Development and Psychometric Validation of a Modified Statistical Anxiety Rating Scale (STARS) Among Postgraduate Medical and Nursing Students in India

Abstract

Background: Statistical anxiety is common among postgraduate health-profession students and may adversely affect research learning, dissertation completion, and evidence-based practice. Although the Statistical Anxiety Rating Scale is widely used, its applicability in the Indian postgraduate health-profession education context has not been established. This study aimed to develop and psychometrically validate a modified version of the STARS among postgraduate medical and nursing students.

Methods: A cross-sectional psychometric validation study was conducted among 627 postgraduate medical and nursing students at a tertiary care teaching institute in India. Following expert-based face and content validity assessment, a modified 31-item STARS was developed. Exploratory factor analysis (EFA; n=327) and confirmatory factor analysis (CFA; n=300) were performed using independent samples. Reliability was assessed using Cronbach's alpha and McDonald's omega, while convergent and discriminant validity were evaluated using Average Variance Extracted and the Heterotrait–Monotrait ratio.

Results: EFA identified a four-factor structure. After removal of six additional items during psychometric evaluation, the final instrument comprised 25 items. CFA confirmed good model fit (CFI=0.92, TLI=0.91, RMSEA=0.06, SRMR=0.05). Standardized factor loadings ranged from 0.66 to 0.90. Cronbach's alpha ranged from 0.79 to 0.89, composite reliability (McDonald's ω) from 0.78 to 0.88, AVE from 0.40 to 0.50, and HTMT from 0.43 to 0.87, supporting satisfactory reliability, convergent validity, and discriminant validity.

Conclusion: The modified 25-item STARS demonstrated satisfactory psychometric properties and provides a concise, reliable, and contextually relevant instrument for assessing statistical anxiety among postgraduate medical and nursing students. Further multi-institutional validation is recommended before broader application.