Hillary Clinton and Donald Trump. (Getty Images; Reuters)

Monday night’s presidential debate — or the “Clash of the Century,” if you prefer — is anticipated to bring a record audience and has political observers speculating endlessly about its potential impact. So what is really important to know? Here are five keys to thinking about presidential debates in general and this debate in particular.

1) Presidential debate history is awash in folklore.

If your thinking about Monday night’s debate is guided by the perennial “six debate moments that mattered” or “five key debate moments that altered the course of a presidential race,” you’re mostly mistaking myth for fact. As I wrote in 2012, the history of presidential debates is filled with alleged “turning points” that didn’t change much at all.

Gerald Ford says there’s no Soviet domination of Eastern Europe, Michael Dukakis is too unemotional in response to a question about whether he’d favor the death penalty if his wife were raped and murdered, George H.W. Bush looks at his watch — all of these are moments that didn’t really affect voters if you examine the polling data around debates.

The folklore typically ignores the polling data entirely in favor of impressionistic assertions about what was “disastrous,” etc. So go ahead and purge the folklore from your mind.

2) Presidential general-election debates can move the polls but rarely decide the winner.

A close analysis of polling data does reveal instances in which the polls shifted after debates. But debates rarely turn the trailing candidate into the winner on Election Day.

One comprehensive study of the 1960-2000 elections by political scientist James Stimson found that debates may provide a “nudge” in close contests (1960 or 2000, say), but that “there is no case where we can trace a substantial shift to the debates.”

A study of the 1952-2012 presidential elections by political scientists Robert Erikson and Christopher Wlezien concludes that “the best prediction from the debates is the initial verdict before the debates.”

It used to be that this take on debates was the buzzkill view of political science. But now it’s widely acknowledged.

There are three reasons why it’s hard for debates to have that large an impact. First, they occur relatively late in the campaign, after most voters have made their decisions. (More on this below.)

Second, a lot of other stuff happens alongside and after the debates that can also move the polls. In 1980, for example, the debate between President Jimmy Carter and Ronald Reagan may have increased Reagan’s lead by a couple of points. But the problem is that all of the following also happened after the debate. Prominent Reagan and Carter aides were forced to resign; inflation rose; the congressional investigation of Carter’s brother, Billy, continued to make news; and Iran again refused to release American hostages. It’s hard to know what mattered the most when so much happened at the same time.

Third, even if one candidate has a “bad” debate, they usually recover in any later debates. Presidential candidates aren’t dumb or untalented. They adapt. So don’t overreact to this first debate. Remember this headline from after the first debate in 2012: “Did Obama just throw the entire election away?”

The crucial question isn’t who wins any one debate, but whether over the course of the debates, one candidate can get a net advantage.

3) But beware: Polls may exaggerate the actual impact of debates.

Lost in the hoopla about that first debate in 2012 was this under-appreciated fact: At least some of the apparent shift in the polls toward Mitt Romney was not the result of people changing their minds. It was the result of Republicans suddenly being more eager, and Democrats less eager, to talk to pollsters. This phenomenon — called “differential non-response” — was documented then by YouGov. Subsequent analysis of the 2012 polls suggested that differential non-response may have exaggerated other shifts as well. The same appears to be true in 2016.

Of course, you know to ignore the “snap” polls of debate “watchers” Monday night. But beware of the later polls as well.

4) Media interpretations of who “won” the debate matter. A lot.

This is a point that I see few in the news media acknowledge. Even if 100 million people watch the debate Monday night, their views will be shaped by the news coverage that comes afterward. Many studies have shown that news coverage helps “frame” or interpret politics for average voters. For better or worse, we “outsource” some of that interpretive labor to reporters and political commentators. Candidates know this, too, which is why they devote so much energy to shaping the expectations of news media actors.

In a study I’ve discussed before, researchers at Arizona State University divided people into three groups. The first watched footage of the third presidential debate in 2004. The second watched that footage plus 20 minutes of post-debate commentary on NBC. The third watched that footage plus 20 minutes’ time to read commentary on CNN.com. So who won the debate, George W. Bush or John F. Kerry?  It depended on whether you watched the news:


Here’s another example: Ford’s famous “gaffe” about Soviet domination of Eastern Europe. The polling actually showed that no one initially perceived this as a gaffe. It was only in the days after the debate that voters “realized” the gaffe once it had been the focus of news coverage.

Now of course, as my Washington Post colleague Paul Farhi points out, there is no single entity called “the media.” There is no cabal that meets in a room to decide who won the debate.

But there is an ongoing conversation among reporters and political observers — now conducted in real time via Twitter — that helps to shape a narrative about who did well and who did poorly. That narrative then affects what voters believe.

This is why, for example, some political scientists and reporters think there is value in having political observers watch the debate without one eye on their Twitter feed.

And this is to say nothing of whether political observers and commentators might be well-served to downgrade the significance of debates, period, and especially to avoid focusing excessively on the “style” of the candidates.

5) Is 2016 different?

This is the big question, of course. I don’t have a prediction, and I’m not going to provide armchair psychology of the candidates. I will note that the number of undecided voters has shrunk of late: Together Clinton and Trump combine for 91 percent in current polls. At this point in 2012, Romney and President Obama combined for 93 percent. For that reason, at least, the debates may once again not be able to provide more than a nudge.

But even a small nudge after Monday night — say, two points — in Trump’s favor would essentially cut Clinton’s lead in half. A similar bump for her would give her a more comfortable lead. The question, then, would be what happens in the later debates.