Graph*of mean and median of two variables over two variables

[Stephen Okiya]

Hi Stata users,

I have two variables that I would like to compare their mean and median - var1 and var2 over two categorical variables - cat1 and cat2

The code am using is

graph bar (mean) var1 (mean) var2 (median) var1 (median) var2, over(cat1) over(cat2)

but am wondering whether there is a better way of presenting this information through graphs

The data example is below

* Example generated by -dataex-. For more info, type help dataex
clear
input float cat1 byte cat2 float(var1 var2)
2 5   3     9
2 2   .   -25
2 5   .     0
2 5   .   298
1 5   .   363
2 5   .    51
2 1 155   331
2 4   .    51
2 4  24    37
. 2   .     4
2 2   0    -5
2 1   9    63
2 2 295   610
2 1   .    21
2 2  18    93
2 5   1     3
2 1 -68   109
2 5 220   302
2 1   .  -120
2 1   8    19
2 3  23    65
2 1   .    72
2 2  14   293
2 1  32    31
2 1   .   692
2 1   .   -28
2 5  65   101
2 4   .     9
1 5   .   163
2 5  44   143
2 4   .     1
2 3  36   192
2 5   .     0
2 5  99   367
2 5   .    47
2 1 295   610
2 4  11     1
2 2 168   270
. 5   .    64
2 2   .    82
2 2  23   102
2 4   .    11
2 1   .    66
2 1   .    67
1 5   .    91
2 1 101    20
2 3 385   847
2 3  49    12
2 3 295   385
. 4   .   -20
2 3   . -1389
2 3   .   539
2 3 299   -19
2 4   .   404
2 5  84   217
1 5   .   382
2 4   .   238
2 2   .   -33
2 4  23    88
2 4   0     0
. 3  40   153
2 2  46     8
2 5  12   100
2 5   6   198
2 5  35    91
2 1  20    43
2 2  60    80
2 2  62   196
2 3  28   129
2 2  62    99
2 4 150   491
2 2  23   136
2 3 433   557
2 3  40    53
2 2 579   364
2 5 114   218
2 1   .   144
2 4 163     .
1 3  10   459
2 1  10    25
2 4  81   116
2 4   .   -12
2 4   .   399
2 1  27   189
2 2   0     .
2 4  45   359
2 1   .   198
2 5   .    92
2 1  14   132
2 2 124   161
2 4   .   -75
2 1   .    25
2 5 -27   -26
2 3   7   123
2 3 350   594
2 5   1     3
2 3 219     .
2 3   .    59
2 3 225   600
2 2 -99   -84
end
label values cat1 cat1
label def cat1 1 "Low", modify
label def cat1 2 "Medium", modify
label values cat2 cat2
label def cat2 1 "Q1", modify
label def cat2 2 "Q2", modify
label def cat2 3 "Q3", modify
label def cat2 4 "Q4", modify
label def cat2 5 "Q5", modify

Thanks in advance!

[Nick Cox]

Many possibilities. Here's one. (Hoping that the graph for the full dataset is more interesting than this!)

foreach v in var1 var2 { 
    foreach s in mean median { 
        egen `v'_`s' = `s'(`v'), by(cat1 cat2)
        label var `v'_`s' "`s' of `v'"
    }
}

gen cat2_1 = cat2 - 0.1 
gen cat2_2 = cat2 + 0.1 

scatter var1_* cat2_2,  ms(Oh X) || scatter var2_* cat2_1, xla(1 "Q1" 2 "Q2" 3 "Q3" 4 "Q4" 5 "Q5") by(cat1, note("")) ms(+ Th)

[Stephen Okiya]

Thanks so much Nick Cox for the proposed approach, I appreciate you taking time and sharing a solution