[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: IIA violations and subsequent steps

From   "Maarten Buis" <>
To   <>
Subject   RE: st: IIA violations and subsequent steps
Date   Mon, 6 Aug 2007 12:14:27 +0200

Maarten L. Buis
Department of Social Research Methodology 
Vrije Universiteit Amsterdam 
Boelelaan 1081 
1081 HV Amsterdam 
The Netherlands

visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434 

+31 20 5986715

--  Douglas Garrett <> wrote:
> I am currently running a multinomial logistic regression model and
> have found that it violates the IIA assumption. <snip> 

--- Xiaoheng Zhang answered:
> I think IIA is irrelavent to the multinomial logistic regression
> because it implies that there is no correlation in terms of latent
> utility between any two alternatives. Just wondering why you think
> in your case it is violated.

IIA is an assumption underlying -mlogit-. It can be tested using
a Hausman test. For a nice example of how IIA is relevant for
-mlogit- see the help file of Jeroen Weesie's -iia- under the 
heading `Some fun ("Substantive background")'. You can install 
-iia- by first typing:

net cd

and than:

net install iia

-- Maarten


*   For searches and help try:

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