Abstract

People can decide whether a person appears either angry or happy in less than 1 second. Despite such speed, research shows that expression decisions are influenced by other facial attributes such as face sex. Nonetheless, specific patterns including participant sex differences remain unclear. Here, multiverse and distributional analyses clarify inconsistent results—participant sex differences for reaction time (RT) analyses were dependent on either an outlier removal method that effectively reduced the skew of the distribution or a specific distribution chosen to model the data. A further finding was that the patten of the Face Sex × Expression interaction effect for female participants differed markedly across the stimulus sets. The Diffusion Model, ex-Gaussian, ex-Wald, shifted Wald, and related distributions are recommended as replacements for analyses of mean RTs, rather than supplementary techniques. An extended analyses using the ex-Gaussian model is provided as an example. (PsycInfo Database Record (c) 2023 APA, all rights reserved)