Advanced Log-Linear Models Using SAS - Tapa blanda

"The intermediate SAS user with a solid background in probability theory will benefit most from this book. Of particular help are the first two chapters, where the author lays the foundation for the discrete distributions and log-linear models. The novice reader who feels comfortable with the topics of probability and statistical distributions will benefit most from this presentation, while those with more experience will benefit from these chapters as review material. The conversational tone makes Advanced Log-Linear Models Using SAS a pleasure to read, and this newest addition to the SAS Press library a useful text for those users who need a reference text and those who need an application-oriented text."

In Advanced Log-Linear Models Using SAS, Daniel Zelterman applies his extensive SAS knowledge and biostatistics experience to illustrate how to use the GENMOD procedure to analyze log-linear models for categorical data. His wide variety of examples illustrate the statistical applications PROC GENMOD can perform. He thoroughly describes the models, provides real data examples, supplies the necessary code, and explains the output from GENMOD. The topics covered include the Pearson goodness-of-fit statistic; tables of categorical data; a review of log-linear model methods for rectangular tables of categorical data; extrapolation methods to estimate population size; new models and distributions for statistical analysis of data; and issues in power analysis and estimating sample size in experiments. The models take advantage of the wide class of generalized linear models and use real data from pharmaceutical studies and epidemiology, wildlife, and government statistics. Statisticians who have a basic understanding both of SAS and of the analysis of categorical data will greatly benefit from this book. The discussion of each model and method emphasizes statistical aspects, such as interpretation of results, rather than programming skills. The numerous examples are used to motivate the theory and methods as they are discussed.