Multilevel evaluation of factors predicting school exclusion among middle and high school students

School exclusion (out-of-school suspension and expulsion) is the most common and serious discipline response to student misbehavior. While most existing studies have considered student and school level risk factors for exclusion separately, multilevel analysis that evaluates student and school level variables together is needed since students are nested within their schools. This article addresses this need by using multilevel modeling to predict the likelihood of school exclusion following the last infraction of the year for 9706 middle and high school students in one school district. Multilevel modeling shows that students' poverty, previous suspensions, and severity of the last infraction predict exclusion while the schools' duplicated suspension rate predicts exclusion at the school level. These results highlight the need for comprehensive interventions that seek to change student behavior as well as the school environment. The role of school administrators, faculty and staff in developing and implementing these interventions and other implications for school personnel are discussed.