Fizzy rascals

Posted on December 22, 2011


Dateline: 24 October 2011 // Posted by: Graham McNamee
Source: Science Daily

Story: “High Fizzy Soft Drink Consumption Linked to Violence Among Teens”

Summary: A study published online in Injury Prevention, a self-described “international peer-reviewed journal for health professionals and others in injury prevention,” found a correlation between amount of fizzy soft drinks (measured in cans of non-diet soda) consumed and aggressive behavior, including the carrying of a weapon and “perpetrating violence against peers and siblings.” The study was conducted on 1,878 teen participants, all 14-18 year olds in the Boston, Massachusetts area.

The teens were asked how many cans of soda they had consumed in the past seven days and were then divided into two groups: low consumption (up to four cans) and high consumption (five or more). They were also asked whether they had engaged in violent behavior in the past year. The measures of “violent behavior” were: being violent toward siblings, peers, or partners, and carrying a gun or a knife. Findings showed a positive correlation in these violent behaviors and consumption of fizzy soft drinks, particularly in the subgroup of those who had consumed fourteen or more cans in the past week. It is also noted that the rise in percentage of “violent acts” mirrors findings of alcohol and tobacco usage among the same demographic.

Why is this PseudoNews? Simply, correlation is not necessarily causation. For all these researchers know, this data could be interpreted in the exact opposite manner: violence causes the consumption of fizzy soft drinks. The evidence for either claim is equally present; that is to say, it is not. Yet the title of the article itself nearly implies that this causation exists, or at least embeds the idea in the mind of the reader. This is problematic without causation established. Essentially, these findings need not be available to the public. The suggestion of this “link” effectively spreads the notion that this causation exists.

At the least, the article is premature. It mentions that the same rise in violent behavior is identically correlated with tobacco and alcohol use. Certainly this is a curious finding: it begs for a likely unstudied variable (or likely variables) to be identified. Could it be stress? Socioeconomic status? Domestic influences? Location? The article makes no mention of any of these specific possibilities or of the significance of their potential influence.

What features of pseudoscience are on show? The idea that this supposed correlation is indeed a cause/effect relationship lacks parsimony. The article cites no past research suggesting that any feature of a fizzy soft drink is linked with violence. If previous studies existed – and were referenced – linking high sugar or caffeine (or even fizziness, perhaps) intake with violent behavior, this idea would be more feasible, or at least more intriguing. Instead, the only mention of previous findings on this subject is a reference to a court case: “US lawyers have successfully argued in the past that a defendant accused of murder had diminished capacity as a result of switching to a junk food diet.” This is far from acceptable support. Lawyers will argue anything to make a case, and they are certainly not scientists.

It also seems as though a bit of confirmation bias is in play here. This is even reflected in the survey: participants were asked about their engagement in violent behaviors over the past year, and about their consumption of soft drinks within the past seven days – perhaps asking a weekly average would have been more informative. Using these different time frames may have contributed to their results. As well, there seems to be a popular conception linking junk food to unhealthy behavior. I am not necessarily arguing against this idea, but it is likely that – especially given the article’s reference to the aforementioned court case involving junk food – the publishers of this article were interested in furthering this perception.

The moral of the story: Just because an interaction between two things seems to “make sense,” it does not mean that it exists. No, nobody would agree that soda is healthy. And yes, the idea that frequent consumption of sugar and caffeine leads to more aggressive behavior would not be surprising. But this study does not show this to be the case, although its publication will likely lead people to believe that it is. The only thing that this article is good for is sparking interest for further research.