Imputed data PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Susan Scott posted on Monday, August 21, 2006 - 7:04 am
Hello Linda,

I am using imputed data, with 5 datasets. Mplus is reading all 5 but only using 3. Is this because the model is not converging in the other 2 datasets, or is there another explanation?

Thank you,
Susan Scott
 Linda K. Muthen posted on Monday, August 21, 2006 - 3:43 pm
This is probably the case. You can add TECH9 to the OUTPUT command to find out.
 Joonmo Son posted on Wednesday, May 05, 2010 - 6:58 am
Hi, Linda, I got a question.

I generated muliply-imputed data in Stata and want to conduct chi-square model fit test in Mplus using difftest. But I got this error message: "The DIFFTEST option is not available for TYPE=MONTECARLO or multiple imputation."

Of course, I used TYPE=IMPUTATION to read in the imputed data sets.

Could you tell me how I can test chi-square model fits between two different models when using TYPE=IMPUTATION?

Thanks in advance.

Joonmo
 Linda K. Muthen posted on Wednesday, May 05, 2010 - 9:51 am
I don't think it is clear how one would do difference testing with multiple imputation. You could do it for each data set but how to combine the information is not clear. This is a research question.
 Joonmo Son posted on Wednesday, May 05, 2010 - 8:53 pm
Thanks. It seems not possible according to your answer.

Then is it doable to impose model constraint (e.g., p1=0) to multiply-imputed data as a detour of chi-square difference test? I tried, but got some error messages too (FYI. I am using MLM estimator).
 Linda K. Muthen posted on Thursday, May 06, 2010 - 8:06 am
I don't think it is clear how to do difference testing for imputed data. If you need help with an error message, send the output and your license number to support@statmodel.com.
 Zhang Rui posted on Monday, September 09, 2019 - 12:14 am
Dear professor,
I am running a cross-lagged model in Mplus, which includes three waves in a year. The data (including sex and child membership) was imputed by multiple imputation. After controlling for age and sex, we examined whether child membership (1=yes, 0=no) moderated the model. However, I got a error. ¡° When using TYPE=IMPUTATION, the number of observations in each group in all data sets must be the same. Problem with group NO
Current data file: SEM1MISSING2.dat
Average number of observations in previous data files: 109
Number of observations in current data file: 110¡±.
When I added ¡°Nobs= 109¡±, the model still did not run and got another error ¡°When using TYPE=IMPUTATION, the number of observations in each group in all data sets must be the same.¡± I also used another method. When I combined the five MISSING1. DAT to Missing5.DAT into one DAT (using copy and paste), the results did not seem right. Is the combing method what I used correct? What can I do in the next step? Thank you very much.
 Bengt O. Muthen posted on Monday, September 09, 2019 - 5:46 am
We need to see your output to answer this - send your first problematic output and data to Support along with your license number.
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: