5 ms·

Scientific Models Too Often Prove Whatever the Grant Provider Wants Proven

41 points by SQL2219 7y ago

teslaberry 7y ago
Yes this is specifically why some old school, perhaps slightly antiquated researchers in various fields, usually Men above the age of 70 , who have had very succesful careers in experimental areas of hard sciences, a NUMBER of them , not all, will often in their graduate office hours use the phrase "SIMULATIONS AREN'T SCIENCE"----becuase the data you get out is only as good as your assumptions, it's not sample data from the real world. and therefore the discussion usually ends with them disapproving of a research dissertation where the results are based on simulation outcomes. you can do or get whatever outcomes you get , whether by bias, by not bias, by laziness, sloppiness, or by over conviction of a particular hidden assumption you hold. Whatever it is is, if you want to prove something or falsify it, you can 100% control ALL you parameters in a simulation , even one with 'RANDOMNESS' built into it, which is bullshit because RANDOMNESS IS A useful TOOL but it's not a bias removal magic button. there's place for simulation. i'm just saying , it's a hallmark of shitty oversold bait science , especially when modelling ANY complex system . that's why much of what passes for cosmology is unfortunately nonsense , or at best , good theoretical math, until it is confirmed with space probe data, or other data sources from real world. Even then, real world cosmology data sometimes fits your model projections in an awkward way, because your model is bullshit.
Tycho 7y ago
I just assume most scientific research is worthless these days.
- everdev 7y ago
  Fake news is working. If you saturate the information channels with enough misinformation it overwhelms people and they give up. One great thing about science is that it encourages reproducing results. Many clickbait articles review un-reproduced or preliminary results. Those feel free to disregard. Science isn't perfect but peer-reviewed, reproducible experiments are still the best data source we have.
  - buzzkillington 7y ago
    >One great thing about science is that it encourages reproducing results. The replication crisis (or replicability crisis or reproducibility crisis) is, as of 2020, an ongoing methodological crisis in which it has been found that many scientific studies are difficult or impossible to replicate or reproduce. The replication crisis affects the social sciences and medicine most severely.[1][2] The crisis has long-standing roots; the phrase was coined in the early 2010s[3] as part of a growing awareness of the problem. The replication crisis represents an important body of research in the field of metascience.[4] https://en.wikipedia.org/wiki/Replication_crisis https://en.wikipedia.org/wiki/Replication_crisis
  - TeMPOraL 7y ago
    > If you saturate the information channels with enough misinformation it overwhelms people and they give up. But that's the whole problem. As the replication crisis is currently demonstrating, some scientific fields are absolutely saturated with misinformation - or, should I say bullshit, under Frankfurt's definition[0]. There isn't any cabal with an agenda trying to confuse people (or at least I believe there isn't) - it's just regular scientists that are incentivized to care less about whether their papers are correct, and more about churning out a lot of them to acquire grants and propel their careers. -- [0] - https://en.wikipedia.org/wiki/Bullshit#Harry_Frankfurt's_concept https://en.wikipedia.org/wiki/Bullshit#Harry_Frankfurt's_con... - tl;dr: bullshit is orthogonal to truths and lies; bullshitters don't care whether what they say is true or not - what matters is whether it serves their interests.
JorgeGT 7y ago
Note that there is absolutely no evidence in the article for the HN title. The only link in that section is to the Bible.
- viraptor 7y ago
  Yes. Also even though we do know of studies which were paid to produce specific results... that proves there is some quality control from wider community. Sure, we may have reproducibility crisis in different areas. But we also caught WV emissions issue and the study regarding MMR - even if they cause lots of issues, not all is lost - we keep advancing.
- solveit 7y ago
  The title also fails to even type-check. Models don't prove things, scientists using models do. While this may sound pedantic, it's not. A good part of the title's shock value disappears if you replace it with "Scientists too often prove...".
- buzzkillington 7y ago
  23. https://www.amazon.com/Overshoot-Ecological-Basis-Revolutionary-Change/dp/0252009886 24. http://www.paulchefurka.ca/Adaptive Cycles.html 25. https://www.youtube.com/embed/JYHX-Ib3Q5Q?version=3&rel=1&fs=1&autohide=2&showsearch=0&showinfo=1&iv_load_policy=1&wmode=transparent 26. https://www.internationalrivers.org/resources/8394 27. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2406852 28. https://history.state.gov/historicaldocuments/frus1969-76v38p1/d46 29. https://www.biblegateway.com/passage/?search=Revelation+18&version=NIV 30. https://en.wikipedia.org/wiki/Thor 31. https://www.bbc.com/reel/video/p07n0ww0/the-creature-behind-uganda-s-moral-monster 32. https://www.amazon.com/Oneness-Great-Principles-Shared-Religions-ebook/dp/B000XUBBR8
