New preprint: powe(R)OC: A power simulation tool for eyewitness lineup ROC analyses
Receiver operating characteristic (ROC) curve analyses have become increasingly common in eyewitness lineup experiments, yet statistical power for these analyses is not well-understood. powe(R)OC is a free, open-source R Shiny web app that allows users with minimal programming knowledge to conduct simulation-based power analyses for eyewitness lineup experiments that use ROC analysis (along with other recognition memory experiments). powe(R)OC uses existing data (either user-uploaded or publicly-available open data) as a basis for simulation, allows powering for analyses of partial area under the curve (pAUC) and deviation from perfect performance (Smith et al., 2019), allows users to specify various simulation parameters (e.g., effect sizes, sample sizes, number of lineups, pAUC truncation, test tails), view simulation results uploaded by other users, view ROC effect sizes in the literature, and download simulation results in a summary report. This report describes ROC analyses, challenges in simulating ROC data, methods implemented in powe(R)OC and their underlying assumptions and limitations, and future plans for the app.
