Using coefplot command for outcomes balance table

[Charlie Groscup]

Hi,

I am trying to generate an outcomes coefficient plot to demonstrate balance across outcomes at baseline using the coefplot command, but I seem to be running into some issues. I'm running several regressions with only two independent variables in each, and want to graph the independent variables on the y-axis with the different dependent variables stacked in-line with their appropriate independent variables.

I've included sample code below. The issue is that I want the mpg and foreign variables on the same plot but overlayed, and want the legend labels to appear on the y-axis instead.

Code:

 sysuse auto.dta, clear

egen price_std=std(price)
egen weight_std=std(weight)
egen length_std=std(length)

reg price_std mpg foreign
est store price

reg weight_std mpg foreign
est store weight

reg length_std mpg foreign
est store length

coefplot price weight length, drop(_cons) xline(0) coeflabel(price="price" weight="weight" length="length") bycoef legend(off)

I've also attached a working paper I found online below with an example that mimics I want to produce (See Figure 3 on page 42). Source: https://www.nber.org/system/files/wo...785/w25785.pdf

[Andrew Musau]

coefplot is from SSC (FAQ Advice #12). Thanks for the reproducible example. Would be easier to include the y-axis labels as marker labels, but here is a way manually calculating the offsets.

sysuse auto.dta, clear
egen price_std=std(price)
egen weight_std=std(weight)
egen length_std=std(length)

reg price_std mpg foreign
est store price

reg weight_std mpg foreign
est store weight

reg length_std mpg foreign
est store length

set scheme s1color

coefplot (price) (weight) (length), drop(_cons) xline(0) ///
ylabel(0.75 "price" 1 "weight" 1.25 "length" 1.75 "price" ///
2 "weight" 2.25 "length", notick) nokey text(1 -1.35 ///
"Outcome 1", orientation(vertical)) text(2 -1.35 ///
"Outcome 2", orientation(vertical)) ysc(outergap(5))

[Charlie Groscup]

Hi Andrew,

Thank you so much for that, that's a huge help!