Bootstrap and SE of STd coefficients PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Jean Frisou posted on Friday, April 04, 2008 - 3:25 am
How can I obtain standard errors of standardized coefficients,with bootstrap option ? Is this possible with Mplus V5 ?
 Linda K. Muthen posted on Friday, April 04, 2008 - 6:26 am
If you use both the BOOTSTRAP and STANDARDIZED options, you will obtain standard errors of the standardized coefficients.
 Jean Frisou posted on Friday, April 04, 2008 - 7:00 am
Thank you very much Linda.
 Simon O. F. posted on Thursday, December 03, 2009 - 7:21 am
Hi,

I'm trying to test a multigroup model with bootstrap. I want to have the standardized solution but I only have the estimates when I use the STANDARDIZED option in the OUTPUT section. Is there something to do to have the std error and the p-value for the standardized estimates ?

Thanks in advance,
 Linda K. Muthen posted on Thursday, December 03, 2009 - 7:30 am
If you are using a version prior to Version 5, standard errors for standardized solutions were not available. If you are using a more recent version, please send the full output and your license number to support@statmodel.com so I can see the exact situation.
 Maren Winkler posted on Monday, January 25, 2010 - 3:18 am
Hi,

I'm using Mplus 5.21.
In my SEM, I've included the following:

... BOOTSTRAP = 1000 (RESIDUAL);
...

OUTPUT:
SAMP STANDARDIZED STD STDYX TECH1 TECH9 CINTERVAL(BOOTSTRAP)RESIDUAL;

I do get S.E. for my estimates and standardized estimates (StdYX Estimate and Std Estimate). However, I don't have S.E. for these standardized estimates.
Am I doing something wrong in specifying what I want Mplus to do? It's the 1st time I use bootstraping.
Or is it not possible to get S.E. for these estimates?

Thanks!
 Linda K. Muthen posted on Monday, January 25, 2010 - 10:26 am
Mplus does not provide standard errors in this case.
 Katrin Lintorf posted on Saturday, October 09, 2010 - 11:56 pm
Hi,

we are conducting a multi-group path analysis with mediation. Because one of our groups has a small sample size (<50), we wanted to bootstrap for the standard errors of the path coefficients as well as for the standard errors of the indirect effects. In the Analysis command we specified "bootstrap=1000" and in the output "stand cinterval(bcbootstrap)" to get standardized parameter results as well as confidence intervals. We are now have the following questions about the output:

1. Under "model results" we received unstandardized parameter results along with standard errors and p values. Are these standard errors and the corresponding p values bootstrapped?

2. For our standardized parameter results, we only received the coefficients and no standard errors or p values. Is there a way we can obtain the standard errors and p values for the standardized coefficients (perhaps with V6)?

3. At the beginning of the output, there are standardized indirect effects with p values given for each group separately, but at the end of the output, there is only one set of standardized indirect effects with confidence intervals given that is not labeled as belonging to either group. Which group do these last indirect effects come from?

Thank you for your help.
 Linda K. Muthen posted on Monday, October 11, 2010 - 11:25 am
1. Yes.
2. No. There is no option for this.
3. Please send the full output and your license number to support@statmodel.com so I can see what you refer to.
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: