Tuesday, January 02, 2007

The road to hell is ...

... due for resurfacing in Summer 2008 or thereabouts, until which it's paved with well-intentioned stories like unto the following. It seems harsh to call it bad journalism, because its heart is in the right place and ultimately it sort of seems to get somewhere near what the researchers might have been close to finding, and the conclusions are probably good for you even if they're reached on specious grounds. Still, you're left wondering -- what if we actually tried to report accurately on this stuff?

Articles on diet bad for teens?
Study links reading on weight loss to eating disorders

The big type is the fault of the member paper, not the AP (with an implied exception or two we're going to get to in a moment), so to the extent this is "bad journalism," the copydesk deserves a chunk of the blame. From the top: The relevant results of this study aren't about body image, they're about behavior. Stories about assertions should never have question heds, and this study doesn't examine what is or isn't "bad for teens" anyway. And it also "links reading on weight loss" to ... exercise! and eating vegetables! And other stuff like that.

Setting aside the inept treatment in the big type, you can get a sense of why this was waved through. Here's the lede:

Magazine headlines entice teenage girls with promises: "Get the body you want" and "Hit your dream weight now!" But a new study suggests reading articles on diet and weight loss could have unhealthy consequences.

I'll take your word for what the headlines sound like. Trouble is, this study doesn't have anything to do with what sorts of headlines are used to entice teenage girls (or, as the article acknowledges later, with how articles about weight loss are illustrated, or what they say, or anything like that). Articles that say "be happy with your body" are the same as "get the body you want" for these purposes.

Teenage girls who frequently read magazine articles about dieting were more likely five years later to practice extreme weight-loss measures, such as vomiting, than girls who never read such articles, the University of Minnesota study found.

Yep. That's one of the things the study found.

... The study, in January's issue of the journal Pediatrics, adds to evidence that girls' attitudes toward their bodies are shaped by popular culture.

Really? As the authors write: "It is somewhat surprising that associations were found with weight-control behaviors but not with body satisfaction" (emphasis mine). The same can be said for self-esteem, only more so; the p value there is .949.

The new findings were based on surveys and weight-height measurements of 2,516 middle school students in 1999 and again in 2004.

That's not true. The 2004 part of the study (the "panel" part, which at least allows a bit of thinking about cause and effect) involved surveys only. The weight-height measurements were done in 1999. Let's not hear any complaints that that's a quibble. This is journalism. You aren't supposed to make any of it up.

About 45 percent of the students were boys.

Only 14 percent of boys reported reading diet articles frequently, compared with 44 percent of girls. For those boys who did read about weight loss, there was no similar lasting effect on behavior. (Which underscores the importance of not raising Stupid Questions about what is or isn't "bad for teens" in the heds.)

In the new study, it was unclear whether it was the diet articles themselves or accompanying photographs of thin models that made a difference. The study didn't ask teenagers which magazines they read, only how frequently they read magazine articles "in which dieting or weight loss are discussed."

It wasn't "unclear." It didn't even come up. The study doesn't ask about illustrations, or which magazines are read, or what the heds look like, or whether the articles encourage positive body images or the "thin ideal" that the study is justifiably concerned with. It asked one question -- "How often do you read magazine articles in which dieting or weight loss are discussed?" -- and allowed four responses: Never, hardly ever, sometimes, or often. Then, five years later, it asked some behavior and attitude questions.

Now, those aren't irrelevant or uninteresting results. But the point is they say only what they say. (One of the things they say is that -- at least up through "sometimes" -- reading articles about weight loss is associated with the weight-loss behaviors the study calls "healthy," like exercising and eating more fruits and vegetables.)

The study was based on students' self-reports about their behavior and, like all surveys, could be skewed by teenagers telling researchers what they think the researchers want to hear, said study co-author Patricia van den Berg.

Time out to complain about inept writing. A survey can't be skewed by "teenagers telling researchers what they think the researchers want to hear" unless it uses teenagers. But it is subject to social desirability bias like any other survey. Meaning you can state the possibility of error caused by self-reporting as a fact, independent of whether the researcher admits it or not. It's good of van den Berg to point it out, in case the AP had overlooked the possibility, but it doesn't need attribution.

Nathan Christopher, a spokesman for Seventeen magazine, said health is important to the magazine's readers. He wouldn't comment specifically on the study because it was unclear which magazines the teenagers read.

Fair enough. Why should he? More to the point, why did we ask him?Again, it's not "unclear" what magazines are involved. It's never even come up.

Thus, much as this one has good intentions and could end up having good outcomes, somebody needs to throw the flag on it. Is it too much to ask for good execution along with the good intentions?


Post a Comment

Links to this post:

Create a Link

<< Home