The Kerry Spot notices some oddness with the newest Pew poll showing a nearly tied race. Any time you get double-digit shifts in your polling between two datapoints, it’s pretty clear something in your methodology is screwed up. If I got such disparate results from an experiment I’d probably have to start again with a new datapoint to confirm which one was the most accurate. I really don’t think that the results shifted that much between the two “waves” of polling, which is why I don’t consider the Pew poll to be accurate or valuable.
Also, the Harris poll is an “interactive” online poll, which are by nature completely worthless. Harris is a decent firm for consumer surveys, but their political polling lags behind.
My personal favorite poll, and one of the most accurate, isn’t a traditional poll at all, but the Iowa Electronic Market. Being a good Hayekian, I tend to believe that distributed systems in which each participant has a personal stake in an accurate outcome tend to be better than opinion polling. The IEM’s Winner Takes All market for this race shows an absolute blowout for Bush and Kerry cratering.
As always, polling is as much an art as a science, and they’re only so useful except in measuring trends. Still, all the signs point to Bush having a major advantage over Kerry – an advantage that will hopefully persist through Election Day.
Dispute the facts all you want. Your boy’s going down. :))