Earwitness meta-analysis

Abstract

The present meta-analysis investigates study and sample characteristics of mock earwitness performance. Based on primary studies we disentangled several a-priori moderators that modulate earwitness performance. Despite heterogeneous results in articles, we found experimental studies investigating effects of stimulus modality, stimulus length, retention interval, familiarity of language, and own-group or gender effects on earwitness performance. Including 33 articles with k = 49 experimental studies we performed a bare-bones and an artefact-corrected meta-analysis across all included primary studies and for five a-priori moderators. The results show a substantial ratio of the population effect size and the standard deviation of the population effect size exclusively for bimodal stimuli and concrete stimuli of the moderator stimulus modality. The fail-safe number was calculated to demonstrate which population effect sizes might be changed to zero depending on the number of unpublished primary studies. We highlight study and sample characteristics that facilitate earwitness performance. In power analyses, we show that the experimental design and individual differences should be taken into account to calculate sample sizes for future earwitness studies. We recommend best-practice strategies to investigate earwitness performance in future experimental studies and in single-person earwitness assessments.