Sex Roles, Vol 25, Nos. 3/4, 1991 
Sex Role Identity and Self-Esteem: A Comparison 
of Children and Adolescents 1 
Betty Allgood-Merten and Jean Stockard2 
University of Oregon 
While data from a sample of fourth-grade children indicate that both self-ef-
ficacy (masculinity) and relationality (femininity) are strongly associated with 
self-esteem for both girls and boys, results from a comparable sample of ado-
lescents aged 14-18 and a subsample of the fourth graders when they were in 
twelfth grade indicate that only self-efficacy is generally associated with self-
esteem. An association between relationality and self-esteem is found in ado-
lescents only for a small, high-achieving, high-SES subset of "androgynous" 
males. 
Literature that examines the relationship between sex role identity and psy-
chological well-being has included three models: the traditional sex-typed 
model (e.g., Abraham, 1911/1949, Kagan, 1964; Mussen, 1969), the androg-
yny model (Bern, 1974, 1979; Spence & Heimrich, 1978), and the mascu-
linity model supported by recent meta-analytic studies (Bassoff & Glass, 
1982; Taylor & Hall, 1982; Whitley, 1983, 1984). Each of these models 
assumes that stereotypic masculinity is measured by traits generally asso-
ciated with self-efficacy and femininity is measured by traits associated with 
expressiveness and relationality. [Given recent criticisms of these scales, and 
especially, questions concerning the extent to which they accurately reflect 
theoretical and empirical differences between males and females at a va-
riety of ages (e.g., Gill, Stockard, Johnson, & Williams, 1987), we use the 
1We are grateful to the Center for the Study of Women in Society for support for the research 
reported in this article, to Jeanne McGee for providing access to part of the data used in 
this study, and to Robert O'Brien for statistical advice. Any opinions expressed in the article, 
however, are our own. 
2To whom reprint requests should be addressed at Department of Sociology, University of 
Oregon, Eugene, OR 97403. 
129 
0360-0025;91/0800-0129$06.50/0 i:J 1991 Plenum Publishing Corporation 
130 Allgood-Merten and Stockard Sex Role Identity and Self-Esteem 131 
term "self-efficacy" to describe scores on the scales traditionally termed theories, leading us to hypothesize that associations of sex-typed attributes 
masculine and relationality to describe traits measured by traditional femi- with self-esteem in adolescents will be similar to those reported in the pre-
ninity scales.] The traditional model assumes that males' identification with ponderance of studies of adults and college students, to wit: that 
efficacy and females' identification with relationality are paramount to self-efficacy is positively related to psychological health in both females and 
mental health. In contrast, the androgyny model asserts that both males males, and that relationality is only minimally related, if at all. 
and females high in efficacy and relationality are more likely to be psycho- Associations of sex-typed characteristics with psychological well-being 
logically well adjusted than either undifferentiated individuals who report in prepubertal children are even less clear. The literature suggests that ado-
low identification with both or those who are identified with only their own lescent patterns of sex differences in psychopathology differ dramatically 
sex-stereotypic characteristics. The masculinity model, on the other hand, from those observed in children (Rutter, 1986). For example, the 2:1 ratio 
views self-efficacy as the critical component in mentally healthy individuals of females to males in depressed adults has also been found to exist in 
of both sex groups. depressed adolescents (Lewinsohn, Hops, Roberts, & Seely, 1988; Schoe-
The evidence in support of a masculinity model in college and adult bach, Garrison, & Kaplan, 1984), whereas in children depression occurs in 
populations is compelling (see Bassoff & Glass, 1982; Taylor & Hall, 1982; boys and girls with about equal frequency (Rutter, 1986). It is also during 
Whitley, 1983, 1984). It is not at all clear, however, whether such a model adolescence that eating disorders develop with an even higher female pre-
is applicable to children and/or adolescents. The present study investigates ponderance-about 9:1. These dramatic shifts in adolescence toward 
the associations of self-reported self-efficacy (masculinity) and relationality female psychopathology suggest that there may be something profoundly 
(femininity) in fourth graders and high school students, and analyzes the unhealthy about the passage to adulthood for young women. 
relation of these variables to self-esteem in each age group. Based on evi- As it exists today, adolescence is a culturally created developmental 
dence showing self-esteem to be a fairly good predictor of mental health period in which young people take on intensified sex-typed roles and ac-
in adults (Coopersmith 1967) and adolescents (Rosenberg 1985), self-es- commodate to the pervasive deference accorded to males in adult 
teem is used as an indicator of psychological well-being. heterosexual relations. Inherent in these roles is a greater valuation, author-
ity, and prestige accorded to masculine qualities and activities, and a 
devaluation of feminine ones (Stockard & Johnson, 1980; Johnson, 1988). 
RELATED LITERATURE The transition from childhood to adulthood appears to involve a transition 
from the relatively egalitarian, although highly segregated, world of child-
Although theory abounds concerning the importance of sex-stereo- hood to a highly unequal adult world where male dominance appears not 
typic characteristics to mental health during adolescence (Erikson, 1950; only in the economic and political world, but also in day-to-day interactions. 
Kohlberg, 1966), little evidence supports this claim. Some researchers have The extent to which these role changes might affect the association of sex-
found measures of both self-efficacy and relationality positively associated typed attributes with psychological well being is not known. We can posit 
with self-esteem in both female and male high school students (Speuce & no specific hypotheses regarding these associations in children except to 
Heimrich, 1978), while others have found self-efficacy to be the more im- suggest that they will be different from the associations in adolescents and 
portant predictor (Lamke, 1982; Lerner, Sorell, & Brackney, 1981; Ziegler, adults. 
Dusek, & Carter, 1984). Wells (1980) reported that for adolescent males 
neither self-efficacy nor relationaiity was significantly related to self-esteem, 
although self-efficacy was predictive for females, and Massad (1981) con- METHODOLOGY 
cluded that the traditional view advocating sex-stereotypic identification 
seems applicable to males whereas an androgyny model seems best suited Participants 
for females. Still others who endorse the desirability of androgyny in adult-
hood propose that healthy adjustment in adolescence is characterized by Two separate studies were conducted for the present investigation. 
adoption of a stereotypic masculine or feminine sex role identification ( e.g., Participants in both were predominantly white, public school students in 
Pleck, 1975). western Oregon. The first was a study of adolescents in grades nine through 
Thus, while the literature on adolescents is mixed and sparse, some- twelve in a largely middle-class community. Cross-sectional data are avail-
what more results seem to support the masculinity model than other able for 799 students (412 girls, 387 boys). The second was a study of fourth 
_________L_ _ -
132 Allgood-Merten and Stockard Sex Role Identity and Self-Esteem 133 
graders in a predominantly working-class community. Cross-sectional data data from the panel study. The items used to measure relationality or 
are available in this second sample for 607 students (300 girls, 307 boys), stereotypical femininity included the following items: (1) "I am a gentle 
with longitudinal data from a subset of 52 students (32 girls, 20 boys) who person," (2) "My art work and my ideas are creative and original," (3) "I 
were first tested in the fourth grade and later tested in the twelfth grade am a very considerate person (thoughtful person)," (4) "I do not help other 
using the same instruments each time. This subset is hereinafter referred people very much" (reversed), (5) "I am kind to other people almost all 
to as the panel study. Students in the panel study have significantly higher of the time," and (6) "I try to do everything I can for the people I care 
fourth-grade achievement test scores (t = 3.33, df = 57, p = .002) and about." Coefficient alpha for this scale was .66 for the total fourth-grade 
mothers with significantly more education ((t = 3.42, df = 68, p = .001) sample, .67 for fourth graders in the panel study, and .77 for the twelfth-
than those in the total fourth-grade group. grade data from the panel study. 
Self-esteem was measured for both groups with the Rosenberg Self-
Esteem Inventory, a widely used self-report measure in which respondents 
Procedures rate their agreement with the statements on a 4-point scale from strongly 
disagreet o strongly agree (Rosenberg, 1965). Responses on the 10-item scale 
After obtaining parental permission, questionnaires were adminis- were summed and averaged, with a higher score indicating a more positive 
tered to the students in their classrooms by trained research assistants. self-concept. 
