Measuring students' school context exposures: A trajectory-based approach

Abstract

Studies of school effects on children's outcomes usually use single time-point measures. I argue that this approach fails to account for (1) age-based variation in children's sensitivity to their surroundings; (2) differential effects stemming from differences in the length of young people's exposures; and (3) moves between contexts and endogenous changes over time within them. To evaluate the merits of this argument, I specify and test a longitudinal model of school effects on children's academic performance. Drawing on recent advances in finite mixture modeling, I identify a series of distinct school context trajectories that extend across a substantial portion of respondents' elementary and secondary school years. I find that these trajectories vary significantly with respect to shape, with some students experiencing significant changes in their environments over time. I then show that students' trajectories of exposure are related to their 8th grade achievement, even after controlling for point-in-time measures of school context.

Keywords: Achievement; Longitudinal latent class analysis; School effects.