Marginsplot for categorical IV and categorical DV?

[Melanie Kolbe-Guyot]

I am trying to figure out how to visually show marginal effects of categorical IVs on categorical DVs.

mlogit cluster4 ib2.dualcitizenship, base(4)
margins r.dualcitizenship
marginsplot, horizontal plotopts(connect(i)) xline(0)

However, in the resulting graph (attached) the lines for the different outcomes overlap. Is there away to offset them just a little to have better visualization or another solution. I would like to see the marginal effects per contrast of the IV on each outcome possibility.

[Alan Neustadtl]

You could plot one predicted outcome at a time and recast your plot to show bars. Maybe something like this:

margins r.dualcitizenship, predict(outcome(1))
marginsplot, recast(bar) plotopts(barw(.8))

Repeat this for your other outcomes and optionally combine the four graphs into one graph using graph combine.

Best,
Alan

[Nicholas Winter]

My -mplotoffset-, available from SSC, is designed for exactly this case.

[Melanie Kolbe-Guyot]

Alan Neustadtl, that's a great alternative! Thank you very much. Works like a charm.

[Melanie Kolbe-Guyot]

My -mplotoffset-, available from SSC, is designed for exactly this case.

Thank you very much. This is exactly what I was looking for.