Embed Size (px)
Citation preview
香字苣中文大學硏究院教育學咅 THE CHINESE UNIVERSITY OF HONG KONG GRADUATE SCHOOL . DIVISION OF EDUCATION
文科教育硬士論文 Master of Arts in Education Thesis
1¾文題目 Thesis Title SEX DIFFERENCE IN ACADEMIC CHOICE: THEIR RELATIONS
WITH SEX-ROLE ORIENTATION AND SEX STEREOTYPES
選科之性別差異-性別角色、性別偏見與它們之關偽
撰作語言 英文 Language Used English
硏究生姓名 鄭慕賢 Name of Student CHENG Mo Yin
專修範圍 教育心理學 Specialization Educational Psychology
論文考試委員會
Thesis Examination Committee
論文導師
Thesis Supervisor Dr. LAU Sing 翻誠，博士 _ 校内委員 Internal Examiner Mr, HAU Kit Tai 侯傑泰f 先生 校内委貢 Internal Examiner Dr. CHEUNG Ping Chung 張炳松博士__
校外委員 External Examiner - Dr. H.W. STEVEHSOH
學部主任 _ Division Head Dr. LAM Man Ping 博士
論文通過日期 。 Date of Approval September 4, 1992
, v
SEX DIFFERENCES IN ACADEMIC CHOICE :
THEIR RELATIONS WITH SEX-ROLE ORIENTATION
& SEX-STEREOTYPES
by
Vivian Mo-yin Cheng
A dissertation submitted in partial fulfillment of the requirements for the degree of
Master of Arts (Education)
in The Chinese University of Hong Kong 1992
i \
TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . ~ . . . . . , . . . . . . . . . . . . i ii iv v v
ABSTRACT ............... 0 • 0 • 0 ••••• 0 •••••• 0 • 0 0 •••• 0 . ,
LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . LIST OF FIGURES ................................ 0 • 0
LIST OF APPENDICES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CHAPTER
10
11.
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Purpose of the Study Significance of the Study
REVIEW OF SELECTED LITERATURE Sex differences Sex-stereotypes Sex-role Orientation Expectancy-value Model
.................
2 2 3
5 5 8
14 19
I I. METHOD 0" 0 •••• 0 • 0 0 ' ~ •• 0 •• 0 •• 0 0 • 0 • 0 0 • 0 0 • 0 0 0 • 0 0 o. 22 Study Overview 22 Subjects 22 Procedure 23 Measures of 25
Academic Choice 25 Acheivement Motives 26 Sex-stereotypes 29 Sex-role Orientation 34 Past School Performances 39
Research Hypotheses 42 Data Analyses 43
IV 0 'RESULTS 0 0 0 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •• 47 Sex-related differences 47 Effects of Sex-role Orientation 50 Effects of Sex-stereotypes 54 Path Model for Academic Choice 56
V. /DISCUSSION ............. , ...... ".,'" .", .. ,.," 61 Summary of Findings 61 Implication ,of Findings 62 Further discussion on Sex-role Orientation 68 Directions for future research trend 72 Limitations of the study 74
APPENDICES ••••••••••••••••••••••••••••••• , I •••• , •• 76
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
ACHNOWLEDGEMENTS
I wish to express my special thank to my supervisor
Dr. · Lau Sing. From the day when I was still studying ln
Postgraduate Diploma of Education, Dr. Lau Sing has
encouraged me to do research on gender. I think, without
his prolonged support and encouragement, I would not
start writing this thesis. I would also like to thank
the other two advisors of my thesis , Mr. Hau Kit Tai and
Dr. Cheung Ping Chung, for their invaluable advice and
support.
I have to thank three of my students in secondary
schools who had spent days to input the raw data into
computer. Thanks are extended to my friends who take the
trouble of conducting the questionnaire in their schools
and collecting the examination results of the whole Form
3. Thanks are also extended to the five schools and all
the students who participate in the questionnaire.
Without the help of all these persons, this research
can't be possible. My friends , my family and these
persons together are .my source of support as I made Iny
way, through ·hardships. I feel ram in debt to all of
them.
i
ABSTRACT
The present study examined the relationships
between sex-role orientation~ sex~stereotypes and .
academic' choice. The principal questions were: (1) Do
males and females differ in a~ademic choice? (2) Do
sex-role orientation and sex-stereotypes have any effect
on academic choice? (3) What factors mediate their
effects on academic choice? The control variables of
this study are past school performances and sex.
The sample included 818 Form 3 Chinese students in 5
co-educational schools of Hong Kong. Data were collected
to measure students I intended choice between Arts and
Science , streams~ their expectancies~ interest and utility
values towards the two streams, their sex-role
orientation and their sex-stereotypes in Arts and
Science. Sex-role orientation was measured by a
simplified translated version of Bemls Sex Role
Inventory. New scales were developed to tap individual IS
sex-stereotypes in Arts and Science studies. The
instruments showed good psychometric properties.
Results demonstrated that: (1) More boys than girls
intended to choose the Science stream even when both had
'the same school performances '. (2) Masculinity favoured
choice towards Science, but no significant effects were
ii
'/
found for femininity and the interaction between
masculinity and femininity. The androgynous group was
not significantly different from the feminine or
masculine groups. (3) Students· stereotyping an academic
stream as suitable for their same sex favoured their
choice towards that stream~ but their stereotyped beliefs
in opposite sex had no significant effect. (4) The
effects of sex-role orientation and sex-stereotypes on
academic choice were mediated by expectancies, interest
and utility values in the academic streams.
The results suggested i) a continuing need for the
study of sex di~ferences in choices made by males and
females, ii) a future trend for studying the effects of
sex-stereotypes on behaviors ,and iii) some possible
improvements in measuring and conceptualizing sex
stereotypes. The results of this study casted some
doubts on the concept and measurement of sex-role
orientation as suggested by Bern (1981, 1985).
i ii
LIST OF TABLES Tables Page
1. Distribution of Sex for different schools and tota 1 samp le. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 24
2. Alpha Coefficients for All Scales in student quest ionnaire ................................. ~ 27
3. Factor Loadings for Achievement Motive Scales 30
4. Factor Loadings for Sex-stereotype Items for two-factor varimax solutions ............... 33
5. Means and Standard Deviations for Own-sex and Opposite-sex Suitabilities 0000 •••••• 00 •••• 0 •••• 35
6. Factor Loadings of BSRI Items for a two-factor varimax so lut ion ............ 0 • 0 ••• 0 0 ' •• 0 • • • • • • •• 38
7. Percentages of Males and Females classified as Androgynous, Feminine, Masculine and Undifferentiated by BSRI ....................... 40
8. Means and F ratios for Academic Choice, and Achievement Motives by Sex ................ ' ,' ... 48
9. Means and F ratios for Academic Choice by Sex-ro le Ty'pe ......... 0 • • • • • • • • • • • • • • • • • • • • • • •• 51
10. Standardized Beta Coefficients and R Square of Hierarchical Regression Analysis of Masculinity and Femininity on Academic Choice ... o •••••••••• 52
11. Standardized Beta Coefficients and R Square of Hierarchical Regression Analysis of Own-sex Suitability and Opposite-sex Suitability on Academi c Cho ice ................................ 55
12. Path Coefficients for the Path Model of Academi c Cho ice ................................ 57
iv
LIST OF FIGURES
Figures Page
1. Path Diagram for Academic Choice ......... 0 .... 0.46
2. Results of Path Analysis for Academic Choice of Boys .............. . ' ............. 0 ••••••• 0 • • •• 58
3. Results of Path Analysis for Academic Choice of Girls ........................................ 59
LIST OF APPENDICES
Appendices
1. Number~ Percentage and Ratio of Males and Females getting A grade in Hong Kong
Page
Certificate of Education Examination 1990 .. 0. 0 00 76
2. Number, Percentage and Ratio of Males and Females Sitting Hong Kong Certificate of Education Examination 1990 . 0 ... 0 . 0 ... 0 0 .... 0 . 0 0 0 77
3. Number, Percentage and Ratio of Males and Females Sitting Hong Kong Advanced Level Examination 1990 0 ...... 0.00 ..... 0 ... 0' 0 ......... 0 78 '
4. Ratio of Males and Females in the Full-time Enrolment in universities of Hong Kong ......... 79
5. Measures of Sex-stereotypes in Arts and Science studies .. 0 ... 0 . o . ••• 0 0 ... 0 .. 0 ............... 0 0 o. 80
6. , Measures of Achievement Motives (expectancies for success, interest and utility values) in Arts and Science, studies . 0 •••••••••••••••••• 0 •• 83
7. Measures of Sex-role Orientation (simplified, translated version of BSRI) ............... oo •• ~ 86
v
Chapter 1
INTRODUCTION
This research was initiated through an interest in
the sex differences found in academic achievement in Hong
Kong (see Appendix 1) and overseas. Several important
meta-analysis studies (Maccoby & Jacklin, 1974; Hyde,
1981; Hyde & Linn, 1988; Hyde, Fennema & Lamon, 1990;
Feingold,1988) had consistently found that there are sex
differences in verbal performance favouring girls and sex
differences in quantitive performance favouring boys, but
these patterns of sex differences emerge only from
adolescent years. Several theorists (Sherman, 1978;
Eccles, 1985; Signorella & Jamison ,1986) had proposed
that the sex differences in achievement was mainly due to
the sex differences in academic choice. Evidence of ·
consistent sex differences in academic choice could be
found both in Hong Kong ( see Appendix 2 ) and overseas.
The discrepancies in participation and achievement of
males and females 1n quantitative or verbal fields
increased as the level of education increased ( see
Appendix 3 & 4 ), and resulted in underrepresentation of
women in many career areas and men in the others
_ Many explanations had been proposed for the sex
differences found in achievement areas, including
biological and social/psychological factors. Among them,
P.1
two important gender-related variables~ sex~role
orientation and sex-stereotypes~ were included in this
study for closer examination.
Purpose of the Study
The major goal of this study was to investigate the
effects of sex-role orientation and sex-stereotype on
academic choice.
In this study~ the dependent variable was academic
choice, which referred to the choice between Arts and
Science streams made by F~rm 3 Chinese students in
Hong Kong. The independent variables in this study were
individuals/sex-role orientation and their sex
stereotypes in the Arts and Science studies. School
performances and sex were introduced as control
variables. The effects of sex-role orientation
and sex-stereotypes on academic choice were hypothesized
to be mediated by expectancies for success and percieved
task values in Arts and Science studies.
The principal research questions of this study were
(1) Do males and females differ in academic choice?
(2) Do sex-role orientation and sex-stereotypes have any
effect on academic choice ?
(3) What factors mediate the effects of sex-role
orientation and sex-stereotyes on academic choice?
P.2
Significance of the Study
Deaux (1984) suggested that differentiations in
many behavioral patterns between women and men are
results of differential choice. Despite the critical ' role
of choice in explaining sex differences in behavioral
pattern, sex differences in choices are rarely studied.
According to Deaux's view, studying sex differences in
academic choice can enhance a better understanding of the
process of how sex differences in achievement pattern is
developed. And with school performances controlled
throughout the study, the academic choices of boys and
girls could be studied as opposed to their capabilities
shown in Arts and Science subjects.
Researchers who stressed the importance of
biological factors had suggested that sex differences in
achievement areas were due to the inherent ability of
males and females . In constrast to these researches,
this study tried to highlight the significance of social/
pychological factors. It examined how academic choice
could be influenced by sex role.
According to Bern's (1985) gender schema theory ,
an individual's sex-role orientation would affect
individual's behavior in culturally sex-typed tasks.
Kagan (1964a,1964b) suggested that sex-stereotypes
would affect one's behavior and performance in sex-typed
P.3
"
achievement areas. As Ar.ts and Science studies were
culturally sex-typed, both sex-role orientation and sex
stereotype were hypothesized to have influence on
students' academic choices.
Besides enhancing a better understanding of sex
differences in academic choice, this study would also
contribute in the measurement and conceptualization
of sex-stereotypes. With reference to Deaux and Lewis
(1983), a new instrument was developed for measuring sex
stereotypes in Arts and Science studies. It had two
main characteristics. First, it adopts probabilistic
judgements (ie. instead of asking respondents whether
Science is suitable for girls, it asks, out of 10 girls,
how many of them are suitable for Science study). Second,
it adopts a two-factor model (ie. instead of
conceptualizing sex-stereotypes as bipolar measures, it
considered stereotypes in males and stereotypes in
females as two independent orthogonal dimensions) .
. The instruments showed g.ood psychometric properties,
and were found to be useful in this study.
P.4
~hapter 2
REVIEW OF SELECTED LITERATURE
Sex differences
In their pioneering review of the literature on sex
differences, Maccoby and Jacklin (1974) concluded that
in cognitive domains, there were only sex differences
favoring girls in verbal ability, and favoring boys in
quantitative and spatial abilities. A series of meta-
analyses (Hyde, 1981; Feingold,1988; Hyde & Linn, 1988;
Hyde, Fennema & Lamon, 1990; Signorella & Jamison,1986
were done to test the size of sex differences in . these
cognitive domains. On the whole they supported Maccoby
and Jacklin1s findings, but the effect sizes Cd) were
found to be small.
These findings grew out of a paradox. As suggested )
by Tittle (1986):
The paradox is that psychological research finds small, persistent gender differences in only a few areas of performance, .... , yet larger and more persistent differences are found in educational outcomes linked to occupational choice and life patterns. (p.1166·)
Here appeared two questions : first, how to settle
this paradox , and second, what were the origins of the
sex differences in cognitive abilities (even though they
P.S \
were small) . . From literature review .. a possible answer
to both questions was found.
In the recent review done by Hyde .. Fennema and Lamon
(1990) on mathematics performance .. they found that
unexpectedly there was a slight female superiority in
performance in the elementary and middle school years ..
but a moderate male superiority emerged in the high
school years (d= 0.29) and it increased in the college
years (d=0.41) .. as well as in adulthood (d=0.59). In
their report .. they highlighted that these sex differences
emerged precisely in the years when students were
permitted to select their own courses. In line with their
results .. reviews of Maccoby and Jacklin (1974) .. Feingold
(1988) .. Sherman (1980) .. Benbow and Stanley (1980),
Signorella and Jamison (1986) all found that sex
differences in cognitive abilities emerged in high school
years (or in adolescence). Therefore .. many psychologist~
( Feingold~ 1988; Kimball .. 1989; Sherman .. 1978; Eccles ..
1985; Signorella & Jamison .. 1986) shared the view that
differential course taking was an important cause for sex
differences in cognitive performances.
These sex-~tereotyped selections existed not only in
high school academic choice. Huston (1983) suggested
that sex-stereotyped selections started from early
childhood. Eccles (Eccles .. 1985; Eccles & Hoffman .. 1984)
P.6
- ~ • •• • ~ .. ". '-" . ' . ... .". .. , .. _ . • •. • • - , , • . ~ " ••• • •• ~ L- . ~ ~ -r t" .· .. _:~ . ~ ... ~ ...... . ..... .
suggested that sex differences in choice patterns not
'only exist in .high school· but mirrored in college and in
the occupational world , and these sex-typed choice
patterns might be one important cause of the persistence
of sex differences in adult earnings. And more
generally, Deaux (1984) suggested that differentiations
in many behavioral patterns between women and men were
results of differential choice. Her comment for future
research on gender was :
To fully deal with the ways in which gender is influential , one must ultimately deal with the process involved ...... A first direction is to consider more carefully the choices that men and women make, as opposed to the capabilities that they show in a particular domain .... (Deaux, 1984, p.113)
According to her view, studying sex differences in
academic choice could enhance better understanding of the
the process in which sex differences in achievement
pattern were developed. And after control for school
performances, these sex differences were studied as
opposed to the capabilities of boys and girls shown in
the achievement areas.
