Stanford  Foothill  Masculinity  .86  .86  Femininity  .80  .82  Androgyny  .85  .86  Social Desirability  .75  .70  i i ) Test-retest r e l i a b i l i t y - Over a four week period, the r e l i a b i l i t y 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 Desirability  = .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 f i r s t 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 f l e x i b i l i t y 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 are the other scales.  of discriminating males from females, as  (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 i s , 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 i t 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 i t is SOMETIMES BUT INFREQUENTLY TRUE that you are sly. Mark a 4 i f i t 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 i t 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  14V  DESCRIBE YOURSELF 1  NEVER OR ALMOST NEVER TRUE  USUALLY NOT TRUE  3 SOMETIMES BUT INFREQUENTLY TRUE  OCCASIONALLY TRUE  OFTEN TRUE  USUALLY TRUE  ALWAYS OR ALMOST ALWAYS TRUE  Self reliant  Reliable  Warm  Yielding  Analytical  Solemn  Helpful  Sympathetic  Willing to take a stand  Defends own beliefs  Jealous  Tender  Cheerful  Has leadership abilities  Friendly  Moody  Sensitive to the needs of others  Aggressive  Independent  Truthful  Gullible  Shy  Willing to take risks  Inefficient  Conscientious  Understanding  Acts as a leader  Athletic  Secretive  Childlike  Affectionate  Makes decisions easily  Adaptable  Theatrical  Compassionate  Individual i s t i c  Assertive  Sincere  Does not use harsh language  Flatterable  Self-sufficient  Unsystematic  Happy  Eager to soothe hurt feelings  Competitive  Strong personality  Conceited  Loves children  Loyal  Dominant  Tactful  Unpredictable  Soft-spoken  Ambitious  Forceful  Likeable  Gentle  Feminine  Masculine  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. 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 f i r s t class b) mostly high second class c) mostly low second class d) mostly pass e) mostly f a i l  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.  In your own  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 a l l  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) i t doesn't matter to me at all  13.  Which statement a) getting b) getting c) getting d) getting e) I don't  best describes your most comfortable or typical position: better marks than everyone else better marks than almost everyone else about the same marks as everyone else slightly lower marks than everyone else care about any particular marks.  145 14.  Academically, do you try to do better than others a) a l l 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 l i f e 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 l i f e are more important e) good marks are the least important area of my l i f e  16.  What level of academic achievement do you aim for? a) mostly f i r s t class b) mostly high second class c) mostly low second class d) mostly pass e) mostly f a i l  CAREER PLANS  (4) 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 Yes  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. Yes  3.  