|
|
 |
| EFA using imputation with categorical... |
 
|
| Message/Author |
|
|
|
|
I am trying to run an EFA using the WLSMV estimator because my data are categorical (0-7). I know that you cannot use FIML with WLSMV, so I am trying to use MI. However, when I run the syntax, I receive the following error message:
TYPE=EFA cannot be used in conjunction with TYPE=IMPUTATION.
Note that TYPE=IMPUTATION is available with the use of EFA factors (ESEM)
declared in the MODEL command.
Can you use MI with EFA? If so, how? |
|
|
|
|
Yes. For example you can run an EFA with 2 factors and 6 indicators using a code like this
data: file = imputeEx76list.dat;
type = imputation;
model: f1-f2 by y1-y6 (*1);
Alternatively you can use the Bayes estimator (FIML based) directly to run the EFA model. See Section 3
http://statmodel2.com/download/HighDimension.pdf |
|
|
| Back to top |
|
|
