How do i visualize the relationship between two ordinal variables ?

[Chanhum Yoon]

Hi,
I have trouble visualizing my ordinal variables.
i hope to show the general relationship between my IV&DV in graph.

My first IV is organized in strongly disagree(score 1) to strongly agree(score 10); and my DV is formed in strongly disagree(1) to strongly agree(5)
and second IV is average of score by adding three variables and divided (Mean: 5.6, range 1-7) and DV is same; strongly disagree(1) to strongly agree(5)

since i've tried various ways to make the best graph, i couldn't figure it out yet.

Any suggestions?

Thank you.

[Nick Cox]

Parallel histograms should help. You don't give a data example (see FAQ Advice #12) but rep78 from the auto data is a 5-level ordinal variable.

sysuse auto, clear
histogram mpg, by(rep78, col(1)) discrete frequency scheme(s1color) subtitle(, pos(9) bcolor(none)  place(e) nobox) yla(, ang(h))

[John Mullahy]

Nick's suggestion is a fine one. A different approach I've used in related work—when only 2 variables are of interest—is to use a twoway scatter graph with marker sizes proportional to joint frequencies. For example, try running something like this.

clear
 
set seed 23
 
set obs 100
 
gen x=rpoisson(1)
gen y=rpoisson(1)
 
preserve
 
contract y x
 
scatter y x [fw=_freq], msize(*.5)
 
restore

[Nick Cox]

See also https://www.statalist.org/forums/for...updated-on-ssc for various plots of ordinal response and ordinal predictor. (Here's a vote against the horrible terms or labels DV and IV.)

[Chanhum Yoon]

Both histogram and scatter graph work well with my data and thanks again for your helpful link attached