Same model, different dependent variables: Single Y label in the combined graph

[Juan Diego]

Dear Statalisters,

I have troubles using graph combine, I don’t know how to get a single y label for a combination of graphs.

I ran the same model for three different dependent variables, then stored the graphs of the coefficients using coefplot. Since I want to get a single y label, I removed the y labels of the second and third graph with yscale(off).

I obtain this combined graph, but it squeezes the first graph.

I’ve tried to modify the first graph with xsize and ysize but it seems that graph combine overrides this adjustment.

Any feedback would be greatly appreciated.

Thank you very much in advance.

[Andrew Musau]

You can alleviate the problem by combining graph 1 with the combination of graph 2 and graph 3.

gr combine two.gph three.gph
gr save four, replace
gr combine one.gph four.gph

Ultimately, if you want perfection, you need to use the -by()- option as Nick Cox argues in https://journals.sagepub.com/doi/ful...36867X20976341. coefplot from SSC does support the -by()- option.

[Juan Diego]

Thanks Andrew!

The result, after some adjustments, is fantastic.



I like perfection, but since I don't master loops in Stata I will keep this result for this moment.

Regards