Smoky Screen

Would eliminating cigarettes from movies slash smoking rates?

Last night I watched Backdraft, the 1991 Ron Howard movie that was condemned by anti-smoking activists for showing firefighters with cigarettes in their mouths. There's an obvious irony when smoke inhalation is both an occupational hazard and a way of unwinding, when men who fight fire also like to play with it. At the same time, the constant puffing by Backdraft's characters makes psychological sense: If you think nothing of plunging into a blazing building, the possibility of lung cancer in 40 years is not likely to faze you.

In other words, the smoking in Backdraft is by no means gratuitous; it's there for a reason. For the film's anti-smoking critics, however, the artistic merits of this choice were irrelevant: Even if the smoking scenes made Backdraft a better movie, giving cigarettes to all those tough, courageous men was socially irresponsible because it set a bad example for our nation's youth.

The campaign to censor cinematic cigarettes got a boost recently from a study described by its authors as providing "strong evidence that viewing smoking in movies promotes smoking initiation among adolescents." Published online by The Lancet last week, the report triggered predictable headlines: "Films Get Teens Smoking, Study Says" (Chicago Sun-Times); "Smoking in Movies Entices Youths, Study Says" (The Boston Globe); "Study: Films Promote Youth Smoking" (Newsday); "Movie Smoking Encourages Kids to Light Up" (New Scientist); "Film Smoking Lures Teens" (Long Beach Press-Telegram).

Glancing at these stories, you may have gotten the impression that the researchers—led by Madeline Dalton, a pediatrician at Dartmouth Medical School—randomly assigned adolescents to groups that were shown either smoky or smoke-free movies, then tracked them for a few years to see how many in each group started smoking. But that is not what Dalton and her co-authors did.

Instead, they presented the subjects, students between the ages of 10 and 14 who had never smoked, with lists of movies and asked them which they had seen. Based on their responses, the researchers classified the kids according to how many "occurrences of smoking" they had witnessed. In interviews conducted 13 to 26 months later, 10 percent of the kids said they had tried cigarettes, and the ones with the highest "movie smoking exposure" were the most likely to have done so.

Anyone who paid attention during Psych 101 could tell you this is hardly an open-and-shut case, as the researchers concede (when they're not talking to reporters). "Our results suggest that viewing smoking in movies strongly predicts whether or not adolescents initiate smoking," Dalton et al. write. "Adolescents who viewed the most smoking in movies were almost three times more likely to initiate smoking than those with the least amount of exposure."

Predicts is not the same as causes, and correlation is not the same as causation. That is why Dalton and her colleagues use cautious phrases such as "If the link…proves to be causal…" and "If the observed association…is assumed to be causal…" After suggesting that "eliminating adolescents' exposure to movie smoking could reduce smoking initiation by half," they concede that "the equation might not be that simple, since many factors affect movie exposure and its effect on adolescent behavior." They also note that "we could not separate the effects of an R-rating and smoking content," since R-rated movies were especially likely to feature smoking.

The crucial question is whether 10-to-14-year-olds who have seen a lot of R-rated movies are different from kids who haven't in ways that might affect their likelihood of trying cigarettes. The researchers found that they were. After Dalton et al. adjusted their results for factors such as "school performance, sensation-seeking propensity, rebelliousness, self-esteem, parent education, authoritative parenting, and perception of parental disapproval of smoking," the association between movies and smoking shrank substantially, with the relative risk for the most avid watchers of R-rated films dropping from more than four to less than three.

Unless you assume that Dalton and her colleagues somehow managed to identify and accurately measure every variation in personality or environment that could have affected both movie watching and smoking, you cannot conclude that the association represents a causal relationship. As Fordham University media theorist Paul Levinson told the Associated Press, "the fact that two things seem to be intertwined doesn't mean that the first causes the second….What we really need is some kind of experimental study where there's a control group."

It would be astonishing if such a study validated Dalton et al.'s suggestion that cigarette-wielding movie stars cause 52 percent of smoking initiation. "This effect is stronger than the effect of traditional cigarette advertising and promotion, which accounts for only 34 percent of new experimentation," says anti-smoking activist Stanton Glantz, who wrote an editorial that accompanied the Lancet study.

So according to Glantz, the combination of on-screen smoking with tobacco advertising and promotion accounts for 86 percent of smoking. That leaves only 14 percent for factors such as personality, education, and the influence of parents, siblings, and peers. Indeed, it may turn out that such factors play no role whatsoever, once you take TV shows and music videos into account.