using table for a three way cross-tabulation

[Maarten Buis]

I am trying out the new table command, to create a three way cross tabulation (is there a difference between East and West Germany in the opinions on the role of women after one controls for religious affiliation). I got far, but there are still two problems where I am stuck.

Within each panel I am showing column percentages. I would like to emphasize this by adding column totals to each panel (they would be all 100%)

The variable labels for relig (religious affiliation) and husb_career (wife should support husband's career) are not necessary as the value labels contain all the relevant information and I would like to remove them from the table

. clear

. input byte(eastwest husb_career relig) int freq

     eastwest  husb_career     relig      freq
  1. 1 0 0  2177
  2. 1 0 1 10582
  3. 2 0 0  2396
  4. 2 0 1   892
  5. 1 1 0   509
  6. 1 1 1  5236
  7. 2 1 0   590
  8. 2 1 1   387
  9. end

. label values eastwest eastwest

. label def eastwest 1 "West", modify

. label def eastwest 2 "East", modify

. label values husb_career husb_career

. label def husb_career 0 "wife can prioritize own career", modify

. label def husb_career 1 "wife should support husband's career", modify

. label values relig relig

. label def relig 0 "non-religious", modify

. label def relig 1 "religious", modify

. label var eastwest "Region of Germany"

. label var husb_career "wife should support husband's career"

. label var relig "religious affiliation"

.
. qui table (relig var)  (eastwest)  [fw=freq], ///
>  stat(fvpercent husb_career) nformat(%9.0f) sformat("%s%%")

.  
. collect style row stack, nobinder spacer

.
. collect preview

-------------------------------------------------------------------
                                           |    Region of Germany  
                                           |   West    East   Total
-------------------------------------------+-----------------------
religious affiliation                      |                       
  non-religious                            |                       
    wife should support husband's career   |                       
      wife can prioritize own career       |    81%     80%     81%
      wife should support husband's career |    19%     20%     19%
                                           |                       
  religious                                |                       
    wife should support husband's career   |                       
      wife can prioritize own career       |    67%     70%     67%
      wife should support husband's career |    33%     30%     33%
                                           |                       
  Total                                    |                       
    wife should support husband's career   |                       
      wife can prioritize own career       |    69%     77%     70%
      wife should support husband's career |    31%     23%     30%
-------------------------------------------------------------------

[German Rodriguez]

Hi Maarten Buis. The -percent- statistic has a -total- option that may be helpful. And -collect style header- may be used to hide the variable labels. Not sure the -binder- was needed, but it doesn't hurt. So using your data I tried the following

table (relig husb_career)  (eastwest)  [fw=freq], ///
  stat(percent, across(husb_career)) nformat(%9.0f) sformat("%s%%")
collect style header relig, title(hide)
collect style header husb_career, title(hide)
collect style row stack, spacer
collect preview

which produces

---------------------------------------------------------------
                                       |    Region of Germany  
                                       |   West    East   Total
---------------------------------------+-----------------------
non-religious                          |                       
  wife can prioritize own career       |    81%     80%     81%
  wife should support husband's career |    19%     20%     19%
  Total                                |   100%    100%    100%
                                       |                       
religious                              |                       
  wife can prioritize own career       |    67%     70%     67%
  wife should support husband's career |    33%     30%     33%
  Total                                |   100%    100%    100%
                                       |                       
Total                                  |                       
  wife can prioritize own career       |    69%     77%     70%
  wife should support husband's career |    31%     23%     30%
  Total                                |   100%    100%    100%
---------------------------------------------------------------

Is this what you meant?

[Maarten Buis]

German Rodriguez 'Yes, that looks good. Thank you.