Latent Independent Variable, Gamma
Message/Author
 Jonathan Cottrell posted on Wednesday, October 02, 2013 - 1:58 pm
Good afternoon,

I am somewhat new to MPlus and need some help with SEM modeling.

I am trying to create a latent independent variable (x) with one indicator. I wish to use that to estimate relationships with other variables, which should be in the gamma matrix, not the beta matrix. However, using the BY statement seems to always turn in into a dependent variable. How do I make it a latent independent variable?

I can use just the indicator and have it go ON another variable to make it an independent variable. However, I will eventually want latent variables, some with multiple indicators, that act as mediators (hence acting both as an independent (gamma matrix) and dependent (beta matrix) variable).

Any help you can provide would be greatly appreciated.
 Linda K. Muthen posted on Wednesday, October 02, 2013 - 3:02 pm
Factor indicators are dependent variables. If you create a factor that is equal to an observed variable, the observed variable will be treated as a dependent variable. The only way for the observed variable to be treated as an observed exogenous independent variable is if you use the observed variable as a covariate. The matrix a parameter appears in is not important. Parameters can be placed in different matrices by different programs with no impact of the value the parameter is estimated at. You can have a model with a combination of factors and observed variables.
 Jonathan Cottrell posted on Thursday, October 03, 2013 - 10:10 am
I see. Thank you.

So if I attempt to use Linear Growth Modeling and regress one intercept onto another intercept, then both will be considered dependent variables, right?

Will doing this cause convergence problems for my model?
 Linda K. Muthen posted on Thursday, October 03, 2013 - 10:27 am
You will need to show what you mean by showing your input.
 Jonathan Cottrell posted on Thursday, October 03, 2013 - 2:18 pm
Never mind, I think I can manage it from here. Thank you for your time!
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: