In the context of a particular policy or program, let's take the example of old age security. That's a universally applied program essentially to alleviate poverty in senior populations. When we're looking at how effective that program is, we pull apart the data to understand how it touches different populations within that—singles, seniors, men, married couples, couples living apart with one in an old age home, etc.
So we pull things apart at that level. I'm not sure....