News for hackers. Stuff that matters.

A comment on Noted study in psychology fails to replicate, crumbles with evidence of fraud. Parent comment

Re:Noted study in psychology fails to replicate, crumbles wi... by anchpop on Sunday August 22, 2021 @10:10AM (#28265598)
Hmm, I wonder if one could use differential privacy to make mathematically certain that no individual data point could leak. Although that may not be a strong enough guarantee
[ Parent | on HN ]
Re:Noted study in psychology fails to replicate, crumbles wi... by CrazyCatDog on Sunday August 22, 2021 @02:09PM (#28267915)
Yes, I’ve seen it on websites (eg GMAC) but the operators available are trivial (e.g. mean, median)… system will not process the request if the sample is too small (I think the minimum sample size is actually a function of how much you pay)
[ Parent | on HN ]