Slashdot Mirror
Why Published Research Findings Are Often False
Hugh Pickens writes "Jonah Lehrer has an interesting article in the New Yorker reporting that all sorts of well-established, multiply confirmed findings in science have started to look increasingly uncertain as they cannot be replicated. This phenomenon doesn't yet have an official name, but it's occurring across a wide range of fields, from psychology to ecology and in the field of medicine, the phenomenon seems extremely widespread, affecting not only anti-psychotics but also therapies ranging from cardiac stents to Vitamin E and antidepressants. 'One of my mentors told me that my real mistake was trying to replicate my work,' says researcher Jonathon Schooler. 'He told me doing that was just setting myself up for disappointment.' For many scientists, the effect is especially troubling because of what it exposes about the scientific process. 'If replication is what separates the rigor of science from the squishiness of pseudoscience, where do we put all these rigorously validated findings that can no longer be proved?' writes Lehrer. 'Which results should we believe?' Francis Bacon, the early-modern philosopher and pioneer of the scientific method, once declared that experiments were essential, because they allowed us to 'put nature to the question' but it now appears that nature often gives us different answers. According to John Ioannidis, author of Why Most Published Research Findings Are False, the main problem is that too many researchers engage in what he calls 'significance chasing,' or finding ways to interpret the data so that it passes the statistical test of significance—the ninety-five-per-cent boundary invented by Ronald Fisher. 'The scientists are so eager to pass this magical test that they start playing around with the numbers, trying to find anything that seems worthy,'"
I'm a scientist myself. It's quite clear from where I'm standing that to get good jobs, research grants, etc one needs plenty of published articles. Whether the conclusions of those are true or false is not something that hiring committees will delve into too much. If you are young and have a family to support, it can be tempting to take shortcuts.
Is it possible that there has always been error, but it is just more noticeable now given that reporting is more accurate?
Precisely. As mentioned in a Scientific American blog:
"The difficulties Lehrer describes do not signal a failing of the scientific method, but a triumph: our knowledge is so good that new discoveries are increasingly hard to make, indicating that scientists really are converging on some objective truth."
It's only lying if you do it intentionally. If ten labs independently and without knowing of each other perform essentially the same experiment, and one of them has a statistically significant result, is that lying? The other nine won't get published because, unfortunately, people only rarely (and for large or controversial experiments) publish negative results, but the one anomalous study will.
The vast majority of science is performed with all the good will in the world, but it's simply impossible for scientists to not be human. That's why we do replicate experiments - hell, my wife just published a paper where she tried to replicate someone else's results and got entirely different ones, and analyzed why the first guy got it wrong.
Did you even read the article?
This is basically about poorly designed clinical drug trials without sufficient controls. Sloppy work, even if it seemed rigorous enough at the time.
The sensationalistic "scientific method in question" stuff is pure BS, but after all this is New Yorker magazine we're talking about, so one wouldn't expect too much scientific literacy. It was the scientific method of "predict and test" that caught these erroneous results, so the method itself is fine. The "scientist" who designed a sloppy experiment is too blame, not the method.
However, I'm not sure that psychiatric drug trials even deserve to be called science in the first place. The principle of GIGO (Garbage In - Garbage Out) applies. This is touchy-feely soft science at best. How do you feel today on a scale of 1-10? Do the green pills make you happy?
nahh, the problem is a misunderstanding of statistics (thinking that post-hoc analysis with this fishing for statistical significance) is as valid as proper hypothesis testing. The proper way is where the hypothesis is fully pre-formed and then tested. The numbers and statistics apply ONLY TO THE HYPOTHESIS being tested, so you cannot hunt for a statistical significance just somewhere in the data and then re-formulate your hypothesis.
The need to publish (a scientist's income relies on what he publishes in most cases) as well as funding issues force scientists to try to find some usable results from their science, and by trawling through their data they can often salvage what would otherwise have been a failed bit of research. Except this salvaging operation may actually be absolutely worthless. This is most often not done on purpose but rather due to only partly understanding what statistics and significance testing tell us.
So, a capitalistic, fully performance based (with results being the performance metric) environment does not seem to work well for science.
Surprised?
Me neither.
The problem is actually the opposite. Note the E.S.P. experiment cited in the article. Rhine's initial experiment suggested to him that E.S.P. was real. Before publishing his results he did the right thing and reran the tests and the results proving E.S.P. were not repeatable.
The next part is his absolute failure to understand the scientific method and statistics. He concluded that "extra-sensory perception ability has gone through a marked decline.” In fact what he experienced was Regression toward the mean.
