Will We Ever Trust Polls Again?

Andrew Sullivan suggests that question today in this very brief post on the numbers for McCain and Obama in South Carolina.

In following the imbedded links and beyond, I stumbled on this mea culpa at Rasmussen Reports. Money quote: “It is hard to remember a time when the polling and expectations were so universally different from what really happened.”

Of course, Rasmussen is not alone in their frustrated attempts to recall such a discrepancy. Granted, there was that Truman-Dewey head-scratcher sixty years ago. But that particular comparison is hardly valid given the strides made in polling science since 1948.

Naturally, some are voicing the almost unthinkable, that it was not a problem with the polls, but with the polling places, i.e., New Hampshire is to the 2008 primaries what Florida was to the 2000 general election. But on that front, I have to agree to with Justin Gardner’s conclusion: Get over yourselves, “not everything is a conspiracy.”

In the end, perhaps the most telling take-away from the Rasmussen sentence excerpted above is this: It’s hard to remember similar discrepancies because they are, in fact, so rare. In other words, the polls and polsters get it right far more often than they get it wrong. And that means we shouldn’t ignore or discount them, rather, we should use them simply for what they are: good but imperfect guides to the largely unpredictable and unfathomable minds of individuals and communities alike.

Besides, where would the fun be, if we could predict the nominees perfectly before the votes were cast? Take it a step further: Why suffer the cost and pain of primaries and elections at all, if Rasmussen, Gallup, Zogby and the rest were able to perfectly divine our souls and tell us what we want?

Polls are polls. Votes are votes. It’s what you do, not what you say. We know this. Next subject, please.