View article

Traits are unevenly distributed in multidimensional personality space, but the extent of trait lumpiness has never been formally assessed. In this brief comment, I offer a definition, a formula, and a cut-off for trait lumpiness. Furthermore, I provide an example of trait lumpiness in two to seven principal components in an existing lexical dataset (N= 400) containing 551 trait adjectives. In support of Lee and Ashton’s HEXACO model, the analyses support a maximum set of six factors that provide a relatively smooth distribution of traits in multidimensional space.© 2020 European Association of Personality Psychology

The outcomes of lexical personality research are usually presented in terms of several (eg five or six) independent dimensions of personality. From a theoretical and practical point of view this makes sense, because independent dimensions provide the Cartesian coordinates to (a) most optimally locate specific traits and (b) obtain the best possible set of independent predictors of criteria. Consequently, personality researchers have long strived to construct personality instruments that align well with the lexical personality axes. But just as with matter in our cosmological universe, traits are not evenly distributed in our personality universe nor are they exactly located at the poles of the lexical axes. For some, this has been a reason to denounce the practice of obtaining principal components in personality research; others have used it as a reason to search for trait clusters, or ‘types’(Asendorpf, 2002; but see Ashton & Lee, 2009a); still others have used it as a reason to construct trait circumplexes (Hofstee et al., 1992; Strus, Cieciuch, & Rowiński …