TOC Electronic Journal: To print this article select pages 67-70.

Gender Roles and Self-Esteem: A Consideration of

Environmental Factors
Jeffery W. Burnett, Wayne P. Anderson, and P. Paul Heppner

Much empirical research has examined the psychological well-being of individuals with varying
gender role orientations. This research has typically shown masculinity to be a strong correlate
of self-esteem and femininity to be relatively unrelated to self-esteem. This research has often
failed to consider the relative environmental influences impacting this process. This study ex-
amines the relationships of masculinity and femininity to self-esteem and environmental presses
for each of these sets of behavioral characteristics. Results indicate a stronger press for mas-
culine characteristics than feminine characteristics. The person-environment interaction suggests
that a masculine environment may place women who are low in masculinity at particular risk
for low self-esteem. Implications for both research and practice are discussed.

hirty years ago, gender roles was a topic barely worthy of gender role measures has been called into question by Pedhazur and

T study. It was simply assumed that men followed one role and
women another. The women’s movement and later the men’s
movement partially resulted in redefining the concepts of gender
Tetenbaum (1979). They argued that the item content of the mascu-
linity scales was more socially desirable than that of the femininity
scales. Although this is still a controversial topic, Taylor and Hall
identity and gender roles. Interest in studying gender roles has to (1982) concluded that there is no systematic difference in the social
some extent followed a growing national trend towards questioning desirability of these two constructs as measured by the two most
assumptions made about appropriate attitudes, behaviors, and person- widely used instruments, the Personal Attributes Questionnaire (PAQ;
ality characteristics for men and women (Mintz & O’Neil, 1990). Spence, Helmreich, & Stapp, 1974) and the Bem Sex Role Inventory
Empirical research on gender roles has often attempted to understand (BSRI; Bem, 1974). Furthermore, Spence and Helmreich (1978)
the consequences of such a redefinition through examining the psy- maintained that the PAQ assesses instrumentality and expressiveness
chological well-being of individuals with varying gender role orien- rather than masculinity and femininity per se. Thus, social definitions
tations (see Cook, 1987; Whitley, 1983). of these constructs may differ.
In the 1970s researchers thought that a greater degree of psycho- With respect to choices of well-being measures used in these stud-
logical well-being would be associated with an androgynous orien- ies, some have argued that the type of well-being assessed in some
tation. Indeed, Bem (1974, p. 162) suggested that androgyny may measures may be biased towards masculinity (see Whitley, 1983).
define a “more human standard of mental health.” Researchers during One recent study (Sharpe & Heppner, 1991) found evidence for two
this time attempted to demonstrate that androgynous individuals independent areas of well-being. The authors labeled these “tradi-
would have greater self-esteem and overall better psychological ad- tional well-being” and “affiliative well-being.” Traditional well-being
justment than would individuals who adhered to more traditional gen- included constructs such as self-esteem, anxiety, and depression and
der roles (see Cook, 1985; Whitley, 1983). Empirical studies, how- was most strongly associated with masculinity. Affliative well-being
ever, did not support such predictions (see Whitley, 1983); instead, a included measures of intimacy and a lack of conflict due to compe-
“masculinity model” was developed to account for the better adjust- tition and emotional restriction and was most strongly associated with
ment associated with masculinity (Adams & Sherer, 1985). In most femininity. Thus, this research has shown preliminary support for the
of these studies, when the gender-by-gender role interaction was con- notion that there may be more than one dimension of psychological
sidered, results usually showed androgynous women as more well- well-being and that these may be differentially associated with mas-
adjusted than feminine women, but androgynous men’s adjustment culinity and femininity. Self-esteem may be more directly associated
as equal to that of masculine men (Silvern & Ryan, 1979). Two meta- with masculinity, whereas femininity is associated with perhaps less
analyses found this pattern to be particularly true for measures of traditional measures of well-being, such as intimacy and relationship
self-esteem (Bassoff & Glass, 1982; Whitley, 1983). This in turn led satisfaction.
researchers to conclude that masculinity was the active factor for both The last explanation, that of a cultural bias toward masculinity,
men and women. Cook later suggested that research showing mas- has been termed the masculine supremacy effect (Cook, 1985; Yager
culinity as more strongly related to self-esteem than femininity was & Baker, 1979). This argument posits that masculinity simply has
“one of the most stable findings to emerge from androgyny research” greater social utility in American culture (Yager & Baker, 1979).
(1987, p. 476, 477). Thus, research has run counter to the initial More specifically, this position suggests a cultural bias toward mas-
hypotheses of gender-role researchers. culinity such that individuals who are masculine receive more positive
The three major explanations offered for masculinity’s greater im- social reinforcement and hence develop higher self-esteem. This last
pact on self-esteem have been (a) biased or inadequate instruments explanation has not been tested, and it is the subject of the current
of operational definitions used to assess gender roles, (b) a possible study.
masculine bias in the choice of global well-being measures, and (c) Until now, the focus of gender role research has been on mascu-
the notion of a large-scale cultural bias towards masculinity. Regard- linity and femininity as personality or trait variables almost to the
ing the first explanation, the social desirability of the subscales of exclusion of environmental or situational influences. In other areas

