Gender Differences in the Development of English and Maths Constructs:
Longitudinal Models of Academic Self-Concept and Achievement
Alexander Seeshing Yeung and Herbert W. Marsh
University of Western Sydney, Macarthur
Paper presented at the annual conference of the Australian Association
for Research in Education in Brisbane, Australia, 30 November to 4 December 1997
Abstract
Gender differences in the development of English and maths constructs
(academic self-concept, academic affect, school grades, standardized
test scores, and coursework selection) were examined using three waves
of data from the large (N = 24,599) nationally representative NELS88
database of the U.S.A. Academic self-concept and academic affects had
significant effects on subsequent school grades, standardized test
scores and coursework selection, and these effects were domain specific
in that English self-concept had positive effects on subsequent verbal
outcomes and maths had positive effects on subsequent maths outcomes.
Girls had higher scores for English constructs and maths school grades,
but lower maths self-concept and affect. In contrast to the gender
stereotypic model, relations between prior English and maths constructs
and subsequent English and maths constructs were similar for boys and
girls, and no evidence of gender differences in the development of
either construct was found.
This study examines relations between growth in academic self-concept
and academic achievement, and gender differences in the development of
maths and English constructs over the adolescent years. More
specifically, we examine whether changes in academic self-concept
leading to changes in academic achievement in English and maths vary
according to gender. First, we focus on the multidimensionality of
self-concept that is based on the Shavelson, Hubner, and Stanton (1976)
model of self-concept. Then, we present longitudinal structural
equation models (SEMs) that combine the causal ordering of academic
self-concept and achievement, the domain specificity of these
relations, and gender differences in the development of maths and
verbal constructs. The data for this study come from the National
Education Longitudinal Survey of 1988 (NELS88), a multiwave,
longitudinal study consisting of a large (N = 24,599), nationally
representative sample of US students who were in 8th grade in 1988 with
follow-up data collections in 1990 and 1992 (Ingles, Scott, Lindmark,
Frankel, Myers, & Wu, 1992). A selection of Self Description
Questionnaire (SDQII) items included as part of the NELS88 study
(Marsh, 1990c, 1994) is an important component of the present
investigation.
Relations Between Academic Self-concept and Academic Achievement
Shavelson et al. (1976) provided a theoretical definition and model of
self-concept (see review by Marsh & Hattie, 1996). In the Shavelson et
al. model, academic self-concept is one component of overall
self-concept, and it is divided into self-concepts in particular
content areas. Support for the Shavelson et al. model requires academic
achievement to be more highly correlated with academic components of
self-concept than with nonacademic components of self-concept, and that
academic achievement in particular domains should be more highly
correlated with academic self-concepts in the matching domain (e.g.,
maths achievement and Maths self-concept) than self-concepts in
non-matching domains. Marsh (1993; also see Byrne, 1996) summarized a
large body of research in support of the domain specificity of academic
self-concept. For example, Marsh, Byrne, and Shavelson (1988) found
that correlations between Maths and English self-concepts were close to
zero, that maths achievement was substantially correlated with Maths
self-concept but not English self-concept, and that English achievement
is substantially correlated with English self-concept but not Maths
self-concept. Whereas such results provided strong support for the
multidimensionality of self-concept, the small sizes of correlations
actually observed between Maths and English self-concepts implied that
any hierarchical structure of academic self-concept must be much weaker
than anticipated. Such complications led to the Marsh/Shavelson
revision of the original Shavelson et al. model and the development of
the internal/external frame of reference (I/E) model (Marsh, 1986,
1993; Marsh, Byrne, & Shavelson, 1988; Marsh & Shavelson, 1985).
The Internal/External Frame of Reference (I/E) Model
According to the I/E model, academic self-concept in a particular
school subject is formed in relation to an external reference in which
students compare their self-perceived performances in a particular
school subject with the perceived performances of other students in the
same school subject and an internal reference in which students compare
their performances in the particular school subject with their own
performances in other school subjects. Hence, students may have a
favorable Maths self-concept if maths is their best school subject
even if they are not particularly good at maths relative to other
students. The joint operation of these two processes is consistent with
the near-zero correlations between Maths and English self-concepts.
