How we teach impacts student learning: peer instruction vs. lecture in CS0

In this paper we look at the impact on student learning of how a class is taught. We compare 2 sections of a non-majors CS0 course offered in the same term, by the same instructor, covering the same content and utilizing the same book, labs and exams. One section was taught using standard lecture practices including lecture from slides, live coding and weekly quizzes. The other section was taught using the Peer Instruction (PI) method that actively engages students in constructing their own learning, instead of absorbing understanding from the instructor's explanations. Using a factorial analysis of variance, we find a main effect of instructional method on final exam grade (F (1,200) = 5.87, p = 0.016) with students in the Peer Instruction section scoring an average 5.7% higher than in the standard lecture practices section. We find no significant interactions among gender and grade or class status (lower or upper division) and grade. In a separate analysis, we also find the interaction of instructional method and high school background to be significant (F (1,147) = 7.48, p = 0.007). In discussion we consider the meaning of these results for educators and describe questions for future work.