Measures Analysis 
The Personal Attributes Questionnaire Short Version (PAQ; Spence 
Heimrich, & Stapp, 1974) was used to assess the degree to which the ado- Descriptive statistics (means, standard deviations, and effect sizes) 
lescent respondents identified with feminine or masculine stereotypes. The and t tests were used to examine the scores of the boys and girls on each 
PAQ Short Version is a widely used instrument composed of 24 bipolar of the variables described above and changes over time for the panel group. 
items describing personal characteristics on which respondents rate them- Pearson product moment correlations were then used to examine the as-
selves on a 5-point Likert scale. The questionnaire is divided into three sociation between the students' perceptions of their sex-typed personality 
eight-item scales labeled Masculinity (M), Femininity (F), and Masculin- traits and self-esteem. Changes in correlations over time for the panel 
ity-Femininity (M-F). Masculine attributes are those typically associated group were assessed with the procedures recommended by Steiger (1980). 
with self-sufficiency and self-efficacy, whereas feminine attributes are those 
typically associated with communality and relationality. For purpose of this 
analysis, only the "masculinity" and "femininity" scales were used. RESULTS 
The fourth graders' sex-typed personality characteristics were meas-
ured by a modified version of the Children's PAQ, which has been found Table I presents the mean scores for females and males on each of 
to be highly correlated with scores on the adult version (Hall & Halber- the variables measured. In all of the groups the males report higher average 
stadt, 1980). All questions required the children to note on a 4-point scale scores than the females on the self-esteem and self-efficacy scales, while 
how true a statement was of them. Scores on each scale were averaged so the females have higher average scores on the relationality scale. These 
that they would be comparable to each other. The items used to measure sex differences are statistically significant for both the adolescent sample 
self-efficacy or masculinity included (1) "I give up easily" (reversed), (2) and the total group of fourth graders while in the panel study only the 
"It is easy for people to make me change my mind" (reversed), (3) "I do masculinity measure yields statistically significant sex differences. However, 
not do well in sports" ( reversed), ( 4) "In most ways, I am better than most the effect sizes for the self-esteem and self-efficacy measures are compa-
of the other kids my age," (5) "I am more active than most kids my age," rable throughout all the groups, suggesting that the lack of significance in 
(6) "I am often the leader among my friends," (7) "I almost always stand differences on these measures in the panel study reflects its smaller sample 
up for what I believe in," and (8) "When things get tough, I almost always size. 
keep trying." Coefficient alpha for this scale was .62 for the total fourth- In all cases females perceive themselves as more relational than self-
grade sample, .60 for the panel fourth graders, and .78 for the twelfth-grade efficacious. (For the total group of fourth-grade girls, the average 
134 Allgood-Merten and Stockard Sex Role Identity and Self-Esteem 135 
Table I. Mean Scores and Standard Deviations, all Variables, Females and Males, Fourth It is interesting to note that the small subset of fourth graders who 
Graders and Adolescents were followed through the twelfth grade appears less sex typed than the 
Females Males overall group of fourth graders, and this seems particularly true for the 
Mean (SD) Mean (SD) df p d" boys. Compared to the total fourth-grade sample, boys in the subset report 
Self-esteem slightly more relationality (t = 1.28, df = 25, p = .21) , girls report slightly 
Fourth graders 2.94 (0.54) 3.07 (0.50) -3.12 602 <.002 -.25 less relationality (t = -.49, df = 40, p = .63), and both report greater 
(Total group) 
Fourth graders self-efficacy (t = 1.194, df = 42, p = .06 for girls; t = 2.04, df = 25, p = 3.11 (0.47) 3.32 (0.43) -1.60 50 .12 -.47 
(Panel study) .05 for boys) and higher self-esteem (t = 2.11, df = 41, p = .04 for girls, 
Twelfth graders 3.12 (0.56) 3.26 (0.55) -.88 50 .38 -.25 t = 3.16, df = 26, p = .004 for boys). 
(Panel study) Both boys and girls in the panel study have slightly higher scores on 
Adolescents 2.83 (0.53) 3.05 (0.48) -6.13 797 <.001 -.43 
Self-efficacy self-efficacy and relationality in the fourth grade than in the twelfth grade. 
Fourth graders 2.76 (0.45) 3.05 (0.48) -7.73 598 <.001 -.62 (for girls the average difference = .22, t = 2.24, p = .03 for relationality, 
(Total group) average difference = .08 t = .88, p = .38 for self-efficacy; for boys average 
Fourth graders 2.88 (0.38) 3.19 (0.44) -2.65 50 .01 -.76 
(Panel study) difference = .18, t = 1.14, p = .27 for relationality, average difference = 
Twelfth graders 2.96 (0.50) 3.23 (0.50) -1.90 50 .06 -.54 .04 t = .35, p = .73 for self-efficacy.) The girls in the panel study, but not 
(Panel study) the boys, are more strongly sex typed in twelfth grade than in fourth grade. 
