>Yes, I've noticed that people tend to have way too much faith
>in the precision of these polls. It's important to keep in
>mind that the pollsters are trying to estimate the voting
>patterns of tens of millions of voters with a sample of
>1000-1200; sampling error is unavoidable here. If the poll
>estimate is within a few percentage points of the actual vote,
>that's actually quite impressive.
This year could prove very interesting - some of the AAPORites seem a little nervous. Cell phones, new registrants, intense feelings, what they deem to be unusual volatility from week to week and an unusually wide spread among pollsters, etc.
Doug