Over at The Monkey Cage, John Sides has a quick post following up on my Q & A with him about the predictive value — or lack thereof — of early polling.
Sides passes on the following graph from the researchers Chris Wlezien and Robert Erikson, which tracks the predictive ability of presidential polls from 300 days before a general election right up to the vote. In the figure, which covers the period 1956-2008, a higher r-squared number means a poll had greater predictive value:
Unsurprisingly, polls generally become better predictors as the election comes nearer. But beyond that, Sides notes:
The graph shows that polls 300 days [before the vote] have little predictive value at all. The r-squared values increase sharply during the next 3 months or so—when the eventual nominees are becoming better known during the primaries—and then increase more sharply again in the 3 months before the election, when the general election campaign is underway.
For those keeping score at home, there are now 530 days until Election Day 2012.
