Multilevel modeling

[Patrícia Costa]

I just realized i send an email message to all of the group members with my question. I am sorry, I meant for it to be posted on the discussion forum :s. Anyway, here it is. Again, thank you for your inputs.

Dear fellow spss users,

I am conducting a multilevel analysis using spss and I get a result that I am unable to explain theoretically and that I don’t really know whether rends the analysis undoable. 
The thing is: with my first, intercept only model (random intercept), the intercept is significant (p=.000). However, when I add (fixed) predictors (even if just 1), the intercept becomes non significant. The predictors are both significant and the model improves (considering the -2LL).

Can anyone help me understand what this means and if I can use the results of this analysis?

Thank you so much in advance!
Patrícia

[Author]

Posts