Mining Text for Bias in Student Evaluations of Teaching

UC Davis, UC Santa Cruz, & UC Riverside, 2021-2022

Project Principal Investigators

Project Abstract

This multi-campus research project has four specific aims: 1) develop a predictive model that efficiently and automatically scans written course comments, and determines the proportions reflecting student satisfaction levels that are positive, mixed, or negative; 2) pilot an implementation of the predictive model at UC Riverside by integrating it into the iEval student teaching evaluation system to assess both practical and cultural implications of augmenting written comments with a summary report showing the proportions of positive, mixed, or negative comments; 3) use the predictive model to investigate the degree of bias in written comments with respect to the gender, ethnicity, and rank of the instructor, and compare the findings to a parallel bias study of the corresponding numerical scores; and 4) evaluate the efficacy of UC Santa Cruz’s recent revision of instructional evaluation questions as an intervention for reducing bias in comments.

Diversity: Faculty and other academic personnel