PRINCETON, NJ -- In 9 of the past 15 U.S. presidential elections, the candidate who was leading in Gallup polling roughly four months before the election ultimately won the popular vote for president. However, narrowing the set of races to the nine that were competitive, the early polling proved prescient in only three of those.
With Barack Obama leading John McCain by no more than six percentage points in Gallup's early July polling, the 2008 race currently fits best into the "competitive" category. Given that assumption, Gallup's election trends from a comparable point in previous presidential election years offer no strong indication of whether Obama or McCain is headed for victory in November.
(In 1976, Jimmy Carter, who won that election, was ahead by 33 points in mid-July, but the race narrowed significantly by Election Day and he won by only two points. Thus, for the purposes of this analysis, the 1976 race is classified as competitive.)
In all other elections since 1948 -- all but one of which involved the successful re-election of an incumbent -- the eventual winner was leading by at least 16 points at this stage of the campaign, and retained a wide lead right through the election. On this basis, these six races can be considered "noncompetitive."
The category of competitive elections (in which the early to mid-July leader was not necessarily the eventual winner) can usefully be grouped into two subcategories, but it is too soon to say which one this year's race falls into:
In four elections -- 2004, 2000, 1980, and 1960 -- either the race remained closely contested for the entire campaign, or the lead switched back and forth between the two major-party candidates through the summer and/or fall.
In three -- 1992, 1988, and 1968 -- one of the parties' national conventions later in the summer proved to be a turning point in the race, vaulting one of the candidates into a lead, however small, that extended through the duration of the campaign.
The 1976 presidential election was in a class all by itself. Jimmy Carter enjoyed large double-digit leads through the spring and summer; then Gerald Ford narrowed the gap in the fall to single digits and almost won the election.
The 1948 election was also a singular case where the candidate leading in nearly all of Gallup's pre-election polls, including the final pre-election poll conducted more than a week before the election, showed the eventual winner (Harry Truman) trailing his opponent. As a result of this failure to provide an accurate picture of the race, Gallup quickly instituted a number of major and important changes to its pre-election polling methods, including identifying "likely voters" and polling right up to Election Day.
Barack Obama has consistently led John McCain in Gallup Poll Daily tracking for the past month, but by an average of only three points among registered voters. His largest lead since July 1 has been six points, although in the latest Gallup report, based on interviews conducted July 3 and July 5-6, it is just four points, 47% to 43%. (For the full Gallup Poll Daily interactive trend since March, click here.)
History provides no clear indication of the relationship between this narrow margin and the eventual outcome in November. The pattern that occurred in several presidential years suggests that the convention period could be crucial for either cementing Obama's slight advantage or establishing McCain as the new front-runner. This year's Democratic convention will be held in Denver from Aug. 25-28, followed by the Republican convention in Minneapolis/St. Paul from Sept. 1-4. If neither convention succeeds in transforming the election, the race could very well remain close right through the home stretch.
To provide feedback or suggestions about how to improve Gallup.com, please e-mail firstname.lastname@example.org.