LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   General (https://www.linuxquestions.org/questions/general-10/)
-   -   Validity of research results. (https://www.linuxquestions.org/questions/general-10/validity-of-research-results-4175694937/)

business_kid 05-09-2021 02:13 PM

Validity of research results.
 
https://science.slashdot.org/story/2...hing-is-a-joke

I started slightly skeptical, because I was an assistant in many Electronic papers where real risks were disguised, problems ignored or skirted around, lies told and so forth. My name was on none of them (I was 'only' a technician at the time) but in one case I flat out contradicted a professor, and won the whole department around to my side inside a few months, as his product bombed exactly where I predicted.

So reading the link above along with other info
  • A paper entitled "Why most publish research is false,"
  • pub peer
  • Retraction Watch
  • The discovery that randomly generated papers had passed peer review.
  • The horrible history of drugs released on an unsuspecting public by rather ruthless drugs companies

Have only added to my skepticism.

The medical practices of the FDA are noteworthy. A manufacturer has to produce two favourable papers. Said manufacturer may plough under or shred any number of negative reviews, which happens. The FDA don't want to see all the research, just the positive papers. The work of doing the "research" in the form of trials is usually contracted out, and the contractor has to do the trials. If a lie is told on the 'randomly' selected line, it would be very difficult to prove. Let's illustrate a sample double blind trial of 100, with half given the drug. So your sample size per trial is 50.

The drug's efficiency is just over 20%, but there are side effects in many. So 11/50 improve. But if you do 5 such trials, then 55 people will improve. Now two "random" trials can be organised, one with 27/50 & the other with 28/50 improving. Severe effects can be listed an minor, and you're good to go with FDA approval. Sales talk will say 'No known side effects.'

And even with a class action lawsuit, the drug company can recover the costs by price increases. All these cures for osteoporosis prescribed have lost class actions costing them $10 - $15 millions in the States alone. The reason is simple: the bones naturally destroy old bone, and build new bone in accord with your exercise pattern. So Space Shuttle astronauts in weightlessness all take a big calcium hit. These osteoporosis drugs can only redistribute existing (and insufficient) existing calcium in bones. There are class actions against them
https://kevinthomas63.wordpress.com/...rage-expanded/
https://www.lawyersandsettlements.co...s-adverse.html

obobskivich 05-10-2021 03:38 AM

In the US at least, I think a big part of why the 'xkcd-style' publications happen is an imperative within universities to 'publish or perish' as a way of showing 'work product' among tenure-track employees. There are various, complicated, reasons for this but most of them circle back to policy choices (which are usually disconnected from the working scientists). However, taking that to a hyperbolic extreme and giving into psuedo-skepticism wrt science is also (sadly) becoming a political force as well - first we couldn't make the scientists (or nature) say what we wanted it to say, so now we just reject reality and substitute it with our own...

business_kid 05-10-2021 04:12 AM

Interesting.

I'm in touch occasionally with a bioligist/genecitist guy in Europe who does his own well researched papers, and he's inclined to be 9-18 months per paper. Each paper is different, research is very complete, with loads of references. He's retired now, but still publishes on awkward subjects. Only once did he do two successive papers on the same subject. My own paper was a good 3-4 months work, not including circuit & pcb design, which pushed it over 9 months.

My opinion was that guys doing 1paper a week are chancers gaming the system.

hazel 05-10-2021 04:40 AM

It's hardly a new problem. I went for a job interview back in 1975 or thereabouts and one of the interviewers asked me what I thought of the information explosion. So I told him what I actually thought: that there wasn't an information explosion at all, there was a noise explosion. Lots of scientists publishing meaningless papers because your publication rate determined your promotion prospects. He mentioned to me a few years later that that was the comment which got me the job.

igadoter 05-10-2021 05:02 AM

At the it about financing scientific research. To convince governments to waste enormous money on sometimes completely strange and useless research. Useless in terms of applications - to create technology, products, etc. Myself I provide following justification: science is like dense fish net - we just want to catch golden fish - but golden fish is really very small - so net has to be very dense. But once we catch it - we will ask our three wishes - and world will change forever. In other words - chance to catch are very small - but no matter we have to try - cause prize will be enormous. I don't know: wish something to yourself - health, clean environment, no poor people, no starvation - yeah just create kind of utopia. Golden fish will assure all this and more.

enorbet 05-10-2021 12:35 PM

In every human endeavor there are more hacks and posers than craftsman. To my way of thinking, that just makes the craft and actual craftsmen more valuable. The same can be said about about the so-called Information Age. Overall it IS just noise, just as if one somehow placed a microphone over a city and could hear all conversations at once - cacophony. Break it down, filter out individual conversations and then sift through the innocuous and voila! - golden fish. If this weren't so nearly every business and government on Earth would be spending huge sums in surveillance and subscription.

business_kid 05-11-2021 05:56 AM

Right, so the genuine paper is a gem amid an ocean of trash.

