help > Effect size in seed-to-voxel multiple regression analysis?
Showing 1-1 of 1 posts
Results per page:
Sep 15, 2019  05:09 AM | Yana Panikratova
Effect size in seed-to-voxel multiple regression analysis?
Dear CONN experts,

I carried out a multiple regression analysis to define effects of a behavioral covariate on seed-to-voxel functional connectivity of a seed. Between-subjects contrast was:

All Subjects, behavioral covariate, other covariates
[0 1 0]

My question is how to define effect size of this effect, which would be interpretable in terms of "low" or "great effect size". When I open effect sizes (Fisher-transformed z-values) in CONN, it is 0.08. But what does it mean? Is it a correlation coefficient between connectivity values and the covariate? Is this effect large, medium or low?

To clarify, I performed post hoc ROI-to-ROI analysis to derive connectivity values between the predefined seed and result cluster; the average connectivity value was -0.05. When I calculate Pearson r between the connectivity values and the behavioral covariate, the r is 0.37. So what is the correct measure of the effect size in this case?

Would be grateful for any help,