Adolescents 3.46 (0.55) 3.75 (0.49) -7.90 797 <.001 -.56 
Relationality Table II presents the correlations between the measures of sex-typed 
Fourth graders 3.13 (0.54) 2.94 (0.52) 4.20 attributes and self-esteem. Self-efficacy, our measure of masculine attrib-598 <.00] .36 
(Total group) utes, is strongly associated with self-esteem for both males and females in 
Fourth graders 3.08 (0.52) 3.07 (0.51) .11 50 <.91 .02 all samples. Moreover, the correlation between self-efficacy and self-esteem 
(Panel study) 
Twelfth graders 3.31 (0.42) increases significantly from fourth to twelfth grade for the girls in the panel 3.25 (0.66) .38 50 .70 .11 
(Panel study) study (z = 2.01, p < .05). 
Adolescents 3.99 (0.47) 3.53 (0.66) 11.17 797 <.00] .80 However, relationality, our measure of feminine attributes, is strongly 
"Cohen's d equals the difference between the two means divided by the common standard associated with self-esteem only for the fourth graders (both boys and girls) 
deviation. Cohen (1977) suggests that effect sizes of .2 could be considered small, those of and the small subset of twelfth-grade boys in the panel study. The corre-
.5 medium in size, and those of .8 or greater to be large. 
Table II. Pearson Product Moment Corrrelations of Self-Esteem 
difference between relationality and self-efficacy scores is .36, t = 10.68, with Self-Efficacy and Relationality, Females and Males, Fourth Graders and Adolescents 
df = 299, p < .001; for fourth-grade panel girls the mean difference is .20, 
= = = Fourth graders Sell'-cfficacv Relationality t 1.80, df 31, p .08; for twelfth-grade panel girls the mean difference 
is .35, t = Females 3.11, df = 31, p = .004; and for the adolescent girls the mean Total fourth-grade sample _44c _45c 
difference is .53, t = 15.14, df = 299, p < .001). In general, males see Panel fourth-grade sample .47b .44b 
themselves as more efficacious than relational, although the differences are Panel twelfth-grade sample .7 lc .01 
Adolescents .49" .07 
smaller than with the females. The major exception are the twelfth-grade 
boys in the panel study who describe themselves as almost equally relational Males Total fourth-grade sample .sic _34c 
and efficacious. (For the total group of fourth-grade boys the average dif- Panel fourth-grade sample .43" .48" 
ference is .11, t = 3.01, df = 304, p = .003; for fourth-grade panel boys Panel twelfth-grade sample _54b .42" 
the mean difference is .12, t = .88, df = 19, p = .39; for twelfth-grade Adolescents .42c .15b 
panel boys the mean difference is .02, t = .18, df = 19, p = .39; and for "p < .05. 
the adolescent boys the mean difference is .22, t = 5.5, df = 306, and p bp < .01. 
cp < .001. 
< .001 ). In all groups the girls are more sex typed than the boys. 
L -
136 Allgood-Merten and Stockard Sex Role Identity and Self-Esteem 137 
lation is statistically significant for the group of adolescent boys, but is sub- girls, while not rejecting relationality, would develop less positive self-con-
stantially smaller than the correlations found with the other groups of boys cepts. 
and might be dismissed as substantively insignificant. The decline in the It is noteworthy that the one group that exhibited a significant cor-
correlation between relationality and self-esteem from fourth to twelfth relation between relationality and self-esteem in adolescence is the group 
grade for girls in the panel study is statistically significant (z = 2.01, p < that would be seen as most "androgynous." In other words, only those ado-
.05). For females in the large adolescent sample and as twelfth graders in lescents who perceived themselves as equally relational and efficacious 
the panel study there is no association between feminine attributes and associated relationality with self-esteem. Thus the androgyny model of psy-
self-esteem, and the association for males in the large adolescent sample chological well-being was supported only among a small group of higher 
is also relatively small. SES and high-achieving adolescent males reporting androgynous personal-
Results not reported here indicate that this pattern of associations ity characteristics. 
occurs when fourth-grade achievement and mother's educational level are Perhaps most interesting is the considerably greater devaluation of 
introduced as control variables for the students in the fourth-grade and feminine attributes by the adolescent girls than by the adolescent boys. 
panel group, and when results are examined separately within each grade Even among the exceptional twelfth grade girls in the panel study, there 
level for those in the adolescent sample. was no association of feminine attributes with self-esteem. In addition, girls, 
but not boys, in the panel group showed a significantly higher correlation 
between self-efficacy and self-esteem in twelfth grade than in fourth grade. 
