Hello David,
The columns of zero's can be safely removed from the design - they don't really add anything useful to the model. Also their presence could potentially conceal other rank-deficiency problems ( should there be any in the design).
Many Regards
Matthew
> HI, this is a simple question but I need to be sure of what is best. can you please advise?
>
> in my design matrix i've included regressors of no interest such as memory misses, memory errors, attention misses and attention errors, some subjects though have no errors so I created several 3 column txt EVs with 0, 0 ,0 for those subjects.
> Obviously those nuisance regressors are not included in the contrasts
>
> the result is that for quite a few subjects there are many of those 0,0,0 EVs ..
>
> attaches is a pic of the design matrix in one extreme case, surely this is causing the rank deffiency warning
>
> is it ok to proceed like this or should I alter the design?
>
> Thanks
> <dm.jpg>
|