THE DYNAMICS OF POLL PERFORMANCE DURING THE 2008 PRESIDENTIAL NOMINATION CONTEST

This analysis focuses on estimation difficulties pollsters had in the primaries in 2008 in light of recent trends in improved polling accuracy in general elections. We consider the series of polls that were conducted in New Hampshire and other states holding primaries, looking at how the dynamics of the primary contest affected polling accuracy in those contests. The data come from published state-level results of public pollsters from the week preceding each primary or caucus for which polls were conducted; all told, we used 258 polls in thirty-six different Democratic events and 219 polls in twenty-six Republican events. The results show that the winner's vote share almost always exceeded the poll share while the race remained competitive, particularly early on in the nomination process. In an unusual perspective made possible by the length of the contest on the Democratic side in particular, this could be observed through most of the primaries; it was not the case in the Republican events after John McCain became the presumptive nominee. The analysis shows there are contextual factors at work that can affect the quality of the estimates that public pollsters make. Measures of momentum and viability affect the estimates differently early in the process compared to later, and there are special factors associated with the insurgent candidacy of Barack Obama that may also have affected the accuracy of the polls. We model these factors, investigate their explanatory power, and discuss the implications for pollsters in future primary sequences.