What's in a Scale - Multidimensionality or Wording Artifact? A Comprehensive Psychometric Analysis of the Rasch-Type...

This study examined the psychometric properties of the widely used 6-item Rasch-Type Loneliness Scale (RTLS-6), including a modified intensity response format. The RTLS-6 is assumed to capture the aspects of emotional and social loneliness. However, the items capturing social loneliness are negatively oriented, raising concerns that the two factors of the scale may be artifactual. Across three studies (N₁ = 3,079, N₂ = 1,233, N₃ = 500), we examined the dimensionality and Rasch-conformity of the original and two manipulated versions of the RTLS-6, featuring uniformly worded items toward loneliness and different numbers of response categories. Overall, the results support a two-factor structure of the scale, which remains robust even when wording effects are ruled out. Multidimensional, polytomous Rasch analyses suggest that a four-point intensity response format is most appropriate. In sum, the original RTLS-6 provides evidence aligning with the assumptions of the polytomous Rasch model, particularly regarding model fit, ordered thresholds, and the absence of differential item functioning across several subgroups. Accordingly, the RTLS-6 appears suitable for use as a two-dimensional scale capturing emotional and social loneliness.