Multiple box plots on one graph by another variable, sequentially

[N Winters]

Hi all,

I am trying to create a figure for box plots that includes multiple variables - but I want to stratify by another variable (whether the response is from a nurse or a doctor).
What I have is this code:
graph hbox Q1_1 Q1_6 Q1_7 Q1_9 Q1_12 Q1_3 Q1_16 Q1_18 Q1_17 Q1_19 Q1_15, by(Doc_nurse) asy noout legend(col(1) ring(0) position(1))

but I don't want a legend for each question, I want these listed on the right by their box plot and a legend for the stratified variable (nurse vs doctor). The resulting graph is shown at the end.

What I want is to have one graph like so

Q1_1 -> nurse response
Q1_1 -> doctor repsonse
Q1_3 -> nurse
Q1_3 -> doc
etc. for the remaining variables

with a legend for just either nurse or doctor response, with the variable name for each on the right hand side.

Is this possible? I have tried using the -combine- command but this has limited options for axis, and basically leads to the same issue.

any help would be appreciated, thank you in advance!
N

[Nick Cox]

Sorry, I can't read your image. What we recommend here for images is .png. See FAQ Advice #12.

A sample dataset, real or realistic, would help greatly. By the sound of it, four variables would be enough to show the principle. See again FAQ Advice #12.

Did you check out this Tip?

SJ-14-4 gr0062 . . . . . . . . . . . . Stata tip 121: Box plots side by side
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . N. J. Cox
Q4/14 SJ 14(4):991--996 (no commands)
tip on how to and which data should be plotted side by
side with box plots

[N Winters]

I don't know why the image didn't show up, I uploaded through statalist and it shows up on mine in the post (and it was a .png). I also can't access that paper its behind a paywall - even through institutional libraries.


example data (Q1_1 etc. are likert scale survey data):

Study	Q1_1	Q1_3	Q1_6	Doc_nurse
1	2	4	7	doctor
2	1	3	10	nurse
3	3	4	2	doctor
4	5	9	3	nurse
5	4	10	4	nurse
6	10	1	5	nurse
7	9	2	10	doctor
8	7	1	3	nurse

[Nick Cox]

It's worth paying for, naturally. The main idea is that you may need a reshape to make progress.

Your data example is helpful. Using dataex (SSC) would have been even more helpful: again please see Advice #12.

Here is a token example.

clear
input Study    Q1_1    Q1_3    Q1_6 str6 Doc_nurse
1    2    4    7    doctor
2    1    3    10    nurse
3    3    4    2    doctor
4    5    9    3    nurse
5    4    10    4    nurse
6    10    1    5    nurse
7    9    2    10    doctor
8    7    1    3    nurse
end

reshape long Q1_ , i(Study) j(Question)
rename Q1_ Answer
graph hbox Answer, over(Doc_nurse) over(Question)