For example, whereas the external comparison process predicts that good
maths skills lead to higher Maths self-concept and that good English
skills lead to higher English self-concept, the internal comparison
process, however, predicts that good maths skills should lead to lower
English self-concept and good English skills should lead to lower Maths
self-concept. In the path model used to test this prediction, the
paths leading from maths achievement to English self-concept and from
English achievement to Maths self-concept are predicted to be negative
(Marsh, 1986, 1990b; Marsh, Byrne, & Shavelson, 1988).
Causal Modeling Studies of Academic Self-concept and Achievement
Self-concept and academic achievement are typically assumed to be
related, and a positive self-concept may foster academic striving
behaviors (e.g. academic choice) that can enhance academic achievement
(Marsh & Craven, 1997). Following the Shavelson et al. (1976) model and
related theoretical perspectives such as social cognition theory
(Bandura, 1986), attribution theory (Weiner, 1992), and particularly
expectancy-value theories (Eccles, 1987a, 1987b; Eccles, Adler, &
Meece, 1984, Eccles, Wigfield, Harold, & Blumenfeld, 1993; Meece,
Parsons, Kaczala, Goff, & Futterman, 1982; Parsons & Ruble, 1977;
Wigfield, 1994; Wigfield, Eccles, MacIver, Reuman, & Midgley, 1991),
Marsh (1990b, 1993) hypothesized that academic self-concept in
particular school subjects influences subsequent task choice,
motivation, sustained effort, and persistence, which in turn lead to
improved academic achievement, academic choice. and subsequent academic
self-concept. Marsh and Yeung (in press) evaluated SEMs of paths
leading from prior school grades and self-concept in specific school
subjects to subsequent coursework selection. Self-concepts were
significantly related to course selection, but school grades did not
contribute consistently to course selection beyond the effects of
self-concept.
Byrne (1984, 1996) noted that much of the interest in the
self-concept/achievement relation stems from the belief that academic
self-concept has motivational properties that will affect subsequent
academic achievement. Marsh (1990a) reviewed three studies that
appeared to be methodologically adequate in addressing the issue using
a CFA approach (Byrne, 1986; Newman, 1984; Shavelson & Bolus, 1982;
also see Marsh, 1988) and suggested that the results seemed to be
consistent with his (Marsh, 1987) earlier suggestion that the effect of
prior academic self-concept on subsequent achievement is likely to be
substantial. Marsh (1990a) tested the causal ordering of academic
self-concept and academic achievement with four waves of data (last 3
years of high school and one year after graduation) from a large,
nationally representative American sample of boys and found some
support for a "reciprocal effects" model in which prior academic
self-concept affected subsequent achievement and prior achievement
affected subsequent self-concept.
Marsh and Yeung (1997; also see Byrne, 1996) extended this earlier
review to include new research (e.g., Helmke & van Aken, 1995; Skaalvik
& Hagtvet, 1990) and found reasonably consistent support for the
reciprocal effects model (but also see Shavelson & Bolus, 1982; Newman,
1984; Byrne, 1986). In the Marsh and Yeung (1997) study of academic
self-concept, school grades, and teacher ratings of achievement in
English, maths, and science in each of three years, although the paths
leading from prior achievement to subsequent academic self-concept
tended to be somewhat larger and more systematic, there was clear
support for both sets of paths.
Gender Differences
Research conducted in the 1970s or earlier provided a reasonably
consistent picture of gender differences in maths and English
achievement-related constructs. For example Maccoby (1966; Maccoby &
Jacklin, 1974) reported gender-stereotypic differences favouring boys
in maths and girls in English. In large survey studies (e.g., Hilton &
Berglund, 1974; Wise, 1985; also see Fennema & Sherman, 1977), gender
differences in maths achievement were found to increase in size during
high school years. Girl's poorer performance in maths was typically
attributed to gender-stereotypic socialisation patterns (e.g., Brophy,
1985; Eccles, 1987a; 1987b; Eccles & Blumfeld, 1985; Fennema & Sherman,
1977; Meece et al., 1982). Brophy, in particular, emphasised that
gender differences in academic achievement were declining and may
eventually disappear altogether. In recent Australian research,
however, MacCann (1995) analyzed trends in gender differences over a
decade in performances on high school certificate tests and found a
steady increase in the performances of girls relative to boys that was
consistent across different school subjects, geographical regions, and
types of schools, leading educators to propose special programs to
improve boys' educational outcomes.
