Variability as a Predictor: A Bayesian Variability Model for Small Samples and Few Repeated Measures

Abstract

Whilst most psychological research focuses on differences in means, a growing body of literature demonstrates the value of considering differences in intra-individual variability. Compared to the number of methods available for analyzing mean differences, there is a paucity of methods available for analyzing intra-individual variability, particularly when variability is treated as a predictor. In the present article, we first reviewed methods of analyzing intra-individual variability as an outcome, including the individual standard deviation (ISD) and some recent methods. We then introduced a novel Bayesian method for analyzing intra-individual variability as a predictor. To make this method easily accessible to the research community, we developed an open source R package, VARIAN. To compare the accuracy of parameter estimates using the proposed Bayesian analysis against the ISD as a predictor in a regression, we carried out a simulation study. We then demonstrated, using empirical data, how the estimated intra-individual variability derived from the proposed Bayesian analysis can be used to answer the following two questions: (1) is intra-individual variability in daily time-in-bed associated with subjective sleep quality? (2) does subjective sleep quality mediate the association between time-in-bed variability and depressive symptoms? We concluded with a discussion of methodological and practical considerations that can help guide researchers in choosing methods for evaluating intra-individual variability.