Single-indicator latent variable for ... PreviousNext
Mplus Discussion > Categorical Data Modeling >
Message/Author
 Page posted on Tuesday, May 10, 2011 - 9:57 pm
Hi,

I have multiple-indicator latent variables derived from a survey of N=1500 (where each continuous latent var has 3 or 4 ordinal indicators which are 7pt likert scale items). I need to convert them into single-indicator latent variables since drastically cutting down on the survey length is a requirement to use it in our next study. I know this can be done for continuous indicators, but how can it be done for ordinal? Namely, the following doesn't seem to completely apply because it requires variance of the indicator and we have ordinal indicators:

f BY x@1;
x@a;

where a = (1– reliability) * sample variance

Thanks for any advice!
 Linda K. Muthen posted on Wednesday, May 11, 2011 - 9:24 am
With categorical outcomes, residual variances are not parameters in the model. I would not recommend trying to correct these variables using reliability.
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: