Knowledge self-monitoring, efficiency, and determinants of self-confidence statement in multiple choice questions in medical students

Abstract

Background: Multiple-choice question (MCQ) tests are commonly used to evaluate medical students, but they do not assess self-confidence nor penalize lucky guess or harmful behaviors. Based on a scoring method according to the appropriateness of confidence in answers, the study aimed at assessing knowledge self-monitoring and efficiency, and the determinants of self-confidence. Methods: A cross-sectional study of 842 s- and third-year medical students who were asked to state their level of confidence (A: very confident, B: moderately confident and C: not confident) during 12 tests (106,806 events). A bonus was applied if the level of confidence matched with the correctness of the answer, and a penalty was applied in the case of inappropriate confidence. Results: Level A was selected more appropriately by the top 20% students whereas level C was selected more appropriately by the lower 20% students. Efficiency of higher-performing students was higher when correct (among correct answers, rate of A statement), but worse when incorrect compared to the bottom 20% students (among incorrect answers, rate of C statement). B and C statements were independently associated with female and male gender, respectively (OR for male vs female = 0.89 [0.82–0.96], p = 0.004, for level B and 1.15 [1.01–1.32], p = 0.047, for level C). Conclusion: While both addressing the gender confidence gap, knowledge self-monitoring might improve awareness of students’ knowledge whereas efficiency might evaluate appropriate behavior in clinical practice. These results suggest differential feedback during training in higher versus lower-performing students, and potentially harmful behavior in decision-making during clinical practice in higher-performing students.