ConcepTests in Undergraduate Real Analysis: Comparing Peer Discussion and Instructional Explanation Settings

Abstract

Peer Instruction, first introduced by Eric Mazur in the late '90 s, is a method aiming at active student participation in lectures. It includes conceptual questions (so-called ConcepTests) presented to the students, who vote on answer alternatives presented to them and then discuss their answers in small groups. As professors have been reported to implement several variants of this method, it is highly desirable to understand the specific effects of the individual elements of the method (tasks, voting, and discussions in small groups). In the present study, we focus on the role of the discussion phase (peer discussion). Our study implemented two conditions: Peer Instruction in classical fashion, and a variant, in which peer discussion was replaced with instructional explanation by a tutor. Students in a course on Real Analysis were randomly assigned to the two conditions for two semesters. As far as learning outcomes are concerned, we do not measure these in terms of voting results within Peer Instruction cycles but we are focusing on transfer in terms of results in the final exams of the two semesters. Interestingly, we found no significant difference between the two conditions. Additionally, we had positive evaluations of the use of Peer Instruction in both variants, with no significant differences between the groups either. Regarding affective variables and learning strategies, no difference in the development could be detected. As an important practical implication, these results show that both implemented variants of the Peer Instruction method are justifiable as far as learning outcomes, measured by exam results, or students’ assessment of the method are concerned. Our results put the widespread belief that it is mainly the peer discussion that accounts for the success of the use of ConcepTests into question.