Statistical software packages seem to get more userfriendly each year, but as anyone who interacts with novice users of statistics knows, that's not necessarily a good thing. "Back in the day," I think in my more curmudgeonly moments, "we had to actually understand what we were doing in order to get an analysis to run". Of course that's not entirely true: it's always been possible to misuse statistics, unknowingly or otherwise, it's just gotten easier and quicker to do it from your laptop.
If Common Errors in Statistics (and How to Avoid Them) has an overriding theme, it's this: don't let the software do your thinking for you. Good and Hardin's popular guide to the major issues of statistics emphasizes logical thinking in the application and interpretation of statistics, and largely leaves the formulas to other texts. In fact, most of Common Errors in Statistics can be understood by the statisticallyuninitiated, who may need to skip over the few sections which are thick with Greek notation. Most of the text provides a concise guide to the basics of statistics, replete with examples, explained in common language with an emphasis on meaning and understanding rather than calculation. It's a valuable reference volume for more advanced statisticians as well: it's worth having simply to remind oneself of the most common statistical pitfalls before undertaking a new project, and for the references included in each chapter.
Common Errors in Statistics covers all the expected topics: hypothesis testing, choosing an appropriate statistic, reporting results, using graphics, etc. It also includes brief discussions of more advanced techniques, such as nonlinear regression and classification and regression trees. In addition, it includes the most useful glossary I have ever seen in a statistics book, which is arranged by what Good and Hardin term "related but distinct terms". If you have a habit of confusing, say, accuracy and precision, this glossary will set you straight.
Good and Hardin teach online courses at statistics.com and are experienced consultants and teachers. Good is also Operations Manager for the statistical consulting firm Information Research and Hardin is an Associate Research Professor in the Department of Epidemiology and Biostatistics at the University of South Carolina. This practical experience is evident through Common Errors in Statistics as they demonstrate their ability to explain statistical concepts to novices and to predict exactly where an analysis is likely to run off the rails. It would not serve as a textbook in statistics, due to the lack of formulas and exercises, but would be a valuable second text in an introductory course.
Sarah Boslaugh, (boslaugh_s@kids.wustl.edu) is a Senior Statistical Data Analyst in the Department of Pediatrics at the Washington University School of Medicine in St. Louis, MO. She wrote An Intermediate Guide to SPSS Programming: Using Syntax for Data Management for Sage Publications in 2005 and is currently writing Secondary Data Sources for Public Health: A Practical Guide for Cambridge University Press. She is also EditorinChief of The Encyclopedia of Epidemiology which will be published by Sage in 2007.
