The Perils of Untested Assumptions in Theory Testing: A Reply to Patrick et al. (2020)

We respond to a critique by Patrick et al. (2020) of our recent study (Roy et al., 2020) that raised questions regarding the three-factor model of the Triarchic Psychopathy Measure (TriPM). Roy et al. demonstrated that a replicable model involving seven unidimensional factors accounted for the TriPM items across North American and European general population samples, as well as European male offenders. Despite having access to large TriPM datasets, the Patrick et al. critique relied on tangential analyses of general personality traits, using a single college sample with TriPM data as supplemental. Thus, Patrick et al. ignored findings highlighting multidimensional TriPM scales and the uncertainty they introduce with respect to the larger nomological network of psychopathy. In our reply, we demonstrate additional problems with the three-factor TriPM model and show that the seven-factor model outperforms the three-factor model in predicting correlates of psychopathy