Re: st: xtmelogit cross-classified groups

["Johanna Bromakow" <Johanna.Bromakow@gmx.de>]

For better reading here text again (I hope it will work)

I m using xtmelogit to estimate a cross-classified random intercept model. The data represent the participation in adult education courses between 2007 and 2010. 
My subjects are participants of integration courses of an adult education centre in Berlin. 
I want to analyze their transition in non-integration courses in dependence of the behavior of other course members.

My problem is that subjects can participate in multiple courses at the same time und cross different integration courses over time 
(meaning that one integration course is followed by another und different kinds of integration courses are combined). 
As a consequence subjects are exposed to time-varying and crossing group influences.

I spent the last two days trying to find a solution for this problem, but did not succeed. 
One possibility of dealing with cross-classified groups seems to create dummies for each group (vgl. Hox) and treat groups as combinations of the dummy variables nested at a higher level.  
But this would not work for my data because of the high number of courses. Is there any possibility in Stata to deal with such complex issues? 
My idea is to reshape spells with multiple courses from wide to long and treat them as nested within ID and time. But I am not sure if this would be adequate …

My data are structured as following, with
k1, k2  … k6 … ID of integration course number 1, 2 …
tid … ID of participant
jahr … year
hj … semester/halfyear
time … time since first since first integration course (until 2010, halfyear-intervall)




-- 
Empfehlen Sie GMX DSL Ihren Freunden und Bekannten und wir
belohnen Sie mit bis zu 50,- Euro! https://freundschaftswerbung.gmx.de
*
*   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/