[Mohamad Cea]

These are different variables. But zeros in the table would be problematic. You have a choice. You can either pursue #10 in which case #11 and #13 still apply; or pursue #14 in which case we need to see (1) the graph code you tried and (2) the results of

contract , zero
dataex

with the two variables in #14. So if the variables were a b (they aren't) you need to go

contract a b, zero
dataex

See also https://www.statalist.org/forums/help#stata for how to present data examples here.

Thanks for your patience with me!

* Example generated by -dataex-. To install: ssc install dataex
clear
input double(Q2061A Q609) int _freq
1 1 3327
1 2 4949
1 3 1144
1 98 124
1 99 41
1 . 172
2 1 1310
2 2 2131
2 3 444
2 98 47
2 99 22
2 . 574
3 1 172
3 2 275
3 3 78
3 98 7
3 99 2
3 . 10
6 1 417
6 2 649
6 3 154
6 98 12
6 99 1
6 . 92
7 1 721
7 2 1128
7 3 206
7 98 18
7 99 8
7 . 70
8 1 280
8 2 357
8 3 99
8 98 11
8 99 5
8 . 13
11 1 666
11 2 705
11 3 254
11 98 18
11 99 4
11 . 9
12 1 911
12 2 1197
12 3 255
12 98 24
12 99 5
12 . 270
13 1 446
13 2 567
13 3 123
13 98 16
13 99 1
13 . 12
14 1 317
14 2 458
14 3 162
14 98 13
14 99 6
14 . 10
15 1 261
15 2 385
15 3 118
15 98 40
15 99 8
15 . 58
98 1 103
98 2 96
98 3 39
98 98 14
98 99 0
98 . 61
99 1 22
99 2 9
99 3 5
99 98 0
99 99 0
99 . 41
. 1 0
. 2 1
. 3 0
. 98 0
. 99 0
. . 0
end
label values Q2061A Q2061A
label def Q2061A 1 "economic situation", modify
label def Q2061A 2 "financial and administrative corruption", modify
label def Q2061A 3 "democracy and representation/governance", modify
label def Q2061A 6 "internal stability and security", modify
label def Q2061A 7 "foreign interference", modify
label def Q2061A 8 "religious extremism", modify
label def Q2061A 11 "fighting terrorism", modify
label def Q2061A 12 "public services", modify
label def Q2061A 13 "security", modify
label def Q2061A 14 "political/party issues", modify
label def Q2061A 15 "other", modify
label def Q2061A 98 "don't know", modify
label def Q2061A 99 "refused", modify
label values Q609 Q609
label def Q609 1 "religious", modify
label def Q609 2 "somewhat religious", modify
label def Q609 3 "not religious", modify
label def Q609 98 "don't know", modify
label def Q609 99 "refused", modify

And here's the code:


. bysort Q2061A Q609: gen freq = _N if Q609 == 1
(70 missing values generated)

. by Q2061A (Q609) : replace freq = -freq[_N]
(14 real changes made, 14 to missing)


. egen group = group(freq Q2061A)
(84 missing values generated)


. labmask group, values(Q2061A) decode
no observations


When I used this command: by Q2061A (Q609) : replace freq = -freq[_N] with a binomial variable like gender, things worked smoothly, but now I end up replacing the variables with missings.

[Nick Cox]

Thanks for the data example, but for anyone watching

(1) the data example is what you would see if you ran contract Q2061A Q609, zero

(2) the code following is for a differentset-up.

As explained in https://www.statalist.org/forums/help#stata data examples should be presented as CODE using the delimiters provided by dataex or as provided by the # button in the forum software.

This code takes (1) and orders the rows of the plot according to the frequencies of people answering 1 to Q609 (highest value at the top).

clear
input double(Q2061A Q609) int _freq
1 1 3327
1 2 4949
1 3 1144
1 98 124
1 99 41
1 . 172
2 1 1310
2 2 2131
2 3 444
2 98 47
2 99 22
2 . 574
3 1 172
3 2 275
3 3 78
3 98 7
3 99 2
3 . 10
6 1 417
6 2 649
6 3 154
6 98 12
6 99 1
6 . 92
7 1 721
7 2 1128
7 3 206
7 98 18
7 99 8
7 . 70
8 1 280
8 2 357
8 3 99
8 98 11
8 99 5
8 . 13
11 1 666
11 2 705
11 3 254
11 98 18
11 99 4
11 . 9
12 1 911
12 2 1197
12 3 255
12 98 24
12 99 5
12 . 270
13 1 446
13 2 567
13 3 123
13 98 16
13 99 1
13 . 12
14 1 317
14 2 458
14 3 162
14 98 13
14 99 6
14 . 10
15 1 261
15 2 385
15 3 118
15 98 40
15 99 8
15 . 58
98 1 103
98 2 96
98 3 39
98 98 14
98 99 0
98 . 61
99 1 22
99 2 9
99 3 5
99 98 0
99 99 0
99 . 41
. 1 0
. 2 1
. 3 0
. 98 0
. 99 0
. . 0
end
label values Q2061A Q2061A
label def Q2061A 1 "economic situation", modify
label def Q2061A 2 "financial and administrative corruption", modify
label def Q2061A 3 "democracy and representation/governance", modify
label def Q2061A 6 "internal stability and security", modify
label def Q2061A 7 "foreign interference", modify
label def Q2061A 8 "religious extremism", modify
label def Q2061A 11 "fighting terrorism", modify
label def Q2061A 12 "public services", modify
label def Q2061A 13 "security", modify
label def Q2061A 14 "political/party issues", modify
label def Q2061A 15 "other", modify
label def Q2061A 98 "don't know", modify
label def Q2061A 99 "refused", modify
label values Q609 Q609
label def Q609 1 "religious", modify
label def Q609 2 "somewhat", modify
label def Q609 3 "not", modify
label def Q609 98 "don't know", modify
label def Q609 99 "refused", modify

* bysort Q2061A Q609: gen freq = _N if Q609 == 1

* by Q2061A (Q609) : replace freq = -freq[_N]
* egen group = group(freq Q2061A)
* labmask group, values(Q2061A) decode

set scheme s1color 

egen tosort = total(cond(Q609 == 1, -_freq, .)) , by(Q2061A) 
egen order = group(tosort Q2061A) 
labmask order, values(Q2061A) decode 

tabplot  order Q609 [fw=_freq], yla(, labsize(small)) bfcolor(none) xla(, labsize(small)) ytitle("") xtitle("") subtitle("") showval(offset(0.2))

[Nick Cox]

#17 can now be reproduced more briefly making use of myaxis from SSC, which will be announced separately.

clear
input double(Q2061A Q609) int _freq
1 1 3327
1 2 4949
1 3 1144
1 98 124
1 99 41
1 . 172
2 1 1310
2 2 2131
2 3 444
2 98 47
2 99 22
2 . 574
3 1 172
3 2 275
3 3 78
3 98 7
3 99 2
3 . 10
6 1 417
6 2 649
6 3 154
6 98 12
6 99 1
6 . 92
7 1 721
7 2 1128
7 3 206
7 98 18
7 99 8
7 . 70
8 1 280
8 2 357
8 3 99
8 98 11
8 99 5
8 . 13
11 1 666
11 2 705
11 3 254
11 98 18
11 99 4
11 . 9
12 1 911
12 2 1197
12 3 255
12 98 24
12 99 5
12 . 270
13 1 446
13 2 567
13 3 123
13 98 16
13 99 1
13 . 12
14 1 317
14 2 458
14 3 162
14 98 13
14 99 6
14 . 10
15 1 261
15 2 385
15 3 118
15 98 40
15 99 8
15 . 58
98 1 103
98 2 96
98 3 39
98 98 14
98 99 0
98 . 61
99 1 22
99 2 9
99 3 5
99 98 0
99 99 0
99 . 41
. 1 0
. 2 1
. 3 0
. 98 0
. 99 0
. . 0
end
label values Q2061A Q2061A
label def Q2061A 1 "economic situation", modify
label def Q2061A 2 "financial and administrative corruption", modify
label def Q2061A 3 "democracy and representation/governance", modify
label def Q2061A 6 "internal stability and security", modify
label def Q2061A 7 "foreign interference", modify
label def Q2061A 8 "religious extremism", modify
label def Q2061A 11 "fighting terrorism", modify
label def Q2061A 12 "public services", modify
label def Q2061A 13 "security", modify
label def Q2061A 14 "political/party issues", modify
label def Q2061A 15 "other", modify
label def Q2061A 98 "don't know", modify
label def Q2061A 99 "refused", modify
label values Q609 Q609
label def Q609 1 "religious", modify
label def Q609 2 "somewhat", modify
label def Q609 3 "not", modify
label def Q609 98 "don't know", modify
label def Q609 99 "refused", modify

set scheme s1color

myaxis order=Q2061A , sort(mean _freq) subset(Q609==1) descending
tabplot  order Q609 [fw=_freq], yla(, labsize(small)) bfcolor(none) xla(, labsize(small)) ytitle("") xtitle("") subtitle("") showval(offset(0.2))