Article 3
Measuring Stress in Caribbean University Students: Validation of the PSS-10 in Barbados
Michael H. Campbell
The University of the West Indies, Cave Hill
Jill Gromer
Florida State University, Tallahassee
Donna-Maria B. Maynard
Maisha K. Emmanuel
The University of the West Indies, Cave Hill
Abstract
This study sought to establish support for research and clinical use of the Perceived Stress Scale-10 (PSS-10) with university students in Barbados. The instrument was administered to undergraduate student volunteers (N = 415, 75% women, mean age = 25.2). Other instruments administered included the Beck Depression Inventory-II (BDI-II; Beck, Steer, & Brown, 1996), the Zung Self-Rating Depression Scale (ZSDS; Zung, 1965), the Ego Resiliency-89 (ER-89; Block, 1989), and the State-Trait Anxiety Inventory – Trait Version (STAI-T; Spielberger, 1983). Response rate (94.9%) and internal consistency (α = .72) were acceptable. We hypothesized that the factor structure of the PSS-10 would be consistent with the oblique, two-factor structure suggested by previous studies of other populations. A confirmatory factor analysis (CFA) revealed acceptable fit on four of five fit indices. The model was modified to eliminate 3 items that fit poorly with their respective factors and negatively influenced item-total correlations. The revised model showed acceptable fit on each fit index, and the revised 7-item PSS showed improved internal consistency. Concurrent validities were demonstrated by significant correlations in the theoretically expected directions. Work to establish the validity of the PSS-10 in Barbados has yielded strong empirical support for the use of a modified 7- item version in clinical and research settings. Given the prevalence of clinical and research use of instruments developed in North American and other Western, Educated, Industrialised, Rich and Democratic contexts (WEIRD; Henrich, Heine, & Norenzayan, 2010), establishing Caribbean psychometric utility prevents misapplication of tools, reducing potential for both empirical and clinical errors.