Wednesday, March 7, 2012

Do Movies Drive Adolescents to Drink?

Just in time for the new movie Project X, a new study comes out that suggests that movies with scenes of alcohol consumption apparently leads to more "binge" drinking among teenagers.  The cross-sectional study, conducted in six European nations (Germany, Iceland, Italy, Netherlands, Poland, and Scotland) did in fact find a statistically significant correlation in five of the six countries, and the mainstream media are really eating it up.

However, correlation is not the same as causation.  Famous 18th century philosopher David Hume's criteria for causation requires three criteria:  1) association (correlation), 2) temporal precedence (which came first), and 3) isolation (from all potentially confounding variables).  While the first one was found, the second was not (a cross-sectional study can never determine temporality), and while the third one was attempted, it is very difficult to do in practice, especially with a single study.   So, even the authors concede that the study does not in itself prove causation due to the lack of data on temporal precedence, which really is the sine qua non of causation.  And while criterion #3 (isolation) is arguably a very stringent standard to apply, this study doesn't even meet many of the Bradford-Hill criteria favored by epidemologists.   So we have good reason to be skeptical of this study.

Due to the lack of data on temporality, reverse causation remains a plausible explanation (i.e. “binge” drinkers are more likely to prefer to watch movies about drinking and partying like Animal House, Van Wilder, Superbad, and the new Project X rather than the other way around). And there is always the possibility of residual confounding.  Interestingly, the study only looked at whether participants ever consumed 5 or more drinks in an evening, not whether they currently do or how often.  Also, one of the countries (Iceland) saw no significant association (in fact, it had the "wrong" sign) after adjustment for confounders.  But we at Twenty-One Debunked must point out even if the relationship is truly causal, it does not follow that censorship is the answer.   Better education about both alcohol AND media literacy seems to be a better solution for a country that is supposed to be a free society.

In the USA and Canada (neither were included in the study), we know that self-reported teen drinking and “binge” drinking (except perhaps for American college students) has significantly declined in the past decade or so in spite of the apparent increase in these types of movies.  In fact, high school drinking is at a record low in both countries as of 2011.  In a similar vein, teen pregnancy rates in the USA, though still the highest in the industrialized world, are also at a record low despite the fact that TV and movies today are by far the raunchiest in history.

Perhaps we should listen to the wisdom of sociologist Mike Males.

No comments:

Post a Comment