Alternatives to Configural Frequency Analysis

  • Peter Ihm
  • Ingeborg Küchler
Conference paper
Part of the Studies in Classification, Data Analysis, and Knowledge Organization book series (STUDIES CLASS)


The aim of Configural Frequency Analysis (CFA) is the search for outliers or ‘types’ (subdivided into ‘types’ and ‘antitypes’) in a sample of d-dimensional finite vectors, generally represented in a d-dimensional contingency table. Type search is done by analysis of residuals. It can be shown, however, that this technique may be misleading. The use of interpolated (deleted) residuals and/or other techniques will give better results. Deletion of entries results in incomplete tables. Expected values can be computed with the aid of Iterative Proportional Fitting (IPF). The analysis of logarithmic expectations leads to equation systems similar to those occurring in log-linear models. There is no restriction to the independence model assumed in CFA. The Markov chain as example of a more general but still simple model is treated in this paper.


Contingency Table Independence Model Indicator Matrix Type Search Structural Zero 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Deming, W. E., Stephan, F. F. (1940): On a least squares adjustment of a sampled frequency table when the expected marginal totals are known. Ann. math. statist, 11, 427–444 CrossRefGoogle Scholar
  2. Hadi, A. S., Simonoff, J. S. (1993): Procedures for the Identification of Multiple Outliers in Linear Models. JASA, 88, 1254–1272 Google Scholar
  3. Ihm, P. (1986):Diagnostics and robust estimation in multivariate contingency tables. In: E. Diday et al. (eds.): DataAnalysis and Informatics. Proc. TVth Symp. on Data Analysis and Informatics, Versailles 1985. North Holland, Amsterdam,429–442 Google Scholar
  4. Kieser, M. (1991): Identifikation von Syndromen und Typen mit Methoden der Kontingenztafelanalyse. Dissertation, Univ. Heidelberg.Google Scholar
  5. Kieser, M., Victor, N. (1991): A test procedure for an alternative approach to Configurai Frequency Analysis. Methodika 5, 87–97. Google Scholar
  6. Krauth, J. (1993): Einfährung in die Konfigurationsfrequenzanalyse (KFA). Ein multivariates nicht- parametrisches Verfahren zum Nachweis und zur Interpretation von Typen und Syndromen. J. Beltz, Weinheim, Basel.Google Scholar
  7. Krauth, J., Lienert, G. A. (1973): Die Konfigurationsfrequenzanalyse und ihre Anwendung in Psychologie und Medizin. K. Alber Freiburg i. Br. Google Scholar
  8. Lienert, G. A., Ploog, W. D., Von Eye, A. (1993): Inverted configurai types derived from incomplete contingency tables: Q-CFA. Biometrical Journal 35, 259–266.Google Scholar
  9. Von Eye, A. (1990): Introduction to Configurai Frequency Analysis. The Search For Types and Anti- types in Cross-Classifications. Cambridge Univ. Press, Cambridge.Google Scholar
  10. Victor, N. (1989):An alternative approach to Configurai Frequency Analysis. Methodika 3, 61–73. Google Scholar
  11. Victor, N., Kieser, M. (1991): Identification of types in contingency tables. Bull. Int. Statist. Inst., 54. 4, 691–692 Google Scholar

Copyright information

© Springer-Verlag Berlin · Heidelberg 1996

Authors and Affiliations

  • Peter Ihm
    • 1
  • Ingeborg Küchler
    • 2
  1. 1.Institute für Medizinische BiometriePhilipps UniversitätMarbugGermany
  2. 2.Institute für Biomathematik und Informatik, CharitéHumboldt UniversitätBerlinGermany

Personalised recommendations