Student evaluations of teachers (SETs) are often used as a measurement of instructional quality in higher education, however, SETs may be influenced by lenient grading practices and they may not accurately reflect student learning. The purpose of this study is to provide estimates of the effect of grading leniency on SETs that account for unobserved teacher productivity that can affect both grades and SETs and to develop an alternative measure of instructional quality that is based on student learning To accomplish this, I use unique data from the University of Washington that includes a measure of students' relative expected grade in a course. I find that instructors have an incentive to grade leniently in order to increase SET scores and that this incentive remains under both instrumental variable and fixed effects estimations. This suggests that adjustments to SET scores for grading leniency are warranted. I also find that the ranking of principles of microeconomics instructors based on their contribution to student learning in intermediate microeconomics courses has zero correlation with their ranking by SETs, suggesting that SETs are poor measurements of an instructor's impact on student learning. Additionally, there is an apparent gender bias in SETs, given that female principles of microeconomics instructors receive lower SET scores than male instructors, holding student learning and grading leniency constant. The results of these essays suggest that the usage of SETs in administrative decisions about instructional quality should be reexamined. |