You are on page 1of 10

P1: GDX

Sex Roles [sers]

pp673-sers-455292

November 8, 2002

16:50

Style file version June 3rd, 2002

C 2002)
Sex Roles, Vol. 47, Nos. 5/6, August 2002 (

Gender-Role Identity, Attitudes Toward Marriage,


and Gender-Segregated School Backgrounds
Emiko Katsurada1,3 and Yoko Sugihara2

The purpose of this study was twofold: (1) to investigate the relationship between genderrole identity and attitudes toward marriage by comparing Bems gender schema theory
and Spences multifactorial model of gender identity; (2) to examine the effects of gendersegregated school backgrounds on gender-role identity and attitudes toward marriage. A total
of 524 male and 696 female Japanese college students completed the Japanese version of
the Bem Sex Role Inventory and a series of questions regarding attitudes toward marriage.
Overall results were more supportive of Spences multifactorial model. The effect of school
background was found only in women; women without any coeducational school background
had relatively strong masculinity and desired to marry at an older age, but tended to have a
conservative opinion about men taking nontraditional roles.
KEY WORDS: gender-role identity; attitudes toward marriage; school background.

INTRODUCTION

tity and attitudes toward marriage in Japanese college


students.
The secondary purpose of our study was to explore environmental influences on gender-role identity and attitudes toward marriage in college students.
Theories of gender-role development emphasize the
importance of environmental influences on genderrole development. Although parental influences on
gender-role development have been studied well
(Fagot & Leinbach, 1989; Haigler, Day, & Marshall,
1995; Jackson, Ialongo, & Stollack, 1986; Lombardo
& Kemper, 1992; Orlofsky, 1979), the influences of
school environment on gender role development do
not seem to have been investigated often. School
is an important place where socialization of children takes place. In adolescence, boys and girls
spend a large amount of time in schools where
the gender-egalitarian principle and sexism coexist (Kimura, 1999). Therefore, school influences on
gender-role development cannot be ignored. In the
present study, we examined whether school backgrounds (coeducational or gender-segregated junior
high and/or high schools) had any effect on the college
students gender-role identity and attitudes toward
marriage.

Perry and Bussey (1984) defined gender-role development as the process whereby children come to
acquire the behaviors, attitudes, interests, emotional
reactions, and motives that are culturally defined
as appropriate for members of their sex (p. 262).
This definition suggests that gender roles include
behavioral, attitudinal, and personality aspects. It
also assumes that these gender aspects are closely
related to each other because they are expected
to be consistent. Culturally, Japan has strong traditional gender stereotypes compared to other countries (Inoue, 1992; Williams & Best, 1990). This fact
suggests that gender-related phenomena in Japan
should be consistent. The primary purpose of the
present study is to examine the interrelatedness of
gender phenomena by focusing on gender-role iden-

1 Akita

University, Akita, Japan.

2 Los Angeles County Department of Mental Health, Los Angeles.


3 To

whom correspondence should be addressed at Department of


Education and Human Studies, 1-1 Tegata-Gakuenmanchi, Akita
010-8502, Japan; e-mail: katsurada@ed.akita-u.ac.jp.

249

C 2002 Plenum Publishing Corporation


0360-0025/02/0900-0249/0

P1: GDX
Sex Roles [sers]

pp673-sers-455292

November 8, 2002

16:50

Style file version June 3rd, 2002

250
Theoretical and Empirical Controversy on the
Interrelatedness of Behavioral, Attitudinal,
and Personality Aspects of Gender Roles
Both Bem (1974) and Spence (Spence,
Helmreich, & Stapp, 1974) adopt the trait approach to measure an individuals masculinity and
femininity. They also agree on the conception that
masculinity and femininity are two independent
unidimensional properties (Spence, 1984). However,
with respect to the interrelatedness of gender phenomena, they have different perspectives. According
to the gender schema theory (Bem, 1981), peoples
gender typing is the result of gender-schematic processing. Gender-schematic persons tend to process
information, including information about themselves,
according to the cultures definitions of masculinity
and femininity (Bem, 1985). The gender schema
theory suggests the interrelatedness of genderrelated phenomena: gender-personality type, gender
attitudes, and gender-related behaviors.
On the other hand, Spence (1984, 1985) proposed
the multifactorial model of gender identity; she suggested that gender-related personality, attitudes, and
behaviors are relatively independent. She stated that
at the level of the individual these different kinds of
gender-related attributes, attitudes, and behaviors do
not necessarily have common developmental histories (Spence, 1984, p. 3). On the basis of her model,
she further claimed that the instruments that measure
individuals gender-role identity such as the Bem Sex
Role Inventory (BSRI; Bem, 1974) and the Personal
Attributes Questionnaire (PAQ; Spence et al., 1974)
were not for assessing individuals broad gender concepts such as gender-typing or gender schematization,
but for measuring individuals desirable aspects of instrumentality and expressiveness (Spence, 1993).
Both of these theoretical perspectives have some
empirical support. Consistent with Bems perspective, Ickes and his colleagues found an association
between gender-role identity and social behaviors
(Ickes & Barnes, 1978; Ickes, Schermer, & Steeno,
1979). Collins, Waters, and Waters (1979) found
that gender-typed men and women had less favorable attitudes toward women managers than did less
gender-typed persons. Frable (1989) also found that
gender-typed male and female undergraduates were
more likely to evaluate gender rules as fair and to dislike gender rule violators than were those who were
not gender-typed.
Several other studies provide partial support
for Bems perspective by indicating a relationship

