Salary studies in the United States and some other countries show a gap between male and female faculty salaries in higher education that widens over time. The present study examines one possible explanation, particularly at research universities, by examining changes in faculty salaries in relation to student ratings of instruction. Student ratings of instruction are often used to evaluate teaching effectiveness. The present study uses data on salaries and student ratings of instruction for every tenured and tenure-track faculty member teaching at least one class during the 2008-2009 academic year at one research university in the Midwest. A regression analysis was conducted regressing faculty salaries on a standardized version of a student rating of instruction variable while controlling for other variables such as market salary, rank, and whether or not the faculty member was in a STEM discipline. As student ratings of instruction increased for male faculty, their salary went up, whereas when student ratings of instruction increased for female faculty, their salary went down. Further information was gathered at the same Midwest university on a work-life survey. Female faculty wanted to spend a significantly lower proportion of their time on teaching than they actually spent compared to male faculty.