Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
By exploiting the correlation structure of individual outcomes in a network, we show that a carefully constructed root estimator can identify peer effects in linear social interaction models, when identification cannot be achieved via variation of group sizes or intransitivity of network connections. We establish the consistency and asymptotic normality of the root estimator under heteroskedasticity, and conduct Monte Carlo experiments to investigate its finite sample performance.