Food for thought, but don’t believe the label

Michael Sykuta —  10 August 2010

“People who read food labels such as the Nutrition Facts Panel, ingredient lists or serving size are more likely to have healthier diets than those who do not read labels, according to a new study appearing in the August issue of the Journal of the American Dietetic Association.”  So reads the opening line of a news report today. The article goes on to describe a number of correlations between gender, education, income and use of food labels. The article concludes by citing the study:

“If food labels are to have greater influence on public health rates of use will likely need to be increased among U.S. adults,” researchers said, adding that “Low rates of label use also suggest that national campaigns or modification of the food label may be needed to reduce the proportion of the population not using this information.”

Food labeling is a big issue, both for public health reasons and for business reasons among agrifood firms. So I was curious to see just what the study actually shows (or doesn’t). Best I can tell from reading the article, the authors do nothing to control for the biggest underlying correlations in their model: Who reads food labels?  And what kinds of consumers are more likely to be health-conscious in their food consumption decisions?

Oddly enough, the same demographic characteristics that feature prominently in the use of food labels (female, highly educated, higher income) also feature prominently in health-conscious behavior to begin with. The researchers seem to ignore this self-selection bias in their research method. So, do consumers who read the nutrition label eat more healthfully because they read the label, or do people who eat more healthfully read the label because the want to eat more healthfully?  As yet, (to quote a candy-peddling owl) the world may never know.

The researchers find that the “proportion of variance in nutrient intake explained by food label use alone was small (0% to 1.99%).” But don’t let little explanatory power get in the way of a policy recommendation!  The authors go on to claim that, “Despite the limited amount of variance explained by label use, mean differences between users and nonusers for several nutrients were of sufficient magnitude to be of potential public health importance.” They go on to list a series of recommendations about possible labeling changes.

Of course, there was no consideration to whether these labeling changes would have a meaningful health benefit (even assuming their findings on nutrient consumption are meaningful), nor any consideration for the costs associated with changing labels on tens of thousands of food products.  Interestingly, however, there is another article in the same journal issue titled “Labeling Law Could Mean New Career Opportunities for RDs” (registered dietitians).

3 responses to Food for thought, but don’t believe the label


    Matt, this is actually more consumer behavior research than it is medical research. It could just as easily have been done (though hopefully better) by economists. In fact, I’d not be surprised if there is some such research. Perhaps I’ll do some searching in between prepping for classes and revising a couple papers.

    That said, I certainly agree with your lament about the media’s indiscriminate reference to things published in journals (regardless the discipline).


    To quote a popular phrase from my teens, “Well, DUH!” I read food labels because I want to know what is in my food and how nutritious it is or isn’t. Shouldn’t that be obvious?


    The Journal of the American Dietetic Association isn’t even worth reading. Is it even peer-reviewed?

    If you are seeking any type of competent medical studies, controlled for kappas and the like, you need to stick with the major medical journals: Lancet, Nature, JAMA, NEJM, Annals of Internal Medicine, Circulation, etc.

    It’s a crying shame the media quotes these incompetent medical journals as if they were reputable.