Comparing Standard-Setting Methods for Determining Minimum Passing Scores in a High-Stakes Veterinary OSCE.

Abstract

Establishing defensible minimum passing scores is a critical challenge in high-stakes objective structured clinical examinations (OSCEs). Direct comparisons of different standard-setting methods and the varying passing and failure rates remain limited in veterinary medical education. This observational, cross-sectional study compared minimum passing scores (MPSs) generated using the Angoff method, Hofstee method, and borderline regression method (BRM) for an integrated, summative veterinary OSCE. Eighty-one third-year veterinary students at Texas Tech University School of Veterinary Medicine completed a 13-station OSCE. Only 10 stations-3 communication and 7 clinical skills-incorporated both a checklist and a global rating scale and were included in these analyses. All stations were assessed by trained examiners. MPS values for Angoff and Hofstee judgments were provided by six external judges who received training on standard-setting methods. Passing thresholds were compared using descriptive statistics and repeated-measures ANOVA with Greenhouse-Geisser correction. Passing scores differed significantly across standard-setting methods (p < .001). Angoff and BRM produced comparable and more stringent passing thresholds across stations, whereas Hofstee consistently generated lower cut-scores and fewer student failures. Pearson (r = 0.60-0.86; p < .001) and Spearman correlation coefficients (ρ = 0.58-0.89; p < .001) demonstrated moderate to strong associations between checklist and global rating scores. These findings indicate that the choice of standard-setting methods substantially influences OSCE passing scores and pass-fail decisions. Careful selection and transparent application of standard-setting methods are essential to ensure fairness and defensibility in veterinary OSCEs.