Gustavo,
In your case 3dRegAna and 3dttest++ should render the same results if you specify the same model. With 3dttest++, you put all the three quantitative variables as three columns in a text file, and then use option -covariates in 3dttest++. You may also consider proper centering if the intercept is of interest. Check out the help of 3dttest++ for more details:
3dttest++ | less
or,
3dttest++ -h_view
A side note: You may label each variable the way you like (e.g., predictor vs. covariate), but all the variables are treated the same in the model regardless each of them is of interest to the investigator.
Gang
Edited 1 time(s). Last edit at 09/17/2014 03:27PM by Gang.