P.7
~ .. ,
.' . ',\ ' ....... : . . . , .. ... . . , " '\. ' ~ . , ... ' .. " ,
~, ' ... " •. . . . ' .. .... " ' . ' .' . '
... . .' .... , ... .. ;. , . ", ',: ~ ' . ~ ... ~: , . .' . :: '. .. . . . . . 1 •• '.
Sex Stereotypes
Sex stereptypes were defined as "the perceptions of
persons, objects, activities, or contepts that are based
on relatively rigid, oversimplified, and over-generalized
beliefs or assumptions regarding the characteristics of
males and females" (Beere, 1979, p.164). A review by
Myers (1988) concluded that strong gender stereotypes
really exi~ted, and, as often happens, both males and
females accepted these stereotypes.
,Cultural sex-typing of academic area
Though sex differences in performance was found to
be rather small , there was evidence that sex stereotypes
in achievement areas were pervasive. Archer and Freedman
(1989) found that students rated science subjects as
masculine and Arts subjects as feminine, and science-art
was significantly correlated with the masculine-feminine
dimension. Stein and Bailey (1973), Stein and Smithells
(1969), Huston (1983) found that adolescents considered
social, verbal, artistic skills as feminine; and
mechanical, spatial and mathematics skills as masculine.
In short, Science was sex-typed as masculine while Arts
was sex-typed as feminine in our culture.
P.s
: - '_ . .. • • • 1 ,. , • • : : ., .; • .-
- " .~ ~ .. - . . .. ..... -'~ ' ' '- • • -, . .. .. . J , " - ' ~ \.. • " ..... ' . " • J'o •• • .. .. . _. ' • • ' . _ : ' , ' ,. ;. . .. . . . ~ • •
' • • J " - " -:' ~ _" .;. • • _ ' • • :, . _ •• _ _ • • . • ~ " _ ' .. . . : .J~ .< .. __ -~ _' , . _. . , •• - ' • • . ' - " , . : ' . : , ', " . . ... ~ . _ .. . ' : '.: • " .. · ••• • i .... · ~ ... . .. ..... . . - . " ' ' '. _ . ... . . . . . ... ; .• ~ • .•• _ ~. ;, ..
Relation with Achievements
Researches relating academic choice and sex
stereotypes of the related achievemen~ areas could hardly
been found , but researches relating sex-stereotypes with
performance and attitudes in academic (or cognitive)
achievement could be found.
Kagan (1964a,1964b) proposed that cultural
sex-role stereotypes would generate individuals' sex-role
standards (ie. individual's conception of whether
certain activity was appropriate to males or females ),
and students' achievement in an academic area was in
congruence with his or her sex-role standard of that
area. With different instruments developed, Stein
(1971), Dwyer(1974), Paulsen and Johnson (1983) tried to
test her hypothesis. Dwyer (1974) found that reading
and arithmetic sex differences were a function of the
child'~ perception of these areas as sex-appropriate or '
sex-inappropriate. Stein (1971) found that children ' s
sex-role standards for different areas of achievement
were related to their attainment values, expectancies and
standards of performance in these areas. Similarly,
Paulsen and Johnso~ (1983) found that the attitude
considering mathemat-ics as not inappropriate for one's
sex , correlated positively with high mathematics
scores . Ehindero (1986) found that the interaction
P.9
· .'
between sex stereotypes on achievement in physics and
gender had a strong effect on individual achievement in
physics. Kelly (1988) found that girls who endorsed sex
stereotypes both in general behaviors and occupations _
and in specific science domains _ did less well and
were less likely to choose science, especially physics.
However, most of the above findings were not strong.
Some other evidence found was even inconsistent. In
contrary to Kagan's hypothesis, the stereotyping of
mathematics as a male domain was found to be positively
related to both female achievement and course taking
(Fennema & Sherman_1977; Sherman & Fennema_1977;
Sherman, 1980; Meyer .. 1985) . Stereotyping mathematical
ability was slightly positively related to mathematics
achievement for girls (Kaczala, 1983). None of these
studies could found correlations for boys. Some other
researches (Handley & Morse, 1984; Smend & Chase, 1981)
also gave inconsistent findings .
Among the above researches.. those using same or
similar kind of instruments tend to give similar results.
Ther;efore.. one might suspect that the design of
instrument for , measur~ng sex-stereotypes might
considerably affect the results obtained.
P .10
Measures of Sex Stereotypes
From literature review, very few instruments were
found to assess sex stereotypes in achievement-related
areas. Only five were found. First, Stein & Smithells
(1969) and Stein (1971) used the "Sex Role Standard
Questionnaire" to assess sex stereotypes . on several
achievement-related activities. Subjects were asked
whether each activity was a more masculine or a more
feminine thing to do. Later, two improved modified
version of Sex Role Standard Questionnaire was
constructed by Dwyer and Paulsen and Johnson. Dwyer
(1974) asked subjects which interests and activities they
thought boys or girls preferred. Paulsen and Johnson
(1983) asked subjects to what degree they felt an
activity was inappropriate for a particular sex. The
fourth instrument found was the "math as a male domain"
scale in Fennema-Sherman Mathematics Attitudes Scales,
adopted in Sherman and Fennema (1977), Fennema and
Sherman (1977), Sherman (1980), and Meyer (1985). In
this test, items like 11 studying mathematics was just as
appropriate for women as for men" , were being rated on a
Likert scale. More recently, Eccles (Eccles et al.,1983;
Kaczala,1983) had developed another instrument for
measuring stereotypes on the utility of math. Subjects
were asked to respond to items like JI how useful do you
P.11
. . . " .- .. . ' --. - .. . •... . , --_. - - .. .. - . • .. - . . ~ i·. - ,' .. . ' - ' ",,;. , . MO _ .. - . :-- ~. - , • ,. • • __ , , • • ~ . L - ~- .. . -' ', ... _ • • _ - _ •• , • • • • _ _ _ _ . _ ....... .. _ . . . . . .... . .. ... .. .. .. . ... , • • I _'~'" .. ~., ,,\ •• • _ .... .... . . Jo: • • '
think women find basic math in their future" jobs?".
For most.of these instruments, evidence for their
reliability and validity was not found. As commented by
Beere (1979) and Ruble & Ruble (1982), past researthes
for sex stereotypes had suffered from a number of
methodological difficulties and the instruments for
assessing sex stereotypes in achievement areas were not
well-developed.
At least three problems were speculated in these
instruments . The first problem was that the instruments
might be too transparent. Respondents were likely to
give socially desirable answers rather than responding
from · their honestly felt stereotypes. As in their own
comment -, Sherman and Fennema (1977) suggested that girls
studied were living in community where feminist movement
had received much publicity, so they would disagree the
stereotypes verbally, but in fact th~ masculine view of
mathematics had already communicated to the girls ,
resulting in the girls's actions contradicting their
words.
The second problem was that all the above
instruments tapped sex stereotypes in an absolute rather
than relative manner. As a result, they produced a list
'of discrete activities, . interests or traits that were
believed to be absolutely characteristic or not
P .12
characteristic of boys or girls. However, as suggested
by Deaux (1984) and Deaux and Lewis (1983), sex
sterotypes usually were relative rather than absolute
assignment of characteristics to men or women. In -many
cases, subjects believed that there was reasonable
probability that women had the same qualiti~s as men.
Deaux and Lewis (1983) had successfully developed a new
instrument in ·which probabilistic judgements were
adopted. In this test, subjects were asked to estimate
the probability that an average man or woman possesses a
particular characteristic. Unfortunately, her instrument
did not measures sex-stereotypes in achievement areas.
The final but most important problem was that all
the above instruments ( except the one constructed by
Paulsen & Johnson,1983 ) conceptualized sex stereotypes
in bipolar terms. That meant, an achievement area was
considered as either stereotyped for male or female.
Its suitability for boys and that for girls were
considered as mutually exclusive constructs, representing
the two endpoints of a single, bipolar continuum. Any
sex-stereotype scores could only stand between these two
extremes, and represented the relative suitability of the
. achievement area between boys and girls.
However, in this way, measures in between the two
extremes were not well-defined, it might represent
P.13
suitable for both sexes or suitable for neither sexes.
And no evidence was found to support their preassumption
that the suitabilities of an achievem,ent area for girls
and for boys were exclusively related . The relation
between individual's stereotypes in males and in females
was unclear. Logically, an individual having cultural
stereotypes in one sex might not necessarily implied that
he/she would have cultural stereotypes in another sex.
Further research need to done to find out which
stereotypes (ie. stereotypes in same sex or that in
opposite sex; or both) would affect individuals'
behaviors.
Sex-role Orientation
A person's sex-role orientation was described by
terms: masculine or feminine. Though commonly used,
masculinity and femininity were rarely defined clearly
\(Spence .. 1985). According to the theory proposed by Bern ..
masculinity and femininity referred to (either the
conscious or unconscious) perception of culturally
defined maleness or femaleness of the self. In essence,
sex role orientation was one aspect of self-concept.
P.14
l,;//
Bem's Gender Schema Theory
According to Bem's gender schema theory (Bem~ 1981~
1985)~ our society had its cultural definition of
femaleness and malesness which in fact comprised a
diverse network of sex-linked associations. A developing
child would learn these sex-linked associations and build
up an evolving gender schema. Because of the salient
role of gender assigned by our culture~ an individual
developed a generalized readiness to process information
on the basis of gender schema. Bern claimed that there
was considerable evidence supporting the existence of
gender schemas (Bem~ 1981~ 1985; Liben & Signorellla~
1987) ' .
In Bem's theory~ schematic individuals would
assimilate their self-concept into the gender schema.
They learned to evaluate their adequacy as a person
according to the gender schema. Internalized motivation
was then created to prompt them to regulate their
behavior so that it conforms to cultural definitions of
femaleness and malesness. Bern and Lenney (1976) , Bem
(1985), Helmreich, Spence and Holahan (1979) did find
that gender schematic individuals were more likely to
adopt gender-appropriate and to avoid gender
iriappropriate behaviors.
To assess individual differenc~s in gender
P.15
\,./
schematicity, Bem' relied on people's scores on the Bem
Sex Role Inventory (BSRI; Bem,1974). Forty culturally
defined desirable personality traits for male and female
were selected to be items in BSRI. Subjects were asked
to rate how much they agree that the traits describe
themselves. Traits for male and traits for female
contribute to two orthogonal scales -- femininity (F) and
masculinity (M) scales. Basing on a median split on M
andF scales, masculine, feminine, androgyuous and·
undifferentiated groups were obtained. Sex-typed persons
(feminine girls and masculine boys ) were proposed to be
gender schematic, while androgynous persons to be gender
aschematic.
According to Bern's theory, sex-typed individuals
would adopt more culturally defined sex-appropriate
behaviors , while androgynous individuals would adopt
less. Therefore. if Arts and Science studies were
culturally stereotyped tasks, then feminine girls and
masculine boys were more likely than their same sex
androgynous peers to choose academic subjects in the
culturally stereotyped directions.
In Bern's conception, androgyny referred to gender
aschematicity and implied a balance between maculinity
artd femininity. In her model, M and F were postulated
to have an interactive effect and androgyny to have
P.16
unique predicability. However, Taylor and Hall (1982)
and Hall and Taylor (1985) had suggested that M and F may
have only main effects but no interactive effects. They
suggested that both models (ie. main effect and
interactive models) could be test simultaneously by 2 X 2
analysis of variance, or by a . multiplicative M X F term
in regression analysis. These statistical methods were
found to be useful in reviewing androgyny research
findings (Hall & Taylor; 1985) .
Relation with Achievement
Consistent with the above sex-role theories, Nash
(1979) concluded from a review of literature that
individuals did perform better on cognitive tasks when
the masculinity and femininity in their self-concepts
was consistent with the gender stereotype of the tasks.
In her own research, Nash (1974) found out that the more
masculine a boy was, the 'better his arithmetic score;
and the more feminine a girl was, the better her reading
score. These masculinity and femininity scores were
measured by asking respondents to rate themselves on
bipolar sex-typed attributes.
Later, a meta-analysis done by Signorella and
-Jamison (1986) on 73 researches also concluded simi.lar
results. For both boys and girls, higher masculinity and
P.17
lower femininity scores were associated with better
performance in spatial and mathematics tasks, but less
consistent findings for verbal tasks and androgyny
individuals. A variety of instruments were used in the
73 studies, all measures were objective tests in which
subjects were asked to describe themselves using sex
stereotyped characteristics.
Though bulk of researches had "been done to
investigate the relationship between sex~role orientation
and cognitive achievement, only very few concerned about
academic choice were found. Eccles et.al. (1983) found
that for girls, masculinity score was significantly
correlated with the intention to take more mathematics
course, but femininity and M X F interaction had no
significant effects, and no evidence for boys. Kelly
(1988) found that girls who saw themselves as masculine
were slightly more likely than other girls to choose
physical science, while girls who saw themselves as
feminine were slightly more likely to choose biology.
Again no evidence for boys. Comparable findings for
feminine academic tasks , such as Arts subjects , had not
been reported. On the whole, past researches had found
little consistent findings for the relation between sex
role orientation and academic choice.
P .18
. Expectancy-value Model
. If the relationships of sex-role orientation and
sex-stereotypes with choice in sex-typed tasks were
proved to exist , then the next question concerned was
how could we explain the effect of sex-role orientation
and sex-stereotypes on the choice , that meant, what
mediated their effects on the choice.
A comprehensive expectancy-value model was developed
by Eccles to explain sex differences in achievement
behavior (Meece, Eccles, Kaczala, Goff & Futterman ,
1982). According to her model_ expectancy for success
and percieved task value were hypothesized to be
mediators of the effect of sex role orientation and sex
stereotypes on academic choice .
Expectancy for success
- Despite the findings that actual sex differences in
mathematics and verbal achievement were mininmal ,
the sex differences in academic self-concepts persisted.
Researches had found that boys had higher mathematic
self-concepts, whereas girls had higher verbal/reading
self-concepts (Meece et.al.,1982; Marsh, 1989a,1989b;
Marsh et.al.,1988,1985).
In our society, there were many culturally defined
stereotypes , for example, females were less competent
P.19
than males; fem~les were culturally stereotyped to have
higher verbal ability and males to have higher
mathematical and mechanical abilities; high-level
professions ,science-related vocational fields and
science course were males activities ; and teaching,
clerical work. language-related vocational fields and
courses were females activities (Nicholls, 1975 ; Eccles
et al.,1976; Eccles,1987; Eccles & Hoffman, 1984; Huston,
1983 ). According to Bem's theory, the sex-typed
individuals would incorporate these stereotypes into
one's self-concept, and, hence, would have lower
expectations in sex-inappropriate activities and higher
expectations in sex-appropriate activies.
Moreover, Deaux (1984) explained the sex differences
in expectancy for success in terms of sex differences in
the causal attribution patterns. Supporting Deaux's
view, Eccles et al. (1983,1984) found that girls rank
. effort and boys rank ability as more important causes of
their mathematics success. As ability was a stable
internal attribution, boys were likely to have higher
expectancy in mathematics.
achievement behaviour. Eccles suggested that the
perceiv~d task value had three major components : 1) the
utility value for future goals; 2) the incentive ( or
interest value) of engaging in the task ; and (3) the
cost of engaging in the activity ( Eccles et al.,1983;
Eccles,1987) .
Eccles (1985) suggested that perceived values of a
task were expected to be related to one's personal values
and life goals. Sex-role orientation and sex
stereotypes would affect the range of careers one would
consider appropriate~ and also affect the importance one
would given to the parenting or family role, instead of
the career role. Baker (1987) had found that females
prefering science-related careers or nontraditional
careers such as police, military and trades were more
masculine, while those preferring traditional careers
such as teacher or hairdresser were more feminine. And
of course, persons with different career preference would
place different value ·on the Arts and Science courses.
