Practical Statistics Simply Explained.

Rating:
Summary: If you only own one statistics book...
Review: If you only own one statistics book, this should be it! In than 20 years of working with laypeople, engineering and biology undergraduates and grad. students and medical students, this has proven to be the most reliable, intelligible and accessible single volume on how to treat experimental and observational data in a sound, statistically valid, way. Of particular value is a table in the back (Guide to Significance Tests) which takes a "leaf identification" approach to selection of the most appropriate test. Many good examples and problems with solutions makes this an excellent self study text.

Jonathan Black, Professor Emeritus, Clemson University.

Rating:
Summary: Simple, clear, and hard-working
Review: Maybe you're just getting started as a stats user, maybe you're only an occasional user who doesn't want to relearn the whole field just to get quick answers. In either case, this book might be one of the best around.

The first half of the book is background: a little probability, a little bit about sampling and experimental design, a little about the most basic and common statistics (means, quartiles, etc).

The second half of the book offers a number of basic parametric and non-parametric significance tests. Langley describes each one, when it is applicable, and how to perform the calculations. He doesn't stray far into the slick computing tricks of the pre-calculator days, so the structure of each calculation stays reasonably clear.

The only real weakness in this book is its lack of index. That is especially incovenient because the tables, a staple of most stats books, are interleaved with the text. The table of contents is descriptive, but doesn't replace an index. The other problem, and not really a flaw in the book, is that it's easy to outgrow this text. Even moderately heavy stats users need a bit more theory and background, to allow meaningful adaptation to new conditions. The author has chosen an audience, though, and has addressed that audience and its needs very well. If your skills are beyond those of the intended reader, that's not a fault of the book.

Basic, clear, and reasonably broad - it's everything that an ordinary, casual stat user could want.