Keele Research Repository

Aim: To evaluate the ability of medical students to accurately predict their performance in written anatomy tests.
Methods: In part 1, 189 Year 1 to 4 students answered 100 best-of-four multiple-choice anatomy questions under examination conditions. In part 2, 58 Year 2 students answered 100 questions of different styles. Immediately after both tests, students estimated their marks. Unpaired t test, ANOVA and Pearson coefficient were used for statistical analysis.

Results: Overall, students underestimated their performance by 16.2±15.2% (mean ± SD), with no improvement in accuracy noted along the years (p = 0.7560). Although there was no significant difference in the mean overall score obtained by gender, females underrated their performance significantly more than males (19.1±13.9% vs 12.8±16.0%, p = 0.0048). Students from other EU counties underestimated their performance significantly more than Maltese students. A weak-to-moderate uphill correlation was noted between improved performance and degree of underestimation (Pearson's r=0.38, 95% CI 0.13 to 0.58, p =0.0033). No difference was found in students' ability to predict their mark in relation to the question type they considered hardest.

Discussion: Underestimation was significantly greater in female students and non-Maltese students from other EU countries, suggesting a lack of confidence disproportionate to their ability. No question type allowed students to better estimate marks, suggesting that accuracy may depend more on exam-taking skills.

Conclusions: Targeted training of self-evaluation skills should be included in the medical curriculum and beyond.