MC Analysis starting values <-> 95% c... PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
 Julia Mang posted on Monday, May 20, 2019 - 4:18 am

I am running different external Monte Carlo analyses with having starting values for the hierachical models.

As I want to report also the 95% coverage of the estimates, I recognized that this is dependent from the starting values as they were assumed as population values.

Am I right, that I have to choose the starting value with the population values in order to get right coverage values? At the moment I don't directly use the population values for the start values, but values slightly below them, so that they also cover the lower possible border.

Thank you for your support and kind regards,

 Bengt O. Muthen posted on Monday, May 20, 2019 - 5:39 pm
The values you give in the Model command will be used for coverage.
 Julia Mang posted on Tuesday, May 21, 2019 - 1:01 am
Thank you for your fast reply!

Just to get sure: Should I always use the population values as starting values in the Model command?

In my understanding, having lower starting points achieves also estimates that are below the population values.

Kind regards,

 Bengt O. Muthen posted on Tuesday, May 21, 2019 - 5:36 pm
If the analysis model is the same as the model that generated the data, you would use the Model Population values. I just copy them into the Model command. But sometimes the analysis model is different and you may want the basis for the coverage report to be some other values than the data-generating pop values - then you use different starting values in the Model command.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message