Monitoring and Removing Scored Test Questions for Biased Functioning: A Macro Ethics Case Study on Social Work Licensure

In August of 2022, the Association of Social Work Boards (ASWB) released data showing that test takers of minoritized races, older ages, and whose first language was not English pass ASWB’s licensing examinations at 2-5x lower rates than dominant groups. Although ASWB maintains that its exams “meet or exceed” psychometric standards and “reflect the highest standards of validity and reliability,” (ASWB, 2022a, p.66), ASWB recently reported that minoritized test-takers were 2-6x more likely to fail the clinical examination (Kim & Joo, 2024). This article will begin with a discussion of how ASWB’s methods deviate from best practices in psychometrics--specifically, how ASWB removes scored exam items if they demonstrate biased statistics without informing examinees their exam’s cut score changed and state boards their automatic licensure decisions were invalid. After exploring rational approaches to discerning whether to monitor for biased functioning in scored items, the article will use Tronto’s (1998, 2013) ethics of care framework to discuss how ASWB should respond. The analysis reveals how ASWB and its member boards deny responsibility for problems in its examinations, leading to care responses that are incompetent, unresponsive to feedback, and untrustworthy.