Convictions Based on Character: An Empirical Test of Other-Acts Evidence

Despite the time-honored judicial principle that “we try cases, rather than persons,” courts routinely allow prosecutors to use defendants’ prior, unrelated bad acts at trial. Courts acknowledge that jurors could improperly use this other acts evidence as proof of the defendant’s bad character. However, courts theorize that if the other acts are also relevant for a permissible purpose—such as proving the defendant’s identity as the perpetrator of the charged crime—then a cautionary instruction will cure the problem, and any prejudice is “presumed erased from the jury’s mind.” We put this judicial assumption to an empirical test. We recruited 249 participants to serve as mock jurors in a hypothetical criminal case. After reading the identical case summary, jurors were randomly assigned to one of two groups, each of which received different evidence on the issue of identity. Group A received conclusive proof, in the form of a stipulation, that if a crime was committed, the defendant was the one who committed it. Group A convicted at the rate of 33.1%. Group B received less certain evidence of identity in the form of the defendant’s somewhat similar, prior conviction, along with a cautionary instruction that this other act may not be used as evidence of the defendant’s character. Group B convicted at the much higher rate of 48.0%. The difference in conviction rates is statistically significant. Further, jurors in Group B were also more confident in their verdicts despite receiving less certain evidence of guilt and a cautionary instruction. These empirical findings demonstrate that cautionary instructions are not effective, and jurors will use other-acts evidence for impermissible purposes including, for example, the forbidden character inference. Given this, we discuss several pretrial strategies for defense counsel to limit the prejudicial impact of other-acts evidence