Performance under pressure: Development and validation of an objective structured clinical examination (OSCE) anxiety scale in pharmacy students
International Journal of Health Sciences · 2026-11-12 · 20 · 269-279
DOI: 10.25259/ijhs_158_2026 ↗Abstract
Objectives: Objective Structured Clinical Examinations (OSCEs) provoke considerable anxiety, yet most studies use general test-anxiety measures not designed for performance-based assessment. This study developed and gathered initial validity evidence for an OSCE-specific anxiety scale (the OSCE Anxiety Scale, OAS) in pharmacy students. Methods: The Test Anxiety Inventory was adapted through expert review, content validity indexing, and piloting; all 20 adapted items were administered, of which 16 form the two hypothesized subscales (Emotionality, Worry). The scale was administered to 608 Doctor of Pharmacy students at a single institution before the OSCEs. Internal structure was examined using polychoric correlations, parallel analysis, and factor analysis; score reliability was assessed using ordinal omega; and relations to performance were examined using Spearman correlations. Robust ordinal estimation (weighted least squares, mean- and variance-adjusted estimator; WLSMV) was applied as a sensitivity analysis. Results: Parallel analysis supported at most two factors, while the exploratory two-factor solution showed weak separation (inter-factor correlation F = 0.71). No confirmatory model met conventional fit criteria; the two-factor model fitted poorly (comparative fit index [CFI] = 0.749; root-mean-square error of approximation [RMSEA] = 0.113) and resembled a one-factor solution (latent inter-factor correlation = 0.99); under WLSMV estimation, approximate fit improved (CFI = 0.902, RMSEA = 0.080) but the two factors remained empirically indistinguishable (latent correlation = 1.00). Ordinal omega was 0.81 (Emotionality) and 0.74 (Worry). OAS scores correlated negatively with OSCE performance (r = –0.27 to –0.30). Conclusion: The OAS showed adequate emotionality reliability and an expected association with performance, but a dominant general dimension in which the hypothesized emotionality and worry factors could not be empirically distinguished, requiring item refinement before interpretive use.
EUVIMED