Hi everyone, I am running some propensity score models.
My y var is a 5 point scale. The treatment is binary.
Afterward, I am running the post estimation predict commands:
My understanding is that the predict commend here would give me the potential outcome means for the matched control (y0) and treatment (y1) groups. Running a sum command at least confirms these expectations.
What I would like to do is plot the potential outcomes on a single figure. I have been playing around with a twoway kernal density plot
The result is not bad, but not what I am after. 
Ideally, I would like something cleaner. Basically just a plot with a smoothed trendline of the two potential outcomes, so that I can point to a shift in the mean.
Any help would be great. Also, I might be way off base, so correct would be welcome.
Code:
teffects psmatch (Y) (treatment dist_control dir_control, logit), gen(match1)
Afterward, I am running the post estimation predict commands:
Code:
predict y0 y1, po
What I would like to do is plot the potential outcomes on a single figure. I have been playing around with a twoway kernal density plot
Code:
twoway kdensity y0 || kdensity y1
Ideally, I would like something cleaner. Basically just a plot with a smoothed trendline of the two potential outcomes, so that I can point to a shift in the mean.
Any help would be great. Also, I might be way off base, so correct would be welcome.
