Association for Public Policy Analysis & Management

Local and national education leaders have broadly encouraged schools to reduce their use of exclusionary discipline like suspensions (U.S. Department of Education, 2014). There are clear correlations between suspensions and future outcomes, both for suspended students and for non-suspended students at schools with large numbers of suspensions (e.g. Balfanz, Byrnes & Fox, 2014; Brownstein, 2009); however, it is difficult to causally distinguish between the impacts of misbehavior or its causes and the impacts of a school’s choice to use of exclusionary discipline in response to those behaviors.

Following national trends, Chicago Public Schools (CPS) adopted policies encouraging reduced use of exclusionary discipline between 2009 and 2014. Schools seem to have responded, often substituting in-school suspensions for out-of-school suspensions (Sartain et al., 2015). We exploit the policy-induced change in suspending behavior to examine how changes in the use of out-of-school suspension impact academic performance, attendance, GPA, and student reports of school climate.

We produce plausibly causal estimates using panel models with school and student fixed effects, as well as controls for student behavior. The student and school fixed effects control for all time-invariant factors at the student and school level. We pursue a variety of strategies to attempt to separate the impact of changes in the use of exclusionary discipline from the impact of changes in student behavior at the school level. In addition to specifications directly controlling for observed events in the administrative data, we are also able to control for student reports of misbehavior from a system-wide student survey. Finally, we estimate specifications using the ratio of out-of-school to in-school suspensions to capture the school discipline regime, out of concern schools under-report disciplinary events which do not result in a suspension of some type. Preliminary results suggest the impacts of changes in the use of out-of-school suspension are small but significant.

It is reasonable to expect school discipline regimes will have heterogeneous impacts by student characteristics and school types. For example, Perry and Morris (2014) find suggestive evidence that extensive use of exclusionary discipline may negatively impact test scores for students who do not misbehave. We examine this heterogeneity by separately estimating impacts for students who are never observed misbehaving compared with those who are; boys compared with girls; and schools with high baseline suspension rates compared with schools with moderate or low rates.