Gender differences in children's academic self-concept and achievement in math and reading

Abstract

The purpose of this study was to explore gender differences in self-concept and achievement in math and reading among elementary-aged students. Research questions were constructed to explore (a) within-group differences of boys and girls in terms of their math and reading self-concepts and achievement scores, (b) between-group differences of boys and girls in areas of math and reading self-concept and achievement, and (c) the predictive utility of gender, math and reading self-concept, and the interaction of gender and math and reading self-concept in relation to math and reading achievement.

The study used archival data of 104 students in the third through the sixth grades, enrolled in a rural elementary school in Arizona. Approximately half of participants were male and half were female. Data were collected via file reviews for gender, demographic, and Terra Nova reading and math achievement scores. Survey research procedures were also utilized via questionnaires to collect self-concept data, using math and reading composite self-concept variables from the Self-Description Questionnaire, developed by Marsh (1990).

Bivariate and multivariate statistical procedures were used within a correlational design, with t test, repeated-measures ANOVA (RM-ANOVA), and hierarchical multiple regression analysis (MRA) procedures. Results of dependent-samples t tests demonstrated no significant differences in math and reading self-concepts or achievement scores for boys. Among girls, the only significant mean-level difference was found in higher reported reading self-concept scores, than math self-concept scores. Girls did not, however, exhibit significantly higher reading than math achievement scores.

Results of RM-ANOVA procedures revealed no significant differences between boys and girls in the domains of reading and math achievement. Similarly, no significant differences were obtained between boys and girls in either domain (math or reading) of self-concept.

Significant multiple regression models were specified for both reading achievement and math achievement. The independent variables were gender and reading self-concept in the first model, and gender and math self-concept for the second model. Dependent variables consisted of matched areas of reading and math achievement.

Although both MRA models were significant, they accounted for only about 9% of the variance in math and reading achievement. The gender by self-concept interaction variables were not significant for either model (math or reading), indicating that the relationships between domain-specific self-concept and achievement were not moderated by gender.