Regression Methods in Biostatistics: Linear, Logistic, Survival, and Repeated Measures Models (Statistics for Biology and Health)

This new book provides a unified, in-depth, readable introduction to the multipredictor regression methods most widely used in biostatistics: linear models for continuous outcomes, logistic models for binary outcomes, the Cox model for right-censored survival times, repeated-measures models for longitudinal and hierarchical outcomes, and generalized linear models for counts and other outcomes.

Treating these topics together takes advantage of all they have in common. The authors point out the many-shared elements in the methods they present for selecting, estimating, checking, and interpreting each of these models. They also show that these regression methods deal with confounding, mediation, and interaction of causal effects in essentially the same way.

The examples, analyzed using Stata, are drawn from the biomedical context but generalize to other areas of application. While a first course in statistics is assumed, a chapter reviewing basic statistical methods is included. Some advanced topics are covered but the presentation remains intuitive. A brief introduction to regression analysis of complex surveys and notes for further reading are provided. For many students and researchers learning to use these methods, this one book may be all they need to conduct and interpret multipredictor regression analyses.

The authors are on the faculty in the Division of Biostatistics, Department of Epidemiology and Biostatistics, University of California, San Francisco, and are authors or co-authors of more than 200 methodological as well as applied papers in the biological and biomedical sciences. The senior author, Charles E. McCulloch, is head of the Division and author of Generalized Linear Mixed Models (2003), Generalized, Linear, and Mixed Models (2000), and Variance Components (1992).

"This book provides a unified introduction to the regression methods listed in the title...The methods are well illustrated by data drawn from medical studies...A real strength of this book is the careful discussion of issues common to all of the multipredictor methods covered." Journal of Biopharmaceutical Statistics, 2005

"This book is not just for biostatisticians. It is, in fact, a very good, and relatively nonmathematical, overview of multipredictor regression models. Although the examples are biologically oriented, they are generally easy to understand and follow...I heartily recommend the book" Technometrics, February 2006

"Overall, the text provides an overview of regression methods that is particularly strong in its breadth of coverage and emphasis on insight in place of mathematical detail. As intended, this well-unified approach should appeal to students who learn conceptually and verbally." Journal of the American Statistical Association, March 2006

"This is a first-class book dealing with one of the most important areas of current research in applied statistics…the methods described are widely applicable…the standard of exposition is extremely high."
--Short Book Reviews from the International Statistical Institute

"The new chapters (10-14) improve an already excellent resource for research and instruction. Their content expands the coverage of the book to include models for discrete level-1 outcomes, non-nested level-2 units, incomplete data, and measurement error---all vital topics in contemporary social statistics. In the tradition of the first edition, they are clearly written and make good use of interesting substantive examples to illustrate the methods. Advanced graduate students and social researchers will find the expanded edition immediately useful and pertinent to their research."
--TED GERBER, Sociology, University of Arizona

"Chapter 11 was also exciting reading and shows the versatility of the mixed model with the EM algorithm. There was a new revelation on practically every page. I found the exposition to be extremely clear. It was like being led from one treasure room to another, and all of the gems are inherently useful. These are problems that researchers face everyday, and this chapter gives us an excellent alternative to how we have traditionally handled these problems."
--PAUL SWANK, Houston School of Nursing, University of Texas, Houston

Popular in the First Edition for its rich, illustrative examples and lucid explanations of the theory and use of hierarchical linear models (HLM), the book has been reorganized into four parts with four completely new chapters. The first two parts, Part I on "The Logic of Hierarchical Linear Modeling" and Part II on "Basic Applications" closely parallel the first nine chapters of the previous edition with significant expansions and technical clarifications, such as:

* An intuitive introductory summary of the basic procedures for estimation and inference used with HLM models that only requires a minimal level of mathematical sophistication in Chapter 3
* New section on multivariate growth models in Chapter 6
* A discussion of research synthesis or meta-analysis applications in Chapter 7
* Data analytic advice on centering of level-1 predictors and new material on plausible value intervals and robust standard estimators

While the first edition confined its attention to continuously distributed outcomes at level 1, this second edition now includes coverage of an array of outcomes types in Part III:

* New Chapter 10 considers applications of hierarchical models in the case of binary outcomes, counted data, ordered categories, and multinomial outcomes using detailed examples to illustrate each case
* New Chapter 11 on latent variable models, including estimating regressions from missing data, estimating regressions when predictors are measured with error, and embedding item response models within the framework of the HLM model
* New introduction to the logic of Bayesian inference with applications to hierarchical data (Chapter 13)

The authors conclude in Part IV with the statistical theory and computations used throughout the book, including univariate models with normal level-1 errors, multivariate linear models, and hierarchical generalized linear models.