causal
Convergent Evidence Without Shared Bias
No single one of the four measures is a knockdown argument and each has flaws, but there is no reason they should all be biased in the same direction; if the hypothesis of declining scientific productivity were false we would expect contradictory data, yet all the analyses point the same way, which is what makes the conclusion worth taking seriously.
causalpending
Speaker
Patrick CollisonEvidence Quote
“I don't think any one category is itself knock down... But, I don't see a reason as to why they should all be biased in the same direction... Whereas, we kind of see all of these analyses pointing in the same direction. And so it's when you take all of them together--and they are all sort of pointing the same way--that is something I think we should really take seriously.”
Created: 6/13/2026, 12:29:12 AM
Your notes
Loading notes...