Wednesday, July 28, 2010

Statistics

The phrase “lies, damned lies, and statistics” has been attributed to various people, most notably to Mark Twain, and some variation upon it gets trotted out every time somebody doesn’t like the direction a statistical line of argument is going. I spent the summer of 2005 in Toronto, three months that the media referred to as “the summer of the gun.” An usual number of gun-related murders occurred in the city that year, and the media (most notably the local commuter rag, The Toronto Sun) took every opportunity it could to prove the old adage that what bleeds, leads. Worst of all were the pundits and radio jocks, both left and right wing, who positively rubbed their hands with glee at every new killing, because every death reinforced, or so they believed, their thesis that society was going to hell (and sometimes quite literally to Hell). The only thing they differed about was the root cause of it all. Some said it was because of a decline in "values." Others said it was because of rising poverty. Some blamed it on guns and the United States. Others, on TV and video games. The consumer culture. Drugs. Alcohol. Masculinity. Rap music. Hotter summers. In more whispered tones some said the culprit was multiculturalism and immigration.

Occasionally, some courageous fool would attempt to inject some rationality into the discussion. On one talk-radio show, up against a notoriously fickle gadfly of a host, a visiting criminologist threw a series of haymakers that should have positively demolished the sense of crisis. He pointed out that, while the individual deaths were often very tragic, it was nonetheless the case that Toronto’s homicide rate was only about average for Canada as a whole, very far below the average when Toronto was ranked alongside American cities, and – this is the crucial point – actually much lower than it was a generation ago, back in the hallowed past when men were real men and families went to church and TV was inoffensive and Canadians didn’t let all these foreigners in. On average, he said, a typical Canadian could expect to be murdered about once every fifty thousand years. As if on cue, the host replied, “You can use statistics to prove anything", which is what people usually say when they can’t.

A few years ago, a scholar named Barry Glassner wrote an important though admittedly rather commonsensical book called The Culture of Fear, in which he pointed out that people are afraid of the wrong things. Take, for example, school shootings. The statistical fact of the matter is that children are safer from violence in school than just about anywhere else you’d expect to find them, including their own homes, and are more likely to die getting to school than in it. And yet Glassner found that something on the order of three-quarters of television media coverage of children concerned incidents of violence. Another example: people are unafraid to drive to the airport but take all kinds of drugs to get themselves calmly into the air, even though the total number of deaths in the history of commercial aviation – that’s globally – is about half of the annual death toll from car crashes in the United States.

We could go on and on. On a basic level, my point is that the media has a way of amplifying the significance of individual tragedies. I don’t think there’s anything conspiratorial about this – not at the level of basic reporting, at any rate. As a famous Monty Python sketch reminds us, it would be absurd for the media to report that nothing happened. Nonetheless, there is something askew about the fact that a single school shooting will produce days of coverage, while the fact that tens of millions of schoolchildren make it home safely alive every day is ignored. Similarly, in 2007 and 2008, commercial airlines in the United States carried 1.5 billion passengers without a single fatality, but that fact will receive far less attention than the next statistically improbable but nonetheless inevitable commercial air disaster.

All this brings me to an appallingly bad column I read in The Globe and Mail on Saturday. It was a small, rather hysterical article about a serious and emotional issue – drinking and driving, in which the authors did what authors of that sort of column always do: put emotion before reason. The authors began by reporting with alarm that drunk-driving citations had increased by nearly three percent the previous year. (We’ll leave aside the fact that the population increased by a little over one percent last year, since the authors did, too.) Very briefly, the authors acknowledged the possibility that this might simply be the result of the police citing more people, and then, having stumbled over the truth – quoting Churchill now – they picked themselves up and carried on as if nothing had happened. What is the source of this alarming increase of drunk driving they asked? The usual suspects were rounded up, and an assortment of experts with hobby horses were quoted. Youth culture is the problem. Alcohol advertising. Tough economic times. (I'm surprised they didn't blame video games and the Internet.) "Social engineering," the authors write, "stops at the bar counter." Uh-huh. Ask the owner of any of today's blessedly smoke-free watering holes if he thinks that statement is true.

The basic fact of the matter is that the authors failed entirely to prove to drunk driving actually is increasing. We have no idea how many drivers are drunk behind the wheel at any moment. We only know how many citations police hand out. These are two entirely different things, and they might not even be related. For all we know, the actual numbers of drunk drivers could be plummeting. Thus did Canada's national newspaper run a column about the cause of something that might not even exist.

Want to make the world a better place? The first step is to suspend judgment before the facts are in. Debates over momentous and complicated issues such the economy, health care, public safety, climate change, and so forth, are extraordinarily complex. Understanding these issues – let alone finding solutions to the problems that attend them – requires careful thought and consideration. In some cases, such as the issue of climate change, it may even require major study. Yet to read the newspaper or listen to talk radio or watch TV pundits is to encounter people who are paid to do opposite. Their job is to express views on matters about which they seldom have expertise, and to do so through the prism of ideology in order to attract comparably minded audiences, and to reassure them that, yes, not to worry, everything is going to hell.