I tend to put things in practical terms, even though I have a major in mathematics and can talk with these people a bit.
The practical issue with a survey is that you don't know where the returns are going to come from; you don't know how representative they are.
Put it this way: Charlottetown, within P.E.I., may not be identified with enough returns to be able to say Charlottetown vis-à-vis P.E.I. Indeed, P.E.I. may not have enough returns to be identified other than in and of the Atlantic or Maritime region. That's what a survey means, as against a census that makes sure there is the distribution of returns to get representation.
In simple terms, that's the difference.