Hello,
I need some help with the following data analysis. I have a dataset of 335
horses trained by 8 trainers, with daily exercise observations for each
horse (min 45, max 716 observations per horse). The aim of the analysis is
to identify exercise variables that are related to the risk of injury (56
injuries in 52 horses). I want to perform Poisson regression analysis, using
-xtpoisson- with trainer as a random effect.
The command
tsset trainer date
tells me there are 'repeated time values within panel', which is fair enough
(multiple horses per trainer).
However, when ignoring this error message, Stata does fit a random effect
Poisson regression model that seems perfectly valid - is it?? Or should I be
using another approach?
I have looked into -gllamm- but this command tells me there are
"insufficient observations"
Using the -cluster- option with -poisson- I sometimes run into trouble when
the number of parameters in the model is at least as large as the number of
clusters (8). Furthermore, I understand that the 'robust' approach is not
optimal when you have a small number of clusters with many observations per
cluster (in my case, 6099 to 42312 observations per trainer).
Many thanks for any help.
Kristien Verheyen
---------------------------------------------------------------
The Animal Health Trust will be at Discover Dogs on 13th and 14th
November. Do come and visit our stand at this amazing canine
event at Earls Court!
----------------------------------------------------------------
#####################################################################################
This e-mail message has been scanned for Viruses and Content and cleared
by MailMarshal
For more information please visit www.marshalsoftware.com
#####################################################################################
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
