chuck weeks posted on Tuesday, March 20, 2012 - 7:52 pm
I'm very new to Mplus and the SEM world as a whole. I'm hoping the community could provide some references and direction.
I'm interested forming a path model using composite indicators, but I can't find any information in the Mplus Users Guide regarding the syntax needed to eliminate the disturbance term in the composite variable.
Have a look at our slides in the Topic 1 handout on our website. Towards the end you will find how to specify formative/composite indicators. You fix the residual variance to zero (f@0) and fix one slope to one (f ON x1@1).
luke fryer posted on Sunday, October 18, 2015 - 7:50 pm
Assistance requested with persistent error.
I am working with a model employing composite variables.
Regardless of how I structure the model, I get this error every time.
THE MODEL ESTIMATION TERMINATED NORMALLY
THE CHI-SQUARE COULD NOT BE COMPUTED. THE CORRECTION FACTOR IS NEGATIVE.
I cannot find anything in the guide book or online to suggest how I might address this problem.