BTW, it was 7-11, not Starbucks.

Unlike Tom, I’m not just picking some random number out of the air in a fit of wishful thinking. There are three factors I think might contribute to a surprise big victory for Barack…

…a surprise to the media anyway, because the first factor is the media’s blinding bias towards reporting elections as horse races. If they’re not close, it’s not news. They’re encouraged by the parties, who want their followers to believe the race is close as a motivator for donations, volunteer work, and votes.

Then there’s the growing doubt that traditional polling is producing credible results any more. We’ve gone way beyond the cell phone problem, in terms of being able to connect to a representative sample of the population. Some people have abandoned phone service with a public number or directory listing, and a phone number is no guarantee of location any more. Expatriate Ohioans who keep their cell phones are no doubt getting calls in the other 49 states and the results are lumped in with Ohio’s predictions.

On top of the difficulty of contacting a representative sample, I think the results are increasingly skewed by people who refuse to talk to pollsters, or deliberately lie. Americans have grown fairly sophisticated about the political uses of polling, and I would assume that any pollster I talk to is conducting a push poll and isn’t really interested in my opinion anyway. I also don’t pick up the phone if I don’t recognize the call-waiting number, and possibly I’ve “missed” the opportunity to participate in one or more political polls. I’m trying to filter out telemarketers, but the same techniques filter out pollsters.

But the most compelling reason I doubt the conventional wisdom of a close election comes from “nontradtional”, or “crowdsourced” polling. I’ve tried and failed to find it again, but last week the NYTimes, I think it was, had a short feature on other predictors of the outcome of presidential races: A Starbucks using Romney and Obama cups, tracking requests for both; several bakeries with candidate-themed or party-themed cookies, tracking sales; Halloween mask sales; book sales on Amazon; and several more.

What they have in common is that they use crowdsourcing to learn about mass attitudes, not by asking, but by observing. People routinely lie or spin when answering a question, but their behavior and the choices they make get a lot closer to the truth.

The results produced by crowd-sourcing overwhelmingly favor Obama, with the sole exception of book sales on Amazon, which is known as a cornucopia of right-wing self-published book-length rants. People prefer Obama coffee cups and Obama and donkey-shaped cookies and Obama Halloween masks over equivalent Romney products two to one.

I think Tom and his “300 points” is going to be the Zone’s “Dewey beats Truman!”

If the same thing were occurring naturally the changes would be so slow they might at most contribute 0.05%. Decades are nothing for nature.

I’m no expert and I have not read any scientific papers in many years.

I believe we are single digit contributors. That is enough to justify efforts to do better but we can not totally control it.