From Incremental Validity to Decision Utility: A Framework for Intelligence Testing in Education

Intelligence tests predict academic achievement, but their use in educational decision-making remains contested. We develop a decision-analytic framework, centered on a staged decision architecture, to determine when, for whom, and for which educational decisions intelligence testing adds value beyond grades, achievement measures, and contextual evidence. Drawing on psychometric scholarship, a generative account of achievement, and illustrative decision scenarios, we distinguish incremental validity from decision utility. Incremental validity refers to the predictive gain obtained by adding cognitive measures, whereas decision utility refers to the net benefit of using those measures once base rates, capacity constraints, error costs, fairness, and legitimacy are considered. We use the framework to identify conditions in which intelligence testing is expected to be most informative, especially educational transitions, contexts with uneven opportunity, and discrepancy-focused decisions such as underachievement or twice-exceptionality. We also specify minimum conditions for responsible use, including intended use, construct representation, reliability or precision, measurement comparability, predictive bias checks, and monitoring of distributional impact. We conclude that intelligence testing should be used conditionally and sequentially, with achievement and contextual indicators used first and cognitive assessment added only when it is likely to change the decision.