Saturday, 7 January 2017

Modern sciences amd Homeopathy and Other Things

A friend in another thread asked why I was sceptical of modern science, and why I criticized those who returned to villages after IIT. Regarding the second, I have only respect for those who return to villages to work. I have only questioned why the fact of their having studied in IIT or in the local Arts college needs to be mentioned. It is only the ethical choice they made that counts. To mention IIT implies that that is a superior caste to the local Arts college. I beg to differ there.
Regarding science - I have no fight with modern science. But I refuse to grant it the position of the ONLY valid system. And I also refuse the demand that other systems, including ayurveda, need to be validated in its language.

Srividhya Gopalakrishnan i sometimes seriously think that there has to be way of collecting back the subsidy offered to them after they graduate and start work
Shyamala Sanyal Not just the subsidy but some continuing return too
Narayana Sarma No need to worry anymore. From next year on fees at IITs are going to be one and a half lakhs per semester. Move on to a new era where IITs become self financing institutions. A move so you won't ever feel guilty later. 
Aparna Krishnan The support a poor country gives thus goes far beyond the money ma. The dreams of the country are invested in them, at the cost of support that otherwise needed to be given to desperate villages. They need to return the debt with sweat and tears and loyalty to the nation that nurtured them. Otherwise let them keep the money as a gift from a poor country that they claimed and took !
T.R. Shashwath All this raises the question - what exactly do you think is "modern" science, and how does it differ from "ancient" science?

To my mind, "science" implies a rational and skeptical outlook. It's the scientific method that is important, not mumbo jumbo.
Aparna Krishnan well. if snake bites respond to mantrams, then mantrams work. if not, otherwise. period. yes, all verification needs to be done.
T.R. Shashwath Here's the thing; once verification is done, you're not allowed to say "but it worked for me" - you need to look at the preponderance of data.

And the mechanism of verification is important too - you can't just say "somebody knows somebody". It has to be a proper study. If possible, double-blind.

Under those parameters, my hypothesis is that your mantrams will not survive. But those are the only parameters that are worth talking about.
Aparna Krishnan yes. agreed. test.
Balachander Swaminathan In the modern world, 'scientific' seems to be what is defined in publication of four publishers...Elsevier, Springer, ACS and ... (one more which I forget).... who publish more than 90% of what is considered scientific publishings....I would never be willing to swear by anything that is controlled by such a small group... Also to all people who swear by what has been taught to us as 'scientific' I like to point to this essay by venki,, where the difference between 'small truth' and 'greater truth' and 'convergent' and 'divergent' problems is explored.... It helps clear our head around the fact that a fixed set of processes and metrics cannot be used to measure everything else.... and if that is used, all other systems would always be found wanting.... example... if instead of short term yield, we were to use 'inches of top soil created' as the metric for agri productivity, where would the 'published studies' on chemical based farming stand?... and with new topsoil created we may even have better yield and nutrition over the long term... Once we marry ourself to fixed metrics like double-blind peer review (and having been involved in research i fully understand how non water-tight those processes are), we constrain our ability to understand what is beyond the boundaries of that system.... and thats what I believe is the main problem with people who swear by allopathy and can't accept other forms of healing... Disclaimer: I have, for several years used both, allopathy and homeopathy and both have worked for me... homeopathy with no side effects.... so now homeopathy is the first line of defence for me and it works 95% of the time... and for the other 5% I'm happy to try out allopathy or anything else

NASA's Marshall Space Flight Center Photo Stream @ FlickrVenkataraman…
Aparna Krishnan Balachander Swaminathan, could you give the critique of the double blind study. I have it thrown at me all the time in connection with our ayurveda.
T.R. Shashwath Balachander Swaminathan Don't get me started on Elsevier! Bloody rent-seekers!

On the other hand, scientific publications *don't make science what it is*. The scientists who publish do. The problem with the publications is that they have very shady copyright rules, not that what they publish is crap.

"Also to all people who swear by what has been taught to us as 'scientific' "

Again, I'm not talking about what we sucked up from textbooks and vomited on exam sheets. I'm talking about the scientific method.

"Once we marry ourself to fixed metrics like double-blind peer review"

If you were really into medical research, you would never have used such an abortion of a phrase. Double-blind studies and peer review are two *entirely different things*.
T.R. Shashwath Aparna Krishnan In case you don't know, double blind studies are those in which there are a large group of subjects, a researcher and somebody who actually administers the study. For example, let's say it's a drug that's being tested. In a double blind study, half the participants would get the actual drug, and the other half would get a placebo (or maybe another older drug, depending on the study). The guy who administers the test doesn't know which is which, and the participants don't know what they're getting. That way, they administer the test, and then gather the results without actually knowing what the medication really was.

The researcher (who knows who got what) can then correlate between the two groups and see what the difference in efficacy was. If the drug was really effective, you would expect at least a 95% confidence that it was more effective than the placebo or other drug. So, if 50% of the control group got better and only 51% of the study group did, that would be considered no better than the placebo effect. However, if 90% of the study group did, confidence levels would be much higher.

For sure, p-values can be hacked using selective reporting and so on. Which is why we have peer review to ensure that the computations were done correctly.

One process that we often miss is reproducibility. It is always best if this kind of study is done by multiple groups, to eliminate researcher bias or just some error in data collection. In their rush to discover new things, most groups don't work on reproducing old results. This is definitely a problem that has been recognised by many people.
T.R. Shashwath I think he means p-hacking. That was found to be a major problem in literally thousands of studies coming out of China for example (not to say that China is all bad - thousands more were absolutely great). They had to throw out lots and lots of data.

Thankfully, in medicine, those were mostly for treatments that were yet to hit the market.

p-hacking means that you fiddle with the data through selective reporting, or just by manipulating the statistics, in order to show a significance to the technique under study that's just not there.

For example, it is possible, using the right dividing line, to get a result of high significance even with a 1% difference between two populations.
T.R. Shashwath More explanation: "p" is the probability of the null hypothesis - that is, the control group shows the same behaviour as the study group. If p < 0.05, it's usually taken as a statistically significant result. But several manipulations by the researcher can actually reduce p quite easily.

Because of this, people have been abandoning pure statistical significance, and using things like confidence levels along with it. And peer review and reproduction do help.
Balachander Swaminathan p hacking is one of the things happenning and happens in medicine as well (google for 'the cholesterol myth'.... deeper subject... don't want to hijack this thread with this issue)... Other issues also like the non p hacking related facts like the Gates foundation recent funding of the Merck cervical vaccine given to minors in Khammam, which resulted in a few deaths and other issues... with the intent of measuring effect of vaccine on minors, but showing it as an act of philanthrophy...... However as I understand, this thread is not about the malpractices, it is about naming something that does not fit into a paradigm which people call 'scientific' as bogus.... With double blind peer review, I was referring to the review mechanism which many journals use....
Balachander Swaminathan I identify with Dr KC Kumarappa's thoughts and would say that his book 'Economy of Permanence', which only recently read, helped clarify a lot of what I was feeling.... For understanding and reflecting on the case of 'Science' Vs 'everything else', I would suggest reading/re-reading the chapter on 'Valuation' (and in fact the whole book) from that book. I firmly believe what we call 'scientific' is just one method (with multiple methodologies within that framework), from which mankind has benefited... Mankind also has benefited immensely from systems which are outside the understanding of what is generally referred to as 'scientific', and it is in our good interests to give it due importance and take the time out to understand them... Sadly our prescribed education system completely ignores these to the detriment of the human race

No comments:

Post a Comment