The Black–White Suspension Gap for Disobedience: The Role of School Size, Racial Composition, Performance, and Resources

Background and Purpose

Despite well documented harms to students’ educational and health outcomes, exclusionary discipline practices remain prevalent in U.S. schools and disproportionately affect Black students. Although prior research has examined school factors associated with these disparities, most studies rely on discipline data aggregated across all infraction categories. This approach may obscure potential differences in how educators respond to subjective versus objective behaviors, particularly given that “disobedient/disruptive” behavior accounts for the majority of suspensions. Understanding racial disparities in this subjective category is important for informing interventions. This study used statewide school- and district- level data to examine how selected school characteristics relate to Black-White disparities in suspensions rates for disobedience.

Methods

We analyzed publicly available 2023–2024 data from a Midwestern state department of education. School-level variables included suspension rates for disobedience, total enrollment, student demographics, and academic proficiency rates. Student–teacher ratio data were obtained from the National Center for Education Statistics. Schools with adequate numbers of Black and White students were included, yielding a sample of 1,022 schools nested within 241 districts. Suspension rates were calculated as the number of suspension incidents within each racial group divided by that group’s enrollment, and the disparity measure was defined as the difference between Black and White suspension rates. We first estimated a null model to calculate the intraclass correlation coefficient (ICC), followed by a random-intercept multilevel model to account for clustering of schools within districts. Analyses were conducted in R version 4.5.1.

Results

The null model ICC indicated that 11.8% of the variance in the suspension rate gap was attributable to differences between school districts. Multilevel model results revealed that a 1 standard deviation (SD) increase in the percentage of Black student enrollment was associated with a .09 (p < .001) SD increase in the suspension rate gap. A 1 SD increase in school enrollment was associated with a .20 (p = .03) SD increase in the gap. Conversely, a 1 SD increase in the pupil–teacher ratio was associated with a .17 (p < .001) SD decrease in the gap; and a 1 DS increase in the school performance index score was associated with a .32 (p < .001) decrease in the gap. The percentage of economically disadvantaged students was not significantly associated with the gap (β = -.06, p = .32)

Conclusions and Implications

These findings suggest that racial disparities in suspension rates are shaped by both school- and district-level contexts. Larger, lower-performing schools and those with higher proportions of Black students exhibited wider Black–White suspension gaps. School-level institutional characteristics, particularly academic performance level, showed stronger associations with disparities than student demographic composition, underscoring the importance of institutional climates. The non-significant effect of the percentage of economically disadvantaged students challenges the differential involvement hypothesis, which attributes discipline gaps primarily to student poverty. Contrary to prior literature, higher pupil–teacher ratio was associated with narrower gaps. Future interventions should prioritize large, low-performing schools with higher proportions of Black students to reduce discretionary disciplinary responses to subjective behaviors.