|
a grasp of some of the problems with this poll, but offhand dismiss them & still believe it to be reprentative of American Democrats?
You dismiss these problems by claiming that other polls are inaccurate. Well even if that is true, it does not make THIS poll accurate.
If you are going to trust that a poll that is subject to freeping, residing on a web site that self identifies as "The Flagship of the Left", only available to self-selecting, left leaning, The Nation reading, internet users as an indicative cross-section of America, you are deluding yourself. The problem here is The Nation readers are not a cross section of America.
A large sample size does not modify a polls margin of error proportionally. The point is bad data in, bad data out. Think of it this way. If you stood at the door of a sports event and asked all people that walked through the door if they enjoyed sports you simply could not make any conclusions about the views on sports of the population at large whether you asked 100 of the attendees or 50,000.
An unscientific biased poll only returns unscientific biased results, without regard to how many other bad polls there are out there.
|