POLLS ARE VARIABLE BUT VOTERS ARE PREDICTABLE.

POLLS ARE VARIABLE BUT VOTERS ARE PREDICTABLE. This article by Jay Cost (link via RealClearPolitics), and this article, discuss “Why Are American Presidential Election Campaign Polls So Variable When Voters Are So Predictable?”, an academic study by Gary King and Andrew Gelman. The study addresses the following problem, as Cost states it: “Political scientists have developed models that do a very good job of predicting presidential elections based upon ‘fundamental’ variables like incumbency, partisanship, and the state of the economy. All of these are available a long time before ballots are cast. Meanwhile, the polls run all over the place prior to Election Day. How to explain this?” I am glad to see it pointed out that a lot of the coverage of the horse race focuses on the temporary. In thinking about this, I had always focused on the journalist’s need for a daily story, much like the daily explanations for minor movements in the stock markets. But why do the polls of prospective voters move? All four writers, Cost, Mark Mellman, King and Gelman give the same answer: people are not very good at predicting how they will feel in the future, in this case, how they will vote on election day. Daniel Gilbert makes a similar argument about people being unable to predict their future feelings in STUMBLING ON HAPPINESS, as posted here and here.