The Next Phase of Competency Assessment: A Research Roadmap for Social Work Licensure

Social work licensing examinations rely on a single response format: standardized multiple-choice items. The August 2026 Association of Social Work Boards (ASWB) redesign improved the blueprint, consolidating four content areas into three and reducing scored items from 150 to 110 on the basis of a 2024 practice analysis, but left the response format unchanged. This paper asks whether entry-level competence might be measured more validly through multiple assessment modalities, and argues that social work has raised this question since at least 2010 without producing evidence to answer it. It reviews that literature, including demonstrations that a language model selects keyed answers well above chance when the question stem is removed, and distinguishes licensure assessment from the educational competency assessment programs conduct under accreditation standards. It examines documented pass rate disparities, declines to treat format change as an equity remedy, and applies the same fairness questions to the multimodal formats it proposes, including measured racial disparities in automated speech recognition. It treats development cost and candidate volume as constraints rather than details. It concludes with a staged research design built to avoid the self-selection confound that would invalidate a candidate-choice pilot, and with an independent replication of the stem-removal analysis against the redesigned item bank, which needs no regulator cooperation and could begin immediately.