I think you'd then be skewing towards a sample that spends a lot of money on lottery tickets (certainly not something I do but maybe I'm the minority), and your

I think you'd then be skewing towards a sample that spends a lot of money on lottery tickets (certainly not something I do but maybe I'm the minority), and your sample size would be pretty small, and you wouldn't see how it affects a community.