How to generate a new binary variable based on individuals having 2 or more of other risk factors

[Dil Singh]

I would like to create a variable based on individuals being classified as "high risk" or "low risk".

This risk stratification would be based on individuals having 2 or more of certain other characteristics or risk factors, such as sex, age, education etc.

Ordinarily if it was looking at 1 or more of any risk factor, my practice would be to generate a new variable and just "replace as I go" with each old variable in turn.

However, how is it possible to do this if an individual is classified as "high risk" based on having 2 or more of a series of variables?

I can't seem to readily find this on this forum or elsewhere online.

[Rich Goldstein]

you don't say enough for a complete answer but here is a sketch:
1. for each such variable, generate a new variable that is equal to 1 if "high" for that variable and 0 otherwise; then
2. use -egen- with the "rowtotal" function to calculate the number of "high's"; if that total is 2 or more, make your new variable high and make it low if it less than 2
see

FAQ
help egen

added in edit: then, of course, you can drop all the interim variables you created if they are no longer needed (though you might want to retain the pen-ultimate variable that adds the number of "high's"

[William Lisowski]

I think you are saying that your normal practice would be something like the following.

generate atrisk = 0
replace atrisk = 1 if sex==2
replace atrisk = 1 if smoker==1
replace atrisk = 1 if education < 12
...

which would leave atrisk as either 1 - if any of the risk factors were met - or 0 - if none were.

Perhps what you seek is

generate atrisk = 0
replace atrisk = atrisk+1 if sex==2
replace atrisk = atrisk+1 if smoker==1
replace atrisk = atrisk+1 if education < 12
...
replace atrisk = atrisk>=2

which would leave atrisk as either 1 - if two or more of the risk factors were met - or 0 - if one or none were.