| Message/Author |
|
|
|
|
Hello Dr. Muthen,
I wonder if I bootstrap a complex survey data set (with sampling weights)in Mplus, will weights be generated automatically while each time I am bootstrapping them?
Thank you.
Best wishes,
Bing |
|
|
|
|
That is correct - see section 2.1
http://www.statmodel.com/download/Resampling_Methods5.pdf |
|
|
|
|
Thank you, Dr. Asparouhov.
But what if I use type=twolevel instead of type=complex in the analysis command? Do the same commands in the analysis part work as well? Will I still get replicate weights if I use the same commands as you provided in the paper?
ANALYSIS: TYPE=TWOLEVEL;
RESPSE=BOOTSTRAP;
BOOTSTRAP=500;
Thank you.
Best wishes,
Bing |
|
|
|
|
Dear Dr. Asparouhov,
What is the difference between type=complex and type=twolevel? My understanding is that we may use different scaling methods in type=twolevel, but not in type=complex. Am I right?
And I tried bootstrapping with type=twolevel. It says that REPSE option and BOOTSTRAP are not allowed with Type=twolevel.
Best,
Bing Tong |
|
|
|
|
Bootstrap is not implemented for twolevel at this point. The differences between complex and twolevel is described in
https://www.statmodel.com/download/SMMuthenSatorra1995.pdf
Weight scaling for type=complex does not effect the results. For type=two-level you can find scaling information here
http://statmodel.com/download/Scaling3.pdf
http://statmodel.com/download/asparouhovgmms.pdf |
|
|
|
|
Thank you very much. They are very helpful.
Best wishes,
Bing |
|
|
| Back to top |
