Psychometric properties of the Stress Mindset Measure (SMM) in the Polish population.

Purpose: The goal of the research was to evaluate the psychometric properties of the Polish adaptation of the Stress Mindset Measure (SMM; general version, SMM-G, and specific version, SMM-S).

Methods: Study 1 was an online survey conducted among 1651 adults (81% women, aged 18-84 years). To assess the theoretical validity of the SMM, the following constructs were also measured: Big Five personality dimensions, positive orientation, self-control, perceived stress at work, depressiveness, assessment of one's own health, and ego-resiliency. Study 2 was a test-retest reliability measurement and took place 10 months later among 344 participants.

Results: A factor validity was examined using exploratory (EFA) and confirmatory (CFA) factor analysis. EFA revealed a two-factor structure for the SMM-G and a one-factor structure for the SMM-S. However, these models obtained unsatisfactory goodness-of-fit indices in the CFA. Among the alternative models, the four-factor hierarchical model was best fitted to the data for both the SMM-G (RMSEA = .038, CFI = .996, TLI = .985) and the SMM-S (RMSEA = .041, CFI = .996, TLI = .990). These results were supported in the test-retest sample (SMM-G: RMSEA = .066, CFI = .990, TLI = .968; SMM-S: RMSEA = .056, CFI = .994, TLI = .983). Thus, four lower-order factors were identified: General, Health and Vitality, Performance and Productivity, Learning and Growth. The reliability of the overall general and specific indices measured with Cronbach's alpha was high and repeatable in both studies (Study 1: SMM-G α = .88; SMM-S α = .91; Study 2 (SMM-G, α = .87; SMM-S, α = .91). The stability for the SMM-G was satisfactory (r = .62; p < .001), and moderate for SMM-S (r = .46, p < .001). The theoretical validity analysis showed low (< |.40|) correlations in the expected directions with the majority of the selected tools.

Conclusion: The Polish adaptation of the SMM has very good psychometric properties. However, the unidimensional character of the original scale is not confirmed, which was also the case in other existing adaptations. The analyses in a sample several times larger than in previous studies revealed a greater complexity of the construct, identifying one higher-order factor and four lower-order factors.