This book is about generalized linear models as described by NeIder and
Wedderburn (1972). This approach provides a unified theoretical and
computational framework for the most commonly used statistical methods:
regression, analysis of variance and covariance, logistic regression,
log-linear models for contingency tables and several more specialized
techniques. More advanced expositions of the subject are given by
McCullagh and NeIder (1983) and Andersen (1980). The emphasis is on the
use of statistical models to investigate substantive questions rather
than to produce mathematical descriptions of the data. Therefore
parameter estimation and hypothesis testing are stressed. I have assumed
that the reader is familiar with the most commonly used statistical
concepts and methods and has some basic knowledge of calculus and matrix
algebra. Short numerical examples are used to illustrate the main
points. In writing this book I have been helped greatly by the comments
and criticism of my students and colleagues, especially Anne Young.
However, the choice of material, and the obscurities and errors are my
responsibility and I apologize to the reader for any irritation caused
by them. For typing the manuscript under difficult conditions I am
grateful to Anne McKim, Jan Garnsey, Cath Claydon and Julie Latimer.
