CATEGORICAL-DATA ANALYSIS IN PUBLIC-HEALTH

A greater variety of categorical data methods are used today than 15 y ears ago. This article surveys categorical data methods widely applied in public health research. Whereas large sample chi-square methods, l ogistic regression analysis, and weighted least squares modeling of re peated measures once comprised the primary analytic tools for categori cal data problems, today's methodology is comprised of a much broader range of tools made available by increasing computational efficiency. These include computational algorithms for exact inference of small sa mples and sparsely distributed data, conditional logistic regression f or modeling highly stratified data, and generalized estimating equatio ns for cluster samples. The latter, in particular, has found wide use in modeling the marginal probabilities of correlated counted, binary, and multinomial outcomes. The various methods are illustrated with exa mples including a study of the prevalence of cerebral palsy in very lo w birthweight infants and a study of cancer screening in primary care settings.