Hi Gwenaelle,
Thanks so much. I'll change the number of contrasts to 2 in the design.con and rerun it. I initially used the design_ttest script to generate the design mat/con files and I modeled the mean with the -m option. I then edited them to include the covariates. I demeaned all the covariates before inclusion in the mat file. Am I correct in saying the effect of initially modeling the mean will still be applied and the effect of the covariates are removed and hence the corrected tcfe stats files 3 and 4 (1 and 2 are for the means) are the files I need to use, as would be the case if I hadn't included the covariates.
Thanks again and very much appreciated.
Kind regards
Erik
|