Re: st: Nested Logit Model

["Alvarez,Sergio" <sergioal@ufl.edu>]

Felipe,

It sounds like you need to create the alternatives.  Remember that you 
need an observation for each choice at the terminal end of the tree.  If 
you give more info on how your data is organized we may be able to help 
some more.

Here's a helpful link:
http://www.stata.com/support/faqs/stat/nlogit.html

Usually the data we have access to records only the actual choice, and 
it is up to the researcher to create the alternative choices, and to 
create a binary variable that indicates whether the choice was made or 
not.

It would also be good to try and run a conditional logit first; you'll 
know if your data are in the right format before you start going into 
setting up the nested logit.

Hope it helps,

Sergio

On Wed, 2 Nov 2011 13:20:02 -0300, Luis Felipe Correa Smits wrote:
> Dear All,
>
> I’m quite new to stata so I don’t know If I have a command or 
> database
> issue. So here I go:
>
> I’m trying to solve a model for my thesis dissertation on the
> probability for a pharmacy to enter and then locate in a Chilean
> district. My ideal tree should be:
>
>
>                  N
>           /              \
>         /                  \
>       /                       \
> Enter                Not Enter
>    |                    /        |    \
>    |                  /          |      \
>    |                /            |        \
>    0     Location=1       …      N
>
> I generated the nlgoit tree with the district=0 for the decision not
> to enter the industry, and all the observed districts for the 
> decision
> to enter (nlogitgen type=district (No: 0, Yes: 1|…|N). Then I
> generated the “chosen” variable based on whether there’s a 
> pharmacy on
> the district or not.
>
> Then I tried to run the regression (nlogit   chosen
> number_of_stores_pharmacy_chain1    number_of_stores_pharmacy_chain2
>  number_of_stores_pharmacy_chain3
> number_of_stores_independent_pharmacies
> type_of_competitor_base_on_distance[inmediate; nearby; far away]
> population_between_0_to_5_years_old  
> population_more_than_65_years_old
>  median_rent  district_area[sq miles] business_density||type:
> income_per_capita urban_population rural_population||district:,
> case(district_id)).
>
> But I get:
> note: 334 cases (1336 obs) dropped due to no positive outcome or
> multiple positive outcomes per case
> no cases remain after removing invalid observations
> r(2000);
>
>
> I don’t know what I’m doing wrong. Can anyone help me please?
>
> Thanks!
>
>
> Best Regards
>
> Felipe Correa
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/

--
Sergio Alvarez
Food and Resource Economics
University of Florida
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/