Early reviews (e.g., Wylie, 1979; Maccoby & Jacklin, 1974) of gender
differences in self-concept have focused primarily on global or total
scores (e.g., general self-concept or self-esteem) and reported little
or no gender differences. Basow (1986) reported that boys tended to
have higher self-esteem than girls. More recently, Feingold (1994)
compared results from three meta-analyses of gender differences in
personality variables, and each demonstrated small differences in
self-esteem favouring boys. Wylie (1979), however, proposed that small
differences in global self-concept may reflect larger,
counter-balancing gender differences in specific components of
self-concept, a proposal that has been supported by a wide variety of
subsequent self-concept-related research (e.g., Basow, 1986; Brush,
1980; Crain, 1996; Eccles, 1987a, 1987b; Eccles & Blumfeld, 1985;
Eccles et al, 1984, Eccles et al., 1993; Fennema & Sherman, 1977;
Marsh, 1990b, 1993; Meece et al., 1982; Wigfield, 1994).
Based on normative archive responses to the three SDQ instruments,
Marsh (1989) reported statistically significant but small gender
differences in most SDQ scales, some favouring girls but more favouring
boys. Total self-concept scores favored boys, though this gender
difference explained only 1% of the variance. The gender differences in
self-concept are broadly consistent with gender stereotypes, but the
small gender differences suggested, perhaps, that this influence on
self-concept is diminishing.
Developmental Models of Gender Differences in Academic Self-concept
In their classic review of gender differences in maths constructs,
Meece et al. (1982) reported few gender differences in Maths
self-concept for elementary school children, but consistent differences
in favour of boys for junior high and high school students. In
subsequent research, Eccles (1987a, 1987b) reported stereotypic gender
differences in maths and verbal achievement that emerged during junior
high and high school years. However, Hyde, Fennema, and Lamon (1990)
found in their meta-analysis very little gender difference in maths
achievement except, perhaps, for differences in problem solving that
emerged in high school and college, and in highly selective samples.
According to Eccles stereotypic gender differences in maths and verbal
areas emerge during early adolescence and grow larger during the
adolescent years.
Ethington and Wolfle (1986), using the nationally representative High
School and Beyond data of the US, reported that higher prior maths
ability and more positive maths affect were related to greater
increases in maths achievement for high school boys than for girls,
whereas the negative relations of prior verbal ability on subsequent
maths affect was more negative for girls than for boys. In further path
analyses of data from the published Ethington and Wolfle research,
Marsh and Yeung (in press) reported that paths leading from maths
affect to maths coursework, after controlling for prior maths and
verbal ability, was significant for both boys and girls. Ethington
(1991) presented path analyses relating maths school grades,
self-concept, and other variables to coursework selection intentions.
Maths self-concept was more highly correlated with intentions than was
prior achievement for both boys and girls. In a further path analysis
of data from her published study based on only prior achievement,
self-concept, and intentions, Marsh and Yeung (in press) reported that
paths from self-concept to intentions were significant and larger than
paths from prior achievement to intentions.
Method
NELS88 Sample and Measures
The present investigation is based on selected variables (Table 1) from
the commercially available NELS88 database (Ingles et al., 1992). The
base year data consisted of responses by the 24,599 students. However,
because the analyses were based on covariance matrices constructed with
pairwise deletion for missing data for T1, T2 and T3 variables, the
weighted N used in the actual analyses was 4,775.
Structural model
In the path model there are 17 latent constructs: gender and 8 pairs of
maths and English constructs collected at Times 1, 2, or 3 representing
school grades (T1, T2, T3), standardized test scores (T1, T2, T3),
affect (T1), or self-concept (T2). The model tested is a "full forward"
model in which path coefficients from each construct lead to all
constructs following it in the proposed causal ordering of the
constructs. There are four indicators for each self-concept construct,
three indicators for each affect construct, and a single indicator for
each of the remaining constructs. The ordering of variables is
determined primarily on the basis of time: following gender, all T1
variables preceded T2 variables, which preceded T3 variables. Within
each wave, school grades refer to previously earned grades (e.g., for
8th grade students, grades refer to grades earned since 6th grade), and
so these come first within each wave. For present purposes,
standardized achievement tests were posited to preceded academic affect
variables (T1) and academic self-concept (T2). Finally, for each of
these major constructs, there were parallel measures for English and
maths for which no causal ordering was proposed (these relations were
merely represented as factor covariances or residual covariances).