SUMMARY AND DISCUSSION This finding is puzzling given the fairly strong positive associations between 
relationality and self-esteem reported by their male peers (the high-achiev-
As expected, the associations of sex-typed attributes with self-esteem ing boys) in both years, and little change in the association between 
for adolescents are consistent with those found in college samples and self-efficacy and self-esteem over the time period. This suggests that young, 
adults, supporting the masculinity model. Self-efficacy, but not relationality, bright, high-achieving women may internalize the cultural devaluation of 
is strongly associated with self-esteem, especially for females. In contrast, the feminine to an even greater extent than males with similar charac-
both self-efficacy and relationality are associated with self-esteem for both teristics, and for them, unlike their male peers, androgyny does not seem 
sex groups in the fourth graders, providing support for the androgyny an option. What this means remains open to speculation. At the minimum 
model. Our results suggest that as children move into adolescence where it suggests that coming of age for young women in postindustrial society 
adult sex-typed roles and attitudes become more salient, relationality is no brings with it a double bind-a marked devaluation of that which is female, 
longer associated with self-esteem, especially for females. The association yet being female-which makes it very difficult to feel good about oneself. 
between relationality and self-esteem remained strong only for the high- It is not far-fetched to wonder whether this core conflict may contribute 
SES, high-achieving twelfth-grade boys in the panel study, a group unique to women's greater vulnerability to the internalizing disorders that emerge 
in its equally high self-ratings on both relationality and self-efficacy. (e.g., anorexia, bulimia nervosa) or increase (e.g., major depression) during 
The association between masculine-typed traits and self-esteem is not this developmental period. Recent evidence suggests that poor body im-
surprising. As Richmond (1984, p. 1031) notes, "the concepts of masculinity age/self-esteem may be a vulnerability factor that sets adolescent girls at 
and self-esteem are both defined by the idea of standing up for oneself, greater risk for depression (Allgood-Merten, Lewinsohn, & Hops, 1990). 
having a distinctive point of view, and being independent and self-sufficient. Further work in this area seems essential if we are to understand 
... [T]he relationship is logical rather than empirical." Thus, a strong as- more about the association of sex-typed personality attributes and mental 
sociation could, by definition, be expected to be both positive and health. It seems especially important to use longitudinal studies, where 
significant, whatever the life stage. Of more interest is the finding that the changes in self-esteem and personality attributes can be observed over long 
ideal stereotypic feminine traits of expressiveness or relationality are asso- periods of time. Additional measures, both of sex-typed traits (see Gill et 
ciated with self-esteem among children but not generally among al., 1987) and of mental health related characteristics are needed. Particu-
adolescents. This suggests that while feminine traits have prestige and im- Llar ly needed are scales that more reliably tap sex-typed traits of children portance in the world of children they are devalued in the world of and that can be used with both children and adolescents. It is possible that adolescence. Thus it is not surprising that adolescent boys would diminish some of the results reported here reflect the relatively low reliability of the the importance of relationality in their self-concepts and that adolescent children's PAQ with the fomth graders and unknown diffecences between 
138 Allgood-Merten and Stockard Sex Role Identity and Self-Esteem 139 
the PAQ and CPAQ. Further work should examine this issue. Finally, given Rutter, M. (1986). The developmental psychology of depression: Issues and perspectives. In 
that most studies of the relationship between sex-typed attributes and men- M. Rutter, C. Izard, & P. Reed (Eds.), Depression in young people: Developmental and 
Clinical Perspectives. New York: Guilford. 
tal health have focused on college-aged people or adults, research that Schoenbach, V., Garrison, C., & Kaplan, B. (1984). Epidemiology of adolescent depression. 
examines other life stages is needed. Studies investigating these associations Public Health Review. 12, 159-189. 
in old age, a life period in which sex role differentiation diminishes Spence, J., Helmreich, R., & Stapp, J. (1974). The Personal Attributes Questionnaire: A 
measure of sex-role stereotypes and masculinity-femininity. ]SAS Catalog of Selected 
(Stockard & Johnson, 1980), could be informative. Documents in Psychology, 4, 127. 
Spence, J. T., & Helmreich, R. L. (1978). Masculinity and femininity: Their psychological 
dimensions, correlates, and antecedents. Austin: University of Texas Press. 
Steiger. J. H. (1980). Tests for comparing elements of a correlation matrix. Psychological 
REFERENCES Bulletin. 87, 245-251. 
Stockard, J., & Johnson, M. M. (1980). Sex roles: Sex inequality and sex role development. 
