Percentage table by two categorical variables

[Chiara Piazzo]

Dear all,

I have a dataset on the number of visits made to doctors by person along with their occupation and the city they come from. I would like to make a table with the percentage of visits made by each occupation for each city, e.g. percentage of people who never visited doctor in each occupation code for each city.

I realize that this might be trivial but I have been having really having trouble with it. A snapshot of the data is below.

* Example generated by -dataex-. To install: ssc install dataex
clear
input byte id long(visit occ_code city)
1 3 1 3
2 1 2 2
3 2 3 1
4 1 3 2
5 2 1 1
6 3 2 3
end
label values visit visit
label def visit 1 "more than once", modify
label def visit 2 "never", modify
label def visit 3 "once", modify
label values occ_code occ_code
label def occ_code 1 "agri", modify
label def occ_code 2 "manufac", modify
label def occ_code 3 "services", modify
label values city citi
label def citi 1 "Canberra", modify
label def citi 2 "Melbourne", modify
label def citi 3 "Sydney", modify

Should I decode the occupation and city variables to string? Thank you!

Sincerely,

Chiara

[Nick Cox]

Thanks for the data example. Note that "more than once" "never" "once" looks to be in the wrong order.

That won't bite with this:

. bysort city occ_code: egen never_pc = mean(100 * (visit == 2)) 


. tabdisp city occ_code, c(never_pc) 

----------------------------------------
          |           occ_code          
     city |     agri   manufac  services
----------+-----------------------------
 Canberra |      100                 100
Melbourne |                  0         0
   Sydney |        0         0          
----------------------------------------

[Bruce Weaver]

User-written program tab3way might also work.

. *findit tab3way
. *net install tab3way.pkg
. tab3way city occ_code visit, colpct


Table entries are cell frequencies and column percentages
Missing categories ignored

----------------------------------------------------------------------------------------------
          |                                 visit and occ_code                                
          | ----- more than once -----  ---------- never ---------  ---------- once ----------
     city |     agri  manufac services      agri  manufac services      agri  manufac services
----------+-----------------------------------------------------------------------------------
 Canberra |                                    1                 1                            
          |                               100.00            100.00                            
          |
Melbourne |                 1        1                                                        
          |            100.00   100.00                                                        
          |
   Sydney |                                                                1        1         
          |                                                           100.00   100.00         
----------------------------------------------------------------------------------------------

. tab3way city occ_code visit if visit==2, colpct


Table entries are cell frequencies and column percentages
Missing categories ignored

------------------------------
          | visit and occ_code
          | ------ never -----
     city |     agri  services
----------+-------------------
 Canberra |        1         1
          |   100.00    100.00
------------------------------

[Chiara Piazzo]

Nick Cox and Bruce Weaver - thank you so much! Both suggestions work great.

Bruce, really useful program.