search

dmzuckerman / Sampling-Uncertainty

Best Practices article intended for LiveCoMS
36 stars 5 forks source link

N vs N-1 #38

Closed dwsideriusNIST closed 6 years ago

dwsideriusNIST commented 6 years ago

From a reviewer: For eqn 8, the authors may wish to discuss when n and n-1 is needed. I find this is a source of confusion with students.

mangiapasta commented 6 years ago

Okay, I dealt with this problem in a footnote. This is a frustrating comment by the reviewer because it immediately opens up a huge can of worms. The stuff about Jensen's inequality is not widely acknowledged, but if we're going to talk about why we have n versus n-1, I think we gotta go there.

dwsideriusNIST commented 6 years ago

I made a very slight text revision to your footnote, 1) to use our macro for s(x), 2) "an" instead of "and", and 3) clarifying that the true standard deviation of the mean is being underestimated.

mangiapasta commented 6 years ago

Looks good. Thanks

dwsideriusNIST commented 6 years ago

Closed.