Comparison of 2 possibilities for extracting information in the analysis of (multi-dimensional) contingency tables. Demonstrated with a medical-sociologic example
Enke, H.
Zeitschrift für die Gesamte Hygiene und ihre Grenzgebiete 35(5): 275-279
1989
ISSN/ISBN: 0049-8610 PMID: 2750236 Document Number: 341531
In several papers the author has stressed the importance of qualitative characters in medical studies. From this follows the necessity of analysing more-dimensional contingency tables. In the first part of this paper the most important results are repeated. Two procedures may be distinguished: 1. A three-step procedure, called Elementary More-Dimensional Contingency Analysis (EMCTA) with and without model fitting 2. A combined Residual and Contrast Analysis (RA/CA). By the aid of a medical example it is shown, in the second part of the paper, that it is possible to extract rather easily all information about interesting dependencies be the aid of RA/CA in the case of one response variable.