Evaluating clinical teachers with the Maastricht clinical teaching questionnaire: How much 'teacher' is in student ratings?

Background: Students are a popular source of data to evaluate the performance of clinical teachers. Instruments to obtain student evaluations must have proven validity. One aspect of validity that often remains underexposed is the possibility of effects of between-student differences and teacher and student characteristics not directly related to teaching performance. Aim: The authors examined the occurrence of such effects, using multilevel analysis to analyse data from the Maastricht clinical teaching questionnaire (MCTQ), a validated evaluation instrument, in a veterinary curriculum. Methods: The 15-item MCTQ covers five domains. The authors used multilevel analysis to divide the variance in the domain scores in components related to, respectively, teachers and students. They estimated subsequent models to explore how the MCTQ scores are dependent on teacher and student characteristics. Results: Significant amounts of variance in student ratings were due to between-teacher differences, particularly for learning climate, modelling and coaching. The effects of teacher and student characteristics were mostly non-significant or small. Conclusion: Large portions of variance in MCTQ scores were due to differences between teachers, while the contribution of student and teacher characteristics was negligible. The results support the validity of student ratings obtained with the MCTQ for evaluating teacher performance.