Posted on 02/25/2017 7:21:05 PM PST by combat_boots
That used to be the way it was done. I don’t know if they still do that. It would appear that they don’t.
well FRiend, i’d have to know you a lot better before i gave you anything, but i’ll continue to hold on to what’s real, true and beautiful, especially my Jesus.
As a good example, take "global warming"; it is merely an unvalidated assertion (not an hypothesis, as (IIRC) observations are required for that).
Too many want fame and are willing to sacrifice their integrity.
ROTFL. Most of what people call science these days is just bull****. Gotta get those government grants.
And that is why climate "science" is being shown up as a modern day medicine show. Scientific Method starts out with a hypothesis which is proved either correct or incorrect by experiments. In climate "science" there is a politically motivated hypothesis but there are no experiments, there are only GIGO infected computer models, thousands upon thousands of GIGO infected computer models which were no better at predicting tomorrow'sweather much less the climate in fifteen or twenty years. The chicanery that followed requests for data was, and continues to be, mind boggling. And the only way to shut down the discussion was to declare that the "science" was settled.
Self styled "scientists" don't seem to know that replication is one of the sine qua non of science.
But the frauds get all the press.
Who knew?
When most folks think of research they think in terms of a closed system. A good example is boiling water in a jar. At sea level, you can count on it boiling at 212 degrees every time. There are only a few things that affect the boiling point. Clean the jar. Use pure water. Get an accurate thermometer. The things to consider are easily identified.
Now try to do cancer research on people and there are hundreds of things to consider. Many of those things people have no clue about and therefore never attempt to control for. For many others, there in no way to know their relative importance. Even if you end up with a good result, was it caused by the drug or did it just happen at the same time? That is the conundrum of an open system.
Open systems can only be analyzed with statistics which gives probabilities and never yes or no answers. The key here is picking the correct method and then not pushing analysis beyond what the math truthfully says. Using good mathematics, it is hard to drag useful information out of an open system.
Here is link to a report which rightfully condemns global warming for the reason I mentioned.
2006 Global Warming Congressional (Statistical Report)
http://www.google.com/search?hl=en&q=ad+hoc+committee+report+hockey+stick+global+climate+reconstruction&btnG=Search
Select the PDF
If the purpose of your error bars is to prove that an effect is real, chances are the experiment will not be reproducible.
If the purpose of your error bars is to show off the thoroughness of your work to the reviewers before publication, it’s probably a reproducible result.
not to mention it’s also a funny verb
“Writing in the latest edition of Nature, he outlines a new approach to animal studies that calls for independent, statistically rigorous confirmation of a paper’s central hypothesis before publication.”
Sounds good. The problem is who is going to pay for the replication of the study? Who is going to spend the time and effort, at the expense of their own work to do it? Who is going to pay for the supplies? Its just not practical. Thats why most science is done on the honor system.
There is less of a reproducibility problem in industrial science because the work we do has a practical purpose. If we develop something it is because someone wants to use or sell it. If it doesn’t work it will become apparent soon enough. Even so I do know of one instance of someone faking data. The guy was fired but quickly found a job at a research hospital. For legal purposes no one could tell his new employer why this guy was suddenly looking for work, so he was free to continue is faux research for a different employer.
The problem with science is that science doesn't work at all with peer review. Only by using the scientific method of having other reproduce your work produces working science. The honor system leads to our current state of lots of money wasted but no real science produced.
Western science took off with the Royal Society requiring all scientific discoveries to reproduced multiple times before being accepted. Yes it's costly, but it triggered the scientific revolution. We've abandoned that system after WW2 in the name of saving money and now we waste 10 times as much money for very little real science.
When it comes down to it you can either have real science with accounting(scientific method) or you can fake science with scientific priests certifying that's it's real when it's not(peer review).
I was in the business. What you say is accurate and true. The sheer number of "certain of these experiments" is mind boggling. The U.S. Government expends billions of tax on grants funding the above, and it is beyond rare that they fund for replicated confirmation from independent researchers.
“When it comes down to it you can either have real science with accounting(scientific method) or you can fake science with scientific priests certifying that’s it’s real when it’s not(peer review).”
The purpose of peer review is not to confirm something is reproducible. It is to insure the scientific methods/statistics are appropriate, the work is original, the proper controls have been done, the conclusions are reasonable, that relevant work in the field is cited, the writing and figures are understandable, and that mistakes are caught and corrected before publication.
Plenty of real science has been done since WWII and the system works quite well when it is not politicized or conducted by ideologues intent on using it to prove one of their pet causes. Most of the problems are in the social sciences/humanities and other areas where opinion is considered good data, proper controls are not possible and IMO politics is rampant. As for reproducibility, that is much less a problem in the harder, more practical sciences. If you publish a new and useful chemical synthesis, someone is going to try and use it. If you say you built an improved biothreat detector, someone is going to test it. If it doesnt work the customer will tell you and his friends in the field.
>The purpose of peer review is not to confirm something is reproducible. It is to insure the scientific methods/statistics are appropriate, the work is original, the proper controls have been done, the conclusions are reasonable, that relevant work in the field is cited, the writing and figures are understandable, and that mistakes are caught and corrected before publication.
So it’s useless in producing working science?
>Plenty of real science has been done since WWII and the system works quite well when it is not politicized or conducted by ideologues intent on using it to prove one of their pet causes.
Sure plenty of real science at the start because peer review started with scientists trained in the old ways. However, the farther along we go the less and less real science gets done. Where as if we returned to the scientific method we’d do nothing but real science. There’s no incentive to cheat or distort findings with non peer reviewed science and lots of incentives to do with peer review.
>Most of the problems are in the social sciences/humanities and other areas where opinion is considered good data, proper controls are not possible and IMO politics is rampant.
Medical science can only reproduce around 40% of published results on a random sample. IE 60% of medical science is garbage. Things are much worse than you think.
>As for reproducibility, that is much less a problem in the harder, more practical sciences. If you publish a new and useful chemical synthesis, someone is going to try and use it.
Sure they try to use it, assuming that you published all the details and data, which rarely happens these days. Then it doesn’t work, you drop the issue and move on. But the original researcher still got credit for scientific fraud and thus the system rewards false results.
>If you say you built an improved biothreat detector, someone is going to test it. If it doesnt work the customer will tell you and his friends in the field.
So? There’s no feed back to the original researcher who got credit for it. With a system like that you’ve effectively front loaded the rewards with no or rare punishment for fraud or bad work. With the scientific method of reproduction credit was never given until it was proven thus giving a massive incentive towards good work and advancement of science.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.