randombio.com | commentary
Monday, September 05, 2016
Science Under Siege, Part IIIUnderstanding what causes bad science is critical to reforming it.
t's generally understood that PC is a form of emotional thinking (see here for a description). But as pervasive as it is, emotional thinking is only one way of perverting the truth. There is also ideological correctness, commercial correctness, and self-aggrandizing correctness. All are ways of misleading people to convince them to accept your conclusion.
As I type this, I am on vacation, and as always happens I am sick as a dog. Last time I needed an emergency root canal; this time the problem is a gastrointestinal bug that I somehow picked up.* Since I am unable to read my book, all I can do is watch TV. At the moment there's a commercial telling us that talcum powder is linked to ovarian cancer. It's a perfect example of commercial correctness.
The idea that talc causes cancer is so preposterous it's not even worth rebutting: there is no asbestos in talc, and it's hard to imagine in any case how talc could get into someone's ovaries. Even if talc did what these ambulance chasers suggest, the FDA already approved it as safe and effective; that such approval does not indemnify the manufacturers throws the FDA's stated gatekeeper role into question. With power comes responsibility, and the fact that the lawyers don't sue the feds shows that the FDA has none.
Maybe we should be grateful that the lawyers on both sides turn to science, regarding it as an arbiter of truth. But as much as some people would like to deny it, science, like my gastrointestinal tract, is having big problems these days, and in much the same way.
Many of the attacks on science are overtly political. Many conservatives despise academics. The global warming scandal, the intolerance, and the anti-Western shenanigans on campus have turned conservatives against the university. Many argue for abolition of tenure.
Liberals feel same thing about industry: to hear them tell it, everyone in industry is corrupt and all their discoveries are fabricated for a profit motive. Industry deeply resents that, so we get people like C.G. Begley and John P.A. Ioannidis attacking academic science.
I discussed here how Begley's claims are unsubstantiated and impossible to verify: an impossible result presented with zero evidence. Ioannidis's claims are plain ordinary nonsense, based on bad statistics. When people claim to believe these guys, they are indulging in ideologically motivated thinking: they repeat the claims, not because they really believe them, but because they're angry and want to produce a specific effect.
Some people, of course, see a genuine problem and are earnestly trying to improve science. But the solutions they typically propose—forcing scientists to keep more detailed records of every experiment, more paperwork, restricting science to practical topics, and so on—will only make the problem worse.
That's not to say there isn't any bad science out there. It's been my misfortune to see a lot of it. One reason is that professors rarely do research these days. They're bogged down in meetings, grant writing, and paperwork. It takes them three months to write a grant, and in many fields only one in ten grants is ever funded. So every funded grant represents, on average, a year and a half of tedious writing and rewriting.
As a result the students, who do the actual research, are often unsupervised. Postdocs and grad students are the ones who do the vast majority of the research in this country. As happens in IT and in engineering, the more senior the person, the more they are punished by having to do managerial tasks instead of meaningful work.
As a result, many students are studying things they barely understand. I know one student who dug up a method in the research literature for a process called LCMS to measure one particular molecule. I've been doing stuff like this for thirty years, and I warned her that this was not the best way: there are 7 or 8 different isomers—forms of the molecule with identical chromatographic properties—that cannot be distinguished by this method.
But I was not her boss, and her professor pretty much ignored the project. So the student is writing a paper whose sole purpose is to get her career started, but contributes nothing to science. Whether the findings are true or not is down to random chance; without good evidence they are not credible, and science, like a vast river, will simply move on, diluting them out.
Fixing this would require changing the entire work model of academic science.
A second reason for bad science is this: instead of being aloof, the professor may dictate what topics everyone in the lab works on, and dictate in advance what the conclusions must be if they wish to stay employed. If the professor's theory happens to be correct, nobody cares; if it's not, none of the staff ever gets a grant and their careers are destroyed. In either case, bad science is the result.
I worked for years in such a lab, and now am sadly watching my H-1B friends panicking at the prospect of going back to their home country. One administrator accused me of having “white privilege” for not having to suffer their fate. I was too polite to say how racist and insulting I found that comment to be.
But hell, I wanted to. Yeah, I was thinking, the goobs invent a law bringing in so many foreign workers it undermines an entire industry, and that makes me a racist for being a white person and not being one of them. But the more I thought about it, I realized that it reveals a third, more fundamental reason for the problems in science. If you want to save science, get rid of the idiot bean counters and the bureaucrats.
* The County sprayed herbicide along the road, leaving a wide swath of dangerously dead trees. They saved money by not calling Asplundh as normal states do, and left me to bear the cost of cutting them down and clearing them by hand. Somehow, I'm not sure how, this gave me a fever, chills, dehydration, and massive GI problems.
Last edited sep 06 2016, 6:33am
Science under siege, Part II
People say there are no jokes in scientific papers. But I found one.
Science under siege, Part I
Bogus claims about the reproducibility of scientific research will not die on their own. We must give them a push.
Corrosive claims about academic research
Industry scientists often claim that academic research is not reproducible. But they never show their data.