Science’s Biggest Fail • Hacker News Board

Re: Science’s Biggest Fail

#181

Post by pjkundert » Mon, Feb 02, 2015, 9:06 PM UTC

api wrote: post #17

mcphage wrote: post #16

Doing science is a lot like making sausages. There's a lot of conjecture, hypothesizing, testing, making observations, generating theories, testing those theories, throwing out the ones that don't hold up to repeatable studies. I don't think nutrition is unique in that matter. Where it differs from other sciences are: 1. The general public is very interested in the results, and so there's a lot of motivation for peop…

Insufficient repetition is a huge problem in medical science. There is no incentive to replicate existing studies, as it doesn't advance your career. A related problem is that there is no incentive to publish negative results.

The incentive to publish negative results is: "to not be labelled a charlatan".

The widespread non-publication of negative studies (and, tragically, their datasets) has, rightly, tarnished the validity of all published scientific results.

You can't blame normal people for doubting "scientific" results. It doesn't take a rocket scientist to deduce the worthlessness of published results, when it is disclosed that negative results have been withheld:

Just do studies until you get the results you want, then publish that one.

This has been de rigeur in medical science for some time.

For shame...

Re: Science’s Biggest Fail

#182

Post by shanusmagnus » Mon, Feb 02, 2015, 9:06 PM UTC

dragontamer wrote: post #165

shanusmagnus wrote: post #158

I think Adams is very wrong here. I'm sympathetic to his reasoning, since I have followed nutrition closely for the last twenty years, and believe the scientific/media consensus to be much as he described, and even worse than that occasionally, where people who should know better (a PhD teaching a class on nutrition in a community college) claim things that are both factually and obviously wrong ("Low carb diets are…

I work with these "creationist idiots" you talk about. And let me tell you something... A creationist engineer codes just as well as a Darwinist engineer. Young Earth creationist nonetheless! He's one of the best students in my class that I've ever seen, extremely good at coding, great at running meetings and organizing people. The fact of the matter is, global warming, pollution, and even creationism / whatever does…

I wouldn't deny that there are lots of smart people who disagree with me on certain foundational issues. I absolutely deny that the majority of those who do are the 'intelligent supporters' you describe. And if smart people are acting in an idiotic manner, the fact that they have mental horsepower to do better is actually worse.

Re: Science’s Biggest Fail

#183

Post by mcfunley » Mon, Feb 02, 2015, 9:09 PM UTC

muaddirac wrote: post #126

ZeroGravitas wrote: post #76

I found out recently that Scott Adams doesn't believe in evolution, which rather colours my response to this headline. http://rationalwiki.org/wiki/Scott_Adams

He has some odd ideas about science. The end of his book "The Dilbert Future" calls for "new ways of looking at existing things." And then claims that all of science is just "looking at things" which, he points out, is pretty unreliable, and that we need to improve our ability to perceive the universe in order to move forward in science. Of note is that he doesn't consider technologies which enhance our perception to…

I'm glad someone else has noticed the profound weirdness of "The Dilbert Future." I stopped taking Scott Adams seriously a few minutes after reading it. I haven't looked at it since high school (in the 90's -- eek), but IIRC he also believes that his thoughts control the direction of the universe if he writes down affirmations in a notebook daily.

Re: Science’s Biggest Fail

#184

Post by suchow » Mon, Feb 02, 2015, 9:10 PM UTC

aetherson wrote: post #108

mikeash wrote: post #86

Earlier quoted context omitted.

If I understand you correctly, you're saying that it's not the media's fault, because the media just blindly repeats whatever they're told. This is exactly why it's the media's fault! If a journalist doesn't have the skills to understand what they're reporting on, then they shouldn't be reporting on it. We wouldn't put up with a report on Afghanistan that says the place is full of Arabs because the reporter was told…

I'm sorry, but you're holding the media to an impossible standard. This isn't about "checking the CIA World Fact Book," it's "Read dozens of dense studies that assume PhD-level study in the relevant fields." The media does light fact-checking, and it's exactly that fact-checking that allows institutions to set an agenda. If a reporter sees a press release about a new study in, say, the field of cancer, and they try t…

> The media does light fact-checking, and it's exactly that fact-checking that allows institutions to set an agenda.

This is true only of some media. For example, The New Yorker has a fact-checking department that holds their journalists to high standards. Reading their long-form articles, it is clear that the writers have taken the time to understand things deeply. Nobody writes a 10,000-word essay on science the night before on a deadline.

It's quite reasonable to expect a journalist reporting on science to read dozens of articles (or review papers) about a field and to interview many scientists. It's not hard to find people with that level of understanding in a field who would be happy to write public-facing magazine articles. Many of the best science writers do.

Re: Science’s Biggest Fail

#185

Post by jeffdavis » Mon, Feb 02, 2015, 9:15 PM UTC

"Scientific consensus" is an oxymoron, because consensus is not a scientific process.

Unless by "consensus" you mean that a lot of scientists have successfully reproduced an experiment. But nobody ever means that.

That being said, scientists are often informed and educated, so their opinions should be valued. But please take off the lab coat and call it an opinion.

