Psychometric functioning, measurement invariance, and external associations of the Relationship Assessment Scale in a sample of Polish Adults

Abstract

The current article reports data from three Polish samples to examine the Relationship Assessment Scale (RAS) with respect to its unidimensionality, invariance across countries, gender, formal and informal relationships, degree of precision (or information) across latent levels of relationship satisfaction, and the functioning of individual items. The analyses of the data from the reference sample (n = 733) confirmed a clear 1-factor structure of the RAS-PL and good internal consistency. Configural, metric, and scalar invariance for countries (Poland, Hungary, USA), gender (women and men) and relationship types (formal and informal relationships) were achieved. Item Response Theory Analysis (IRT) suggested that the RAS-PL assesses relationship satisfaction most reliably at low to average levels. Analyses of the data from validation samples (n = 203 and n = 209) confirmed the convergent and divergent validity by weak, medium, and large correlations of the RAS-PL with measures of other theoretically related constructs. Concurrent criterion validity was demonstrated by a strong positive correlation between the RAS-PL and the intent to continue the current relationship. This investigation provides considerable psychometric information about the items and scale of the RAS-PL.