A Simulation-Based Approach to Teach Interaction Effects in Postgraduate Biostatistics Courses

Abstract

Simulation-based teaching can be a valuable method for learning statistical concepts. Its practical implementation for health-related subjects is seldomly evaluated. We propose a simulation-based approach to teach interaction effects in a postgraduate biostatistics course. We describe the steps involved in organizing and implementing a simulation-based activity and evaluate its execution. Mainly master and doctoral students in medical sciences (public health/epidemiology) participated to a 3-hr long online workshop on interaction effects. We presented one main learning activity, broken down into six progressive steps, from visualizing an interaction effect to investigating samples leading to erroneous inference. A total of 85 people attended of which 53 filled out an evaluation survey. Most (89%) of the respondents reported that the proposed activities helped them better understand interaction effects. A qualitative content analysis of free-text answers indicated that learning activities focusing on interaction effects in this way were useful for developing an understanding of both the meaning of statistical models and their interpretations. We conclude that a simulation-based approach can be a useful and interactive tool to create a valuable learning experience. The materials provided in this article can serve as a starting point for teachers who wish to implement such a method.