System variables in eyewitness identification: Control experiments and photospread evaluation¹

Abstract

An incident was staged in front of approximately 140 college students Three months later, 48 of them tried to identify the person of the incident (the target) in a nineperson photospread (Experiment 1) Assisted with a verbal description of the target, 77 less confident students similarly tried to identify him (Experiment 2) Additionally, 69 students from another class tried to choose the target with the verbal description only (Experiment 3 a mock-witness control experiment) Results of Experiment 1 showed that although more than half of the participants correctly identified the target, a similarly high proportion of control participants falsely identified a specific foil as the target. In the three experiments. correct and false identification patterns were similar, and the variables of target presence and distracter similarity independently affected probability estimates for selected photographs to be the target's The implications for actual eyewitness identification are discussed

Testing the reflection assumption

“…While the police may not extensively question or ask the witness to make an identification if they did not attend to the crime, in ‘live’ staged crime experiments, the results are rarely conditioned on such factors. Additionally, these results from Kurosawa (1996) and Riske et al (2000) suggest that our crime simulations might not be as impactful as real criminal events; to our recollection, we coded only three studies (Hollien et al, 1983; Hosch and Bothwell, 1990; Leippe et al, 1978) reporting that subject witnesses to a staged crime attempted to intervene on behalf of the victim. While for obvious ethical reasons we cannot expose our participants to real criminal activity, studies such as these suggest that we should generalise laboratory results more cautiously, and that we should take seriously the issue of whether our simulations allow participants to be representative of real-world witnesses.…”

“…Whether we are appropriately simulating witnessing is a reasonable question to ask, for, unlike the real-world cases, experimenters do not typically ask their participants how closely they paid attention, or screen out people who did not pay attention to or remember the crime (see Yuille and Cutshall, 1986, for an interesting discussion relating to how investigators might screen out witnesses who had a poor vantage point). To illustrate, Kurosawa (1996) used a typical staged crime scenario in a lecture hall setting and found that 33% of the participants could not recall seeing the staged incident. Additionally, 40-50% of the students were not able to say what the intruder did.…”