Journal of Counseling & Development l January/February 1995 l Volume 73 323

Burnett, Anderson, and Heppner

of psychological research, it is readily assumed that environmental scales were rated as ideal for both men and women, but significantly
influences play a large part in determining behavior. With respect to more typical of one sex than the other (Spence & Helmreich, 1978).
gender role research, Cook (1985) suggested that an analysis of per- Thus, the IM scale assesses stereotypical masculine qualities such as
son-by-environment interactions should replace a focus on personal independence and competitiveness, and the IF scale assesses stereo-
variables alone. More specifically, this would suggest that individual typical feminine qualities such as warmth and devotion to others
masculinity and femininity be studied within their social context. (Spence & Helmreich, 1978). The M-F scale has not been widely
Given that the environment provides powerful cues about how an used in research and was not scored in this study.
individual ought to behave, as well as the fact that masculinity and The PAQ has adequate internal consistency, with alpha coefficients
femininity are largely social constructions (Scher & Good, 1990), an of .85, .82, and .78 for the M, F, and M-F scales (Wilson & Cook,
examination of environmental presses for masculinity and femininity 1984). In this study, alpha coefficients for the M and F scales were
could provide additional information about the relationship between .74 and .75. Test-retest reliability is approximately .60 over a 2.5-
gender roles and psychological well-being. The term environmental month period (Yoder, Rice, Adams, Priest, & Prince, 1982). Construct
press is defined here as the implicit and explicit cues an individual validation of the PAQ has usually consisted of correlating scores on
receives from others about how he or she should think, feel, and the IM scale to typical masculine traits such as competitiveness
behave. In the context of this investigation, those cues will be directly (Spence & Helmreich, 1978) and the type-A behavior pattern (Ste-
related to the constructs of masculinity and femininity. vens, Pfost, & Ackerman, 1984); scores on the IF scale have been
The purpose of this study was to examine the interaction between related to typical feminine traits such as nurturance (Bem, Martyna,
personal masculinity and femininity and social (or environmental) & Watson, 1976) and placing value on the socioemotional aspects of
presses for masculinity and femininity. Specifically, the study ex- life (Shichman & Cooper, 1984). The PAQ also shows moderate cor-
amined whether participants would report feeling greater social pres- relations with other measures of masculinity and femininity (see
sure to demonstrate masculine characteristics than feminine charac- Cook, 1985).
teristics. In addition, because the study examined both personal and The Personal Attributes Questionnaire-Environmental form (PAQ-
environmental variables, the mutual influence of these two variables env) was employed to assess environmental presses for masculinity
was studied, particularly in line with Cook’s suggestion of the person- and femininity. Instructions on the PAQ-env stated, “The following
by-environment interaction. In accordance with previous research, the is a list of the same characteristics you just completed about yourself
first hypothesis predicted that individual masculinity would be more . . . now you are to rate those same characteristics according to how
strongly related to self-esteem than individual femininity. Further- you feel other students encourage or reward you to exhibit each of
more, as theorized by the masculine supremacy effect (Cook, 1985), these characteristics.” Internal consistency for the modified PAQ was
