|
You're right: if it's within the margin of error, it could swing in any direction of that given margin (usually it's about 3%, but could be as much as 5.5%).
The purpose of these polls is supposed to show a trend. If you look at it that way, they can be quite useful. If you think it's an exact predictor, as most folks seem to, the poll will have less value than used toilet paper. Take Colorado, for an example: the latest polls seem to show Bush ahead, but I believe in my heart -- because of long- and short-term trends -- that Kerry will take it. He won't even need the "Electoral Vote Division" referendum to pass -- which likely won't anyway...
Another example: Ohio. A couple of days ago, no less than THREE polls showed Kerry slightly ahead, but within the MOE. It was still heartening to Kerry supporters because he had been behind there for most of the general election campaign. Today, another poll is out showing Kerry ahead 50-44. That's OUTSIDE the MOE, which suggests Kerry's strength there is VERY real.
That actually leads me to my next point. Kerry's lead in states like Ohio seems to have been built gradually, which is something a person like me can trust more. But look at the national polls that recently had Bush jumping out to 8-13 point leads for NO APPARENT REASON AT ALL! I just can't trust unresonable bounces like this. Directly after the GOP Convention, yes; but NOT directly after the President took CLEAR ass-whoopings in the debates that millions of Americans watched with our own eyes. We can't be THAT blind AND stupid, can we? So now you can get an idea why so many DU-ers were pissed off with pollsters like Gallup: it seems they have an agenda in their polling, and it isn't "to inform the people..."
|