|
|
Hi all,
This is a common problem I know, but I just haven't been able to set up
user-defined contrasts in UNIANOVA appropriately. Hopefully someone here
can help me out.
The design has one fixed factor (island - 4 levels) with a random factor
(nest) nested within it. I'm trying to get a contrast that will compare
levels 1 & 2 of 'island' to levels 3 & 4. I've been trying various
LMATRIX settings but am obviously not structuring it in the right way as
I get the 'L Matrix is not estimable' error.
Here is the existing basic syntax:
UNIANOVA
eggweight BY Island nest
/RANDOM = nest
/DESIGN = Island nest(Island) .
Thanks for any pointers.
Cheers,
Kylie.
|
Locked
