Achievement

Three of our IGERT trainees (Charles Gibbons, Juan Carlos Suarez, and Mike Urbancic) have been working on a methodological study documenting the common misinterpretation of the instrumental variables (IV) estimator in social science research. In the paper ``LATE for School'' they find that comparing estimates from IV to ordinary least squares (OLS) estimates leads to conclusions that are logically flawed. In the paper ``Broken or Fixed Effects'' they find that fixed effects fail to control for the correlation between demographic characteristics and the variable of interest. In both cases they use the literature estimating the private (wage) returns to schooling as well as the public value of schooling (through its effects on incarceration) as case studies where the extant literature is guilty of malpractice.