Katsurada and Sugihara


between gender-role identity and gender attitudes
in only one gender. Abraham, Feldman, and Nash
(1978) found such a relation only in women; womens
modern gender-role attitudes were correlated positively with their masculinity scores and negatively
with their femininity scores on the BSRI. Andersen
(1978) found that masculine men tended to have
discriminatory attitudes toward women more frequently than others, but such a tendency was not
found among feminine women. In a Japanese study,
Onodera, Furukawa, and Aoki (1995) also reported
some relationships between gender-role identity
and the attitudes toward babies; androgynous men
had more positive attitudes and masculine women had
more negative attitudes toward babies.
Spence (1993) also found a statistically significant
correlation between gender-role identity and genderrole attitudes but only in men. However, because the
magnitude of the correlation was small and none of
the other correlations were statistically significant,
she concluded that gender-role identity and genderrole attitudes were relatively independent. Similarly,
Orlofsky, Cohen, and Ramsden (1985) found a moderate association between gender-role identity and
gender-related behaviors. However, because the association was far from perfect and differences in
gender-role behaviors were more strongly explained
by biological sex than by gender-role identity, they
concluded that both Bems and Spences theoretical perspectives were partially supported. Results of
other studies by Orlofsky and his colleagues were
more supportive of Spences model and indicated no
significant relation between gender-role identity and
gender-role attitudes (Orlofsky, Aslin, & Ginsburg,
1977; Orlofsky & OHeron, 1987).
Thus, the two conflicting perspectives (i.e., Bems
schema theory and Spences multifactorial gender
identity theory) have some empirical support. However, overall empirical support seems to be inclined
toward Bems theory. A majority of previous studies indicated that there was a certain degree of interrelatedness among gender phenomena. Therefore,
we would expect consistency between gender-role
identity and gender-related attitudes, although the
magnitude of the cohesion might be weak.
Traditional Gender Roles in Japan
In Japan the gender-role ideology that a man
should be a breadwinner and a woman should be
a housekeeper still exist quite obviously. An international comparison revealed that Japanese women

P1: GDX
Sex Roles [sers]

pp673-sers-455292

November 8, 2002

16:50

Style file version June 3rd, 2002

Gender-Role Identity and Atitudes


were ranked second only to Philippine women in their
approval of traditional gender-role ideology (Inoue,
1992). Also, the results of a survey conducted by a
major Japanese newspaper company in 1986 revealed
that 50% of men and 40% of women agreed that
married women could work outside the home as long
as they also worked at home as housewives (Arichi,
1993). As Arichi (1993) stated, the traditional idea
of gender roles is still strongly rooted in Japanese
society.
The traditional gender-role ideology is also reflected in Japanese womens working patterns. The
number of women in the labor force according to age
is characterized by an M-curve, which represents a
decrease in the number of women in the labor force
between the ages of 25 and 35 when marriage, childbirth, and child rearing take place, then later the
number of women in the labor force increases again
(NHK & Daiwa Soken, 1995). Another survey of public opinion conducted in 1990 showed that the average time Japanese men spent on household chores in
a day was 33 min, which was one ninth of the time
Japanese women spent on house work (Sechiyama,
1993). These reports indicate that a strong gender-role
ideology is reflected in Japanese society. Although the
personality aspect of gender roles was not included
in the above reports, they clearly show an association
between gender-role ideology and gender-role behaviors in Japan. Therefore, Bems theoretical perspective seems culturally more plausible than Spences
perspective for Japanese people.
On the basis of Bems perspective, the empirical
studies, and the cultural background reviewed above,
we hypothesized that gender-role identity would be
related to gender-role attitudes. More specifically, we
expected that men and women who had traditional
gender-role identity would be more likely to have
traditional gender-related attitudes than would those
with nontraditional gender-role identity.
In this study, we focused on the gender-role attitudes related to marriage. Ohashi (1993) stated that
marriage is a system established on the assumption of
a division of labor based on gender-role stereotypes.
Therefore, an individuals expectations for marriage
would reflect his/her acceptance of traditional gender roles. Also, by being asked about their future expectations regarding marriage, the participants could
consider it as their own personal issue rather than
as a larger social issue. These indicators should reveal a participants gender-role attitudes more honestly than his/her answers to general questions about
gender roles.

