PRINCETON, NJ -- U.S. presidential candidates historically have seen a median increase of five percentage points in their support in preference polls among registered voters after their party's nominating convention. The average is slightly higher, six points, due to the record 16-point increase for Bill Clinton after the 1992 Democratic convention.
The conventions, and their subsequent effects on voter preferences, are an expected and anticipated part of each presidential campaign. Nearly all presidential candidates have seen at least a minimal increase in their support when one compares the last Â鶹´«Ã½AV poll conducted before the convention and the first poll conducted after it. The only exceptions were George McGovern in 1972 and John Kerry in 2004. (See page 2 for complete pre- and post-convention estimates for all candidates.)
Clinton's 16-point bounce came after he was running third behind Republican George H.W. Bush and independent Ross Perot in polls in the spring and early summer of 1992. Perot dropped out of the race during the Democratic convention, at which Clinton and the Democrats successfully got voters to focus on the poor state of the economy and to blame Bush for it. The 16-point bounce estimate is not a function of Perot's departure because it is based on a comparison of vote preference questions before and after the convention that asked voters for their preference between Bush and Clinton only.
The historical typical increase of five points applies regardless of the particulars of each convention, such as whether it was the first or second convention in a given election year, whether it nominated an incumbent or non-incumbent candidate, and whether it was a Democratic or Republican convention. To the extent the medians or averages exceed five points, it is due solely to the influence of the 1992 Clinton convention on those estimates.
What to Expect in 2012
Everything else being equal, Romney and Obama could each expect to see a five-point increase in his support after the coming political conventions, which could change voter preferences at least in the short term. has shown Romney and Obama generally within two points of each other in registered voters' preferences. Thus, if Romney gets a typical bounce, he could lead Obama after the GOP convention, although Obama would essentially erase that lead if he, too, gets a similar bounce. If one of the candidates gets a bigger bounce than the other, that candidate could well establish a lead in the race after both conventions are over.
There are parallels between the 2012 and 2004 contests that suggest voter preferences may not be affected much by this year's conventions, as was the case in 2004. Both the 2012 and 2004 elections involved incumbent presidents seeking re-election with . The approval ratings in the high 40% range for both Bush and Obama may be a function of of them by party, generally a greater partisan divide than for other presidents.
Also, in both elections a high percentage of Americans were to the campaigns before the conventions. In other words, if voters are already tuned in to the election and already have made up their minds on whether Obama deserves a second term, neither he nor Romney may see his support change much after the convention.
The 2008 election was another contest in which voters were engaged in the campaign early but in which there were still typical bounces for both and . That may be because it pitted two non-incumbents against each other, both of whom were , so voters may not have been as locked into their preferences as they were in 2004, when voters were deciding whether to give Bush a second term in office.
Â鶹´«Ã½AV will continue to track voter preferences through the election, and will report on convention bounces in the days after each convention.
