Hello everyone,
Is there a way to define a missing value in the design.mat file for
randomise (e.g. "NaN")? I am working on a large data-base and it would be
nice not to need a different 4D image file for each test I am doing. Also,
it would be possible to analyise several independent variables in the same
model, even if the data are incomplete for one of them.
If I cannot put in the "NaN"-like coding in the variable column itself,
maybe there is another way of doing it? Like inserting in an extra binary
column in which e.g. all cases with 0 in this column are excluded from the
analyis? Or even by selecting only certian images from the 4D file when
calling randomise?
Thanks,
Emma
p.s. I know someone in 2003 said that this kind of missing value solution
was on the todo-list but I cannot find anything about it since then.
|