Beyond Performance in Higher Education: A Longitudinal Study of Confidence–Performance Alignment in Undergraduate Dental Education

Abstract

Confidence is often treated as an indicator of readiness in clinical education, although it may not develop in parallel with assessed performance. This non-randomized longitudinal cohort evaluation followed 90 senior dental students in peer-assisted learning (PAL; n = 45) and faculty-led (n = 45) cohorts across three course-embedded assessment points. Confidence was measured with an eight-item course-specific 1–5 rating and assessed performance with structured pediatric dentistry tasks scored 0–6 using a prespecified rubric. Both measures were rescaled to their theoretical 0–100 ranges. Signed and absolute alignment gaps were analyzed with participant-clustered generalized estimating equations (GEE) and robust standard errors; a 10,000-sample participant-cluster bootstrap and a post hoc baseline-adjusted GEE were used as sensitivity analyses. Baseline confidence and performance differed little between cohorts, whereas the signed-gap standardized mean difference was larger because within-cohort variability was limited. From T1 to T3, the PAL cohort showed larger observed increases in confidence (+7.9 percentage points relative to faculty-led; 95% CI [5.4, 10.4]) and assessed performance (+6.8; 95% CI [4.9, 8.7]). Between-cohort differences in change were not statistically significant for the signed gap (+1.2; 95% CI [−1.0, 3.4], p = .28) or absolute gap (+2.21; 95% CI [−0.43, 4.85], p = .10). In the baseline-adjusted signed-gap sensitivity model, the PAL-minus-faculty difference was +11.59 points at T2 (95% CI [8.71, 14.47], p < .001) and +0.77 at T3 (95% CI [−1.70, 3.24], p = .54), preserving the interpretation of marked midpoint divergence but no detectable endpoint difference. These findings indicate that performance level, discrepancy direction, discrepancy magnitude, and longitudinal trajectory provide different educational information. Confidence and numerical gap scores should therefore be interpreted alongside repeated evidence of assessed performance rather than used as substitutes for achievement.