Use of Exploratory Graph Analysis in Inspecting the Dimensionality of the Revised University of California Los Angeles (R-UCLA) Loneliness Scale Among Older Adults

Abstract

Exploratory graph analysis (EGA) based on network theory has been introduced as a highly reliable and effective method to assess scales’ dimensionality. We estimated the dimensional network structure of the Revised University of California Los Angeles Loneliness Scale using EGA among a cross-sectional cohort of Korean older adults living alone (N = 1,041). We also evaluated the stability of estimates using a bootstrap version of EGA (bootEGA) and verified the overall fit structure using confirmatory factor analysis (CFA). EGA revealed a two-dimensional structure of the scale initially. The bootEGA result revealed that Item 4 (“I do not feel alone”) did not sufficiently load on any dimension, and Item 20 (“There are people I can turn to”) was replicated in two or more dimensions. Removing these items resulted in better stability of the dimensions, leading to excellent structural consistency. CFA confirmed a satisfactory fit of the improved structure.