Two very old, often ineloquent white men will get into a lengthy argument in Atlanta on Thursday night — and, quite possibly, change the course of American history.

Some may doubt that the first debate of the 2024 presidential election has such high stakes. There are indeed many reasons to think that this week’s oratorical showdown between Joe Biden and Donald Trump will be of no real consequence.

For one thing, some academic research suggests that presidential debates have little to no impact on election outcomes. A 2019 paper from economists at HEC Montréal and Harvard Business School tried to gauge the electoral relevance of 56 debates televised during 31 election seasons in the US, Canada, and five other countries. To do this, they analyzed data from surveys that asked voters whom they supported shortly before or after the debates, and then contacted those same voters after the election to find out whom they ultimately backed. The researchers found that, across all political systems they studied, there was no evidence that debates led undecided voters to settle on a candidate, or other voters to switch allegiances.

This is not entirely surprising. Previous research from the political scientists Robert S. Erikson and Christopher Wlezien also found that, in every election featuring a debate from 1960 through 2008, the impact of debates on the respective candidates’ poll numbers appeared small to nonexistent.

There is also a theoretical reason to doubt the influence of debates on election outcomes: The voters most likely to watch a presidential debate — or to consume large amounts of news coverage about that debate — are those who are most interested in politics. Voters who consume lots of political media tend to be strong partisans, and therefore have little interest in reconsidering their loyalties.

In certain respects, this Thursday’s debate might seem liable to matter even less than a typical presidential debate. After all, both candidates are already extremely well-known to the electorate,having already debated twice in 2020. And given the unusual timing of this year’s first debate — which comes before the major party conventions instead of after, as had heretofore been standard — whatever impact the event does have on voter impressions could fade out long before Election Day.

Recall that, just one month before the 2016 election, the Washington Post published a video in which Trump bragged about grabbing women by the genitals. Hillary Clinton’s advantage over the Republican in RealClearPolitics’s polling average jumped from 4.7 points to 7.1 points within a week only to crash back down to 3.2 points by Election Day, a finding that overstated her ultimate share of the national popular vote.

But wait! Don’t decide to tune out Vox’s comprehensive coverage of the 2024 presidential debates just yet! If there is cause to believe that Thursday’s proceedings won’t matter, there are also reasons to think that they will.

Specifically, there are (at least) four reasons to suspect that Trump and Biden’s impending rematch could be one of the rare debates that actually does have a significant impact on voters’ decisions.

First and foremost, the debates do not have to matter much to influence the election’s outcome. As of this writing, in FiveThirtyEight’s polling average, Biden leads Trump by a margin of 40.8 percent to 40.7 percent — a statistical tie.

Polls in Michigan, Pennsylvania, and Wisconsin — the three most closely contested battleground states, which could collectively determine the Electoral College’s victor — are similarly tight. In RealClearPolitics’s polling average, Trump leads Biden in Michigan by 0.2 percent while besting him in Pennsylvania by 2.8 percent. In Wisconsin, meanwhile, the two candidates are tied.

In this context, an event that changes the voting intentions of a tiny sliver of the electorate could be of decisive importance. And debates can sometimes achieve that much, according to some researchers. As the communications scholars Mitchell S. McKinney and Benjamin R. Warner note in a 2013 review of the empirical literature on debate effects, while there is generally “very little change in voting intention following exposure to general election presidential debates,” some studies “have found that among undecided, conflicted, or weakly committed voters, debates do help form voting preference or even change candidate selection.”

These studies appear less comprehensive than the more recent research cited above, as they focus on a much smaller number of elections. Nevertheless, they suggest that there may be some contexts in which debates can have a real impact on outcomes.

And the 2024 election could be one such context because it features a plenty of weakly committed voters. Many Americans disapprove of both Trump and Biden. In the New York Times’s polling, there was a small but significant group of Biden 2020 voters who now supported Trump over the president — but said they would support Kamala Harris over Trump.

It is not hard to imagine that such voters might revert back to their conventional partisan allegiances in the face of a strong debate performance from Biden, or conversely, that they might grow firmer in their support for Trump, were the presumptive GOP nominee to dominate Thursday’s event.

Since the 2020 election, Donald Trump has fomented an insurrection against the US government, been found civilly liable for sexual abuse, faced indictment on a wide range of criminal charges, and been convicted of 34 felonies. And yet, he is more popular with the American public today than he was for much of his presidency.

Given this reality, it’s tempting to conclude that more or less nothing can poison Trump’s image with voters who remain in his corner. If Trump didn’t lose these folks by siccing a mob on his own vice president, how could he lose them with a subpar debate performance?

Yet in recent weeks, we’ve seen that novel information can put a dent in Trump’s poll numbers — albeit a small one.

