Monday, July 18, 2011

Also, bad journalism may cause ignorance? Fisking "MSG May Cause Weight Gain—Chinese Food Worse Than McDonald's"

Lolcat

Nothing irritates me quite so much in lifestyle journalism than the breathless Consumption Red Alert feature — that "investigative health reporting" staple in which a statistics-blind editorialist takes a survey or study out of context, misinterprets or inappropriately frames its conclusions, and backs it up with irrelevant quotes and facts or deliberate distortions in order to warn of some terrifying phenomenon. This article is a prime example. As Simone Walters writes for FYILiving: OH NOES MSG MAY CAUSE WEIGHT GAIN! CHINESE FOOD "WORSE THAN MCDONALD'S!"

How many ways can I rip this article a new one? Let's go line by line:

  • "Many items on a typical Chinese food restaurant menu are more than 1,000 calories. That’s half your caloric intake per day in just one meal": First of all, you're not talking about Chinese food, you're talking about Chinese American food — takeout stuff that bears the same resemblance to "typical" Chinese food as McRibs do to, well, ribs. But even giving Walters the benefit of the doubt here, Chinese food is not meant to be single-serving by definition. It's eaten family style, which means a few containers should serve four people. Only Americans embrace a one-dish-per-person concept for Chinese food, but I suppose the headline "Typical Manner in Which Americans Eat Fake Chinese Food May Cause Weight Gain" doesn't really grab pageviews, does it?
  • "According to a recently published study, MSG (monosodium glutamate), a flavor enhancer most often associated with Chinese food, may also contribute to weight gain.": First of all, MSG is in virtually every processed, packaged food, under names like glutacyl, autolyzed yeast extract, calcium caseinate, sodium caseinate, hydrolyzed vegetable protein, textured vegetable protein, plant protein extract and yeast extract. And while Chinese (and Japanese, and Koreans, and many Southeast Asians) have used it as a seasoning for centuries, the quantity usually used is minimal — it's a finisher, not a primary taste ingredient. There's usually less MSG in even crappy Chin-eez™ takeout than there is in Cheetos, for crying out loud. (Note: Link is to an MSG hysteria site.)
  • "The study measured the dietary MSG intake of 10,095 Chinese adults in order to determine if there was an association between MSG intake and obesity. The study found that participants who consumed the most MSG were 28 percent more likely to be overweight compared to those who consumed the least amount of MSG.": The first rule of Statistics Fight Club? Don't talk about Statistics Fight Club. The second rule? "Correlation does not imply causation." Repeat after me: CORRELATION DOES NOT IMPLY CAUSATION. There may well be an "association" between MSG intake and obesity, but is it because obese people eat more crappy junk food that contains MSG, or is it because MSG causes obesity?

    It gets worse. In FYILiving's own reporting on the study, linked to by Walters, "A survey in the 1990s estimated an average intake of 580 mg/d [of MSG[ for the general population in the United Kingdom. It was, however, two-folds higher in Japan and Korea." The U.K. is 28th on Forbes's list of the fattest countries in the world; Korea is 123rd and Japan is is 163rd. So does this imply that MSG consumption prevents obesity? Nope. But that would be an equally valid (that is to say, invalid) conclusion based on the evidence cited by FYILiving.

    But again, the study itself doesn't suggest that MSG causes obesity — merely that high levels of MSG in the diet of some Chinese consumers is associated with obesity. Now, let's consider what this means. The study was performed in China — an emerging market undergoing a rapid and wholesale societal transformation, in which a rise in affluence is an enormous factor and changing diet is a significant outcome. Given that MSG is in nearly all processed foods, and consumption of processed foods in emerging markets is directly correlated with rising household income, wouldn't it be more likely that the study was measuring the degree to which wealthy people in China can afford to eat junk food? Because, in direct contradiction to Simone Walters' ultimate conclusion, the less affluent people in that study are more likely to be eating Chinese food (as opposed to Western food) — not less. Chinese food is what Chinese people eat when they can't afford McRibs. And in China, as in most emerging markets, chronic obesity is positively correlated with household income (note: PDF link to Youfa Wang's seminal work on Chinese obesity and socioeconomic status). 

  • The final kicker? "A major limitation of this study was regarding the accuracy of the MSG measurement. Like other food additives the intake of monosodium glutamate was difficult to measure accurately. The fact that only a few people had a high body mass index made it inconvenient to establish a straightforward link between obesity and MSG consumption.": In short, the study was drastically flawed in both input and outcome, and the fundamental characteristics of the pool used for research made it "inconvenient" to establish the very conclusion Simone Walters is asserting in blaring headline font in her article. So, move along, nothing to see here, people, but thanks for the clickthrough.
  • It would be remiss of me not to suggest how one might actually design a study to test a direct causative link between MSG consumption and obesity, so here goes: Take several groups of healthy adults of similar age and body mass index but diverse racial/ethnic background; make one a control group given a healthy but otherwise unrestricted diet (e.g., eat as much as you want), and give the other ones an identical diet, except add MSG to their dishes in varying controlled dosages. Keep them on their diets for six months. If there's more weight gain among the MSG groups than the control group, you have a good case for correlation. if there's a scalable effect (e.g. groups consuming more MSG had greater weight gain) then you have a good case for causation. Is that so hard, people? 

I should note that I saw this article on the website of the Wall Street Journal and clicked on it mistakenly thinking it was actual WSJ content. It's not — it's "sponsor content," which nevertheless appears in the same presentation format and font as actual WSJ articles, but represents fodder from "the OneSpot Content Exchange, a paid-distribution network of sponsors" whose "publishers have paid to have their links appear."

Whether it's a good idea for the Journal to be renting out its diminishing credibility to content farmers is a blogpost for another day. 

Posted via email from OriginalSpin

0 Comments:

Post a Comment

Links to this post:

Create a Link

<< Home