Abstract
Accurate assessment of dysfunction is central to clinical psychological science, essential for valid conclusions about prevalence, risk, and appropriate intervention. Measures applied without adjustment across diverse racial/ethnic groups may risk errors if measurement equivalence has not been established. We tested this possibility in the domain of youth depression, applying item response theory (IRT) and differential item functioning (DIF) analyses to reports by White, Black, Latino, and Asian youths (N = 2,335) on the most widely-used measure of symptoms, the Children's Depression Inventory (CDI). Analyses revealed that 77% of CDI items were non-equivalent indicators of symptom severity across groups. CDI sum scores exhibited marked over-estimations of group differences and inappropriate classification as "clinically-elevated" for 29% of Latino, 23% of Black, and 10% of Asian youths. Applying DIF adjustment corrected these errors. The study demonstrates a useful strategy for ethnically sensitive assessment, applicable to other symptom domains and ethnic groups.