the second hypothesis predicted that participants would rate the en- assessed in this study. Alpha coefficients on the PAQ-env M and F
vironmental press for masculinity as stronger than the press for fem- scales were .78 and .84.
ininity. Finally, the third hypothesis predicted that there would be a The Coopersmith Self-Esteem Inventory (SEI; Coopersmith,
person-by-environment interaction such that participants low in per- 1967) is a 25-item, self-report measure of self-esteem. Self-esteem is
sonal masculinity living in an environment high in demand for mas- defined here as an individual’s evaluation of him- or herself as a
culinity would have the lowest self-esteem. Given that the relation- person, and is an expression of general approval or disapproval of
ship between individual femininity and self-esteem is usually not the self (Coopersmith, 1967). Participants respond to each item on a
statistically significant, no predictions were made about a person-by- 2-point scale as like me or unlike me. Scores on the SEI range from
environment interaction with respect to femininity. Because the re- 0 to 100, with higher scores indicating greater self-esteem. The adult
sults for the three hypotheses might differ for men and women, anal- form, which was used in this study, is a modified version of the more
yses were conducted on the total sample as well as separately for broadly used school form, from which most of the psychometric es-
each sex. timates have been developed. The correlation between these two
forms is approximately .80 (N = 647; Coopersmith, 1967). The alpha
coefficient for the adult form is .75 (Ahmed, Vallient, & Swindle,
METHOD
1985). The alpha coefficient obtained in this study was .81. Test-
Participants retest reliability is approximately .88 over a 5-week period (Coop-
ersmith, 1967). Convergent validity for the SEI was demonstrated by
Participants for the study consisted of 236 undergraduate students (90
showing a positive correlation with other self-report measures of self-
male and 146 female) at a large midwestem university. Participants
esteem and with peer ratings (Demo, 1985). Additionally, scores on
were mostly freshmen and sophomores, and 90% were between the
the SEI were found to correlate positively with ego strength and neg-
ages of 17 and 22 years. Students took part in the study either for
atively with anxiety, providing support for both convergent and dis-
extra credit or in partial fulfillment of a course requirement.
criminant validity (Kawash, 1982).
Instrument Procedure
The Personal Attributes Questionnaire (PAQ; Spence, Helmreich, & Participants completed the PAQ, the PAQ-env, and the SEI in groups
Stapp, 1974) was employed to assess participants’ individual sex-role of 25 to 50. Each participant received a packet containing the instru-
orientations. Instructions on this form state, “The items below inquire ments and a consent form. Administration time was approximately
about what kind of person you think you are . . . . You are to choose 30 minutes. After participants had completed the instruments, they
the number which you think best describes where you fall on the received a page of debriefing information.
scale.” The PAQ yields three scores based on eight items each: an
individual masculinity (IM) score, an individual femininity (IF) score,
RESULTS
and a masculinity-femininity (M-F) score. Items on each scale are
scored on a 5-point Likert-type scale, from 0 to 4, such that scores Descriptive statistics on all variables are presented in Table 1. The
on each scale may range from 0 to 32. All items for the IM and IF means suggest that overall, participants’ appraisal of their individual

