Mid-Proof Peer Instruction: the diagnostic potential of voting questions in the midst of proofs

Abstract Proofs play a central role in university mathematics, as they fulfil important functions in generating and communicating mathematical knowledge. However, both the experience of many lecturers and empirical studies show that students often do not benefit from proof presentations in lectures to the extent intended. Here we examine the question of what specific diagnostic potential the method of Mid-Proof Peer Instruction can offer in this context. Our study was conducted in courses on real analysis and business mathematics. It shows what valuable information about students’ understanding of proof can be gained from the voting results obtained when using the method. Based on the results, we discuss whether additional targeted input by the instructor might contribute to making peer discussion more productive in specific situations.