Random Item Effects Models
Cluster-speci_c item e_ects parameters are introduced that are assumed to vary over clusters of respondents. The modeling of cluster-speci_c item parameters relaxes the assumptions of measurement invariance. Item characteristic di_erences are simply allowed, and it is not necessary to classify items as being invariant or noninvariant. Tests and estimation methods are discussed for item response models with random item e_ects parameters.
KeywordsMeasurement Invariance Threshold Parameter Item Parameter Item Characteristic Item Discrimination
Unable to display preview. Download preview PDF.