Much of the math behind surveys is based on finite samples taken from infinite populations. It scales really, really well. You just have to make sure your samples aren't totally biased, which is a little more difficult. But there are ways to reduce those problems, if you're aware of how the bias is introduced.
Much of the math behind surveys is based on finite samples taken from infinite populations. It scales really, really well. You just have to make sure your samples aren't totally biased, which is a little more difficult. But there are ways to reduce those problems, if you're aware of how the bias is introduced.