Moreover, Feather (1984, 1988) found that perceived
value of Mathematics and English was related to
individuals' personal values, which in turns, were
correlated with mas~ulinity and femininity .Therefore~
in support of Eccles' view, sex role might influence
one's · perceived task values through its impact on
personal values. P.21
Study Overview
Chapter 3
METHOD
The major goal of this study was to investigate the
effects of sex-role orientation and sex-stereotype on
academic choice. In this study, the dependent variable
was academic choice, the predicting variables were sex
role orientation and sex-stereotypes on the Arts and
Science studies. School performance and sex were
introduced as control variables, while the achievement
motives (ie. expectancies for success, interest and
utility values) of Arts and Science studies were
introduced as mediating varibles of the effects of sex
role orientation and sex-stereotypes.
Subjects
Subjects included 818 Form 3 Chinese students from 5
selected schools in Hong Kong. The schools chosen were
coeducational, government subsidized secondary schools
with only Arts and Science streams in Form 4 & 5. These
schools had adopted the usual method in allocating
students to Arts/Science streams. That meant , they
would allow their students to make their choice before
the final examination in Form 3. And after the
examination, the students were allocated first according
P.22
to their choice ~ but when places were not enough~
students were allocated according to their performance in
Arts and Science subjects examination. Moreover~ these
schools were not those with especially high or low
academic standard. Table 1 summarized the number of
males and females sampled at each school.
Procedure
After reviewing related literature~ instruments for
meausuring academic choice, expectancies for success ,
percieved task values, sex-stereotypes on Arts/Science
studies were developed in this study. On the other hand,
instrument for measuring sex-role orientation was adopted
from Bern's Sex Role Inventory (BSRI). Four interviews and
two pilot studies were conducted to revised all
instruments. The whole questionnaire was shown in
Appendix 5, 6 & 7.
The questionnaire was done at least two months
before students made their final academic choice to
schools 1 • In order to avoid students adjusting their I
stereotypes to fit their choice, questions for measuring
sex-stereotypes were arranged before other questions in
Note 1. This arrangement was trying to avoid the influence on
the answers by a need to be consistent with their formal choice, which might be a decision made under school pressure or actually just their parents' decision instead of their own choice.
P.23
TABLE 1
Distribution of Sex for different schools and total sample
Male Female Total
School 1 92 95 187
School 2 82 83 165
School 3 61 72 133
School 4 99 87 186
School 5 55 92 147
Total 389 429 818
P.24
the questionnaire.
The procedures for questionnaire administration were
straight forward. The whole set of questionnaires were
sent to the 5 schools. Teachers were instructed to let
students complete the questionnaire in class. Copies
of mid-term examination results of all Form 3 students
were collected from each school.
Measures
Academic Choice
The academic choice measured in this study was '
the intended choice for arts and Science streams of Form
3 students. They were asked if they were free to choose,
which 'stream they would prefer.
Four questions were designed for this variable
(see Appendix 6_ question no. 13 to 16). Two tapped the
preference for Arts, while the other two for Science.
Because most students not yet had a definite academic
choice, 5-point bipolar rating scales were adopted
Only one final score was obtained in this test,
namely the academic choice. In real situation, students
could either choose Arts or Science, but not both. So
preference for Arts and Science were not analysed
separately in this study. The score of academic choice
was computed by subtracting each score for Arts choice
P.25
from their corresponding score for Science choice and
then summing up the differences with equal weights. A
high academic choice score represented a preference
for Science while a low score represented a preference
for Arts. Internal consistency of the scale was found to
be very high ( see Table 2).
Achievement Motives
Expectancies for success. The variables measured
here were the expectancies for success in future
Arts/Science studies. Each measure consisted of 2 items
asking students to rate how well they expected to do in
future Arts/Science stUdies, and what result they
expected to get in future external examination HKCEE (see
Appendix 6, question no. 1, 2, 7 & 8). Each item
involved a 5-point rating scale.
Two scores expectancies for success in Arts and ·
Science studies were obtained here. Each was obtaine~
by summing up the self-ratings on the correspOnding items
with equal weights. A high score represented a high
expectancy for success. Internal consistencies of the two
scales were found to be quite high (see Table 2).
Perceived task values. The variables measured here
were the perceived task values in Arts and Science
P.26
TABLE 2
Alpha Coefficients for All Scales in student questionnaire
Cronbach's alpha No.of
Scales Whole ·items sample boys girls
Sex-role Dimensions (BSRI) Masculinity .87 .87 .88 17 Femininity .84 .85 .80 17
Sex-stereotypes Own-sex Suitability .73 .78 .88 6 Opposite-sex Suitability .72 .76 .87 6
Achievement Motives
Expectancy for Science .86 2 Expec-tancy for Art .83 2
Interest for Science .84 2 Interest for Art .83 2
Ut i 1 i ty Va·l ue for Science · .69 2
Utility Value for Art .65 2
Academic Choice .93 2
P.27
studies. In Eccles J (1985,1987) conceptual analysis of
task value, it was suggested that task value consisted of
three major components: intrinsic or interest value,
utility value and cost. In this study, only the interest
values and utility values were involved.
With reference to the instruments used by Eccles et.
al. (1983), Meece, Wigfield and Eccles (1990),Kaczala
(1983) and Feather (1988), the scales for the interest
values and the utility values in Arts and Science studies
were constructed. Each measure consisted of 2 items in
a 5-point rating scale. Two items asked students to rate
how important Arts/Science study was for their future
occupational career; while the other two items asked them
to rate how interested they were in Arts/Science course
(see Appendix 6, question no. 3, 4, 9 & 10 for interest
values; and question no. 5,6, 11 & 12 for utility
values)
Four stores -- interest values and utility values
for Arts and Science studies -- were yielded here. Each
was obtained by summing up the self-ratings on their two
corresponding items with equal weights. A high score
represented a high percieved task value. Internal
consistencies of all interest scales was found to be
high, but that of utility scales was only moderate (see
Table 2).
P.28
All scales for measuring student achievement motives
towards Arts and Science ,studies (ie. scales for
expectancies for success, interest and utility value)
were factor analyzed to test the existence of one or more
overriding constructs guiding these attitudes. Three
successive factor analyses were done: one on all items,
one on Science items only, and another one on arts items
only. Varimaxsolutions were found and shown in Table 3 .
They were consistent with the theoretical basis for the
scale construction.
Sex-stereotypes
This study measures the individuals' sex stereotypes
of Form 3 students in Arts and Science studies . '
With reference to the instrument developed by Deaux
and Lewis (1983), a new instrument for measuring these
stereotypes was developed in this study. Measures here
consisted of sex stereotypes in three different domains
(ie. the utility values, interest values, and
expectancies for success in Art and Science studies).
The test asked students to rate how many boys/girls could
study well, get a good job or to have interest in Arts
and Science studies. , There were all together 24 items, 6
for stereotypes in Arts for boys; 6 for stereotypes in
Science for boys; 6 for stereotypes in Arts for girls and
P.29
TAB
LE
3
Facto
r L
oad
inq
s o
f A
chie
vem
ent
Mo
tiv
e Scale~
Two
facto
r "s
olu
tio
n
Th
ree
facto
r so
luti
on
T
hre
e fa
cto
r so
luti
on
fo
r all
it
em
s fo
r S
cie
nce
item
s fo
r A
rts
item
s It
ems
Scie
nce
·Art
s E
xp
ecta
ncy
In
tere
st
Uti
lity
v
alu
e
Ex
pec
tan
cy
Inte
rest
Uti
lity
v
alu
e
facto
r fa
cto
r fa
cto
r fa
cto
r fa
cto
r fa
cto
r fa
cto
r fa
cto
r
1 -.
02
.7
8
.86
.2
9
.15
2
.75
·-
.08
.8
7
.20
.1
9
~
3 -.
26
.7
7
.32
.8
3
.21
w
4
.79
-.
22
.4
2
.80
.1
6
0 5
. -.
09
.6
6
.12
.2
7
.82
6
.66
-.
02
.3
2
.08
~
7 -.
02
~
.78
.2
8
.33
8
.79
-.
10
.7
8
.37
.1
9
9 -.
21
.7
4
.24
.8
8
.18
10
.7
6
-.2
1
.20
.8
8
.25
11
-.
00
.6
6
.28
.0
9 .
.8
2
12
.66
.0
3
.06
.3
0
.82
No
te.
The
u
nd
erl
ined
lo
ad
inq
s in
dic
ate
th
e
co
nst
ruct
to
wh
ich
th
e
scale
co
ntr
ibu
ted
to
.
6 for stereotypes in Science for girls. Simplifying the
scale used by Deaux, a ll-point rating scale was
developed (see Appendix 5 ).
The measures developed here had two - characteristics.
First was the use of probabi 1 isti-c judgments in the
assessment of stereotypes (ie. instead of asking students
whether boys like studying Arts, - they were asked to
rate, out of 10 boys, how many of them like studying
Arts). This viewpoint was adopted because sex
stereotypes were believed to be relative rather absolute
assignement of characteristics to males and females
(Deaux, 1984) .
The second characteristic was adopting the two
factor model of sex-stereotypes. That means, instead of
conceptualizing sex-stereotypes as bipolar measure,
stereotypes in own sex and opposite sex are considered as
two orthogonal independent dimensions. Here the
suitability of an academic study for same sex and that
for opposite sex were measured separately in the -
instrument, and they were analysed as independent
variables throughout the study.
Two scores were obtained from this test. They were
own-sex suitability (ie. the suitability of Science for
own sex) and opposite-sex suitability (ie. the
sUItability of Science for opposite sex). The former one
P.31
was obtained by su~tracting the scores of Arts items for
their own sex from their corresponding Science items, and
then taking the mean of these differences. A high score
indicated that the student perceived Science as more
suitable than Arts for their own sex. The score for
opposite-sex suitability was obtained similarly from
items for opposite sex. A high score indicated ' that
student perceived Science as more suitable than Arts for
their opposite sex .
A student was considered as highly sex-stereotyped
if he/she believed Science was for males and Arts was for
females. That means, a girl was considered as highly sex-
stereotyped if she perceived high opposite-sex
suitability and low own-sex suitability. A boy was
con~idered as highly sex-stereotyped if he perceived high
own-sex suitability and low opposite-sex suitability.
Internal consistencies of the two sex-stereotype
scales was found to be quite high (see Table 2) . In
factor analysis, score of each Arts item was subtracted
from its corresponding Science item to obtain 12
subscales, and then two-factor varimax solution was
computed ' on these 12 subscales. The solutions were shown
in Table 4. Own-sex suitability and opposite-sex
suitabilities emerged as the two constructs guiding
respondents' judgements. All subscales load heavily on
P.32
, . .. ' , .,\,. t ' . "\·
- ~ ---": "::~
~
w
w
(Scie
nce
item
s -
Art
it
em
s)
1 -
14
3
-1
6
5 -
18
6
-2
0
9 -
22
1
0 -2
4
2 -
13
4
-1
5
6 -
17
8
-1
9
10
-
21
11
-
23
Tab
le
4
Facto
r L
oad
ing
s o
f S
ex
-ste
reo
typ
e
Item
s fo
r tw
o-f
acto
r v
ari
max
so
luti
on
s
Bo
ys
I _ re
sp
on
se
Gir
ls'
Op
po
sit
e-s
ex
O
wn
-sex
O
wn
-sex
S
uit
ab
ilit
y
Su
itab
ilit
y
Su
itab
ilit
y
Facto
r F
acto
r F
acto
r
-.0
6
.64
-.
00
-.
15
.7
1
-.1
4
. 11
.6
6
.06
-.
10
.J
1
.01
-.
17
.4
9
-.1
0
.12
.7
0
.10
.6
4
-.1
5
~6
-.!.-71
. -.
11
.7
2
.71
-.
07
.6
9
.61
.0
7
.73
.5
7
-.1
0
.51
.7
1
.10
.6
9
resp
on
se
Op
po
sit
e-s
ex
S
uit
ab
ilit
y
Facto
r
.66
.7
3
.65
.7
0
-.!.M
.7
0
-.1
0
-.1
0
.10
-.
01
-.
09
.1
6
No
te.
Th
e u
nd
erl
ined
lo
ad
ing
s
ind
icate
·th
e
co
nstr
uct
to
wh
ich
th
e
scale
co
ntr
ibu
ted
to
.
the constructs which they were designed to measure.
The mean and standard deviations of all the
subscales were shown in Table 5. In boys 'response, the
mean scores of all own-sex suitability subscales were
positive, while the mean scores of all opposite-sex '
suitability subscales were negative. Exactly opposite
results were obtained in girls' ' response. That meant ,no
matter in terms of expectancies, interest, or utility
value, both boys and girls thought that Science was more
suitable for boys than Arts ; while Arts was more
suitable for girls than Science. These showed all
subscales were stereotyped in a direction which was
consistent with our cultural sex-stereotypes. The sex
stereotypes on Arts and Science were found to be quite
pervasive. ' These results did lend partial support to the
validity of the instrument.
Sex Role Orientation
In this study, Bem Sex Role Inventory (BSRI) was
chosen to measure sex role orientation because the
theoretical framework of this study was based on Bern's
gender schema theory and BSRI was both widely used in
overseas and in Hong Kong . The instrument had been
translated into Chinese and gave satisfactory reliability '
and validity results in previous studies with (Tse, 1988;
P.34
TABL
E 5
Mea
ns
an
d
Sta
nd
ard
D
ev
iati
on
s
for
Ow
n-s
ex
an
d O
pp
osit
e-s
ex
S
uit
ab
ilit
ies
Gir
ls'
resp
on
se
Bo
ys'
re
sp
on
se
;
Item
s (S
cie
nce
-A
rt)
Fo
r O
wn
-sex
F
or
Op
po
sit
e-
Fo
r O
wn
-sex
F
or
Op
po
sit
e-
sex
se
x
1.
Lik
e t~
e le
sso
ns
M
-1.3
3
.5
3.3
-1
.7
SD
(2.4
) (2
. 1
) (2
.4)
(2.5
)
~
2.
Stu
dy
w
ell
M
-2
.2
1.6
1
.3
-2.0
w
SD
(2
.0)
(2 .
1)
(2.3
) (2
.2)
U1
3.
Get
a su
itab
le
M
-1.7
.8
1
.0
-1.8
jo
b
SD
(2.4
) (2
.1)
(2.3
) (2
.4)
4.
Feel
the
su
bje
ct
M
-. 7
,3
.9
3.6
-
. 9
inte
resti
ng
SD
(2
.8)
(2.3
) (2
.5)
(4.1
)
5.
Get
go
od
resu
lt
M
-1.4
1
. 1
1.0
-1
.3
in
H.K
.E.E
. SD
(1
.8
) (1
.9
) (2
.1)
(2.2
)
6.
Hav
e a
go
od
M
-
' .8
1
.2
1.2
~1.2
pro
sp
ect
SD
(2.0
).
(2.0
) (2
. 2
) (2
.3)
Ov
era
ll
Av
era
ge
M
-1.4
2
.1
2.0
-1
.5
SD
.( 1
. 5 )
(1.4
) (1
.5)
(1 .
7)
Wong, 1987; Lau, 1989). Internal consistency of the
Chinese version of BSRI was high (Cronbach's alphas for
the masculinity and femininity scales were .80 and .70
repectively in Lau & Wong, in press; and .87 and .73
respectively in Lau, 1989). Validity evidence was also
available. Keyes (1980,1984) and Lau & Wong (in press)
had asked Chinese youths to freely produce trait items
that described masculine and feminine qualities, and
Cheung (1986) had asked students to rate typical male and
female on an adjective checklist. The responses obtained
in all three studies were similar to those in the BSRI.
In this instrument, only the masculine items and the
feminine items in BSRI were used. Among them, several ·
items were deleted 2 to lower the total number of items
from 40 to 34, leaving only 17 masculine and 17 feminine
items. And some difficult items were rewritten in a
simpler way for better understanding of Form 3 students.
