Correct me if I’m wrong. But it would seem to me that all these polls that have this thing close or BO ahead are probably not using correct samples. Don’t most of them question more Democrats than Republicans?
Let’s say they randomly question 2000 in a night based on other demographics like sex, age, income, etc..
They then determine which are republicans, democrats, and others.
Then they take their weightings, in this case 37, 33, 29 and randomly select 37% republicans, 33% democrats, and 29% others until they reach their desired sample size. I believe Rasmussen’s rolling is 1500.
That’s my understanding of how this works. I’m no pollster, so I’m open to correction.