Hi Isaak,
In the RUM-consistent nested logit model which is the default in Stata
10 the dissimilarity/IV parameter is not identified when the nest is
degenerate, i.e. has only one alternative. Since this is the case in
the 'yes' nest in your model, Stata constrains the IV parameter for
this nest to 1. This is explained in the help file (Remark on
degenerate branches) where its suggested that you use the
-constraints()- option to set unidentified IV parameters to 1.
Hope this helps.
Arne
On 16/08/07, Isaak Bergmann <[email protected]> wrote:
> Dear all,
> I'm using STATA10 and have problems with the nlogit command. I kindly ask for your help:
> My data comes from a choice experiment, in which subjects were asked to choose between two doctors (described in terms of gender and age) or to "opt-out", that is, chose neither of the two. I want to model this data as a nested logit model in which the choice among alternative doctors is nested within the choice to see a doctor at all.
>
> I created an dummy variable "nodoc" that indicates whether subjects choose a doctor or not (the nest). The data is organized as follows:
>
> Id Alt Gender Age Choice inc educ nodoc
> 1 1 -1 25 0 1 5 0
> 1 2 1 45 0 1 5 0
> 1 3 0 0 1 1 5 1
> 2 1 -1 60 1 4 1 0
> 2 2 -1 45 0 4 1 0
> 2 3 0 0 0 4 1 1
> 3 1 1 45 0 2 3 0
> 3 2 -1 25 1 2 3 0
> 3 3 0 0 0 2 3 1
>
> My command in STATA10 is:
> . nlogitgen nodoc = alt(no:1 | 2, yes: 3)
> . nlogit Choice Gender Age || nodoc: inc educ || alt:, noconstant case(id)
>
> STATA returns
> note: branch 2 of level 1 is degenerate and the associated dissimilarity parameter ([yes_tau]_cons) is not defined;
>
> Can someone please help me to understand:
> - I really wonder whether I specified the code correctly to represent my model?
> - How to interpret the above return message?
> - And does that explain why there are only dots in the SE of the IV value?
>
> I tried to figure it out, but all examples in the manuals etc. are structurally different (the nest is no opt-out as in my model).
>
> Thank you very much in advance for your help!
> Isaak
>
> RUM-consistent nested logit regression Number of obs = 774
> Case variable: id Number of cases = 258
>
> Alternative variable: alt Alts per case: min = 3
> avg = 3.0
> max = 3
>
> Wald chi2(5) = 77.80
> Log likelihood = -210.15028 Prob > chi2 = 0.0000
>
> ------------------------------------------------------------------------------
> choice | Coef. Std. Err. z P>|z| [95% Conf. Interval]
> -------------+----------------------------------------------------------------
> alt |
> gender | .6883889 .2263547 3.04 0.002 .2447419 1.132036
> age | .1485912 .1587822 0.94 0.349 -.1626163 .4597986
> ------------------------------------------------------------------------------
> nodoc equations
> ------------------------------------------------------------------------------
> no |
> inc | (base)
> edu | (base)
> -------- ----+----------------------------------------------------------------
> yes |
> inc | .0323411 .1466224 0.22 0.825 -.2550336 .3197158
> edu | -.1373286 .169304 -0.81 0.417 -.4691584 .1945011
> ------------------------------------------------------------------------------
> dissimilarity parameters
> ------------------------------------------------------------------------------
> nodoc |
> /no_tau | 1.089185 .3503095 .4025913 1.775779
> /yes_tau| 1 . . .
> ------------------------------------------------------------------------------
> LR test for IIA (tau = 1): chi2(1) = 0.07 Prob > chi2 = 0.7970
> ------------------------------------------------------------------------------
>
> _____________________________________________________________________
> Der WEB.DE SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen!
> http://smartsurfer.web.de/?mc=100071&distributionid=000000000066
>
> *
> * For searches and help try:
> * http://www.stata.com/support/faqs/res/findit.html
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
