Mitigating Bias in Student Evaluations of Teaching

• In Event: Sheraton Roundtable Session Seventeen
  In Roundtable Session: Data-Driven Instructional Design (Table 2)

Abstract

This paper discusses a method for examining bias in student surveys of teaching effectiveness and using such analyses to revise instruments in order to mitigate bias. The researchers examined differences in student ratings by instructor race/ethnicity and gender, controlling for other instructor variables, student demographics, and course characteristics, in order to test the scale for potential race/ethnic and gender bias. The results of these analyses then informed revisions to problematic items, which were then piloted, leading to the replacement of four items. The revised scale was then examined in the same way and showed no evidence of racial/ethnic or gender bias. Proposed recommendations for testing for and mitigating bias on student evaluations of teaching are discussed.

Authors

• Elizabeth K. Lawner, Association of College and University Educators
  Presenting Author

• Meghan Snow, Association of College and University Educators
  Presenting Author

• Theo Pippins, Association of College and University Educators
  Presenting Author

• Jonathan Gyurko, Association of College and University Educators
  Non-Presenting Author