putting multiple graphs for quantitative vs categorical variables in one graph

[Ashley Liu]

Hi,

I have 7 variables, which are actually for one survey question that asks respondents to rank the importance of 7 options. The data were entered in a way that separates the seven choices as 7 variables and each row was a ranking for the separate variables.

the data looks like this:

variable name:	health	repro	breed	meat	nutr	care	rankoth
variable label:	Animal Health and Well Being	Reproduction	Animal Breeding/Genetics	Meat Science	Nutrition	Management and daily care of animals	Others

ID: 1	3	1	2	6	4	5	7
ID: 2	3	1	2	7	5	4	6

That way, I need all seven variables on one graph for it to make sense. For instance, the simplest graph I want is just displaying the means of these 7 variables on one graph, probably as a bar chart.
Further analysis would require a side-by-side bar chart that displays the mean ranking (y) for each of the seven variables (x) by gender (so there are two bars).

In short, I just want to put multiple 2-way graphs between "1 categorical variable and (mean of) 1 quantitative variable" of multiple quantitative variables on one graph.

This is something I want:

is there any way to do that in stata? Thanks!

[Nick Cox]

Please read and act on https://www.statalist.org/forums/help#stata in presenting data examples. Yours required a lot of surgery and a new variable to get to this:

* Example generated by -dataex-. To install: ssc install dataex
clear
input byte(id health repro breed meat nutr care rankoth) float girl
1 3 1 2 6 4 5 7 0
2 3 1 2 7 5 4 6 1
end
label values girl girl
label def girl 0 "boy", modify
label def girl 1 "girl", modify

statplot from SSC can help here.

ssc install statplot

statplot health-rankoth , over(girl) recast(bar) varnames

[River Huang]

Dear Nick, Following your example above, is it possible to have different colors for boy/girl? Thanks.

[Andrew Musau]

The usual way to do this in graph bar works.

statplot ... ,asyvars showyvars leg(off)

[River Huang]

Dear Andrew, Many thanks for this helpful suggestion.

[River Huang]

Dear Andrew, By the way, I have this dataset and code

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(lly btot privo) byte income4
 41.0236  83.9456  65.4788 1
 66.9684  95.7224 57.85528 2
  78.022  90.3459 55.01267 2
 74.4912    92.72 62.01511 3
 53.3381  74.3563 36.72324 3
125.9392  96.7181 128.3783 3
 70.3053  92.7402 65.04914 3
 54.7421  94.7271 49.14149 3
 51.9463  84.2815 37.43197 3
77.47569  87.7728 59.08989 3
 31.7611  88.9397 34.79329 3
 67.4966  98.4359 65.29842 4
  45.349  97.2227 51.78182 4
 49.0197  92.0092 25.64779 4
 54.0413  90.0224 81.62302 4
   63.37  96.5411 75.46823 4
71.41078  98.0962 86.69409 4
57.46272   97.571 76.45579 4
 49.4787  88.0999 42.44699 4
 48.6267  83.5528 46.31212 4
56.50429   88.997 60.86423 4
 53.4937  88.9401 89.11234 4
 51.7266  92.6655 54.81786 4
  49.626    82.43 37.59236 4
123.4055  98.9879 141.2944 4
 62.1195  93.1142 113.0658 4
143.4259   92.565 43.97008 2
 63.7392  96.4534 46.89995 2
 19.1631  61.6758 21.44617 2
 51.4366  94.7736 71.93586 3
 51.5927  91.8502 40.64384 3
 25.5744  69.3824 22.89418 3
 46.8707  82.2124 24.36066 3
 22.9597  52.8085  27.8104 3
 29.4668  59.2383 21.21116 3
 18.3381  74.1596 15.68565 3
 37.4615  91.7183 31.39167 4
 36.8425  90.8998 33.12155 4
 47.7859  84.6563 47.27738 1
 27.5044  81.4047 26.97613 1
 16.3877  31.5974 13.14186 2
 17.6203  65.2917 14.51989 2
 20.5826  73.3352 19.11292 2
 31.0542   89.117  20.8444 2
 30.3415  57.7968 16.20979 2
 26.9358  72.0441 22.84953 2
 20.0486  62.1616 17.98507 2
 43.3227  46.9346   8.8252 2
 33.3723   71.971 40.22082 2
 52.9648   58.018  20.5166 2
 20.2198  75.1588 13.31726 2
 22.4096  80.1646 22.08222 2
 36.8505  78.0906 24.54591 2
 18.5189  86.0365 13.32469 3
 29.3846  72.8176 21.77378 3
 37.4766  96.9463  23.6507 3
    32.9  82.0514 21.88465 1
 16.0277  28.5825  4.08484 1
 14.4336  83.8875 13.05017 1
 20.2658  57.8637  7.72334 1
 38.6757  67.8879 20.76761 1
 35.7355  81.2692 22.78104 1
  9.7347  37.9002 10.16035 1
 32.9494  63.6846 19.53084 1
 16.8288  49.5504  5.06637 1
  17.577  39.7841  5.07367 1
 22.6035  23.7174   7.7135 1
 24.6914  86.3752 13.54584 1
46.92189 75.80044 23.03637 1
 23.0376  76.5371  23.8559 1
 22.7968   84.538 27.50557 1
end
label values income4 temp
label def temp 1 "Low", modify
label def temp 2 "Lower Middle", modify
label def temp 3 "Upper Middle", modify
label def temp 4 "High", modify

statplot lly btot privo, over(income4) recast(bar)  asyvars showyvars leg(off)

with graph

My question is: how to deal with the unreadable labels on the X axis. (One way is below, but how to do that in Stata?)

[Nick Cox]

Two solutions:

statplot lly btot privo, over(income4) recast(bar)  asyvars

statplot lly btot privo, over(income4) asyvars showyvars legend(off)

Arguably this is a more mature graph yet:

collapse lly btot privo, by(income4)

twoway connected lly btot privo income4, xla(, valuelabel) xtitle("")

with extra work needed to add direct labelling | labeling, as discussed e.g. in https://www.stata-journal.com/articl...article=gr0080

[River Huang]

Dear Nick, Thanks a lot. It really helps.

[Ivan Manhique]

Dear Nick, Following your example above, is it possible to have different colors for boy/girl? Thanks. [ATTACH=CONFIG]n1550351[/ATTACH]

Dear all,
Hope you are all fine.

Following the example, is there a way for me to include an additional bar in each category denoting the overall mean (both boys and girls) of each?
Any tip will be well received.

Best regards,
Ivan Manhique

[Ivan Manhique]

Dear all,
Hope you are all fine.

Following the examples, is there a way for me to include an additional bar in each category denoting the overall mean for those graphs?
If any doubt what I want to produce is a graph like the attached to this reply.
Any tip will be well received.

Best regards,
Ivan Manhique