Rasch analysis on OSCE data: An illustrative example
Publication Details
Tor, E., & Steketee, C. (2011). Rasch analysis on OSCE data: An illustrative example. Australasian Medical Journal, 4(6), 339-345. http//dx.doi.org/10.4066/AMJ.2011.755
Abstract
Background: The Objective Structured Clinical Examination (OSCE) is a widely used tool for the assessment of clinical competence in health professional education. The goal of the OSCE is to make reproducible decisions on pass/fail status as well as students’ levels of clinical competence according to their demonstrated abilities based on the scores. This paper explores the use of the polytomous Rasch model in evaluating the psychometric properties of OSCE scores through a case study.
Method: The authors analysed an OSCE data set (comprised of 11 stations) for 80 fourth year medical students based on the polytomous Rasch model in an effort to answer two research questions: (1) Do the clinical tasks assessed in the 11 OSCE stations map on to a common underlying construct, namely clinical competence? (2) What other insights can Rasch analysis offer in terms of scaling, item analysis and instrument validation over and above the conventional analysis based on classical test theory?
Results: The OSCE data set has demonstrated a sufficient degree of fit to the Rasch model (χ2 = 17.060, DF=22, p=0.76) indicating that the 11 OSCE station scores have sufficient psychometric properties to form a measure for a common underlying construct, i.e. clinical competence. Individual OSCE station scores with good fit to the Rasch model (p > 0.1 for all χ2 statistics) further corroborated the characteristic of unidimensionality of the OSCE scale for clinical competence. A Person Separation Index (PSI) of 0.704 indicates sufficient level of reliability for the OSCE scores. Other useful findings from the Rasch analysis that provide insights, over and above the analysis based on classical test theory, are also exemplified using the data set.
Conclusion: The polytomous Rasch model provides a useful and supplementary approach to the calibration and analysis of OSCE examination data.
Keywords
peer-reviewed
