Some believe that elections are decided by independent voters, those who are not loyal to a political party. But even as the ranks of those independents have grown, most of them vote for candidates of the same party time after time, political scientist Christopher Wlezien said during a recent presentation at the Federal Reserve Bank of Atlanta's Public Affairs Forum.

"What has grown are the independent 'leaners,' who vote very much like partisans, and the percentage of 'pure' independents actually has shrunk," said Wlezien, a University of Texas professor who cowrote the 2012 book, The Timeline of Presidential Elections.

During his January 20 talk at the Atlanta Fed's headquarters, Wlezien discussed the reliability, or unreliability, of pre-election polls and the factors that motivate voters to cast ballots for particular candidates.

Trial polls taken 300 or more days before an election—such as those positing a theoretical general election of Candidate X vs. Candidate Y before either has won a party's nomination—"have no predictive power," Wlezien said. The pivot timeframe is 200 days before an election. From that point, the accuracy of polls historically has improved dramatically, Wlezien and coauthor Robert Erikson found in studying nearly 2,000 national polls from every presidential election from 1952 through 2008.

Other nuggets Wlezien shared include:
  • In the final weeks before an election, big leads shrink; small leads don't change much.
  • Debates generally do little to change general election outcomes.
The Federal Reserve is a non-partisan organization. The views expressed by the Public Affairs Forum speaker are his own, not those of the Federal Reserve Bank of Atlanta or of the Federal Reserve System.