Soda and the Scientific Pile-On Effect

We should be skeptical that any one phenomenon causes everything bad

  • Share
  • Read Later
Spencer Platt / Getty Images

Drinking soda is known to contribute to obesity, leading to efforts to remove soda from schools and to Mayor Bloomberg’s recently overturned ban on large-sized sodas in New York City. Last week, new research by Shakira Suglia, Sarah Solnik and David Hemenway in the Journal of Pediatrics correlated soda with aggression and attention problems in young children. Solnik and Hemenway had already published research last year showing a correlation between soda and aggression in teens.

So, is soda leading to the decline of civilization as we know it? Well maybe, but maybe not.

The Journal of Pediatrics study is actually well done. They use a clinically validated measure of aggression (surprisingly rare in aggression studies) and control for numerous other variables. The effects are small, but not insignificant. However, all data were collected from a single source (the mothers of the children in the study), which can sometimes cause spurious correlations; and the authors themselves note they cannot determine cause and effect. But what is interesting to me is how we are beginning to move beyond the obvious consequences of drinking soda to less obvious ones. In their article, the authors suggest they have forthcoming data linking soda to suicidal thoughts in teens.

Soda sounds like some powerful stuff.

In our society’s worries about obesity, soda has occupied a prime bad-boy spot. We love soda, of course, but we sense things have gotten a little nuts with 64-ounce “big gulps.” And once something is identified in society as “naughty,” it becomes a convenient target for scientists to pile on more bad news. The result is sometimes the opposite of “cure all” snake oil, with naughty “bad boys” being linked to every negative outcome imaginable. But just as we should be skeptical that any one medicine can cure everything, we should be skeptical that any one phenomenon causes everything bad as well.

Consider the case of spanking. The issue of whether open-hand disciplinary swats on the behind are appropriate has long been controversial. Spanking has even been banned in many countries, despite the fact that research linking spanking to aggression in children has been inconsistent. Yet, at least the hypothesis that spanking and aggression in children may be linked makes some kind of sense. However, a recent study by Tracie Afifi  and colleagues at the University of Manitoba, in the journal Pediatrics, was reported in the press as saying that spanking in childhood might be linked to negative adult health outcomes like heart disease and arthritis.

Actually reading the Afifi article, they ran eight separate correlations for spanking and medical outcomes, half of which were significant, half of which were not. But news reports highlighted mainly the significant results. The effects in this case were tiny, based on self-report from participants only rather than participants and their doctors (which, as mentioned above, can cause spurious correlations) and not actual diagnoses . And there is no good theoretical reason why childhood spanking should lead to adult arthritis.

This is the “pile on” effect you can sometimes see in science. Something is viewed as naughty and scientists, usually acting in good faith, pile onto it with more and more bad news, sometimes to the point of absurdity. Why? Because Thing X (soda, spanking, caffeine, marijuana, etc.) is our latest cultural demon—and it’s hard to get grant funding, news headlines and professional acclaim by arguing that Thing X may be no big deal after all. This is particularly true given some of the weaknesses of social and medical science. As I discussed in a previous article, statistics can allow scientists to even unconsciously convert non-significant into significant findings. Significant findings are easier to publish than non-significant ones (the long-known “publication bias” problem in science). So the Suglia article linking soda to aggression is interesting, but how many studies finding no links between soda and aggression simply never get published or reported in the news?

I suspect on some level we scientists get a perverse pleasure out of warning people of dangers or perhaps in being sanctimonious about how people “should” live. But I suspect if we go warning crazy, people simply may pay less and less attention to important warnings. Soda and obesity go hand-in-hand in the societal data, but aggression and bullying in youth have been plummeting for decades despite the ubiquity of soda. If society sees that one set of warnings do not match our popular perception of reality, people may experience warning overload and begin to ignore real issues of concern. The temptation to “pile on” and move from cautious science to moralistic finger-wagging can do more harm than good.