Quote:
Originally Posted by Flint
Step one: major polling aggregates predict Clinton victory as about 90-100% certain. (With the exception of Nate Silver's 538 which gave Clinton a 71.4% chance.)
|
To be fair, only an aggregator who said 100% Clinton was actually
wrong. "So you're telling me there's a chance!"
The pre-aggragation polls are more checkable, as their percentage can be compared to counted votes.
You could probably judge the aggregator's results somewhat, by comparing the actual difference between each poll and relevant vote count with the assumed error range and bias applied to it in the aggregator's formula, and see how much that affects the results.
The polls got the nationwide stats pretty accurately, but not some key swing states. Though, if the difference was within their margin of error, there's not much they could have done about that.
Of course, none of that matters on the gut level.