251
Effects of Gender-Segregated School
on Gender-Role Development
Along with the new educational system, coeducational schools in Japan were started after World
War II (WW II). According to the annual survey
on schools conducted by the Ministry of Education
(1999), 14% of high schools are gender segregated.
Four percent of them are only for boys, whereas 10%
of them are only for girls. Although similar statistics
on junior high schools were not available, the ratio
of the gender-segregated junior high schools is probably smaller than that of high schools because compulsory education for children in Japan is enforced
through junior high school. Thus, the majority of current junior high and high schools in Japan are coeducational. Hashimoto (1992) stated that after WW II the
coeducational system had tremendous influence on
changes in Japanese peoples gender-role perceptions
and attitudes. People in Japan generally think that
the gender segregated school system enhances traditional gender role identity and attitudes. In fact, some
people have claimed that gender-segregated schooling would be better particularly for women because
it would boost their femininity (Hamao, 1972). However, Kimura (1999) claimed that although school culture on the surface seemed egalitarian, there were
many hidden aspects that transmitted sexist ideology
to children. She further stated that the sexist ideology could be conveyed more effectively in coeducational schools where gender comparisons can be
done easily and clearly. Similarly Fujita (1993) recognized that underlying the problem of the coeducational school system was the hidden curriculum that
cultivated traditional gender-role norms in children.
Thus, it is implied that Kimura and Fujita believe
that the coeducational school system in Japan does
not necessarily produce more egalitarian gender-role
identity and attitudes.
There are few empirical studies of the influence
of gender-segregated schools on gender phenomena.
A study conducted in England showed that students
aged 1415 years who attended coeducational schools
reported more stereotypical perceptions of genderlinked traits, general abilities, school subjects, and
career choices than did those in gender-segregated
schools (Lawrie & Brown, 1992). In Japan, Ito (1997,
1998) examined the influences of gender-segregation
in schools on gender conceptions and gender-related
awareness among high school students. She found
that male high school students in boys only schools
had more stereotypical gender conceptions than their

P1: GDX
Sex Roles [sers]

pp673-sers-455292

November 8, 2002

16:50

Style file version June 3rd, 2002

252
counterparts in coeducational schools, whereas such
a difference between schools was not found among
girls. The results of these two studies are inconsistent
and reflected the debate on the effect of coeducational
schooling.
In both of the previous studies the researchers
examined concurrent effects of school setting in the
different ages groups, which may contribute to the
inconsistent results. In the present study, we examined the participants history of school environment in
terms of gender and how this variable might influence
college students gender-role identity and attitudes toward marriage. In this way, we would be able to see
the gender-segregated school effect from a developmental viewpoint.
In general, school culture reflects the genderrole structure in Japanese society (Fujita, 1993), so
in gender-segregated schools students have to assume both roles. Therefore, we hypothesized that people who had graduated from gender-segregated junior high and high schools would have nontraditional
gender-role identity and more egalitarian attitudes toward marriage than would those who had graduated
from coeducational junior high and/or high schools.

METHOD
Participants
The final 1,220 participants in this study consisted
of 524 male and 696 female unmarried college students. Married participants (n = 10) were excluded
from the analysis. The average age of participants was
19.5 years old (SD = 1.6), with a range of 1831 years
of age. The students majors were diverse, including
arts and humanities, education, economics, agriculture, engineering, medicine/dentistry, law, natural science, and marine science.

Measures
The Japanese version of the Bem Sex Role Inventory (Katsurada & Sugihara, 1999; Sugihara &
Katsurada, 1999) was employed to measure individuals gender-role identity. A validation study of the
BSRI in Japanese culture (Sugihara & Katsurada,
2000) indicated that 12 masculine and 7 feminine
items of the original BSRI possessed the etic (universal) aspect of gender-related personality traits in
Japan. In the present study, the average score on these

Katsurada and Sugihara


12 masculine items (Defend my own beliefs, Independent, Assertive, Strong personality, Forceful, Have
leadership abilities, Willing to take a risk, Willing
to take a stand, Self-reliant, Competitive, Ambitious,
and Act as a leader) was used for the individuals masculinity score and that of the 7 feminine items (Affectionate, Eager to soothe hurt feelings, Tender, Love
children, Gentle, Cheerful, and Soft-spoken) was used
for the individuals femininity score. Cronbachs alpha
for this modified BSRI was .62 for both masculinity
and femininity scales.
Attitudes toward marriage were measured by
three questions regarding (1) marriage (expectations
and desired age), (2) women working after marriage,
and (3) men taking nontraditional gender roles. For
the attitudes toward marriage, the participants were
asked to choose one from the following items: I definitely want to marry, I want to marry, I dont mind
being single for life, or I dont want to marry at all,
and were asked to write the age when they would like
to marry. As for the question about women working
after marriage, the choices provided for men were I
want my future wife to quit her job when we marry,
I want my future wife to quit her job when we have
a child, I want my future wife to continue working even after we have a child, and It is up to my
future wife to work or not. For women, the choices
were I want to quit my job when I marry, I want
to keep my job until I have a child, and I want to
keep my job even after I have a child. The participants were asked to select one of the above choices.
As for men taking nontraditional roles, both men and
women were asked whether they agree or do not
agree with the idea of house husband, that is, a
man who stays at home and takes full responsibility
for domestic duties.
Amongst the demographic questions, we asked
whether participants went to coeducational, male
only, or female only junior high schools and high
schools. On the basis of their answers we divided the
students into three groups: both junior high and high
school were coeducational, either one was coeducational, and neither of them were coeducational.

