I very nearly didn't review this book. I'm not really qualified to comment on a book with this title but, as it happens, the book isn't about geostatistics! Instead, it is about the slightly different topic of statistical methods in Earth Sciences where I do have some background (I teach this to undergraduates). That brings me to a second reason why I nearly turned down the review; who needs another ‘stats for geologists’ book when we already have the excellent Statistics and Data Analysis in Geology (Davis, Reference Davis2002)?
McKillup & Darby Dyar tackle this second objection head-on in the book's Preface where they specifically mention the Davis book and explain that their aim was to produce something a bit less daunting (my words not theirs). The result is a book that won't tell you how to produce much statistics from scratch but which will help you understand the results of computer statistical packages and which will help you decide which analyses to choose. If this is what you need then the book succeeds extremely well.
Some readers may be irritated by my pedantic insistence that this book is about statistics in geology rather than geostatistics. The term ‘geostatistics’ did indeed originally mean what you might expect but, over the last 30 years, it has come to refer specifically to spatial statistical methods (e.g. interpolation) and has become a very specialized subject. Geostatistics Explained barely touches on geostatistics in this narrower sense and therefore has a potentially very misleading title. The reason for this faux-pas becomes apparent once the author's biographies are examined. Steve McKillup is a biologist who has previously published a statistics textbook for life scientists. He clearly decided that there was scope to do something similar in geology and turned for help to Melinda Darby Dyar who, like me, is an earth scientist who happens to teach and use statistical methods. Hence, neither author is a geological-stats specialist and they overlooked the problem with the book title. Nevertheless, between them, they have come up with a charming text which succeeds very well in achieving their aims.
What the book does is to provide a friendly introduction to the standard statistical tools from univariate and bivariate methods, for both parametric and non-parametric cases, through to multivariate methods, time series and simple spatial statistics. The first five chapters are a gentle introduction to scientific methodology with the next four providing a grounding in statistical inference. The book then has fourteen chapters giving succinct introductions to the most commonly used statistical methods. I particular enjoyed the explanation of multi-dimensional scaling (MDS) since I've never used this approach and came to it with the eyes of a student trying to grasp a new concept. I'm pretty sure I picked up the main ideas without a struggle and this has to be a good sign!
In summary, reading this book will stop students, researchers and working geologists from using statistics packages as ‘black boxes’. There's now no excuse for not having a basic grasp of what you're doing, why you're doing it and how to interpret (but not over-interpret) the results. However, those using statistical procedures more intensively might want something a bit meatier and, in that market, Davis' book is still the best bet.
