Influence of the wording of evaluation items on outcome-based evaluation results for large-group teaching in anatomy, biochemistry and legal medicine.

Student learning outcome is an important dimension of teaching quality in undergraduate medical education. Measuring an increase in knowledge during teaching requires repetitive objective testing which is usually not feasible. As an alternative, student learning outcome can be calculated from student self-ratings. Comparative self-assessment (CSA) gain reflects the performance difference before and after teaching, adjusted for initial knowledge. It has been shown to be a valid proxy measure of actual learning outcome derived from objective tests. However, student self-ratings are prone to a number of confounding factors. In the context of outcome-based evaluation, the wording of self-rating items is crucial to the validity of evaluation results. This randomized trial assessed whether including qualifiers in these statements impacts on student ratings and CSA gain. First-year medical students self-rated their initial (then-test) and final (post-test) knowledge for lectures in anatomy, biochemistry and legal medicine, respectively, and 659 questionnaires were retrieved. Six-point scales were used for self-ratings with 1 being the most positive option. Qualifier use did not affect then-test ratings but was associated with slightly less favorable post-test ratings. Consecutively, mean CSA gain was smaller for items containing qualifiers than for items lacking qualifiers (50.6±15.0% vs. 56.3±14.6%, p=0.079). The effect was more pronounced (Cohen's d=0.82) for items related to anatomy. In order to increase fairness of outcome-based evaluation and increase the comparability of CSA gain data across subjects, medical educators should agree on a consistent approach (qualifiers for all items or no qualifiers at all) when drafting self-rating statements for outcome-based evaluation.