Modeling Passing Rates on a Computer-Based Medical Licensing Examination: An Application of Survival Data Analysis |
| |
Authors: | André F de Champlain Marcia L Winward Gerard F Dillon Judy E de Champlain |
| |
Institution: | Andrd F. De Champlain is a Senior Psychometrician, National Board of Medical Examiners, 3750 Market Street, Philadelphia, PA 19104;. His areas of specialization are dimensionality, cross-cultural psychometrics, item response theory, and validity.;Marcia L. Winward is a Measurement Associate, National Board of Medical Examiners, 3750 Market Street, Philadelphia, PA 19104;. Her areas of specialization are sociology of education and survey scaling.;Gerard F Dillon is a Measurement Associate, National Board of Medical Examiners, 3750 Market Street, Philadelphia, PA 19104;. His areas of specialization are standard setting and validity.;Judy E. De Champlain is a Biostatistician, Theradex Corporation, Princeton, NJ 08543;Judy. . Her areas of specialization are clinical trials and biostatistics. |
| |
Abstract: | The purpose of this article was to model United States Medical Licensing Examination (USMLE) Step 2 passing rates using the Cox Proportional Hazards Model, best known for its application in analyzing clinical trial data. The number of months it took to pass the computer-based Step 2 examination was treated as the dependent variable in the model. Covariates in the model were: (a) medical school location (U.S. and Canadian or other), (b) primary language (English or other), and (c) gender. Preliminary findings indicate that examinees were nearly 2.7 times more likely to experience the event (pass Step 2) if they were U.S. or Canadian trained. Examinees with English as their primary language were 2.1 times more likely to pass Step 2, but gender had little impact. These findings are discussed more fully in light of past research and broader potential applications of survival analysis in educational measurement. |
| |
Keywords: | CBT longitudinal performance modeling survival analysis |
|
|