Graphing different variations of x variable age [age_squared; ln(age); age] on one graph

[Nae Khar]

Hello,

Does anyone know how to recreate this graph (image attached) through Stata?

I am basically trying to plot the three regressions (that have a different variations of the predictor X variable age) on one graph:


Regression 1, which is represented by the black line on the graph:

reg ln_AHE age if female==0 & bachelor==0, robust level(95)

Regression 1 line equation: ln_AHE = 1.9 + .027age



Regression 2, which is represented by the blue dashed line on the graph:

reg ln_AHE ln_age if female==0 & bachelor==0, robust level(95)

Regression 2 line equation: ln_AHE = -0.003 + 0.80ln_age



Regression 3, which is represented by the red line on the graph:

reg ln_AHE ln_age age_squared if female==0 & bachelor==0, robust level(95)

Regression 3 line equation: ln_AHE = 2.7 + -0.029age + .001age_squared



I attempted this:

graph twoway (scatter ln_AHE age) (lfit ln_AHE age if female==0 & bachelor==0) (qfit ln_AHE ln_age if female==0 & bachelor==0) (qfit ln_AHE age age_squared if female==0 & bachelor==0)

But this results in only one of the lines being correctly graphed (line 1 from regression 1) because the other two regressions require a different variation of the x axis age (e.g. ln_age).

Any advice would be highly appreciated.

Thanks