The Effect of Peers on Effort and Achievement

Overview

While existence of peer effects in a variety of educational settings is well-established in the literature, the channels through which peer effects arise are less explored. However, understanding the root causes of peer effects is essential for effective education policy design. In this project, we estimate causal effects of peers on effort and achievement using data on 382 students randomly assigned to 107 groups within three large undergraduate economics courses that regularly incorporate small-group classroom activities and discussions.

To quantify the degree of in-class exposure students have to one another, we use unique attendance data based on in-class polling questions. In-course study effort is measured using student reports of weekly and dedicated pre-exam study hours submitted as a part of each midterm and the final exam. Final exam scores and performance on low-stakes standard assessments serve as measures of student learning. Student data collected via end-of-semester surveys is further enhanced by administrative data, including start-of-semester GPA and demographic characteristics.

Our preliminary results show that female peers both decrease female students’ study hours and improve their performance on final exams. Further, we see the importance of heterogeneity in peer effects. High-GPA students assigned to groups with higher proportions of low-GPA peers, see a reduction in weekly effort levels and large gains in test performance, explained through learning-by-teaching. This effect is not symmetric: Low-GPA students in groups with more high-GPA peers are also less likely to exhibit high weekly effort, but they experience no change in assessment performance. Finally, when high-GPA students are grouped together, they seem to improve the learning experience for each other, performing better on low-stakes assessments and final exams, at least partially through increasing each other’s study effort.