m3kw9 7y ago
Probably in the contract it will say if you can prove it, it triggers another $
everdev 7y ago
TLDR; Religious person thinks religion is helpful and renewable energy isn't: https://ourfiniteworld.com/author/gailtheactuary/ https://ourfiniteworld.com/author/gailtheactuary/
age_bronze 7y ago
The most frequent misunderstanding of science, (and it applies to many other fields such as cyber security security, politics, etc.) is that an imperfect solution is worthless, that if a model is incomplete it's as good as no model, etc. Perfect is the enemy of good. At the core of science is settling for the least errors model. You suggest solutions because doing them is better than doing nothing at all. Of course they will be partial if the problem is of such a large scale as climate change. That's not a reason not to do them. So what if all the politicians are dirty scumbags? Voting for none is still worse than voting for the least shitty politician. You need to accept you can't have a perfect solution. I find a perfectionist attitude to problems more common with religious people. Because once you already have one single generalizing solution to so many questions (god / religion), it is just so tempting to believe the same applies to other problems.
dane-pgp 7y ago
The article says: 'Australia’s attempt to put renewable electricity on the grid has sent electricity prices skyrocketing and resulted in increased blackouts. It has been said that intermittent electricity has “wrecked the grid” in Australia.' whereas reality is somewhat more nuanced: https://www.theguardian.com/australia-news/2019/dec/24/south-australias-clean-energy-shift-brings-lowest-power-prices-on-national-grid-audit-finds https://www.theguardian.com/australia-news/2019/dec/24/south... 'Hugh Saddler, the author of the monthly audit and an associate professor at the Australian National University, said wind and solar power supplied more than 50% of the electricity generated in South Australia for most months over the past two years. “That’s made electricity in SA the cheapest in the national electricity market and dramatically increased reliability.”' To give an idea of the quality of the analysis from "Our Finite World", consider this post from 2017: https://ourfiniteworld.com/2017/01/10/2017-the-year-when-the-world-economy-starts-coming-apart/ https://ourfiniteworld.com/2017/01/10/2017-the-year-when-the... "Things could start falling apart badly in 2017, or alternatively, major problems may be delayed until 2018 or 2019." or this post from 2015: https://ourfiniteworld.com/2015/08/26/deflationary-collapse-ahead/ https://ourfiniteworld.com/2015/08/26/deflationary-collapse-... "The big thing that is happening is that the world financial system is likely to collapse."
- buzzkillington 7y ago
  >'Hugh Saddler, the author of the monthly audit and an associate professor at the Australian National University, said wind and solar power supplied more than 50% of the electricity generated in South Australia for most months over the past two years. “That’s made electricity in SA the cheapest in the national electricity market and dramatically increased reliability.”' [...] the combination of significant network investment over the past decade, recent increases to gas prices, more concentrated wholesale markets, and the transition from large scale synchronous generation to variable and intermittent renewable energy resources has had a more pronounced effect on retail prices and number of offers in South Australia than any other state in the National Electricity Market. http://theconversation.com/factcheck-qanda-are-south-australias-high-electricity-prices-the-consequence-of-renewable-energy-policy-93594 http://theconversation.com/factcheck-qanda-are-south-austral...
- 7y ago
caterama 7y ago
Prompting purely from the HN headline "Scientific Models Too Often Prove Whatever the Grant Provider Wants Proven" (to quote in case it changes), I wanted to provide two small anecdotes. Preface: science rarely claim to prove anything. 1) It is often the case that researchers "suspect" the result before they have formally been given a grant to investigate. This ranges from forming a hypothesis based on other experience to having stumbled upon something interesting, and then afterward making a proposal for funding to thoroughly investigate. Both of these cases might look like the scientist is proving what the grant provider wants. 2) As @age_bronze hints at, models are often built to explain or reproduce observations. Think: "all models are wrong, some models are useful". Modern science tends towards "X is understood if X can be explained". Having a model that reproduces X is the correct direction to understand X, so the observation that "Scientists tasked with understanding X tend to produce models demonstrating X" is... _shocking_?

