Monday, November 26, 2018

Tell me the story in means, deviations and effect sizes.

Effect size, and the unreporting of them, is a longstanding plaint of mine. Stephen Vaisey points out another aspect which is also equally valid - the relative significance in differences between group effects. Or more specifically, the relative insignificance of differences at the mean.



Heh. The comment string is excellent, including this equally pertinent graph:



Heh, heh.

Means, standard deviations, effect sizes - tell me those things and I begin to understand.

Outage about group differences almost always ends up being outrage at what is happening in the tails where the instances are dramatically fewer anyway. You can make much ado about the uniqueness of a few things at the end of the tails or you can celebrate the great commonality in the middle. Your response depends on you and not the data per se.

No comments:

Post a Comment