Re: st: Questions on -mlogit-

[Ada Ma <pec187@abdn.ac.uk>]

Hi Clive,

If your base category is Labour (i.e. Labour=0) then the Conservatives are MORE 
likely to win if celtic==1, and to calculate the the probability of con==1 given 
that celtic==1 (and ignoring everything else), the function you should use is:

exp(1.96e-17)/(1+exp(1.96e-17))

Not 1/1.96e-17 as you have used.  I don't know anything about the rest of your post.

Ada Ma


Clive Nicholas wrote:


> All,
>
> Just a couple of questions after (re-)running some -mlogit- models.
>
> First, after running one such model (modelling party victories at British
> by-elections), one of the coefficients returned looked a bit odd:
>
> -----------------------------------------------------
>              |               Robust
>        bewin |        RRR   Std. Err.      z    P>|z|
> -------------+---------------------------------------
> con          |
>       celtic |   1.96e-17   2.21e-17   -34.21   0.000
>
> The base category against the Conservatives here is Labour. What this is
> saying, if I'm interpreting this (quasi-)odds ratio correctly, is that,
> after I
>
> . display 1/0.0000000000000000196
> 5.102e+16
>
> the model is saying that the odds of Labour winning a by-election in
> either Scotland or Wales are 51,020,000,000,000,000 times greater than the
> Conservatives, Now, I'd certainly expect Labour's odds to be higher in
> this region simply because they have been the strongest party there for
> many decades, but not quite _this_ high! What to do about this? This model
> was estimated along with seven other factors. No more factors could be
> added without getting back an -mlogit- model returned with blanks all over
> the place.
>
> Second, Dow and Endersby (2004: 117) have recently reported MNL (and MNP)
> models together with log_10 hessian condition numbers, which I think would
> be useful to report, for the reasons they mention. -search-ing and
> -findit- only produced the -hessian- option in -ml-, which isn't what I'm
> looking for. Is there a way of reporting these after -mlogit- in Stata
> 8.2?
>
> Ta.
>
> CLIVE NICHOLAS        |t: 0(44)191 222 5969
> Politics Building     |e: clive.nicholas@ncl.ac.uk
> School of Geography,  |f: 0(44)870 126 2421
>  Politics & Sociology |
> University of         |
>  Newcastle-upon-Tyne  |
> Newcastle-upon-Tyne   |
> NE1 7RU		      |
> United Kingdom	      |http://www.ncl.ac.uk/geps
>
> Dow JA and Endersby JW (2004) "Multinomial probit and multinomial logit: a
> comparison of choice models for voting research", ELECTORAL STUDIES 23(1):
> 107-22.
> *
> *   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/
--
Ada Ma
Research Assistant
Department of Economics
University of Aberdeen
Edward Wright Building F55
http://www.abdn.ac.uk/economics/
http://www.abdn.ac.uk/~pec187/firstpage.htm
Tel:   +44 1224 273417
Fax:   +44 1224 272181
Email: a.ma@abdn.ac.uk

*
*   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/