The point becomes, 'how do you isolate the gem from things that try so hard to imitate it?'

And do you see why I treat with skepticism every assertion that because of the belief that if something is "scientifically proved" it's gospel, or fact which must be accepted? In the case of some subjects, papers espousing an opposing viewpoint cannot be published in peer reviewed journals.

Can you imagine if a vaccine maker employed a scientist to examine their vaccine, and his conclusions were "This vaccine sucks, has horrible side effects, and noxious chemicals capable of causing long term damage in sensitive people?"

No, any scientist is given his conclusions and told to do research to find them.

business_kid 05-11-2021 06:51 AM

Just musing here... I was looking for a random generator of medical papers and trying to get it talking about vaccines. But I didn't find it. It could be very dangerous if it existed.

Academics can lean on their students to game the system. So a paper with multiple authors need involve no work from anyone, except those seeking their degree. Surely every lecturer can get on every student's paper, and it's counted to their credit. I did what I regard as my single paper under 4 or 5 lecturers, all of whom had PhDs, and could have gone on in front of me with my humble B.E. But I was also required to submit 2 papers every year with a project. Usually my Xmas papers were unusable, as I spent too much time slagging them or myself off, much to everyone's amusement. But the summer ones had to be straight for external examiners. So that meant 70 unpublished papers just from my year, many of which could be published. It's sad to think science has sunk to this.

enorbet 05-11-2021 08:46 PM

There is a whole world of difference between Pure Science and what amounts to Commercial Science. Just as 2 opposing lawyers, DA and Defense, on the same case can certainly find some "expert" willing to accept financial reward to fit into that side's view of the case, so can corporations, at least proportional to how much money is at stake.

That's VERY different from research in areas that yet have no commercial application, so one way we can sift through research to separate wheat from chaff is "Follow the money". Even if it can be concluded that a breakthrough might be very lucrative in just a few years, there are so many that will almost assuredly be profitable in days or weeks, most profiteers and their backers and facilitators, prefer to deal in those "surefire" schemes. That's where one's skepticism should be on high alert...., not some broad brush prejudice against research in general.

Regarding vaccines, and I suppose you are leaning heavily on Covid vaccines here, the numbers are already in and growing by the day. The US has something like 60%-70% deployment and things are opening up with 3 weeks of no serious climb in rates, either of new infection or deaths even in LA and NYC, while India, with somewhere considerably less than 25% is in severe crisis.

There's no such thing as Zero Risk overall in Life. All we can hope for is good odds. 1,000,000 to 1 qualifies in my book.

To attempt top clarity, certainly their is a lot of money to be made by creating and selling Covid vaccine...BUT there is also HUGE risk if that vaccine does little or nothing with Covid and/or has a high percentage of bad reactions. It is probably likely that there is some corporate warfare going on to spread FUD about competitor's product (J&J comes to mind) but that can't detract from the real world validity of the Moderna, Pfizer, and Astra Zeneca (and probably others unfamiliar to me) numbers.

business_kid 05-12-2021 09:47 AM

First of all, I generally agree with your comments on commercial science and it's corruptibility. I wasn't thinking only of Covid Vaccines, but you are entitled to make your assessment of vaccines and your risk/benefit analysis as you see fit. I think it is shallow to take death as the sole criterion for that analysis. There are actually worse things than dying, such as severe autism for a lifetime, which destroys more than one life. I feel, nevertheless, that you are entitled to your POV.

What interests me most in your post in the designation of "Pure Science." I take this to mean Scientific research for purely Altruistic and no financial motives, advancing the understanding of some concept. With the exception perhaps of some solutions to long standing problems, I would challenge you to produce a few modern recent examples of such research, as I feel it has basically died out. I can point to 19th century examples, where men such as Snell (of Snell's Law), Kelvin (of Kelvin's Laws), naturalists, and so forth. In Ireland, the third Earl of Rosse had a 72" refractor telescope built in Birr Castle. He discovered the Spiral nature of galaxies, and anyone wishing to observe such until 1905 had to come to Birr in Ireland to view them. The telescope was still working when I visited less than 10 years ago.

Such pioneers mainly were self financing, or relied on funding from such organisations as
  • The Smithsonian Institute
  • The Royal Asiatic Society in the UK.
  • Individuals (landed gentry usually) in the various 'Gentlemen's clubs' in various capitals, or rich estate holders.
  • Bequests from Governments or Sovereigns.

These days, all of these have vanished as sources of funding, and expenses have skyrocketed. The best place for "Pure Research" is Academia, but Universities have their financial realities, too. Universities in the main are under-resourced, overcrowded, and subject to the realities of the Academic year. No project is undertaken without consideration of the prospect of a reward - the speaking tours, the job offers, the book, the professorship, etc. That's the reality of a Capitalist system.

So where is Pure research? Let's have some concrete recent examples.

enorbet 05-12-2021 03:42 PM

The first satellites went into orbit starting with Sputnik in 1957. For at least a decade in the US more exploration rockets failed than succeeded... total losses. It took over a half century before satellites would become remotely commercial. Voyager, conceived in 1968 didn't launch for almost 10 years... 10 years of developmental expense. Now, 40 years later both still send back scientifically useful data that has (as of yet) zero commercial application. Of course we can add the exploration of planets, our Moon, asteroids and comets these days.

It's great that FermiLab is still producing useful data but it pales compared to what the SuperCollider could have accomplished. Thankfully the Large Hadron Collider took up the slack and then some, but though they employ thousands of top scientists, the multi-billion dollar costs are viewed as a waste of time by commercial interests. Also thankfully, they are already planning on a bigger one. None of them profited in dollars and cents.

The LIGO gravity wave observatories herald a whole new branch of Astronomy that, since light could not escape the hot dense early Universe but gravity did, we will be able to "see" back far closer to the moment of the Big Bang as well as other remote events around us, obviously of zero commercial value.

Much less in the public eye but amounting to huge time effort and money spent, that maybe just maybe might someday have commercial value is research into Physics, Biology, Botany, Chemistry, Archaeology, Materials, and every other branch of Science exactly because knowledge is power. Certainly the more abstract has difficulties as it always has but it most certainly hasn't stopped.

A child of 9 began building a Fusion Reactor in his parents' garage, finally successfully completed at age 14. With a handful of off-the-shelf parts amounting to less than $100 any person handy with electronics can build a Tunneling Electron Microscope in his kitchen, basement, or bedroom, an accomplishment costing tens of thousands of dollars and a college education just 20 years ago. A scientist wishing to model Black Hole radiation built a bona fide supercomputer with several gaming consoles and accomplished serious research.

Exploration for exploration's sake is thankfully part of the Human Condition. It didn't and is not going to just "die on the vine"

business_kid 05-13-2021 04:29 AM

The actions of Governments are not motiveless. "Sure, let's all get together and throw billions into this idea." I think not.

enorbet 05-15-2021 07:51 PM

What actions by any collective OR individual are motiveless? Given how low the percentage of GNP is spent on Pure Science (I wish it was 10 times as much) I'm just glad there are at least a few wise supporters in our government. It is a very low percentage of citizens who recognize all the spinoffs from even just the Cold War "Space Race". I care little that it was prompted by a nasty idealogical conflict spun as altruism and human spirit. The results are the same - some learned that eventually exploring pays off, even if "eventual" might be measured in decades or even a hundred years, and as counter-intuitive as that is to common business practice, it's good that this gets lauded and recognized.

Elon Musk has at least 4 enterprises, each of which is not motivated by immediate ROI, and it's not to get rich since he was already there and even risked it all - put his money where his heart is - and all of us on Earth are and will continue to be better off because of it. Yes he is motivated but not by greed. He just wants to improve quality of Life for everyone everywhere and I'm overjoyed it has begun to earn profits. Because of that and his example, he is not alone. Humanity needs such examples.

Too bad Venezuela didn't notice in time. Government might be classifiable as a necessary evil but there certainly are degrees.

ntubski 05-20-2021 10:12 PM

Quote:

Originally Posted by business_kid (Post 6249968)
Just musing here... I was looking for a random generator of medical papers and trying to get it talking about vaccines. But I didn't find it. It could be very dangerous if it existed.

Not exactly what you're looking for, but related and kind of amusing: a game where you're shown 2 title+abstracts and you guess which is fake. I find "hard mode" is quite difficult. I got 11 correct out of 20, which is just barely better than a coin flip...

https://www.stefanzukin.com/enigma/

business_kid 05-21-2021 04:42 AM

I gave up earlier - I got 4-0 correct on hard mode, lost the concentration and went to 4-3. You need to look up the gobbledygook, and I couldn't be bothered.


All times are GMT -5. The time now is 09:44 PM.