by David Kopel
July 28, 2002, Rocky Mountain News
In the sports pages of the Rocky Mountain News and The Denver Post, you will never, ever read a single word from an author who isn't exactly sure about the rules of sport, nor will you ever see deliberately fraudulent misuse of statistics, or the use of statistics that are self-evidently far- fetched. If only the newspapers had such high standards for science articles.
Consider the recent study about teen-age suicide, which was reported by the Associated Press in both papers on July 15. The study asserts that 1 million youths aged 12 to 17 attempt suicide every year.
Yet according to the National Vital Statistics Report (Oct. 12, 2001), in 1999 there were 1,857 suicides by persons aged 10-19. The study's data, if true, would imply that only about 1 in 1,000 teen-age attempted suicides results in a death - a figure far below previously published research on the issue.
Gary Kleck - a professor at Florida State University who has written on technical problems in suicide statistics and who is an editorial consultant for 16 sociology and medical journals - said he thought the study's figure about suicide attempts was unrealistically high. He suggested that one reason might be that "to claim a suicide attempt is to claim adult status as a person with problems serious enough to motivate suicide, and adolescents are obsessed with claiming adult status." Kleck added that "even when teen respondents were reporting actual experiences, many of the acts would be better described as 'suicide gestures,' i.e. not attempts to kill themselves but rather dramatic means for communicating to those around them how unhappy they are." At the least, the papers should have put the study in context by including the actual number of completed suicides for the age group.
The newest thing to worry about, we are told, is hormone replacement therapy. Both dailies covered the story with several commendably calm articles by staff writers. Yet the local stories, like the national stories, were credulous in accepting reports that a recent study by the Women's Health Institute had proven that HRT is dangerous.
Iain Murray, director of research for the nonpartisan Statistical Assessment Service, in a commentary distributed by United Press International, explained that the study revealed much less than was claimed: "These studies rely on something called hazard ratios to sort out actual effects from what might be the results of pure chance. A heavy smoker, for instance, has a hazard ratio of 30-to-1 for the risk of contracting lung cancer compared with a nonsmoker. Generally speaking, risk ratios of under 3-to-1 are regarded as suspect. The closer one gets to 1-to-1, the more likely some random or unsuspected factor is to blame rather than the factor being studied. The risk factors for heart attack, stroke and breast cancer identified in this study were 1.3-, 1.4 and 1.26-to-1 respectively. The risk factors for death from these causes were even smaller. Moreover, when the ratios were adjusted to take account of the long time period of the study, they all ceased to be statistically significant."
The Post (June 16) ran a major New York Times story by Timothy Egan claiming that temperatures in Alaska have risen 7 degrees in the last three decades, and blaming global warming. On July 11, the Times published a correction (which did not run in the Post) reducing the temperature change to 5.4 degrees. On July 19, the Post printed a Washington Post story with the same 5.4 degree Alaska factoid.
As detailed by former New Republic editor Andrew Sullivan ( www.andrewsullivan.com ), the 5.4 degree figure is created by contrasting 1966 (one of the coldest years in recorded Alaskan history) with 1995 (one of the warmest). These two extreme years do not represent a trend. According to the Alaska Climate Research Center ( climate.gi.alaska.edu ), about 2.7 degrees of warming took place in Alaska in the 1960s and 1970s, but since then "the temperature in the last 20 years has not changed much in Alaska."
Sullivan concludes that the Times has a practice of "lying" about climate change.
Two weeks ago, the News printed a letter from an anti-smoker group chastising me for citing Jacob Sullum's book For Your Own Good. The letter called Sullum, a writer for the Reason Foundation in California, "part of a stable of scientists and reporters paid by the tobacco industry to downplay the risks of secondhand smoke." Sullum, supposedly, "got his from R.J. Reynolds." Similarly, a Speakout column in the News (July 19) asserted that Sullum has "longtime ties to the tobacco industry."
The truth is that in 1994 Sullum wrote an Op-Ed which was published in The Wall Street Journal. R.J. Reynolds paid him for permission to reprint the article. That one reprint was Sullum's only financial dealing with a tobacco company.