Psychometric evaluation of the Perceived Stress Scale (PSS-10) in Ecuadorian adults: evidence from a bifactor model and measurement invariance

Abstract

Background Perceived stress is a key transdiagnostic construct associated with multiple indicators of psychological maladjustment. Although the Perceived Stress Scale (PSS-10) has previously been validated in Ecuador, evidence has been limited to university students. This study examined the psychometric properties of the PSS-10 in a large community sample of Ecuadorian adults. Methods A cross-sectional instrumental study was conducted with 16,074 Ecuadorian adults (mean age = 30.91 years, SD = 10.05; 46.8% women). The sample was randomly divided into exploratory (n = 8,019) and confirmatory (n = 8,055) subsamples. Factorial structure, bifactor reliability, measurement invariance across sex, and validity based on relationships with loneliness, psychological inflexibility, depressive symptoms, and impulsivity were evaluated. Results Exploratory analyses supported a two-factor structure (Negative Stress and Coping), whereas confirmatory analyses indicated that a bifactor model, comprising one general perceived stress factor and two specific factors, provided the best fit among the competing models (CFI = 0.990, TLI = 0.982, SRMR = 0.043), although RMSEA remained elevated (0.098; 90% CI [0.096, 0.101]). Bifactor indices supported the predominance of a general factor (ECV = 0.561, ωH = 0.645, H = 0.940), while the Coping factor retained meaningful reliable variance beyond the general factor. The total score showed excellent reliability (ω = 0.945) and moderate-to-strong positive associations with psychological inflexibility (r = 0.711), depressive symptoms (r = 0.698), loneliness (r = 0.582), and impulsivity (r = 0.579). Measurement invariance analyses provided limited evidence of equivalence across sex, and women reported higher observed stress scores than men (F = 1,438.81, p < 0.001, η2 = 0.152). Conclusions The Ecuadorian PSS-10 provides reliable and valid scores for assessing perceived stress in adults. The bifactor model supported the use of the total score as an overall indicator of perceived stress, while the Coping factor may provide complementary information. These findings should be interpreted considering the overall psychometric evidence and the limited support for measurement equivalence across sex.