Dear FSL experts,
I am doing a two-sample t-test group comparison using 20 patients and 60
controls, and would like to use randomise to do permutation test as to
derived corrected p value statistical maps.
I read on the FSL official website for randomise, it stated " ..... For
large sample sizes, with 50-100 or more degrees of freedom, the P-values
should be highly accurate. When the sample size is low and there are many
nuisance regressors, accuracy could be a problem."
My group size is large (N=80), however, it consists of relatively small
number (N=20) of patient and a relatively large number (N=60) of control.
Will the discrepancy in group size lead to serious statistical problems?
Will that be a better idea if I only pick those controls who are best match
to patients, to reduce the groups size differences and obviate the need to
use nuisance covariates.
Thank you in advance for your help.
Best,
Yi-Shin
|