Re: Science’s Biggest Fail

#186

Post by npsimons » Mon, Feb 02, 2015, 9:16 PM UTC

muaddirac wrote: post #126

ZeroGravitas wrote: post #76

I found out recently that Scott Adams doesn't believe in evolution, which rather colours my response to this headline. http://rationalwiki.org/wiki/Scott_Adams

He has some odd ideas about science. The end of his book "The Dilbert Future" calls for "new ways of looking at existing things." And then claims that all of science is just "looking at things" which, he points out, is pretty unreliable, and that we need to improve our ability to perceive the universe in order to move forward in science. Of note is that he doesn't consider technologies which enhance our perception to…

And then claims that all of science is just "looking at things"

Well, at least he got one thing correct. Mostly. Science is a system for generating a model of the world. It just also happens to be the best one we have, insofar as it has generated the most accurate model of the universe to date. Is it perfect? No. Should it be criticized? Yes, otherwise it won't improve. Should we throw it out altogether? Perhaps that question should be put to the literally billions of people whose lives have been saved and improved by the results of science. Or not. Those same billions are probably very badly informed about science, which is what Adams should be railing against, not science itself.

Re: Science’s Biggest Fail

#187

Post by shanusmagnus » Mon, Feb 02, 2015, 9:18 PM UTC

msandford wrote: post #162

shanusmagnus wrote: post #158

I think Adams is very wrong here. I'm sympathetic to his reasoning, since I have followed nutrition closely for the last twenty years, and believe the scientific/media consensus to be much as he described, and even worse than that occasionally, where people who should know better (a PhD teaching a class on nutrition in a community college) claim things that are both factually and obviously wrong ("Low carb diets are…

>It is not an accident that the vocal opponents of things like climate change, global warming, animals welfare, pollution, etc., are the same folks who are 100% convinced that they sometimes get personal messages from angels. Their idiotic worldviews do not arise as a result of Bayesian discounting based on having received bad advice, but from intellectual laziness and an inability (or unwillingness) to look in the m…

First, I'd never say that not agreeing with me makes anyone an idiot. Not agreeing with thousands of people from around the world whose entire professional lives revolve around investigating these very issues, though, requires a higher standard of evidence than your steely gaze and good old horse sense.

But speaking of standards of evidence:

>>>> For example, there was a significant hiatus from warming for the last 10 years or so when there "shouldn't have been" (whatever that means) and it's only recently that some scientists have proven why The fact that you can't write down a model that predicts an extraordinarily complicated system down to the last detail is a very poor standard of evidence to require. It's the same flavor of argument that Big Tobacco used for all those years to squirm through legal loopholes about whether or not they knew that smoking killed people. I don't follow this literature but I'm not surprised that nobody can explain the handful of 100 year old ladies who smoked their whole lives and are as healthy as horses. The 'can you explain _this_ then???' defense isn't much of a defense when viewed in context.

Re: Science’s Biggest Fail

#188

Post by SilasX » Mon, Feb 02, 2015, 9:27 PM UTC

enraged_camel wrote: post #143

SilasX wrote: post #119

Earlier quoted context omitted.

Adams wasn't criticizing science's updating of beliefs on new evidence. He was criticizing the practice of shouting "Hey! We're ultra-certain about this! Very important! You've got to do this!" when the truth was more like, "We kind of have some preliminary evidence this might be bad for you, stay tuned." (And I don't buy the claim that it's just the media overblowing this; why weren't the national academies of scien…

>>He was criticizing the practice of shouting "Hey! We're ultra-certain about this! Very important! You've got to do this!" when the truth was more like, "We kind of have some preliminary evidence this might be bad for you, stay tuned." This is almost always a result of the "Science News Cycle." http://www.phdcomics.com/comics.php?n=1174 Aside from that, there's a lot of pressure on scientists to produce actionable r…

If it were just an issue of media hype, that would be a valid explanation for the food pyramid failure.

But it's not just that. It's the fact that a) government experts believed it b) with high confidence that c) isn't justified, and d) are so confident that they plaster it in every classroom with all their authority behind it, and e) no scientific authorities go on record saying the real state of the research in case anyone bothers to get the real story.

You can't blame the media for those parts.

Re: Science’s Biggest Fail

#189

Post by snowwrestler » Mon, Feb 02, 2015, 9:28 PM UTC

People trust science implicitly 1000 ways every day: when they get in their car, when they check Facebook, when they step into an airplane, when they take some Advil, when they eat a Snickers bar, when they take an antibiotic, when they take Viagra, when they make a phone call, etc.

So the question is not "why don't people trust science," the question is "why do people very selectively mistrust small segments of science?"

A plausible answer is because there are people and organizations who are encouraging them to mistrust those small segments of science--by purposefully feeding bad information into the marketplace of ideas.

I think Adams is making a fundamental error of attribution, blaming good actors (real scientists) for the actions of bad actors. He's basically arguing that unless scientists can stop all bad information from anyone, they can be blamed for the bad information. Doesn't seem fair or sustainable.

Re: Science’s Biggest Fail

#190

Post by angelbob » Mon, Feb 02, 2015, 9:28 PM UTC

DanBC wrote: post #109

exelius wrote: post #104

Earlier quoted context omitted.

We actually know shockingly little about a lot of topics relative to our bodies. One big reason for this is that, ethically, we can't just go experiment on live humans. But genetics and human biological systems may only play a part of how nutrition affects us. As an example, science is just now coming around to is that the bacteria in our intestines has a huge impact on how our body reacts to food - and that those ba…

> One big reason for this is that, ethically, we can't just go experiment on live humans The US has an enormous prison population. They get fed for $2.50 per day. Is it possible to recruit prisoners ethically to take part in diet research?

Current ethics boards say "no".

It's a debatable point -- a lot of things that ethics boards forbid might be, in some way, reasonable to study in some specific cases.

But you have to draw the line somewhere, and ethics boards (correctly) draw simple lines in a place where they'll forbid a very high percentage of the bad stuff and a small-ish but not very small percentage of the good stuff.

You're going to get errors when you draw simple lines. Ethics boards are conservative about where the lines go. They should be.