Drake Steed posted on Monday, April 21, 2014 - 12:26 pm
I have a question. I am trying to compute a chi-square difference test, comparing the loglikelihood chi-square values of algebraically constrained versus nonconstrained estimations. My understanding is that upon algebraically constraining one of the variables in the analysis, the degrees of freedom changes as compared with the nonconstrained analysis. However, in my case, the degrees of freedom are the same for both the constrained and unconstrained loglikelihood chi-squares. I tried the analysis using ML as well as MLR. Is this an aberrant finding or should I understand this outcome as typical? Thank you!