Collapse a dummy variable

[fai ot]

Dear All,
I am trying to collapse my variable of interest by country and year. I have a dummy variable that takes a 1 if the variable of interest is grant and 0 if it's a credit. After collapsing data , the dummy variable gets values of 0.97 ... 0.5 .... . Is there a different way to keep values 0/1.

This is the code:

collapse (sum) Var of interest (mean) DummyVar, by(Year Country)

Many thanks,
Fairouz

[Andrew Musau]

The mean of a 0/1 variable will be zero if all observations in the group are 0 and one if all observations are 1. Otherwise, it will be a fraction between 0 and 1. If you want to impose a rule where the result is one if at least one observation equals one, and zero otherwise, you can use:

collapse (sum) Var of interest (max) DummyVar, by(Year Country)

[Nick Cox]

Perhaps like Andrew Musau I am not clear quite what you were expecting from a collapse of a {0, 1} variable. I will add that if you keep a count too, then the mean tells you how many 0s and 1s you have in the data. Further, keeping the minimum too tells you if there are any zeros.