Procedure
We solicited cooperation from colleagues in universities and colleges in various areas of Japan. Colleagues in 14 universities and colleges agreed to collect data. They administered questionnaires during
their class time. Participation was on a voluntary basis;

P1: GDX
Sex Roles [sers]

pp673-sers-455292

November 8, 2002

16:50

Style file version June 3rd, 2002

Gender-Role Identity and Atitudes


students neither got any extra credit nor were penalized for not answering the questionnaire. Anonymity
was kept because the whole procedure did not involve
taking any individual identification.

RESULTS
The medians of masculinity and femininity scores
were 3.80 and 4.33 for men and 3.60 and 4.33 for
women. Using the median split method, the participants were categorized into four gender-types. Men
and women similarly fell in the Bems four groups.
Eighteen percent of men were identified as Masculine,
19% as Feminine, 33% as Androgynous, and 30% as
Undifferentiated. Of the women, 19% were identified
as Masculine, 20% as Feminine, 31% as Androgynous,
and 30% as Undifferentiated. To make the examination of hypotheses easy, these four types were rearranged into two groups: traditional and nontraditional
groups. Masculine type was considered traditional for
men and Feminine type for women; the rest were considered nontraditional. These two gender types were
used to examine the hypotheses in this study.

Relationships Between Gender-Role Identity


and Attitudes Toward Marriage
To check the reliability of three questions about
the attitude toward marriage, correlations of each
question were examined. For the association involving mens attitudes toward womens work after marriage, a chi-square analysis was conducted because the
response choices for this question were not strictly
considered as continuous. Statistically significant but
weak correlation coefficients and a weak phi indicated
that three questions about the attitude toward marriage were generally consistent but measured different attitudes. For example, women who had a strong
desire for marriage wanted to marry at younger age
(r = .359, p < .01), were less motivated to work after marriage (r = .208, p < .01), and more likely
to disagree with the idea of house husband (rpb =
.239, p < .01). Also, men who had a strong desire
for marriage wanted to marry at younger age (r =
.301, p < .01), were less liberal about women working after marriage, 2 (9, N = 509) = 16.71, p = .053,
= .181, and more likely to disagree with the idea of
house husband (rpb = .136, p < .01).
To investigate the relationship between genderrole identity and attitudes toward marriage, t-tests

253
Table I. Means, Standard Deviations, and t-Values of Traditional
and Nontraditional Womens Attitudes Toward Marriage
Traditional

Desire for marriage


Age they want
to marry
Desire for work
after marriage
p

Nontraditional

Mean

SD

Mean

SD

3.02
25.96

0.52
2.97

2.88
26.75

0.57
4.28

2.66
2.02

2.38

0.69

2.54

0.64

2.59

< .05.
< .01.

were conducted. For the relationship involving mens


attitudes toward womens work after marriage, a chisquare analysis was conducted. As shown in Table I,
traditional women had stronger desire for marriage
than did nontraditional women. On the other hand,
nontraditional men had stronger desire for marriage
than did traditional men (see Table II). As for the age
they want to marry, a significant difference between
groups was found only in women; traditional women
wanted to marry at younger age than did the nontraditional group (see Table I).
To the question about women working after marriage, the majority of men students (77.4%) answered
It is my future wifes decision, 11.5% of them answered I want my future wife to work until we have a
child, and the rest answered I want my future wife to
quit her job when we marry (5.3%) and I want my
future wife to continue working even after we have
a child (5.8%). The chi-square analysis, however,
did not show significant association between these responses and gender types.
More than half of the women students (59.7%)
answered I want to continue working even after I
have a child, one third (31.7%) answered I want to
work until I have a child, and the rest (8.6%) answered I want to quit my job when I marry. The result of the t-test indicated that nontraditional women
had stronger desire to work after marriage than did
nontraditional women (see Table I).
Table II. Means, Standard Deviations, and t-Values of Traditional
and Nontraditional Mens Attitudes Toward Marriage
Traditional

Desire for marriage


Age they want
to marry
p

< .01.

Nontraditional

Mean

SD

Mean

SD

2.69
27.72

0.63
5.45

2.88
27.61

0.58
3.87

2.90
ns

P1: GDX
Sex Roles [sers]

pp673-sers-455292

November 8, 2002

16:50

Style file version June 3rd, 2002

254

Katsurada and Sugihara


Table III. The Means (and Standard Deviations) of Womens Masculinity and Femininity
Scores Divided by Educational Background
Educational background

Masculinity

Femininity

Both junior high and high


schools were coeducational
Either junior high or high
school was coeducational
Both junior high and high
schools were for women only

430

3.60a (0.90)

4.26 (0.90)

134

3.52a (0.82)

4.31 (0.89)

34

4.00b (1.04)

4.62 (1.04)

F(2, 594) = 3.92

F(2, 594) = 2.56

Note. Means in the same row that do not share subscripts differ at p < .05. in the HSD
comparison.
p < .10. p < .05.

