I have done a 3 group comparison with age (mean centered), sex (two covariate variables 1 for female, 1 for male)  and site (two covariate variables 1 for site 1, 1 for site 2).  How do I interpret the REX chart vs. my chart of the mean import values by group.  These charts are slightly different.    

REX = the adjusted means (i.e., estimated regressor coefficients of second level model) with confidence intervals estimated at the zero level of covariates.  Is there an easy way to actually get these values out of REX?  

Average group values of output of import values = average connectivity between seed and cluster (averaged across all voxels in that cluster) for each subject by group - WITHOUT ANY EFFECT OF COVARIATE MODERATING THIS AVERAGE?  

Just wish to make sure I have this correct when presenting these finding. 

Thank you, 

Jennifer