Replies

I can't see how Rasmussen could be accurate. We're tracking 100m voters over 6 months and the results keep rolling back and forth. One candidate has a one pt lead for 6 weeks, then the other assumes a two pt lead for a month. The challenger is up by 3 for three weeks, then the incumbant turns it around & looks like he may pull away...then uh oh the thing evens out again. This poll drives me nuts.

The polls have to be fairly close, but remember they are designed to make news for the most part and not give us a truly accurate picture.

Rassmusen has established a pretty strict determination of who qualifies as a likely voter. The problem with that model is it is static. The actual voters in this election will not be the same as those Rassmusen is using. Other polls are attempting to catch changes in who is likely to vote and they seem to be determining that more GOP s will turn out and less rats will turn out. That seems plausible at this time.

The problem with establishing a too strong test of who a likely voter is, you can end up with samples that contain very strong representation from either end of the spectrum and none from the middle. Imagine running a poll that was comprised of equal parts of Freepers and DU'ers each day. That sample would not fluctuate at all.