324 Journal of Counseling & Development l January/February 1995 l Volume 73

 Gender Roles and Self-Esteem: A Consideration of Environmental Factors

TABLE 1 (high versus low press for masculinity) analysis of variance (ANOVA)
was conducted. A significant main effect was obtained for individual
Means and Standard Deviations for All Variables masculinity, F (1,231) = 33, p < .001, but not for the environmental
press for masculinity, F (1,231) = 2.57, p = .11. Moreover, the in-
Men Women
teraction was also significant, F (1,231) = 4.10, p < .05; hence main
Variable n M SD n M SD effects will not be further interpreted. Planned contrast statements on
Individual the total sample revealed that those low in individual masculinity who
masculinity reported a strong environmental press for masculinity had the lowest
Individual femininity self-esteem, t (228) = 2.56, p < .01. Follow-up analyses with separate
ANOVAs conducted for each sex revealed that the individual by en-
Press for masculinity
Press for femininity vironment interaction was significant for women, F (1,143) = 5.27,
Self-esteem p < .05, but not for men, F (1,87) = .27, p < .05. Again the same
pattern was observed; women low in masculinity reporting an envi-
Note. Individual masculinity and femininity = Personal Attributes Ques- ronment that encouraged high amounts of masculinity, had the lowest
tionnaire (PAQ, Spence et al., 1974); press for masculinity and feminin-
ity = Personal Attributes Questionnaire-Environment (PAQ-env); self- self-esteem. Thus, Hypothesis 3 was partially confirmed. The 2 (high-
esteem = Coopersmith Self-Esteem Inventory (SEI, Coopersmith, versus-low individual femininity) × 2 (high-versus-low press for fem-
1967). Higher scores indicate greater amounts of masculinity, feminin- ininity) ANOVA revealed no significant main effects or interactions,
ity, environmental presses for masculinity and femininity, and self- p < .05.
esteem.

DISCUSSION

masculinity (M = 20.93), individual femininity (M = 24.24). and This study found evidence in support of a masculine bias in American
self-esteem (M = 43.33) were typical of other college student samples society, such that individuals who possess a larger amount of mas-
(see Sharpe & Heppner, 1991; Zeldow, Clark, & Dougherty, 1985). culine characteristics such as decisiveness, independence, and com-
Examination of the correlation matrix (represented in Table 2) petitiveness report greater self-esteem than do those with less of those
revealed that individual masculinity was significantly correlated with traits. Consistent with past research (e.g., Bassoff & Glass, 1982;
self-esteem for the total sample (r = .55, p < .001) as well as within Whitley, 1983), individual masculinity was significantly correlated
each sex (men: r = .58, p < .001; women: r = .52, p < .001). with self-esteem for both men and women, but individual femininity
Individual femininity scores, however, were not significantly corre- was not significantly related to self-esteem in either sex. This pattern
lated with self-esteem across the total sample or within either sex (all of results support what has been termed the masculinity model (Cook,
p < .05). Thus, as predicted, the relationship of masculinity to self- 1987; Whitley, 1983) as an explanation for the relation of gender
esteem was significantly higher than that of femininity to self-esteem, roles to psychological well-being.
t(232) = 5.81, p < .01. Thus, the first hypothesis was confirmed. The correlates of masculinity—goal directedness, high achieve-
Examination of the means for environmental presses for mascu- ment motivation, competitiveness, and assertiveness—are traits
linity and femininity (see Table 1) revealed that participants reported highly valued in this culture (Kenworthy, 1979; Locksley & Colten,
feeling a greater press for masculine characteristics (M = 22.51) than 1979). The results of this study confirm those findings on a more
for feminine characteristics (M = 20.83). T tests conducted on the personal level. Specifically, the results suggest that for both men and
means for the environmental press variables have revealed that these women, environmental presses to meet or live up to this definition
differences were statistically significant across the total sample, of masculinity were significantly greater than for femininity. Even
t(230) = 4.13, p <.001, as well as for men, t(87) = 3.61, p < .001, though participants in another study rated the ideal person as more
and women, t(142) = 2.28, p < .05, separately. Thus, the second feminine than masculine (Silvern & Ryan, 1979), participants in this
hypothesis was accepted. study felt that their peers encouraged achievement, competition, and
Finally, to test the combined influence of environmental factors independence more so than sensitivity, emotional expressiveness, and
with trait variables, a 2 (high versus low individual masculinity) × 2 satisfaction in relationships.
Moreover, within this high press for masculinity, the person-by-
environment interaction suggests that, especially for women, those
low in individual masculinity were at a particular risk for decreased
TABLE 2
self-esteem. When in an environment considered high in masculinity,
Correlation Matrix of Self-Esteem Inventory With these women reported the lowest self-esteem scores of any group.
PAQ and PAQ-env Because the culture places a high value on masculine or instrumental
characteristics, those individuals who show these traits perhaps re-
Self-Esteem Inventory ceive more respect from others and hence enjoy greater self-esteem
than individuals who display fewer of these masculine traits.
Total Men Women
PAQ (N = 236) (n = 90) (n = 90) In short, this study suggests that gender roles, which were previ-
ously studied in isolation from environmental influences, are embed-
Individual masculinity ded within a cultural context that can interact with one’s personal
Individual femininity traits. Counselors involved in exploring gender issues with their cli-
Press for masculinity ents may also want to attend to the characteristics of the environment
Press for femininity within which their clients function. Our results suggest that it is per-
Note. See Table 1 note. haps most important to do so with women. Counseling orientations
*p < .05. **p < .01. such as systems theory and feminist theory place a strong emphasis

