On bifactor model generation PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Gabriella Fu posted on Monday, November 19, 2018 - 5:27 am
Dear Professor,

I am using M+ to generate a bifactor model. Ideally, I want the general factor load to be 0.5, item1 to item5 load 0.4 only on the first specific factor f1, item8 to item10 load 0.4 only on the second factor f2, item6 and item7 have cross factor loading, both of them load 0.3 on f1 and 0.4 on f2.All factors are uncorrelated. I set the error variance according to 1 - (f1 square loading) - (f2 square loading). 100000 data sets are generated according to the above command.

However, when I tried the command in UG(exp5.29) using the generated data (rotation=bi-geomin), the model parameters are not the same as I set up. The general loadings were higher (over 0.6), all f2 loadings are smaller (around 0.1). I tried the command in exp5.30 and target rotation, giving a model command after model population, all of them showed the model cannot be computed. I waas wondering is there something wrong with the parameter I originally set up? Or the model just cannot repeat itself?

Many thanks,
Gabby
 Bengt O. Muthen posted on Monday, November 19, 2018 - 11:14 am
The problem with this setup has been answered by Mplus Support.
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: