Academic Test Bias

The claim that a test of academic achievement is biased is often equated with the claim that the test is unfair. From the perspective of Deaf Studies, the bias or unfairness charge typically depends on the examinee’s identity or status group—that a test is seen to favor hearing examinees and disadvantage Deaf examinees. A commonly cited example is when a test includes an item that depends on distinguishing or recognizing sounds in order to answer it correctly. Though such items are rare, they tend to exhibit unequal group-based performance: Hearing examinees outperform Deaf examinees. This group-based difference may signal bias, which points to measurement concerns. Where such a test is used to select, place, or certify examinees, this group-based difference in performance may have an adverse impact, which raises questions of fairness. In the field of educational measurement, the terms bias and fairness are related but distinct. Bias refers to group-based differences in performance between examinees who are otherwise equally proficient, knowledgeable, skilled, or competent. Fairness refers to the consequences of test score use. The performance differences that establish bias, by themselves, do not mean that a test is unfair. Bias becomes unfairness only when the explanation for the differences is unrelated to the content of the test. Three concepts are required to determine whether academic achievement tests are biased or unfair: construct measurement, access and opportunity, and use and consequences. In the explanations that follow, the relationship between these three testing concepts and those of hearing status and Deaf identity is considered. In this discussion, hearing status (e.g., hearing, hard of hearing, or deaf) refers to examinees’ relationship to sound, which includes both environmental sound and spoken language. Deaf identity refers to examinees’ relationship to a sociolinguistic community in which a signed language is a central feature.