merging variables

[Stacy Williams]

Hello. I have 5 variables in my data file which are binary (dummy variables =1 or 0). I want to put them together by creating a single variable called Medical_visits. So when I will tabulate Medical_visits, I want all those 5 variables to show up. can anyone help me?

[Andrew Musau]

See

help reshape long

e.g.,

clear
set obs 20
set seed 01042022
forval i=1/5{
    gen var`i'= runiformint(0,1)
}
gen obs_no=_n
reshape long var, i(obs_no) j(which)
tab var which

Res.:

. tab var which

           |                         which
       var |         1          2          3          4          5 |     Total
-----------+-------------------------------------------------------+----------
         0 |         9          8         10          8         13 |        48 
         1 |        11         12         10         12          7 |        52 
-----------+-------------------------------------------------------+----------
     Total |        20         20         20         20         20 |       100

[Stacy Williams]

Thank you Andrew! sorry if I was unclear, but I actually want to run a multinomial regression. So I want those 5 variables to be 5 levels (categories) of a new variable called Medical_visits.

[Øyvind Snilsberg]

egen Medical_visits = group(dummy*)

[Nick Cox]

You have two guesses already from highly competent users and I could guess at yet another answer. It would be better if you gave us a data example following

https://www.statalist.org/forums/help#stata

As flagged there, if your data are confidential, a fake dataset with the same structure is more than fine as an example.