Maybe someone can explain to me how these polls sample is determined.
If a pollster randomly calls 1,000 people and of that they get 500 who identify as Dems, 400 who indentify as GOP and 100 as indies, do they then report what the candidate preference was of those 1,000 and then list the breakdown as 50D, 40R and 10I?
Or do they go back and then say they think the electorate this fall will be 45 D, 40R and 15I and then take that percentage of people from the 1,000 thus the result reported may actually only reflect 900 of the people they called?
“Or do they go back and then say they think the electorate this fall will be 45 D, 40R and 15I and then take that percentage of people from the 1,000 thus the result reported may actually only reflect 900 of the people they called?”
That is exactly what they do, they also weigh it by race and other factors.
There are polls that do just that.
There are also pollsters who base their sample on the last Presidential Election year turn out. This is part of where you get the Democrat oversampled. 2008 was an abnormally good year for Dems. A reasonably conservative estimate is a 36D-32R-32I sample is reasonable for a national poll
All of these polls have biases factors. They are weighted by gender, race, age, cell phone use etc.Pollsters who do it for a living try really hard to find a rational sample mix. But even the best polling is an educated guess.
Media pollsters do not bother to try to be accurate. Their polling is push polling designed to generate a headline or validate a preconceived story line. No one goes back and tracks how far off reality they were.
The problem is too many Americans define reality by what the polls tell them to belive. The Leftists know this and exploit it.