The Bayesian borderline regression method: Identifying pass marks for small cohorts

摘要

Dear Sir The borderline regression method is commonly used in Objective Structured Clinical Examinations (OSCEs) to identify pass marks for OSCE stations. It is a simple linear regression with an independent variable (i.e. a global rating). Like other statistical procedures, some practical issues which surround the assumptions of regression analysis should be considered. One of these issues is the sample size. The required sample size for statistical significance is influenced by the alpha level, the desired power, the effect size and the number of predictors (in borderline regression there is only one predictor, the global rating).