Similar to BSRI, a 7-point scale was adopted.
The test yields two scales, masculinity and
femininity, which were referred to as "sex-role
dimensions" in later parts of this study. The
masculinity score was the mean of the self-ratings on the
Note. 2. Six items were deleted because the item analysis of
Lau's study (1989) found that the reliability of the instrument would be considerably raised if each of the items was deleted.
P.36
17 masculine items; the femininity score was the mean of
the self-ratingS ' on the 17 feminine items. The possible
range of scores of each 'item was 1 to 7.
Sample medians for the femininity· and masculinity
scales were 3.25 and 3.01 respectively. Students whQ
score above the median of both the masculinity and
femininity scales were classified as androgynous;
students who score below the median of both scales were
classified as undifferentiated; students who score above
the median of only femininity were classified as
feminine; and students who score above the median of only
masculinity were classified as masculine. The four
groups -- androgynous, masculine, feminine and
undifferentiated -- were referred to as "sex-role types"
in later parts of this study. Results of this study were
analysed both in terms of sex-role dimensions and sex
role types.
Internal consistencies of the femininity and
masculinity scales were found to be high (see Table 2).
Factor analysis was also computed. Factor loadings of a
two-f~ctor varimax solution was shown in Table 6 .
Macul ini ty and, feminini ty factors emerged in the
sblution. All, except one of the items (ie.flatterable)~
load heavily on the ' factor which they were designed to
.measure.
P.37
TABLE 6
Factor Loadings of BSRI Items for a two-factor varimax solution
Masculine factor
1. ambitious 2. gentle 3. aggressive 4. affectionate 5. analytical 6. childlike 7. compassionate 8. defends own beliefs 9. eager to soothe hurt feelings 10.has leadership abilities 11.independent , self-reliant 12.love children 13.makes decisions easily 14.sensitive to the needs of o~rs i5.strong personality 16.willing to take a stand 17.soft spoken
18.willing to take risks 19.sympathetic 20.competitive 21.tender 22.understanding 23.dominant 24.self-sufficient 25.warm 26.yielding 27. individualistic 28.gullible 29.acts as a leader 30.flatterable 31.assertive 32. loyal 33. forceful 34.shy
.60
.10
.44
.19
.56 -.05
.20
.48
.24
.62
.58
.05
.58
.39
.69
.65
.06
.58
.22
.54
.15
.30
.46
.45
.36
.10
.49 -.03
.66
.34
.65
.25
.72 -.12
Feminine factor
.11
.61
.34
.60
.24
.51
.55
.19
.60
.21
.06
.51
.07
.52
.06
.13
.63
.01 . . 67 .04 .65 .57 .10 .14) .57 -.53
-.06 · .33 .15 .20 .13 .49 .10 .38
---------------------------------------------------------: Note. The underlined loadings indicate the construct to
which the scale contributed to.
P.38
And Table 7 showed the number of males and females
classified as androgynous~ masculinity, femininity and
undifferentiated. It showed that more boys than girls
were classified as masculine, and more girls than boys
were classified as feminine by the instrument. And for
both males and females, number of cross-sex typed person
was lowest among the four groups. These results lent
some support to the validity of BSRI.
School Performances
A copy of the mid-term examination results of all
Form 3 students were obtained from the 5 schools. First,
two scores were computed for each student, namely, Arts
performance and Science performance. Arts performance
score was the standardized Z-score of the sum of the
scores forEnglish~ Chinese, geography, history, Chinese
history and economics and publ ic affairs, . wi th Engl ish
and Chinese carrying double weightings 3 . Science
performance score was the standardized Z-score of the sum
of the scores for mathematics, physicsI' chemistry and
biology, with mathematics carrying double weightings 3 .
The standardized Z-scores of each stUdent were computed
within their own school 4
Note. 3. It is because all th~ five schools would put heavier
weightings on English, Chinese and mathemati~s when computing students I marks for stream allocatlon.
P.39
TABLE 7
Percentages of Males and Females Classified as Androgynous (A), Feminine (F), Masculine (M) and
Undifferentiated (V) by the Bem Sex Role Inv~ntory (BSRI)
BSRI Categorization
Sex A M F u Total
Male 30.6% 23.6% 10.696 35.1% 100%
Female 31.4% 16.9% 23.6% 28.1% 100%
Total 31.0% 20.1% 17 . 596 31.4% 100%
P.40 '
Two more scores were computed from the examination
results~ namely~ total performance and performance
difference. The total performance score of each student
was the sum of the standardized Arts performance score
and the standardized Science performance score. The
performance difference score of each student was obtained
by subtracting the standardized Arts performance score
from the standardized Science performance score. Overall
performance and performance difference scores were used
as regressors in regression analyses~ instead of Arts
performance and Science performances ~ .
Note. 4. The standardized Z-scores of each student were
computed within their own school because students were going to compete with their own schoolmates for Arts or Science places in Form 4.
5. It was because the total variance accounted for by the two sets of performance variables would exactly be the same ~ but the correlation between Arts and Science performances (r=.76) were much higher than that between overall performance and performance difference (r=-.02). It was better to choose independent variables as rearessors to avoid the problem of mult-icollinearity in regression analyses.
P.41
Research Hypotheses
In light of the research reviewed in the last
chapter, the following research hypotheses were offered
for study:
Hypothesis 1. Males and females would differ in their
academic choices and achievement motives toward Arts and
Science studies (ie. expectancies for success, interest
and utility values), even after school performances were
controlled.
Hypothesis ~ A student's sex-role orientation would be
related to his/her academic choice~, even after school
performances were controlled.
a) Sex-role type :
i) Masculine boys were more likely than androgynous boys to choose Science;
ii) Feminine girls were less likely than androgynous girls to choose Science.
b) Sex-role dimensions :
i) Masculinity would have positive effect on choosing Science;
ii) Femininity would have negative effect on choosing Science;
iii) Maculininity and femininity would have an interactive effect.
Hypothesis ~ Student's sex-stereotype in Arts and
Science studies would be related to his/her academic
P.42
choice, even after school performances were controlled.
i) Students who perc,e i ved higher own-sex sui tabi 1 i ty in Science would have higher preference for Science.
ii) Students who perceived higher opposite-sex suitability in Science would have lower preference for Scierice .
Hypothesis ~ The effects of sex-role orientation and
sex-stereotype on academic choice were mediated by three
achievement motives ~- expectancy for success, interest
and utility value.
Data Analyses
Hypothesis 1. Sex differences were tested with one-way
multivariate and univariate analyses of variance. Sex
was the independent variable, while academic choice,
expectancies. interests and utility values for Arts and
Science were the dependent variables in the analyses.
Arts and Science school performances were controlled as
covariates.
Hypothesis ~ a . To test the possible effects of sex-role
type, students were classified into 4 groups of sex-role
·personality type by using the median split method
P.43
outlined by Bem (1974). Two-way analysis of variance were
then computed on academic choice with sex-role type and
sex as independent variables. Arts and Science school
. performances were controlled as covariates. Scheffe1s
multiple comparisons of group means were done separately
for each sex.
Hypothesis 2 b . To test for the effects of sex-role
dimensions~ hierarchical regression analyses were
computed on academic choice for each · sex. Predictors
were entered in this order --- step 1. overall
performance and performance difference between Arts and
Science; step 2. masculinity and femininity;
step "3. cross-product of masculinity and femininity 6
Hypothesis 3 . To test the effect of sex-stereotypes~
hierarchical regression analyses were computed on
academic choice for each sex 7 • Predictors were entered
: in this order --- step 1. overall performance and
performance difference; step 2. own-sex suitability and
opposite-sex suitability.
Note. 6. M X F is included in the regression equation because
its effect is postulated by Bem1s theory. Other interaction terms are not included because their effects are not postulated in this study.
P.44
Hypothesis 4 . The mediating role of the achievement
motives was tested by computing path analysis based on
the causal model , depict'ed in Figure 1. Analyses were
done on separate sex 7 •
This model was an example of recursive model where
the causal links were strictly hierarchical. School
performances, sex-role dimensions and sex-stereotype
variables were hypothesized as causes of the achievement
motives which, in turn, were causes of academic choice.
This model did not consider the possibility of causal
paths in the reverse direction, or any other
non-mentioned the causal paths. Path coefficients of each
causal path were computed by multiple regression
analysis. The path coefficients ,were the regression
coefficents obtained ' when all the variables hypothesized
as causes were included in regression equation.
Note.
7. The regressions analysis for hypotheses 3 and 4 were done on separate sex. Sex was not included in the regression equations as control variables because prelimary analysis had found that the correlation of sex with own-sex suitability (r=.75, p<.OOl) and opposite-sex suitability (r=-.74, p<.OOl) were very high; and the beta coefficients of these sexstereotype variables increase when sex was entered into the regression equation. According to theory described by Cohen & Cohen (1983), cooperative suppress~on had occurred between sex and these sexstereotype variables. The results obtained in regression analysis which include both Sex and these sex-stereotype variables as predictors might be highly misleading.
P.45
Fig
ure
1
Path
D
iag
ram
fo
r A
cad
emic
C
ho
ice
3
Ex
og
en
ou
s V
ari
ab
les
Ach
iev
em
en
t M
oti
ves
1 .
Ov
era
ll
Perf
orm
an
ce
2 1
. E
xp
ecta
ncy
fo
r S
cie
nce
1 ~
2.
Perf
orm
an
ce D
iffe
ren
ce
> 2
. E
xp
ecta
ncy
fo
r A
rt
>1
Acad
em
ic
~
3.
Mascu
lin
ity
3
. In
tere
st
for
Scie
nce
Ch
oic
e
~
4.
Fem
inin
ity
4
. In
tere
st
for
Art
0
).
5.
Ow
n-s
ex S
uit
ab
ilit
y
5.
Uti
lity
V
alu
e
for
Scie
nce
6.
Op
po
sit
e-s
ex
S
uit
ab
ilit
y
6.
Uti
lity
V
alu
e
for
Art
No
te.
Th
e p
ath
d
iag
ram
is
a
sim
pli
fied
v
ers
ion
.
Arr
ow
1
rep
resen
ts
6 p
ath
s
lin
kin
g
each
ach
iev
em
en
t m
oti
ve
to
acad
em
ic
ch
oic
e
.
Arr
ow
2
rep
resen
ts
6 X
6
path
s
lin
kin
g
each
ex
og
en
ou
s v
ari
ab
le
to
each
ach
iev
em
en
t m
oti
ve.
Arr
ow
3
rep
resen
ts
6 p
ath
s
lin
kin
g
each
ex
og
en
ou
s v
ari
ab
le
to
acad
em
ic
ch
oic
e.
·Chapter 4
RESULTS
Hypothesis 1. Sex-related Differences
In multivariate analysis ~ significant effects of
sex were found on the set of dependent variables
(academic choice~ expectancies, interest and utility
values for Arts and Science) ( F(7,796) =11.9 , p< .001).
Results ofunivariateanalyses on each of the
dependent variables were shown in Table 8 ·. For the
academic choice, significant effects of sex were found
(F(1,806)=20.4, p<.OOl). The mean scores result showed
that both sexes perferred Science rather than Arts, but,
on the average, boys perferred Science to a much stronger
degree than girls (boys, M=3.49; girls; M=.64) a .
For the three achievement motives (expectancies,
interests and utility values). signigicant effects of sex
were also found. Among them, greatest effect was found . )
Note . 8 . According to the scoring method of academic choice, a
score greater than 0 indicated higher preference towards Science, while a score less than 0 indicated higher preference towards Arts. Because the mean scores of both · sexes were above 0, so on the average, both sexes perferred Science rather than Arts . However, the mean score of males was much higher than that of females, 'so, on the average, boys perferred Science to a much . stronger degree than girls . . These mean scores were obtained before controlling school performances.
P.47
Table 8
Means and F ratios for Academic choices~ and Achievement Motives by sex
Dependent Variables
Academic choice
Expectancy for Science
Expectancy for Art
Interest for Science
Interest for Art
Uti 1 i ty value for Science
Utility value for Art
sex
M F
M F
M F
M F
M F
M F
M F
3.49 3.85 381 90.3*** .64 4.58 427
6.45 1.56 381 · 134.0*** 5.15 1.61 427
5.36 1.54 381 39.9*** 5 . 98 1.27 427
7.55 1.64 381 42.9*** 6.74 1.81 427
5.05 1.77 381 31.4*** 5.71 1.61 427
7.14 1.48 381 16.5*** 6.70 1.62 427
5.86 1.55 381 41.0*** 6.53 1.45 427
20.4***
59.8***
3.8
5.7*
1.7
2.9
12.9***
Note. F- represent analyses with Arts and Science performances controlled.
-- Q < .05 - * Q < .01
- _.... Q < .001
P.48
in Expectancy for · Science ( F(1~806)=59.8~ p<.OOl ) .
9 By looking at the mean scores of boys and girls on the
~xpectancy for Science (boys~ M=6.45; girls~ M=5.15 )~ it
showed that~ on the average~ boys had positive expectancy
for Science while girls had negative expectancy for
Science.
The above significant sex differences were found
after Arts and Science performances were controlled.
That meant~ even when boys and girls had the same school
performances, the boys still had higher expectancy for
Science. However, no similar sex differences was found
on expectancy for Arts, interest for Arts and utility
value for Science after control.
In summary .. the overall result supports hypothesis 1
with only some minor exceptions.
Note. 9 . According to the scoring method of the variables ..
scores above 6 represented positive responses, while that below 6 represented negative responses. As mean score of boys was 6.45 and that of girls was 5.15, so boys had positive expectancy for Science while girls had negati.ve exp~ctancy for Science .. These mean scores were obtained before controlllng school performances.
P.49
Hypothesis ~ Effects of Sex-role Orientation
Sex-role Type
Main effects of sex-role type were found .
all ln
tests (see Table 9) . In boys~ F(3,381)=4.4 .. p< .01 .. ,and
in girls, P·(3 .. 423)= 3.5)' p< .05. This showed that
stUdents in different sex-role type did differ in
thier academic choice.
However, in contrary to hypothesis 2a, the group
means of masculine boys and feminine girls were not
higher than that of their androgynous same sex peers .
That meant, this study could not prove that masculine
boys or feminine girls were more likely than their
androgynous peers to choose Science. In Scheffe1s
multiple comparisons of group means .. no significant
difference between androgyny girls and feminine girls
and between androgyny boys and masculine boys was Dound.
Sex-role Dimensions
The results of . regression analyses were shown in
Table 10 . Masculinity had significant main effects in
all tests (in boys .. b=.16 at p<.Ol; in girls)' b=.20 at
p<.OOl). That meant, for both boys and girls)' the more ,-
masculine he/she was)' the more likely he/she would choose
Science)' even after control for school performances.
However)' femininty showed nO significant effects (except
P.50
' Table 9
Means and F ratios for Academic choice by Sex-ro 1 e Type .
Sex-role Type
Androgynous Masculine Feminine Undifferentiate
Androgynous Masculine Feminine Undifferentiate ·
Androgynous Masculine Feminine Undifferentiate
M
2.56 · 2.93
.58 1.52
4.13 4.07 2.78 2.67
1.18 1.50
- .31 .23
SD N
Whole Sample
4.41 4.50 4.73 4.23 ·
Boys
3.54 3.89 4.12 4.00
Girls
4.65 4.83 4.69 4.14
252 163 142 255
118 91 41
135
134 72
101 120
F
7.1***
4.3**
3.3*
F*
7.6***
4.4**
3.5*
Note. F* represent 'analyses with Ar~and Science performances controlled.
