Korean J Med Educ.  2010 Sep;22(3):215-223.

The School Effect on the Reliability of Clinical Performance Examination in Medical Schools

• Yim MK
• Lee GM

Abstract

PURPOSE
The purpose of this study is to test the reliability of the clinical performance examination (CPX) using Generalizability theory (G-theory). Through G-theory, the effects of not only students and tasks but also the school will be analyzed as primary sources of error, which can affect the interpretation of the reliability of the CPX.
METHODS
One thousand three hundred nineteen students from 16 medical schools that participated in the Seoul-Gyeonggi CPX Consortium 2008 were enrolled. In our research design, we suppose that student is nested within school and crossed with task. Data analysis was conducted with urGenova.
RESULTS
According to our analysis, the percentage of error variance was 6.2% for school, 14.9% for student nested within school, 14.4% for task, and 3% for interaction between school and task. An effect of school on students was observed, but the interaction between task and school was insignificant. When student is nested within school, the universe score decreased and the g-coefficient was less than the g-coefficient of the p x t (p: studentm, t: task) design.
CONCLUSION
The results show that generalizability theory is useful in detecting various error components in the CPX. Using the generalizability theory to improve the technical quality of performance assessments provides us with greater information compared with traditional test theories.

Keyword

Clinical performance examination; Generalizability theory; Reliability

MeSH Terms

Humans
Research Design
Schools, Medical
Statistics as Topic