| |
|
Science Forum Index » Statistics - Math Forum » Align Friedman ranks for univariate repeated measures
Page 1 of 1
|
| Author |
Message |
| Ollie |
 Posted: Sat Dec 30, 2006 3:03 am |
|
|
|
Guest
|
Hello,
I read this from an article from Beasley:
"Data from a split-plot design have three nuisance parameters that must be removed in order to align the scores for ranking and subsequent analysis of interaction effects. Specifially, the three nuisance parameters from model are the repeated measures main
Effect, the between-subjects main effect and subjects’ individual differences effect that is nested in the between-subjects factor."
The method to align ranks is then:
Y'ijk = [Yijk - meanY..k - meanYij. + grandmeanY...]
However, in my study I do not wish to account for between subjects variations, is the following way of aligning ranks sufficient then:
Y'jk = [Yjk - meanYj. - meanY.k + grandmeanY..] ????
Thanks for answering and best wishes!
Ollie |
|
|
| Back to top |
|
| |
|
Page 1 of 1
All times are GMT - 5 Hours
The time now is Thu Jul 24, 2008 4:18 am
|
|
