Just some fun cross tabs:
Trump wins 18-39 year olds
Trump wins 40-54 year olds
The black vote breaks down 73/18 for Clinton - really?
Clinton only wins 77% of Democrats - really?
They are tied using the most strict voter screen (extremely likely): 46-46, but when you look open it up to very likely, Clinton 50-37 and somewhat likely, Clinton 50-16.
Trump winning college graduates 54-39
What kind of bullsh**t LV screen is this?
How is the difference that big between "extremely likely" and "very likely"?
No, in the sample, extremely likely voters are 46-46 - I think that's what they are reporting. But the very likely sample is Hillary 50-37, and somewhat likely sample is 50-16 Clinton...I think they excluded very likely and somewhat likely and just used the extremely likely sample
It's obvious that the "extremely likely" voters are, well, more likely to vote than the other two categories. But it's ridiculous to think that
nobody in the lower categories will vote. Perhaps they should weight the results by likelihood, e.g. 1.0 for extremely likely, 0.7 for very likely, 0.3 for somewhat likely, etc. (Don't quibble about the specific weights; I just picked them for example purposes.)