Hi,
How can one report the effect size in randomize results on TBSS? I mean the
statistical effect size, e.g. Cohen's d. I would like to see whether the
effect size of a simple 2-group comparison is larger in some regions than
in the others.
I find it difficult to follow the instructions from 26 Jan 2007 on the size
of the effect "before_ normalising" by any measure of noise:
>>>
>>> randomise does not output cope images, just the raw tstat and various
>>> forms of p-values. To get the PE or COPE you would need to use
>>> film_gls - to get the same results as randomise, use OLS, ie the -
>>> noest option.
Could you give some more details about the procedure? (in case stat effect
size doesn't work)
Thank you,
Best,
Aga
|