Log-linear modeling is a popular statistical tool for analysing a contingencytable. This presentation focuses on an alternative approach to modeling ordinal categoricaldata. The technique, based on orthogonal polynomials, provides a much simplermethod of model fitting than the conventional approach of maximum likelihood estimation,as it does not require iterative calculations nor the fitting and re-fitting to searchfor the best model. Another advantage is that quadratic and higher order effects canreadily be included, in contrast to conventional log-linear models which incorporate linearterms only.The focus of the discussion is the application of the new parameter estimation techniqueto multi-way contingency tables with at least one ordered variable. This will alsobe done by considering singly and doubly ordered two-way contingency tables. It willbe shown by example that the resulting parameter estimates are numerically similar tocorresponding maximum likelihood estimates for ordinal log-linear models.
展开▼
