dfmlogit multinomial fractional model compositional data analysis

[Don Vicendese]

Hi,

Using Stata 13, I am running dfmlogit on the Stata dataset citybudget.dta available here
use http://fmwww.bc.edu/repec/bocode/c/citybudget.dta, clear

I am running the help file example

fmlogit governing safety education recreation social urbanplanning, ///
eta(minorityleft noleft houseval popdens)

It works fine. I am estimating marginal effects with

margins, dyex(houseval)

This also works fine but I only get marginal effects for the governing outcome and not the others -
safety, education, recreation, social, urbanplanning.


The help file uses dfmlogit but this command has been retired I believe and the following Stata error message suggests to use the margins commands.

dfmlogit only exists for older versions of Stata
use margins instead


How do I get the marginal effect for the other outcomes?

Thank you,

Don

[Maarten Buis]

use http://fmwww.bc.edu/repec/bocode/c/citybudget.dta, clear
fmlogit governing safety education recreation social urbanplanning, ///
eta(minorityleft noleft houseval popdens)
margins, dydx(*) predict(outcome(governing))
margins, dydx(*) predict(outcome(safety))
margins, dydx(*) predict(outcome(education))
margins, dydx(*) predict(outcome(recreation))
margins, dydx(*) predict(outcome(social))
margins, dydx(*) predict(outcome(urbanplanning))

[Don Vicendese]

Thanks very much for taking the time to respond Maarten. That works.

Don

[Jimmy Yuan]

use http://fmwww.bc.edu/repec/bocode/c/citybudget.dta, clear
fmlogit governing safety education recreation social urbanplanning, ///
eta(minorityleft noleft houseval popdens)
margins, dydx(*) predict(outcome(governing))
margins, dydx(*) predict(outcome(safety))
margins, dydx(*) predict(outcome(education))
margins, dydx(*) predict(outcome(recreation))
margins, dydx(*) predict(outcome(social))
margins, dydx(*) predict(outcome(urbanplanning))

Hi Maarten

Thanks for the new code provided !
But I am still confused about why I get same results of average marginal effects when I transfer the regression
from：fmlogit governing safety education recreation social urbanplanning, eta(minorityleft noleft houseval popdens)
to： fmlogit urbanplanning governing safety education recreation social , eta(minorityleft noleft houseval popdens)
Thank you for all your assistance