Histogram with multiple variables

[Esraa Dagli]

Hello everyone,

I have a problem creating a histogram for my experimental master's thesis. I need to do a chat data analysis for the arguments used during the chat by group members. I created six dummy variables for the ideas related to risk, honesty, and dishonesty. I want to create the following graph, but I can't find a way to make it. Basically, I want to show the share of those arguments used during the chat.

Thank you for your time.

[Ken Chui]

This is perhaps one of the many ways to approximate the example graph, but you'll need to create that summary first that looks like the sample data using collapse. Basically, just need to layer on with over() and by().

clear
input str5 reporting pc subvar y
A 1 1 0.4
A 1 2 0.05
A 1 3 0.02
A 1 4 0.02
A 1 5 0.08
A 2 1 0.36
A 2 2 0.15
A 2 3 0.15
A 2 4 0.0
A 2 5 0.05
B 1 1 0.08
B 1 2 0.08
B 1 3 0.16
B 1 4 0.03
B 1 5 0.0
B 2 1 0.06
B 2 2 0.2
B 2 3 0.06
B 2 4 0.06
B 2 5 0.02
end

label define l_pc 1 "GroupPC" 2 "GroupNoPC"
label values pc l_pc
label define l_subvar 1 "Money" 2 "Honesty" 3 "Insecurity" 4 "Rules" 5 "Others' behavior"
label values subvar l_subvar

graph bar y, over(subvar, lab(angle(45))) over(pc) by(reporting)

[Esraa Dagli]

Thank you soo much! It worked perfectly.

[Nick Cox]

I agree with Ken Chui that there are many ways to plot such data. Here I steal Ken's data example and use tabplot from the Stata Journal. For an overview, see https://www.statalist.org/forums/for...updated-on-ssc

set scheme s1color 
tabplot subvar pc [iw=y], by(reporting, note("")) showval(format(%03.2f)) separate(subvar) xtitle("") ytitle("")