Moving beyond the mean in studies of mental chronometry: The power of response time distributional analyses

Abstract

Although it is widely recognized that response time (RT) distributions are almost always positively skewed and that mathematical psychologists have developed straightforward procedures for capturing characteristics of RT distributions, researchers continue to rely primarily on mean performance, which can be misleading for such data. We review simple procedures for capturing characteristics of underlying RT distributions and show how such procedures have recently been useful to better understand effects from standard cognitive experimental paradigms and individual differences in performance. These well-studied procedures for understanding RT distributions indicate that effects in means can be produced by (a) shifts of RT distributions, (b) stretching of slow tails of RT distributions, or (c) some combination. Importantly, effects in means can actually be obscured by opposing influences on the modal and tail portions of RT distributions. Such disparate patterns demand novel theoretical interpretations. © The Author(s) 2011.