Abstract
BACKGROUND: In modern industrialized countries, loneliness has been declared an epidemic. The present paper aimed to replicate previous findings about the dimensionality of the UCLA Loneliness Scale-3 version (LS-3) in a sample of Italian adults using Exploratory Structural Equation Modeling (ESEM). METHODS: In Study 1, 1676 participants (M(age)= 31.15; SD = 11.89) completed the LS-3. The three-factor ESEM model was performed to replicate the previous Italian version, and a Confirmatory Factor Analysis compared the three- and ten-item short versions. In Studies 2, 3, and 4, we administered measures of anxiety, depression, stress, internet addiction, self-esteem, and social interaction anxiety to 3513 participants (M(age)= 30.81; SD = 11.77) to assess the convergent and discriminant validity of LS-3. RESULTS: The ESEM 3-factor model was replicated using a Target rotation (X(2) = 847.597; df = 133; CFI = 0.942; TLI = 0.917; RMSEA = 0.062), whereas the already-known LS-3 short version raised psychometric concerns. Positive relationships with anxiety, depression, stress, internet addiction, and social interaction anxiety, as well as negative association with self-esteem were found. CONCLUSION: In conclusion, our findings support the use of the LS-3 in the Italian sample and discourage the administration of the current short forms. The implications for research and clinical practice have been discussed.