,. 12 < .05 * * 12 < .01
*** 12 < .001
P.51
~
U1
N
Tab
le
10
S
tan
dard
ized
B
eta
C
oeff
icie
nts
an
d
R
Sq
uare
o
f H
iera
rch
ical
Reg
ress
ion
A
naly
sis
o
f M
ascu
lin
ity
an
d
Fem
inin
ity
o
n
Aca
dem
ic
Ch
oic
e
Wh
ole
S
amp
le
Bo
ys
Gir
ls
Pre
dic
tin
g
beta
R2
R2
b
eta
R2
R2
b
eta
R2
R2
S
tep
V
ari
ab
les
incre
men
t in
cre
men
t in
cre
men
t
1 SE
X
-.3
1*
**
.0
98
**
*
.09
8
2 M
.1
8*
**
.1
7*
**
.2
0*
**
2
F N
S N
S -.
10
*
.12
7*
**
.0
28
**
*
.03
9*
**
3
M X
F
NS
NS
NS
.12
7*
**
.0
29
.0
30
**
*
.03
0
.04
0*
**
.0
40
4 SE
X
X M
N
S 4
SEX
X
F
NS
.13
0*
**
5
SEX
X
M X
F
NS
.13
1*
**
.0
04
1 SE
X
-.3
1*
**
.0
98
**
*
.09
8
2 O
vera
ll p
erf
. .1
2
.15
**
.1
0*
*
2 P
erf
. d
iff.
.49
.3
9*
**
.5
3*
**
.3
10
**
*
.21
2
.17
4*
**
.1
74
.2
97
**
*
.29
7
3 M
.1
7*
**
.1
6*
*
.20
**
*
3 F
NS
NS
-.0
8*
.3
36
**
*
.19
9*
**
.3
29
**
*
4 M
X F
N
S N
S N
S .3
36
**
*
.02
6
.19
9*
**
.0
25
.3
29
**
*
.03
2
5 SE
X
X M
N
S 5
SEX
X
F
NS
.33
8*
**
6
SEX
X
M X
F
NS
.33
9*
**
.0
03
No
te.
M =
Mascu
lin
ity
, F
= F
em
inin
ity
, N
S =
No
t S
ign
ific
an
t ..
~
<
.05
, •
It
~
< .
01
••
• 2
< .
00
1.
I
a weak and marginal effect In girls) . And no
interactive effect between femininity and masculinity was
found in all tests.
The insignificant main effect of femininity and the
insignificant interactive effect together implied that
the effect of sex-role type (androgynous .. masculine~
femininie and undifferentiated) on academic choice found
previously was simply due to the effect of masculinity
(ie. M X F and F had no contribution) . According to this
result, androgyny would have no unique predicability on
academic choice .. but it's effect could be represented
simply by that of masculinity.
Results also showed that sex-role dimensions had
significant predictive power on academic choice .. even
after control for school performances. The result in
Table 10 showed that sex-role dimensions significantly
increase the variance accounted for in academic choice
They accounted for 2.596 and 3.2% of the variance in
academic choice of boys and girls respectively .
In summary .. the results failed to support hypothesis
2a .. 2b(ii) and 2b(iii) .. but provided support to
hypothesis 2b(i).
P.53
Hypothesis 3. The Effect of Sex-stereotypes
Significant effects of own-sex suitability on
academic choice were found in all tests (Table 11). In
males, beta coefficients were .32 ; in females, beta
coefficients were .30 , all at p<.001 level. For both
boys and girls, if they believed that students of their
same sex were suitable for studying Science, then they
would more likely choose Science. This effect persisted
significantly even after control for school performances~
However, no significant main effect of opposite-sex
suitability was found. That meant, students' stereotypes
in opposite sex did not affect their academic choice,
though these stereotypes did exist.
The sex-stereotype variables had significant
predictive power on academic choice. The sex-stereotype
variables together accounted for 10% of variance in
academic choice of males; and 9.1% in females. Compare
with the predictive powers of sex-role orientation (2.5%
& 3.2%), that of sex-stereotypes were consistently
larger. - Sex~stereotypes had stronger effects on academic
choice than sex-role orientation. And since opposite-sex
suitability had no significant influence, all the effects
of sex-stereotypes actually came from one single
- variable, the stereoytpes in same sex.
In summary. the results provided support to
P.54
t-c.1
Ul
Ul
Tab
le
11
S
tan
dard
ized
B
eta
C
oeff
icie
nts
an
d
R
Sq
uare
o
f H
iera
rch
ical
Reg
ressio
n A
naly
sis
o
f O
wn
-sex
S
uit
ab
ilit
y
an
d
Op
po
sit
e-s
ex
S
uit
ab
ilit
y
on
A
cad
emic
C
ho
ice
Bo
ys
Gir
ls
Pre
dic
tin
g
beta
R2
R2
b
eta
R2
R2
S"te
p V
ari
ab
les
incre
men
t in
cre
men
t
1 OW
N .3
6*
**
.4
0*
**
1
op
p
NS
NS
.13
1*
**
.1
62
**
*
2 OW
N X
op
p
NS
NS
.13
3*
**
.1
33
.1
63
**
*
.16
3
1 O
vera
ll p
erf
. .1
5*
*
.10
*
1 2 2 3
No
te.
Perf
. d
iff.
OWN
OPP
OWN
X O
PP
OWN
NS
~
.39
**
*
.17
4*
**
.1
74
.3
2*
*
NS
.27
3*
**
N
S .2
74
**
*
.10
0
= O
wn
-sex
S
uit
ab
ilit
y,
OPP
=
N
ot
Sig
nif
ican
t Q
<
.0
5
~*
Q
<
.01
.53
**
*
.29
7*
**
.2
97
.3
0*
**
N
S .3
88
**
*
NS
.38
8*
**
.0
91
= O
pp
osit
e-s
ex
su
itab
ilit
y
.
~ ~ ~
Q
<
.00
1
hypothesis 3(i) but not 3 ,(ii).
Hypothesis 4. Path Model for Academic Choice
Estimates of the path coefficients were presented In
Table 12. Results showed that all achievement motives
(expectancies for success _ interest and utility values
in Arts and Science studies) had significant direct
effects on academic choice . but school performances_
sex-role dimensions _ sex-stereotype variables had no
significant direct effect on academic choice (see Table
12). That meant_ the six paths represented by arrow 1
in the path model depicted in Figure 1 were all
significant, but that represented by arrow 3 were all
insignificant. These results implied that, for all
exogenous variables in the path model, if they did have
any effect on academic choice, then all their effects on
academic choice were mediated by the achievement motives.
The effects of sex-role orientation and sex
stereotypes on the achievement motives were also
studied. Results (Figure 2 & 3) showed that own-sex
suitability had consistent and significant effects on all
six achievement motives. For both boys and girls, if
they believed that students of their same sex w~re more
suitable for Science_ then they would have higher
P.56
~
Ul
,]
Dep
en
den
t V
ari
ab
le
Ch
oic
e
E.S
. E
.A.
I.S
. I.
A.
U. S
. U
.A.
Ch
oic
e
E.S
. E
.A.
I.S
. I.
A.
U. S
. U
.A.
No
te.
Tab
le
12
P
ath
C
oeff
icie
nts
fo
r th
e
Path
M
od
el
of
Aca
dem
ic
Ch
oic
e
Pre
dic
tin
g V
ari
ab
le
--------~------------------------------------------------------------------------------------
Ov
era
ll
Perf
.
.04
.2
5*
**
.1
3*
*
.14
**
.0
8
.13
**
.0
5
.02
.2
4*
**
.2
2*
**
.1
8*
**
.1
0*
.1
3*
**
.0
7
Perf
. D
iff.
.01
.2
9*
**
-.
34
**
*
.26
**
*
-.3
2*
**
.0
8
-.1
9*
**
.04
.4
2*
**
-.
36
**
*
.41
**
*
-.3
6*
**
.2
6*
**
-.
20
**
*
Has
=
Masc
uli
nit
y;
Mas
Fe
rn
.03
-.
02
.2
8*
**
.0
9
.15
**
.0
9
.19
**
*
.07
.0
1
.07
.1
6*
*
.17
**
.0
6
.13
*
.02
-.
03
.2
4*
**
-.
01
.0
3
.09
*
.19
**
*
.01
-.
05
.1
3*
*
.17
**
*
.04
.0
6
.12
**
Ow
n =
Ow
n-s
ex S
uit
ab
ilit
y;
E.S
. =
Ex
pecta
ncy
fo
r S
cie
nce
I.S
. =
In
tere
st
for
Scie
nce;
U.S
. =
Uti
lity
V
alu
e
for
Scie
nce;
• ~
<
.05
•
* 2
<
.01
Ow
n O
pp
E.S
. E
.A.
I.
S .
I.A
. u
. S.
U.A
.
Bo
ys
.03
.0
4
.30
**
*
-.2
3*
**
.2
3*
**
-.
25
**
*
.14
**
*
-.1
0*
**
.1
6*
**
-.
04
-.
26
**
*
.03
.2
3*
**
-.
09
-.
26
**
*
.05
.1
9*
**
.0
9
-.3
2*
**
-.
06
Gir
ls
.04
-.
04
.3
5*
**
-.
18
**
*
.24
**
*
-.1
9*
**
.1
3*
**
-.
10
**
*
.20
**
*
.07
-.
15
**
*
-.0
6
.23
**
*
.05
-.
20
**
*
-.0
9*
.2
4*
**
.1
3*
*
-.2
5*
**
-.
16
**
*
Fern
=
Fem
inin
ity
; O
pp
=
Op
po
site
-sex
S
uit
ab
ilit
y;
E.A
. =
Ex
pecta
ncy
fo
r A
rt;
I.A
. =
In
tere
st
for
Art;
U
.A.
= U
tili
ty
Valu
e
for
Art
. ••
• 2
<
.00
1
t-cl
U1 ro
Fig
ure
2
Resu
lts
of
Path
A
naly
sis
on
A
cad
emic
C
ho
ice
of
Bo
ys
Ov
era
ll
Perf
orm
an
ce
Ex
pecta
ncy
in
S
cie
nce
-----------
Perf
orm
an
ce
Dif
fere
nce
Ex
pecta
ncy
in
A
rts
Masc
uli
nit
y
Inte
rest
in
Scie
nce
Aca
dem
ic
I Fem
inin
ity
In
tere
st
in
Art
s C
ho
ice
Ow
n-s
ex su
itab
ilit
y
Uti
lity
V
alu
e in
S
cie
nce
I~ O
pp
osi
te-s
ex
s~
itab
ilit
;1
Uti
lity
V
alu
e in
A
rts
No
te.
1.
All
th
e
path
co
eff
icie
nts
sh
own
are
at
p <
.0
01
. 2
. T
he
path
co
eff
icie
nts
o
f "o
vera
ll
perf
orm
an
ce"
and
"p
erf
orm
an
ce d
iffe
ren
ce"
are
o
mit
ted
b
ecau
se
these
re
su
lts
are
n
ot
the
maj
or
co
ncern
o
f th
is
stu
dy
. 3
. O
ther
deta
ils
of
the
path
an
aly
sis
can
b
e fo
un
d
in
Tab
le
12
.
~
U1
\0
.... \.
Fig
ure
3
Resu
lts
of
Path
A
naly
sis
o
n
Acad
em
ic Choic~
of
Gir
ls
r-o
vera
ll
Perf
orm
an
ce
Ex
pecta
ncy
in
S
cie
nce
Perf
orm
an
ce
Dif
fere
nce
.Ex
pecta
ncy
in
A
rts
Mascu
lin
ity
In
terest
in
Scie
nce
Aca
d.em
i c
Ch
oic
e
Fem
inin
ity
In
terest
in
Art
s
Ow
n-s
ex su
itab
ilit
y
Uti
lity
V
alu
e
in
Scie
nce
Op
po
sit
e-s
ex
sU
itab
ilit
yl
Uti
lity
V
alu
e
in
Art
s
No
te.
1.
All
th
e
path
co
eff
icie
nts
sh
ow
n are
at
p <
.0
01
. 2
. T
he
path
co
eff
icie
nts
o
f "o
vera
ll
perf
orm
an
ce"
an
d
"p
erf
orm
an
ce d
iffe
ren
ce"
are
o
mit
ted
b
ecau
se
these resu
lts are
n
ot
the
majo
r co
ncern
o
f th
is
stu
dy
. 3
. O
ther
deta
ils
of
the
path
an
aly
sis
can
b
e
fou
nd
in
T
ab
le
12
.
expectancies, interest and utility values in Science but
lower achievement 'motives in Arts than their peers, even
,though they might have the same school. performances.
Similar to previous result~, opposite-sex suitabilit~ had
insignificant and inconsistent effects. That meant,
stereotypes in opposite sex did not affect their
achievement motives.
Again similar to previous results, masculinity had
much more consistent and significant effects on the
achievement motives than femininity. Students with higher
masculinity would have high expectancies, interest and
utility values in Science. Interestingly, nearly all
significant effects of masculinity -were exerted on the
achievement motives of Science, but not on that of Arts.
An explanation to this result had not yet been thought
of.
In summary, hypothesis 4 was fully supported.
P.60
: ~ r '. • •
Chapter 5
DISCUSSION
Summary of Findings
To explain sex differences in academic choice, ,this
study evaluates the effects of sex-role orientation and
sex-stereotype on academic ch~ice, and how these effects
are mediated'. The resul ts do successfully show that ,
even after controlling for school performances,
i) sex differences exist in academic choice ;
ii) masculinity and own-sex suitability' for Science
. favour choosing Science ; and
iii) their effects are mediated by expectancies,
interests and utility values of the academic
studies.
On the other hand, this study found no consistent
significant results for
i) the main effect of femininity, and the interactive
effect between masculinity and femininity on
academic choice ;
ii) the effect of opposite-sex suitability on academic
choice .
P.61
Implication of Findings
This study differs from other study of sex
differences in achievement pattern primarily in its focus
on choice ~ its control for past performances~ and its
concern for the effect of sex-stereotypes.
Academic Choice
Many past researches focus their attention on the
sex differences in achievements. Their reports remind
people of the existing gender structure of our society~
and~ to certain extent~ reinforce reader's sex
stereotypes.
Instead of focusing on sex differences in
achievements~ this study tries to direct peoples'
attention towards sex differences in academic choice. It
highlights the critical role of choice in the process of
how sex differences are developed. This study do
successfully show that ~ even with the same past
performances~ boys are more likely than girls t6 choose
Science. As some girls deprive themselves from the
chance of studying science~ therefore~ in later years~
girls will have much less opportunity in science-related
educational and occupational fields~ and are likely to be
weaker in science-related cognitive areas. In this way~
sex differences in achievement patterns may derive from
P.62
equal early performances. One important implication of
this study is that boys and girls differ in achievement~
. career or even life patterns because they make different
choices.
Past performance
Another contribution of controlling past
performances is to provide a more vigorous proof for the
causal effects of sex-role orientation and sex-
stereotypes. As past performances are believed to be an
important factor for academic choice~ if it is not
controlled, one may suspect that the effects of
sex-role orientation and sex-stereotypes are only
spurious, past performances may be .a common factor
affecting both the predicting variables and the academic
. choice. However, this study succeeds in finding that the
effects of sex-role orientation and sex-stereotypes
persist even after past performances are partial led ~ut.
For individuals with same academic performances, those
higher in masculinity or thinking that Science is
suitable for their own sex~ are more likely to choose
Science. These two psychological factors are then
believed to play substantial roles in shaping the sex
differentiated achievement pattern .
P.63
Sex-role Orientation
This study finds that~ for both boys and girls ~
masculinity favour Science choice. As~ . on the average~
boys are more masculine that girls~ the result of this
study implies that masculinity creates sex differences in
academic choice. Other implications of the results of
sex-role orientation are discussed in separate section.
