"Understanding peer effects is critical to evaluating the impact of de facto public school segregation on the achievement of white and nonwhite students. Using a unique panel data set of North Carolina public elementary school students, I estimate a model of achievement production that incorporates heterogeneous responses by students at different points of the achievement distribution, while also allowing for peer spillovers to vary across races and for the formation of different race-based reference groups within the classroom. I find evidence of stronger peer influences within reference groups than across reference groups, the magnitude of which varies substantially across the percentiles of the achievement distribution. I apply my results to evaluate the efficiency and distributional effects of alternative classroom assignment policies. Diversifying peer groups leads to small but fairly uniform improvements in the achievement gap across the achievement distribution."
PDFs can be opened with the free Adobe Acrobat Reader.
Comments