Re: st: margeff after mlogit

[Richard Williams <richardwilliams.ndu@gmail.com>]

A few other thoughts:

I am surprised it won't run with only a 1000 
cases and one variable. That makes me wonder if 
there is something weird about the data or the 
model, e.g. does one category only have 3 cases? 
But, if it is a problem with margeff, maybe switching to margins will fix it.

Consider (at least for exploratory purposes) 
estimating the marginal effects at the means 
(MEMs) rather than the AMEs. This should be far 
less computationally intensive. If problems 
persist, that again makes me wonder about the data or model.

One of the nice things about margeff is that a 
single command will generate all the marginal 
effects. With margins, you have to have a 
separate command for each outcome, e.g. if y has 
4 categories you need to do something like

mlogit y i.x1 i.x2
margins, dydx(*) predict(outcome(#1))
margins, dydx(*) predict(outcome(#2))
margins, dydx(*) predict(outcome(#3))
margins, dydx(*) predict(outcome(#4))

Having margins do all the outcomes at once after 
things like mlogit and ologit is on my ongoing 
wish list for each new version of stata.


At 09:54 AM 12/12/2013, Maarten Buis wrote:
> With Stata 12 you can use the official -margins- command instead of
> the user-written -margeff- command to get the average marginal
> effects. You would have to use factor variables instead of manually
> creating the indicator variables; see -help fvvarlist-.
>
> Hope this helps,
> Maarten
>
>
> On Thu, Dec 12, 2013 at 3:11 PM, LIECHTI Lena <lena.liechti@unifr.ch> wrote:
> > Dear STATA-List ;I am trying to compute 
> average marginal effects via margeff command 
> (example: margeff, dummies(educ_* child* age* 
> zivl* lang* pop*) replace) after a multinomial 
> logistic regression. Stata (SE 12.1) seems to 
> compute something but even after several days, 
> the process is still going on.;I have a huge 
> data set (around 700´000 observations). The 
> dependent variable has three possible outcomes 
> and I include 18 independent variables. But a 
> trial with a much smaller data set (1´000 
> observations) and only one independent variable 
> also takes ages without providing any 
> results.;Does anybody know, why this happens 
> and how to overcome this problem?;Thanks and best regards;Lena
> >
> >
> >
> > *
> > *   For searches and help try:
> > *   http://www.stata.com/help.cgi?search
> > *   http://www.stata.com/support/faqs/resources/statalist-faq/
> > *   http://www.ats.ucla.edu/stat/stata/
>
>
>
> --
> ---------------------------------
> Maarten L. Buis
> WZB
> Reichpietschufer 50
> 10785 Berlin
> Germany
>
> http://www.maartenbuis.nl
> ---------------------------------
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/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:  Richard.A.Williams.5@ND.Edu
WWW:    http://www.nd.edu/~rwilliam


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/