Student Poster Award 2023

Abstract: Police lineups usually consist of one suspect and multiple fillers; fillers are people who match the physical description of the suspect but are known innocents. Currently, there is no consensus regarding the ‘optimal’ level of suspect-filler similarity. However, recent research grounded in diagnostic-feature-detection theory (Wixted & Mickes, 2014) suggests that to maximise witnesses’ ability to discriminate between innocent and guilty suspects, lineups should consist of fillers who match the witness’ description, but otherwise look dissimilar to the suspect (Colloff et al., 2021). To investigate the effect of suspect-filler similarity on eyewitness discriminability, we used a multiple trial simultaneous lineup task. Participants (N = 646) were presented with 6 randomly selected to-be-remembered target faces – each of a different physical description – and were asked to identify each face from 6 photo lineups. Half of the lineups contained the target, and the remaining half contained a medium-similarity target replacement. Lineups were randomly generated to contain fillers that were either low, medium, or high similarity to the target or target-replacement. Similarity measurements for each face pair were obtained using a separate spatial arrangement task. Unexpectedly, we found that discriminability decreased with lower similarity. Upon further exploration, we discovered that due to lineup randomisation, the suspect stood out in some lineups due to lighting and physical description issues. To understand whether this a theory issue or an experimental design issue, we plan to run a second experiment with stricter parameters on the filler pool (e.g., all faces would be ‘reasonably described’ in the same way) to ensure that all lineups are fair, and accurately represent low, medium, or high similarity.