Table 1
Summary Statistics For Boys and Girls
_
Time/Factor Boys Girls Effect Size
M SD M SD r d
T1 English grade 3.78 .98 4.17 .87 .20* .41
T1 maths grade 3.92 1.01 4.01 .97 .05* .10
T1 English test 25.83 8.56 27.78 8.53 .11* .22
T1 maths tests 35.97 12.04 35.74 11.65 -.01 -.02
T1 English affect 2.48 .87 2.68 .83 .12* .24
T1 English affect 3.08 .72 3.12 .72 .03* .06
T1 English affect 3.03 .81 3.23 .73 .13* .26
T1 maths affect 2.62 .90 2.56 .87 -.03* -.06
T1 maths affect 3.05 .75 2.99 .80 -.04* -.08
T1 maths affect 3.33 .76 3.24 .75 -.06* -.12
T2 English grade 5.46 1.82 6.16 1.66 .20* .41
T2 maths grade 5.50 1.89 5.64 1.86 .04* .08
T2 English test 29.36 10.22 31.35 9.62 .10* .20
T2 maths test 43.58 14.18 43.28 13.51 -.01 -.02
T2English self-concept 4.43 1.32 4.69 1.25 .10* .20
T2English self-concept 3.67 1.61 4.09 1.63 .13* .26
T2English self-concept 4.21 1.49 4.61 1.39 .14* .28
T2English self-concept 4.88 1.40 5.21 1.23 .12* .24
T2 Maths self-concept 4.09 1.72 3.68 1.83 -.12* -.24
T2 Maths self-concept 4.19 1.62 3.87 1.73 -.10* -.20
T2 Maths self-concept 4.26 1.63 4.05 1.72 -.06* -.12
T2 Maths self-concept 4.44 1.58 4.22 1.67 -.07* -.14
T3 English grade 6.37 2.69 7.57 2.62 .11* .22
T3 maths grade 6.12 2.71 6.69 2.66 .11* .22
T3 English test 31.70 10.43 33.97 9.73 .11* .22
T3 maths test 48.61 14.72 47.44 14.07 -.04* -.08
* p<.05 (two-tailed tests of statistical significance for gender
differences )
Results
Mean Gender Differences
Consistent with earlier research, gender differences for English
variables were all statistically significant and favored girls,
whereas those for maths variables were mixed (Table 1). For maths,
girls earned higher school grades, but had significantly lower scores
for T1 maths affect items and T2 maths self-concept items. The only
other statistically significant gender difference favouring boys was
for T3 maths test scores. Across all variables, there were consistent
patterns whereby gender differences favouring girls for English
variables were more positive than the corresponding gender differences
for maths variables, and gender differences favouring girls were larger
for school grades than for standardized test scores. The most directly
comparable scores from T1, T2, and T3 were the standardized test scores
and these showed reasonably stable gender differences for both maths
and English. The gender differences in school grades were quite
similar over time. However it is also important to note that none of
the differences was large. The only differences to exceed |d| = .3 were
the differences favouring girls in English grades.
Longitudinal Causal Models
The longitudinal causal model provided a very good fit to the data (TLI
= .968) based on conventional standards (e.g., TLI > .9). Factor
loadings were consistently substantial for English self-concept (.77,
.80, .87, .64) and Maths self-concept (.88, .86, .88, .62), but were
somewhat lower for English affect (.79, .26, .49) and Maths affect
(.75, .28, .41).
