Hi
I am trying to calculate the effect size from one of my group contrasts.
I tried to use fmripower but it doesn't seem to like my gfeat.
I have already extracted the mean parameter estimate for each group for the contrast in question and so have the mean and standard deviation of each group. However, if I put those values in an online effect size calculator (http://www.uccs.edu/~lbecker/), I get an effect size of r = .9. This seems very large given its only a small effect (only seen with a low threshold). But, are the mean parameter values from featquery the correct figures to use for this calculation, or are they 'biased' by the contrast they were set up to investigate?
Another option on this calculator is to input between groups t value and dof. I looked up the tstat in the contrast for the peak coordinate from the thresh_zstat image, and also the dof from that file. When I input that into the calculator I get an effect size of r = .44 - which seems more likely. Is this an appropriate way to calculate effect size?
Any advice would be most appreciated,
With thanks
Ella Hinton
|