Election polls and the primaries

Jay Cost's blog alerted me to this article in the LA Times...

The title of the Thursday op-ed peice is "The Heisenberg Primaries." The authors write:

Professor Werner Heisenberg postulated that "the more precisely the position is determined, the less precisely the momentum is known." Applied to the presidential race, this suggests that the more we measure how the candidates stand now, the less we may know about where things are going to end up — because the measurement itself can render the findings inaccurate.

This is the way polls always are. They are a stopped clock that's right twice a day. The only accurate polls are the ones taken right before election day, and the pollsters use these to prove themselves. Opinion polls of course can never be tested. And election polls that are taken early in a campaign have no predictive value for where things will end up on election night.

The authors continue:

Consider this: More than two-thirds of the Democrats who voted in the 2004 Iowa caucuses didn't decide who to vote for until a month before the caucuses. Four in 10 decided in the last week. In 2004, 54% of New Hampshire Democrats decided within a week of the primary. It's no surprise, then, that in the 2004 election, John Kerry was lagging in third place until only a few weeks before the Iowa caucuses. Kerry then more than doubled his vote in Iowa and nearly quadrupled it in New Hampshire -- all in less than 20 days.

Iowa's Republican caucus-goers are no different. In 1996, nearly a quarter chose their candidate on caucus night or in the preceding two days; fully 42% decided in the last 10 days. And in New Hampshire, only 12% of Republicans decided in 2000 who they would support in the primary before Jan. 1 of election year.

Amazing, yet to be expected. But then what's the point of paying attention this early? Jay Cost has a brilliant reply to this question:

My intuition is that the polls are simply reflecting the elite dialogue in the nation. Average voters hear that Hillary or Rudy is "up" amidst the dribs-and-drabs of news analysis they acquire, and they toss their support to one of them when queried by a pollster. The source of the response is real - it is just not from the respondent. It is a "sampling" of the elite dialogue. So, I don't think that the poll results are "illusory." They are just something other than what they appear to be at first blush. They do not reflect the views of the electorate per se. Rather, they reflect the elite dialogue on the candidates.

Nevertheless, elite analysis is not idle. It has real value. Political elites of all stripes are trying to gauge these candidates to see who will be the most competitive in both the primary and the general elections. In so doing, they are narrowing the choices down for the voters. This is a needed and valuable civic service.

I don't think that the early contest is for the sake of entertaining the elites. Rather, I think it is a campaign designed specifically for their consumption because they serve an important function in this - and any - presidential election. They decide who is, and who is not, politically viable. They set the agenda. This is not to say that I think this uniquely early start date is an efficient way for elites to accomplish this civic task. They do not need so much time to make up their minds. However, I think it is appropriate and inevitable for the campaign season to begin at least a few months before the average voter starts to give a damn. Elites have a role to play - and their role comes before the average voter's role.

This is really good analysis. Cost is not condoning the choices elites make, but just pointing out the necessity of it. You can't literally have a free-for all democratic selection process, and if you did, it would be heavily weighted towards celebrity and wealth, even more than it is now. Elites of course can make bad choices of candidates, and everyone has to suffer for that, and it can be very frustrating. But the answer is not to cut out the elites; not because it wouldn't make things better but because it is impossible. There will always be the equivalent of party bosses (now its the political elite class in the media rather than the parties) to narrow the selection down to an understandable level. A hundred-plus candidates running is too confusing, and the Arnold Swarcheneggers of the world would come out on top.