Taking a well understood principle, renaming it with a term that suggests an action is taking place, then arguing that you have found some new phenomenon that proves science doesn't work is not critical information about anything.
It is ignorance that will be dismissed for obvious reasons. Too much time and energy is wasted repeatedly addressing these attacks on science by people who want so badly for their pseudo-science or supernatural beliefs to be true. In a perfect world when somebody stumbles upon regression to the mean without knowing it they would do additional research to understand what it is they are observing rather than conclude that their initial experiment was correct and the supernatural ability they detected was "declining" rather than accept the alternate, it was never there in the first place.
Are you serious? Many thousands of people are dead simply because a few people were trying to stay gainfully employed to support their families?
I am truly sorry if this comes off as offensive as I think it does but if you believe there would be mass suffering from unemployment if we did not bomb the shit out of Iraq and that was the basis for the lies that resulted in many thousands losing their lives then you are seriously deluded.
As a U.S. citizen I found Clinton's actions and lies embarrassing, but the lies from Bush transferred billions, if not trillions, of public funds into the hands of a few and resulted in the deaths of many thousands of people.
Comparing lies about a blow job to lies resulting in debt and death is absurdity on a grand scale.
nahh, the problem is a misunderstanding of statistics (thinking that post-hoc analysis with this fishing for statistical significance) is as valid as proper hypothesis testing. The proper way is where the hypothesis is fully pre-formed and then tested. The numbers and statistics apply ONLY TO THE HYPOTHESIS being tested, so you cannot hunt for a statistical significance just somewhere in the data and then re-formulate your hypothesis.
This significance of this fundamental mistake cannot be overstated. It seems to be prevalent in medical literature and there was a doctor doing the rounds lecturing about this a couple of years back. I wish I could recall exactly which podcast but he covered all sorts of common fundamental errors in medical research statistics and did it in a very accessible way. The key thing to remember is that if you have enough variables there WILL by complete coincidence be correlation between some of them in any given sample. So to test a hypothesis properly, not only must you formulate it in advance without looking for any correlation within the data, but you must look at more than one data set to verify your findings.
These posts express my own personal views, not those of my employer
No, that doesn't solve the problem, it increases it.
The consistent lack of results is a result, and a very useful one too.
The logical next step is to ban marketing of humbug until and unless the snake oil sellers can show valid scientific theories and peer reviews for their remedies.
Likewise, capitalist-funded research needs to stop rewarding findings, but start treating all results as equally valid science, and stop punishing scientists who produce negative and inconclusive results. That's good science, which is what they should pay for.
Consistently publishing more results than randomness would dictate is a clear indication of bad science, and should be punished, not rewarded.
I'm a biochemist. After earning my PhD five years ago I've been working in academia, but my funding's about to run out and I'm applying for jobs at biotech and pharmaceutical companies. Do you think I had the empathy and morality centers of my brain removed or something? Do you think that every single person working in those sectors underwent the same procedure or were blessed from birth with complete amorality? The reality is that science is hard. The reality is that science is expensive. The reality is that our knowledge is incomplete and we do the best we can with the limited resources at our disposal. If we're lucky, that means we can turn a life-destroying illness into something treatable. Take cancer, for example. There's no magic pill to take it away and probably never will be, but it's because cancer is a large family of disease caused by different breakdowns of cellular mechanisms, many mechanisms that we don't understand very well and that are very hard to tease apart. That's why cancer, and diseases in general, tend to end up with treatments and not one-pill cures, not because big pharma's hiding it.
My brother went through 10 months of chemotherapy. 10 months of being nauseous, 10 months of not wanting to eat, 10 months without a sense of smell, 10 months with no sense of taste, 10 months of physical weakness, 10 months of diminished mental capacity, 10 months of needles, 10 months of IVs full of chemicals that burned when they went in, 10 months of doctors prodding and poking. He's now cancer-free and has been for 12 years. Back when he went through that his odds of surviving Hodgkin's lymphoma were about 80%. Current treatment has reached 90%, and a recent experimental treatment is at 98%. They're all still unpleasant and take months. Do you honestly think that if I had the ability to jump in with a magic pill and spare my brother those 10 months I wouldn't do it because it might hurt the corporate bottom line? Fuck the bottom line. Fuck having a job if it came to it. That's the prevailing attitude in biotech and pharmaceutical companies because they're made up of people like me, people who have seen loved ones go through horrible illness, and not the monsters your fantasy requires.