How Can You Tell When to Trust a Statistic?

Newspaper and magazine writers often cite statistics in their articles about crime. But where do these statistics come from, and how can you tell if they’re valid or answer your question?

Chapter 7 of Measuring Crime: Behind the Statistics lists eight questions you can ask to judge the quality of a statistic, and applies them to a published statistic about fraud. The questions, based on fundamental statistical principles, are not limited to the subject of crime—they work with statistics about unemployment, public opinion, health, or any other topic you’re interested in.

You can buy the book at Amazon or Barnes & Noble, or directly from CRC Press. Members of the American Statistical Association (ASA) can get a 20% discount by logging into their ASA account: Look for “Discounts & Offers” in the “Members Only” tab.

The endnotes (103 pages of sources and additional material; updated in April 2019) and links to data sets are available here.

Editorial Reviews:

“This book is an excellent primer on handling the mass of data and information researchers are faced with. While it is geared toward followers of criminal justice information, much of the book is a very good introduction to survey techniques discussing their strong and weak points. Most importantly, there are very good guidelines and questions that one should employ before citing any data or using data for policy decisions or for reporting on data such as journalists do. The book is written in a non-technical manner and does a very good job of explaining the nuances in reviewing data. Any researcher who utilizes data would find this valuable. While it has specific examples in the criminal justice field, it really is quite useful for any user of data." (Barry Nussbaum, 2017 President of the American Statistical Association)

"The book aims to achieve two goals: introduce statistical ideas to a general audience and provide an overview of US crime statistics. These are disparate topics, but in the way they are approached here, there is a strong synergy that reinforces both aspects. One the one hand, the reader's natural curiosity about crime (what is it, how are crime events classified and reported, how reliable are the numbers you see in the newspaper, etc.) will help him/her become interested in the statistical issues and learn these concepts in a practical and concrete setting. And on the other hand, by reading about the statistical issues surrounding crime data, he/she gains a better appreciation for the complexities of crime statistics, eventually acquiring a deeper understanding of them. As a statistician myself, I learned interesting facts about the types of crime, their nomenclature and the possible confusion surrounding them, and how the data are collected and reported. Overall, I think the combination is effective and very well developed in this book." (Jean Opsomer, Westat)

Sharon Lohr