The role of learning in returns to college major: evidence from 2.8 million reviews of 150,000 professors

Why do some college majors have much higher returns than others? I ask if differences in returns are due to differences in quality of education across majors. I use a novel dataset in which college students rate courses and professors on difficulty, the level of effort required to obtain an A. Major difficulty correlates positively with study hours, implying students respond to grading standards when choosing study time. Using the American Community Survey, I show that harder majors earn more, under a variety of specifications. To deal with selection concerns I use the College Scorecard to compare graduates from the same university-major but exposed to harder or easier professors due to being in different graduation year cohorts. Those exposed to harder professors earn more one year after graduation. I also use the NLSY97 panel to estimate an event study, finding that difficulty causes lower earnings while in college but higher earnings after graduation and provides access to higher skilled occupations. I estimate that one-third of the variance in returns to major can be explained by differences in learning as proxied by difficulty.