Hi Statalist,

I have a survey with two sample populations. One is representative of the general population (frequency 2029), one is representative of a specific target population (frequency 5934). I have to create very simple tabulations of variables that capture questions asked of BOTH the general and target populations. However, I have never done such tabulations including two different weights. Any advice on how I can integrate two different weights into one tabulate command?


Code:
. tab sample

Data Only Variable: |
             Sample |      Freq.     Percent        Cum.
--------------------+-----------------------------------
            General Pop |      2,029       25.48       25.48
Target Pop |      5,934       74.52      100.00
--------------------+-----------------------------------
              Total |      7,963      100.00
Naturally, each sample population comes with its own weights, namely 'weight_gp' for the general population, and 'weight_target' for the target population

Code:
. codebook weight_gp // for general pop analyses

---------------------------------------------------------------------------------------------
weight_gp                                    Post-Stratification weight for Gen Pop (n=2,029)
---------------------------------------------------------------------------------------------

                  type:  numeric (double)

                 range:  [.212,3.052]                 units:  .0001
         unique values:  1,129                    missing .:  5,934/7,963

                  mean:         1
              std. dev:   .328805

           percentiles:        10%       25%       50%       75%       90%
                             .6517     .7857     .9459    1.1671    1.4248


. codebook weight_target // for target pop analyses

---------------------------------------------------------------------------------------------
weight_target              Post-Stratification weight for Qualified Targeted Sample (n=5,934)
---------------------------------------------------------------------------------------------

                  type:  numeric (double)

                 range:  [.0655,7.6575]               units:  .0001
         unique values:  1,647                    missing .:  2,029/7,963

                  mean:         1
              std. dev:   1.27466

           percentiles:        10%       25%       50%       75%       90%
                             .1665      .293     .5625    1.1167    2.1893
I am now looking at a survey question that was asked of both the general and target populations. The question is captured in the variable 'mleave_pew'. codebook & dataex:

Code:
mleave_pew   Following the birth or adoption of a child, do you think MOTHERS should be able 
---------------------------------------------------------------------------------------------

                  type:  numeric (byte)
                 label:  Q14A

                 range:  [1,99]                       units:  1
         unique values:  3                        missing .:  0/7,963

            tabulation:  Freq.   Numeric  Label
                         7,802         1  Yes, mothers should be able to
                                          take leave
                           115         2  No, mothers should not be able
                                          to take leave
                            46        99  Refused


* Example generated by -dataex-. To install: ssc install dataex
clear
input byte mleave_pew
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
end
label values mleave_pew Q14A
label def Q14A 1 "Yes, mothers should be able to take leave", modify
label def Q14A 2 "No, mothers should not be able to take leave", modify
Again my question is, how can I create a simple weighted table using both weights?