Purpose: During the high-stake examinations such as OSCE(Objective structured clinical examination) or CPX(clinical performance exam), test security is generally accepted as a major concern for test validity. This study was conducted to investigate the effect on examinee's scores of repeated, serial administrations of essentially the same standardized patient(SP)-based performance exam. Methods: A performance-based examination using eight SP cases was administered to 123 senior medical students at Hanyang University School of Medicine. Students were randomly assigned to one of 16 groups of 8 students each. Three groups were tested serially each day, requiring 5 days for the complete administration of the examination. We compared the mean scores of the five groups of the examinees tested on different days with ANOVA and linear trends with multiple regression analyses. Results: For both checklist scores and written scores during the interstation work, the mean scores of the first day groups were significantly lower compared to subsequent groups. And, there were slight linear trends in the scores over the five days. Scores related to case-specific history taking, information sharing, and clinical courtesy were significantly affected by the sharing of information between students. Scores related to patient satisfaction, physical exam, and physician-patient interaction were not influenced by the same pattern of behaviour. Conclusion: Test security may be violated during SP-based performance exams even though the checklists are not accessible to the examinees. It would be desirable for the test-givers to prepare alternative forms of cases for maintaining the validity of SP-based performance exams.

Purpose: During the high-stake examinations such as OSCE(Objective structured clinical examination) or CPX(clinical performance exam), test security is generally accepted as a major concern for test validity. This study was conducted to investigate the effect on examinee's scores of repeated, serial administrations of essentially the same standardized patient(SP)-based performance exam. Methods: A performance-based examination using eight SP cases was administered to 123 senior medical students at Hanyang University School of Medicine. Students were randomly assigned to one of 16 groups of 8 students each. Three groups were tested serially each day, requiring 5 days for the complete administration of the examination. We compared the mean scores of the five groups of the examinees tested on different days with ANOVA and linear trends with multiple regression analyses. Results: For both checklist scores and written scores during the interstation work, the mean scores of the first day groups were significantly lower compared to subsequent groups. And, there were slight linear trends in the scores over the five days. Scores related to case-specific history taking, information sharing, and clinical courtesy were significantly affected by the sharing of information between students. Scores related to patient satisfaction, physical exam, and physician-patient interaction were not influenced by the same pattern of behaviour. Conclusion: Test security may be violated during SP-based performance exams even though the checklists are not accessible to the examinees. It would be desirable for the test-givers to prepare alternative forms of cases for maintaining the validity of SP-based performance exams.

Clinical competence, Performance, Evaluation, Standardized patients, Validity