Table 2
Path Coefficients
Sex Egrd1 Mgrd1 Etst1 Mtst1 Eaff1 Maff1 Egrd2 Mgrd2 Etst2
Mtst2 Eslf2 Mslf2 Egrd3 Mgrd3 Etst3 Mtst3
Sex ---
Egrd1 .20* ---
Mgrd1 .04* 0 ---
Etst1 .04* .30* .16* ---
Mtst1 -.08* .28* .30* 0 ---
Eaff1 .11* .42* -.05* -.02 -.12* ---
Maff1 -.05* -.07* .54* -.18* .05* 0 ---
Egrd2 .10* .26* .16* .17* .09* .08* -.01 ---
Mgrd2 .02 .11* .21* -.04* .29* .01 .09* 0 ---
Etst2 .01 .02 .00 .60* .23* .02 -.04* .07* 0 ---
Mtst2 -.03* .02* .07* .14* .71* -.01 .00 .02* .07* 0
---
Eslf2 -.02 .02 -.02 .04* -.07* .17* .05* .66* -.14* .12*
.04 ---
Mslf2 -.06* -.04* .03* -.07* .06* -.01 .24* -.14* .64* -.07*
.19* 0 ---
Egrd3 .10* .13* .07* .02 .06* -.04* -.06* .36* .15* .08*
.20* .05* -.05* ---
Mgrd3 .07* .09* .05* -.01 .09* -.03 -.07* .18* .31* .04*
.23* -.09* .17* 0 ---
Etst3 .03* -.01 .00 .16* -.03 -.01 -.03* -.01 .02 .51*
.21* .03* -.02 .06* -.01 ---
Mtst3 -.05* .01 .02* .02* .06* -.03* .02 .01 -.01 .07*
.72* .00 .06* .05* .04* 0 ---
Residual Factor Variance/Covariances
Var 1.00 .96 1.00 .84 .77 .83 .73 .69 .76 .32
.21 .46 .39 .36 .41 .29 .14
Covar .37 .53 .22 .22 .10
.07 .21 .07
Note. Egrd1 = T1 English grade; Mgrd1 = T1 maths grade; Etst1 = T1
English test; Mtst1 = T1 maths test; Eaff1 = T1 English Affect; Maff1 =
T1 Maths affect; Egrd2 = T2 English grade; Mgrd2 = T2 maths grade;
Etst2 = T2 English test; Mtst2 = T2 maths test; Eslf2 = T2 English
self-concept; Mslf2 = T2 maths self-concept; Egrd3 = T3 English grade;
Mgrd3 = T3 maths grade; Etst3 = T3 English test; Mtst3 = T3 maths test.
(2 (df) for this model is 1652.77 (195) and for the corresponding null
model is 82287.13 (351). For this solution, TLI = .968 and RNI = .982.
* p < .05
Consistent with observations based on previous research with the I/E
model, English and maths test scores were very highly correlated (.71,
.76, .74) in each of the three waves, English and maths school grades
were substantially correlated (.38, .45, and .74), but correlations
were comparatively small between T1 maths and English affects (.21) and
particularly T2 English and Maths self-concepts (.14). This
demonstrated that students differentiated between Maths and English
self-concepts and between Maths and English affects to a much greater
extent than could be explained in terms of the substantially more
positive correlations between maths and English test scores and school
grades.
T1 English and Maths affects were substantially related to prior school
grades, but had surprising little relation with T1 test scores. The
path coefficients leading from school grades to matching academic
affects were substantial and positive, whereas those leading from
non-matching constructs (e.g., English grades to Maths affect) were
small and significantly negative. T1 academic affects also had small,
but significantly positive paths leading to matching T2 school grades
but non-significant paths leading to matching T2 test scores. T1
affects also had significant paths to matching T2 academic
self-concepts, beyond the indirect relations mediated by intervening
school grades and test scores. Furthermore, the relations between
academic affects and academic self-concept were very domain specific in
that path coefficients leading to matching domains (.17 and .24; Table
2) were substantially larger than the small or non-significant paths
leading to non-matching domains (.05 and -.01).
The largest paths leading to Maths and English self-concepts
respectively (Table 2) were matching T2 school grades, matching T1
academic affects, and matching T2 test scores. The pattern of path
coefficients was very domain specific in that paths leading from
non-matching domains tended to be small or, particularly for school
grades, even negative. The pattern of path coefficients leading from
school grades to academic self-concepts provided additional support for
predictions from the I/E model; paths for matching domains were
substantial and positive (.66 and .64), whereas those for non-matching
domains were significantly negative (-.14 and -.14).
Paths leading from T2 self-concepts to T3 school grades, and T3 test
scores in matching domains were all significantly positive. This
finding shows that prior self-concept is associated with improved
school grades and test scores even after controlling for the effects of
prior school grades and test scores. The results show the remarkably
strong domain specificity for T2 Maths and English self-concepts in a
way that was not possible to evaluate in previous research that only
considered a single component of academic self-concept. Whereas all
paths leading from Maths self-concept to T3 maths outcomes and from
English self-concept to T3 English outcomes are significant and
positive, all paths leading from Maths self-concept to T3 English
outcomes and from English self-concept to T3 maths outcomes are either
non-significant or significantly negative.
