STATS ARTICLES 2010
Sticky Study: Chocolate-depression link ignores actual chocolate content
Rebecca Goldin Ph.D, April 30, 2010
Snickers has almost no real chocolate; Lindt Excellence has lots; so which did the depressed chocoholics actually eat?
Chocolate is a national, if not international, obsession. Two years ago, news stories abounded about the heart health benefits of antioxidants in chocolate; now, however, chocolate is coming under fire as a new study links it to depression.
Mood and chocolate have been associated with one another in countless films and images that pervade the American psyche – the awkward boy gives a box of chocolates to his crush, or the heart-broken young woman eating chocolates in bed as she cries over being jilted. No doubt the cultural obsession with chocolate (and its relationship to that our emotional lives) spurred on researchers to consider the possibility of a statistical association between chocolate and depression.
The study, published this week in the Archives of Internal Medicine, found a strong correlation: people who consume more chocolate are more depressed. The study was reported by the Los Angeles Times with some caveats about causality; even the researchers acknowledge that there is limited knowledge of whether eating chocolate leads to depression, depression leads to eating chocolate, or some third factor (such as stress) leads to both depression and chocolate consumption. Impressively, the author of the Los Angeles Times article noted the difficulty in making causal conclusions.
But a closer look at the study suggests that the results have little if anything to do with chocolate. Perhaps the LA Times should have had someone at the food desk take a look.
Chocolate consumption was based on self-reporting on questionnaires. The first questionnaire used to evaluate chocolate intake, called the Statin Study Questionnaire (SSQ), simply asks “How many times a week do you consume chocolate?” The second one, the Fred Hutchinson Food Frequency Questionnaire (FFQ), similarly asks how much chocolate participants consumed, and specifies that a “medium serving” of chocolate is 28 grams or one ounce of chocolate candy. This is about half a candy bar.
Participants’ moods were independently assessed using the Center for Epidemiologic Studies Depression Scale (CES-D). A positive depression screening result was taken as a score above 16, while a “probable major depression” diagnosis was given to those who scored above 22. Those who scored under 16 were considered to have a negative depression screen. While CES-D is not a diagnostic for clinical depression, the authors use this screen as a proxy for actual depression.
Where’s the Chocolate?
Never mind the fact that “chocolate candy” is a euphemism for chocolate-like candy that does not contain much (if any) chocolate. The Food and Drug Administration specifies that to be called “chocolate,” a candy bar must have at least ten percent cacao. Hershey’s bars, for example, have eleven percent according toThe Washington Post (Hershey’s does not advertise its chocolate content).
And many other chocolate products in the United States have even less. Consider a Snicker’s or Kit Kat bar -- by most Americans’ definition, these are considered “chocolate bars” but they are really only a coating of milk chocolate (with low cocoa content) around a non-chocolate product.
On the other hand, fine dark chocolate has upwards of 85 percent chocolate. The total chocolate percentage is the sum of the cacao solids and the cacao fat (also called cocoa butter). Even the notion of “chocolate” is not a clearly defined food, since the ratio of solids to fat can change. And with white chocolate, at best you get some cocoa butter, but there is no cocoa in it.
Why the chocolate content matters
The difference in purported chocolate consumption among depressed and not depressed people was impressive. Depressed people consumed more than 50 percent more chocolate than those not depressed, and (depending on which groups were compared using which of the surveys) the differences were as high as three times as much.
The overall purported chocolate consumption was not that high; the average participant consumed six servings per month (about three chocolate bars). Those who scored low on the depression test consumed just under five and a half servings per month. Those in the “depressive” category consumed about nine servings per month, and those in the “highly depressive” category consumed 12-16 servings per month, depending on the survey used to evaluate.
Translated into actual chocolate amount, the study has a major problem: what if those who are less depressed eat five and a half servings of chocolate with high chocolate content, and those in the most depressive category eat 16 servings of low-chocolate-content food? The vast range of chocolate content, (from virtually no cocoa in a Snickers bar to 85 percent cocoa in some of the Lindt Excellence bars) easily spans an order of magnitude (a factor of ten) while the range of difference in number of servings per week/number of times per week ranges by about a factor of 3. The error in actual chocolate measurement is so high that it seems unreasonable to make a conclusion comparing heavy-chocolate consumers to less-chocolate consumers. The data is simply too messy.
So what might be going on? Why would people who eat more chocolate candy (containing varying amounts of chocolate) also have more signs of depression? It may reflect on the mental associations we have with chocolate. If it is viewed as a consolation for a bad day, we may just turn to it, even if it has at most a palliative effect.
While we at STATS often advocate that news agencies have experts in health journalism write about scientific studies regarding health, this study might have benefited from a chocoholic’s discerning palate.