Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | "Pesola, Francesca" <francesca.pesola@kcl.ac.uk> |
To | "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu> |
Subject | st: xtmixed and vce(robust) |
Date | Fri, 4 Apr 2014 15:24:40 +0000 |
Hi there, I am using xtmixed in Stata 12 to analyse trial data where I am comparing follow-up scores for 2 intervention groups adjusting for baseline scores and accounting for clustering. Some of the outcome variables are skewed but the residuals are normal . The ladder command suggests that no transformation is able to improve the distribution for these measures SO I decided to use the vce(robust) option. Is this a valid approach? Using robust standard errors has meant I now found an effect of intervention which was not there when not using this option. Grateful for any insight on why this might be the case and what the best approach is - F * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/faqs/resources/statalist-faq/ * http://www.ats.ucla.edu/stat/stata/