slowcat95 wrote:stan wrote:
Read the climategate emails to get a real good sense of just how pathetic science has become.
Read Tetlock's work. Or Future Babble written about it by a NY Times science reporter. Read John Ioannidis stuff. Most papers are flawed. The stats are bungled. p values are crap. And no one can predict the future.
In my experience, science isn't pathetic. It's difficult, and the best research rarely gets seen publicly. But that doesn't mean science is pathetic. Indeed, the vast majority of science is good, but the vast majority of science isn't what the public The public sees the flamboyant pieces, and those are usually crap. As scientists, we need to do a better job of disseminating our work and making sure its relevant. But we can't help it when bad eggs shamelessly promote bad work (and when the media jumps on it).
You are generalizing from a crappy subset.
If no one ever replicates, how do we know? That's the point. There is zero quality control. None. And no, peer review is not quality control.
There's also no accountability. Quality always suffers in the absence of accountability.
For the most relevant example regarding quality control, let's examine the most famous science study of our generation -- Mann's hockey stick. Not interested in getting to a war about global warming, let's just stay focused on the issue of quality control in science.
We now know that no one ever checked his work. He botched his data. Used data that he knew was inappropriate, used data upside down. We know that rather than use a standard, commercially available stats package to do his PCA, he wrote his own fortran program and gooned it. Badly. And we know that he completely screwed up his verification stats as well.
And yet, the scientific establishment the world over endorsed his study, represented it as the best science on the subject, and relied on it to make demands on governments. Lots and lots of scientists put their own reputations on the line and committed negligent, even reckless, misrepresentation.
This is not how quality is done. This is just reckless and stupid.
But it does give us a good idea how the govt food science scandal could unfold over decades. And how the GRIDs fraud went down. And the govt 2d hand smoke fraud. And how scientists for the govt got caught planting evidence in the forest.
For an example of just how egregiously bad science can be, get a gander of a study that the govt has relied on in various federal court cases - Charles Monnett's polar bear study. One day, out of 15 years of flying the coast, he saw four dead polar bears in the water below. He guessed they drowned. Doesn't know for sure cause he didn't examine them. Might have been shot.
Who knows. [Drowned polar bears are pretty rare. They've been monitored swimming over 400 miles of open ocean. They don't drown even when they swim out to a piece of ice to play and get their picture taken.] There was a storm a few days earlier. And he'd read a science article speculating that global warming would make storms more intense in the future. So -- the fact that one time in 15 years some polar bears were observed from a plane and seemed to have died and a storm happened a few days earlier proves scientifically that global warming is wiping out polar bears. The federal govt has used his study to prove that in court on repeated occasions.
If scientists actually cared about quality, they would be an uproar over stupid crap like this. Instead, we got 'em lined up at the trough to feed.
