Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
| From | Mark Livingston <Mark.Livingston@glasgow.ac.uk> |
| To | "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu> |
| Subject | st: Spatial Panel analysis for Negative binomial or Poisson models |
| Date | Wed, 14 Aug 2013 12:27:42 +0100 |
Dear Statalisters, I am running a series of fixed effects panel (balanced) negative binomial models for data at a small area scale (in Stata IC version11 64 bit using xtnbreg after the xtset command) . I have been including a variable in the model to try and account for spatial autocorrelation ( The median of the dependent variable in the contiguous areas). However on checking for Moran's I test I find significant spatial autocorrelation left in many of the models. I need to either run spatial error or lag models to account for this. I can find various options for running spatial models for panel data (such as xsmle) but as far as I can see they all are based on OLS regression with the presumption of normal distribution in the dependent variable which is not appropriate for my data. Does anyone know of an option for doing spatial panel analysis for either negative binomial or Poisson models in Stata. Regards Mark Dr Mark Livingston School of Social and Political Sciences University of Glasgow * * 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/