On the day that he was convicted in a New York court for falsifying business records, Trump led Biden by 1.7 points in FiveThirtyEight’s national polling average. Ten days later, that lead had fallen to 1 percent. Now, it has been completely erased.

Recontact surveys — in which pollsters re-interview respondents they’d polled shortly before the conviction, to see if Trump’s legal woes had changed their minds — yielded similar results, with Trump’s margin over Biden shrinking by 2 percentage points in polls from the New York Times and Echelon Insights.

Correlation is not causation. It is conceivable that other, less conspicuous factors eroded Trump’s support over the past month. Regardless, it is apparent that a potentially decisive segment of voters lacks firm allegiances to either candidate and are liable to shift their preferences in response to major news events.

And this week’s debate will be one of the biggest political news events of the year thus far. All major networks will broadcast Biden and Trump’s rematch, and news outlets will spend days picking over its most memorable moments.

Were voters completely dug into partisan camps and/or impervious to new information about the candidates, then the substance of the debate coverage wouldn’t matter. But the polling shift since Trump’s conviction suggests that this is not the case.

In recent polls, Biden has performed significantly worse than down-ballot candidates in his party. This suggests that Biden’s personal attributes — as opposed to his policy positions — are leading some independent voters to withhold their support.

The president’s advanced age — he would be 82 years old at his second inauguration — appears to be one such attribute. In an ABC News poll from February, 86 percent of voters said that Biden was too old to serve another term. That is broadly in line with other pollsters’ findings on the subject.

Biden has had other opportunities to demonstrate his vitality. As the political blogger Kevin Drum has noted, the president’s State of the Union address in March appeared to abruptly dampen search interest in questions about his age.

But the State of the Union attracts a much smaller audience than general election debates: Biden’s address attracted 32 million viewers while his first debate with Trump in 2020 drew 73 million. More critically, responding extemporaneously to questions and attacks in the form of clear and concise soundbites fit for clipping is a much more cognitively demanding task than reciting a pre-written speech.

This is not to say that a candidate’s capacity to deftly navigate a televised debate is a good proxy for their capacity to serve as president. The skills necessary for shepherding legislation through Congress or brokering international alliances and those required for dispensing well-calibrated talking points in 90-second batches don’t necessarily overlap. Biden has never been an especially eloquent politician. But he has proven effective at enacting bipartisan legislation and orchestrating international cooperation on the war in Ukraine.

Nevertheless, were Biden to sputter, serially misspeak, ramble, or otherwise project an aura of senescence during Thursday night’s debate, he would very likely reinforce the public’s sense that an octogenarian should not be given a four-year lease on the White House.

Conversely, were Biden to deliver a performance on par with his drubbing of Paul Ryan in 2012, he might alleviate swing voters’ concerns about his cognitive well-being and embarrass Republican spin doctors who’ve been falsely accusing him of senility.

In the wake of the January 6 insurrection, major social media companies banned Trump from their platforms.

This may have actually redounded to his benefit. One of Trump’s greatest liabilities as a politician was his compulsion to broadcast unvetted thoughts directly to the public over Twitter. This provided the mainstream press — and thus, swing voters — with unmediated access to the demagogue’s (frequently cruel, racist, and authoritarian) thoughts.

Now that the ex-president’s posts are confined to his personal social media platform, Truth Social, his incendiary tantrums are much less visible to the general population. Voters who pay little attention to politics outside of presidential election years — a bloc that is exceptionally likely to be undecided — may also lack familiarity with Trump’s recent stump speeches.

Which is almost certainly a good thing for Trump’s campaign. Over the past two years, the Republican’s rhetoric has grown increasingly radical and deranged. He has taken to deriding his political opponents as “vermin” and migrants as “animals,” while lauding January 6 rioters as “great patriots.” His descriptions of American life have grown apocalyptic, recently declaring, “Our cities are choking to death. Our states are dying, and frankly, our country is dying.”

Above all, though, his extemporaneous remarks have become deeply strange. Trump’s speeches have always been digressive, but lately, the 78-year-oldcandidate’s digressions often fail to sync back up with his actual point, leading instead to still further digressions.

The typical undecided voter has not been exposed to this angry, authoritarian, and scatterbrained version of Trump in a long time, perhaps ever. And in the absence of the actual man, it is surely much easier for Americans to project onto Trump the candidate they wish he was: an unscrupulous but ruthlessly effective businessman who knows how to bring down inflation, get tough on immigration, and run the government more competently than Joe Biden.

Therefore, to the extent that the debate reacquaints the electorate with the actual Donald Trump, it could shift undecided voters in Biden’s direction.

All that said, there is certainly no guarantee that the debate will make a durable impression on the electorate. But given the extraordinary stakes of the 2024 race, Thursday night’s proceedings could mark a turning point in the history of our republic — or, failing that, an engaging (if almost certainly exasperating) spectacle.