Exit Polls, etc

The blogs are rife with speculation that the exit polls in New Hampshire predicted an Obama landslide and that Hillary, utilizing voting-machine hackery, stole the primary.

I've written elsewhere about my belief that such shenanigans occurred in 2004, when exit polls in Ohio and many other states showed Kerry winning but the states went to Bush. Nationally, the exit polls were 6-8 percent off, and in some states as much as 15 percent.

Exit polls, it bears reiterating, are taken AFTER the person has voted. It's "What did you do?" not "What do you think you will do?" Theoretically, they should be accurate, but because they use sampling (13,000 nationally in 2004, under 2,000 in NH last night), they can't be perfect. So there can be mistakes. 2004 doesn't pass the smell test because the mistakes weren't random -- they all favored Bush.

So the bottom line in New Hampshire last night? The exit polls were accurate. The sample stated a 39-37 preference for Hillary, and that's exactly how the real vote turned out.

So we can put away our tinfoil hats, or redeploy them to art projects, until fall.