st: margins

[Chiara Mussida <cmussida@gmail.com>]

Dear All,
I estimated mlogit model as follows:

xi: mlogit utr male_north male_centre male_south fem_north fem_centre
age2 age3 age4 italian famsize primary diploma durric gdp_growth
i.year if age>=15 & age<=55, b(3)

xi is for year dummy variables, gender and area of residence variables
as covariates (male and female; north, centre and south), dummies for
individual age, citizenship (whether italian or not), # household
members (famsize), educational attainment level, and two continuous
variables, for unemplyment duration and gdp growth, respectively.

I'm interested in computing and plotting the marginal effects of
unemployment duration (variable durric) for my interactions gender and
area of residence. I tried - without success - two postest commands:

1. mfx compute, predict(outcome(1))

In this case, the mfx are not saved and therefore I cannot plot them;

2.  margins durric

The error message: 'durric' not found in list of covariates
r(322);

As an alternative I tried with the command:
margins, dydx(durric)
marginsplot

but this does not satisfy my requirement, since it computes marginal
effects only at the mean.

By looking at the Stata archive, at this link:
http://www.stata.com/capabilities/overview/marginal-analysis/ ,

I found some suggestions on margin and marginplot but I only obtained
error messages (e.g., r(322) above). I would be pleased to obtain a
plot like the one in the example in the archive for the margins of
durric for each gender and area of residence in my data.

Thanks for your help and assistance,

-- 
Chiara Mussida
Assistant Professor
Università Cattolica Sacro Cuore
Piacenza - Italy

*
*   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/