Thursday, May 28, 2015

Health Reporters Easily Punked by Chocolate Study

Did you read about the new study showing chocolate helps you lose weight? I'm sure regular readers here weren't taken in, but you might want to show something to your friends who keep up with health "news." The authors of that study just revealed that it was a hoax to shine a light on the sloppiness of the health media. The study was real and the authors didn't lie about anything but their credentials, they just did a poor experiment, sent out press releases and paid the impressive-sounding journal The International Archives of Medicine 600 euros to publish it.

The study really did show greater weight loss in the chocolate group than the non-chocolate group and the control group, but...

Here’s a dirty little science secret: If you measure a large number of things about a small number of people, you are almost guaranteed to get a “statistically significant” result. Our study included 18 different measurements—weight, cholesterol, sodium, blood protein levels, sleep quality, well-being, etc.—from 15 people. (One subject was dropped.) That study design is a recipe for false positives.... With our 18 measurements, we had a 60% chance of getting some“significant” result with 0.05. (The measurements weren’t independent, so it could be even higher.) The game was stacked in our favor. It’s called p-hacking—fiddling with your experimental design and data to push p values under 0.05—and it’s a big problem. (Emphasis added.)

In other words, the weight loss results could have been luck--for instance, the women might have been on a certain place in their cycles. If the result had been lower blood sugar, some people in the control group might have been getting a cold or under stress--both things raise blood sugar. As for sleep and well-being, a thousand different things are more likely to affect those than a few bites of chocolate.

But the reporters--and even the prestigious sounding International Archives of Medicine, which purports to "rigorously review" articles--asked about none of this. Even a "fact checker" from Shape didn't ask for many details about the study. Quite a bit of nutritional science is like getting a diploma from a mail-order college in the Caribbean: answer a few questions, write a check, and you're official.  Health reporters recycle the articles until junk science becomes conventional wisdom, the same conventional wisdom you'll get from most medical professionals, because that's mostly where they (in general) learn about nutrition.

But...the chocolate is still good for you, isn't it? They selected bitter chocolate because it's a favorite of "whole food fanatics." Never mind that chocolate is highly processed--you can't even make a proper chocolate bar at home--or that the stuff they used (81% is bitter and chalky as aspirin) isn't the milky, sugary junk food the magazines and web sites showed.

For the full story, read this: http://io9.com/i-fooled-millions-into-thinking-chocolate-helps-weight-1707251800.

10 comments:

Galina L. said...

Once , when I arrived to my mom in Moscow, I found her cupboards well-stocked with a marmalade. She said to me that there were a lot in news about healthy benefits of a marmalade - candies made out of natural fruit juices(wonderful antyoxidants) with a lot of pectin (which magically lowers cholesterol and prevents a colon cancer), which supposed to turn it into a healthiest candy, especially if you remember that sugar is necessary for a brain health. It reminds me the situation in US with a chocolate and a red vine. There are a lot in news and even health blogs what makes candy-eaters feel better about their comfort food. While candies are candies and a vine is a vine.

Larcana said...

I read about this on Gawker...but I think Suppversity put this up as a real well done study, too. I'm not that surprised...I like chocolate but alas, it doesn't like me so I didn't change anything I would normally do. Kinda funny!

Lori Miller said...

And don't forget--marmalade is low in fat!

I'm beginning to think statistics needs to be a required course to get a degree in anything.

Lori Miller said...

I just talked to a Russian passer-by who recognized the service berry bushes in my front yard and said her mother made service berry jam (she had a Russian name for the bushes, though).

Galina L. said...

The fruit preserve made from that berry is very tasty. There are much more eatable berries , especially in the form of preserves than it is used as eatables in US. What is not a fantastic jam, could be used in a creation of vodka-infused drink. Here in US only a cranberry-infused vodka is known. A Hippophae berry is a good example. It is absolutely wonderful, but I saw it in Edmonton only as a decorative bush, no one thought about eating it or even making it into a drink.

Lori Miller said...

I've seen hippophae (sea buckthorn) around, but didn't know the fruits were edible.

I'm not much of a fruit eater, but even like a few service berries right off the bush when they turn black. Maybe rose hips and choke berries (aronia), which are prolific in my yard) could be used for a vodka infusion. I don't know about the choke berries since they fall off the bush, but even the birds won't eat the rose hips. I guess they prefer seeds to fruit--kind of like humans.

Galina L. said...

As far as I know, aronia (if the google translator is correct) is not eatable, in Russia people mix it with a sugar and let it ferment into a homemade alcoholic beverage. Rose hips are not used for eating in normal circumstances because in the middle of the hip are seeds surrounded by a lot of small thin very invasive needs which are better not to touch at all, left alone swallow. The thin layer of flesh is eatable and rather tasty(sweet+sour), but it is very thin.It is high in a vitamin C. In Russia people often collect rose hips, dry it uncrushed and use as a fruit tea.
In Russia a sea buckthorn is considered to be a very, very important medical plant, mostly due to a buckthorn oil which is used for hard to treat burns, trophic ulcers, radioactive burns. The oil people derive from seeds, but tasty berries contain it too. The oil and berries are used in many cases when their anty-inflammatory properties could be desirable. A vodka infusion made with the berries is a nice drink, better than made with cranberries. It is extremely hard to remove berries from a wild plant without a tree damage, people mostly deal with a cultivated variety which was changed from a wild one in a way which made a berries collection possible, or wait till frosty weather which makes possible to detatch berries from a branch without crushing. There are frozen buckhorn (oblepikha) berries for sale in some ethnic eastern European stores in US.

Lori Miller said...

Chokeberries aren't really edible, IMO, either. My mother used to make jelly out of them. It was awful and mostly went uneaten.

As you probably know, rose hips are a common ingredient in fruity teas here, too, especially up in Boulder at Celestial Seasonings where they make a lot of funky teas.

I had no idea sea buckthorn was medicinal. I just knew it as a hardy xeriscape plant.

Lori Miller said...

I just realized I've been using the wrong name for the bushes. They're not service berries, they're golden currants.

Galina L. said...

No I didn't know rose hips were used here in fruit teas. Somehow a fruit tea is the product I don't buy. My mom dry rose hips when the season,and drink it at winter time, but I don't share her liking for it. Like many Russian families, me and my husband take tea drinking too seriously, I buy a loose-leaf tea mostly from middle-eastern stores. Sometimes we add to our tea cranberries, or thinly sliced apple, or lemon.