Reexamining the Evidence of Validity for Office Discipline Referrals

Abstract

This mixed methods research synthesis study reexamined the evidence of validity for identifying individual students using their office discipline referrals (ODRs). ODRs range from severe (e.g., weapons) to subjectively determined (e.g., disruption) problem behavior suggesting two content domains. There are no studies showing teachers can reliably identify student behavior by ODR content. Proposed ODR cut score intervals (i.e., 0-1, 2-5, and > 6) were not linked to teacher behavior rating scale scores across the proposed percentile ranges (i.e., 0-80%, 81-94%, and > 95%). Convergent and discriminant correlations between ODRs and behavioral observations or rating scales showed small effect sizes (on average 7% and 4% explained variance, respectively). ODR data tend to require specialized statistical analyses because of their distributional properties; yet, these analyses were not used in the studies reviewed. Odds- and risk-ratios show that students of color2 receive relatively more ODRs than do students who are White, and for different ODR content and contexts, and for relatively more out-of-school suspensions and expulsions. Because there is limited evidence of validity and unfair practices for students of color, we question identifying individual students using subjectively determined ODR categories. In conclusion, the American Psychological Association Council of Representative’s apology to persons of color and call for the dismantling of racist practices suggest we take immediate action to end unfair assessment practices.