Did any of yall catch the story on NPR's Weekend Edition this morning re polling? They talked to a guy who knows the polling racket.
The 1st thing to watch is who was polled. Eligible voters, registered voters, likely voters. The last is probably the most significant.
Then, there should be this disclosure, but there usually is not unless you go searching: The poll is based on a sample of xxxx (likely voters etc). We are yy% sure that the outcome will be candidate A will win over B by a margin of z%. A whole lot of variables as our friend Nimh, who knows about all of this, has talked about on other threads.
Dewey over Truman. A headline based on a bad poll. Dewey "won" based on a telephone poll conducted before the paper went to press. But many Truman supporters did not have telephone service at the time.
An amusing similarity this time around. Many young people no longer have land lines. Only cell phones. They can not be polled. The folks in that racket claim that they have made adjustments to account for that. We shall see.