Computerized Adaptive Test Cronbach's Alpha

Cronbach's alpha applied to computerized adaptive test (CAT) data estimates internal consistency reliability under the special condition that different examinees receive different subsets of items. Because the classic formula assumes every respondent answers the same items, its direct application to CAT data violates core assumptions and typically underestimates or misrepresents true reliability, requiring careful adaptation or replacement with IRT-based reliability indices.