“…At Time 1, participants reported on demographic characteristics, including their sex (0 ϭ girl, 1 ϭ boy), their age in years, and their racial/ethnic background (White, African American, Latina/o, Asian, American Indian or Alaska Native, other, and multiethnic). Total number of people in the household was used as a proxy for SES, with higher numbers of people indicating crowding and therefore serving as a proxy indicator of lower SES (Hardiman et al, 2007). Finally, as some students were in single-sex classrooms, we created a dummy variable for students' inclusion in any single-sex classrooms (0 ϭ no single-sex classrooms, 1 ϭ one or more single-sex classrooms).…”