Drawing a line between points in coefplot

[Lisa Acker]

Graph.gph

I have the following graph where I plot the Coefficients of a Difference in Difference analysis (using coefplot). To make the graph look nicer, I would like to draw a line between all red and all blue points respectively. How would I do that?

I used the following command to plot the graph:

```
*Running the regressions
reg deal_rate dummy_plus_7 dummy_plus_6 dummy_plus_5 dummy_plus_4 dummy_plus_3 dummy_plus_2 dummy_plus_1 cut_dummy dummy_lag_* if Single_Collateral == 1 , vce(robust)
estimates store Single

reg deal_rate dummy_plus_7 dummy_plus_6 dummy_plus_5 dummy_plus_4 dummy_plus_3 dummy_plus_2 dummy_plus_1 cut_dummy dummy_lag_* if Single_Collateral == 0, vce(robust)
estimates store Pool

*Plotting the graph
coefplot Single Pool, vertical drop(_cons) yline(0) xlabel(, angle(90))

````

[Andrew Musau]

Reproducible example? coefplot is from SSC. Also note the advice to upload images in PNG format (FAQ Advice #12) .

[Lisa Acker]

Okay then here we go: I took a random dataset from Stata. The regression does not make any sense but it shows what I want to do with my data.

The code is:
```
ssc install coefplot

sysuse emd, clear

*Creating Random Dummy
gen dummy = 1 if income_pc > 6000
replace dummy = 0 if income_pc <= 6000

*Having two random regressions
reg ln_income median_age fips hhsize if dummy == 1
estimates store model1

reg ln_income median_age fips hhsize if dummy == 0
estimates store model2

coefplot model1 model2, vertical drop(_cons) yline(0) xlabel(, angle(90))

```

I want to connect the blue and red points with a line each.

[Lisa Acker]

And here in picture format

[Andrew Musau]

Thanks for the data example. Does this do it?

webuse emd, clear

*Creating Random Dummy
gen dummy = 1 if income_pc > 6000
replace dummy = 0 if income_pc <= 6000

*Having two random regressions
reg ln_income median_age fips hhsize if dummy == 1
estimates store m1

reg ln_income median_age fips hhsize if dummy == 0
estimates store m2

coefplot m1 m2 , vert drop(_cons) yline(0) xlab(, angle(90))  recast(connected)