Summary
This investigation extended previous longitudinal studies of the causal
ordering of academic self-concept and achievement in different domains,
and the development of gender differences in these domains. The
critical findings were the gender differences, the relative sizes of
paths associated with school grades and test scores, the paths leading
from academic affect and particularly academic self-concept to
subsequent grades and test scores, the domain specificity of these
relations, and the support for predictions based on the I/E model.
Consistent with previous research, gender differences favored girls
for all English constructs and were mixed for maths constructs. Paths
leading from T2 academic self-concept to subsequent school grades, and
test scores were significantly positive. For both academic affects and
academic self-concepts, relations were stronger with school grades than
test scores. Consistent with the I/E model, the paths leading from
prior school grades to subsequent academic self-concept were very
domain specific in that the paths from grades to matching areas of
self-concept were substantial and positive, whereas the paths leading
from grades to non-matching areas were smaller and significantly
negative. A similar pattern based on smaller paths was evident for the
relations of prior school grades and academic affects.
Marsh (1987, 1990b, 1993; Wylie, 1979) hypothesized that academic
self-concepts should be more strongly related to school grades than
standardized tests because school grades are a more salient source of
feedback to students with motivational properties. Thus, academic
self-concept should be more positively correlated with school grades
than standardized test scores and paths leading from prior self-concept
to subsequent achievement should be larger for school grades than test
scores. The results of the present study support this prediction.
Previous longitudinal causal modeling studies (see review by Marsh &
Yeung, 1997) have typically considered either maths or a global
academic construct that did not distinguish between specific academic
domains. Marsh and Yeung did, however, note two studies that included
separate analyses of maths and English constructs within the same
study. In each case there were stronger, more systematic relations
between prior self-concept and subsequent achievement in maths than in
English. In the present investigation, there is some support for this
trend in path coefficients relating T2 Maths and English self-concept
and T3 outcomes. There is, however, not sufficient evidence to draw
firm conclusion based on the relatively small differences observed
here.
The results of this study found little support for gender stereotype or
differential socialisation models of gender differences in academic
outcomes. Although there were gender stereotypic differences in Maths
and English self-concepts, girls scored as well or better than boys on
other maths outcomes (school grades and, perhaps, test scores) and did
better than boys in all English constructs. There was, perhaps, some
support for the gender stereotype perspective in that the advantages
favouring girls were stronger for English constructs than for maths
constructs.
Whereas girls tended to outperform boys on school grades based on
self-reports and official school transcripts, girls seemed to fare
relatively less favorably on standardized test scores. Inspection of
the gender effects (Table 1) demonstrated that girls consistently
scored much better than boys on English grades but only slightly better
on English test scores, whereas they scored slightly better than boys
on maths grades but did slightly worse than boys on maths test scores.
These gender differences were consistent across the three waves of
data.
In conclusion, it is important to emphasize the importance of combining
within a single study the rigour of the longitudinal causal modeling
studies of academic achievement and academic self-concept, the
theoretical emphasis on the multidimensionality of academic
self-concept embodied in the Marsh/Shavelson and I/E models,
longitudinal studies of the development of gender differences in
English and maths constructs, methodological advances in the
application of SEM, and the evaluation of these issues with a very
large, nationally representative, longitudinal database. By bringing
these different approaches together within a single study, we
demonstrated good support for models in which there are significant
paths leading from academic self-concept to subsequent academic
outcomes -- particularly school grades, a remarkably strong domain
specificity of this pattern of relations, and the similarity of these
patterns of relations for boys and girls.
References
Bandura, A. (1986). Social foundations of thought and action: A social
cognitive theory. Englewood Cliffs, NJ: Prentice-Hall.
Basow, S. A. (1986). Gender stereotypes: Traditions and alternative
(2nd ed). Monterey, CA: Brooks/Cole.
Brophy, J. (1985). Interactions of male and female students with their
male and female teachers. In L. C. Wilkinson & C. B. Marrett (Eds.),
Gender influences in classroom interaction (pp. 115-142). New York:
Academic Press.
