Review

Stat-Spotting is a basic guide to recognizing questionable statistics. The author, Joel Best, is a professor of sociology and criminal justice and the author of two previous books on the misuse of statistics. He focuses mainly on the simple statistics found in news reports, so don’t expect a detailed treatment of experimental design, regression analysis, or analysis of variance. This is simply a guide to identifying numbers that don’t make sense, or that are reported without enough context to make sense of,or that are presented in a way that is biased or misleading.

Since his interest is mostly in sociological statistics he begins by laying out the background data that can put those numbers into their context: US population, number of births and deaths per year, leading causes of death and their frequencies, and so on. He then points out that, for the most part, extreme outcomes tend to occur less frequently than moderate outcomes.

He then describes many ways in which dubious data can make its way into publication: mistakes, guesses, overly broad or narrow definitions, sample bias, etc.

This is an interesting and useful book, requiring no mathematical background, and a good antidote to numerical gullibility.

Metadata