Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: st: Problems with Margeff


From   Neli Demireva <[email protected]>
To   "[email protected]" <[email protected]>
Subject   RE: st: Problems with Margeff
Date   Mon, 1 Nov 2010 16:58:06 +0000

Dear Richard,

This is a great guide! Thank you!

Neli 


Dr. Neli Demireva
Leverhulme Post-Doctoral Researcher
Nuffield College Research Fellow
Department of Sociology
Manor Road
OX1 3UQ
Oxford
Tel: 01865 286177
http://oxforddiversityproject.weebly.com/neli-demireva.html
________________________________________
From: [email protected] [[email protected]] On Behalf Of Richard Williams [[email protected]]
Sent: Monday, November 01, 2010 1:06 PM
To: [email protected]; [email protected]
Subject: Re: st: Problems with Margeff

Assuming you have Stata 11, why don't you use the new margins command
and see what it gives you?  Older commands tend to give wrong answers
because they don't know when variables are interdependent, e.g. they
will plug in a 1 for one of the dummies and mean values for the
others, when the others should be set to zero. This handout discusses
the problems with older commands like mfx and adjust when you have
multiple dummies, interaction terms and squared terms.  Older
commands also don't let you use the new -asobserved- option which
many feel is superior to plugging in means in your calculations.

http://www.nd.edu/~rwilliam/xsoc73994/Margins01.pdf

At 07:34 AM 11/1/2010, Neli Demireva wrote:

>Dear Stata Users,
>
>I am looking into migrants' transitions from unemployment to employment.
>
>I am using a mulltinomial logistic regression for this purpose with
>the following syntax line:
>xi:mlogit activity age agesq  i.famstat i.housing i.isced i.natz
>i.country i.anyclaim Ldurun Lmethod w1yr  if L1.activity==2 & sex==1
>[pw=wt], cluster(persid) baseoutcome(1)
>
>Ldurun and Lmethod are lagged variables. The cluster option is
>needed because individuals are represented with more than one observation.
>
>If I use the 'mfx' command to compute marginal effects, I get
>different signs of the coefficient and the m.e. estimate. I thought
>I will use margeff. Neither margeff, nor margeff, percent seem to be
>working (although I have margeff installed). Only margeff, outcome
>(2) produces m.e. estimates still different in sign from the coefficients.
>
>Any ideas as to why this is happening and suggestions how to proceed?
>
>With best wishes,
>
>Neli
>
>
>
>
>Dr. Neli Demireva
>Leverhulme Post-Doctoral Researcher
>Nuffield College Research Fellow
>Department of Sociology
>Manor Road
>OX1 3UQ
>Oxford
>Tel: 01865 286177
>http://oxforddiversityproject.weebly.com/neli-demireva.html
>*
>*   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/

-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
HOME:   (574)289-5227
EMAIL:  [email protected]
WWW:    http://www.nd.edu/~rwilliam

*
*   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/
*
*   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/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index