Gender Differences in a Structural Model of Mathematics Achievement

The present study examined the expectancy-value model of math achievement using TIMSS 1999 (Third International Mathematics & Science Study) data. The model, originally developed by Eccles and her colleagues (e.g., Eccles et al., 1983), was suggested as a general theoretical model to explain sex differences in mathematics achievement behaviors. However, literature has shown mixed evidence about the validity of the model, especially when achievement behaviors were defined as mathematics achievement (as measured standardized math scores). In this study, a structural equation model based on the expectancy-value framework was estimated using 239 U.S. eighth graders. Sex differences in the structural model across each gender group were further tested using multi-sample analysis technique, as suggested by Joreskog and Sorbom (1996). Results of the study showed that mathematics values did not significantly influence mathematics achievement, whereas there was a significant direct effect between mathematics self-efficacy and mathematics achievement both for boys and girls. It was also found that the value-expectancy model of mathematics achievement, as utilized in the current study, was invariant across boys and girls.