Gender Gaps in Student Academic Achievement and Inequality

Abstract Results from international large-scale assessments, such as PISA surveys, suggest that boys do better in math and science, whereas girls do better in reading. How do gender gaps vary across subjects, when estimated simultaneously? Building on the work of Tsai, Smith, and Hauser (2017), we answer this question by applying a multilevel-MIMIC model that enables us to estimate gender gaps in two ways: gender differences in the effects of observed family and school factors on math, science, and reading scores; and the “adjusted” gender gaps in test scores across all three subjects after controlling for observables. We apply the model to 2012 PISA data of students aged 15–16 and enrolled in 9th or 10th grade in three East Asian (Japan, South Korea, and Taiwan) and three Western countries (USA, Germany, and the Czech Republic) that represent both similar and different types of school systems. Our findings indicate that the gender gap in math or science achievement in Western countries, favoring boys, does not necessarily apply to the East Asian countries examined here, while all three East Asian countries exhibit similar features of gender reading gaps in the 10th grade. There is evidence indicating that observed background and school factors impact boys’ and girls’ achievement in a similar way in USA, Japan, Korea, Taiwan, and the Czech Republic, but not in Germany. Overall, gender differences in family and school influences do not account for gender differences in academic achievement in any of the six countries.