Posted on 11/04/2004 12:20:44 PM PST by yldstrk
exit polls.
We benefitted from this attempt. Even the reasonable pundits say the first exit polls should be ignored, that they are extremely unreliable. IMO, it motivated our base.
1. They're, um, POLLS. Polls have sampling errors or they wouldn't be polls.
2. They are subject to any number of biases: geographic (which locations did they happen to exit-poll?), personal (they polled way more women than men for some reason), time (an exit poll in the a.m. is by definition polling only the people who vote in the morning, not those who vote after work), etc.
3. They will necessarily be based on some *model* of the overall electorate. If they poll some blacks and their *model* says blacks are 60% of all voters then they'd always show the (D) way on top. But that model would be wrong. It looks like the model was wrong yesterday too - not as wrong as mine, but still wrong.
So my solution: disregard "exit polls" in the first place. They don't mean squat.
I guess it all depends on whether exit polls are believed to be the last word, or just an initial fuzzy estimate. I took them to be the latter. I guess there might be some who think of them as the former, but perhaps they're more likely to be Democrats... because you have to be kind of stupid to place so much stock in an a.m. exit poll ;-) In other words, if this *was* a (D) trick, I'm not sure it didn't backfire on them....
I read that the polls in NC showed women were 69% of voters at one point. Then the dems wondered why the gap in the state was so much larger than what the exit polls had predicted.
Has college education really fallen off that much since I took statistics in grad school? You'd think a basic reality check would be in order?
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.