We use a sampling methodology that is scientific and probability-based but is intended to be representative nationally, not at the state level. It can be used to study different sub-populations (education, race, gender) at the national level if there are enough cases...