Brush, L. R. (1980). Encouraging girls in mathematics: The problem and
the solution. Boston: Abt Associates.
Byrne, B. M. (1984). The general/academic self-concept nomological
network: A review of construct validation research. Review of
Educational Research, 54, 427-456.
Byrne, B. M. (1986). Self-concept/academic achievement relations: An
investigation of dimensionality, stability, and causality. Canadian
Journal of Behavioural Science, 18, 173-186.
Byrne, B. M. (1996). Academic self-concept: Its structure, measurement,
and relation to academic achievement. In B. A. Bracken (Ed.), Handbook
of self-concept (pp. 287-316). New York: Wiley.
Crain, R. M. (1996). The influence of age, race, and gender on child
and adolescent multidimensional self-concept. In B. A. Bracken (Ed.),
Handbook of self-concept: Developmental, social, and clinical
considerations (pp. 395-420). New York: Wiley.
Eccles, J. S. (1987a). Gender roles and achievement patterns: An
expectancy value perspective. In J. M. Reinisch, L. A. Rosenblum, & S.
A. Sanders (Eds.), Masculinity/femininity: Basic perspectives (pp.
240-280). New York: Oxford University Press.
Eccles, J. S. (1987b). Gender roles and women's achievement-related
decisions. Psychology of Women Quarterly, 11, 135-172.
Eccles, J. S., Adler, T. F., & Meece, J. L. (1984). Sex differences in
achievement: A test of alternative theories. Journal of Personality and
Social Psychology, 46, 26-43.
Eccles, J. S., & Blumfeld, P. (1985). Classroom experiences and student
gender: Are there differences and do they matter. In L. C. Wildinson
and C. Marrett (Eds.), Gender influence in classroom interaction (pp.
79-114). Hillsdale, NJ: Erlbaum.
Eccles, J., Wigfield, A., Harold, R.D., & Blumfeld, P. (1993). Age and
gender differences in children's self- and task perceptions during
elementary school. Child Development, 64, 830-847.
Ethington, C. A. (1991). A test of a model of achievement behaviors.
American Educational Research Journal, 28, 155-172.
Ethington, C. A., & Wolfle (1986). A structural model of mathematics
achievement for men and women. American Educational Research Journal,
23, 65-75.
Feingold, A. (1994). Gender differences in personality: A
meta-analysis. Psychological Bulletin, 116, 429-456.
Fennema, E., & Sherman, J. (1977). Sex differences in mathematics
achievement, spatial visualization, and affective factors. American
Educational Research Journal, 14, 51-71.
Helmke, A., & van Aken, M. A. G. (1995). The causal ordering of
academic achievement and self-concept of ability during elementary
school: A longitudinal study. Journal of Educational Psychology, 87,
624-637.
Hilton, T. L., & Berglund, G. W. (1974). Sex differences in mathematics
achievement: A longitudinal study. Journal of Educational Research, 67,
231-237.
Hyde, J. S., Fennema, E., and Lamon, S. J. (1990). Gender differences
in math performance: A meta-analysis. Psychological Bulletin, 107,
139-155.
Ingles, S. J., Scott, L. A., Lindmark, J. T., Frankel, M. R., Myers, S.
L., & Wu, S. (1992) National education longitudinal study of 1988.
First follow-up: Student component data file user's manual (Vol. 1).
Washington, D.C.: U.S. Department of Education Office.
MacCann, R. (1995). A longitudinal study of sex differences in HSC and
SC. Australian College of Education New South Wales Chapter Newsletter
(June, 1995, Supplement). pp. 8-9.
Maccoby, E. E. (1966). The development of sex differences. Palo Alto,
CA: Stanford University Press.
Maccoby, E. E., & Jacklin, C. N. (1974). The psychology of sex
differences. Stanford: Stanford University Press.
Marsh, H. W. (1986). Verbal and math self-concepts: An
internal/external frame of reference model. American Educational
Research Journal, 23, 129-149.
Marsh, H. W. (1987). The big-fish-little-pond effect on academic
self-concept Journal of Educational Psychology, 79, 280-295.
Marsh, H. W. (1988). Causal effects of academic self-concept on
academic achievement: A reanalysis of Newman (1984). The Journal of
Experimental Education, 56 100-103.
