Individual differences in cognitive paradigms are increasingly employed to relate cognition to brain structure, chemistry, and function. However, such efforts are often unfruitful, even with the most well established tasks. Here we offer an explanation for failures in the application of robust cognitive paradigms to the study of individual differences. Experimental effects become well established – and thus those tasks become popular – when between-subject variability is low. However, low between-subject variability causes low reliability for individual differences, destroying replicable correlations with other factors and potentially undermining published conclusions drawn from correlational relationships. Though these statistical issues have a long history in psychology, they are widely overlooked in cognitive psychology and neuroscience today. In three studies, we assessed test-retest reliability of seven classic tasks: Eriksen Flanker, Stroop, stop-signal, go/no-go, Posner cueing, Navon, and Spatial-Numerical Association of Response Code (SNARC). Reliabilities ranged from 0 to .82, being surprisingly low for most tasks given their common use. As we predicted, this emerged from low variance between individuals rather than high measurement variance. In other words, the very reason such tasks produce robust and easily replicable experimental effects – low between-participant variability – makes their use as correlational tools problematic. We demonstrate that taking such reliability estimates into account has the potential to qualitatively change theoretical conclusions. The implications of our findings are that well-established approaches in experimental psychology and neuropsychology may not directly translate to the study of individual differences in brain structure, chemistry, and function, and alternative metrics may be required. Electronic supplementary material The online version of this article (doi:10.3758/s13428-017-0935-1) contains supplementary material, which is available to authorized users.
