Classroom Peer Effects and Student Achievement

The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.

The text below is an excerpt from the complete document. Read the full report in PDF format.

Abstract

Using a unique longitudinal dataset from Florida, we analyze the impact of classroom peers on individual student performance. Focusing on the influence of peers' fixed characteristics on individual test score gains, we control simultaneously for student and teacher fixed effects. We find some sizable, significant peer effects within nonlinear models, but not with linear specifications. We find peer effects depend on a student's own ability and on the ability of the peers under consideration. Peer effects tend to be smaller when teacher fixed effects are included, a result that suggests co-movement of peer and teacher quality within a student over time. We also find that peer effects tend to be stronger at the classroom level than the grade level.

Introduction

The potential for peers to affect individual achievement is central to many important policy issues in elementary and secondary education, including the impacts of school choice programs, ability tracking within schools, “mainstreaming” of special education students, and racial and economic desegregation. Vouchers, charter schools and other school choice programs may benefit those who remain in traditional public schools by engendering competition that leads to improvements in school quality, but may also harm those left behind by diminishing the quality of their classmates (Epple and Romano 1998; Caucutt 2002). Grouping students in classrooms by ability can likewise have significant impacts on student achievement, depending on the magnitude of peer influences (Epple, Newlon, and Romano 2002). The effect of desegregation policies on achievement depends not only on potential spillovers from average ability, but on whether different peers exert different degrees of influence on individual outcomes (Angrist and Lang 2004; Cooley 2007; Fryer and Torelli 2005).

Despite the importance of these issues for American education policy, there are relatively few empirical studies of the magnitude and structure of peer effects on academic achievement in U.S. primary and secondary schools. A number of recent studies have attempted to estimate peer effects in the K–12 education context, yet most have been hampered by data limitations that constrain the scope of their analyses and the estimation techniques they are able to employ. With a unique panel data set encompassing all public school students in grades 3–10 in the state of Florida over the period 1999/00–2003/04, we have unprecedented resources with which to test for peer effects in the educational context. Unlike any previous study, we simultaneously control for the fixed inputs of students, teachers and schools in measuring peer influences on academic achievement. These controls sharply limit the scope for biases from endogenous selection of peers and teachers and permit a sharper estimate of the influence of classroom peers (as opposed to grade-level-at-school peers), than previous studies. Further, unlike previous work, which focuses almost exclusively on peer effects in elementary school, our data allow us to compare the impact of peer influences on math and reading achievement in elementary, middle and high school.

In addition to exploiting an extremely rich data set, we also employ a new analytical technique, adapted from Arcidiacono et al. (2005), that alleviates a number of problems associated with using student performance to measure peer influence. Typically, past research uses contemporaneous or lagged peer outcomes to measure peer ability. This can lead to a number of related estimation problems, such as simultaneity bias, measurement error bias, and biases caused by regression to the mean. Because observed academic outcomes, whether current or lagged, constitute a noisy measure of a student’s fixed inputs, measures of peer group influences based on such performance measures will be noisy and peer effects estimates may be biased downward. To better capture peer group characteristics, we estimate “peer fixed effects” simultaneously with individual fixed effects. The method has been shown to perform well even with a small number of observations per student. We extend the work of Arcidiacono et al. by estimating models which allow peer effects to operate through multiple moments of the distribution of the peer-group’s fixed effects and in which the effects of peer-group ability depend on individual ability.

(End of excerpt. The entire report is available in PDF format.)

The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.

Usage, posting and reprint of materials on the UI web site:

Most publications may be downloaded free of charge from the web site in PDF format. This information may be used and copies made for research, academic, policy or other non-commercial purposes. Proper attribution is required.

Copyright of the written materials contained within the Urban Institute website is owned or controlled by the Urban Institute. Posting UI research papers on other websites is permitted subject to prior approval from the Urban Institute—contact paffairs@urban.org.

If you are unable to access or print the PDF document please contact us or call the Publications Office at (202) 261-5687.