To the question about men taking nontraditional


gender roles, more than half of men and women (57%
for both) agreed with the idea of house husband.
However, the association between gender types and
the responses to this question was not significant in
either men or women.
Effects of Gender-Segregated Educational
Background
The majority of men (83%) and women (72%)
went to coeducational junior high and high schools.
Only 7% of men and 6% of women went to both
gender-segregated junior high and high schools. The
rest (10% of men and 22% of women) went to either
gender-segregated junior high or high schools.
Chi-square analyses indicated that the association between gender types and educational
backgrounds was not statistically significant in either
men or women. Gender-segregated educational
backgrounds did not affect the formation of traditional or nontraditional gender identity. However, it
may be possible that gender-segregated educational
backgrounds influence an individuals masculinity
and femininity differently, hence dividing people
into two gender types would make it harder to
detect its influence. Therefore, we also examined its
effect on the masculinity and femininity scores using

one-way ANOVA. The results showed that women


who graduated from womens junior high and high
schools had significantly higher masculinity scores
than those from coeducational junior high and/or
high schools (see Table III). For men, however, there
were no significant differences in either masculinity
or femininity scores.
In the relationships between school backgrounds
and attitudes toward marriage, statistical significance
was found only in womens desired age for marriage and their attitude toward men taking nontraditional gender roles. Table IV indicates the result of
the one-way ANOVA. Women who had only gendersegregated school background wanted to get married
at an older age than did those who had coeducational
school backgrounds (see Table IV). Also, more than
half of the women who graduated from coeducational
junior high and high schools agreed with the idea of
house husband. On the other hand, more than half
of the women who had gender-segregated school experience in junior high and/or high school disagreed
with this idea (see Table V).
DISCUSSION
The primary purpose of the present study was
to examine the interrelatedness of gender phenomena by focusing on gender-role identity and attitudes

Table IV. The Means (and Standard Deviations) of Womens Desired Age for Marriage Divided by
Educational Background
Educational background

Age they want to marry

Both junior high and high schools were coeducational


Either junior high or high school was coeducational
Both junior high and high schools were for women only

408
128
32

26.49a (2.86)
26.63a (5.49)
28.41b (9.65)
F(2, 565) = 3.08

Note. Means in the same row that do not share subscripts differ at p < .05 in the HSD comparison.
p < .05.

P1: GDX
Sex Roles [sers]

pp673-sers-455292

November 8, 2002

16:50

Style file version June 3rd, 2002

Gender-Role Identity and Atitudes

255

Table V. Frequency Distribution of Womens Attitudes Toward the Idea of House Husband
According to the Educational Background
Idea of house husband
Educational background
Both junior high and high schools were coeducational
Either junior high or high school was coeducational
Both junior high and high schools were for women only

Agree

Disagree

264 (63%)
57 (44%)
16 (47%)

158 (37%)
74 (56%)
18 (53%)

2 (2, N = 587) = 16.42


Note. Figures in ( ) indicate the percentage of each educational background.
p < .05.

toward marriage among Japanese college students.


On the basis of the previous studies and cultural background, we hypothesized that in Japan gender-typed
men and women would have more traditional attitudes toward marriage than androgynous, undifferentiated, and cross-gender-typed individuals.
The results of the analysis of data from women
generally supported our hypothesis. Women who had
traditional gender-role identity were more likely to
desire marriage, want to marry at a younger age, and
were less motivated for work after marriage than
were their nontraditional counterparts. For men, however, a significant association was found only between
their desire for marriage and gender type, and it was
somehow inconsistent with our hypothesis. Men who
had traditional gender-role identity were less likely to
desire marriage than were nontraditional men. This
result may suggest that traditional men would feel
strong responsibility in marriage, and they are not
ready to take such responsibility because they are still
university students. As mentioned before, in Japan
men are traditionally expected to be the sole breadwinner after marriage. Although the desire for marriage means accepting traditional gender roles for
men and women, because mens and womens roles
in marriage are different, the readiness for accepting
those roles may vary in men and women. The gender
difference observed in this study may reflect different readiness for marriage between men and women
students.
This mens result, however, is consistent with
Spences (1985) claim. She stated that masculinity
and femininity as personality traits measured by
the PAQ and the BSRI would successfully predict
other gender-related phenomena that were influenced by self-assertive and interpersonally oriented
characteristics, but would not predict a broad genderrelated construct. Mens traditional gender-role identity (strong masculinity) was not related to their desire
for marriage, which reflected interpersonally ori-

