Interpretation of statistical evidence in criminal trials: The prosecutor's fallacy and the defense attorney's fallacy.

Abstract

In criminal cases where the evidence shows a match between the defendant and the perpetrator on some characteristic, the jury often receives statistical evidence on the incidence rate of the “matching” characteristic. Two experiments tested undergraduates' ability to use such evidence appropriately when judging the probable guilt of a criminal suspect based on written descriptions of evidence. Experiment 1 varied whether incidence rate statistics were presented as conditional probabilities or as percentages, and found the former promoted inferential errors favoring the prosecution while the latter produced more errors favoring the defense. Experiment 2 exposed subjects to two fallacious arguments on how to interpret the statistical evidence. The majority of subjects failed to detect the error in one or both of the arguments and made judgments consistent with fallacious reasoning. In both experiments a comparison of subjects' judgments to Bayesian norms revealed a general tendency to underutilize the statistical evidence. Theoretical and legal implications of these results are discussed.