THE UCLA LONELINESS SCALE: INVARIANCE OF SOCIAL STRUCTURAL CHARACTERISTICS

Abstract This study examines the structure and external validity of a reduced form of the revised UCLA Loneliness Scale. With a random sample of adults from a large metropolitan area ( N = 619), we used structural equation modeling to test the adequacy of a seven-item version of the UCLA Loneliness Scale assessing feelings of friendship and to analyze a hierarchy of factorial invariance hypotheses across a range of population groups. Our findings indicated that the abbreviated loneliness scale represents a good global fit and that the component fit measures are also highly satisfactory. Further, there were no differences in the conception of loneliness represented by the scale across the different population groups defined by gender, race, marital status, employment status, income, age and education. There were mean differences for some variables, indicating generally that those of lower social economic status (blacks, women and those with lower income) and those who were married were more likely to feel lonely. We conclude that the causes and consequences of differential loneliness based on social structural situation needs to be more cohesively analyzed, and that the seven-item univariate loneliness scale invariant across gender, race, marital status, employment status, income, age and education, as presented in this research merits further use.