Hillel Bavli on "Stereotypes as Evidence"

Hillel Bavli has published a draft of Stereotypes as Evidence on SSRN. This article is forthcoming in volume 77 of the Stanford Law Review in 2025. It analyzes how the admission of profile evidence "involves substantial risks for the aims of fairness and equal treatment based on race, gender, economic status, and other personal characteristics."

Baserate evidence connects a defendant to an act through the defendant’s membership in a certain population. It includes evidence arising from forensic analysis, criminal profiling, statistical analysis, artificial intelligence, and many other common and emerging scientific methods. But while this evidence is prevalent in civil and criminal trials, it is poorly understood, and there is little predictability in how a court will decide its admissibility or even what standard the court will apply.

 

In this article, I show that although some forms of baserate evidence are desirable and even critical to achieving an accurate case outcome, a common form of baserate evidence called profile evidence constitutes unrecognized character evidence—evidence that a defendant acted in accordance with a certain character trait and that is prohibited by federal and state evidentiary rules. To show this, and to describe precisely the relationship between baserate evidence and character evidence, I draw on an area of statistics called Bayesian inference to define a new concept that I call population-propensity evidence. It describes a behavioral propensity of a population to suggest that an individual member of the population acted in accordance with this propensity. I show that this evidence—a form of baserate evidence that involves behavioral stereotyping—relies on impermissible character reasoning and therefore determines whether baserate evidence constitutes character evidence.

 

Finally, I discuss critical implications of my analysis. First, I show how an understanding of population-propensity evidence contributes descriptively to resolving longstanding confusion and inconsistency surrounding baserate evidence and profile evidence in particular. I then demonstrate that applying the rule against character evidence to determine the admissibility of profile evidence is essential to achieving correct and predictable evidentiary decisions, to minimizing the influence of implicit biases based on race and other personal characteristics of a defendant, and to reaching accurate verdicts.