Examining Structural and Contextual Factors That Predict High Expulsion Rates in California Schools

Background: Research has shown that low-income students, students with disabilities, and students of color are disproportionately targeted by exclusionary discipline, pushed out of schools and toward the justice system (e.g. Skiba & Knesting, 2001; Skiba, Arredondo, & Williams, 2014).  Reducing such disparities requires a more nuanced knowledge base about school contextual levers that predict exclusionary discipline. The current study examined the extent to which expulsion rates were predicted by school type, prevalence of socioeconomic disadvantage among student body, and charter status.

Methods: The study used the California Longitudinal Pupil Achievement Data System (CALPADS) to examine expulsion records from 1,586 local educational agencies and charter schools from 51 counties during the 2016-17 academic year.  Multiple linear regression was used to estimate differences in mean expulsion rates by school type and charter status. Multiple logistic regression was used to investigate whether the odds of high expulsion rate depend on the type of alternative school and the percent of enrolled socioeconomically disadvantaged students.  All analyses controlled for total enrollment at schools.

Results: Mean estimates of expulsion rates were statistically significantly lower at elementary (b=-1.43, p < 0.001), middle (b=-1.23, p < 0.001 ), and high schools (b=-0.92, p < 0.001 ) in comparison to alternative schools.  In fact, the odds of high expulsion rates were estimated to be 7.8 and 5.8 times as great for alternative schools and continuation high schools, respectively, compared to non-alternative school types (alternative: OR: 7.83, p < 0.001; continuation: OR: 5.83, p < 0.001).  Each percentage increase in the population of socioeconomically disadvantaged students at schools was associated with a 12% increase in the estimated odds of high expulsion rates (OR: 1.12, p < 0.001).  After controlling for percent of socioeconomically disadvantaged students and total enrollment at schools, the odds of having high expulsion rates were estimated to be 9.1 and 5.4 times as great for alternative school types and continuation high schools, respectively, compared to traditional schools (alternative: OR: 9.06, p < 0.001; continuation: OR: 5.35, p < 0.001).  Charter schools had a statistically significantly lower mean estimate of expulsion rates in comparison to non-charter schools (b=-0.56, p < 0.001 ).

Conclusions and Implications: The findings of the study suggest that specific contextual factors, namely school type, socioeconomic disadvantage, and charter status, predict high expulsion rates in California schools.  Alternative schools and schools that serve low-income populations appear most motivated to invoke expulsion as a disciplinary intervention. This study implicates future research to examine the characteristics of schools nested in districts and counties in California to better understand how the indicators discussed in this paper relate to a more rich understanding of expulsion, and to determine the how school-based interventions may best contribute to the safe and supportive inclusion of students at schools.