Survival of the Wrongest | Hacker News Reader

Survival of the Wrongest

cjr.org

65 pointsanupj27 comments

I can't speak for the rest of the article, but this guy's take on low carb diets is a bit...odd:

>Unfortunately, it’s an approach [Low carb diets] that leaves the vast majority of frontline obesity experts gritting their teeth, because while the strategy sometimes appears to hold up in studies, in the real world such dieters are rarely able to keep the weight off—to say nothing of the potential health risks of eating too much fat.

So he's saying that studies do support low carb diets having high efficacy rates (for weight loss), but "in the real world" this doesn't pan out. What is this unbiased, statistically significant source of data on "the real world" that he's relying on? He's ruled out studies, and it's obviously not anecdotes because that would be even worse. That basically just leaves intuition. He then goes on to cite the risks of a high protein diet (while linking it as high fat only) by linking to a WebMD article which appears to be written by an unnamed author who has poorly collated the research on the issue and left out major advances of our understanding of issues like cholesterol levels.

Practicing weight maintenance before temporarily switching to a low-calorie, weight-loss diet helps dieters keep weight off. http://med.stanford.edu/ism/2012/october/stability.html

Low-calorie diets are good for losing weight, but physical activity is more effective at maintaining a target weight. https://www.mayoclinic.com/health/weight-loss/AN01619

Very-low-carb diets result in higher metabolism but might also increase stress and risk of heart problems. Low glycemic index diets might be healthier for long-term weight maintenance. http://www.nih.gov/researchmatters/july2012/07162012weight.h...

I think he means that they work under controlled conditions, but it's hard to replicate them in the real world environment. Can't go to most restaurants, can't make a lot of dishes, can't do social gatherings, definitely need to cook yourself everything you eat, and so on.

That applies to any diet though, if the advice is good, but people don't follow it, it's not the fault of the diet. The point is he is singling out low carb diets as if it's the diet itself that is at fault, not just the common problem of people not following through on a diet regimen.

It doesn't apply to any diet. For example, a simple caloric restriction diet might simply require that you leave half of everything you're served on your plate and never have seconds. This would allow you to eat anything in any social context, just not as much. A low carb diet would require you to forego the cake and ice cream entirely at your own child's birthday, etc.

Having a bite of cake is still low carb.

Is it any harder than having a strong food allergy? It's only difficult when you're not used to it. (55 lbs on Paleo, kept it off, zero calorie counting, in fact downright gluttonous sometimes.)

It is far easier to psychologically comply with a concrete immediate severe allergy diet than a vague "analog" weight loss diet.

I understand why he is saying that.

I for example, "re-invented" a atkins or paleo diet of sorts (ie: I on my own drifted to that sort of died before knowing those terms or news on that).

And it is working...

But it is working not because I avoid carbs. But because I avoid non-meaningful calories, I avoid carbs, and most of the fat.

The difference is: When I think I need meat, cheese, etc... to reach a certain protein goal, I don't freak out because there are fat on it.

But I don't go around using mayo, huge cheese-only sandwiches, bacon like if there was no tomorrow...

I eat bacon? Yes... ALso eggs! In fact frequently I eat bacon+egg sandwich. But I do that keeping track of the fat I am eating and not eating too much fat (and keeping carbs very low as I can).

So, the study the guy cites work in theory? Yes, because it is tighly controlled, in real life people understand "you can eat fat" as "fat make you thin".

You know, like those people that buy dieting shakes, and drown themselves in shakes and manage to consume 1000+ calories in a day only in shakes, thinking that drinking the shake that will make them lose weight.

It's all well and noble to say "health journalism is flawed, everything is breathlessly reported as a breakthrough", but when you put Tara Parker-Pope and Gary Taubes in the same category, you're committing more of the same mistakes. There's nothing noble at all in shooting down all of health journalism.

He goes on to say:

"Worse still, health journalists are taking advantage of the wrongness problem. Presented with a range of conflicting findings for almost any interesting question, reporters are free to pick those that back up their preferred thesis."

It appears that this author's preferred thesis is that when presented with conflicting evidence, one should throw one's hands up in despair and do whatever you want ("apply common sense liberally"), instead of some kind of analysis of the evidence to find which side of the conflict is more reliable.

The previous submission of the canonical URL

http://news.ycombinator.com/item?id=5019442

received no discussion, but I think that's too bad, and I would love to see discussion of this interesting article begin here.

The problem with personal health news is that is is never news. A single study that contradicts conventional wisdom or purports a new finding is not meant to be digested by the general public.

Ideally such a study would be published in a medical journal, and other researchers would then perform studies to corroborate or disprove the original; After a sufficient amount of evidence is gather a systematic review would be published, and the results of that would be distributed to doctors and health advice reporters, who would present the information to the general public in a clear, unified and easy-to-understand manner.

In a perfect world.

The reporting of every study only leads the general public to believe that nothing is certain (because dissenting views are published) and not take any of the advice seriously. It is almost as bad as politics.

Ideally everyone would be rational, and honest, and have aligned interests, and trust eachother enough to all outsource their critical thinking to experts. Except... that sounds quite boring and I rather like being a somewhat-independent individual instead of a single-purpose cell in a larger collective organism/society.

The reporting of every study only leads the general public to believe that nothing is certain (because dissenting views are published) and not take any of the advice seriously.

Since things are in fact not certain, I would think this is a good thing?

I think this is an important article for the HN crowd. Educated STEM people usually have a strong devotion to science, but science very often fails at revealing the truth (dangerously, in a misleading manner). The article points out the problem isn't malice or incompetence, it's bad expectations.

To me, a critical analysis of research is often more valuable than the research itself.

> To me, a critical analysis of research is often more valuable than the research itself.

Critical analysis of research is science. I'm perpetually frustrated with articles like this that seem to portray science as a failure by citing faulty studies, ignoring the fact that everything is working as it should.

> Critical analysis of research is science

You missed the point in the name of semantics, but I'll bite. Finding flaws in a scientific paper my be taking part in the scientific process, but it is by no means a science in itself - e.g. as a stats person, I can point out when a p value is misapplied, but I would never claim I just 'used science'.

>I'm perpetually frustrated with articles like this that seem to portray science as a failure by citing faulty studies

It's worth separating the philosophy of science from its application. This article is purely about the application and what has gone wrong. This is why I think this article is important for the STEM crowd: we are taught that science is an infallible philosophy (which I agree with), but it's easy to forget the garbage-in garbage-out nature of experimentation.

This is the Columbia Journalism Review. It's not saying science is stupid, it's saying that reporters are too quick to publish study results, when not enough research has been done on the subject topic.

JohnsonB's comment[0] draws the same conclusion I did. The article seems to suggest that studies aren't working and we need to start listening to anecdotal evidence. I could have misinterpreted that passage, though.

[0] http://news.ycombinator.com/item?id=5080094

The idea that scientific research is infallible is exactly what the article is arguing against - in fact, it claims this attitude is a bigger problem than the scientists or the media.

Human-performed studies with limited budgets and unrealistic controls should not supersede expert opinion per se. When there is a disconnect between the two it is irresponsible not to question the scientific studies (in addition to the experts, of course).

It bothers me to no end when people discuss "science" as if it were just one of many options available. Science is the pursuit of truth and the pursuit of truth is science. There are no alternatives.

Science is a means of pursuing the truth. People who read tea leaves also believe they're pursuing the truth, but nobody would say that's science.

Is logic science?

> Educated STEM people usually have a strong devotion to science, but science very often fails at revealing the truth

The article criticizes journalism for not properly reflecting science and scientific opinion, so it implicitly accepts science as the gold standard. I think the lesson is that there's more demand for scientific novelty and certainty than science itself can satisfy, so science journalism focuses on the conjecture, controversy, preliminary results, and premature conclusions that arise in the practice of science instead of focusing on the emergence of consensus among scientists. The job of a science journalist like Tara Parker-Pope or a pop science author/advocate like Gary Taubes is akin to the job of a prosecuting attorney: to tell a single story as compellingly as possible. The article is about cases where journalists succeed in telling a simple, intuitive, compelling story while science itself is telling a collection of messy, incomplete, counter-intuitive, and/or contradictory stories.

I know many "educated STEM people," but I don't know anybody who does any significant reading of scientific literature outside their field. I've read a lot of abstracts (and a few articles, when they're available) while trying to look one layer deeper than the news, but I don't have the background to understand a lot of what I read, and I've never gone to the library and reviewed dozens of related papers to get context for the paper I'm reading, like a real scientist would. Virtually every vote and consumer choice influenced by science is influenced through news media, science journalism, or classroom education, not by direct exposure to the work of practicing scientists.

I think you missed the second half of the article.

The article is about "How personal-health journalism ignores the fundamental pitfalls baked into all scientific research." The second half of the article is about biases that influence individual scientists and journals. Scientists acknowledge such factors and do their best, within the limits of their own bias, to take all limiting factors into account when evaluating research. Where the influence of the scientific community coincides, you can expect to see bias shared by all scientists, and where powerful institutions can unduly sway community opinion, bias will bubble up into the community consensus, but scientific consensus is still much more reliable than the results of individual papers or the opinions of individual scientists, because the interests of individual scientists vary and often compete. The article makes the point that journalism should take skepticism within the scientific community about specific claims more seriously and could serve as an outside critical perspective on science, but in fact reads the scientific literature more naively than scientists themselves do.

Ben Goldacre's recent book "Bad Pharma" is all about how easy it is for pharmaceutical research to go astray (and/or be led astray). I would recommend it for anyone interested in the details of this thorny subject.

If a p-value of .05 means a study's results are significant, then 20% of studies will arrive at an incorrect conclusion through chance! If counterintuitive conclusions make interesting articles, but negative results are somewhat boring even to professional scientists, is it any real surprise that the end result is a steady stream of dubious advice?

This argument was shamelessly paraphrased from XKCD [1].

[1] http://xkcd.com/882/

Reply on news.ycombinator.com