Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | "JVerkuilen (Gmail)" <jvverkuilen@gmail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | Re: st: Three level mixed random and crossed effects model |
Date | Thu, 13 Sep 2012 09:00:31 -0400 |
On Wed, Sep 12, 2012 at 10:59 PM, Mrs Mohana Rajmokan < mohana.rajmokan@uqconnect.edu.au> wrote: > Dear List, > > I am trying to fit a 3 level random and crossed effects model to a > continuous outcome to estimate the respective variance components of each > level of the 3 (for doctors) > > <snip> At first blush nothing appears wrong with your specification, though I thought you needed to use the _all keyword in the specification of a crossed random effect. I've not fit a level 3 model with crossed random effect though. General advice whenever fitting a complex model is to fit simpler ones nearby even if you don't really want to interpret them. So I'd try ignoring the third level to see how things go. In short, simplify the model to see where things blow up when you add another level. You should also consider turning difficult on (maximization options). You don't say what your Ns at different levels are. If there aren't many doctors that could create problems, but in that case you might well be better off simply putting in dummies for them, even though it's not the model you truly wanted to fit. Jay -- JVVerkuilen, PhD jvverkuilen@gmail.com "Out beyond ideas of wrong-doing and right-doing there is a field. I'll meet you there. When the soul lies down in that grass the world is too full to talk about." ---Rumi * * 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/