Obama 58%, Clinton 41% in Wisconsin

Close race, eh? What? You think a 17-point margin is not close? Well, the political prognosticators were calling it even one day before the primary. One poll had Obama with a 5 point lead, another had Clinton with the same. The margin of error was +/- 3% or so, so they called each a tie.

Graciously giving the benefit of the outside range to the poll that predicted Obama, it would have shown 51% to 39%. That's 12 points. That poll could not even entertain the POSSIBILITY that there would be a 17-point spread.

You know, maybe these pollsters aren't looking at the right things, or asking the right questions, or talking to the right people. Or, the pollsters too-confidently think they have great insight and don't inspect their own processes.

I think they are trying to influence the results and/or provide employment for themselves. Releasing any "data" will influence a few people to change their behavior ("why bother voting", "damn! I have to vote to change that!", etc.). Showing a race is close will more likely have the candidates and media coming back for more polls. But showing a race is lop-sided will make it look like there is no longer any drama. Fewer polls are done just to show a huge gap in opinion.

Exit polls were wrong in 2004, and pre-election polls failed in the Wisconsin Primary. Time to stop wasting our time on speculation. Actually, that time has long passed. Now it's time to run the pollsters out of town on a rail. But not before we poll the populace on whether we should first use tar and feathers.