I’ve been out for most of the day (I’m having one-hour day surgery on Wednesday so you have to go through, like, 20 hours of pre-surgery, time-wasting, nonsense….all at $35 a pop co-pay. It’s a racket, I tell ya. A racket. But I digress….), so I’m just now getting caught up on stuff.
First (and last, since I have nothing else to offer) on the polls. I chatted with Michael Brodkorb a bit ago who noted — justifiably so — his pride at going through years of newspapers to document the record of the Minnesota poll.
My gripe with his graph — which I may not have explained properly — was with the word “prediction.” In fact, I followed a link that someone had posted in the comments section, and then did a Google search, and it really is remarkable how often the word “poll” and “predict” come in the same sentence.
Here’s the thing: I “hate” the concept of polls predicting anything. I don’t think they do. I think they show you what some people said at a particular time and those people — if the methodology is correct — represent a broader spectrum.
I think if you’re really good at reading polls, you might be able to predict something, but I don’t believe the poll itself does that.
Specifically, Michael’s graph does note that it represents the final poll before Election Day, and I’ll grant you that the final poll before election day is likely to have a better relationship with the final electoral results than one a month before, or two months before etc.
I think the best way to use a poll — any poll — is to look at it as a snapshot in time and then compare it to the same poll, which is a snapshot of another time, and look for movement.
For example, at one point, Tim Penny was “leading” in the race for governor, according to a poll. Some months later, he was third. What happened? Tim Penny didn’t do anything wrong, but voters — and this is my interpretation — didn’t want to vote for an Independent because they thought the eventual winner was going to be a Republican or a DFLer and they didn’t want to allow the (pick name of party you don’t like and put it here) Party into office with their “wasted” vote.
That poll showed movement. (And I also think the fact the poll showed movement actually did influence the election because I think it told other people “jump off the Penny wagon bandwagon because other people are”).
The interesting thing about the Senate race polls, whether you embrace Zogby or Rasmussen or the Minnesota poll is this: they ain’t budging. The numbers might be different when compared with one another, but they’re not changing when compared to the same poll.
What does that tell us in this race? Oh, what? You think I know?