Graph of multiple coefficients

[Heike Vethaak]

Hello all,

I have estimated an age-period-cohort model:

reg y ib1.age year_3-year_n ib1.cohort

This is a simplification of the model, where age is a factor variable indicating the age, year_i are dummies of the years (periods), and cohort is a factor variable of the cohorts.
Now I want to make a graph of the all age coefficients, all year coeffients and all cohort coefficients, as in Tunali et al. (2017):

In addition, I also want to have the confidence intervals in the graph. How do I make a graph of multiple coefficients from one regression?

[Scott Merryman]

See -margins- and -marginsplot-
For example:

sysuse auto,clear
reg price c.mpg c.weight
margins, at(mpg =(10(2)50)) atmeans
marginsplot, recastci(rline) name(gr1,replace)
margins, at(weight = (1500(250)5000)) atmeans
marginsplot, recastci(rline) name(gr2,replace)
graph combine gr1 gr2,row(1)

[Heike Vethaak]

Thank you very much for you response!

Unfortunately, this does not exactly solve my problem completely. I do not want to do any post-regression estimation. I only want to use the actual coefficients from the regression and not estimate the margins first. Maybe you (or anyone else) also know how I can do this?

[Scott Merryman]

Your example graphs display predicted values, however, if you just want to plot the estimated coefficients I would take a look Ben Jann's -coefplot -(ssc desc coefplot-).