Monte Carlo -style analyses query PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
Message/Author
 Chris Stride posted on Wednesday, September 21, 2005 - 7:09 am
I'm looking to run a series of analyses as follows.

I have a very large data set. I want to take 1000 random samples from it and run the same CFA on each one, saving the fit indices to an ASCII file.

I have MPlus 2 at present and I know this would be possible in a very roundabout and timeconsuming way by first saving 1000 different samples via another package and then using the RUNALL facility.

But is there an easier way in subsequent versions of MPlus, where the package will take the random sample itself (some variant of the NOBSERVATIONS command maybe?), fit the model, save the required indices, then loop back to the start, take another random sample from the data set, etc, etc.

cheers
Chris
 Linda K. Muthen posted on Wednesday, September 21, 2005 - 7:29 am
Mplus does not take random samples of the analysis data. If you create the 1000 data sets, you can use the external Monte Carlo facility of Version 3 to run all of the analyses. It summarizes your results as in a Monte Carlo study. You can also save all of the results using the SAVEDATA command.
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: