Donald Trump won’t win because of the presidential debate. But neither will Hillary Clinton.

A dashing John F. Kennedy smiling at a disheveled Richard Nixon. Gerald Ford denying Soviet control of Eastern Europe. Michael Dukakis fumbling a question about the hypothetical murder of his wife.

These presidential debate images are seared in our memory, so we assume they influenced the outcomes. But we’re wrong. While debates can influence a candidate’s image, they alone rarely change anyone’s mind.

We tend to forget that debates are a recent phenomenon, dating to the dawn of TV. The first time two general-election presidential candidates faced off was the Kennedy-Nixon debate of 1960, when JFK’s easy manner and good looks allegedly helped him nose out Nixon in the contest.

And yes, Kennedy cut a very different figure on TV than Nixon. JFK wore a form-fitting dark suit and looked at the camera, creating a personal connection with viewers. Nixon sweated in his loose gray suit.

So that’s why Nixon lost the debate, right? Not necessarily. That well-worn myth is based on a study showing that people who heard the debate on the radio said Nixon won, while those who watched it on TV gave the edge to JFK. But the finding was based on 282 radio listeners, who mostly hailed from rural areas and were probably predisposed to favor Nixon anyway.

But after the myth took hold, it was hard to dislodge it. Nixon apparently believed it, refusing to debate Democratic opponents in 1968 and 1972. The next presidential debate didn’t happen until 1976, when Ford declared that there was “no Soviet domination of Eastern Europe.”

Newspapers made a big deal over the gaffe, which helped reinforce the image of Ford as a genial bumbler. Again, though, it didn’t seem to affect the electoral outcome. Polls showed Ford was gaining ground on Carter before the debate and continued to close the gap after that, suggesting viewers didn’t care as much about Ford’s mistake as the media did.

Ditto for Dukakis’ performance in 1988, when he said he would oppose the death penalty for someone who raped and killed his wife. Media commentators pronounced Dukakis’ epitaph, noting his impassive manner on such an emotional matter, but a Gallup study showed the debate had “little to no impact on voter preferences.” He was already dropping in the polls, and there’s no evidence that his remark sank him further.

But the myth of the decisive debate continues, which helps explain why candidates spar over the rules. Ford and Jimmy Carter battled over the height of their lecterns, eventually agreeing the lecterns would intersect their torsos at their belt buckles so neither man would appear taller than the other. Bill Clinton and George H.W. Bush fought over whether they would have drinking water on stage.

But nobody has charged that the other team tried to rig a debate by holding it at the same time as a football game. That’s what Donald Trump claimed, even though this fall’s debates were scheduled by a nonpartisan commission a year ago. Nor did Trump receive a letter from the NFL complaining that two of the debates conflict with games, as he falsely asserted this summer.

Trump is also apparently the first presidential candidate to poll-test his debate strategy via email, as he did in a message to supporters last week.

“Do you think Trump should refer to Hillary as ‘Crooked Hillary’ on stage?” asked one of the questions. “Should Trump call out Hillary’s reluctance to say ‘radical Islamic terrorism’ and her underlying fear of being politically incorrect?” asked another.

Whatever Trump and Clinton do onstage, though, they probably won’t sway voters one way or another. Most people who tune into these events already have made up their minds; they’re looking for validation, not information. And that might be the most politically incorrect fact of all.

Jonathan Zimmerman teaches education and history at the University of Pennsylvania.