No  No  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 r e l i a b i l i t y 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 r e l i a b i l i t y coefficient =.88  for the Total Positive score.  For another sample, Fitts reports a  r e l i a b i l i t y 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 B i l l s ' 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. F3  Gr. Fl  Gr. F2  Gr. M3  Gr.M2 Agric.  2.3333  5.5333  6.6667  7.3333  Gr. F3 Arts  -  3.200  4.334  5.0000  8.200*  1.1334  1.800  5.000  -  0.666  3.866  Gr. M2  -  Gr.Fl Teach. Gr. F2 Nurse.  -  Gr.M3 Math  Gr. Ml * 10.53  3.200  -  Gr. Ml Science  Minimum difference required = 7.64 * difference is significant to = .05 level. a  151  APPENDIX C  Tables Corresponding to "Additional Recommendations for Further Study." PAGE Table XIX - Means & Standard Deviations of Low Self-Esteem Third of each Career Group on remaining variables. Table XX  - Intercorrelations among three Components of Inter/Personal Self-Esteem and original BSRI  Table XXI - Intercorrelations among three Components of Inter/Personal Self-Esteem and Four Factors of BSRI. Table XXII  Means and Standard Deviations of Career Groups on "Identity" Component of Self-Esteem Score and BSRI Scales.  Table XXIII - Univariate & Multivariate Analysis of Variance of Career Groups on "Identity" Component of Self-Esteem Score and BSRI Scales. Table XXIV  Table XXV  Table XXVI  Table XXVII  152  153  154  155  156  Means and Standard Deviations of Career Groups on "Behavior" Component of Self-Esteem Score and BSRI Scales.  157  Univariate & Multivariate Analysis of Variance for 6 Career Groups on "Behavior" Component of Self-Esteem Score and BSRI Scales.  158  Means and Standard Deviations of Career Groups on "Self Acceptance" Component of Self-Esteem Score and Two BSRI Factors.  159  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.  54.20 9.03  44.60 8.82  48.40 5.90  46.20 4.32  44.60 7.70  37.80 7.95  54.40 3.78  52.20 8.53  51.00 7.25  53.00 3.16  50.20 6.87  42.40 7.63  28.80 5.17  30.00 4.80  26.00 6.00  28.80 5.97  30.80 2.68  28.00 7.65  25.20 1.92  24.20 4.32  26.80 7.79  27.20 3.34  25.20 2.59  24.40 6.19  56.80 5.89  61.20 10.83  55.80 7.40  58.00 2.35  52.80 7.19  57.20 8. 01  93.20 7.82  100.20 7.40  92.20 5.07  92.20 10.47  94.20 7.83  99.00 4.64  37.40 5.18  43.00 3.46  38.00 5.70  38.60 1.34  37.20 5.81  31.40 5.73  Dominance X  SD Nurturance X  SD Independence X  SD Passivity X  SD Self Concept Abi 1. X  SD Soc. Desir. X  SD Self.Crit. X  SD  TABLE XX Intercorrelations Among 3 Components of Inter/Personal Self-Esteem and original BSRI  1  Variable  2  4  3'  :  5  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  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*  7 Masculinity  0.430**  0.324**  0.269*  0.052  8  0.295*  0.289**  0.019  0.075  0.011  0.036  Femininity  9 Androgyny  6  7  8  9  1.000  1. 000  0.261* -0.030  1.000  -0.130  0.132  -0. 206  0.053  1.000  -0.041  0.045  -0. 