You need 16 times the sample size to estimate an interaction than a main effect

Comments

from Hacker News http://ift.tt/2tMlmsO
via

Comments

Popular posts from this blog