January 08, 2004

Contra polls

With the election season heating up, I thought it might be a good time to criticize the predictable obsession with polling that seems to infect everyone before each round of actual voting.

Polls seem to be an especially attractive target for the best bloggers. Here's The Decembrist; here's Talking Points Memo; here's Calpundit. Conservative bloggers love 'em too: here's Instapundit and Andrew Sullivan.

What do these polls really mean? Not much. They're undoubtedly useful for campaign managers as they try to assess whether voters might respond to their candidate's message. In the worst cases, a candidate may even choose to change her position on an issue or two in order to get a bump in the polls. But if you're not a campaign manager or a candidate, these political polls are like the daily odds line for sporting events in the newspaper. In other words, they're pretty much meaningless.

Poll results don't determine anything. They're the classic example of "made-up" news; something to talk about when you're impatient for the real election and want to root for your candidate like you root for a boxer, or for a racehorse.

Sometimes, though, polls can be damaging. The metaphor of a political campaign as a horse race is usually harmless, except when people forget that it's a metaphor. Campaigns aren't horse races, they're an exposition and an explanation of a candidate so that voters can decide whom to vote for in an election.

Until the actual election, voters might be smitten first with this candidate and later with that one, but none of these opinions really matters: only when the voter walks into the booth and realizes that he only has one vote, and that he has to synthesize all his previous infatuations with his rational calculations in order to cast that one vote correctly, does his professed support for a candidate actually matter.

But these calculations are too often swayed by otherwise meaningless poll information. The voter really likes candidate A, but he's read in the papers again and again that the polls show candidate B with a "commanding lead" over A. Mistaking the horse-racing metaphor for reality, the voter decides not to "waste" his vote, and votes for B. All on the basis of meaningless poll data.

The poll has perversely made itself meaningful by masquerading as a tangible fact about the campaign. It has erroneously proclaimed one candidate "far ahead" of another, when the candidates were never actually "racing."

In our politically tuned-out society, with rates of voter participation among the lowest of any democracy in the world, this distorting emphasis on polls becomes even more destructive. Many voters choose to stay home when they hear of a poll showing their candidate either "way ahead" or "way behind." But in fact, all candidates are standing in the same place until the actual election pulls one of them out from the crowd and crowns her an "elected official." There never was any such thing as a "front-runner."

But yet, thanks to an obsession with polls, many voters have altered their votes or decided not to cast a vote at all because of meaningless poll data. All the bloggers cited above know this.

I don't want to say that polls aren't interesting; I would get impatient if I couldn't get some hint of whether or not Clark's tax plan was "catching on," or if Dean was "inspiring people," or if Lieberman was "putting voters to sleep." But let's remember that polls aren't very meaningful, and they can be dangerous. For the sake of running a good election, let's not give any more weight to pre-election polls than they deserve.

Posted by Carey at January 8, 2004 11:26 PM