The 10% Fail: Dieting Girls Statistics (Case Study)

February 23, 2017


About once or twice a month (though sometimes once or twice a week, depending on how much I’m reading at the time), I come across an article or blog that makes some important point that I agree with. Maybe it’s about the need for skepticism, or about politics, or anything else. I’m reading along, nodding in approval in paragraph after paragraph (or assertion after assertion), pleased at thinking about those it might educate.

And, just as my finger is reaching to share or like the post, I wince. The writer or commenter stumbles, making a gaffe or mistake that I can’t in good conscience implicitly endorse. It’s frustrating because I agree with the overall point, and think it merits a wider audience.

It’s like some well-intentioned skeptic writing a piece about why the evidence for Bigfoot (or recovered memories, or alien visitation) is poor, and giving two solid, accurate reasons–followed by a third which is flat-out wrong, or an argument whose premise is embarrassingly flawed. The phenomena happens regularly enough that I’ve taken to describing it (to myself anyway) as The 10% Fail. Ninety percent of it is on target, but the last ten percent undermines the author’s credibility in some way. This issue is a common lament among professional skeptics: a well-meaning but inexperienced skeptic goes on television or gives an interview–ostensibly representing organized skepticism–in which he or she misspeaks or mangles some salient fact in the process of debunking some bogus claim, and that error is then seized upon by opponents as proof that skeptics (writ large) don’t know what they’re talking about.

Whether that bad 10% is enough to contaminate the rest of the person’s opinion or blog is of course a subjective question that varies from person to person. In today’s world of fake news, “alternate facts,” and other misinformation, I take a dim view of it. For me it’s very often a deal killer because I can probably find and share someone else’s viewpoint or post on a similar topic that doesn’t contain the error. In this way–ideally at least–diligent journalism and well-considered commentary rises to the top and is shared and rewarded, while poor fact-checking and sloppy thinking remains unseen. In the real world, of course, there are far more salient factors that make a post go viral, including how much a person agrees with the view expressed in it.

I don’t mean to suggest that any news story or point of view which is not completely supported by hard evidence and airtight logic shouldn’t be considered or shared; we’re all human and everyone makes mistakes. Imprecision in some minor details is often a necessary part of journalism. For example if a half-dozen tornadoes hit northern Texas but one of them briefly crosses the state line into Oklahoma (doing little or no damage), it’s okay for a journalist to generalize for the sake of clarity and brevity that the events happened in northern Texas. Though perhaps technically not completely true, it’s close enough (and the part that’s not is not significant enough to undermine the larger point of the piece).

People are understandably reluctant to point out errors in their friends and colleague’s work (and more generally in points they agree with), but it’s an important part of critical thinking. Skeptics and scientists reject dogmatism for exactly that reason, and understand that offering constructive criticism is a sign of respect, not personal grievance. The goal is to achieve a better understanding of facts and truth. Pointing out and acknowledging errors (by qualifying or removing them, for example) strengthens arguments.

In future columns I will highlight a series of these “10% Fails,” taken from a variety of sources and contexts. To be clear: I generally agree with the larger points being made by the authors in the pieces I quote, and I highlight their errors with the expectation that the mistakes in them have likely been overlooked. I hope it will encourage readers to think more deeply and critically about the information they see and share-especially when it confirms their point of view. It’s much easier and more intuitive to find errors and identify faulty logic and unsound assumptions in positions we disagree with than in those we approve of. Confirmation bias is one of the most difficult psychological errors to detect, and these examples may help identify and correct it.

Kazoo’s Dieting Girls Statistics

Last year a crowdfunding project helped launch a new magazine, Kazoo, to empower girls and (in part) help steer them toward STEM careers. Kazoo focuses on girls and women, according to its website: “All of our stories are either developed or inspired by top female artists, explorers, scientists, chefs, athletes, activists, writers and others. Regular features include: science experiments; comics; art projects; recipes; interviews with inspiring women from Olympic athletes to astronauts; and fun activities, like secret codes, jokes, mazes, search-and-finds and more…. It will feature some of the most powerful and inspirational women in their fields, thus giving girls a more well-rounded sense of the world and the possibilities within it.”

Founder Erin Bried explains that she and her five-year-old daughter were looking for a magazine they could enjoy together but were dissatisfied with what was available. Bried drew upon nearly twenty years of experience in high profile magazine including Self and Glamour, and launched a Kickstarter campaign “with hopes that other people would also be as interested in a magazine that doesn’t tell girls how to look or act, but instead inspires them to be strong, smart, fierce and, above all, true to themselves. Within 30 days, Kazoo became the most successful journalism campaign in crowdfunding history.”

I went to the Kickstarter campaign’s page to learn more about it. I read the copy and watched the accompanying video. I cheered at the comments about encouraging girls to be independent, inquisitive, and go into STEM fields. I smiled at the ideas for the magazine, remembering my own boyhood reading materials. All was fine until I was confused by a statistic that appeared in both the video and in the campaign copy; I wasn’t sure if it was a mistake or a typo or if I was misunderstanding something.

The copy read, “by age 11, 30 percent of them have already put themselves on a diet,” and the video narration says “She [the girl featured in the video] doesn’t know what a diet is, much less that by age eleven, 30% of her peers will try one” (at 1:45). It sounds dangerous and ominous–surely most people would be alarmed to think that about a third of young girls have tried a diet. But as a journalist who often covers health issues–and a person whose Masters thesis researched dieting and eating disorders–I sensed that part of the story was missing.

So I did some research. The study linked to in the website explains the context to that statistic in the first paragraph under the Results section: “About 30% of girls reported early dieting (by 11 years old); … 30% of the girls were overweight at 9 years old.”

Thus it’s clear from the study why 30% of the girls around that age were on a diet: they were overweight. Diet and exercise, of course, are the universally recommended ways to combat overweight, obesity, and their associated health problems including juvenile onset diabetes. According to the CDC the percentage of kids on a diet almost exactly reflects the percentage of children and
adolescents who are overweight or obese
 and should be on a health regimen of diet and exercise, as recommended by pediatricians.

So the data shows that most or all of the girls who are dieting are doing so because they are overweight and for health reasons, but the way the statistic was presented in the Kazoo video and information made it seem as if the girls are dieting unnecessarily.

Again, to be clear: I love the idea of Kazoo. I donated money to make it a reality and have told people about it online and in an upcoming issue of Skeptical Inquirer magazine. But their use of misleading statistics is troubling, especially for a magazine dedicated to factual information. It was not merely an incidental factoid mentioned in passing; instead it was important enough to include in the final voice-over narration. Fact-checking is more important than ever in this age of “fake news” and misleading information–and perhaps especially so in a magazine aimed at young women.

The theme of Kazoo’s most recent issue (Winter 2016/2017) is architecture, and features blueprints for making a snow fort and a bridge made of candy; a comic about the Brooklyn Bridge, a city scavenger hunt, ice science experiments, a banana bread recipe, and more. Kazoo, which carries no advertising, is only available in screen-free print form and costs $50 per year for four issues; subscriptions are available!