Bayesian predictive Configural Frequency Analysis

Configural Frequency Analysis is a method for cell-wise inspection of cross-classifications. CFA searches for patterns of variable categories that occur either more often or less often than expected from a given base model. In this paper, we propose and discuss an alternative notion of types and antitypes that focuses on the likely values of the cell frequencies in future experiments, as opposed to the average values of such frequencies. The idea is developed from a Bayesian point of view.