Know-Nothing Know-It-All/Quotes: Difference between revisions
Content added Content deleted
No edit summary |
mNo edit summary |
||
| Line 13: | Line 13: | ||
there is a gem in the next paragraph: |
there is a gem in the next paragraph: |
||
{{indent}} Cox, a physicist who works on the Large Hadron Collider where the Higgs boson was discovered, said that 95% certainty in science is effectively total. |
{{indent}} Cox, a physicist who works on the Large Hadron Collider where the Higgs boson was discovered, said that 95% certainty in science is effectively total. |
||
Wow. If he actually were a competent particle physicist rather than a pathetic environmentalist clown pretending to be a caricature of [[Richard Feynman|a scientist]], he would know that the 95% confidence – also known as 2-sigma evidence – is almost equivalent to no evidence at all<ref>aka "[https://duckduckgo.com/?q{{=}}%222 |
Wow. If he actually were a competent particle physicist rather than a pathetic environmentalist clown pretending to be a caricature of [[Richard Feynman|a scientist]], he would know that the 95% confidence – also known as 2-sigma evidence – is almost equivalent to no evidence at all<ref>aka "[https://duckduckgo.com/?q{{=}}%222+sigma+evidence%22+%22marginal%22 marginal]"</ref>. [...] |
||
As all actual particle physicists would agree, 95% is just an insignificant hint. The hints only become evidence worth discussing when the confidence approaches the 99.7% or 3-sigma level and an effect is only claimed to be discovered when the confidence reaches 99.9999% or 5-sigma. This convention of hard science is no excessively demanding dream. If a new effect is actually there, collecting the 5-sigma evidence takes just slightly (6.25 times) more data than accumulating the 2-sigma evidence. If a 2-sigma hint refuses to grow to 5 sigma for a decade or two, you may be pretty sure that it was a fluke and the effect doesn't exist. And indeed, most effects that are, at one point, observed to be there at the 95% confidence level ultimately go away. |
As all actual particle physicists would agree, 95% is just an insignificant hint. The hints only become evidence worth discussing when the confidence approaches the 99.7% or 3-sigma level and an effect is only claimed to be discovered when the confidence reaches 99.9999% or 5-sigma. This convention of hard science is no excessively demanding dream. If a new effect is actually there, collecting the 5-sigma evidence takes just slightly (6.25 times) more data than accumulating the 2-sigma evidence. If a 2-sigma hint refuses to grow to 5 sigma for a decade or two, you may be pretty sure that it was a fluke and the effect doesn't exist. And indeed, most effects that are, at one point, observed to be there at the 95% confidence level ultimately go away. |
||
| "[https://motls.blogspot.com/2014/09/brian-coxs-incompetence.html Brian Cox's incompetence]" by '''Luboš Motl''' on ''The Reference Frame'' blog }} |
| "[https://motls.blogspot.com/2014/09/brian-coxs-incompetence.html Brian Cox's incompetence]" by '''Luboš Motl''' on ''The Reference Frame'' blog }} |
||