|
|
Chi-Square Diff Test value higher tha... |
|
Message/Author |
|
Evan posted on Friday, August 03, 2018 - 10:24 am
|
|
|
I am doing a 1 degree of freedom Chi-Square Difference test between two models and I am getting a value that is higher than both of the comparison models separate values combined and seems much higher than expected. This is for categorical data and I am using MLR. Some other information about the model comparison. difftest = mydiff.dat MODEL= nocovariances param=theta Let me know if you need any other information to help diagnoses this. |
|
|
Difftest and Param=Theta go with WLSMV estimation, not MLR. For a 1-df test, you can simply square the printed Est/SE ratio to get an approximate chi-square value. |
|
Back to top |
|
|