Margins - Marginsplot - Manipulation of color for each estimate

[ema baroni]

Hello, I think I might be getting lost over a somewhat simple matter, and I apologize for the disturbance. After a regression and a margins operation, I want to create a graph, but I want to differentiate the color of the estimates. Below you'll find an example (short one) of the code.

reg test gender income geo
margins gender, dydx(income)
marginsplot, scheme(s1mono)

Note gender (0) boys (1) girls
Any suggestion?

[Andrew Musau]

You can use the options -plotdimension()- and -xdimension()- so that the plot is not rendered as one. If you need the connecting line, you can add it using -addplot()-.

webuse nhanes2, clear
regress bpsystol i.sex##c.bmi
margins sex, dydx(bmi)
*NOTE SEX IS CODED 1=male 2=female (MARGINS HIGHLIGHTED)
set scheme s1color
marginsplot, xdim(sex) plotdim(sex) ///
addplot(scatteri 1.446317 1 1.758291 2, lc(gray) lw(medium) recast(line) xlab("") xtitle("") leg(order(3 "Male" 4 "Female")))

Res.:

. margins sex, dydx(bmi)

Average marginal effects                                Number of obs = 10,351
Model VCE: OLS

Expression: Linear prediction, predict()
dy/dx wrt:  bmi

------------------------------------------------------------------------------
             |            Delta-method
             |      dy/dx   std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
bmi          |
         sex |
       Male  |   1.446317   .0771793    18.74   0.000     1.295031    1.597603
     Female  |   1.758291   .0527283    33.35   0.000     1.654933    1.861649
------------------------------------------------------------------------------