Tech 4 output PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Richard E. Zinbarg posted on Tuesday, March 17, 2009 - 11:00 am
Hi Linda,
We have just noticed something that puzzles us a bit about the Tech 4 output for one of our models. It is a complicated model with a bifactor structure for one set of indicators (a general factor plus several group factors orthogonal to each other and the general factor) and a fairly simple unidimensional model for a second set of indicators (with a couple of correlated errors). Let's call the first set of indicators X and the second set Y. We want the correlations of each of the Y indicators with the X latent variables and whereas Tech 4 does show us the correlations of each of the X indicators with both the X and the Y latent variables, it doesn't show us the correlations of the Y indicators with the X latent variables. Any thoughts you might have would be most appreciated.
 Thuy Nguyen posted on Tuesday, March 17, 2009 - 1:29 pm
I can't get my head around this without seeing exactly what you are doing. Please send the full output and your license number to support@statmodel.com.
 Benjamin Miller posted on Wednesday, July 03, 2013 - 11:27 am
Hello,

I recently updated to Mplus version 7.11. The website indicates that p-values and standard errors for estimates of latent variable covariances are available with the Tech4 command in this new version. However, when I specified Tech4 in the Output section, all I saw were matrices of latent variable means, covariances, and correlations, no p-values or standard errors.

Example: "output: tech4;"

Is there something else I need to do to obtain p-values and standard errors? Does this option not work under certain conditions?

Thanks in advance for your help.
 Bengt O. Muthen posted on Wednesday, July 03, 2013 - 11:34 am
There are some conditions for which the TECH4 SEs are not yet implemented.
 Benjamin Miller posted on Wednesday, July 03, 2013 - 1:19 pm
Hello,

Thank you.

In that case, I just want to confirm that it will be appropriate to obtain zero-order correlations between latent variables by using the "on" command, modeling a one predictor regression equation?
 Bengt O. Muthen posted on Thursday, July 04, 2013 - 1:42 pm
If those are the only 2 latent variables in the model and their variances are 1.
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: