Analyzing the Gender Gap in Math Achievement: Evidence from a Large-Scale US Sample

Abstract

The US portion of the Program for International Student Assessment (PISA) 2003 student questionnaire comprising of 4,733 observations was used in a multiple regression framework to predict math achievement from demographic variables, such as gender, race, and socioeconomic status, and two student-specific measures of perception, math anxiety and math self-efficacy, while controlling for age, psychological disability, and language proficiency. The results confirmed the existence of persistent racial and socioeconomic gaps in math achievement, showed that both anxiety and self-efficacy contribute significantly towards explaining variation in math achievement, and provided evidence that the gender gap disappears once important predictors of math achievement, such as math-specific self-efficacy and anxiety, are controlled for.