Peer turnover and student achievement: Implications for classroom assignment policy

The role of testing noise in the estimation of achievement-based peer effects

“…(ii) adopted here are compatible with the heterogeneous findings of Luppino (2015) because not only they allow the expectations of the mean cognitive ability of former peers (a 0 (−i)cs,old ) and the unobserved influences (µ ics ) to be correlated with p ics within schools, but also the correlations may vary across schools. Specifically, the conditional expectations of a 0 (−i)cs,old and µ ics given p ics and s can be expressed as their school-specific linear projections on p ics as follows:…”

“…Gibbons and Telhaj (2011) show that peer mobility has a detrimental effect on incumbent students in England's primary schools. However, using the Project STAR data, Luppino (2015) finds that high levels of classmate turnover generate positive externalities to students in center city schools but negative externalities to students in non-center city schools. 10 Assumptions (i) and…”