Missing the Trees for the Forest: The Importance of Context and Limits of Generalizability in the Racialization of Dis/abilities

• In Event: 77.057 - MTCC Roundtable Session 35
  In Roundtable Session: 77.057-12 - Critical Disability Studies and Intersectionality

Abstract

This study used data mining approaches to explore associations between schoolwide racial disproportionality in special education identification and publicly available measures of social context and demographic composition. The sample included all high schools in Arizona and Virginia with at least 10 Black and/or Latinx students in 2015-2016. Using regression trees and random forest models, results indicated that across the two states and for the two student subgroups, different contextual predictors influence disproportionality. In most cases, nonlinear relationships emerged that would not be easily detected by traditional statistical modeling. This study advances work on disproportionality by illustrating the complex role of context, and suggests that traditional statistical models may not be the best approach for understanding this complex phenomenon.

Authors

• Michael Broda, Virginia Commonwealth University

• Adai A. Tefera, Virginia Commonwealth University

• Eric Ekholm, Virginia Commonwealth University