Introduction Assessment of preclinical operative skills is often limited by subjectivity and inconsistent grading. Traditional glance-and-grade methods provide minimal formative feedback. Criteria-based rubrics may improve objectivity, reliability, and learning outcomes in preclinical dental education. The aim of the present study was to test the reliability of a novel, criteria-weightage-based rubric for the evaluation of Class II restorations in preclinical settings. Methods A purposive sample of 10 subject-matter experts and 10 non-subject-matter experts was selected, and the rubric document was shared with them. Face Validity Index (FVI) and Content Validity Index (CVI) were calculated. The validated rubric was pilot-tested to evaluate preclinical competency in Class II cavity preparation (N = 36 BDS second-year students). Agreement between scores of two sets of evaluators using the traditional and rubric methods was used to measure inter-examiner reliability (consistency), and agreement between scores of the same examiners evaluating the same preparation after a seven-day washout period was an indicator of intra-examiner reliability (reproducibility). Results The rubric exhibited an adequate FVI of 0.96 and CVI of 0.98. The rubric's inter-examiner reliability and intra-examiner reliability were significantly higher than those of traditional methods. For the rubric method, the intraclass correlation coefficient (ICC) showed moderate agreement (ICC = 0.526), while the Pearson correlation coefficient indicated a statistically significant positive correlation between observers (r = 0.584, p < 0.001). Overall, the rubric-based method produced consistently higher ICC values for intra-examiner reliability (0.573-0.600) compared with the traditional method (0.265-0.396). Conclusion Based on the findings of the pilot study, the rubric-based assessment method is recommended as the primary tool for evaluating preclinical cavity preparations, as it demonstrated comparatively better inter- and intra-observer reliability than the traditional glance-and-grade approach.
Bains et al. (Thu,) studied this question.