GERAGHTY: When a politician criticizes your polls, and your poll ends up being pretty close to the final result, do you ever feel like calling him up and saying, “Neener-neener-neener”? It seems that no matter how closely you track to the actual election results, every year it’s back to square one, ”Oh, we can’t trust that guy.”
RASMUSSEN: You have to have a thick skin to be in this business. You have to realize it’s nothing personal. In 2006, [then-senator] Conrad Burns [of Montana, a Republican] attacked our credibility because we showed him in trouble in January. He just didn’t believe it. Then, of course, it turned out to be right. We had the same thing happen with [North Carolina Republican senator] Elizabeth Dole in 2008.
It is part of the process, and you have to understand that people don’t care about polls, they care about what we poll about. The politicians are not truly commenting on the credibility of the poll; they’re trying to spin their version of the story any way they can. If attacking the pollster helps them out, well, that’s what they’ll do.
GERAGHTY: Pollsters weight by age, race, and gender, and sometimes by geography. How important is it to weight by party?
RASMUSSEN: There is a legitimate industry discussion on that. We all know that if you can tell what the partisan makeup [of the electorate] is going to be, it is a great indicator of how the results will turn out. What most pollsters struggle with is, what are the appropriate targets? . . . We do a lot more polling than just about anybody else, so we have enough data that we can provide some ongoing estimates of where the party trends are.
Whether or not you weight by party, it is certainly reasonable for an analyst to look at the results and say, “You know, I just don’t believe that there will be more Democrats turning out in 2010 than in 2008, and I ought to look at that when I consider the results.”
GERAGHTY: Have you ever re-polled a race after getting results that didn’t sit well with your gut?
RASMUSSEN: We release the data and then we go poll it again if something is going on. There are two races right now where I am very curious about what our next polls will show, one in West Virginia and one in Alaska. In both cases, I can come up with a logical argument as to why the numbers are the way they are; I can also come up with a logical argument as to why they show the race as closer than it really is. But we want to get the information out there and let other people engage in that discussion, and we’ll poll again and see where it ends up.