Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Maarten buis <maartenbuis@yahoo.co.uk> |
To | statalist@hsphsun2.harvard.edu |
Subject | Re: st: Dirifit / ddirifit: "maximum number of iterations exceeded" |
Date | Sun, 25 Apr 2010 15:17:19 +0000 (GMT) |
--- Martin Gelter wrote: > > Depending on which control variables we use, we > > sometimes do not get results at all <snip> We > > suspect that one of the problems could be that > > we have many values in the matrix that are very > > close to zero. --- On Sun, 25/4/10, Mark Nichols wrote: > I'd suggest you try linear models (ols). Normally I am not so sceptical about using simple methods (even though I co-wrote -dirifit-), but in this case it is probably a bad idea. The reason is that linear models don't do well near boundaries. In most data that is not a big deal as most of the observations are usually far enough away from the boundary to sorta ignore that issue. However, Mark explained explicitly that this is not the case in his data. Anyhow, if Mark desides to use -regress- he should definately specify the -robust- option, as there is certainly going to be heteroskedasticity. Hope this helps, Maarten -------------------------- Maarten L. Buis Institut fuer Soziologie Universitaet Tuebingen Wilhelmstrasse 36 72074 Tuebingen Germany http://www.maartenbuis.nl -------------------------- * * 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/