Aimed at ‘students and researchers’, this book starts from scratch and covers all the topics in applied statistics you would expect from its title, up to, but not including generalized linear models. There is a lot of text but the writing is clear and relatively jargon-free. The statistical approach is frequentist with emphasis on concepts, how to do the arithmetic by hand (where feasible) and how to use statistical software. Apart from model definitions the use of mathematics is almost completely avoided. There are many fully worked examples but no exercises. As well as being a guide to exploiting statistical software, the book is designed to be a self-contained manual with a full complement of statistical tables.

So what distinguishes this textbook from the many other introductory texts? Firstly, it assigns greater importance to the design and analysis of experiments, 25% of the book's content, and interactions in particular are nicely explained. Secondly, every example includes a ‘Conclusions’ paragraph which converts the technicalities of the statistical analyses into plain English. Thirdly, two styles of statistical software are illustrated, the specialist package Genstat and the more general Microsoft spreadsheet statistical functions, along with the Microsoft Excel Analysis Toolpak.

Although I felt uncomfortable with recommended analysis for ‘regression with replicated values’, where the concept of pure error was ignored, thought too much emphasis was given to chi-squared tests and disliked the use of GLM as an abbreviation for General Linear Model, I would happily recommend this volume to its target readership either as a textbook or reference book.