Instructors' Teaching Performance and Students' Grade Expectations : An Application of the Stochastic Frontier Model

Student evaluation of teaching (SET) scores have been used widely to assess instructors' performance in teaching, and it has been argued that instructors may "purchase" better evaluation scores by inflating students' grade expectations. In this paper, using a stochastic frontier model, we explicitly control for the grade expectation not only as a regressor but also as a factor in instructors' inefficiency in reaching the frontier. This method permits a separation of unobserved factors that are distributed randomly ("luck") from an instructor's unobservable inherent teaching skill (efficiency or lack thereof). After controlling for various instructor, student and course characteristics we find that, with students' higher grade expectations, instructors earn higher SET scores and instructors' inefficiencies decrease