Predicting and postdicting eyewitness accuracy and confidence

Abstract

A primary consideration regarding the admissibility of expert testimony in criminal trials is whether particular evidentiary issues are intuitively understood by jurors. Experiment 1 assessed the eyewitness identification accuracy and confidence of 287 retail store clerks, half of whom knew, while interacting with the target, that they would later be tested on their ability to identify that person. Two weeks later, each clerk/witness was tested on both a target absent and a target present lineup. In Experiment 2, 32 undergraduate students were given a detailed description of Experiment 1 and asked to postdict the clerks’ accuracy and confidence. Although individual students often erred in their absolute estimates, their postdictions were sensitive to the effects of the warning and lineup manipulations; on average their estimates largely approximated the direction and magnitude of the effects observed in the clerks’ data.