ented characteristics, but womens traditional genderrole identity (strong femininity) was. Also, men and
womens traditional gender-role identity was not related to their attitude toward men taking nontraditional gender roles, which reflected neither selfassertive nor interpersonally oriented characteristics.
Another association found in this study was between
womens nontraditional gender-role identity (strong
masculinity) and their motivation to work after marriage, which reflected self-assertive characteristics.
Thus, the present results seem to sustain Spences theory more strongly than Bems gender schema theory.
The secondary purpose of this study was to examine the effect of gender-segregated school environment on gender-role identity and attitudes toward marriage with a developmental perspective.
We expected that gender-segregated school backgrounds would enhance nontraditional gender-role
identity and liberal attitudes toward marriage. When
we examined the hypothesis in terms of traditional
and nontraditional gender types, there was no effect
of gender-segregated school backgrounds on gender
types. However, when we examined it in terms of
masculinity and femininity scores, the results indicated that the women who went to gender-segregated
junior high and high school had higher masculinity
scores than those who had coeducational school backgrounds. They also tended to have higher femininity
scores than those who had coeducational school backgrounds.
As mentioned before, school life in junior high
and high school requires both masculine and feminine characteristics. Actually, students perceive various roles in school according to gender. For example, they consider that the president of the student
body or a leader of a class should be a boys job and
the secretary of the student body or a sub-leader of
a class should be a girls job (Kimura, 1999). Female
students in gender-segregated junior high and high
school assume all of these roles. Therefore, women

P1: GDX
Sex Roles [sers]

pp673-sers-455292

November 8, 2002

16:50

Style file version June 3rd, 2002

256
who spent 6 years in a gender-segregated school environment have developed their masculine personality traits as well as their feminine personality traits,
whereas women who had coeducational school backgrounds may not have developed theirs. Also, if the
sexist ideology is conveyed more effectively in coeducational settings as Kimura (1999) stated, women
who had a longer coeducational school experience
may have received stronger negative messages about
women. Therefore, such educational experience may
have inhibited womens development of masculinity.
Another explanation for high masculinity of
women with gender-segregated school backgrounds
is self-selected effect; masculine girls might have selected gender-segregated schools, thus their masculinity scores could have been high before they entered
a junior high and high school. It is generally believed that gender-segregated schools are for feminine women or masculine men in Japan. In accordance
with the different gender-role expectations for men
and women in Japanese society, different education
was provided to men and women through gendersegregated schools before World War II. From this
historical background it is unlikely that masculine
women chose gender-segregated schools. However,
some gender-segregated schools in Japan have high
prestige with respect to competitiveness. Therefore,
it is possible that competitive girls chose the school
because of its high ranking rather than because it
was a gender-segregated school. Because the competitiveness ranking of the schools is unknown in the
present study, we cannot deny the possibility of the
self-selected effect. Future research on the effect of
school environment should include this aspect as well.
Regarding the effect of gender-segregated school
backgrounds on womens gender-related attitudes,
the results indicated that women who went to gendersegregated junior high and high school wanted to
marry at an older age and tended to disagree with the
idea of house husband. Women who had gendersegregated school backgrounds are more liberal in
terms of their own age for marriage, but more traditional in terms of their attitudes toward men taking nontraditional gender roles. Gender-segregated
school backgrounds seem to have different influences
on the attitudes toward womens own gender roles
and those toward mens.
Despite the differences in the age of the participants and the targeted gender phenomena, the
present results of the gender-segregated school effects appear to be consistent with Lawrie and Browns
findings (1992) to a certain degree; gender-segregated

Katsurada and Sugihara


schooling, rather than coeducational schooling, has
some effects on enhancing womens nontraditional personality aspects (masculinity) and prolonging womens desired age for marriage. However,
the present results are inconsistent with previous
Japanese studies (Ito, 1997, 1998). Ito found the effect of gender-segregated schooling among only boys,
whereas no effect was found among men in the
present study. Ito focused on high school students
gender construct and gender awareness, therefore the
discrepancy may be due to the different gender phenomena investigated in each study. Or, it may be due
to the difference between short-term and long-term
effects of the school environment. She investigated
the concurrent effect of gender-segregated schooling, whereas our investigation focused on the developmental effect. Future research employing the
longitudinal method would clarify this speculation.
A limitation of this study is related to use of
the Japanese version of the Bem Sex Role Inventory
(BSRI). Frable (1989) stated that the BSRI is an
appropriate measurement instrument for studies trying to link gender personality and ideology (p. 106).
However, employing the BSRI with the Japanese samples has certain limitations because the BSRI measures only common aspects of gender-role identity
based on American norms. It excludes the culturally
specific aspect of gender-role identity in Japan. In a
study (Sugihara & Katsurada, 2001) designed to identify socially desirable masculine and feminine characteristics in Japanese society, we found many personality traits that were not included in the BSRI. For
example, characteristics such as Have a broad perspective and Play fair were identified as masculine,
whereas characteristics such as Polite, Calm, and
Neat were identified as feminine. When we examine gender identity using an instrument that includes
those culturally specific traits, different results may
be produced. We recommend that the measurement
used in future research in Japan include culturally
specific aspects of gendered personality traits as well.
Another limitation is the lack of information
about participants sexual orientation. Although it is
culturally inappropriate to ask such a private question
in a survey in Japan, participants sexual orientation
could be important in relation to their desire for marriage. In this study, it was thought that an individuals
desire for marriage would reflect his/her acceptance
of traditional gender roles. This assumption is based
on the heterosexual relationship. Although the population of homosexuals in Japan is reported to be between 3 and 4% (Ito, 1996, cited in McLelland, 2000),

