Wed, 08/27/2014 - 18:37

Good day,

For my data set, I have to generate confidence bands for a number of parameters. I already did empirical bootstrapping via the function umxCI_boot from the package 'umx'. My data set contains a number of missing values (with FIML used for the model estimation) so when I did empirical bootstrapping, it stated a warning indicating the presence of missing values in the observed covariance matrix, making it non-positive-definite so the empirical bootstrapping did not proceed.

My question is: What would be the next sensible thing to do? Use parameteric Monte-Carlo bootstrapping instead based on the expected covariance matrix? With the observed covariance matrix containing missing values, doing parametric Monte-Carlo bootstrapping on it might be futile.

Thank you for your feedback.

metavid

Good day,

I will consider it given that it takes the missing values into account, instead of using the expected covariance matrix for the parametric bootstrapping. Thank you!

metavid

Is there a reason you can't use the mxCI() function? It computes confidence intervals based on the profile likelihood. That is, it varies the free parameters and monitors the change in likelihood to get the confidence intervals based on changing the likelihood by a certain amount. It handles missing data just fine because it's using FIML, the same as the estimation routine.