Measuring higher education students' perceived stress: An IRT-based construct validity study of the PSS-10. (December 2019)

• Record Type: Journal Article
• Authors: Nielsen, Tine; Dammeyer, Jesper
• Is Part Of: Studies in educational evaluation. Volume 63(2019)
• Page Start: 17
• Page End: 25
• Publication Date: 2019-12
• Subjects: Higher education -- Stress -- Perceived Stress Scale -- Rasch model -- Construct validity; Curriculum evaluation -- Periodicals; Educational tests and measurements -- Periodicals; Programmes d'études -- Evaluation -- Périodiques; Tests et mesures en éducation -- Périodiques; 375.006
• DOI: 10.1016/j.stueduc.2019.06.007
• Languages: English
• ISSNs: 0191-491X

Abstract

Highlights: The perceived Stress Scale (PSS-10) consist of two separate subscales. Subscales measuring perceived stress and perceived lack of control. Subscales both function differentially across academic disciplines. Perceived stress subscale function differentially across gender. Use of the PSS-10 with higher education students should be studied further. Abstract: The Perceived Stress Scale (PSS-10) is widely used. Factor-analytic studies have reported a two-dimensional structure (Perceived Stress and Perceived Lack of Control subscales), however, few studies have employed Item Response Theory (IRT) or focused on the issue of differential item functioning (DIF). The Rasch Model (RM) for ordinal data and graphical loglinear Rasch models (GLLRM) were used with data from 1, 552 technical and psychology university students. After elimination of item 6 the two subscales were found each to fit a GLLRM. However, lack of conditional independence of the items (LD) and equal item functioning (DIF) were found. The findings point to the need for further construct validity studies of the PSS-10 subscales focusing on the issues of DIF and LD.