Hello Dr Anderson (& TOM),
I had a doubt regarding use of randomize after dual regression. Generally for randomize permutations are 5K
If I have two groups having 10 samples in each group and also If I don't want to see the effects of nuisance variables but need results to be corrected for the variables and by looking at the table for counting permutations in Randomise/Theory fslwiki page. This test lies under two sample t test, the number of permutations required would be (n1+n2)!/n1!Xn2! this comes equal to 184756
I just want to confirm that 5k permutations will fulfill the requirement or we need to modify the number of permutations?
If yes then how much? and please if you can provide the reason also.
Himanshu Joshi
NIMHANS, Bangalore
|