3 ms·
We can only wish that is true, we are all just humans in the end: http://www.economist.com/news/briefing/21588057-scientists-think-science-self-correcting-alar
by Balgair 13y ago
We can only wish that is true, we are all just humans in the end:
http://www.economist.com/news/briefing/21588057-scientists-think-science-self-correcting-alarming-degree-it-not-trouble http://www.economist.com/news/briefing/21588057-scientists-t...
Also, I vaguely remember a study in 'some journal' that looked at all it's P values[0] and found an alarming number of them parked right at the limit of acceptance (.05), more so than chance could assume. If anyone remembers this study and can provide actual proof (not my terrible memory), I would be very thankful.
[0]https://en.wikipedia.org/wiki/P-value https://en.wikipedia.org/wiki/P-value
- betenoire 13y agoI don't think he's commenting on statistical analysis, but rather saying that a scientist should follow the clues where ever it may take them, instead of looking for the clues that will get them where they want to be. And of course they fuck up, too.
- judk 13y agohttp://en.m.wikipedia.org/wiki/John_P._A._Ioannidis http://en.m.wikipedia.org/wiki/John_P._A._Ioannidis "Most published findings are probably false."
- lomnakkus 13y agoI can't claim to understand much of the statistics, but according to [1], p-values are pretty sketchy as a measure of "is this likely to be true?" [1] https://www.youtube.com/watch?v=ez4DgdurRPg https://www.youtube.com/watch?v=ez4DgdurRPg