Expulsion in Early Schooling: Unpacking Disparities and Predictors in California's Public Elementary Schools

Background

Research problematizes race, socioeconomic, gender, and other disparities in exclusionary school discipline methods such as suspension and expulsion (e.g., Okonofua, Walton, & Eberhardt, 2016; Wallace et al., 2008; Skiba, Michael, Nardo, & Peterson, 2002), which persist even in prekindergarten years (Gilliam, 2005). Expulsion is of particular concern because it is theoretically reserved for the most serious disciplinary infractions and is, in some states, accompanied by legal protections for students. Despite the fact that districts are involved in setting guidelines for disciplinary procedures, district-level variations in expulsion remain under-investigated. Moreover, there is a dearth of empirical work on disparities in exclusionary discipline in early education. Therefore, the current study examines how expulsion rates among elementary schools vary within and across districts in California, with particular focus on school-level socioeconomic and racial/ethnic composition after controlling for structural covariates.

Methods

This study is a secondary analysis of discipline data collected across all California public schools from the 2016-2017 school year. Data were obtained from a state-level administrative archive, linkable through unique school, district, and county numeric identifiers. The sample for the current study included all 5,640 elementary schools nested in 683 districts in the state. School expulsion rate was the dependent variable; school and district characteristics included racial/ethnic compositions, number of students enrolled, charter status, and district type. Regression analyses and two multilevel models (linear random-intercept and random coefficient) were used.

Results

Estimates indicated significant variation between districts, with intraclass correlations ranging from 0.28 with the random intercept model and 0.19 after using a random coefficient model by district type. The Hausman test indicated weak evidence for model misspecification between fixed and random effects models. Between-district effects were significantly predicted by proportions of populations of students of color at school and district levels. Within-district effects were statistically significantly predicted by the proportions of students of color as well as socioeconomically disadvantaged, foster, and English learner populations.

Conclusions and Implications

The current study provides new insight about expulsion in California elementary schools. Though the story of expulsion is that it is invoked in response to extraordinary circumstances, the results indicated that disparities in expulsion are not unique to individual elementary schools. Rather, elementary school expulsion rates were significantly explained by racial/ethnic economic, and other compositional factors at the district level, suggesting that the propensity to surveill and enforce punitive recourse is elemental across administrative and organizational governance of education — even in early schooling years. It is critical to build a knowledge base about the contexts of schools where stark discipline disproportionalities exist. Rigorous examination of factors predictive of high expulsion rates may lead to the identification of creative entry points for school social workers and allied professionals to raise consciousness about systemic bias, to implement inclusive interventions, and to influence district and school policies.