Meta-analytic factor analysis of the UCLA loneliness scale

Abstract

AbstractThe UCLA (University of California, Los Angeles) Loneliness Scale is frequently employed instrument for assessing the structure of loneliness in psychology studies. Although many studies have tested the scale’s factor structure, the literature has produced inconsistent findings or different results. Some studies support a unidimensional structure, while other studies support a multidimensional structure including two-factor, three-factor, five-factor models. Moreover, these dimensions are denoted differently across studies. Hence, this study employs meta-analytical structural equation modeling (MASEM) to investigate the factor structure of the UCLA Loneliness Scale, MASEM provides more precise parameter estimations than conventional structural equation modeling and allows for the synthesis of conflicting findings concerning factor structure. Furthermore, MASEM provides an examination of the factor structure by utilizing correlation matrices from studies in which the scale has been employed but without a specific examination of its factor structures. Consequently, this study analyzed the correlation matrices from 52 studies encompassing 52 correlation matrices. The results of the meta-analysis revealed that the two-factor and second-order three-factor models provided the best fit for the pooled correlation matrix. Even when considering subgroup analyses based on sample size and sample age, which are possible variables that can explain heterogeneity, the three-factor structure of the scale remained consistent. This suggests that different variables may account for the observed heterogeneity.