Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

AW: st: negative binomial model

From   "Mareike" <>
To   <>
Subject   AW: st: negative binomial model
Date   Sun, 20 Jun 2010 22:04:08 +0200

Thank you, Maarten, for your hint. 
Now that I tried to run both the Poisson model with fixed and random
effects, with the RE specification actually the same problem occurred again.
This probably implies that there is really no other choice but to dropp some
of the explanatory variables?!?


-----Ursprüngliche Nachricht-----
[] Im Auftrag von Maarten buis
Gesendet: Sonntag, 20. Juni 2010 19:04
Betreff: Re: st: negative binomial model

Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen

--- On Sun, 20/6/10, Mareike wrote:
> I want to analyze count data. Given that for several
> reasons the Poisson analysis showed to be inappropriate,
> my idea was to run a negative binomial model(using the
> command xtnbreg because I have panel data). But what
> happened was that Stata was giving me one iteration after
> the other (what appeared in the result window lokked like
> the following: 
> Iteration 14721:log likelihood = -53.558237  (not
> concave)
> and after 3 hours and ten thousands of iterations I stopped
> the program without having gotten any result.
> Does anybody of you have an idea of what happened? I didn't
> have the problem when running the Poisson model using the
> same dataset.

It means that the model is too complex for your data, so you
need to simplify your model, either move back to -xtpoisson-
or remove some explanatory variables.

Hope this helps,


*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index