Sex Stereotypes
This study show that sex-stereotypes in Art and
Science studies really exist and are accepted by both
boys and girls. It also shows that sex-stereotypes
significantly affect academic choice. Because boys
believe that boys are more suitable for Science while
girls believe that girls are more suitable for Arts, so
more boys are going to choose Science than girls, even
they might have the same past performances. In this
way~ sex-stereotypes create sex differences in ' academic
choice which, in turn, result in later sex differences in
achievement and career patterns. Sex-stereotypes have
acted an important mechanism through which theexistinq
gender structure of our society is reoroduced.
In most of the past researches~ sex-stereotypes are
conceptualized in bipolar terms. They usually refer to
how far a quality is stereotyped to boys as compared with
P.64
girls~ or vice versa. However~ this study found that
only own-sex suitability has significant effect on
academic choice~ but opposite-sex suitability has no
effect. Therefore~ one important implication of th~
result is that the perception in the same sex rather than
the relative perception between the two sexes influence
respondents' behaviors. Conceptualizing sex-stereotypes
in two independent factors (ie.stereotypes in own sex and
stereotypes in opposite sex) can provide us a more
precise understanding of the effect of sex-stereotypes.
In this study~ a new instrument which adopts
probabilistic judgement in assessing sex-stereotypes
in Arts and Science studies are developed. This
instrument is believed to be quite successfully developed
because it shows good psychometric properties. This
result implies that people are not absolutely assigning
certain activities or characteristics to boys/girls~ but
rather believe that there are reasonable probability of
boys and girls sharing similar qualities. Tapping sex
stereotypes in an absolute manner is not recommended .
Overall Remarks
As an overall remark to this study~ the results of
this study are encouraging. There are two reasons.
First~ the results suggest a more Dositive view of
P.65
women's achievement motivation than those inherent in
other theories. Many popular explanations of sex
differences in achievements are based on deficit models
of female achievement orientation. For example, sex '
differences in achievement patterns have been attributed
to females' learned helplessness, low self-concepts, low
expectancy attributional patterns, and fear of success
(Eccles,1985). Each of these theories suggests that
females are deficient in some critical components of
achievement motivation. They provide not much solution to
the question of what cause these deficiencies in
achievement motives, and how they can be improved. This
study. succeeds in finding that these achievement motives
are affected by sex-role orientation and sex-stereotypes.
Females are deficient in the expectancies, interest and
utility values of Science studies simply because they are
less masculine and they stereotype girls to Arts rather
than Science. In this study, sex differences in
achievement motives are not considered as causes of sex
differences in academic choice. but only as mediators of
the effects of 'sex-role orientation and sex-stereotypes
on academic choice.
Second, the results highlight the importance of
modifiable factors in creating sex differences. First of
all, academic choice which is believed to be more
P.66
modifiable than academic achievements, is ch6sen to be
the dependent variable of ,this study. And instead of
studying the effect of school performances on academic
choice, this study focuses on two more modifiable
psychological factors -- sex-role orientation and sex
stereotypes. Among them, the effects of sex-sereotypes
are found to be consistently larger. This result is
encouraging because sex-stereotypes in Arts/Science
studies are only personal beliefs in specific domains,
they are expected to be more easily modified than
masculinity.
Since students' sex-stereotypes significantly affect
their expectancy, interest and utility values in Arts/
Science (see figure 2 & 3), one may suspect that students
are having sex-stereotypes in these three domains. For
example, if some girls believe that girls do not have the
ability to study Science; girls have no interest in
Science; and girls are not suitable for Science-related
careers, then they the~selves would tend to have lower
expectancy for success" interest and utility values in
Science. Some measures should be implemented to modify
students' sex-stereotypes in these three areas. studies
of intervention prog~ams (Eccles & Hoffman,1984 ;
Casserly, 1980) have demonstrated that the s~x
stereotypes of students can be modified by
P.67
appropriate role models, information, career guidance and
teachers' and parents' attitudes.
Further Discussion
on Sex Role Orientation
Though the hypothesis of this study is based on
Bern's gender schema theory, its results do not support
Bern's theory. Androgyny is found to have no unique
predicative power, and the effect of sex-'role orientation
all comes from a single masculinity scale in BSRI.
Its Relation with sex-stereotypes
This study has tried to find evidence for B'em' s
theory in another area but in vain again. According to
her theory, a gender schematic individual would be one '
who readily search for and assimilate incoming
information into their gender schema. Therefore,
schematic individual would be highly sex-stereotyped in
the cultural directions. Bern's theory may then predicts
that masculine boys and feminine girls would have greater
sex-stereotypes in Arts/Science stUdies than their
androgynous same sex peers. However, in further data
analyses, neither sex-role type nor sex-role dimensions
are found to have significant relations with any sex-
stereotype 'vari ab I es . P.68
Spence's Gender Identitv theory
In search for an explanation for the results of sex-
role orientation in this study, Spence's gender identity
theory is reviewed.
Spence (1985) defined gender identity as a
fundamental existential sense of one's maleness or
femaleness. She suggested that this gender identity
guided the acquisition of gender-congruent attributes and
behaviors in very young childhood~butits guiding role
would become minimized after gender identity is
established.
The main contrast between Bem's and Spence's theory
is that, the former exists a single central construct
which c~n explain the acquisition of all gender-related
charactertics, while the latter does not has such a
construct. As commented by Spence & Helmreich (1978), the
implicit assumption of Bem's theory is that all gender~
related phenomena contributed to a single bipolar factor,
(either termed as sex-role orientation or gender
schematicity), so that 'the assessment of an individual's
masculine and femininie personality traits can be used to
infer his or her posltion on the hypothetical bipolar
continuum. This position can, in turn, predicted most of
his/her gender-related behaviors or characteristics.
However, Spence postulated that gender identity only
P.69
had a less active role. She suggested that gender
related differences were results of multiple sets of
variables, besides gender identity. In her theory,
gender-differentiating phenomena are considered to be
multidimensional. An individual's standings on each'
dimensions must be independently assessed rather than
measured by any specific collection of gender-related
characteristics, or by any all-purpose instrument. Thus,
individuals' self-descriptions measures in personality
traits ,like BSRI, would have little gender-related
implications, but instead would reflect possession of
instrumental and expressive personality traits . It
cannot be use to predict activities beyond instrumental-
expressive dimensions. (Spence, 1984, 1985 ; Spence &
Helmreich,1978). Signorella et.al. (1989) have proyided
some evidence for Spence's theory. And there is a growing
. body of evidence for the multidimensionality of gender
relatedphenonema (Orlofsky ,1981; Orlofsky,Cohen &
Ramsden, 1985; ·Deaux and Lewis .. 1983 .. 1984).
The results of this study support Spence' theory
more than Bern's theory. According to Spence's theory, M
and F scales will .have no interactive effect but only
main effects .. and M and F scales in BSRI will not have
any relation with sex. stereotypes in Arts/Science studies
( because the M and F scales are only measures of
P.70
instrumentality and expressiveness). Moreover~ the
results of this study are in line with some past
researches. The reviews 'of Taylor and Hall (1982)~ Hall
and Taylor (1985) ~ Whitley (1983) and Lau1s research
(1989) all provide support to the strong effect of M, and
favour the "masculine ll model .
In Summary
The result of this study casts doubts on the Bem1s
gender schema theory~ and on what BSRI is really
measuring. It suggests a necesscity for t'urther research
in this question. And if Spence1s viewpoint is adopted in
this study, then instrumental personality, rather than
masculinity~ is considered to have significant effect on
academic choice.
P.71
Directions for future research trend
Sex-stereotypes . '
After reviewing research on gender in the past
decade, Deaux (1984) has summarized t~at , research on
gender can be divided into three major approaches: a) sex
as a subject variable; b) individual differences in
masculinity, femininity, and androgyny; and c) sex as a
social category. The first approach looks at the
differences between men and women, the second approach
looks at the differences between individuals differing in
masculinity, femininity and androgyny, and the last
approach look at how sex stereotypes affects observers'
judgements or individuals' behaviors .
She comment that :
Main effect differences of subject sex are found to be surprisingly small in most cases, and the status of androgyny is uncertain, limiting the potential of the first two approaches. The impact of sex as a social category is considerable , but more detailed research is needed . . (Deaux, 1984, p.10S)
The result of this study support her view. It cast
doubts on the conception and the measurement of
masculinity, femininity and androgyny. And it does not
support the unique predicability of androgyny which is
proposed by Bern . On the other hand, this study finds
that sex-stereotypes , in academic studies are prevasive
and have sigificant effect on academic choice . . These
results suggest that the third approach is promising.
P.72
Studies which examine the precise content and structure
of individuals' sex-stereotypes, and how these
stereotypes affect individuals' behaviors. are ln need.
For measurement and conceptualization of sex-stereotypes,
probabilistic judgement and two-factor model are
recommended .
Sex differences in Choices
In her review, Deaux' (1984) further suggested that
Yet an understanding of the role of gender in our society ·cannot rest on stereotypes alone .... To fully deal with the ways in which gender is influential, one must ultimately deal with the processes involved. (p.113) .
In line with her view, this study does suggest that
studying choices made by males and females would lead to
better understanding of the process of how sex difference
are developed. Boys and girls with same past
performances have different academic choice , creating
larger sex difference in later educational outcomes and
career patterns. The choice studied in this research is
only one of the numerous choices which boys and girls
have to make. From the choice of toys to the choice of
careers and family roles, sex role has exerted its great
influence on us. Td understand the process how gender
influence our society. longitudinal study of choices made
by males and females on a wide variety of sex-typed tasks
are required. P.73
-Limitations of the study
Caus~-and-effect relatioriships.
Since the present study is a corr~lational study, it
is difficult to establish cause-and-effect relationships
between sex-role" orientation , sex-stereotypes and
academic choice. In fact, through interviews~ it is
found that many students already have some preference
between art and science studies in very early years.
These students may adjust their stereotypes to fit their
preference, or diversify their efforts in studing to Arts
and Science. These stereotypes and the differentiation
in performance resulted may then further reinforce their
original choice. It is highly likely that sex-stereotypes
academic preference and academic performance are a web of
interacting variables. A "longitudinal study on choice,
performance and stereotypes are suggested for a better
understanding of their causal relationships.
Intended academic choice.
Another point need to emphasize is that the academic
choice measured in this study is only the intended choice
of students . The intended choice is used instead of the
formal choice made to school because the former is more
likely to be a free choice than the latter. There is two
reasons. First, when some students are aware that their
P.74
poor examination perform~nce has ~lre~dy ruled out their
chance of being allocated to a specific stream~ they may
adjust their choice to fit their performance. Second~
·the formal choice to school of some students may not
actually be their own choice~ but the choice of their
parents 10 . The educational implication of this study
would depend on the correlations between the intended
choice tapped in this study and the formal choice made to
school ~ which can only be found in follow-up studies.
Effect of sex-role socialization.
How sex role influences academic choice is the major
concern of this study. However, only two of its
related psychological factors are involved. Socialization
factors like the expectations and attitudes of pa~ents
and teachers are not examined in this study. These
socialization factors are found to have important
influence on students'achievement attitudes (Eccles,
Adler & Kaczala,1982; Eccles,1987). The effects of these
sex-role socialization factors need further investigation
by other studies.
Note. lO.However J this number of student is believed to be only
very small. From 'a pilot study done ,only two out of 57 students let their parents make the decision of academic choice.
P.75
~
"-J
0'.
Subje~t
Math
em
ati
cs
En
gli
sh
Ch
inese
Ap
pen
dix
1
Nu
mb
er,
Perc
en
tag
e
an
d
Rati
o
of
Male
s an
d
Fem
ale
s g
ett
ing
A
g
rad
e
in
19
90
H
on
g
Kon
g Ce~tificate
of
Ed
ucati
on
E
xam
inati
on
Nu
mb
er
23
14
46
3
57
9 M
ale
s
% b
ase
d
on
to
tal
no
. o
f m
ale
s sit
tin
g
the
su
bje
ct
ex
am
.
6.2
%
1.2
%
1.5
%
Fem
ale
s
Nu
mb
er
81
7
82
4
12
30
% b
ase
d
on
to
tal
no
. o
f fe
male
s sit
tin
g
the
su
bje
ct
exam
.
1.9
%
1.8
7%
2.7
8%
Rati
o
% o
f m
ale
s %
of
fem
ale
s)
3.3
1
1 1
.6
1 1
.9
No
te.
Data
are
ta
ken
fr
om
1
99
0
Ho
ng
K
on
g C
erti
fic
ate
o
f E
du
cati
on
E
xam
inati
on
A
nn
un
al
Rep
ort
,
pu
bli
sh
ed
b
y
Hon
g K
on
g
Ex
am
inati
on
s A
uth
ori
ty
'l::J
'-J '-J
'1
Ap
pen
dix
2
Nu
mb
er,
Perc
en
tag
e
an
d
Rati
o
of
Male
s an
d
Fem
ale
s sit
tin
g
19
90
H
on
g
Kon
g C
erti
fic
ate
o
f E
du
cati
on
E
xam
inati
on
----
----
----
----
----
----
----
----
----
----
----
----
----
----
----
----
----
----
----
----
----
---
Su
bje
ct
Male
s
% b
ase
d
on
to
tal
Nu
mb
er
no
. o
f m
ale
s sat
sit
tin
g
the
H.K
.C.E
.E.
(Scie
nce
Su
bje
cts
)
Ph
ysi
cs
21
59
2
54
.7%
C
hem
istr
y
20
11
5
50
.9%
B
iolo
gy
1
90
77
4
8.3
%
(Art
s su
bje
cts
)
His
tory
6
66
9
16
.9%
G
eo
gra
ph
y
10
38
7
26
.3%
E
con
om
ic
11
36
7
28
.8%
C
hin
ese
H
isto
ry
38
40
9
.7%
Fem
ale
s
% b
ase
d
on
to
tal
Nu
mb
er
no
. o
f fe
male
s sat
sit
tin
g
the
H.K
.C.E
.E.
97
75
9
67
3
12
57
0
14
73
8
20
33
9
21
28
4
65
02
20
.7%
2
0.5
%
26
.6%
31
.2%
4
3.1
%
45
.1%
1
3.8
%
Rati
o
% o
f m
ale
s %
of
fem
ale
s)
2.6
2
.5
1.8
1 1 1 1
1 1 1 1.8
1
.6
1.6
1
.4
--------------------------------------------------------------------------------------
No
te.
Data
are
ta
ken
fr
om
1
99
0
Ho
ng
K
on
g C
ert
ific
ate
o
f E
du
cati
on
E
xam
inati
on
A
nn
un
al
Rep
ort
,
pu
bli
sh
ed
b
y
Hon
g K
on
g
Ex
am
inati
on
s A
uth
ori
ty
t"(J
'-l
CO
Su
bje
ct
Ap
pen
dix
3
Nu
mb
er,
Perc
en
tag
e
an
d
Rati
o
of
Male
s an
d
Fem
ale
s sit
tin
g
Ho
ng
K
on
g
Ad
van
ced
L
ev
el
Ex
am
inati
on
1
99
0
Nu
mb
er
sat
Male
s
% b
ase
d
on
to
tal
no
. o
f m
ale
s sit
tin
g
the
H.K
.A.L
.E.
Fem
ale
s
% b
ased
o
n to
tal
Nu
mb
er
no
. o
f fe
male
s sat
sit
tin
g
the
H.K
.A.L
.E.
(Scie
nce
Su
bje
cts
)
Ap
pli
ed
M
ath
. 2
49
9
35
.0%
3
73
5
.6%
P
ure
M
ath
. 2
84
3
39
.8%
6
49
9
.8%
P
hy
sic
s
48
27
6
7.6
%
17
32
2
6.0
%
Ch
em
istr
y
39
21
5
4.9
%
16
44
2
4.7
%
Bio
log
y
19
65
2
7.5
%
10
86
1
6.3
%
(Art
s su
bje
cts
)
Ch
inese
1
14
2
16
.0%
3
17
5
47
.7%
H
isto
ry
67
1
9.4
%
17
14
2
5.7
%
Ch
inese
H
isto
ry
10
31
1
4.4
%
21
72
3
2.6
% .
