CodingInLondon on Nostr: I found out about this while learning about statistics 11y ago as part of a finance ...
I found out about this while learning about statistics 11y ago as part of a finance certification.
The 5% of surveys with wrong results among a group of surveys with a 95% confidence level are more likely to be published than the other 95%. That’s because among those 95%, you find mostly ‘inconclusive’ stuff which is not interesting, specially not to the people who financed those surveys.
Who cares if a survey is wrong when the result is exciting and has a 95% confidence level?
The 5% of surveys with wrong results among a group of surveys with a 95% confidence level are more likely to be published than the other 95%. That’s because among those 95%, you find mostly ‘inconclusive’ stuff which is not interesting, specially not to the people who financed those surveys.
Who cares if a survey is wrong when the result is exciting and has a 95% confidence level?