41 points by SQL2219 7y ago

Yes this is specifically why some old school, perhaps slightly antiquated researchers in various fields, usually Men above the age of 70 , who have had very succesful careers in experimental areas of hard sciences, a NUMBER of them , not all, will often in their graduate office hours use the phrase "SIMULATIONS AREN'T SCIENCE"----becuase the data you get out is only as good as your assumptions, it's not sample data from the real world. and therefore the discussion usually ends with them disapproving of a research dissertation where the results are based on simulation outcomes. you can do or get whatever outcomes you get , whether by bias, by not bias, by laziness, sloppiness, or by over conviction of a particular hidden assumption you hold. Whatever it is is, if you want to prove something or falsify it, you can 100% control ALL you parameters in a simulation , even one with 'RANDOMNESS' built into it, which is bullshit because RANDOMNESS IS A useful TOOL but it's not a bias removal magic button. there's place for simulation. i'm just saying , it's a hallmark of shitty oversold bait science , especially when modelling ANY complex system . that's why much of what passes for cosmology is unfortunately nonsense , or at best , good theoretical math, until it is confirmed with space probe data, or other data sources from real world. Even then, real world cosmology data sometimes fits your model projections in an awkward way, because your model is bullshit.

Tycho 7y ago

I just assume most scientific research is worthless these days.

everdev 7y ago
Fake news is working. If you saturate the information channels with enough misinformation it overwhelms people and they give up. One great thing about science is that it encourages reproducing results. Many clickbait articles review un-reproduced or preliminary results. Those feel free to disregard. Science isn't perfect but peer-reviewed, reproducible experiments are still the best data source we have.
- buzzkillington 7y ago
  >One great thing about science is that it encourages reproducing results. The replication crisis (or replicability crisis or reproducibility crisis) is, as of 2020, an ongoing methodological crisis in which it has been found that many scientific studies are difficult or impossible to replicate or reproduce. The replication crisis affects the social sciences and medicine most severely.[1][2] The crisis has long-standing roots; the phrase was coined in the early 2010s[3] as part of a growing awareness of the problem. The replication crisis represents an important body of research in the field of metascience.[4] https://en.wikipedia.org/wiki/Replication_crisis https://en.wikipedia.org/wiki/Replication_crisis
- TeMPOraL 7y ago
  > If you saturate the information channels with enough misinformation it overwhelms people and they give up. But that's the whole problem. As the replication crisis is currently demonstrating, some scientific fields are absolutely saturated with misinformation - or, should I say bullshit, under Frankfurt's definition[0]. There isn't any cabal with an agenda trying to confuse people (or at least I believe there isn't) - it's just regular scientists that are incentivized to care less about whether their papers are correct, and more about churning out a lot of them to acquire grants and propel their careers. -- [0] - https://en.wikipedia.org/wiki/Bullshit#Harry_Frankfurt's_concept https://en.wikipedia.org/wiki/Bullshit#Harry_Frankfurt's_con... - tl;dr: bullshit is orthogonal to truths and lies; bullshitters don't care whether what they say is true or not - what matters is whether it serves their interests.

JorgeGT 7y ago

Note that there is absolutely no evidence in the article for the HN title. The only link in that section is to the Bible.

viraptor 7y ago
Yes. Also even though we do know of studies which were paid to produce specific results... that proves there is some quality control from wider community. Sure, we may have reproducibility crisis in different areas. But we also caught WV emissions issue and the study regarding MMR - even if they cause lots of issues, not all is lost - we keep advancing.
solveit 7y ago
The title also fails to even type-check. Models don't prove things, scientists using models do. While this may sound pedantic, it's not. A good part of the title's shock value disappears if you replace it with "Scientists too often prove...".
buzzkillington 7y ago
23. https://www.amazon.com/Overshoot-Ecological-Basis-Revolutionary-Change/dp/0252009886 24. http://www.paulchefurka.ca/Adaptive Cycles.html 25. https://www.youtube.com/embed/JYHX-Ib3Q5Q?version=3&rel=1&fs=1&autohide=2&showsearch=0&showinfo=1&iv_load_policy=1&wmode=transparent 26. https://www.internationalrivers.org/resources/8394 27. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2406852 28. https://history.state.gov/historicaldocuments/frus1969-76v38p1/d46 29. https://www.biblegateway.com/passage/?search=Revelation+18&version=NIV 30. https://en.wikipedia.org/wiki/Thor 31. https://www.bbc.com/reel/video/p07n0ww0/the-creature-behind-uganda-s-moral-monster 32. https://www.amazon.com/Oneness-Great-Principles-Shared-Religions-ebook/dp/B000XUBBR8

