|
|
 |
| CFA bi-factor model and centering |
 
|
| Message/Author |
|
|
| Ina Prokjev posted on Tuesday, September 23, 2014 - 12:57 am
|
|
|
Dear Linda, dear Bengt,
I want to run a CFA for a bi-factor model with correlated latent factors. One latent factor (general achievement; GenAch) is indicated by school grades (v1-v6)and I want to center these indicators at the cluster mean. cluster is "class". And I want to take into account that there is a hierarchical multilevel structure of data (pupils in classes).
Is this how I can do it?
Variable:
[...]
cluster=class;
centering=groupmean (v1-v6);
Analysis:
type=complex;
model:
[...]
GenAch by v1-v6*;
thanks for your help! |
|
|
|
|
You have centered and you have taken clustering into account. I don't see a bi-factor model. And if you free all factor loadings as you have done,
GenAch by v1-v6*;
you must fix the factor variance to one.
GenAch@1; |
|
|
| Ina Prokjev posted on Tuesday, September 23, 2014 - 12:32 pm
|
|
|
thanks for your help. Yes, I know about fixing to one. I just send an excerpt of the syntax. the full model is specified as a bi-factor model. my main concern was to clarify the centering question.
thanks! |
|
|
| Back to top |
|
|
