Psychometric analysis of the three item loneliness scale in the Czech Republic

Abstract

There is a growing importance of loneliness measurement through valid and reliable instruments. However, to establish valid and reliable measures, there is a need to explore their psychometric properties in different research settings and language environments. For this reason, this study aimed to validate the Three Item Loneliness Scale (TILS) in the Czech Republic within a Slavonic language environment.MethodsA sample of Czech adults (n = 3236) was used consisting primarily of university students. We utilized Classical Test Theory to assess TILS internal consistency, temporal stability, factor structure, and measurement invariance. Item Response Theory (IRT) was used to estimate the discrimination and difficulty of the TILS items and to estimate the measurement precision of the whole scale. Convergent validity was explored through the Spearman correlation coefficient using personality traits, depression, and anxiety.ResultsThe results showed satisfactory reliability and validity of the TILS in the Czech Republic. The scale's internal consistency and temporal stability were found to be satisfactory (Cronbach’s α = 0.81, McDonald’s ω = 0.82, r = 0.71). The parallel analysis supported the unidimensionality of the TILS with a high factor loading of items (λ = 0.61 - λ = 0.85). Moreover, measurement invariance indicated that the TILS measures loneliness equivalently between genders. The IRT results indicated that the highest measurement precision was reached in individuals with lower and above-average levels of loneliness. Significant correlations between the TILS scores, anxiety, depression, and personality traits supported the convergent validity of the scale.