Hi Onur,


This is a problem that has arisen before:
https://www.nitrc.org/forum/forum.php?th...

My understanding is that if you want your group to have an impact in the independent column you are doing an independent test with you will have to specifically model that (see Martin response to lara in the above link). I believe that if you are testing just interaction test you will test the independent variable relationships with the raw 3d coordinates, and yes, ignore the group variable. You will have to model it as suggested by Martin above, if you want to test interactions with group and independent column.


Best,
Beatriz