Anonymous grading, suggested solution to professor grading bias

By JAKE KALKSTEIN

Recent studies suggest that grading biases are pervasive throughout higher education. A myriad of variables such as a student’s race, ethnicity, socioeconomic status, age and sex, as well as a student’s personal relationship with instructors, performance on previous work, class attendance and class participation can all elicit grading biases in an academic setting.

Thus, a grading bias does not necessarily have to be driven by a particular animus. Positive feelings toward an individual or group can also evoke a grading bias. In the majority of these cases, however, the influence of grading biases are subtle and nearly impossible for instructors to admit, even to themselves.

One study, published in 2014 by John Malouff, a neuroscientist, found that prior experience with a student biased the grading of written work product.

In the experiment, 159 faculty members or teaching assistants across substantive disciplines were randomly assigned to grade either a poor oral presentation or a good oral presentation, respectively, given by the same university student. Subsequently, all graders assessed an essay written by that student. Analysis showed that graders assigned significantly higher scores to the essay following the good oral presentation.

Another recent study, published in 2015 by Tobias Rausch, examined the effect of personality similarities between students and instructors on judgmental biases of forthcoming student achievements. Analysis showed that students who displayed similar personality traits to their instructor were judged significantly more positively than students who displayed dissimilar personality traits.

If undergraduate institutions are to be bastions of academic integrity, they must be pragmatic by acknowledging grading biases and urging its faculty to implement more objective grading methods.

Grading work without knowing who completed it, also referred to as anonymous grading, is commonly known as the most effective mitigation tool. In fact, anonymous grading can be found ubiquitously in American law schools as well as many graduate programs. Matthew Davis, a visiting UMW economics instructor, has utilized an anonymous grading policy since he first started teaching.

As a TA, Davis realized that students’ previous performance had the potential to bias his grading on subjective assignments such as essays. “If a particular student who had always done well received a poor grade on an assignment, I found myself looking harder to make sure I didn’t miss anything, that maybe they really did have the right idea in there somewhere; to sort of give them that extra attention that I wouldn’t necessarily give a C student…I realized that this slightly differential grading was not fair,” Davis said.

Going forward, Davis felt obligated to implement an anonymous grading policy in which students write their Banner ID numbers on their assignments instead of their names. “It’s very low cost. My students write their IDs on their tests and all I have to do is match their IDs to a roster.”

Question by question grading, another mitigation tool, is also highly effective. As opposed to grading one test at a time, question by question grading prevents the expectation that students’ performance on previous questions will be indicative of their performance on upcoming questions.

Daniel Kahneman, a famous psychologist and Nobel Prize winner, writes on the subject of question by question grading by saying, “the first question I scored had a disproportionate effect on the overall grade…if I had given a high score to the first essay, I gave the student the benefit of the doubt whenever I encountered a vague or ambiguous statement later on.”

Biased grading is also preventable by using detailed rubrics. Rubrics provide an objective and rigid form of grading criteria. This approach works best if the rubric is shared with students. Of course, there are limitations with respect to the efficacy of these grading methods. For example, when students consult an instructor regarding unique assignment topics it will be difficult for students to remain anonymous.

Furthermore, these grading methods will only be useful when assignments or testing involves subjectivity, (e.g. essays or short narrative answers). On the other hand, it is virtually impossible to grade multiple choice, fill-in-the-blank and matching questions in a biased manner because answers can only be right or wrong. Thus, instructors must use discretion.

There is no room for irrelevant factors like pro-activity, age, gender or race to result in one student to be treated differently than the others. Grading has to be unbiased so that all students are held to the same standards and treated fairly. Instructors have a moral obligation to promote objectivity and to that end, to avoid applying bias and other subjective criteria.