Amgen, the worlds largest independent biotech firm, ran a replication survey on fifty three widely cited landmark cancer research papers. Only six were replicable.
If conservatives are losing faith in official science, maybe that is an indicator that conservatives are in contact with reality.
Sweet, statistics. I was just talking about independent lines of evidence, too…
“He charges that as much as 90 percent of the published medical information that doctors rely on is flawed.”
Amgen shouldn’t have been astonished. Ioannidis already told them it would happen.
“His model predicted, in different fields of medical research, rates of wrongness roughly corresponding to the observed rates at which findings were later convincingly refuted: 80 percent of non-randomized studies (by far the most common type) turn out to be wrong, as do 25 percent of supposedly gold-standard randomized trials, and as much as 10 percent of the platinum-standard large randomized trials.”
Oh good, your p-values are below 0.05. So it wasn’t due to chance. Unfortunately you had so many other sources of error that they completely swamped statistical error.
–
Gee guys, where did I get the idea to trust my own observations over that of scientific studies? For all nutritional ‘science’: does it predict what happens in my kitchen? (No.) For all fitness science, does it predict what happens during workouts? (The record isn’t good.) And so on.
Physicists seem to be more reliable not because they design better experiments, but simply because it is much harder to neglect your own physical observations.
Physicists deal with stuff where it is easier to do reproducible experiments.
If there is little risk of your experiment being replicated, people will not try very hard to make their experiments reproducible.
Richard Feynman discussed this in the context of experiments with rats running mazes. It turns out it is quite difficult to reproduce a maze so that it is replicated from the point of view of a rat. And it also turns out that in the entire field, only one guy tried.
If you are not trying very hard to make your experiments reproducible, you are not doing science, and if the field tolerates irreproducible results the field is not science.
The nutrition and exercise stuff is in fact amenable to reproducible experiment. It’s just that rather than spend the research money on a few very expensive metabolic ward studies and such-like, it gets blown on a whole lot of utterly worthless survey based studies.
Physics has been in a rut since the 70s. String theory become the big field from almost 30 years without producing one damn thing. There’s a book on the subject called “The Trouble With Physics: The Rise of String Theory, the Fall of a Science, and What Comes Next ” by Lee Smolin that covers it.
When you pay even Physicists big bucks for nothing more than untestable theory with fun theory names, then your can wreck even the hardest of sciences.
It’s not clear to me how they determined that 11% of results in large randomized controlled trials are “wrong.” Does that just mean that these results disagreed with subsequent large randomized controlled trials?
Smolin’s book was a really interesting treatment of string theory and the sociological problems around it. Does anyone know of a similar treatment of dark matter/energy? This area smells like epicycles to me (our theories don’t work based on what we can observe; therefore, there must be a bunch of unobservable stuff distributed just so which make our theories work again), but I’m not willing to put in the time to genuinely understand it.
“Does that just mean that these results disagreed with subsequent large randomized controlled trials?”
Yes. They used the law of non-contradiction.
“but I’m not willing to put in the time to genuinely understand it.”
I was willing to put in moderate time, and it looked the same way to me until I saw this picture. As measured by gravitational lensing, the normal matter stopped and the dark matter kept going. It’s actually a separate substance, that interacts less via electromagnetism and suchlike.
On the other hand, the description of dark matter is still, “Some stuff we found that has mass.” It’s not exactly in the Standard Model…
That’s pretty interesting, thanks.
The Food Pyramid and Ancel Keys’s saturated fat heart disease theory is one of the better illustrations. Health conscious people of all kinds of political persuasions are now keenly aware the government and “cathedral” at large spent decades giving dietary advice that amounted to lies. They might have produced the obesity epidemic in the process.
It might be interesting to ask the smarter “warmists” if they eat their three servings a day of whole grains.
Consensus always winds up dominated by the evil, the insane, and the evil and insane. In the case of the food pyramid, the animal rights activists.
The animal rights crowd do not like animals any more than Marxists like proletarians. They do not have pets, they wanted to kill Knut, the cute cuddly polar bear cub, because he was raised by humans. Marx’s personal behavior suggested that he hated proletarians with an almost insane hatred, and analogously with animal rights. They are crazy. They love animals, but dislike each particular specific individual animal.
In consensus, the sane shift, the insane don’t shift, and the evil lie about what they believe and claim to believe whatever is a convenient path to maximizing power, money, and murder. Because the sane shift, they are in the long run irrelevant to the outcome of the consensus.
And note that Marx’s philosophy was extremely good at wrecking the lives of proles. This is not a coincidence.
Proggie philosophy is extremely good at wrecking the lives of the not-powerful. Guess what I think proggies truly hate.
The food pyramid actually works fine. Americans eat more in every category, not just grains, hence they are fatasses.
People simply eat too much and move too little and there’s nothing the government can do about it. You can eat 100% protein or 100% animal fat or both and still put on fat. Ask anyone who’s done the Anabolic Diet or any Atkins “Dieter.”
In general, though, I agree with the sentiment in this thread.
Shooting from the hip…
Liberals are unwittingly promoting things that are not Science. Probably because of educational culture and ideology. Thus where Liberals tread political ground in Science is exactly where Conservatives get skeptical and speak out. This now happens to be all the time.
Conservatives are losing faith in the Sciences because it may be convenient to do so – or, they see Liberals as complete economic dunces and their inclination to argue from authority as nigh on laughable. And it is.
I agree, there is tons of dubious stuff in the literature, some minority of papers in biomed are true. Science sort of triumphs at the last, though, because you can’t build on people’s work if you can’t reproduce it. It may not be the most efficient.
I think everyone knows though that the traditional p < 0.05 standard of biomed is ridiculous. Surely it should be cranked down to p < 0.02 to cut down just slightly on the noise and just f**ing clear out people's minds, or 0.015 even, but what can the individual do about it — insurmountable costs to doing it unilaterally.