The preface to the first edition of this book claims that it aims to give an exposition of (linear) statistical modelling complete with a necessary statistical background to allow analysis of a wide spectrum of practical problems by applying the statistical package GLIM (Generalized Linear Interactive Modelling). The second edition is an extended reconstruction based on the facilities of GLIM4, the most recent release of GLIM (by the Statistical Computing Group of the Royal Statistical Society). The authors describe GLIM as a command driven package; commands for manipulation, transformation, display and fitting of data may be entered in any order and saved for later use. GLIM is primarily designed for interactive modelling but may also serve for standard time consuming procedures such as the fitting of models to extended sets of data.

The book covers the following subjects: regression and analysis of variance; binary response data; multinomial and Poisson data; survival data; finite mixtures models; random effects models; variance component models; random coefficient models; variance component model fitting; autoregressive random effect models; IRT models; spatial dependence; multivariate correlated responses. This book will be appreciated by graduates, PhD students and professional statisticians as a tool that provides a comprehensive treatment of the statistical theory with an emphasis on application. A wide range of case studies and a gentle self-contained presentation of GLIM4 complete the monograph, making it an extremely useful publication.

Reviewer:

jste