Can someone explain me the rationale of MCT use in a situation where two groops are compared regarding a number of variables (using apropriate parametric and non-parametric methods)?
So there are several different tests performed, to test the difference of means (t), median(U) or frequenies (chi-square) between two indipendent sets of data.
Question is: Is this considered multiple comparison and which "protection" method to use and how?
I understand the omnibus-ANOVA and post hoc t-tests situation...
Last edited by Erzherzog; 04-05-2008 at 06:39 PM.
Advertise on Talk Stats