Board 56: Using anonymous grading for high-stakes assessments to reduce performance discrepancies across student demographics

Abstract

Classroom assessments like exams are frequently used metrics for evaluating student course performance. However, evaluation methods like exams or quizzes may suffer from implicit bias like the halo or horn effect, introducing grade discrepancies. Moreover, a student's anxiety about being judged by the instructor or teaching staff can cause suboptimal performance. Anonymous grading can help mitigate these issues. Our hypothesis is that by implementing anonymous grading, there will be a reduction, if not an elimination, of implicit bias during grading. An added benefit is the enhancement of the fairness perception among students. The objective of this study is to implement anonymous grading for in-person exams or quizzes in engineering. The first part of the project is on developing a tool for easy implementation of anonymous grading. The software tool uses the class roster and generates unique barcodes corresponding to each student. The barcodes in the form of stickers can be attached to the paper exam. The barcodes and the grades can be scanned using a mobile application. The system automaps the grades to the student providing completely anonymous grading. The second part of the project would focus on data collection. We plan to collect data from several classes, both at the graduate and undergraduate levels. We will perform statistical analysis on the data collected and historical data on non-anonymous grading. This analysis will help us determine if there are grading differences across student demographics. The third part of the project will focus on collecting qualitative and quantitative feedback from students and other faculty members. This will be done through surveys and focus groups. This will help us understand how the students' perception has changed due to anonymous grading.