About the event

How Artificial Intelligence Can Help Mitigate Bias in Student Evaluations of Teaching

Abstract: This talk addresses bias and classification challenges in the analysis of written comments from student evaluations of teaching (SET). While bias in numerical SET scores is well-documented, less attention has been given to potential biases in open-ended written responses. Written comments from two campuses within the University of California system are analyzed by gender, ethnicity, tenure status, and course type (STEM vs. non-STEM). Findings are that while some combinations of factors favor higher proportions of positive comments, these advantages do not consistently align with gender or race alone. These results suggest that written comments may be less biased and thus more suitable for evaluating instructors compared to numerical scores. To facilitate their use, a machine learning algorithm is developed that predicts whether a written comment is positive, negative, or mixed. Through the incorporation of a neutral zone to address conditional misclassification rates, the algorithm provides a risk-controlled method for efficiently extracting insights from the often substantial volume of comments. Together, these studies contribute to advancing discussions on equitable and effective teaching evaluation practices in higher education.

Speaker Bio: After finishing his MS and PhD degrees in Statistics at Iowa State University, he went to work at AT&T Bell Laboratories.  He started as a member of technical staff, became a distinguished member of technical staff, and then moved into leadership as a technical manager.  He worked with engineers, computer scientists, physicists, and chemists and helped AT&T build highly reliable next generation telecommunications equipment. For most of his time at Bell Labs, he was also a part-time lecturer in the Department of Statistics at Rutgers University. After 17 years at Bell Labs, he decided to switch careers and move to the Department of Statistics at the University of California, Riverside (UCR). He has been a Professor in UCR Department of Statistics for 17 years, director of the UCR Statistical Consulting Collaboratory for 14 years, chair of UCR department of statistics for 7 years, and is the current Vice Provost of Academic Personnel at UCR. A more detailed biography and list of his history can be found on his website.

The Calvin and Jean Long Distinguished Lecture in Mathematics

An endowed fund that brings internationally renowned mathematics scholars to the WSU campus to discuss research and current topics. A public lecture geared toward community members and members without a deep mathematical background is paired with a more in-depth colloquium with faculty and students.

The lecture honors Calvin Long, professor emeritus (1956–92) and former department chair (1970–78) and his wife, Jean.