Marsh, H. W. (1989). Age and sex effects in multiple dimensions of
self-concept: Preadolescence to Early-adulthood. Journal of Educational
Psychology, 81, 417-430.
Marsh, H. W. (1990a). The causal ordering of academic self-concept and
academic achievement: A multiwave, longitudinal panel analysis. Journal
of Educational Psychology, 82, 646-656.
Marsh, H. W. (1990b). A multidimensional, hierarchical self-concept:
Theoretical and empirical justification. Educational Psychology Review,
2, 77-171.
Marsh, H. W. (1990c). Self Description Questionnaire (SDQ) II: A
theoretical and empirical basis for the measurement of multiple
dimensions of adolescent self-concept: An interim test manual and a
research monograph. San Antonio, TX: The Psychological Corporation
(Republished in 1992, Publication Unit, Faculty of Education,
University of Western Sydney, Macarthur.).
Marsh, H. W. (1993). Academic self-concept: Theory measurement and
research. In J. Suls (Ed.), Psychological perspectives on the self
(Vol. 4, pp. 59-98). Hillsdale, NJ: Erlbaum.
Marsh, H. W. (1994). Using the National Longitudinal Study of 1988 to
evaluate theoretical models of self-concept: The Self-Description
Questionnaire. Journal of Educational Psychology, 86, 439-456.
Marsh, H. W., Byrne, B. M., & Shavelson, R. (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., & Craven, R. (1997). Academic self-concept: Beyond the
dustbowl. In G. Phye (Ed.), Handbook of classroom assessment: Learning,
achievement, and adjustment (pp. 131-198). Orlando, FL : Academic
Press.
Marsh, H. W., & Hattie, J. (1996). Theoretical perspectives on the
structure of self-concept. In B. A. Bracken (Ed.), Handbook of
self-concept (pp 38-90). New York: Wiley.
Marsh, H. W. & Shavelson, R. (1985) Self-concept: Its multifaceted,
hierarchical structure. Educational Psychologist, 20 107-125.
Marsh, H. W., & Yeung, A. S. (1997). Causal effects of academic
self-concept on academic achievement: Structural equation models of
longitudinal data. Journal of Educational Psychology, 89, 41-54
Marsh, H. W., & Yeung, A. S. (in press). Coursework selection:
Relations to academic self-concept and achievement. American
Educational Research Journal.
Meece, J. L., Parsons, J. E., Kaczala, C. M., Goff, S. B., &
Futterman, R. (1982). Sex differences in math achievement: Toward a
model of academic choice. Psychological Bulletin, 91, 324-348.
Newman, R. S. (1984). Achievement and self-evaluations in mathematics.
Journal of Educational Psychology, 76, 857-873.
Parsons, J. E., & Ruble, D. N. (1977). The development of
achievement-related expectancies. Child Development, 48, 1075-1079.
Shavelson, R. J., & Bolus, R. (1982). Self-concept: The interplay of
theory and methods. Journal of Educational Psychology, 74, 3-17.
Shavelson, R. J., Hubner, J. J., & Stanton, G. C. (1976). Self-concept:
Validation of construct interpretations. Review of Educational
Research, 46, 407-441.
Skaalvik, E. M., & Hagtvet, K. A. (1990). Academic achievement and
self-concept: An analysis of causal predominance in a developmental
perspective. Journal of Personality and Social Psychology, 58, 292-307.
Wigfield, A. (1994). Expectancy-value theory of achievement
motivation: A developmental perspective. Educational Psychology
Review, 6, 49-78.
Wigfield, A., Eccles, J. S., MacIver, D., Reuman, D. A., & Midgley, C.
(1991). Transitions during early adolescence: Changes in children's
domain specific self-perceptions and general self-esteem across the
transition to junior high school. Developmental Psychology, 27,
552-565.
Weiner, B. (1992). Human motivation: Metaphors , theories, and
research. Newbury Park, CA: Sage.
Wise, L. L. (1985). Project TALENT: Mathematics course participation in
the 1960s and its career consequences. In S. F. Chipman, L. R. Brush, &
D. M. Wilson (Eds.), Women and mathematics: Balancing the equation (pp.
25-58). Hillsdale, NJ: Erlbaum.
Wylie, R. C. (1979). The self-concept. (Vol. 2) Lincoln: University of
Nebraska Press.