This report is based upon the book “Damned Lies and Statistics: Untangling Numbers from the Media, Politicians, and Activists”, written by Joel Best and published by University of California Press in 2001. Joel Best, a professor of sociology and criminal justice at the University of Delaware, has written a highly readable treatise on statistics, and how we can become better consumers of the statistical information that permeates the environment in which we live.
Joel Best is a sociologist and, as a result, this is not a book about the mathematics of statistics, but about its sociology.
That is, a book about the ways in which bad statistics are generated and spread through society. The title of the book comes from Mark Twain’s famous phrase “lies, damned lies, and statistics”, which is usually interpreted as grouping statistics with lies. A more critical title would come from the phrase “figures don’t lie, but liars can figure”.
Despite its cynical title, Best’s book is one of the best ways to learn how to cease being awestruck by statistics, and to start critically evaluating them.
In straightforward prose filled with real world examples, Joel Best deconstructs the processes by which social statistics are created and take on a life of their own, primarily through blind and unquestioning repetition by the media. He also delineates how such statistics are sometimes mutated, misinterpreted, misapplied, and manipulated.
In his view, there are no perfect statistics, just better or worse ones. Every statistic involves human choices: defining what to measure, determining how to measure it, deciding whom to count or how to count it, and choosing how to deal with unreported cases (the dark figure) of whatever is being counted. Not only does every statistic contain identifiable, though generally unrecognized strengths, weaknesses, and dark figures, but many of the most controversial and heavily publicized statistics are created by people in advocacy positions.
Social statistics – statistics about social problems, such as prostitution or suicide – are often produced by activists who are concerned about the problem, and may exaggerate it. When not produced by activists, statistics are often a product of government, which may be motivated in the opposite direction of the activists, to play down a problem. A quick summary of the issues and topics in this book offers a good overview of clear thinking on statistical issues. Chapter 1, “The Importance of Social Statistics”, explains where statistics come from, how we use them, and why they are important.
Chapter 2, “Soft Facts”, discusses sources of bad statistics. Guessing, poor definitions, poor measures, and bad samples are the primary sources of bad statistics. Good statistics require good data; clear, reasonable definitions; clear, reasonable measures; and appropriate samples. Chapter 3, “Mutant Statistics”, describes the methods for mangling numbers. Most of these arise from violating the four requirements of good statistics, but a new problem arises here. While it is relatively easy to spot bad statistics, mutant statistics require a second level of understanding.
As statistics mutate, they take on a history, and it becomes necessary to unravel the history to understand just how and why they are mutant. Transformation, confusion, and compound errors create chains of bad statistics that become difficult to trace and categorize. Chapter 4, “Apples and Oranges”, discusses the dangers of inappropriate comparison. Dangers arise when comparisons over time involve changing and unchanging measures, and projections. Comparisons among places and groups lead to problems not merely in the data measured, but in the ways the data may be gathered and collated.
Comparison among social problems also creates unique difficulties. Best offers logic of comparison to help the reader understand how to make sense of good comparison and bad comparison. Chapter 5, “Stat Wars”, describes the problems that arise when advocated use questionable numbers to make a case. Chapter 6, “Thinking About Social Statistics”, sums up Best’s advice on understanding statistics – don’t be awestruck in the face of numbers, and don’t be cynical about them, he suggests, be critical and thoughtful.