Script concordance testing: more cases or more questions?

摘要

Script concordance test (SCT) is a case based assessment format of clinical reasoning in which questions are nested into several cases. Recent results using Q4 format suggest that nested questions contribute more to reliability of measure than cases. The present study aims at documenting variance components associated with SCT cases and nested questions and to determine what are the optimal number and combinations of cases and nested questions. Data from SCT in three different fields are presented. G study and D study methodology are used to estimate variance component and to determine optimal number and combinations of cases and questions. Questions nested into cases contributed a large amount of score variance (more than 70%). D studies with varying samples show that, depending on the reliability of the test, an optimal number of 2–4 questions nested into 15–25 cases represents the best combination. Nested questions contribute to a significant portion of score variance, with the implication that formulation of up to 5 questions per case is an efficient way to optimize the reliability of SCT scores.