Facets of measurement error for scores of the Big Five: Three reliability generalizations

Abstract

Measurement error in self-reports of personality consists of multiple facets that include random, transient, item- and scale-specific error components. Different reliability coefficients reflect different facets of measurement error. This study presents three reliability generalizations for measures of the Big Five based on 71 independent samples (total N = 38,944) that derived estimates for five types of reliability. The median aggregated coefficient of equivalence for the five traits was .82, the median coefficient of stability fell at .84, and the respective value for the generalized coefficient of equivalence was .74. The four facets of measurement error accounted for up to a half of the variance in observed scores. Estimates of different reliability coefficients are presented that can be used in future artifact corrections to derive construct-level relationships for the Big Five of personality. (C) 2014 Elsevier Ltd. All rights reserved.