The Reliability and Medical Students’ Appreciation of Certainty-Based Marking

Abstract

Certainty-Based Marking (CBM) involves asking students not only the answer to an objective question, but also how certain they are that their answer is correct. In a mixed method design employing an embedded approach with a quasi-experimental design, we have examined the use of CBM during a 5-week Gynaecology and Obstetrics course. The study was conducted as a non-mandatory revision exam with two additional questionnaires on Moodle. Majority of students perceive CBM as fair (78%) and useful (94%). Most students would immediately want CBM to be used for revision exams, but more practice would be needed for CBM to be used in graded exams. The lowest self-evaluation of knowledge was mostly seen by worst (less than 70% Accuracy) and best achievers (more than 90% Accuracy); the worst achievers probably have knowledge gaps, and the best achievers probably correctly guessed at least one question. Our findings conclude that CBM does not discriminate any learner type (p = 0.932) and does not change the general distribution of the exam scores, since there is no significant differences between Certainty-Based Score (M = 80.4%, SD = 10.4%) and Accuracy (M = 79.8%, SD = 11.1%); t(176) = 0.8327, p = 0.4061. These findings are widely applicable, as learner type study models are used extensively in education. In the future, larger samples should be studied and the implementation of CBM on question types other than MCQ should be investigated.