Contrasts using multinomial logistic PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
 Anonymous posted on Saturday, January 08, 2005 - 11:21 am

I wish to perform a multinomial logistic regression using Type = Complex. I created a 5 level MANIFEST and nominal DV (0-5) and have categorical and continuous manifest predictors. I know that the 0 level will automatically become my reference group and understand the code to obtain this analysis. I, however, wish to examine contrasts. For example, level (1 and 2 collapsed) vs. level (3 and 4 collapsed); level (3 and 4 collapsed) vs. level (0,1,and 2 collapsed). I was wondering how I would run these contrasts. Also, I know if you are using latent classes you can receive odds ratio estimates by specifying a tech option. Is there a tech option to receive OR for a manifest nominal dv analysis?

 bmuthen posted on Saturday, January 08, 2005 - 5:26 pm
These contrasts and ORs are not currently printed out. Something to put on our list. This can be computed from the output, but it takes effort.
 monica oxford posted on Thursday, May 18, 2006 - 4:10 pm
I ran a multinominal SEM model, I have a categorical outcome predicted by latent factors. How do I caculate the Odd's Ratio in this case.
 Linda K. Muthen posted on Friday, May 19, 2006 - 1:33 pm
I don't think the fact that the covariates are latent changes how the odds ratio is computed.
 Michael Lorber posted on Thursday, April 12, 2007 - 12:26 pm
Going back to the 1/8/05 post, how can ORs be computed from the output of multinomial regression? Are there instructions in the May, 2006 manual?
 Bengt O. Muthen posted on Friday, April 13, 2007 - 9:01 am
Yes, see chapter 13.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message