[Limdep Nlogit List] SEs around market shares

[Vera Vennedey]

Dear all, 

Below you see an extract from an output of a discrete choice analysis. For the variables CO300 and CO100 I (dummy coded cost variables) I get no confidence intervals   (only ****). As the standard error is very large for these Variables (.5932D+07) I assume it means that the confidence intervals are infinite, is this true? 

I would very much appreciate  your comments on this. 

Greetings, 
Vera Vennedey


        |                  Standard            Prob.      95% Confidence
  CHOICE|  Coefficient       Error       z    |z|>Z*         Interval
--------+--------------------------------------------------------------------
        |Random parameters in utility functions
  TOOTH1|   -1.74063***      .17235   -10.10  .0000    -2.07843  -1.40283
  TOOTH2|   -4.65532***      .35318   -13.18  .0000    -5.34754  -3.96310
   CO500|    -.28489*        .16493    -1.73  .0841     -.60815    .03837
   CO300|    -.08466      .5932D+07      .00 1.0000 ***********  ***********
   CO100|     .08466      .5932D+07      .00 1.0000 ***********  ***********


Institut für Gesundheitsökonomie und Klinische Epidemiologie (IGKE) Klinikum der Universität zu Köln (AöR) 
Cologne Institute for Health Economics and Clinical Epidemiology, The University Hospital of Cologne (AöR) 

Gleueler Strasse 176 - 178 / II
D - 50935 Köln/ Cologne

Email:  vera.vennedey at uk-koeln.de
Tel.:     +49 (0) 221 4679-133
Fax:      +49 (0) 221 4302304






-----Ursprüngliche Nachricht-----
Von: limdep-bounces at limdep.itls.usyd.edu.au [mailto:limdep-bounces at limdep.itls.usyd.edu.au] Im Auftrag von Samuel Kaufman
Gesendet: Dienstag, 3. Mai 2016 22:04
An: Limdep and Nlogit Mailing List
Betreff: Re: [Limdep Nlogit List] SEs around market shares

I have tried several times to be removed from this list, and every time it has had no effect.

-----Original Message-----
From: limdep-bounces at limdep.itls.usyd.edu.au
[mailto:limdep-bounces at limdep.itls.usyd.edu.au] On Behalf Of William Greene
Sent: Monday, May 02, 2016 10:57 PM
To: Limdep and Nlogit Mailing List
Subject: Re: [Limdep Nlogit List] SEs around market shares

Matt.
;SIMULATE can't compute a measure of variability around the fitted market shares. Since they have to sum to 1, there is no way to form a confidence interval for each one while being internally consistent.
/Bill Greene


On Mon, May 2, 2016 at 1:14 PM, Matt Quaife <matthew.j.quaife at gmail.com>
wrote:

> Hi All,
>
> Hopefully a quick and easy one this. I am seeking a measure of 
> variability around a set of market share predictions as produced by 
> the ;simulate command, preferably standard errors.
>
> If this is achievable - either through a workaround or explicitly in 
> NLOGIT, I would appreciate any guidance.
>
> Best wishes
> Matt
> _______________________________________________
> Limdep site list
> Limdep at limdep.itls.usyd.edu.au
> http://limdep.itls.usyd.edu.au
>



--
William Greene
Department of Economics
Stern School of Business, New York University
44 West 4 St., 7-90
New York, NY, 10012
URL: http://people.stern.nyu.edu/wgreene
Email: wgreene at stern.nyu.edu
Ph. +1.212.998.0876
_______________________________________________
Limdep site list
Limdep at limdep.itls.usyd.edu.au
http://limdep.itls.usyd.edu.au
_______________________________________________
Limdep site list
Limdep at limdep.itls.usyd.edu.au
http://limdep.itls.usyd.edu.au