P1: GDX
Sex Roles [sers]

pp673-sers-455292

November 8, 2002

16:50

Style file version June 3rd, 2002

Gender-Role Identity and Atitudes


it is possible that a significant number of homosexuals
were included in our participants. In that case, their
desire for marriage may not really represent their acceptance of traditional gender roles. To exclude this
possible contamination, we should dare to ask participants sexual orientation in the future research.
Despite the limitations, there are some important implications of this study. First, the interrelatedness of gender phenomena is complicated and sensitive. What kind of gender-related attitudes, such as
general or personal, or whether the attitude is toward the persons own gender or the other gender
seems to have different association with their genderrole identity. Also, because of the sensitive nature
of the interrelatedness, employing the gender type
as a unit of analysis sometimes disguises existing effects. Some researchers have criticized dividing gender types based on the BSRI scores (Bernard, 1980;
Pedhazur & Tetenbaum, 1979), and the present study
also shows the limitation of using gender types. Future
researchers should keep this methodological point in
mind.
The second implication is based on the conflicting influences of the coeducational school system on gender phenomena. Kimura (1999) and Fujita
(1993) stated that the coeducational school system
had both egalitarian and sexist aspects. The results of
the present study endorse their proposition and suggest that a coeducational school environment might
enhance egalitarian gender-role attitudes in women
as indicated by their positive attitude toward men
taking nontraditional gender roles, but that it does
not necessarily produce more androgynous women in
terms of personality traits. School background, however, did not influence mens gender personality traits
and gender attitudes. This differential effect of school
backgrounds on men and women may suggest that the
dominant characteristics in school culture are masculine. In another words, school environment seems
to encourage the instrumental aspect of gender roles
such as competitiveness more strongly than the interpersonal aspect. More empirical studies are needed
for more conclusive effects of coeducational/gender
-segregated school systems on childrens gender role
development.

ACKNOWLEDGMENTS
The authors thank the students who answered
the questionnaire for this study and the friends who
collected data for us. We also thank Roberta Golliher

257
and Gene Pleish for editing this paper. Our appreciation extends to the anonymous reviewers for their
constructive comments.
REFERENCES
Abraham, B. S., Feldman, S., & Nash, S. C. (1978). Sex role selfconcept and sex role attitudes: Enduring personality characteristics or adaptations to changing life situations. Developmental
Psychology, 14, 393400.
Andersen, S. M. (1978). Sex-role typing as related to acceptance of
self, acceptance of others, and discriminatory attitudes toward
women. Journal of Research in Personality, 12, 410415.
Arichi, T. (1993). Kazoku wa kawattaka [Have families changed?].
Tokyo: Yuhikaku.
Bem, S. L. (1974). The measurement of psychological androgyny.
Journal of Consulting and Clinical Psychology, 42, 155162.
Bem, S. L. (1981). Gender schema theory: A cognitive account of
sex typing. Psychological Review, 88, 354364.
Bem, S. L. (1985). Androgyny and gender schema theory: A conceptual and empirical integration. In T. B. Sonderegger (Ed.),
Nebraska symposium on motivation: Psychology and gender
(pp. 179226). Lincoln, NE: University of Nebraska Press.
Bernard, L. C. (1980). Multivaritate analysis of new sex role formulations and personality. Journal of Personality and Social
Psychology, 38, 323336.
Collins, M., Waters, L. K., & Waters, C. W. (1979). Relationships
between sex-role orientation and attitudes toward women as
managers. Psychological Reports, 45, 828830.
Fagot, B. I., & Leinbach, M. D. (1989). The young childs gender
schema: Environmental input, internal organization. Child Development, 60, 663672.
Frable, D. E. S. (1989). Sex typing and gender ideology: Two facets
of the individuals gender psychology that go together. Journal
of Personality and Social Psychology, 56, 95108.
Fujita, H. (1993). Kyoikuni okeru seisa to jenda [Gender differences
and gender in education]. In Seisa to bunka (pp. 257294).
Tokyo: Tokyo University Press.
Haigler, V. F., Day, H. D., & Marshall, D. D. (1995). Parental attachment and gender-role identity. Sex Roles, 33, 203220.
Hamao, M. (1972). Onnanoko no shitsuke kata: Yasashii kodomo
ni sodateru hon [How to discipline female children: A book
for raising gentle girls]. Tokyo: Kobunsha.
Hashimoto, N. (1992). Danjo kyougaku no shiteki kenkyu [A historical study of coeducational system]. Tokyo: Ohtsuki Shoten.
Ickes, W., & Barnes, R. D. (1978). Boys and girls together- and alienated: On enacting stereotyped sex roles in mixed-sex dyads.
Journal of Personality and Social Psychology, 36, 669683.
Ickes, W., Schermer, B., & Steeno, J. (1979). Sex and sex-role influences in same-sex dyads. Social Psychology Quarterly, 42,
373385.
Inoue, T. (1992). Joseigaku heno shyotai [Invitation to womens
study]. Tokyo: Yuhikaku.
Ito, Y. (1997). Koukousei niokeru seisakan no keiseikannkyou to
seiyakuwari sentaku [The formative factors of gender conception and its relationship with a selection of gender roles in
adolescents]. Kyouiku Shinrigaku Kenyu, 45, 396404.
Ito, Y. (1998). Koukousei no jender wo meguru ishiki [Relationship
between gender conception and gender-related awareness and
experiences in adolescents]. Kyouiku Shinrigaku Kenyu, 46,
247254.
Jackson, L. A., Ialongo, N., & Stollak, G. E. (1986). Parental correlates of gender role: The relations between parents masculinity, femininity, and child-rearing behaviors and their childrens
gender roles. Journal of Social and Clinical Psychology, 4, 204
224.

