We really don't know who voted or why... but that won't stop the fine-tooth analysis

[cthulu2016]

Last edited Mon Nov 12, 2012, 08:31 PM - Edit history (1)

States do not record how women, African-Americans, and people making $30K-$50K voted. There are no official demographic results.

Our sense of who voted, and why, is based entirely on the exit polling done by news organizations, and there is only one exit poll—a single polling effort funded by a consortium of news organizations. And that has been pared back considerably due to cost.

In 2008 hispanics were 9% of the electorate. In 2012 they were 10% of the electorate. We all "know" that... but it is ridiculous to form conclusions from a 1% difference in a sub-group of a poll... even a large poll.

When there is only one source of information we treat it as gospel. But if there were three exit polling efforts we would compare the three, see how they disagreed, and recognize that the things are useful but not precise in the way the US census is precise.

Again, when there is only one source of information it appears authoritative because it cannot be contradicted.

What percentage of the 2012 electorate was white? We do not know. Our best information is that it was 72% in 2012 and that it was down from 74% in 2008. But we don't know that. Could it have been 73% both years? Sure.

Aside from exit polls, the only source of information on who voted, and how, it post-election polling. Not only is polling sometimes unreliable (eg. Gallup), but we know that people do not tell the truth in such polls. The number of people who tell pollsters they voted for the winner is always different than the actual vote count, so it is garbage-in, garbage-out. Since we know that the universe of people who tell pollsters they voted for candidate X is distorted there is little sense in expecting demographic information about that unreliable population to be magically reliable.

We can see some broad strokes about the electorate but anything that gets down to a few points is probably just guessing.

[cthulu2016]

1. .

[customerserviceguy]

2. Are the results of the exit polls on a searchable database?

I'm specifically interested in the proportions of fundie voters. I'd specifically like to see if there is a "didn't vote in that race" selection for President.

I've long predicted that the evangelicals feared eight years of a Mormon president far more than they feared four more years of the present one. I'd like to look for evidence that I was either right or wrong about this, particularly in the battleground states.

[cthulu2016]

3. I don't know if they are right now

Since they are a 'product' of the consortium I assume they try to make money from the data-seet somehow. (Just being cynical, but why not?)

But maybe the whole thing does become public at some point. I don't know.