> > On Feb 21, 2018 6:24 PM, "Biju Chacko" <biju.cha...@gmail.com> wrote: > > > > On Wed, Feb 21, 2018 at 2:31 PM, Srini RamaKrishnan <che...@gmail.com> > wrote: > > > > > > Paying respect to science is good form, but doesn't always mean it's an > > > indication of quality. Neither is questioning science inherently a bad > > > idea. > > > > Erm, there's a hell of a difference between questioning specific > > studies or hypotheses and dismissing established scientific consensus. > > > Science is largely determined by the people doing the science and > their human failings. There's no protection against human nature. > > A good rabbit hole is the Google search term, "half of all science is > wrong", which is a paraphrase of the words of Richard Horton > (www.thelancet.com/journals/lancet/article/PIIS0140-6736(15)60696-1/fulltext), > editor of The Lancet, who admitted rather timidly or bravely > (debateable) in his editorial that "much of the scientific literature, > perhaps half, may simply be untrue." Also in his words, "science has > taken a turn toward darkness."
<snip> > Most are silent, and the few defenders seem to come out with silly > excuses that would make politicians blush. > > Here's one - > > https://www.newscientist.com/article/dn7915-most-scientific-papers-are-probably- > wrong/ > > << But Solomon Snyder, senior editor at the Proceedings of the > National Academy of Sciences, and a neuroscientist at Johns Hopkins > Medical School in Baltimore, US, says most working scientists > understand the limitations of published research. I think you misunderstand the scientific process. Scientists may do research and publish a paper detailing their hypothesis, method, results, study numbers etc. and draw their conclusions. But it is not 'established scientific consensus' until it is reproduced by many different groups, many times with outliers in the data accounted for. One scientific paper does not science make. And moreover, the person you quote above makes it clear that most scientists understand that that is how science works. It is the mainstream media and press that latch on to a single research paper and claim it is 'science' that do not understand how science works. A single paper may draw an incomplete or incorrect conclusion due to; study groups being too small, errors in statistical methods, errors in equipment or apparatus, human bias or indeed by outright deceit. But to claim that half of science is false due to human failings is somewhat of a stretch. Of the problem of the mainstream media reporting a single study as 'science' reminds me of what Professor Spiegelhalter [1] said on More or Less this week, that if something is being reported as it is usually contrary to the established view and is therefore extraordinary. And that extraordinary claims require extraordinary proof. Which is never usually the case of a single research paper. So yes, half of all scientific papers in the last twenty years may not be repeatable but what we can then draw from that is that their conclusions and hypotheses are incorrect. That is the scientific method is it not? And it has worked remarkably well. <snip> > > > And in either case, any serious disputation demands support of > > objective evidence. Unless, of course, you're saying the scientific > > method itself is questionable -- in which case I'd humbly ask for your > > alternative way of understanding reality. > > > Heh, why do you go opening that can of worms? That wasn't something I > said, still if one goes there it soon begs a metaphysical question on > the nature of reality itself. I don't think Silk is a medium built for > that kind of debate. However there are some jumping off points for > those interested, > > 1. Quantum events are not deterministic, but probabilistic, which > requires reworking Francis Bacon's assumptions that experiments are > always repeatable. Such debates are currently only being held in > philosophy departments, and not in physics departments. > > Yet makers of still more accurate atomic clocks, random number > generators all run into this sooner or later. > What problems regarding atomic clock accuracy are you referring to? An article in this week's New Scientist mentions an atomic clock that is accurate to one second in a billion billion, it would be out by one second in 32 billion years. [2] The utility of the result in the framework to which it applies matters. An accuracy of an atomic clock that is greater than the age of the universe is probably fine for the uses to which it would be put. Also, consider the radioactive decay of phosphorous-32. It has a half-life of 14.29 days. This is well established and repeatable. We know what will happen to 10kg of the stuff in a period of time. However, the decay of individual atoms is indeed a matter of statistical probability. We are unable to determine which atoms will decay and which ones won't. But in the objective framework of real-world utility, we care not about individual atoms in this case but of the 10kg as a whole. I don't think that it is fair that you use an effect like quantum probability to make an assertion that experiments are not repeatable. <snip> > For most things it doesn't matter, but in sufficiently complex > problems like cancer research where 80% of research is not repeatable, > it matters. Scientists could only brush it aside as placebo effect for > so long. > What is your source that 80% of cancer research is not repeatable? > Science is built on the closed world assumption, which isn't the case > with the human body at the cellular level, where an open world > assumption makes more sense. > I don't understand what you refer to regards open world and closed world assumptions. Can you please elaborate? Cheers - Keith [1] http://www.statslab.cam.ac.uk/Dept/People/Spiegelhalter/davids.html [2] https://www.newscientist.com/article/2160951-super-accurate-atomic-clock-used-in-real-world-for-first-time/ ________________________ This email was scanned by Bitdefender