P1: GDX
Sex Roles [sers]

pp673-sers-455292

November 8, 2002

16:50

Style file version June 3rd, 2002

258
Katsurada, E., & Sugihara, Y. (1999). A preliminary validation of
the Bem Sex Role Inventory in Japanese culture. Journal of
Cross-Cultural Psychology, 30, 641645.
Kimura, R. (1999). Gakko bunka to jenda [School culture and gender]. Tokyo: Sokeisha.
Lawrie, L., & Brown, R. (1992). Sex sterotypes, school subject preferences, and career aspirations as a function of single/mixedsex schooling and presence/absence of an opposite sex
sibling. British Journal of Educational Psychology, 62, 132
138.
Lombardo, J. P., & Kemper, T. R. (1992). Sex role and parental
behaviors. Journal of Genetic Psychology, 153, 102113.
McLelland, M. J. (2000). Male homosexuality in modern Japan.
Richmond: Curzon Press.
Ministry of Education. (1999). Gakkou kihon chosa houkokusho
[Report on the basic survey of schools]. Tokyo: Author.
NHK International Department Economic Project, & Daiw Soken
Economic Department. (1995). A bilingual guide to the
Japanese economy. Tokyo: Kodansha International.
Ohashi, T. (1993). Mikonka no shakaigaku [Sociology of singlization]. Tokyo: NHK books.
Onodera, A., Furukawa, M., & Aoki, K. (1995). Kazokukeisei ni
kansuru kenkyu II [Study on family formation, II]. Bosei Eisei,
36, 274279.
Orlofsky, J. L. (1979). Parental antecedents of sex-role orientation
in college men and women. Sex Roles, 5, 495512.
Orlofsky, J. L., Aslin, A. L., & Ginsburg, S. O. (1977). Differential
effectiveness of two classification procedures on the Bem Sex
Role Inventory. Journal of Personality Assessment, 41, 414
416.
Orlofsky, J. L., Cohen, R. S., & Ramsden, M. W. (1985). Relationship between sex-role attitudes and personality traits
and the revised sex-role behavior scale. Sex Roles, 12, 377
391.
Orlofsky, J. L., & OHeron, C. A. (1987). Stereotypic and nonstereotypic sex role traits and behavior orientations: Implications

Katsurada and Sugihara


for personal adjustment. Journal of Personality and Social
Psychology, 52, 10341042.
Pedhazur, E. J., & Tetenbaum, T. J. (1979). Bem Sex Role Inventory:
A theoretical and methodological critique. Journal of Personality and Social Psychology, 37, 9961016.
Perry, D. G., & Bussey, K. (1984). Social development. Englewood
Cliffs, NJ: Prentice-Hall.
Sechiyama, K. (1993). Tassei no kanatae: Feminism wa mou furuika
[Beyond the accomplishment: Is feminism already old?]. In
S. Kato, K. Sakamoto, & K. Sechiyama (Eds.), Feminizumu
Korekusion: Seido to tassei (pp. 361390). Tokyo: Keiso Shobo.
Spence, J. T. (1984). Masculinity, femininity, and gender-related
traits: A conceptual analysis and critique of current research.
Progress in Experimental Personality Research, 13, 297.
Spence, J. T. (1985). Gender identity and implications for concepts
of masculinity and femininity. In T. B. Sonderegger (Ed.), Nebraska symposium on motivation: Vol. 32. Psychology and gender (pp. 5996). Lincoln, NE: University of Nebraska Press.
Spence, J. T. (1993). Gender-related traits and gender ideology:
Evidence for multifactorial theory. Journal of Personality and
Social Psychology, 64, 624635.
Spence, J. T., Helmreich, R., & Stapp, J. (1974). The personal attributes questionnaire: A measure of sex-role stereotypes and
masculinity and femininity. JSAS: Catalog of Selected Documents in Psychology, 4, 4344.
Sugihara, Y., & Katsurada, E. (1999). Masculinity and femininity
in Japanese culture: A pilot study. Sex Roles, 40, 635646.
Sugihara, Y., & Katsurada, E. (2000). Gender role personality traits
in Japanese culture. Psychology of Women Quarterly, 24, 309
318.
Sugihara, Y., & Katsurada, E. (2001, August). Developing gender
role scale in Japanese culture. Poster session presented at the
annual convention of the American Psychological Association,
San Francisco, CA.
Williams, J. E., & Best, D. L. (1990). Sex and psyche: Gender and
self viewed cross-culturally. Newbury Park, CA: Sage.