025  -0.354  0.229  * = significant to <* .05 level ** = significant to <= .01 level (leading decimal points are omitted)  1.000  TABLE XXI Intercorrelations Among 3 Components of Inter/Personal Self-Esteem 4 Factors of Bern Sex Role Inventory Variable  2  1  3  4  5  6  7  1.000  1  Identity  2  Self Acceptance  .636**  3  Behaviour  .040  .016  1.000  4  Dominance  .176  .198  -.058  5 Nurturance  .257*  .072  -.138  .286**  1.000  6  Independence  .382**  .281**  .094  .325**  .163  1.000  7  Passivity  -.323**  -.061  .219*  -.320**  -0.199  1.000  1.000  -.062  * = significant to«:.05 level ** = significant to <* .01 level (leading decimal points are omitted)  1.000  TABLE XXII  Femininity  Androgyny  Social Desirability  Self Concept of a b i l i t y  Self Criticism  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  4.86 0,75  4.91 0.58  0.25 104.73 2.34 8.50  126.73 9.44  59.47 7.74  35.87 7.00  X  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  4.85 0.61  5.07 0.56  0.40 101.60 1.87 10.25  125.60 13.06  63.33 6.97  35.07 6.06  Agriculture n = 15  X  4.61 0.65  4.66 0.39  0.06 100.93 1.51 10.53  122.27 12.24  52.80 7.01  34.20 5.48  Mathematics n = 15  X  4.68 0.81  4.75 0.70  0.30 102.20 7.28 2.51  126.33 8.41  60.13 6.63  33.00 4.55  Career Groups Teaching n = 15  X  Nursi ng n = 15  X  Arts n = 15  SD  SD  SD  SD  SD  SD  Identity  Masculinity  Means and Standard Deviations of Career Groups on "Identity" Component of Self-Esteem Score and BSRI Scales  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  Androgyny  Social Desirability  Behavior  Self Concept of A b i l i t y  Self Criticism  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  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  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  Masculinity  Femininity  Means and Standard Deviations of Career Groups on "Behavior" Component of Self-Esteem Score and BSRI Scales  Career Groups Teaching n = 15  X  Nursing n = 15  X  SD  SD  Arts n = 15  X  Science n - 15  X  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  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  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  SD  SD  SD  SD  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  49.53 7.32  56.67 6.67  97.67 9.45  Nursing n = 15  X  48.87 55.07 9.40 . 7.90  104.73 8.50  Arts n = 15  X  46.33 10.72  51.20 5.75  Science n = 15  X  50.47 6.91  Agriculture n = 15  X  Mathematics n = 15  X  SD  SD  SD  SD  SD  SD  Self Criticism  Social Desirability  X  Self Concept of a b i l i t y  Nurturance  Teaching n = 15  Career Groups  Self Acceptance  Dominance  Means and Standard Deviations of Career Groups on "Self Acceptance" Component of Self-Esteem Score and Two BSRI Factors  111.27 • 58.33 12.73 6.26  36.00 5.55  117.27 19.53  59.47 7.74  35.87 7.00  96.33 8.61  111.13 12.24  55.13 8.30  36.07 3.90  55.00 6.69  101.60 10.25  114.93 18.68  63.33 6.97  35.07 6.06  46.87 7.88  51.33 7.42  100.93 10.53  115.80 19.97  52.80 7.01  34.20 5.48  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  Dominance  Univariate F (5,84)  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. 1928 p < 0.2291  APPENDIX D TABLE XXVIII CORRELATIONS AMONG ITEMS OF THE BSRI  162  TABLE XXVIII Correlations* among items of the BSRI 1 X  5.7444  S.D. 1.2958 . n Var. 1. 