Earwitness meta-analysis

Based on primary studies we hypothesized effects of a-priori moderators on mock earwitness performance. Including 46 articles with k = 66 experimental studies we performed a bare-bones and a random-effects, artefact-corrected meta-analysis. The results show a substantial ratio of the population effect size and the standard deviation of the population effect size (/SD) for bimodal stimuli of the moderator stimulus modality and for gender of listener. The fail-safe number demonstrates which population effect sizes might be changed to a small effect size d = 0.10 (e.g., retention interval) depending on the number of unpublished studies. The experimental design should be taken into account to calculate sample sizes for mock earwitness studies. Moreover, study characteristics like number of stimulus repetitions, individual differences and cognitive functions should be investigated in experimental mock earwitness studies. Single-person earwitness assessments in court should be substantiated by reality monitoring and disentangled from hearsay evidence.