Association for Public Policy Analysis & Management

Several recent papers find that boys’ and girls’ academic and non-cognitive skills respond differently to human capital inputs. This is observed throughout adolescence, and repercussions appear to persist into the labor market.  Recent literature finds a socioeconomic status (SES) gender gap gradient, whereby boys’ academic achievement and behavior suffer more than girls from socioeconomic disadvantage (Autor et al. 2016; Bertrand and Pan 2013). The most disadvantaged boys have particularly diminished future employment prospects (Chetty 2016). Authors have found, moreover, that economically disadvantaged boys benefit more from better schools than girls both in terms of behavior and academic achievement (Autor et al. 2016). On the opposite end of the SES distribution are further interesting patterns. Results suggest that girls see lower returns to inputs at the high end of the SES spectrum (Fryer and Levitt 2010)—with larger gender disparities existing in private schools than in public schools. 

Using a nationally representative dataset, this paper extends these studies to explore how high school students’ academic achievement and self-perceptions of ability correspond with SES levels and student attendance at public or private schools. The High School Longitudinal Study (HSLS) collected information about students’ schools and asked students and their parents about future educational expectations. It asks the students to assess their math and science identity and to complete a quantitative test. The first wave was administered in 2009 to over 20,000 9^thgraders across 944 schools. The second wave followed up with the students in 2012, and, more recently, students were surveyed during their first year after high school.

Consistent with the previous studies, analysis finds evidence for the relationship between SES and math achievement. Boys have a stronger associated relationship between SES and math scores—however, this gender difference in the relationship looks more pronounced in private schools. The relationship between SES and math identification demonstrates gendered patterns as well. Boys have a stronger associated relationship between SES and math scores—however, this gender difference in the relationship looks more pronounced in private schools. Private school attending girls exhibit little relationship between SES and math identification, while private school attending boys exhibit a strong positive relationship between socioeconomic background and math identification.

The results here suggest that the most disadvantaged boys benefit by having more resources at home and enrolling in private schools. The seemingly most advantaged girls (those from high SES backgrounds that attend private schools) end up exhibiting a much lower affinity towards STEM fields than analogous boys. One might ask whether the gender differences in outcome might, in part, reflect differences in how a student’s gender influences parents’ decision to enroll him/her in private school. Preliminary evidence suggests that there are some notable demographic differences between low SES private school boys versus girls.