Confirmatory factor analysis and item response theory : two approaches for exploring measurement invariance

摘要: This study investigated the utility of confirmatory factor analysis (CFA) and item response theory (IRT) models for testing comparability psychological measurements. Both procedures were used to investigate whether mood ratings collected in Minnesota China comparable. Several issues addressed. The first issue was that establishing a common measurement scale across groups, which involves full or partial invariance trait indicators. It is shown using CFA IRT models, test items function differentially as indicators groups need not interfere with comparing examinees on same dimension. Second, model fit proposed person-fit statistics be judge practical models. Finally, topics future research are suggested. Much debate has been motivated by question how establish measures dimension, way, when administered two more qualitatively distinct (e.g., men women). can also posed follows: Are scores individuals who belong different examinee populations comparable scale? objectives this review linear analysis' (CFA; Long, 1983) (IRT; Lord, 1980) approaches addressing important suggest, way real-data application, advantages disadvantages each approach.