Payback Time : An Experimental Analysis of Teaching Evaluations and Grades

We tackle the identification problem of the causal effects of grades on teaching evaluations by implementing a quasi-experimental design on a sample of 100 students of an undergraduate economics class at NYU. A control group has to fill out the evaluations before getting the exams, and the treated group follows the reverse order. This way we can estimate average treatment effects, and in particular, estimate the causal effect that grades have on the evaluations. We find a strong negative effect, and we argue that this is indeed the lower bound of the real unobserved effect