Psychometric comparison of two short versions of the Perceived Stress Scale (PSS-4) in a representative sample of the German population

Abstract

Perceived stress is a construct of crucial importance to health and well-being, necessitating the provision of economic, psychometrically sound instruments to assess it in routine clinical practice and large-scale survey studies. Two competing short versions of the Perceived Stress Scale (PSS), each consisting of four items, have been proposed. In the present study, we compare the two in a sample representative of the German general population (n = 2,527). Our analyses show that both versions are sufficiently reliable and valid, given the right measurement model. Specifically, the original PSS-4 by Cohen et al. suffers from response style effects, which we remedied using random intercept factor analysis. With the addition of the method factor, it is a highly reliable and valid scale. The PSS-2&2 by Schäfer et al. is more complex in its interpretation since it is split into two facets which cannot be summarized into a single score. Specifically, the Helplessness subscale correlates with related constructs very similar to the original unifactorial model but its reliability is lackluster. In contrast, the Self-Efficacy subscale is reliable but diverges in terms of its correlational pattern. In sum, both versions can be recommended for research designs in need of a brief measure of stress and offer unique contributions.