Exit Polls, and Why the Primary Was Not Stolen From Bernie Sanders

[brooklynite]

The allegations are remarkably consistent. They go like this: Mr. Sanders did better in the early exit polls than he did in the final result. Therefore, Mrs. Clinton probably stole the election. The exit polls are a sufficient basis to make this determination, in the eyes of the conspiracists, because exit polls are used internationally to detect fraud. They’re supposedly very accurate and “well controlled” (where this phrase comes from, I don’t know). Furthermore, they say, the exit polls were right on the G.O.P. side — confirming the underlying validity of the methodology and raising suspicions about the Democratic vote count.

All of this starts with a basic misconception: that the exit polls are usually pretty good.

I have no idea where this idea comes from, because everyone who knows anything about early exit polls knows that they’re not great.

We can start in 2008, when the exit polls showed a pretty similar bias toward Barack Obama. Or in 2004, when the exit polls showed John Kerry easily winning an election he clearly lost — with both a huge error and systematic bias outside of the “margin of error.” The national exits showed Kerry ahead by three points (and keep in mind the sample size on the national exit is vastly larger than for a state primary exit poll) and leading in states like Virginia, Ohio and Florida — which all went to George W. Bush.

https://www.nytimes.com/2016/06/28/upshot/exit-polls-and-why-the-primary-was-not-stolen-from-bernie-sanders.html