Abstract
A general model for prediction and association is described for the situation in which both criterion and predictor(s) are discrete variables. The model is based on the concept ofproportion of error reduction. The related measure of association is the Goodman and Kruskal L B. The approach can be employed for the general case involving any number of predictors and the related measures of multiple and partial association are described. Although the exact sampling distribution of L B is unknown, the asymptotic standard error has been derived and can be employed for appropriate tests of significance. The prediction model is based on the use of configural patterns. A general comparison of this procedure to alternative procedures for contingency tables is discussed in terms of the purpose of the analysis
Get full access to this article
View all access options for this article.
