Hi, I followed your conversation, because I have a similar analysis (patients and controls, one variable with values only in patients). I run a linear regression with only 1 EV, demeaned within patients and zeros for the controls. I didn't use the -D option of randomise; now I re-run the analysis with the -D option, but the results is exactly the same. Also, the result is the same if I do not demean at all. I guess, this is due to the fact, that I only have values for patients and not for controls? And I also would like to ask, how I can test for a quadratic effect, as I have read this in a paper - just square my EV? Many thanks for all your help! Best wishes, Viola