2,  3 4 5 6 7 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  90  2  3  6.0000  5.7111  0.9068 90 .0851  5  6  7  8  3.8222  4.5778  5.2667  4.1111  5.2889  1.0354 1.7737  1.4831  1.3233  1.3699  1.2583  90  90  90  90 .4253 .0710  4  90 .1011 .1520 .1777  90 .2618 .3305 .2968 .1658  .1888 .4260 .3806 .1906 .4900  .0786 .4473 .2655 .1956 .5919 .6211  .2838 .0487 .1152 -.0566 .1725 .1806 .1554  163  TABLE XXVIII (cont'd) 9  10  11  12  13  14  15  16  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  90  90  90  90  -.1176 .2302 .0102 .2419 .1521 .1700 .2224 -.0008 .2965 .1700 .0043 -.1443 .3327  .0773 .5212 .2115 .0094 .3884 .5615 .4416 .2720 .4580 .2003 .2609 .1492 .3321 .1981  -.0644 .4235 .1579 .2222 .4405 .5565 .6633 -.0521 .5200 .2233 .0200 .0918 .6768 .3020 .4308  X n Var. 1 2 3 4 5 6 7 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  90  90  90  .2707 .4133 .3060 .2700 .4352 .5377 .4989 .1496  .2622 .1532 .4144 .2844 .3531 .3702 .2692 .2112 .3242  .3247 -.0084 .2980 .1241 .1866 .2484 -.0590 .2184 .1185 .2204  90  .5019 -.0160 -.0337 .3080 .6102 .1917 .0873 .1582 . 1768 .3992 .2742 . 5233 .1122 .6882 .1608 .0583 .2805 .5114 .2807 .2123 .2751 -.0673 .1010  164  LE XXVIII (cont'd) 17 X-  18  19  20  21  22  23  24  3.7667  5.4444  .3444  5.2889  4.1778  5.3889  3.7000  5.3889  S.D. 1.3170  1.0553  .4465  1.3017  1.4266  0.9750  1.5737  1.2710  n Var. 1 2 3 4 5 6 7 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  90  90  90  90  90  .1734 .3815 .2276 .3152 .4729 .4119 .4085 .1211 .7413 .3105 .1116 . 0940 .5344 .3254 .4296 .5250  .4243 .0232 .4633 .0838 .2406 .3368 .2041 .1460 .2205 .1764 .2178 .4670 .1184 .0849 .2312 .1540 .0906  .0825 .0593 .1629 .0542 .1040 .1668 .1769 .2200 .1974 .2765 .0939 .0629 .1913 .1177 .2245 .1671 .1763 .1472  .0174 .1506 .2186 .1474 .1265 .2069 .2499 .0033 .1575 .2457 -.0275 .0929 .2419 .2398 .1494 .2133 .2143 .1330 .3779  .0536 -.1890 -.0555 .1662 -.1326 -.3076 -.1011 .0147 -.0159 .1225 -.0588 -.1157 -.1597 -.1212 -.2877 -.1732 -.0844 -.2296 .0242 .1399  90  .1051 .0755 .2216 .3037 .2675 .4281 .2841 -.0282 .4986 .1325 .2230 .3591 .2295 -.0644 .2482 ..2540 .2353 .0480 .1573 .2006 .0062  90  -.1193 -.1246 -.1282 -.1783 -.3970 -.3244 -.1701 .0325 -.4530 -.2104 -.2513 -.2036 -.2096 -.1658 -.3403 -.4044 -.3769 -.2676 -.0522 -.0228 .1326 -.2936  90  .0064 .0578 -.0328 .2278 .1225 .2819 .1539 -.1189 .2860 .1259 .0934 .1392 .1105 -.0867 .2367 .2525 .2202 -.0377 .1749 .1537 .0477 .3984 -.2861  165  TABLE XXVIII (con 25  -26  -27  28  d) 29  30  31  32  4.0000  6.1222  5.0889  5.7222  5.5222  5.5778  5.3222  5.5778  S.D. 1.4682  0.8797  1.2261  .9193  1.0773  .8942  1.0732  1.1448  90  90  90  .1562 .1782 .0603 .1628 -.0423 .0670 .0564 .1084 .2480 .1713 -.0124 .1567 .0458 -.0130 .1488 .0270 .1994 -.0484 -.0078 .1430 .2069 .1757 .0363 .2227 .1016 .0797 .1356 .4251 .5634  .0352 .2512 .0738 -.1217 .0576 .2759 .2477 .1039 .2526 .0997 .0428 .1206 .1173 -.1174 .2907 .1884 .1239 -.0382 .1003 .2198 -.0011 .2840 .0046 .3317 .0705 .2643 .2485 .4737 .4791 .4428  X n Var. 1 2 3 4 5 6 7 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  90 -.0759 .2253 .1096 .2517 .1276 .1430 .2486 -.1684 .3385 .2051 -.1651 -.1247 .1440 -.0582 .0738 .2059 .3103 -.0287 .1465 .2035 .1963 .2331 -.1491 .2501  90 -.0798 .2229 -.1808 .0780 -.0115 .1247 .1731 -.0319 .1322 .0812 -.0580 -.1420 .1105 .1480 .1017 .1685 .0342 .0013 -.0593 .0953 .0712 -.0425 -.0538 -.0346 -.1118  90 .1961 - .0300 .1953 .0226 - .0588 .0744 .0140 - .0311 .0340 - .0452 .0400 .2314 - .0929 - .4652 .1458 - .0908 .0197 .0897 .0266 .0788 .1307 .1944 .1060 .1489 .1420 - .0719  90  - .0409 .0267  - .0960 - .1461  - .2653  - .0944 - .0373  .1558 .0202 - .1591 - .1428 - .0744 .0105 - .1474 .0095 - .1694 - .0251 - .0904 .0135 .1599 .1054 .1703 .2727 .0164 .0823 .1794 .1994  90 .1274 .1024 .0854 -.0153 -.1471 .2452 .0962 .1346 .2744 .0104 . 0342 .1579 .0734 -.1170 .2220 .0557 .1015 .0206 .0842 .1380 .0336 .3147 -.0321 .2655 .0773 .1788 .1920 .3821  .0353 - .0214 - .0935 .0779 - .1246 .0450 - .0055 .1155 .1154 .0191 - .1156 .1668 - .0368 - .1593 .0216 .0496 .0968 - .1298 .0811 .1490 .3929 .3366 .0407 .3190 .2446 .0402 .2167 .5115 .3950 .4880 .4634  _  166:  TABLE XXVIII (cont'd) 33  34  35  36  37  38  39  40  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  90  90  90  90  90  90  90  -.2124 .0763 -.0365 -.1347 .1128 -.0854 .1023 .1661 -.0482 -.0021 .0076 -.0982 -.0558 -.0445 .3548 -.0341 -.0284 .2533 .2044 -.1120 .1516 -.1438 .1160 .0431 -.0525 -.0871 .0723 -.2153 -.2157 -.2485 -.1258 -.0182 .1818 -.1313  .1613 -.0334 -.2538 -.3192 .1383 .1239 -.2008 .0852 -.0428 .0064 .0737 .0634 -.1361 .1814 .1877 -.1687 -.0482 .3079 .0507 -.1908 -.0925 -.0961 -.4298 -.0474 -.1380 .2105 -.2629 .0088 -.0507 -.1096 .0347 .2118 .0251 -.1050 .0951  .0158 -.0913 .0058 .1399 -.2300 -.0432 -.1383 -.0052 .0417 . 1076 .3497 .1112 -.0522 .0557 -.0440 .0239 -.1482 .0167 .0124 -.0398 -.2772 .2649 .0336 .2060 -.2471 .0240 -.0835 -.1676 .0254 -.0841 -.0237 -.1249 .0330 .1040 .0380 -.1258  -.1291 -.0049 -.1095 .1893 .0386 -.0076 -.2239 -.1863 .1052 .0360 -.1190 .0088 .0685 .1400 .0545 .0186 .1987 -.0504 .0422 -.0719 .0218 .0199 -.1932 .1552 -.1327 -.1850 .1012 -.3648 -.0580 -.0626 -.1540 -.0319 .0562 -.0751 .1191 .0137 -.1287  -.0169 -.0741 -.0859 .0719 -.0865 .0931 -.1773 .0684 -.1175 .0082 .2868 -.1244 -.0587 .1042 .0841 -.0695 .1261 .0132 .0544 -.1795 -.0139 -.0696 -.0376 -.2523 -.1546 .0986 -.1248 -.0061 .1797 .0070 -.1868 .1863 -.1750 -.3471 -.1569 .2261 -.0126 -.0165  .0140 -.0691 .0681 -.1604 .1023 .0583 .1847 -.0835 -.0496 -.0647 -.1190 -.0399 .0140 -.3042 -.1415 .0779 .0780 .1445 .1019 .1617 -.2295 -.0803 -.1208 -.1075 .2214 -.0033 .0597 -.0402 -.2778 -.3125 .0168 -.0843 .0201 -.0584 .0246 .1540 -.0266 -.0059 -.1702  X n  90  Var. 1 2 3 4 5 6 7 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  -.1824 .1875 .-797 -.0452 .2271 .2141 .0218 -.1458 .1944 -.0623 -.0053 .0082 -.0438 -.1581 .1405 -.0596 .0337 -.0045 .0305 -.0785 -.1160 .1357 .1241 .2248 .0944 .1084 -.0219 -.2861 -.1782 -.2913 .0488 -.2022  -.0402 -.0476 .2253 -.0276 .1295 .0897 -.1766 -.0009 -.0234 -.0695 -.3279 .1234 .0440 .0681 -.1577 -.0359 -.1034 -.0218 -.2467 -.0489 -.0983 -.0232 .0798 -.0645 -.0435 -.0408 -.0248 -.0289 -.1618 .0953 .0366 -.1205 .0574  


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