random cases within the same variable

Hi everyone,

I'm hoping there's a better way of doing what I want to do. Specifically, I have a variable that differentiates between 3 groups, i.e. a, b, & c.

However, I just want to draw a 10% random sample of the "a's" while still keeping b & c in the same file and variable, and without having to create different datasets and then merging them. In other words, is there a more convenient way of drawing a 10% random sample of a's, while still keeping the 100% of b's and c's?

I greatly appreciate any advice/suggestions.

Regards,

Grigoris Argeros
PhD Candidate
Fordham University