I'have fitted a mixed model in SPSS with fixed and random effects, and I would like to know what do the Fixed effects contrasts (type III) mean... In my model they are significant (e.g F=23.9; p=.012), but in the parameters estimates table the fixed variables are not significant... How should I interpret it? Thank you so much!!

  • $\begingroup$ They are testing different hypotheses, so the $p$-values do not have to agree. This issue has been discussed on this site on several occasions in various contexts (e.g. stats.stackexchange.com/questions/16665/…). $\endgroup$ – Macro Jun 1 '12 at 18:33
  • $\begingroup$ @Macro Don't you think your remark sounds a little harsh given that you are addressing a newcomer? $\endgroup$ – Michael R. Chernick Jun 2 '12 at 2:52
  • $\begingroup$ @MichaelChernick, I wasn't trying to be harsh. I was just trying to be helpful by pointing out that Mike may find searching previous questions helpful. $\endgroup$ – Macro Jun 2 '12 at 14:57
  • $\begingroup$ Thank you, Macro and MichaelChernick, for your comments. I didn't find it that harsh; I'm sure I should have checked better the previous posts related to my question. It's just I'm new in statistics and not sure I'm understanding well some concepts... Do you think it is correct if I interpret that the independent variable is not significant as predictor, but there is a significant difference in the de dependent variable values related to the independent variable? Is it possible? $\endgroup$ – Mike_999 Jun 2 '12 at 15:42
  • $\begingroup$ Sometimes I am sensitive to comments. Most people here are not sensitive and no one is intentionally harsh. $\endgroup$ – Michael R. Chernick Jun 2 '12 at 16:28

If you ask to print the Contrast coefficients matrix in the Statistics dialog, you can see the linear combinations of parameters tested by the various tests of fixed effects.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.