sugie wrote:
> > I agree that this approach is a little bit weird.
> > But the result looks like what I expect.
> > Is this approach seriously flawed?
> > Any comment or better ideas?
sugie,
Before carrying on any estimation you have to ask your self what is it that
you want to estimate. Every model has a theoretical background, regardless
of yielding a desired outcome or not, you have to constraint your analysis
to this framework. Multinomial logit--as Clive mentioned it--is used in
another context. If you estimate -mlogit- your dependent variable is
"cities", and the outcome of such an estimation makes no sense at all.
You results from the mlogit will read like this: "an increase in the
population density changes the probability (or the log odds ratio) of city
being "i" relative to being "j", where "j" is the base outcome". And this is
not what you want; there is not a decision process taking place!
If the variable of interest is Population Density then this has to be the
right hand side, dependent variable.
Rafa
----- Original Message -----
From: "SR Millis" <[email protected]>
To: <[email protected]>
Sent: Friday, June 18, 2004 2:54 PM
Subject: RE: st: Multinomial Logit vs. Regression with dummy
> Scott Menard's book, "applied logistic regression
> analysis (2nd ed)" published by Sage provides an
> excellent discussion of the potential limitations of
> the multinomial logistic regression model---and
> discusses alternatives.
>
> SR Millis
>
>
> --- Sugie Lee <[email protected]> wrote:
> > It seems like that -discrim- can only deal with a
> > binary dependent
> > variable(0/1).
> >
> > .discrim CITY INDEP1 INDEP2 INDEP3 INDEP4
> >
> > This command did not work because CITY has three
> > groups (city A, city B,
> > city C).
> >
> > I looked at previous postings in the STATA Archive
> > about discriminant
> > analysis.
> > I think we have two ado files for discriminant
> > analysis: -discrim-
> > and -daoneway-.
> > But -daoneway- is not running on the STATA 8 of
> > which I am using. It's for
> > STATA 7.
> >
> > Nick suggested multinomial logit analysis for this
> > kind of case(see the
> > below posting on Dec. 11th 2002).
> > His suggestion seems like what I was supposed to do
> > with multinomial logit
> > analysis.
> >
> > .mlogit CITY INDEP1 INDEP2 INDEP3 INDEP4
> >
> > I agree that this approach is a little bit weird.
> > But the result looks like what I expect.
> > Is this approach seriously flawed?
> > Any comment or better ideas?
> > Thanks,
> >
> > sugie
> >
> >
> >
> >
> >
> > st: RE: discriminant function analysis
> >
> --------------------------------------------------------------------------
--
> > ----
> > From "Nick Cox" <[email protected]>
> > To <[email protected]>
> > Subject st: RE: discriminant function analysis
> > Date Wed, 11 Dec 2002 23:01:53 -0000
> >
> --------------------------------------------------------------------------
--
> > ----
> > Sheela Athreya
> > > Does anyone know if it is possible to perform the
> > following
> > > analysis on
> > > STATA:
> > >
> > > I have a set of 35 fossils, from known taxonomic
> > groups
> > > (1,2,3). There
> > > are five measurements per fossil. I then have a
> > single
> > > newly discovered
> > > fossil, and want to find out how it will be
> > classified
> > > based on these
> > > measurements.
> > >
> > > I wanted to use discriminant function analysis,
> > but I'm not
> > > sure if a)
> > > that is the right approach, and b) how I can do it
> > using the command
> > > -discrim-, which requires that the grouping
> > variable be
> > > only 0/1, instead
> > > of the three groups I have defined a priori.
> >
> > With two categories discriminant analysis
> > is typically inferior to logit, but as
> > you have three, this sounds much more like -mlogit-
> > in Stata [sic]. However, I yield to experts here.
> >
> > Nick
> > [email protected]
> >
> --------------------------------------------------------------------------
--
> > ----
> >
> >
> >
> >
> >
> > -----Original Message-----
> > From: [email protected]
> > [mailto:[email protected]]On
> > Behalf Of SR Millis
> > Sent: Thursday, June 17, 2004 2:53 PM
> > To: [email protected]
> > Subject: Re: st: Multinomial Logit vs. Regression
> > with dummy
> >
> >
> > Another alternative is discriminant function
> > analysis.
> >
> > SR Millis
> >
> > --- Sugie Lee <[email protected]> wrote:
> > > Hello!
> > > I want someone to help me on a following question.
> > > Let's suppose we have three cities (city A, city
> > B,
> > > city C).
> > > And we have just one variable which is population
> > > density(POPDEN)
> > >
> > > I may try regular regression as follows:
> > > .reg POPDEN dummy(city A) dummy(city B)
> > >
> > > What if I use multinomial logit?
> > > In this case, the dependent variable is
> > > "CITY"(A,B,C)
> > >
> > > .mlogit CITY POPDEN
> > > .listcoef
> > >
> > > I want to see differences of population density
> > > between cities.
> > > "listcoef" command immediately after "mlogit" will
> > > give me these
> > > differences.
> > >
> > > My question is whether I can use mlogit for this
> > > case?
> > >
> > > If mlogit is possible for this case, I will do
> > > analysis with more
> > > independent variables as follows:
> > > .mlogit CITY POPDEN INDEP2 INDEP3 INDEP4
> > >
> > > Any comments and suggestions will be helpful to
> > me.
> > > Thanks.
> > >
> > > Sugie
> > >
> >
> >
> > Not if your real research question is "what affects
> > population density?",
> > which I suspect is what you really want to ask. If
> > so, using -mlogit- with
> > cities as the dependent variable strikes me as a
> > bizarre approach: asking
> > the question "does density affect the probability of
> > a city being a city
> > (or being the city that it actually is)" makes
> > little sense, to me at
> > least. In my opinion, the OLS formulation you
> > provided is much more
> > sensible, so try that and work from there.
> >
> > CLIVE NICHOLAS |t: 0(044)191 222 5969
> > Politics |e: [email protected]
> > Newcastle University |http://www.ncl.ac.uk/geps
> >
> > *
> > * 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/
> >
> >
> >
>
>
> =====
> Scott R Millis, PhD, MEd, ABPP (CN & RP)
> Director, Office of Clinical Trials
> Kessler Medical Rehabilitation Research & Education Corp
> 1199 Pleasant Valley Way
> West Orange, NJ 07052
>
> *********************************************************
> This electronic message may contain information that is confidential
> and/or legally privileged. It is intended only for the use of the
> individual(s) and entity named as recipients in the message. If you are
> not an intended recipient of this message, please notify the sender
> immediately and delete the material from any computer. Do not deliver,
> distribute or copy this message, and do not disclose its contents or
> take any action in reliance on the information it contains. Thank you.
> *
> * 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/
