Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, is already up and running.

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

Re: st: nbreg with fixed effect vs xtnbreg,fe

From   Richard Williams <>
Subject   Re: st: nbreg with fixed effect vs xtnbreg,fe
Date   Tue, 07 Feb 2012 11:06:19 -0500

At 03:55 PM 2/6/2012, Shikha Sinha wrote:
Hi All,

I am trying to model count data with county fixed effect. It is a
cross-sectional data. I want to know if

xi: nbreg DV IV control i.county

is same as

xtnbreg DV IV control, fe i(county)

I get different results. Please advise what is the best way to model
count data with fixed effect.

I haven't been following this thread, but just as a sidelight Paul Allison claims that the method used by -xtnbreg- does not produce a true fixed effects regression model. He proposes an alternative approach that is much slower but that he claims is correct. See pps. 62-64 of his 2009 Sage Book "Fixed Effects Regression Models." In fact, if I am reading both you and Allison correctly, I think he would say to do things the way you are doing it in your first command, i.e. create a bunch of dummies for county. (But I won't swear to it.)

I'd be interested to hear if StataCorp violently disagrees or has taken some steps to correct the problem, which Allison first pointed out in a 2002 paper.

Regardless of whether Allison is right on this particular point, it is a great book and well worth the $15 or so that it costs.

Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
HOME:   (574)289-5227
EMAIL:  Richard.A.Williams.5@ND.Edu

*   For searches and help try:

© Copyright 1996–2015 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index