Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | jovogt@rumms.uni-mannheim.de |
To | statalist@hsphsun2.harvard.edu |
Subject | st: Fixed effects XTREG, Multiple Imputation MIM and Bootstraping VCE(BOOTSTRAPPING) |
Date | Sun, 16 May 2010 23:34:13 +0200 |
HiI have a panel data set (220 groups, 8 years) and I used the user written multiple imputation command ?ICE? to fill up missing observations. I am furthermore applying XTREG, FE to estimate a fixed effect model using the generated data. The XTREG command works well with ?MIM? (user written command to estimate models based on imputed data-sets).
(mim: xtreg depvar indep.vars, fe)I am however furthermore trying to estimate confidence intervals based on bootstrapping
(mim: xtreg depvar indep.vars, fe vce(bootsrap)This also doesn?t lead to any complications without any further modifications. It is however not possible to set a higher number of bootstrapping cycles:
Eg.: (mim: xtreg depvar indep.vars, fe vce(bootsrap). Reps(400) Error: ?invalid reps? Or to apply (estat bootstrap, all).May somebody assess whether applying vce(bootstrap) on multiple imputed data sets can lead to biases in general and/or does somebody know how to change the number of bootstrapping cycles / and/or can tell me how get the post-estimation bootstrapping information (to get asymptotically refined confidence intervals)?
Thanks & best regards! * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/