Journal of Counseling & Development l January/February 1995 l Volume 73 325

Burnett, Anderson, and Heppner

on situational and cultural factors, and hence may provide useful Cook, E. P. (1985). Psychological androgyny. Elmsford, NY: Pergamon.
interventions for dealing with clients who experience distress due to Cook, E. P. (1987). Psychological androgyny: A review of the research. The
a lack of congruence with their environment. The practice of a new Counseling Psychologist, 15, 471-513.
form of counseling, called gender aware therapy (see Good, Gilbert, Coopersmith, S. (1967). The antecedents of self-esteem. San Francisco:
& Scher, 1990) may be particularly useful for this type of work. Freeman.
Demo, D. (1985). The measurement of self-esteem: Refining our methods.
Several limitations should be considered in generalizing the results
Journal of Personality and Social Psychology, 48, 1490-1502.
of this study. First, the methodology for assessing environmental
Good, G. E., Gilbert, L. A., & Scher, M. (1990). Gender aware therapy: A
presses was new and merits additional examination and replication.
synthesis of feminist therapy and knowledge about gender. Journal of
The modification of the PAQ to assess participants’ perceptions of Counseling and Development, 68, 376-380.
what other students encouraged or rewarded in them may have only Kawash, G. (1982). A structural analysis of self-esteem from pre-adolescence
narrowly tapped a broader, more general environmental press (a col- through young adulthood. Journal of Clinical Psychology, 38, 301-311.
lege environment is a fairly instrumental environment where goal Kenworthy, J. A. (1979). Androgyny in psychotherapy: But will it in sell in
directedness and achievement is directly rewarded). Given the situa- Peoria? Psychology of Women Quarterly, 3, 231-240.
tional variability of human behavior, other modifications of the PAQ Locksley, A., & Cohen, M. E. (1979). Psychological androgyny: A case of
to assess a broader spectrum of environments may produce different mistaken identity? Journal of Personality and Social Psychology, 37,
results. Future research in this area should perhaps assess environ- 1017-1031.
mental factors across several situations (e.g., work, home, school) as Mintz, L. B., & O’Neil, J. O. (1990). Gender roles, sex, and the process of
well as across different populations and cultures. Second, it is also psychotherapy: Many questions and few answers. Journal of Counseling
and Development, 68, 381-387.
unclear to what extent freshmen, new to the university (48% of our
Pedhazur, E. J., & Tetenbaum, T. J. (1979). Bem Sex Role Inventory: A
sample), can accurately and reliably assess environmental presses as
theoretical and methodological critique. Journal of Personality and Social
they were asked to do in this study. We did several post-hoc analyses, Psychology, 37, 996-1016.
however, and found that separating the sample into freshmen and Scher, M., & Good, G. E. (1990). Gender and counseling in the twenty-first
upperclassmen did not significantly alter any of the reported results. century: What does the future hold? Journal of Counseling and Develop-
Third, the operational definitions of masculinity and femininity used ment, 68, 388-391.
in this study represent only one conceptualization of these emotion- Sharpe, M. J., & Heppner, P. P. (1991). Gender role, gender role conflict, and
ally laden constructs (see Spence & Helmreich, 1978). These defi- psychological well-being in men. Journal of Counseling Psychology, 38,
nitions deserve more attention, which may lead to a more complex 323-330.
conceptualization of sex-related constructs. Finally, no account was Shichman, S., & Cooper, E. (1984). Life satisfaction and sex-role concept.
