Types or antitypes are based on significant deviations from the null hypothesis. Therefore, the meaning of types or antitypes depends on how the null hypothesis is specified. Here, three different base models for the analysis of change using configural frequency analysis (CFA) are presented, two traditional models and one new model: (1) directed configural frequency analysis (DCFA), (2) prediction configural frequency analysis (PCFA), and configural frequency analysis of change (Change-CFA). The new model is the Change-CFA, which is based on the concept of marginal homogeneity in contingency tables; the method uses a probability model that considers equal pre-treatment and post-treatment marginals. Types or antitypes are interpreted as shifts from pre- to posttest. All methods are applied to a dataset from psychopharmacological treatment; the paper supports the notion of flexibility of CFA in testing different hypotheses.