Ignore epidemiology claims that chocolate is healthy. It is not, claiming it is requires the same suspect correlation that "suggests" weedkillers causes human cancer and acupuncture prevents COVID-19. No science involved. Mars, Incorporated was funding the Chair in Nutrition at UC Davis when a lot of those claims came out and while it's not the case that academics are creating results-for-hire any more than industry scientists are, it is the case that government and companies only fund people whose work they like.

There is no way that eating more chocolate will reduce your risk of diabetes when weight is the absolute highest risk factor for getting it.


Eat that every day and you will no longer be a model.

Just because believers in chocolate promote health claims that make no sense - "antioxidant" is the Smurf of nutrition, it can be anything supplement salespeople want it to mean - does not mean you should abstain, like you should cigarettes or alcohol, it just means it should remain a treat. If you are already obese, discipline is more important than chocolate. For many, a cheat day turns into two and then they give up.

Before you buy into claims about chocolate health, understand what epidemiological correlation is, and how it is not science. It is instead EXPLORATORY. Everyone has read "correlation is not causation" but a cliché doesn't always hit home in recognition. Epidemiology can show you are more likely to want to have a baby since Trump got elected, that organic food will give that baby autism, and if you feed them infant formula they will get child lower grades in school. 

Everyone can see what the confounders are in that. Yet those are all just as valid as claims chocolate lowers heart disease. They rely on epidemiologists who have fetishized statistical significance beyond all comprehension and, in the case of food, influencers and academics who desperately want their contrarian hot take to sound legitimate.

Epidemiology can show that coin flips are prejudice against heads. Or tails. Or six-sided die in Dungeons  &  Dragons hate you. All it takes is enough rows and columns and you are guaranteed to get enough data clumps you can declare statistical significance. Henry Miller, MD, and Stan Young, PhD, did just that with coin flips.



They could get into any epidemiology journal in America because they got the p-factor that epidemiologists think makes a result legitimate.

What it makes it scientists and statistics experts crazy when they see it.



Epidemiology is why I was a signatory on a paper in Nature asking journals to stop using statistical significance to "suggest" a result is valid. It is not. 



Have some chocolate on Valentine's Day buy enjoy it as a treat, not a healthy choice. Anyone who claims its antioxidant, mineral, or nutrient value outweighs the negative effects of the calories may be selling you something.

NOTES:

(1) About a decade ago, while running the American Council on Science and Health nonprofit, I wrote an article that was not meant to be serious, but actually turned out to be a harbinger of claims. What is even better is this "study", published by the International Journal of Behavioral Nutrition and Physical Activity, was produced by an outfit called University of Traditional Chinese Medicine.