James Fallows claims in The Atlantic that "moments" from televised general-election debates have "figured in the ultimate outcome" in the presidential elections of 1960, 1976, 1980, 1984, 1988, 1992, 2000, and 2004:
There have been nine series of televised general-election debates. These started with Kennedy-Nixon in 1960, resumed with Ford-Carter in 1976, and have been a campaign fixture ever since. In all but one election, the debates produced a moment that figured in the ultimate outcome. (The exception was Clinton-Dole in 1996, when neither man said anything that changed a voter’s mind.) The dramatic exchanges that made a difference—Ronald Reagan’s amused and dismissive “There you go again” against Jimmy Carter in 1980, Michael Dukakis’s too-composed look when asked in 1988 how he would react if his wife were raped, George H.W. Bush’s desperate “when will this end?” glance at his wristwatch during a town-hall session with Bill Clinton and Ross Perot in 1992, Al Gore’s operatic sighs about George W. Bush in 2000—would have passed unnoticed in a transcript. The transcript conveys only part of, for example, the alarming meandering in Ronald Reagan’s soliloquy at the end of his second 1984 debate with Walter Mondale. Reagan, looking confused and forgetting his point, was rescued only when the moderator, Edwin Newman, announced that time was up: “Mr. President, I’m obliged to cut you off there, under the rules of the debate. I’m sorry.” Mondale should have been sorry, too.
But as I pointed out on Monday, journalists tend to construct post hoc narratives that attribute election outcomes to dramatic visuals like negative campaign ads and debate exchanges. Here, by contrast, is what the eminent political scientist James Stimson concludes about the effect of debates in a passage from his book Tides of Consent that I highlighted back in January (my emphasis):
What we have seen is perhaps some influence. The evidence is inconclusive to say either that debates matter or that they do not. But if they do matter at all, their influence is vastly smaller than, say, the conventions. The reelection landslides [1964, 1972, 1984, 1996] show that once voters have decided, debates will not change the outcomes.
There is no case where we can trace a substantial shift to the debates. But in elections that were close at debate times, there are cases (1960, 1980, 2000) where the debates might have been the final nudge. As to why they are so often featured as the central story line of a presidential election campaign, I lean to the idea that they are conveniently available TV footage.
The issue is that the debates happen so late in the campaign that the popular vote winner has generally already taken an insurmountable lead, as this graphic (which combines the trajectories of 1976, 1980, 1988, 1992, and 2000) illustrates:
Here are the individual trajectories of the competitive races:
Stimson also makes several points about specific elections that conflict with Fallows. First, despite Reagan's shaky performance in his first debate with Walter Mondale in 1984, he regained his form and the debates ended up being irrelevant (Reagan won in a landslide). Also, Gerald Ford and George H.W. Bush actually gained ground on their opponents after the debates despite committing much-hyped gaffes. Finally, it's worth noting that Stimson points to a different Reagan line from 1980 than Fallows ("Ask yourself if you were better off than four years ago"), which highlights the subjectivity of debate interpretation.