Group results by 2 categorical variables using dtable

[Meghan Bezerra]

Hello,

I'm using Stata 18 on Windows 10, and trying to make a Table 1 using dtable that will group my results by 2 categorical variables.

Here is a sample of my data:

* Example generated by -dataex-. For more info, type help dataex
clear
input float(pid VISIT age sex_male haz waz bmi bmiz no_scan baseline_age20)
 1 1  16.67123 1  -.9703893 -1.0016696 19.330885  -.6707739 0 0
 1 2  18.10304 1  -.3789278 -.24052253  21.55985 -.13239644 0 0
 1 3 19.247486 1 -.52851653 -.26594642 22.311863 -.10378502 0 0
 2 1 14.658868 1 -1.0376383  -1.283121 17.134506 -1.1481297 0 0
 2 2 16.230312 1  -.6941634  -.7813833 19.127207  -.6346363 0 0
 2 3 17.010616 1  -.7218754   -.743744  19.89619   -.516387 0 0
 3 1  17.78282 0   .3419666   1.509787 28.253456  1.4325557 0 0
 3 2 19.247486 0  .23503217  1.5211222   29.2352   1.449112 0 0
 4 1  17.04621 0  -.7144825  1.4355024 29.809776   1.662931 0 0
 4 2 18.472544 0   -.631711  1.3682972  29.66877  1.5481687 0 0
 4 3  19.40891 0  -.8471276  1.0972363  28.71389   1.378275 0 0
 5 1 12.071432 0  -.5816514 -1.2927265  15.23467 -1.3934557 0 0
 5 2 13.604546 0  .23995127 -.29156068 17.553335  -.6020396 0 0
 5 3 14.691495 0   .8709965   .2140548 19.101557  -.2229791 0 0
 6 1 12.449377 1 -1.6601788  -.8783625 18.626694  .21731845 0 0
 7 1  17.37761 1   .7562768   .4490592  21.61769   .0525817 0 0
 7 2 19.135233 1    .471106  .22970334 22.222223 -.11331387 0 0
 7 3 20.101713 1   .5853121   .4060313 23.015173 .014880114 0 0
 8 1 21.123066 1    2.56719  1.0560222 22.301117  -.2259649 1 1
 8 2  22.69725 1   2.723965  1.9364756 26.472355   .9306833 1 1
 8 3 23.655516 1   2.723965  1.2119988  22.72774 -.07948803 1 1
 9 1  16.60552 0 -1.6689416   .0778455  23.97853   .8446057 0 0
 9 2 17.563787 0  -1.596563  .11658029  24.35961   .8301668 1 0
 9 3  19.11059 0 -1.6119874   .8026524  28.52512  1.3751934 1 0
10 1 14.431622 1  2.3732917  1.0140684  19.12094 -.12046545 0 0
10 2 16.019493 1  2.0195916  1.0367329 20.937077  .13912234 0 0
10 3 16.958595 1   1.991364  1.1388828 22.157425   .3221904 0 0
11 1 16.977762 1  1.0389788 .018491875 19.383274  -.7352784 0 0
11 2 18.565634 1  1.4909267  .11204115 19.903343  -.9232472 0 0
11 3 19.504736 1   1.672129  .05816798  19.84821 -1.1675317 0 0
end
label values VISIT VISIT
label def VISIT 1 "Visit 1", modify
label def VISIT 2 "Visit 2", modify
label def VISIT 3 "Visit 3", modify
label values sex_male sex
label def sex 0 "Female", modify
label def sex 1 "Male", modify
label values baseline_age20 baseline_age20
label def baseline_age20 0 "20 & under", modify
label def baseline_age20 1 "Over 20", modify

And here is the code I've used to make a Table 1 with results grouped by visit:

dtable age i.sex_male haz waz bmi bmiz if no_scan == 0, ///
    by(VISIT, nototals) ///
    sample(, statistics(freq) place(seplabels)) sformat("(n=%s)" frequency) column(by(hide)) ///
    define(minmax = min max, delimiter("; ")) ///
    continuous(age haz waz bmi bmiz, stat(median minmax)) ///
    nformat(%7.2f median minmax) sformat("(%s)" minmax) ///
    title(Table 1. Participant Characteristics)

What I'd like to do, if possible, is have the results shown in the table by visit, by baseline_age20.

Thanks in advance.
Meghan

[Andrew Musau]

You can use the -group()- function of egen to create a single variable that combines these two variables.

egen wanted= group(visit baseline_age20), label
dtable ..., by(wanted) ...

[Meghan Bezerra]

Thanks Andrew!