Hello all...trying to enter syntax to do multi-level modeling in SPSS and I'm getting annoyed. I need help please.

Level 1 variables (within cluster) Surface Acting

Level 2 variable (between cluster) Political Skill

DV = Surface Acting (ELSurfaceActing_gmc)

IV = Political Skill

Cluster = Person (number assigned to each subject)

This is syntax from a source that i have used previously with great success:

MIXED ELSurfaceActing_gmc WITH PolSkill
    /FIXED= PolSkill
    /RANDOM=INTERCEPT ELSurfaceActing_gmc | SUBJECT(Person) COVTYPE (UN)
    /METHOD=ML
    /PRINT = SOLUTION.

This is the error message...what am I doing wrong?

Thank you...I want to cry...this shouldn't be that difficult.

Robyn

Hello Robyn.  According to what you wrote, Surface Acting (SA, the DV) is the only Level 1 variable.  If Political Skill (PS) is purely a Level 2 variable, that means that everyone within each Political Skill cluster has the same score.  Is that really true?  Or does each person also have their own individual PS score?  And how many observations of the DV do you have for each person?

I think it would help forum members to understand your problem if you shared a small "example" dataset showing the structure of your data, and if you explained in words what question you are trying to address with your mixed model.  Thanks.

Bruce

PS- One reason you had a problem is that by including SA on that RANDOM sub-command, you asked SPSS to estimate the variance of the random slopes for the linear relationship between SA and SA.  And that makes no sense.