Key points are not available for this paper at this time.
Longitudinal causal models of growth in math and English constructs (school grades, standardized tests, academic self-concept, affect and coursework selection) were based on three waves of data from the large (N = 24,599), nationally representative National Education Longitudinal Study of 1988. Math and English self-concepts had significant path coefficients leading to subsequent school grades, coursework selection, and standardized test scores. Unlike previous studies that did not consider math and English constructs in the same model, we found these relations to be very domain specific (e.g., there were significant positive paths from math self concept to subsequent math outcomes but not to subsequent English outcomes). Girls had higher scores for all English constructs and math school grades, but they had lower math self-concepts. Whereas similar studies conducted over the past 20 years found diminishing gender differences, these data show relative gains for girls in achievement and coursework selection for both mathematics and English. Path coefficients relating prior math and English constructs to subsequent outcomes, however, were similar for boys and girls. Hence, the extreme domain specificity of relations between prior self-concept and subsequent outcomes was similar for boys and girls.
Marsh et al. (Thu,) studied this question.