G
eo
gra
ph
y
11
09
1
5.5
%
23
87
3
5.9
%
Eco
no
mic
1
54
1
21
.6%
2
90
7
43
.7%
Rati
o
% o
f m
ale
s %
of
fem
ale
s)
6.3
1
4.1
1
2.6
1
2.2
1
1.7
1
1 3
.0
1 2
.7
1 2
.5
1 2
.3
1 2
.0
-----------------------------_
._-------------------------------------------------------
~~~_~
_ D
ata
are
ta
ken
fr
om
H
on
g
Ko
ng
A
dv
an
ced
L
ev
el
Ex
am
inati
on
A
nn
un
al
Rep
ort
1
99
0
" p
ub
lish
ed
b
y
Ho
ng
K
on
g
Ex
am
inati
on
s A
uth
ori
ty
,
t-cI
'-l
\()
App
end:
ix
4
Rati
o
of
Male
s an
d
Fem
ale
s in
th
e
Fu
ll-t
ime
En
rolm
en
t in
u
niv
ersit
ies
of
Ho
ng
K
on
g
in
19
89
Co
urs
e
/Facu
lty
%
o
f b
oy
s %
of
gir
ls
Den
tal
Stu
die
s
98%
2%
E
ng
ineeri
ng
97
%
3%
Scie
nce
79%
21
%
Med
icin
e
78%
22
%
Arc
hit
ectu
re
69%
31
%
Art
s
28%
72
%
Bu
sin
ess
Ad
min
str
ati
on
38
%
62%
S
ocia
l S
tud
ies
43%
57
%
Law
45
%
55%
Rati
o
(%
of
bo
ys
% o
f g
irls
)
49
.0
1 3
2.3
1
3.
8 1
3.5
1
2.2
:
1
1 2
.6
1 1
.6
1 1
.3
1 1
.2
----------------------------------------------------------~-----------------------
~9~~~
Data
are
ta
ken
fr
om
H
on
g
Kon
g A
nn
ual
Dig
est
of
Sta
tisti
cs,
19
90
, p
ub
lish
ed
b
y
Cen
sus
an
d S
tati
sti
cs
Dep
art
men
t,
Hon
g K
on
g.
Appendix 5
Measures of Sex-stereotypes
in Arts and Science studies
P.BO
The following Questions ~sk you ~o g~ess hON ~any boys/girls are suitable for art/science study. Please answer each questlon by clrcllng suitable nUlber.
Eg. You think that how ~any ' boys would like to do science experiments? Nearly quite half quite nearly
none few lany all o 1 2 3 4 5 6 7 8 9 10
The answer implies that generally, out of 10 boys, about 8 of them like to do science experilents.
(Science questions) You think that generally,
how many girls like to do science experiments?
1. how many boys like to have science lessons? 2. how lany girls like to have science lessons?
3. how .any boys can study science well? 4. how many girls can study science Hell?
5. how many boys in science class can get a suitable job ? 6. how .any girls in science class can get a suitable job?
7. hON .any boys feel science interesting? 8. how many girls feel science interesting?
Nearly Quite half quite nearly none few many all o 1 2 3 4 5 6 7 8 9 10
o o
2 3 4 5 6 7 B 9 10 2 3 4 5 6 7 8 9 10
o 1 2 3 4 5 6 7 8 9 10 o 1 2 3 4 5 6 7 8 9 10
o o
o o
2 3 4 5 6 7 8 9 10 2 3 4 5 6 7 8 9 10
2 3 4 5 b 7 8 9 10 2 3 4 5 b 7 8 9 10
9. hON lany boys can get good results in maths. in H.K.C.E.E.? 0 10.how many girls can get good results in maths. in H.K.C.E.E.? 0
2 3 4 5 6 7 8 9 10 2 3 4 5 6 7 8 9 10
. It.hoN many boys in science class would have good prospect? 0 2 3 4 5 6 7 8 9 10 2 3 4 5 6 7 8 9 10 12.how many girls in science class would have good prospect? 0
(Art questions) You think generally,
13.hoN many girls like to have art lessons? 14.how many boys like to have art lessons?
15.how many girls can study art subjects well? lo.how many boys can study art sujects well?
17.how .any girls in art class can get a suitable job? 18.hoN many boys in art class can get a suitable job?
19.hoN .any girls feel art subjects interesting? 20.hoN lany boys feel art subje~ts interesting?
Nearly quite half quite nearly none few many all o 1 2 3 4 5 6 7 8 9 10 o 1 2 3 4 5 6 7 8 9 10
o o
o o
o o
2 3 4 5 6 7 8 9 10 2 3 4 5 6 7 8 9 10
2 3 4 5 b 7 B · 9 10 2 3 4 5 b 7 8 9 10
2 3 4 5 b 7 8 9 10 2 3 4 5 b 7 8 9 10
21.hoH many girls can get good results in English in H.K.C.E.E.? O· · 22.hoH many boys can get good results in English in H.K.C.E.E.? 0
2 3 4 5 b 7 8 9 10 2 3 4 5 b 7 8 9 10
23.how lany girls in art class would have good prospect? 24.hoN .any boys in art class would have good prospect?
P.81
o o
2 3 4 5 b 7 a 9 10 2 3 4 5 6 7 8 9 10
以下是訧你估訂有幾多男./女生是適合中四時升諛文/理科.囲上逋畨的数宇作答.
例題： 你涊為有幾多男生含喜揿做科學赏驗？
幾顿有 m -半 m w所有
0 1 2 3 4 5 6 7 ( ^ ) 9 10
答茱表示：
普通來說• 10阔男生中,大約有8脑會喜駄做科學S驗.
(理抖問題）你涊為普通來說：
幾乎沒有頦少 一半 頦多幾乎所有 有幾多玄J^含喜歡做科學筲驗？ 0 1 2 3 4 5 6 7 8 9 10
有幾多•营喜撳上理科的課？ 0 1 2 3 4 5 6 7 8 9 10
有 幾 多 t 甚 歡 上 理 抖 的 課 7 0 1 2 3 4 5 6 7 8 9 10
有 幾 多 是 能 夠 諛 好 理 科 ？ 0 1 2 3 4 5 6 7 8 9 10
有 幾 多 是 能 夠 諛 好 理 抖 ？ 0 1 2 3 4 5 6 7 8 9 10
有幾多諛理科的能找到適合自己的橄業？ 0 1 2 3 4 5 6 7 8 9 10
有幾多諛理枓的能找到適合自己的呦業？ 0 1 2 3 4 5 6 7 8 9 10
有 幾 多 會 覓 得 理 科 有 趣 ？ 0 1 2 3 4 5 6 7 8 9 10
有幾多玄會兌得理科有趣？ 0 1 2 3 4 5 6 7 8 9
有圾多M是能夠在曾考的数學抖中収搰好成頜？ 0 1 2 3 4 5 6 7 8 9 10
有幾多M是能夠在會考的数學抖中収得好成頜？ 0 1 2 3 4 5 6 7 8 9 10
有幾多諛理抖的•會有好前途？ 0 1 2 3 4 5 6 7 8 9 10
有 幾 多 諛 理 科 的 會 有 好 前 途 ？ 0 1 2 3 4 5 6 7 8 9 10
(文科問題）你認為普通來说： 《 m m + , ^ 1 幾乎沒有頗少 一 半 頗 多 幾 乎 所 有
有幾多•會喜酞上文科的課？ 0 1 2 3 4 5 6 7 8 9 10-
有幾多•會喜歡上文科的现？ 0 1 2 3 4 5 6 7 8 9 10
有幾多女生是能夠諛好文科？ 0 1 2 3 4 5 6 7 8 9 10
有幾多•是能夠諛好文科？ e 1 2 3 4 5 6 7 8 9 10
有幾多諛文科的能找到適合自己的轍業？ 0 1 2 3 4 5 6 7 8 9 10
有幾多謓文抖的^^能找到適合自己的橄菜？ 0 1 2 3 4 5 6 7 8 9 10
有幾多玄一圭吉凭得文科有趣？ 0 1 2 3 4 5 6 7 8 9 • 10
有幾多H會凭得文科有趣？ 0 1 2 3 4 5 6 7 8 9 10
有幾多M是能夠在言考的英文科中取得好成頜？ 0 1 2 3 4 5 6 7 8 9 10
有幾多@是能夠在會考的英文抖中収得好成頜？ 0 1 2 3 4 5 6 7 8 9 19
有幾多UI文抖的玄L圭會有好前途？ 0 1 2 3 4 5 6 7 8 9 19 .
有幾多謓文抖的M會有好前途？ 0 1 2 3 4 5 6 7 8 9 10
P.82
Appendix 6
Measures of Achievelent Motives
(expectancies for success, interest and utility values)
in Arts and Science studies
P.B3
Please circle a suitable anSMer for . each question.
1. If you study art streal, how well do you think your r.esult will be ?
very poor
2. If you study science stream, how Mell do you think your result will be ? 1 2 1 2
very good
345 345
3. Do you like to have art lessons? 4. Do you like to have science lessons?
5. If you study art streal, do you think you can get a favourite job? 6. If you study science stream, do you think you can get a favourite job?
very dislike 123 123
lost probably
can 1 1
2 2
very poor
3 3
7. If you study art streal, how well do you think your H.K.C.E.E. result will be ? 1 2 B. If you study science streal, how well do you think your H.K.C.E.E. result will be? 1 2
very dull
very 1 He
4 5 4 5
lost probably cannot
4 5 4 5
very good
345 345
very interesting
9. DD you feel art subjects interesting ? lO.Do you feel science subjects interesting?
1 2 3 3
4 5 1 2 4 5
very poor
11.If you study art streal, how well do you think your future prospect will be ? 1 2 12.If you study science stream, how well do you think your future prospect will be? 1 2
13.According to your own will, do you want to prolote to art stream? 14.According to your ONn will, do you want to prolote to science streal ?
surely don't want
123 123
surely
very well
345 345
surely want
4 5 4 5
15.If you would be allocated absolutely according to your wish, would you choose art stream ?
would not 123
surely would
4 5
16.1 you would be allocated absolutely according to your wish, would you choose science stream?
P.B4
2 3 4 5
譆圄上適當的答案.
很差頗差平平頗好很好
如果你升讀M,你佔計你將來的成结會怎榇？ 1 2 3 4 5
如果你升謓皿,你佔計你將來的成结會怎樣？ 1 . 2 3 4 5
很不喜歡 頦不喜歡 普通 頗喜歡很甚歡
你是否喜歡上^iJ的堂？ 1 2 3 4 5
你 是 否 喜 歡 上 的 堂 ？ 1 2 3 4 5
多势 戎If 多數 不 f 夠 不 能 夠 沒 意 見 能 夠 i 毙
如果升你估計你將來能否找到自己喜歡的锘業？ 1 2 3 4 5
如果升讀里你估計你將來能否找到自己喜歡的職業？ 1 2 3 4 5
很差頗差平平頗好很好
如果你升謓你估計你會在會考中収得计麼成頜？ 1 2 3 4 5
如果你升謓你估計你會在會考中取得什麼成緝？ 1 2 3 4 5
很 沉 悶 頗 沉 悶 普 通 頗 有 趣 很 有 趣
你 覓 得 有 沒 有 趣 味 ？ 1 2 3 4 ’ 5
你覓得Sif有沒有趣味？ 1 2 3 4 5
很差頗差平平頗好很好’
如果你升謓tg,你估計你將來的職業出路會怎榇？ 1 2 3 4 5
如果你升謓里你估計你將來的職業出路會怎樣？ 1 2 3 4 5
很 不 想 頗 不 想 沒 意 見 頗 想 很 想
根據你値人的志願•你想中四時升諛ML嗎？ 1 2 3 4 5
根據你値人的志顋•你想中四時升談U!嗎？ 1 2 3 4 5
首定可能 ^ ^ — •^能1定 不會不會未決定 含 會
如果學抆完金依照你的志願分班•你會選擇复丑嗎？ 1 2 3 4 5
如果學抆完金依照你的志顋分班•你會選擇U!嗎？ 1 2 3 4 5
P.85
Appendix 7
Measures of Sex-roie Orientation
(simplified, translated version of
Be.'s Sex Role Inventory)
P.86
Please use the following adjectives to describe yourself. Circle the number which best describes you •
al.ost alllost never always true true
1. a.bitious · .......... , ......... 2 3 4 5 6 7 2. gentle I ••••••••••••••••••••••• 2 3 4 5 6 7 3. aggressive •••••• I •••••• I •••••• 2 3 4 5 6 7 4. affectionate .................. 2 3 4 5 6 7 5. analytical .................... 2 3 4 5 6 7 6. childl He • •••••••••••••••• I ••• 2 3 4 . 5 6 7
7. co.passionate , ................ 2 3 4 5 6 7 B. defends own beliefs • I •••••••• 2 3 4 5 6 7 9. .eager to soothe hurt feelings •• 2 3 4 5 6 7 10. has leadership abilities • I •••• 2 3 4 5 b 7 11. independent , self-reliant .... 2 3 4 5 6 . 7
12. love children ................. 2 3 4 5 " 7 13. lakes decisions easily ........ 2 3 4 5 6 7 14. sensitive to the needs of others 2 3 4 5 b 7 15. strong personality ........... 2 3 4 5 6 7 16. willing to take a stand ....... 2 3 4 5 " 7 17. soft spoken .................. 2 3 4 5 6 7
almost allost never always true true
lB. willing to take risks •••••••••• 1 2 3 4 5 6 7
19. sYlpathetic • ••••• I •••••••••••• 1 2- 3 4 5 " 7 20. cOlpetitive · .................. 1 2 3 4 5 6 7
21. tender I ••••••••• I •• " •••••••••• 2 3 4 5 b 7
22. understanding • ••• I •••••••••••• 2 3 4 5 6 7
23. dominant • ••••••••••••••••• I ••• 2 3 4 5 6 7
24. sel f-sufficient • •••• , •••••• I •• 1 2 3 4 5 6 7
25. war • ....... , ................... 1 2 3 4 5 6 7
26. y ie I d i ng ••••••••••..••••.•••.•• 1 2 3 4 5 6 7
27. i ndi vi dualistic . .............. 1 2 3 4 5 b 7
2B. gullible . ..................... 1 2 3 4 5 6 7
29. acts as a leader ••••••••••••••• 2 3 4 5 6 7
30. flatterable · ......... '-......... 2 3 4 5 6 7
31. assertive · .................... 2 3 4 5 b 7
32. loyal . ......................... . 1 2 3 4 5 b 7
33. forceful . ...................... 1 2 3 4 5 6 7
34. shy ••••••••.•••••••••••••••••• I 1 2 3 4 5 6 7
P.Bl
請你)U以下的形容詞來描述你自己•圚上最能形容你的数宇.
