Estimating Overdispersion Parameter in Neg. Binomial Model

Welcome to the forum for MLwiN users. Feel free to post your question about MLwiN software here. The Centre for Multilevel Modelling take no responsibility for the accuracy of these posts, we are unable to monitor them closely. Do go ahead and post your question and thank you in advance if you find the time to post any answers!

Remember to check out our extensive software FAQs which may answer your question: http://www.bristol.ac.uk/cmm/software/s ... port-faqs/
Post Reply
ktw08
Posts: 2
Joined: Tue Sep 14, 2010 12:41 am

Estimating Overdispersion Parameter in Neg. Binomial Model

Post by ktw08 »

Hello all. I am running a two-level model using count data which shows evidence of overdispersion. I have settled on a Negative Binomial Model in which I would like to also get an estimate for the overdispersion parameter. While all the other estimates appear and are estimated, I can see the overdispersion term at the bottom, but do not receive an estimate for it. I believe the correct model name is multilevel negative binomial regression with extra negative binomial variation (Tseloni, 2000). According to this and a second piece (see references below) it is possible to compare the estimated overdispersion parameter between modes and estimate the unexplained heterogeneity between level one units which is attributed to their characteristics. I am just interested in the mechanics of getting MLwin to produce that estimate so I can interpret it in that fashion. Any information would be greatly appreciated. Thanks! Kevin Wolff

Tseloni, Andromachi 1999. COMPARING MULTILEVEL AND SINGLE-LEVEL NEGATIVE BINOMIAL REGRESSION MODELS
OF PERSONAL CRIMES: EVIDENCE FROM THE NATIONAL CRIME VICTIMIZATION SURVEY.

Tseloni, Andromachi 2000. Personal Criminal Victimization in the United States: Fixed and Random Effects of Individual and Household Characteristics. Journal of Quantitative Criminology. 16:4 415-442.
Post Reply