On Fri, Apr 12, 2013 at 5:44 AM, Geomina Turlea <geomina@yahoo.fr> wrote:
> I know, but - mi impute- does not support glm either
Best advice I have is to logit transform the proportions and then use
PMM, or just use PMM directly. You probably should try several
different sets of assumptions about the imputation model, though. Nick
is right, MAR is a stronger assumption than many people realize and
harder to make plausible than one would suppose. Outliers are also a
problem with MI approaches. MI spreads the effect all throughout the
dataset.
*
* 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/
