The 2024 election has now entered the most topsy-turvy part of the electoral calendar: the convention period. Around this time, election observers tend to look out for just how much of a “convention bounce” a presidential nominee gets \u2014 that is, how much or little the candidate gains in the polls after what is effectively a four-day-long infomercial for the party and its presidential ticket.<\/p>\n\n\n\n
Of course, the 2024 story became more complicated following the assassination attempt against former President Donald Trump just days before this week’s Republican National Convention. But while the shooting will make it more difficult to interpret the causes behind any polling shifts after the RNC, this isn’t the first time that a major event roughly coincided with a national party convention. One way or another, how the polls change will be noteworthy regardless of why<\/em> it happened.<\/p>\n\n\n\n The bottom line is that the weeks surrounding the two major-party conventions have traditionally constituted<\/a> the most volatile polling period in presidential campaign cycles. In the week or two following their party’s conventions, nominees have typically enjoyed some improvement in the share of the vote they garner in surveys, though the size and longevity of these convention bounces has varied greatly. In recent years, however, the magnitude of poll movement around the conventions has shrunk. Increased polarization within the electorate has reduced the universe of swing voters, while the major parties’ recent pattern of holding conventions in back-to-back weeks has also potentially stifled convention bounces.<\/p>\n\n\n\n That trend makes 2024 an even more interesting test of whether we might see some meaningful fluctuations around the conventions. On the one hand, Trump and President Joe Biden’s high familiarity to voters has helped produce a fairly stable polling environment<\/a>. Yet the assassination attempt against Trump could serve as a shock to the political system that affects how the public views the candidates. And beyond that, other factors could make this a more dynamic convention period than in recent years, including the uncertainty surrounding Biden’s status<\/a> as his party’s standard-bearer, the unpopularity<\/a> of the presumptive nominees, the potential appeal<\/a> of third-party candidates and the month-long distance<\/a> between the conventions.<\/p>\n\n\n\n Yet it’s also clear that recent elections have featured less air in the convention bounce. From 1968 to 1996, more than 3 in 5 nominees enjoyed an increase of around 3 percent or more. But in the six election cycles from 2000 on \u2014 the start of the more highly competitive “red state\/blue state<\/a>” era \u2014 just 1 in 6 nominees saw similar improvement, and less than half saw a boost of even around 2 percent.<\/p>\n\n\n\n Polarization has likely contributed to the reduced magnitude of convention bounces by keeping more voters in one camp or the other. Although more voters today identify as independent than with either of the two major parties, most independents lean toward one party<\/a>. And compared with the more distant past, the two parties have become increasingly ideologically polarized<\/a> over the past few decades. The electorate is more likely to view the parties as significantly different<\/a> from each other and hold negative views of the opposing party<\/a>. Collectively, these trends have made it less likely for voters to break from the party they tend to identify with and has also reduced the share of swing voters in the electorate<\/a>.<\/p>\n\n\n\n Another factor might be the timing of recent conventions. From 2008 to 2020, the parties held their events in back-to-back weeks, meaning the second convention undoubtedly affected the amount of attention paid to the first gathering (and perhaps vice versa). Across those four election cycles, candidates averaged a 1 percent convention bounce. By contrast, from 1968 through 2004, the start dates for each convention stood at least two weeks apart, leaving more time afterward for voters and the media to focus on what was said and done by each party. Perhaps not coincidentally, the average bounce across that period was closer to 3 percent.<\/p>\n\n\n\n That effect seems to be disproportionately deleterious to the party that doesn’t control the White House, which since the 1950s has consistently held<\/a> its national convention before the incumbent presidential party does. That means that from 2008 to 2020 the White House party may have stepped on the out party’s potential bounce by holding its convention the very next week. In those cycles, the out party averaged almost no bounce, while the White House party averaged about a 2 percent gain in the polls. But from 1968 to 2004, each<\/em> party averaged nearly a 3 percent gain after its respective convention.<\/p>\n\n\n\n This summer, the Republican and Democratic gatherings are separated by about a month, a change from the back-to-back pattern of recent elections. That could provide more space for a potential convention bounce to play out after the GOP convention without the Democrats’ event stepping on it.<\/p>\n\n\n\n Changes in the polls after a convention don’t just reflect what’s happened inside the arena, but also events that took place around the same time. With that in mind, it will be particularly difficult to disentangle to what extent any changes after the 2024 Republican convention have to do with the assassination attempt against Trump, the gathering to renominate him or the Democrats’ Hamlet act over whether Biden will remain the party’s nominee. Although we don’t have an obvious apples-to-apples comparison for the 2024 situation, we do have precedent for events outside a convention likely having a big influence on voters’ attitudes.<\/p>\n\n\n\n At the Democrats’ 1968 convention in Chicago, the events that played a part were practically right outside the convention hall<\/a>. Amid rising social conflict and division over the Vietnam War, the event turned into a disaster for the Democrats, as millions of Americans watched split-screen images<\/a> of a police riot and violence in the streets<\/a> of Chicago while the highly fractured Democratic Party angrily debated its future inside the convention hall. Tellingly, the post-convention polls found Democratic nominee Hubert Humphrey lost 2 percent following the convention, the only clear convention downturn in the 1968 to 2020 period. Humphrey never fully recovered and went on to narrowly lose to Nixon that November.<\/p>\n\n\n\n Around a quarter century later, Democrats enjoyed the largest bounce in this time period following their July 1992 convention, but that wasn’t entirely down to a marvelously-run event. On the morning of the final day of the Democrats’ convention, independent candidate Ross Perot announced his withdrawal from the race<\/a>, saying he chose to do so<\/a> in part because “the Democratic Party has revitalized itself.” At the time, Perot was polling above 20 percent in national surveys, not far behind Democratic nominee Bill Clinton or incumbent President George H.W. Bush, and Clinton gained in the wake of Perot’s departure<\/a> as voters dissatisfied with the status quo preferred him in a head-to-head race against Bush. (Of course, Perot rejoined the race in October<\/a> and won about 19 percent of the popular vote, one of the strongest third-party performances<\/a> in the history of U.S. presidential elections.)<\/p>\n\n\n\n It’s anyone’s guess as to how much the polls will move during the convention period this cycle, but there were a few reasons to expect somewhat more volatility than in recent years even before the assassination attempt added in additional uncertainty. As we saw in the aftermath of Biden’s rough debate performance in late June, the polls can<\/em> move \u2014 Trump’s position has improved by about 2 percentage points in margin since then.<\/p>\n\n\n\n First, a pinch of doubt<\/a> remains about Democrats’ likelihood of renominating Biden<\/a>. Needless to say, if Biden were to step aside for, say, Vice President Kamala Harris, that would represent a seismic event that would at least partially reset the election<\/a> and the polling of the contest. Although Biden and Harris haven’t polled that differently in recent surveys testing each of them against Trump, surveys of Harris’s standing still involve a hypothetical situation \u2014 for now at least.<\/p>\n\n\n\n But assuming we do get a Biden-Trump rematch, another wrinkle is that both candidates remain unpopular<\/a> with the public. In 538’s polling averages, Biden has a 39 percent favorability rating<\/a> while Trump’s standing is at 42 percent<\/a>. As a result, much of the potential electorate views each of them negatively \u2014 about 1 in 5 likely voters in our 538\/Ipsos debate surveys<\/a> were “double haters.” This historic level of unpopularity could introduce additional uncertainty into an otherwise stable polling environment, with low voter enthusiasm affecting both candidate preference and, critically, turnout.<\/p>\n\n\n\n Some of those double haters might also be more likely right now to say they’re going to vote for a third-party contender even if they ultimately vote for Biden or Trump, another potential reason behind convention-season polling fluctuations. We know that supporters of those candidates are far more likely<\/a> to say they could change their minds about who they’re voting for than supporters of Trump or Biden, so their preferences may also be more apt to shift in response to the convention messaging and other coverage of the campaigns during this period.<\/p>\n\n\n\n Now, it’s hard to imagine a truly dramatic shift in a Biden-Trump contest because the overriding forces of polarization, which have likely played a role in shrinking convention bounces, remain in effect. While it wouldn’t be a surprise if Trump enjoyed a convention bounce of sorts, given his recent polling gains and the timing of this year’s conventions, we’d issue the usual note of caution that it’s best to look at the polls with an acceptance of their inherent uncertainty and know that every election differs. We can look to past examples for clues as to how things might play out, but while history will sometimes rhyme, it rarely repeats itself.<\/p>\n\n\n\n