Graphing more than 1 variable in cibar (is this possible?)

[Zach Goldberg]

Greetings,

I'm running Stata 15.1 on OSX. I've recently installed the cibar package as other methods for generating bar graphs with confidence intervals seemed too laborious. However, cibar appears to only allow you to graph a single variable. If I try to include more than one, only the first in the series is graphed. Is there an option I'm missing here or is this merely the limitation of using cibar? Thanks for your help!

[Nick Cox]

cibar is from SSC. Its declared syntax is just

cibar varname [if], over1(varname) [options]

so using it to get confidence intervals for several variables would require a change of data structure so that different variables become distinct groups of observations. You're right that it allows but ignores variables other than the first-named. For why, you'd need to ask the author.

Although it's been abandoned by its author, ciplot (SSC) supports multiple variables. It even supports, by accident rather than design, the awful dynamite/detonator/plunger plot style:

sysuse auto, clear
ciplot turn trunk
ciplot turn trunk, recast(bar) bfcolor(none ..) base(0)

It's usually better to show much more than means and their CIs.

[Zach Goldberg]

Hey Nick,

This is very helpful. Is it possible to also have the percentages/figures labeled above each bar as is done with the blabel() command in 'graph bar'?

-Zach

[Nick Cox]

Which percentages or figures would that be? With two or more variables the number of observations used will by default be the same for all variables -- precisely those observations for which all variables plotted have non-missing values.

[Zach Goldberg]

Hey Nick,

Here's some sample data to better illustrate what I mean

* Example generated by -dataex-. To install: ssc install dataex
clear
input byte(courtesy disrespect) float hispanic
1 1 0
2 2 0
4 3 0
3 3 1
4 3 1
3 3 0
2 4 1
3 3 1
2 2 1
4 4 1
2 2 1
1 1 1
1 1 1
1 1 1
2 2 1
1 1 0
2 2 0
1 1 1
1 1 1
4 2 1
3 3 1
2 2 1
2 3 1
3 2 1
3 3 0
2 2 0
1 1 0
1 1 0
3 3 1
2 1 1
1 1 0
3 3 0
1 1 0
4 4 1
3 3 1
1 1 1
3 3 1
2 2 1
3 3 0
6 3 0
3 3 0
2 2 1
4 4 1
2 2 1
3 5 1
2 2 1
4 2 0
1 1 0
4 4 1
2 5 0
3 3 1
2 2 1
4 4 1
1 1 1
5 5 1
2 2 0
1 1 1
1 1 1
4 3 1
1 1 1
5 2 1
1 1 1
1 2 1
1 1 1
2 2 0
3 2 1
1 1 1
1 1 1
3 1 1
1 1 1
1 1 1
1 1 0
3 3 0
1 1 1
1 1 1
1 1 1
5 5 1
3 2 1
1 1 1
1 1 1
3 3 0
1 1 0
1 1 1
1 1 1
2 3 1
3 1 1
5 4 1
1 1 1
2 1 1
1 1 1
4 4 1
4 4 1
1 1 1
3 2 1
5 5 1
2 1 1
1 1 1
2 2 1
1 1 1
1 1 1
end
label values courtesy revV06538
label def revV06538 1 "(IF VOL) NEVER", modify
label def revV06538 2 "LESS THAN ONCE A YEAR", modify
label def revV06538 3 "A FEW TIMES A YEAR", modify
label def revV06538 4 "A FEW TIMES A MONTH", modify
label def revV06538 5 "AT LEAST ONCE A WEEK", modify
label def revV06538 6 "ALMOST EVERYDAY", modify
label values disrespect revV06539
label def revV06539 1 "(IF VOL) NEVER", modify
label def revV06539 2 "LESS THAN ONCE A YEAR", modify
label def revV06539 3 "A FEW TIMES A YEAR", modify
label def revV06539 4 "A FEW TIMES A MONTH", modify
label def revV06539 5 "AT LEAST ONCE A WEEK", modify

The variables here are ordinal. Thus, in addition to the rcap, I wanted to know if it's possible to have the mean for each 'over' group (in the above, hispanic or non-hispanic) labeled above the bars. Thanks again!

[Zach Goldberg]

By mean for each over group I'm referring to the groups means on the y variable (not the number of observations per group)

[Nick Cox]

That's not wired in. As perhaps implied in the help, ciplot is not a command I have interest in extending as I believe that there is much more flexibility in other approaches. I will be away from my computers for a while, but if you don't get another better answer I will try to look at this next week.

[Zach Goldberg]

Got it. What's the easiest approach you'd recommend (you can just link me to the relevant thread if you don't feel like elaborating)?

[Nick Cox]

The help for ciplot (SSC) points to my paper on "The statsby strategy" (now accessible at https://journals.sagepub.com/doi/pdf...867X1001000112) which is a natural here.


With your dataset example, I proceed (and naturally all kinds of different small choices are possible):

save tokendata, replace 

statsby , by(hispanic) : ci means courtesy
gen which = "courtesy" 
save ciresults, replace 

u tokendata 
statsby , by(hispanic) : ci means disrespect 
gen which = "disrespect" 
append using ciresults 

encode which, gen(Which) 

gen toshow = string(mean, "%3.2f") 
gen yshow = 2.9 

set scheme s1mono 
twoway scatter mean hispanic, xsc(r(-0.2 1.2)) yla(, ang(h)) ///
|| scatter yshow hispanic, ms(none) mlabpos(0) mlabsize(medium) mla(toshow) ///
|| rcap ub lb hispanic , by(which, note("") legend(off)) xla(0 "no" 1 "yes")