Ball on Causal Inference Under the California Racial Justice Act

W. David Ball (Santa Clara School of Law) has posted MODELING MEANING: CAUSAL INFERENCE UNDER THE CALIFORNIA RACIAL JUSTICE ACT on SSRN. Here is the abstract:

 

In order to evaluate claims arising under the California Racial Justice Act (RJA), judges and attorneys need to learn how to draw inferences about racial disparity from data-and, equally importantly, to learn how to avoid drawing inaccurate inferences from data. The key questions in many RJA claims are, first, how to determine what constitutes "defendant[s] who have engaged in similar conduct and are similarly situated" and, second, what might supply race-neutral reasons for those disparities. In other words, how can practitioners learn to distinguish between permissible disparities driven by offender characteristics and conduct and impermissible disparities driven by factors the RJA seeks to eliminate: implicit bias, explicit bias, and structural factors? 
This Article seeks to provide judges and attorneys with such a guide.

It proceeds in three parts. First, any inferences drawn from data depend on how one models the real world. Judges and attorneys need to ask about these real-world models and treat them like any other evidence, evaluating them for their plausibility, coherence, and so on. Second, the Article explores two simple real-world models: one that is race-inflected, where system actors' treatment of race explains any observed racial disparities, and another where any disparities are the result of real-world racial differences in offender characteristics and conduct. Courts use both of these models to explain the data. The question, then, is which model is more plausible. Third, given that causal inference requires a model, I discuss the dangers of not making these models explicit. Particularly with models that assert real-world racial differences, there is a risk that eugenic tropes, particularly of blackness and criminality, will fill the gap.