Categorical Classification of Indepen... PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Robert Morris posted on Sunday, November 12, 2006 - 4:42 pm
Dear Linda & Bengt,

I've been battling with a modeling issue for some time now and have read through many posts from your wonderful forum but am still a bit confused. I am relatively new to latent variable modeling and could use some help.

I have a model (n>600) with two continuous dependent latent constructs (X3 and x4) and two independent ordinal latent factors (x1 and x2 ; 7 indicators each w/5 response categories each). All variables are somewhat non-normally distributed. I am controlling for 3 observed variables.

x1 is exogenous and leads to all three other factors;

x2 leads to both x3 and x4;

and x3 leads to x4.

Again, x3 and x4 are metric, but x1 and x2 are categorical.

1. Since x1 and x2 are IVs, at least a mediating dv through x1, should I classify them both as categorical? I've read in the forum that important issue is the DVs, not nessesarily how the IVs are classifed if they are ordered.

2. Or, does making an IV endogenous, if only as a mediator, turn it ino a DV which should subsequently be classfied as categorical (this is what is making me confused)?

I'm getting mixed opinions from colleuges.

I would greatly appriciate your help.

Thanks!
 Linda K. Muthen posted on Sunday, November 12, 2006 - 5:17 pm
I think I know what you are asking. Let me know if this does not answer your question. Only endogenous variables should appear on the CATEGORICAL list. Exogenous variables should not. The scale of exogenous variables does not affect model estimation.
 Robert Morris posted on Sunday, November 12, 2006 - 7:19 pm
Thanks, I appreciate your help and your very quick reply.

Take care.
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: