Is this family wise error?

#1
Hi all,

I would like to confirm if this is family wise error.

I have Parameter A, B, C and D and I want to test if each of their distribution (they are categorical) differs between Group 1 and 2. i.e. chi-square test (or fisher's exact test if they aren't normal) for Parameter A for Group 1 and 2, Parameter B for Group 1 and 2, Parameter C for Group 1 and 2 etc.

I usually know family wise error as multiple comparisons between groups. ie. Parameter A for group 1 and 2, Parameter A for Group 1 and 3, Parameter A for group 2 and 3.

Is it family wise error for my case?

Thanks for your help in advance!
 

Dragan

Super Moderator
#2
Yeah, it is. You could also think of Family-Wise error in the context of testing underlying assumptions e.g. normality, equal variances, etc. That is, testing these underlying assumptions increases the usual Type I error rate of 0.05 - for an omnibus F-test associated with a one-way ANOVA. Some people would even go as far as to say that even looking at a histogram of your data of concern increases Family-Wise error.