Abstract
Anonymous student evaluations of teaching (SETs) are used by colleges and universities to measure teaching effectiveness and to make decisions about faculty hiring, firing, re-appointment, promotion, tenure, and merit pay. Although numerous studies have found that SETs correlate with various teaching effectiveness irrelevant factors (TEIFs) such as subject, class size, and grading standards, it has been argued that such correlations are small and do not undermine the validity of SETs as measures of professors' teaching effectiveness. However, previous research has generally used inappropriate parametric statistics and effect sizes to examine and to evaluate the significance of TEIFs on personnel decisions. Accordingly, we examined the influence of quantitative vs. non-quantitative courses on SET ratings and SET based personnel decisions using 14,872 publicly posted class evaluations where each evaluation represents a summary of SET ratings provided by individual students responding in each class. In total, 325,538 individual student evaluations from a US mid-size university contributed to theses class evaluations. The results demonstrate that class subject (math vs. English) is strongly associated with SET ratings, has a substantial impact on professors being labeled satisfactory vs. unsatisfactory and excellent vs. non-excellent, and the impact varies substantially depending on the criteria used to classify professors as satisfactory vs. unsatisfactory. Professors teaching quantitative courses are far more likely not to receive tenure, promotion, and/or merit pay when their performance is evaluated against common standards.