幾乎從來馆常往往偶然往往常常路
不是 不是不是是 是 是 是
1 .有雄心壯志 . . . . . . . . . . . 0 1 2 3 4 5 6
2 .斯文 0 1 2 3 4 5 6
3 . 楨 搔 迆 収 > 1 2 3 4 5 6
4 .溫情 0 1 2 3 4 5 6
5.笞於分 析 0 、 1 2 3 4 5 6
6.像孩子般純真 ••••••• 0 1 2 3 4 5 6
7 .惣情瑝富 . .”. . . .•. . 0 1 2 3 4 5 6
8.維諶自己的信念....…，.0 1 2 3 4 5 6
热充於安慰偽心的人.，..0 1 2 3 4 5 6
10.有領汕才能 0 1 2 3 4 5 6
11.性格攔立•不肷賴別人..0 1 2 3 4 5 6
12.焚小孩子..•，.••••••• 0 1 2 3 4 5 6
13.決斷 . . . . . . . . . . . . 0 1 2 3 4 5 6
14.笞於了解別人的慼受••• 0 1 2 3 4 5 6
15.脑性堅 強 0 1 2 3 4 5 6
16.對琪俏•有自己的立埸..0 1 2 3 4 5 6
1 7 . 說 話 锉 菜 0 ' 1 2 3 4 5 6
幾乎從來常常往往偶然往往常常1
不是 不是不是是 是 足 是
1 8 . 願 意 窗 險 0 1 2 3 4 5 6
19.富间情心.‘•••••••••• 0 1 2 3 4 5 6
20.好與別人競爭...‘.，..0 1 2 3 4 5 6
21.溫柔•...•..•••••••• 0 1 2 3 4 5 6
22.諄解別人.••..•••••• 0 1 2 3 4 5 6
23.有支§2他人的傾向...• 0 1 2 3 4 5 6
24.自足•不須要別人的斛肋 0 1 2 3 4 5 6
25.對人親切，熱俏....... 0 1 2 3 4 5 6
26.赏退謓••..，.•..•.:••. 0 1 2 3 4 5 6
27.我行我索,以自己為中心 0. 1 2 3 4 5 6
28.易聰信別人……，… 0 1 2 3 4 5 6
29.行為像値領袖 ••••••• 0 1 2 3 4 5 6
3 0 . 易 披 人 華 承 . . . . . . . . . 0 1 2 3 4 5 6
31.政於表現自我 ••••••• 0 1 2 3 4 5 6
32.對人忠心耿耿 . . . . . - 0 1 2 1
• 4 5 6
53.性格強而有力 ••••••• 3 1 2 3 4 5 6
3 4 . 省 羞 0 1 2 3 4 5 6
P.88
REFERENCE
Archer, J. & Freedman', S. (1989). Gender-stereotypic perceptions of academic disciplines. British Journal of Educational Psychology, 59, 306-313.
Baker, D. R. (1987). The influence of role-specific self, concept and sex-role identity on career choices in science. Journal of research in science teaching, 24 .. 739-756 .
Beere, C. A. (1979). Women and Women's Issues: A handbook of tests and measures. Jossey-BassPublishers.
Bern, S. L. (1974). The measurement ' of psychological androgyny. Journal of Consulting and Clinical Psychology, 42, 155-162.
Bern, S. L. (1981). Gender schema th.eory: A cognitive account of sex-typing. Psychological Review,88, 354-364.
Bern, S. L. (1985). Androgyny and gender schema theory: A conceptual and empirical integration. In' T.B. Sonderregger (Ed.), Nebraska Symposium on Motivation: Psycholoay and gender. (pp. 179-226). Lincoln: University of Nebraska Press.
' Bern, S. L., & Lenney, E. (1976). Sex typing and the avoidance of cross-sex behavior. Journal of personality and Social Psychology, 33, 48-54.
Benbow,C.P.,& Stanley,J.C. (1980). Sex differences in mathematical ability: Fact or artifact? Science, 210, 1262-1264.
Casserly,P. (1980). An assessment of factors affecting female participation in advanced placement programs in mathematics .. chemistry, and physics. Report to the National Science Foundation. In L.H.Fox, L.Brody,& D.Tobin(Eds), Women and the mathematical mystique. Baltimore: Johns Hopkins University Press.
Cheung .. M. C. F. (1986). Development of gender stereotype. Educational Research Journal (Hong Konq) , 1, 68-73.
Cohen,J.~& Cohen,P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ:Erlbaum.
Deaux .. K. (1984). From individual differences to social categories .. analysis ofa decade's research on gender. American Psychologist, 39, 105-116.
Deaux, K., & Lewis, L. L. (1983). Assessment of gender stereotypes: ~· Methodology and components . Psychological Documents, 13, 25.
Deaux .. K., & Lewis, L. L. (1984). The structure of gender ' stereotypes: Interrelationships among compoe~ts and gender label. Journal of Personality and Soclal Psychology, 46, 991-1004.
P.89
Dwy er ,. C. A. ( 1974). In flu en c e 0 f chi 1 dr en I s sex r 0 1 e standards on reading and arithmetic achievement. Journal of Educational Psychology .. 66. 811-816.
Eccles .. J. (1987). Gender roles and achievement patterns: an expectancy value perspective. In Reinisch .. J.M ... Rosenhlum .. L.A. & Sanders .. S.A. (Ed.) .. Masculinity / femininity: Basic perspective. New York .. Oxford University Press.
Eccles .. J. (1985). Sex differences in achievement patterns. In T.B. Sonderegger (Ed.) .. Psychology and gender ~ Nebraska symposium on motivation .. 1984. , Lincoln: University of Nebraska Press.
Eccles (Parsons) .. J ... Adler, T. F ... Futterman .. R., Goff .. S. B., Kaczala .. C. M ... Meece, J. L., & Midgley, C. (1983). Expectations .. values and academic behaviours. In J.T. Spence (Ed.) .. Perspective on achievement and achievement motivation. San Francisco: W.H. Freeman.
Eccles (Parsons) J ... Adles .. T. & Kaczala .. C~M. (1982). Socialization of achievement attitudes and beliefs: Parental influences. Child Development, 53, 310-321.
Eccles (Parsons) J., Adler, T ... & Meece .. J. L. (1984). Sex differences in achievement: a test of alternate theories. Journal of Personality and Social Psychology, 46 .. 26-43.
Eccles. J ... & Hoffman, L. W. (1984). Sexroles~ socialization. and occupational behavior. In H~W. Stevenson & A.E. Siegel (Eds.), Research in child development and social police (Vol.l). Chicago: University of 'Chicago Press.
EcclesCParsons),J ... Ruhle .. D.N .. Hodges,K.L ... & Small,A.W. (1976). Cognitive-developmental factors in emerqinq sex differences in achievement-related expectancies. Journal of Social Issues, 32, 47-61.
Ehindero, O~ J. (1986). Correlates of Physics ' achievement: the role of gender and non-induced student expectations. The Journal of Experimental Education , , 54, 189-192.
Feather .. N. T. (1984). Masculinity, femininity, psychological androgyny .. and the structure of values. Journal of Personality and Social Psychology .. 47, 604-620.
Feather .. N. T. (1988). Values, valences and course enrollment: Testing the role of personal values within an expectancy-valence framework. Journal of Educational ,Psychology, 80 (3) ~ 381-391.
Feingold .. A. (1988). Cognitive gender differences are disappearing. American Psychologist, 43 .. ?5~103.
Fenn~ma, E. & Sherman, J. (1977). Sex-related dlfterences in mathematics achievement, spatial visualization and affective factors. American Educational
P.90
Research Journal~ 14 ~ 51-71. Hall~J.A. & Taylor,M~C. (1985). Psychological Androgyny
and the Masculinity X Femininity Interaction. Journal of Personality and Social Psychology. 49~ 429-435.
Handley .. H. M.& Morse~ L. W. (1984). Two-year study . relating adolescents·self-concept and gender role
perceotions to achievement and attitudes toward science. National Association for Research in Science Teaching .. 599-607
Helmreich~ R. L.~ Spence, J. T. & Holahan, C. K. (1979). Psychological androgyny and sex role flexibility': A test of two hypotheses. Journal of Personality· and Social Psychology, 37, 1631-1644.
Huston~ A. C. (1983). Sex-typing. In P.Mussen & E.M. Hetherington (Eds.) .. Hanndbook of child psychology. Vol. 4. New York: Wiley. Hyde~ J. S. (1981). How large are cognitvie gender differences? American Psychologist, 36, 892-901.
Hyde,J.S. (1981). How large are cognitive gender differences? American Psycholoaist~ 36(8), 892-901.
Hyde .. J. S., Fennama~ E. & Lamon .. S. J. (1990). Gender differences in mathematics performance: A metaanalysis. Psycholoaical Bulletin, 107, 139-155.
Hyde, J. S. & Linn .. M. C. (1988). Gender differences in verbal ability: A meta-analysis.Psychological Bulletin" 104. 533-569.
Kagan .. J. (1964). The child1s sex role classification of school objects. Child Development .. 35. 1051-1056.
Kagan~ J. (1964). Acquisituon 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 Foundtion .. 1964.
Kaczala" C. M. (1983). Sex role identity and its effect on achievement attitudes and behaviors. Unpublished doctoral disseration, University doctoral ' dissertation, University of Michigan" Ann Arbor.
Kelly, A. (1988). Sex stereotypes and school science: a three year follow-up. Educational Studies .. 14, 151-163.
Kelly, A. & Smail .. B. (1986). Sex stereotypes and attitudes to science among eleven-year-old children. British Journal of Educational Psychology" 56, 158-168.
Keyes .. S . . (1980) .. Sex differencs in cognitive ab,il~ties , and sex-role stereotypes in Hong Kong. UnpunlIshed
doctoral dissertation. University doctorial . dissertation, Harvard University.
Keyes, S. (1984). Measuring sex-r~le stereotypes: Attitudes amonq Hong Kong chInese adolescents and
P.91
the development of the chinese sex-role inventory. Sex Roles .. 10 .. 129-140.
Kimball .. M.M. (1989). ·A New Perspective on Women's Math Achievement. Psychological Bulletin e 105. 198-214.
Lau# S. (1989). Sex role orientation and domains of selfesteem. Sex Roles .. 21. 411-418.
Lau .. S. & Wong~. Y. (in-Press). Value and sex role orientation of Chinese adolescents. International Journal of Psychology. 26 .
Liben,L.S . .. & Signorella. M.L. (Eds.). (1987). Children ) s Gender Schemata. San Francisco: - Jossey-Bass.
Maccoby. E. E .. & Jacklin, C. N. (1974). Psychology of sex differences. Stanford: Standford University Press.
Marsh. H. W. (1987). Masculinity. femininity and androgyny: Their relations to multiple dimensions of self-concept. Multivariate behavioural research. 22, 91-118.
Marsh, H. W. (1989a). Age and Sex effects in Multiple dimensions of self-concept: Preadolescence to early adulthood. Journal of educational psychology. 81 ll.l .. 417-430.
Marsh, H. W. (1989b). Sex differences in the devekopment of verbal and mathematical constructs: The High School and Beyond study. American Educational Research Journal .. 26 (2), 191-225.
Marsh. H. W .. Byrne. B. M. & Shavelson. R. J. (1988). A multifaceted academic self-concept: Its hierarchical structure and its relation to academic achievement. Journal of Educational Psychology. 80 , 366-380.
Marsh. H. W .. Parket, J., & Barnes. J. (1985). Multidimensional ,adolescent self-concepts: Their relationship to age. sex and academic measures. American Educational Research Council. 22,422-444.
Meece. J. L ... Eccles. J .. Kaczala, C. M ... Goff. S. B. & Futt~rman. R. (1982). Sex differences in math achievement: Toward a model of academic choice. Psychological Bulletin. 91, 324-348.
Meece. J. L .. Wiafield. A. & Eccles. J. S. (1990). Predictors of Math Anxiety and its influence on Younq Adolescents' Course Enrollment Intentions and Performance in Mathematics. Journal of Educational Psychology. 82 .. 60-70.
Meyer. M. R. (1985). The prediction of mathematics achievement and participation for females and males: a longitudinal study of affective variables. Unpublished doctoral disserat~on. ~niversi~y . doctoral dissertation. the Unlverslty of WlsconSln-Madison.
Myers D.G. (1988). Social Psychology (2nd ed.). McGraw-Hi 11 .
P.92
Nash, S. R. C. (1974). Conceptions and concomitants of sex-role stereotypincr. Unpublished doctoral -disseration. ' University doctoral dissertation, Columbia University .
Nash~ S. C. (1979). Sex rore as a mediator of intellectual functioning. In M. A. Witting & A.C. Petersen (eds.), Sex-related differences in cognitive functioning ..!.. Developmental issuess. New York: Academic Press.
Nicholls, J. G. (1975). Causal attributions and other achievement-related cognitions: Effects of task outcomes, attainment value, and sex. Journal of Personality and Social Psychology, 31 , 379-38~
Orlofsky, J. L. (1981). Relationship between sex role attitudes and personality traits and sex role behavior scale-I. a new measure of masculine and feminine role behaviors and interests. Journal of Personality and Social Psychology, 40 , 927-940.
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.
Paulsen. K. & Johnson_M. (1983). Sex role attitudes and mathematical ability in 4th-, 8th-. and 11th-grade students from a high socioeconomic area. Develoomental Psycholoay ,. 19 , 210-214.
Ruble.D.N. & Ruble.T.L. (1982). Sex stereotypes. In A.G.Miller (Ed.), In the eye if ' the beholder: Contemporary issues in stereotypes (pp. 188-251). New York: Praeger.
Sherman, J. (1978). Social determainants of sex-related cognitive differences. In Sherman. J., Sex-related cognitive differences: An essay on theory and evidence. Springfield, Ill.: Charles C Thomas.
Sherman, J. (1980). Mathematics. spatial visualization. and related factors: Changes in girls and boys. grades 8-11. Journal of Educational Psychology. 72 476-482.
Sherman. J. & Fennema. E. (1977). The study of mathematics by high school girls and boys: Related variables. American Educational Research Journal, 14 . 159-168.
Signorella. M. L .. & Jamison. W. (1986). Masculinity. femininity. androgyny and cognitive performance: A meta-analysis . Psychological Bulletin, 100 ~. 207-228.
Signorella. M.L .. & Jamison. W. & Krupa. M. H. (1989). Predicting spatial performance form gender , stereotyping in activity preferences and in selfconcept. Deve 1 opmenta 1 Psycho 1 Ogy. 25 ,. 89-95.
Smead, V., & Chase, C. (1981). Student expectations as P.93
they relate to achievement in 8th qrade mathematics. Journal of Educational Research, 75, 115-120.
Spence, J. L. (1984). Masculinity, femininity, and gender related traits: A conceptual analysis and critique of current research. In B.A. Maker & W.B. Maker (eds.), Progress in Experimental Personality Research, Vol 13. New York: Academic Press.
Spence J. T. (1985). Gender ident i ty and i ts ilnp 1 i ca t ions for the concepts of masculinity and femininity. In T . B . Sonderregger (Ed.). Nebraska sympos i un) on motivation: Psychology and gender. (pp. 59-95). Lincoln: University of Nebraska Press.
Spence .. J. T. & Helmreich, R. L. (1978). Mascultinity and femininity: their psychological dimensions. correlates, and antecedents. Austin: University of Texas Press.
Stein, A. H. (1971). The effects of sex-role standards for achievement and sex-role preference on three determinants of achievement motivation. Developmental Psychology, 4 , 219~231.
Stein, A. H. & Bailey, M. M. (1973). The socialization of achievement orientation in females. Psychological Bulletin, 80, 345-366.
S t e in, A. H., Po h 1 y, S. R . & Mu ell er. E. ( 1 971 ) . Th e influence of masculine, feminine and neutral tasks on children's achievement behavior, expectancies of success. and attainment values. Child Development, 42, 195~207.
Stein, A. H. & Smithells. T. (1969). Age and sex differences in children's sex-role standards about achievement. Developmental Psychology, 1. 252-259.
Taylor,M.C. & Hall J.A. (1982). Psychological Androgyny: Theories, Methods and Conclusions. Psychological Bulletin ,92. 347-366. .
Tse, C. M. (1988). Sex role orientation and students' perceptions of the purposes of education. Unpubli'shed master dissertation. The Chinese University of Hong Kong, Graduate School, Division of Education.
Whitley,B.E. (1983). Sex role orientation and selfesteem:A critical metaanalytic review. Journal of Personality and Social Psychology, 44, 756-778.
Wong, K. Y. (1987). Adolescents'conceDtion of success: A personal construct approach .. Unpu~lish~d master dissertation. The Chinese Unlverslty ot Hong Kong, Graduate School,Division of Education.
P.94
CUHK L; bra r; es
1111111111111111111111111111111111111111111111111111111
000360265
