|
|
 |
| CFA Model Comparison with WLSMV Estim... |
 
|
| Message/Author |
|
|
|
|
Is there any way to compare non-nested CFA models statistically when a WLSMV estimator is used across models? I do not know of any statistical tests available, but thought I would check here just to be sure that is the case.
Do any rules of thumb (i.e., CFI change < .01 provides support for the more parsimonious model) that apply for nested models apply in this case that I could use?
In the following example, these two models would not be nested, correct (if X = loading and 0 = non-loading)?
Model 1:
X 0
X 0
X 0
X 0
0 X
0 X
0 X
0 X
Model 2:
X 0 0
X 0 0
X 0 0
X 0 0
0 X 0
0 X 0
0 0 X
0 0 X
Thank you very much! |
|
|
|
|
Q1: No
Q2: I am not enamored with such an approach.
Q3: Right
Just run it with ML and use BIC to decide on the model. |
|
|
| Back to top |
|
|