m3kw9 7y ago

Probably in the contract it will say if you can prove it, it triggers another $

everdev 7y ago

TLDR; Religious person thinks religion is helpful and renewable energy isn't: https://ourfiniteworld.com/author/gailtheactuary/ https://ourfiniteworld.com/author/gailtheactuary/

age_bronze 7y ago

The most frequent misunderstanding of science, (and it applies to many other fields such as cyber security security, politics, etc.) is that an imperfect solution is worthless, that if a model is incomplete it's as good as no model, etc. Perfect is the enemy of good. At the core of science is settling for the least errors model. You suggest solutions because doing them is better than doing nothing at all. Of course they will be partial if the problem is of such a large scale as climate change. That's not a reason not to do them. So what if all the politicians are dirty scumbags? Voting for none is still worse than voting for the least shitty politician. You need to accept you can't have a perfect solution. I find a perfectionist attitude to problems more common with religious people. Because once you already have one single generalizing solution to so many questions (god / religion), it is just so tempting to believe the same applies to other problems.

dane-pgp 7y ago

The article says: 'Australia’s attempt to put renewable electricity on the grid has sent electricity prices skyrocketing and resulted in increased blackouts. It has been said that intermittent electricity has “wrecked the grid” in Australia.' whereas reality is somewhat more nuanced: https://www.theguardian.com/australia-news/2019/dec/24/south-australias-clean-energy-shift-brings-lowest-power-prices-on-national-grid-audit-finds https://www.theguardian.com/australia-news/2019/dec/24/south... 'Hugh Saddler, the author of the monthly audit and an associate professor at the Australian National University, said wind and solar power supplied more than 50% of the electricity generated in South Australia for most months over the past two years. “That’s made electricity in SA the cheapest in the national electricity market and dramatically increased reliability.”' To give an idea of the quality of the analysis from "Our Finite World", consider this post from 2017: https://ourfiniteworld.com/2017/01/10/2017-the-year-when-the-world-economy-starts-coming-apart/ https://ourfiniteworld.com/2017/01/10/2017-the-year-when-the... "Things could start falling apart badly in 2017, or alternatively, major problems may be delayed until 2018 or 2019." or this post from 2015: https://ourfiniteworld.com/2015/08/26/deflationary-collapse-ahead/ https://ourfiniteworld.com/2015/08/26/deflationary-collapse-... "The big thing that is happening is that the world financial system is likely to collapse."

buzzkillington 7y ago
>'Hugh Saddler, the author of the monthly audit and an associate professor at the Australian National University, said wind and solar power supplied more than 50% of the electricity generated in South Australia for most months over the past two years. “That’s made electricity in SA the cheapest in the national electricity market and dramatically increased reliability.”' [...] the combination of significant network investment over the past decade, recent increases to gas prices, more concentrated wholesale markets, and the transition from large scale synchronous generation to variable and intermittent renewable energy resources has had a more pronounced effect on retail prices and number of offers in South Australia than any other state in the National Electricity Market. http://theconversation.com/factcheck-qanda-are-south-australias-high-electricity-prices-the-consequence-of-renewable-energy-policy-93594 http://theconversation.com/factcheck-qanda-are-south-austral...
7y ago

caterama 7y ago

Prompting purely from the HN headline "Scientific Models Too Often Prove Whatever the Grant Provider Wants Proven" (to quote in case it changes), I wanted to provide two small anecdotes. Preface: science rarely claim to prove anything. 1) It is often the case that researchers "suspect" the result before they have formally been given a grant to investigate. This ranges from forming a hypothesis based on other experience to having stumbled upon something interesting, and then afterward making a proposal for funding to thoroughly investigate. Both of these cases might look like the scientist is proving what the grant provider wants. 2) As @age_bronze hints at, models are often built to explain or reproduce observations. Think: "all models are wrong, some models are useful". Modern science tends towards "X is understood if X can be explained". Having a model that reproduces X is the correct direction to understand X, so the observation that "Scientists tasked with understanding X tend to produce models demonstrating X" is... _shocking_?