Comparing models with binary DV PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Tobias Ludwig posted on Wednesday, April 19, 2017 - 9:45 am
Hello,
I regress a binary DV on four independent continuous latent variables. I a first step I calculated 4 single logistic regressions. Afterwards I estimated a multiple regression model where all four predictors explain the binary DV.

a) How can I compare models? I think that BIC is not appropriate because the number of parameters change between the single and multiple regression model, correct?
b) Can I select the models based on the LL values in this case? How is this done?
c) Is there a reference on the fact that Mplus does not report relative model fit for models with logistic regressions? I know why that happens, but I do want to make that clear to potential reviewers.

Thanks in advance!
 Bengt O. Muthen posted on Friday, April 21, 2017 - 6:03 pm
a) BICs are comparable because you have the same DV conditioned on different covariates.

b) Use BIC

c) No.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: