It's the movies, stupid
Author: John Luik
Article Published: 10/03/2004
Thanks to visiting US academic Stanton Glantz we at last know why young people smoke. In Toronto last week to tell the Ontario Film Review Board that movies containing smoking should receive the 18 A rating (that would prevent anyone under 18 from seeing them unless accompanied by an adult), Glantz told the Post’s James Cowan that seeing on screen smoking is the main reason why teens start smoking. No wonder the Post ran the Glantz story on the front page. Since almost everything about youth smoking, but especially what initiates it, is an immensely complicated and controversial issue, it is nice to know that this difficult issue finally has a neat and straightforward solution: just prevent kids from seeing films with people smoking and they will not smoke.
Glantz’s “answer” to what makes young people smoke must come as a surprise to any number of the dozens of serious researchers around the world who study youth smoking and who turn out hundreds of academic articles a year seeking to understand why, despite everything that young people know about smoking, they still start to smoke. For these researchers, the idea that any single thing, let alone something like seeing someone smoke in a film, could lead over half of the children, for instance, who became smokers in one study, to start smoking, is something of a stretch.
But to be fair to Glantz, he is not making these claims on his own, he is simply the front man for a group of US researchers, Dalton et al, whose shocking study about youth smoking and the movies was published in The Lancet (July 26, 2003). The study looked at 3547 adolescents aged 10-14 who claimed to be nonsmokers. After making an estimate of how many instances of film smoking the young people would see, the researchers contacted 2603 of the participants from 13-26 months latter to determine whether they had started to smoke. What they found was that adolescents with the highest exposure to movie smoking were 2.71 times more likely to start smoking than those with the lowest exposure. As they note, while 10% of the young people started smoking during the follow-up period, “52.2% of smoking initiation in this cohort can be attributed to exposure to smoking in movies. If the observed association with smoking initiation is assumed to be causal, reducing movie smoking exposure in this study to the lowest quartile would have reduced the proportion who initiated smoking during the follow-up from 10.0% to 4.8%.”
Now what makes this study a shocking piece of scientific fluff is not simply its conclusion- which can’t withstand careful scrutiny, but its junk science- that’s the only term for it- epidemiology. Let’s begin with the epidemiology. At its very best, epidemiology provides evidence of a statistical association, not causality. A is associated with B 19 times out of 20, NOT A causes B. There is a strong association in some European countries in certain years between high numbers of storks and increased birth rates but the one is not the cause of the other. Honest practitioners always acknowledge this, particularly when their work is likely to be reported to a lay audience who might read too much into their findings. Notice, however, what these researchers say “If the observed association with smoking initiation is assumed to be causal”. But wait a moment, this is patently dishonest since the observed association can never be shown to be causal and therefore there is no reason to assume it to be causal. Causality can’t come in through the back door.
The problems, however, don’t stop there. As virtually every researcher on youth smoking knows, the risk factors- the things that lead kids to begin smoking- are very numerous, with one well regarded researcher noting over 200 factors. This means that unless you take these factors into account (control for is the technical term) in the design of your research you can never be certain that you have even found a genuine association. Dalton, however only takes account of 14 of the dozens of factors that are suspected to lead to youth smoking and not even, according to many researchers, the 14 most significant. In other words she is saying that of the possibly hundreds of things that influence kids to being smoking, that for the young people in her study it was seeing smoking in movies and only seeing smoking in movies that got them started smoking.
Clearly this claim is false since she didn’t bother to look at most of the other factors to see whether they- not the movies- influenced her kids. It’s the equivalent of saying X might be caused by A, B, C, D, E, F and G and we have only studied A and B but we nevertheless “know” that they cause X. Ten percent of the kids in Dalton’s study became smokers over a two year period, a figure slightly less than average. Of this 10%
Then too the
Youth smoking has suffered from twenty years of bad research and simplistic policy prescriptions pushed by an anti-smoking movement that often appears not only uninformed but perversely resistant to anything other than its own tired dogmas. Part of the way forward to finding an answer to why kids smoke is not to buy into anything as alluringly simplistic and wrong as the Glantz-Dalton show.
-- John Luik