Enter the latest, greatest application for your handheld or phone, "Red or Blue,"which uses Global Positioning Systems to tell you if you’re in a Republican or Democratic neighborhood. It will also tell you how much the area has donated to either party or pull up information about the top 100 donors surrounding where you stand. It's obvious this would be a great tool for canvassers and fundraisers. And journalists probably won’t have to spend so much time knocking on doors or sorting through spreadsheets to find that one archetypal Republican or typical Democrat for their stories.
But look, my life is full of strangers and the unexpected and I’m stunned by the data gluttons who'd use this. It seems like this country has developed a whole religion for those who want to know everything about other people, clump those strangers together in predetermined groups, and then avoid the groups they don’t like. The world is addicted to prediction (even though it mostly ends up with simple-minded prejudices). And, there’s something creepy when a creative mind comes up with a “Geiger counter” for people, and something wrong about classifying the hell out of the human race.
Polls are a dangerous addiction. Like any other kind of statistics, you can find numbers that clearly support any given position. So as the election gains momentum the polls become ever more constant, more hyped, and (sometimes) more misleading.
With any poll there are questions that must be asked to determine reliability: how many people were polled? Who paid for the poll? Who conducted it? And beyond those basics are countless arcane details that social scientists have devised over the years to ensure non-leading questions and attempt to keep pollster bias out of the results.
So it is certainly worth noting that Gallup, arguably the most trusted of the major polls, uses a skewed sample for all of its polls. Especially when Gallup which has given Bush the largest and most sustained approval "bounce" since his nomination last month, is getting substantial press coverage of these poll results.
While this is an issue of nothing more than the sample makeup, it may have repercussions far beyond fast-fading news headlines. In a post today on left-wing blog The Left Coaster, Steve Soto shows that 40 percent of Gallup's sample is Republican and 33 percent Democrat. Gallup confirms this skew, saying it is based on a prediction of partisan voter turnout on November 2nd. As fellow pollster John Zogby writes in his rebuttal to this methodology, "there is no evidence anywhere to suggest that Democrats will only represent 31% of the total vote this year."
What matters more than methodological discussions is the power these polls have to discourage voters from going to the booths on Election Day. Just as television stations don't air vote projections until after polls have closed nationwide, polls that inaccurately report a widening gap between candidates when most polls show a statistical dead heat can end up becoming a self-fulfilling prophecy. As the papers cover these faulty poll results, readers believe what they're reading, accept the conclusion that inaccurate polls supply, and before long the race is over in the minds before it's even begun in the voting booths. Instead, it's clear that with 46 days until the election, there is still no way to call a winner in this race, despite what the polls may predict.