Do Boys Benefit Disproportionately From "Good" Schools? How Gender Disparities in Achievement and Attainment Vary Across High Schools

• In Event: Promising Scholarship in Education: Dissertation Fellows and Their Research

Abstract

In recent years, gender disparities in academic success have sparked public debates, policy conversations, and academic research on the “rise of women” and the “trouble with boys” (Bertrand and Pan 2013; DiPrete and Buchmann 2013; Rosin 2012). To better understand the extent to which women are “rising” in relation to men in different educational contexts in the U.S., I examine how gender inequalities in educational outcomes vary across high schools and the types of schools, and school-based resources, related to greater gender inequality across outcomes.

A small body of research has explored school characteristics associated with gender differences in achievement. Using reading data from fifth-graders in Berlin, Legewie and DiPrete (2012) found that boys learn more in classrooms with higher average SES; these authors also documented a similar pattern using math and reading data from fourth-grade students in Chicago (Legewie and DiPrete 2011). In another study, Legewie and DiPrete (2014) found that, net of students’ characteristics prior to high school, schools with stronger math and science curricula have smaller gender differences in students’ interest in STEM careers. Additionally, Ma (2008) found that U.S. schools with greater teacher absenteeism and more principal-reported teacher shortages had larger gender differences in math and science favoring males. Overall, prior research provides some evidence that gender inequalities in achievement vary across schools, but we do not know much about the school characteristics associated with that variation or about how gender inequalities in attainment vary.

I use data from the Education Longitudinal Study of 2002 (ELS) to examine three outcomes: twelfth-grade math achievement, high school graduation, and college enrollment. To measure the extent of variation in gender inequalities across schools, I use multilevel random intercept, random slope models. Because the sample is nationally representative, the standard deviation of the gender slope measures variation in gender inequalities across U.S. high schools. To take advantage of the ELS’ rich but noisy measures of school resources, I employ latent class analysis (LCA), which offers an empirical method of clustering schools and exploring school resources in combination rather than isolation. First, I use LCA to measure schools’ levels of five types of resources: instructional, teacher, school physical resources, student-staff, and student-peer relationships. Then I create a multidimensional measure of school type that integrates the individual resource measures (see Figure 1). Finally, I use a slopes-as-outcomes approach to model the relation between schools’ levels of resources and variability in outcomes by gender across schools (see Figure 2).

Gender differences in math achievement, high school graduation, and college enrollment vary across schools (see Figures 3, 4, and 5). On average, schools with higher math achievement have larger gender inequalities favoring males (see Table 1); in contrast, schools with higher graduation and college enrollment rates have smaller gender inequalities in these outcomes (meaning that female students’ advantage relative to male students is reduced). Male students’ advantage in math is particularly large in schools with the most academically-oriented instructional resources, positive student-staff relationships, and academically-oriented peers (see Figures 6 and 7). Male students’ average disadvantage in high school graduation is erased in schools with the most positive student-staff relationships (see Table 2); in contrast, male students’ relative disadvantage in graduation is higher in less well-maintained but academically advantaged schools, schools with more experienced but less satisfied teachers, schools with the most physical resource problems, and schools with less positive student-staff relationships (and Figures 8 and 9). The results are not conclusive regarding what types of schools or resources are associated with the extent of gender inequalities in college enrollment (see Figure 10).

Improving student-staff relationships may be particularly beneficial for boys given that, for both math achievement and graduation, boys’ performance relative to girls’ increases more when attending schools with positive student-staff relationships. Currently, however, many reform efforts focus on instructional, rather than relational, resources (Spillane, Parise and Sherer 2011). Consistent with research by DiPrete and colleagues, I conclude that boys benefit disproportionately from “good” schools (DiPrete and Buchmann 2013; Legewie and DiPrete 2012). However, DiPrete and colleagues focused exclusively on outcomes where female students perform better on average, concluding that increasing school quality will decrease gender gaps across the board. In contrast, my findings suggest that increasing school quality may also exacerbate gender gaps for outcomes where male students perform better. In sum, I conclude that the extent to which “[e]ducational institutions still work as engines of gender inequality” (Barone 2011: 157) varies across high schools, and that schools may be able to improve gender equality across outcomes by changing their resource allocations.

Author

• Hannah K. Miller, University of Wisconsin - Madison