I worked in my study with six judges, that evaluated subjects concerning their pronunciation of 10 arabic words. As it turns out two of them rated significantly different depending on the day (I met with them 3 times). Here is my ANOVA and the Scheffe Test:
https://www.dropbox.com/s/5rlja8zj3i45r ... .docx?dl=0
I'm probably gona narrow it down to the first three "Beurteiler A-C = judge A-C" since my reliability analyses tell me so. So my question ist:
Can I standardise judge A and B somehow without loosing their meaningful values of means. I can only thing of a z-transformation in a data file that is differentiated by the 3 days (="Termin").
But then I only have these meaningless Z-Values....
How would you solve that problem? And also does that even work with the Z-transformation?
Really hope you guys can help me
(I'm an Austrian student, so sorry for any language confusion)