www.cmm.bristol.ac.uk/forum

Hello,
I am fitting a fixed factor model and the variance at level 2 is negative. Would any of you know what this means, and how I compute an ICC ? Shall I use the absolute value when computing it ? If so, this would underestimate the total variance : varL2negvalue / varL2neg_value + VarL1
Thank you in advance for your help !

With best wishes,
Eric

Hi Eric,

Could you post details of your model?
You should not get a negative variance, so that means something has gone wrong somewhere.
Have you checked if your data are correctly sorted.

Best,

Camster