Smoking Jackets

Probably the most important thing I picked up from all those psychology courses I took in college can be summed up in four words: Correlation is not causation. As Basic Psychology put it, "The fact that two variables are correlated says nothing about the underlying causal relationship between them." It's a lesson that many journalists have yet to learn.

Consider how the press covered a a study reported this month in Archives of Pediatric and Adolescent Medicine. The researchers surveyed about 1,200 students in grades six through 12 and found that kids who owned cigarette promotional items such as jackets and backpacks were four times as likely to smoke as those who did not. It hardly seems surprising that owners of tobacco-related merchandise tend to be smokers, and this fact alone tells us nothing about why people smoke.

Yet newspaper coverage portrayed the survey results as evidence that kids smoke because of promotional merchandise. "Tobacco Gear a Big Draw for Kids," announced the headline in The Boston Globe. The story began, "If tobacco manufacturers hope to promote smoking by producing clothing or accessories emblazoned with cigarette logos, research by Dartmouth Medical School suggests that the tactic works well."

Under the headline, "Study: Logos Foster Smoking," Newsday reported that "children who own cigarette promotional items…are far more likely to smoke." The story did not note that this observation is open to more than one interpretation, but it did quote the lead researcher, James D. Sargent, who said, "We found that, in effect, children are being used to market cigarettes to their peers."

The Los Angeles Times headline was more cautious: "Study Links Cigarette Gear, Youth Smoking." The lead said, "Schoolkids who sport clothing and gear emblazoned with cigarette names and logos are four times more likely to smoke than other children, according to a new study suggesting that such promotional items may foster youth smoking." The reporter waited until the 14th paragraph to tell us that Sargent admitted "a survey cannot prove that the promotional items caused the children to start smoking."

A rather important point, you might think. Yet all of the stories I saw either ignored or downplayed it. The Globe, for instance, quoted Sargent's complaint that "the tobacco industry is always going to question the causality. That's pretty much what they did with the link between cigarette smoking and cancer."

Sargent thus suggested that only self-interested nitpickers would be so bold as to point out the glaring weakness in his study. Yet he and his colleagues themselves conceded, "The finding of an association between CPI [cigarette promotional item] ownership and being a smoker could easily be an expression of an adolescent who acquired these items after having made the decision to become a smoker." Not to put too fine a point on it, but duh.

Later in the article, Sargent et al. wrote, "Our study and others published to date are subject to the usual limitations inherent in cross-sectional studies, in that we are unable to infer a direction between the exposure (ownership of a CPI) and smoking behavior, limiting our ability to invoke a causal relationship between CPI ownership and smoking." Translation: We would like to say that promotional items make kids smoke, but this study doesn't show that.

The researchers did not let that quibble get in the way of their policy recommendation: "All CPI distribution should end immediately." And in case you missed the point, an editor's note in a box at the beginning of the article said: "If CPIs were bacteria, no one would be opposed to eradication. So why can't we eliminate this pathogen?"

Cigarette promotional items, of course, aren't bacteria. They are a form of speech that offends many people--including, apparently, the researchers who did this study and the editor of the journal that published it.

In this country, however, we do not usually ban speech just because it's offensive. The purpose of studies like this one is to convince people that Marlboro caps and Joe Camel T-shirts are somehow worse than racist fulminations, flag burning, nude dancing, pornography, and sacrilegious art. That's a tall order, so it helps to have a credulous audience that does not remember much from Psych 101.