Abraham, K. (1949). Notes on the psycho-analytic investigation and treatment of Englewood Cliffs, NJ: Prentice-Hall. 
manic-depressive insanity and allied conditions. In E. Jones (Ed.), Selected papers of Karl Taylor, M. C., & Hall, J. A. (1982). Psychological androgyny: Theories methods, and 
Abraham (D. Bryan & A. Strachy, tr.). London: Hogarth. (Original publication 1911.) conclusions. Psychological Bulletin, 92, 347-366. 
Allgood-Merten, B., Lewinsohn, P., & Hops, H. (1990). Sex differences and adolescent Wells, K. (1980). Gender-role identity and psychological adjustment in adolescence. Journal 
depression. Journal of Abnormal Psychology, 1, 55-63. of Youth and Adolescence, 9, 59-73. 
Bassoff, E. S., & Glass, G. V. (1982). The relationship between sex roles and mental health: Whitley, B. E., Jr. (1983). Sex role orientation and self-esteem: A critical meta-analytic review. 
A meta-analysis of twenty-six studies. The Counseling Psychologist, JO, 105-112. Journal of Personality and Social Psychology, 44, 765-778. 
Bern, S. L. (1974). The measurement of psychological androgyny. Journal of Consulting and Whitley, B. E., Jr. (1984). Sex-role orientation and psychological well-being: Two 
Clinical Psychology, 42, 155-162. meta-analyses. Sex Roles, 12, 207-225. 
Bern, S. L. (1979). Theory and measurement of androgyny: A reply to the Pedhazur-Tetenbaum Zeigler, C. B., Dusek, J. B., & Carter, D. B. (1984). Self-concept and sex-role orientation: 
and Locksley-Colten critiques. Journal of Personality and Social Psychology, 37, 1047-1054. An investigation of multidimensional aspects of personality development in adolescence. 
Cohen, J. (1977). Statistical power analysis for the behavioral sciences. New York: Academic Journal of Early Adolescence, 4, 25-39. 
Press. 
Coopersmith, S. (1967). The antecedents of self-esteem. San Francisco: W. H. Freeman and 
Co. 
Erikson, E. (1950). Childhood and society. New York: Norton. 
Gill, S., Stockard, J., Johnson, M., & Williams, S. (1987). Measuring gender differences: The 
expressive dimension and critique of androgyny scales. Sex Roles, 17, 375-400. 
Hall, J. A., & _Halberstadt, A. G. (1980) Masculinity and femininity in children: Development 
of the children's personal attributes questionnaire. Child Development, 16, 270-280. 
Johnson, M. M. (1988). Strong mothers, weak wives. Berkeley: University of California Press. 
Kagan, J. (1964). Acquisition and significance of sex-typing and sex-role identity. In M. L. 
Hoffman & L. W. Hoffman (Eds.), Review of child development research, Vol. 1, New 
York: Russell Sage Foundation. 
Kohlberg, L. (1966). A cognitive-developmental analysis of children's sex role concepts and 
attitudes. In E. Maccoby (Ed.), The development of sex differences, Stanford, CA: Stanford 
University Press. 
Lamke, L, K. (1982). The impact of sex-role orientation on self-esteem in early adolescence. 
Child Development, 53, 1530-1535. 
Lerner, R., Sorell, G., & Brackney, B. (1981). Sex differences in self-concept and self-esteem 
of late adolescents: a time-lag analysis. Sex Roles, 7, 709-722. 
Lewinsohn, P., Hops, H., Roberts, R., & Seeley, J. (1988). Adolescent depression: Prevalence 
and psychosocial aspects. Paper presented at the American Public Health Association 
annual meeting, Boston, MA, November 14. 
Massad, C. M. (1981). Sex role identity and adjustment during adolescence. Child 
Development, 52, 1290-1298. 
Mussen, P. H. (1969). Early sex-role development. In D. A. Goslin (Ed.), Handbook of 
socialization theory and research. Chicago: Rand-McNally. 
Pleck, J. H. (1975). Masculinity-femininity: current ~nd alternative paradigms. Sex Roles, 1, 
161-177. 
Richmond, P. G. (1984). An aspect of sex-role identification with a sample of twelve year 
olds and sixteen year olds. Sex Roles, 11, 1021-1032. 
Rosenberg, M. (1965) Society and the adolescent self-image. Princeton, NJ: Princeton University 
Press. 
~--·----------111111111111111• -