Abstract
Lienert's (1973) original approach to comparing groups with Configural Frequency Analysis (CFA) cannot straightforwardly be generalized to the comparison of multiple groups. The present article proposes a new base model for group comparison with CFA. This model allows researchers to compare multiple groups, to evaluate overall model fit, to take covariates into account, and to conduct exploratory and confirmatory analyses. In confirmatory group comparisons, base models need to be specified in which particular configurations are blanked out, and other configurations are explicitly set equal. Reference is made to existing base models, e.g., the configural model of axial symmetry. Data examples are provided in which individuals are compared. Extensions of the new models are discussed.