Deskmates: Reconsidering Optimal Peer Assignments Within the Classroom

• DOI: 10.2139/ssrn.3537509
• Publication Date: 2020-02-19T11:32:39Z
• Authors: Li Li; Eric Mak; Chunchao Wang
• Keywords: 4. Education; 0502 economics and business; 05 social sciences

We find that knowledge spillover in the classroom depends on how peer personalities differ. In addressing the issue of within-group sorting, we conducted our field experiment within primary school classrooms where a pair of deskmates forms the smallest possible designated peer group. We further measured the non-cognitive peer difference between a pair of deskmates, defined as the averaged difference in their responses to Big Five personality questionnaire items, in order to understand the social process within each designated peer group. Our estimation results find that knowledge spillover between deskmates works only when their non-cognitive peer difference is relatively small, but a detrimental effect would be obtained otherwise. After taking non-cognitive peer difference into account, optimizing peer assignments yields a 2 − 6% gain in academic achievement after a semester. Recurring reassignments are needed to sustain this optimization gain because assimilation or differentiation takes place between deskmates over time.