Hello!

The following technical issue:

i have a hierarchical model with two random factors (experiment is top level and subjects belong to one of two experiments as the bottom level factor) plus a continuous within-subject predictor x

The full syntax that causes the error is:
------
GLM depvar BY experiment subject WITH x
/RANDOM= experiment subject
/METHOD=SSTYPE(3)
/INTERCEPT=INCLUDE
/CRITERIA=ALPHA(.05)
/DESIGN=experiment subject(experiment) x x*subject(experiment) experiment*x experiment*subject(experiment) experiment*x*subject(experiment).
------

The error i get is "The factor experiment is specified more than once in the effect. It can be specified only once."

If i remove the two items below from the syntax, the syntax runs fine
experiment*subject(experiment)
experiment*x*subject(experiment)

But then i cannot assess teh interactions im interested in, namely whether experiment interacts with subject , and whether experiment interacts with x*subject

Is the problem that SPSS thinks im trying to do experiment*experiment when im actually trying to do experiment*subject(experiment)? how to get around that?

Many thanks!

Best Wishes,

Kristjan