Configural Frequency Analysis: the Search for Extreme Cells
Configural Frequency Analysis (CFA) asks whether a cell in a cross-classification contains more or fewer cases than expected with respect to some base model. This base model is specified such that cells with more cases than expected (also called types) can be interpreted from a substantive
perspective. The same applies to cells with fewer cases than expected (antitypes). This article gives an introduction to both frequentist and Bayesian approaches to CFA. Specification of base models, testing, and protection are discussed. In an example, Prediction CFA and two-sample CFA are
illustrated. The discussion focuses on the differences between CFA and modelling.
Keywords: Bayesian CFA; Configural frequency analysis (CFA); Dirichlet distribution; antitypes; base models; contingency table; extreme cells; frequentist CFA; protection; types
Document Type: Research Article
Affiliations: 1: Michigan State University, USA 2: Autonomous National University of Mexico
Publication date: 01 October 2004
- Access Key
- Free content
- Partial Free content
- New content
- Open access content
- Partial Open access content
- Subscribed content
- Partial Subscribed content
- Free trial content