An Excellent Diagnosis of the Problem of Deceptive Statistics
One of the most infuriating aspects of the “information age” is that it makes it easy for those who want to undermine freedom to get their way by peddling falsehoods. Statistics can be among the most damaging of falsehoods. Undoubtedly you’ve seen them at work on behalf of meddlers who want something from government: statistics on Americans without health insurance; statistics on homelessness; statistics on deaths of children due to firearms, and so on. The well-documented “innumeracy” that characterizes much of our populace helps the meddlers because few people are able to see when they’re being deceived.
Joel Best’s Damned Lies and Statistics is a level-headed look at the problem of the use (mostly misuse) of statistics in America. He observes that “bad statistics are potentially important: they can be used to stir up public outrage or fear; they can distort our understanding of our world; and they can lead us to make poor policy choices.” The book proceeds to show how bad statistics come into existence, how they take on a life of their own, and how people can spot them.
Best, professor of sociology and criminal justice at the University of Delaware, begins his examination of the problem by citing what he regards as the worst social statistic ever. When serving on a dissertation committee, he read a student’s prospectus that stated, “Every year since 1950, the number of American children gunned down has doubled.” Best checked the citation, an academic journal, thinking that the student had copied the quotation incorrectly. But no, the quotation was accurate. The problem was obvious to anyone with a familiarity with mathematics and the powers of two. By 1980 the number of children gunned down would be two to the power thirty–more than one billion! Obviously and fantastically wrong, there it was, being cited by people interested in propounding the notion that huge numbers of children are killed by guns every year.
The root of the problem is that it is so easy to use statistics to hoodwink people into believing things that government activists want them to believe. Best shows the several ways in which numbers can be cooked to convince people that “Like wow! The government better do something!” His third chapter, “Methods for Mangling Numbers” does a wonderful job of explaining the tricks of the trade. One is the use of questionable definitions to inflate minimal problems into headline makers. A good example is the frenzy over church arson of 1996. By defining as a “church arson” every instance of church fire, including those that could not conceivably have been racially motivated, the crisis-mongers managed to convince Congress and the media that the nation was in the throes of a wave of Klan-like violence directed at black churches. Once its political usefulness had been mined out, the issue quietly went away.
Another reason for bad statistics is bad sampling. The often-repeated “fact” that 83 percent of Internet traffic is pornography is a good example. The researcher who compiled that statistic drew his sample from, as Best writes, “precisely that portion of the Internet where pornographic images were concentrated; it was anything but a representative sample.”
It’s rather easy to create a bad statistic, or to transform a good one into a deceptive one. Then a process kicks in that often results in the widespread dissemination of that statistic:
Once someone utters a mutant statistic, there is a good chance that those who hear it will accept it and repeat it. Innumerate advocates influence their audiences: the media repeat mutant statistics; and the public accepts–or at least does not challenge–whatever numbers the media present. A political leader or a respected commentator may hear a statistic and repeat it, making the number seem even more credible. As statistics gain wide circulation, number laundering occurs. The figures become harder to challenge because everyone has heard them.
The numerous groups that desire to manipulate public opinion in order to get favors know exactly how to put those wheels in motion.
In the author’s view, most Americans are either “awestruck” or “naïve” when it comes to statistics. He would like to see far more of us become “critical,” which is to say, “appreciating the inevitable limitations that affect all statistics, rather than being awestruck in the presence of numbers. It means not being too credulous, not accepting every statistic at face value (as the naïve do).” He does not, however, suggest any means by which we might make any improvement in our ability to understand and analyze statistics. Given that our government-run education system has a strong incentive to keep the citizenry credulous, it’s not easy to see how we can make headway against this educational deficiency.
Nevertheless, the book provides an excellent diagnosis of the problem.
