Reducing the Impact of Bias in Oral Assessments

Abstract

Oral assessment is an important method to evaluate the learning outcomes of scientific courses. However, there are certain limitations when oral assessment is applied. One of these limitations, which is not apparent in anonymous written exams, is the existence of biases which could lead to unfair (positive or negative) outcomes of the assessment. This could lead to decreased motivation and sense of belonging among minority student groups, potentially upholding or even increasing inequalities. The major issue with biases is that most of them are unconscious, meaning that it is very tough to mitigate. In this manuscript, we analyze the emergence and effect of expectancy-based bias through personal construct theory in order to find approaches to reduce the influence of bias in oral assessment. As such, we address biases existing before interaction with the student (stereotype), emerging from interaction with the student (halo bias), and how these can contribute to a biased idea of the student in the evaluator’s mind. We finally discuss how this can cause cognitive dissonance and biased assessment when student performance is not in line with the teacher's cognitive model of the student and propose solutions on how to deal with this to minimize bias in oral assessment.