EFA with twin data PreviousNext
Mplus Discussion > Exploratory Factor Analysis >
Message/Author
 Abdel posted on Monday, May 19, 2008 - 10:48 am
Since it is not possible to use the cluster option with EFA in Mplus 4.1, is there another option to analyse dependent data, like twin data in my case?

Thanks in advance!
 Abdel posted on Monday, May 19, 2008 - 10:52 am
The data is dichotomous by the way..
 Linda K. Muthen posted on Monday, May 19, 2008 - 2:11 pm
Mplus Version 5 introduced TYPE=COMPLEX EFA.
 Michael Baum posted on Sunday, December 21, 2008 - 3:16 am
I'm looking for a worked example of EFA for longitudinal twin data (multiple ordinal indicators). I have failed to locate such an example (likely a result of my relative inexperience with the website/software) and was hoping you could point me in the right direction.

Thanks in advance for your help.
 Linda K. Muthen posted on Sunday, December 21, 2008 - 12:52 pm
The MODEL commands for multiple indicator growth for CFA for categorical outcomes are shown in the user's guide on pages 546-547. Multiple indicator growth is also described in short course Topic 4 beginning with Slide 77. For EFA, you can use the new MODEL command language described in the Version 5.1 Examples and Language Addendums which are on the website with the user's guide.
 Michael Baum posted on Monday, December 22, 2008 - 7:49 pm
Thank you for point me to these materials. I was able to conduct the EFA. Now when I try to move on to a CFA framework, I get the following error message:

Unknown variable(s) in a BY statement: (1)-


I'm following the example in Topic 4 of your slides and although I have been able to fit a model where loadings vary by time, I would like to follow that with a model where loadings are invariant across time. However, when I attempt to follow the example I get the error message above. (Data consists of three time points, 2 observations per cluster and 7 ordinal/5-point indicators). Below are my analysis and model statements:

ANALYSIS:
type = complex ;
ESTIMATOR IS WLSMV;
rotation = geomin;

MODEL: f1to7_3 BY fq3_1 fq3_2-fq3_7 (1)-(6);
f1to7_7 BY fq7_1 fq7_2-fq7_7 (1)-(6);
f1to7_14 BY fq14_1 fq14_2-fq14_7 (1)-(6);

! below allows factors to be correlated over time;
f1to7_3 with f1to7_7 f1to7_14;
f1to7_7 with f1to7_14;
 Linda K. Muthen posted on Monday, December 22, 2008 - 8:20 pm
The specification should be (1-6).
 Michael Baum posted on Monday, December 22, 2008 - 9:32 pm
my mistake and apologies for not trying that, but is there a reason why slide 86 of Topic4 course handout has the following form (or was it a typo/change in command syntax)?

MODEL: f12a BY bru12
fig12-yot12 (1)-(6);
f22a BY bru22
fig22-yot22 (1)-(6);
f32a BY bru32
fig32-yot32 (1)-(6);
f42a BY bru42
fig42-yot42 (1)-(6);
 Linda K. Muthen posted on Monday, December 22, 2008 - 10:48 pm
Somebody altered the slide incorrectly.
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: