Considering Approaches to Pervasiveness in the Context of Personality Psychology

In experimental psychology, the term “pervasiveness” has been introduced to refer to the proportion of participants who exhibit an effect. This paper is a review of existing methods that could be used to extend the idea of pervasiveness to cross-sectional correlational data. More specifically, I consider how bivariate or multivariate effect sizes can be expressed or contextualized through proportions or percentages of individuals via the following methods: Kendall’s Tau, Common Language Effect Size indicators (Dunlap’s CLr, Li and Waisman’s B P , Rosenthal and Rubin’s BESD and Mõttus’ TACT), rule-based methods (mostly Association Rule Mining, but Comparative Configural Models such as QCA and CORA are briefly discussed) and Observation Oriented Modeling. I illustrate the use of these methods in personality psychology using data from Soto's LOOPR project and highlight strengths and limitations for the different methods. I show that an Observed Percentage of Concordant Pairs (OPCP), easily obtained from Kendall’s Tau, can be used as a general-purpose common language effect size that also provides information on pervasiveness. The ‘pervasive’ R package, that provides the OPCP and information on a set of rules for use in a simple or multivariate regression setting, is introduced.