Working Paper: NBER ID: w14354
Authors: Anna Aizer
Abstract: Recent work shows that peers affect student achievement, but the mechanisms are not well understood. I show that peer behavior is an important mechanism, perhaps more so than ability, by exploiting exogenous timing in diagnosis/treatment of ADD among peers that improves peer behavior while holding peer achievement constant. Improvements in peer behavior increase student achievement. Moreover, resources mitigate the negative effects of peer behavior. These findings imply that the optimal response in the presence of peer effects is not necessarily to reorganize classrooms. Rather, existing institutions can modify peer effects by improving behavior and/or mitigating the impact of poor behavior.
Keywords: peer effects; human capital; ADD; education policy; student achievement
JEL Codes: I21; I18; I20
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| children with undiagnosed ADD (I12) | decrease in reading and math test scores of peers (I21) |
| children with diagnosed ADD (J13) | improvements in behavior (D91) |
| children with diagnosed ADD (J13) | no corresponding improvement in academic achievement (I24) |
| improvements in peer behavior (C92) | better academic outcomes (I24) |
| health insurance expansions (I13) | enhance peer behavior (C92) |