made for self-presentation or social desirability effects, particularly Sex Roles, 11, 227-240.
with respect to self-esteem. Given that participants were assessing the Silvern, L. E., & Ryan, V. E. (1979). Self-rated adjustment and sex-typing: Is
masculinity the primary predictor of adjustment? Sex Roles, 5, 739-763.
effects of their environment, it may make sense to control for sen-
Spence, J. T., & Helmreich, R. L. (1978). Masculinity and femininity: Their
sitivity to these pressures in future studies.
psychological dimensions, correlates, and antecedents. Austin, TX: Uni-
versity of Texas Press.
CONCLUSION Spence, J. T., Helmreich, R. L., & Stapp, J. (1974). The Personal Attributes
Questionnaire: A measure of sex role stereotypes and masculinity-feminin-
This study provides preliminary confirmation of a masculine bias in ity. JSAS Catalog of Selected Documents in Psychology, 4, 43.
American culture. Initial investigation of environmental factors sug- Stevens, M. J., Pfost, K. S., & Ackerman, M. D. (1984). The relationship
gests they may be useful sources of information when examining sex between sex-role orientation and type A behavior: A test of the main effect
roles and psychological well-being. Our results suggest a study of hypothesis. Journal of Clinical Psychology, 40, 1338-1341.
both the person and the environment may offer a more complete view Taylor, M. C., & Hall, J. A. (1982). Psychological androgyny: Theories, meth-
ods, and conclusions. Psychological Bulletin, 92, 347-366.
than a focus on the individual alone. We concur with Cook (1985)
Whitley, B. E. (1983). Sex role orientation and self-esteem: A critical meta-
in saying that an analysis of person-environmental contingencies and
analytic review. Journal of Personality and Social Psychology, 44, 765-778.
interactions should perhaps replace a focus on the role of personality Wilson, F. R., & Cook, E. P. (1984). Concurrent validity of four androgyny
variables or traits alone. instruments. Sex Roles, 11, 813-837.
Yager, G. G., & Baker, S. (1979, September). Thoughts on androgyny for the
REFERENCES counseling psychologist. Paper presented at the annual meeting of the
American Psychological Association, New York.
Adams, C. H., & Sherer, M. (1985). Sex role orientations and psychological Yoder, J. D., Rice, R. W., Adams, J., Priest, R. F., & Prince, H. T. (1982).
adjustment: Implications for the masculinity model. Sex Roles, 12, 121-128. Reliability of the Attitudes Toward Women Scale (AWS) and the Personal
Ahmed, S. M. S., Vallient, P. M., & Swindle, D. (1985). Psychometric prop- Attributes Questionnaire. Sex Roles, 8, 651-657.
erties of the Coopersmith Self-Esteem Inventory. Perceptual and Motor Zeldow, P. B., Clark, D., & Daugherty, S. R. (1985). Masculinity, femininity,
Skills, 61, 1235-1241. Type A behavior, and psychosocial adjustment in medical students. Journal
Bassoff, E. S., & Glass, G. V. (1982). The relationship between sex roles and of Personality and Social Psychology, 48, 481-492.
mental health: A meta-analysis of twenty-six studies. The Counseling Psy-
chologist, 10, 105-112.
Bem, S. L. (1974). The measurement of psychological androgyny. Journal of Jeffery W. Burnett is a doctoral intern at the University of Missouri Coun-
Consulting and Clinical Psychology, 42, 155-162. seling Center and Wayne P. Anderson and P. Paul Heppner are professors
Bem, S. L., Martyna, W., & Watson, C. (1976). Sex typing and androgyny: of psychology at the University of Missouri-Columbia. Correspondence re-
Further explorations of the affective domain. Journal of Personality and garding this article should be addressed to Wayne P. Anderson, 209 McAlester
Social Psychology, 34, 1016-1023. Hall, University of Missouri-Columbia, Columbia, MO 65211.

326 Journal of Counseling & Development l January/February 1995 l Volume 73