|
|
Chi-square LISREL vs. Mplus |
|
Message/Author |
|
Eric Teman posted on Friday, April 25, 2014 - 5:10 pm
|
|
|
I am doing an observed path analysis in Mplus and LISREL. I am trying to reproduce the independence null (baseline) chi-square that LISREL reports. Please tell me if this is the right way to have Mplus do it for the following model: AM ON SES@0; GPA ON AM@0 IQ@0; SES; IQ; SES WITH IQ@0; This allows estimation of all variances of DVs and IV's and fixes the covariance of IV's to 0. However, I am unable to reproduce this in LISREL. I am using the following model commands: MO NY=2 NX=2 PH=DI,FR PS=DI,FR GA=FU,FI BE=FU,FI Shouldn't the chi-squares match up? |
|
|
Make sure that you have have the same number of parameters in each analysis. Mplus may correlate the residuals of am and gpa as the default. |
|
Back to top |
|
|