This year's presidential debates are being predictably portrayed as potentially pivotal. Is it so? Do debates change things?
Answer: Maybe — but not always, and when so, more indirectly than directly.
We at ABC News will be looking for answers starting with one of our patented instant debate-reaction polls tonight; more on that below. First, though, a bit of background.
Debates have been held in eight presidential contests since 1960. We find just one after which the lead changed hands by a meaningful margin: In 1980, when Ronald Reagan uttered his "Are you better off?" line. He gained seven points in a post-debate poll.
There are other cases in which debates (or post-debate evaluations) may have had a subtler, less measurable effect on the dynamics of the race. They are, after all, an essential window on the candidates' styles. After Richard Nixon's pasty-faced performance in the 1960 debates, John F. Kennedy went from 46 percent support to 49 percent; Nixon slid from 47 percent to 45 percent.
Those changes are not large enough to be significant given polling tolerances (as we're fond of saying, this ain't laser surgery). Nonetheless, collective memory maintains that the debates spelled Nixon's narrow defeat (that, or Richard Daley).
Nor did polls show significant movement immediately after the 1976 debate in which Gerald Ford said Poland was free (and news reports pounced on the misstatement); Jimmy Carter gained a single point, Ford lost three. Again, though, the gaffe may have had the more subtle effect of halting what had been a slide in Carter's advantage.
Our polls in 2000 showed no significant movement around the debates. Ditto for 1996. In 1992 there was more movement, though never enough to change the lead: Ross Perot gained 11 points through the three debates, moving from a distant third place to a still-distant third place. Bill Clinton and the first President Bush popped up and down by five or six points, possibly accommodating the Perot movement.
Note that when change does occur, it isn't necessarily directly debate-related.
Candidate support sometimes shifts with no debate to pin it on. If it happens around a debate, that's said to be the reason — causality is assumed, like someone guessing what moved the stock market.
Measurable effects show that debates — at least in immediate reactions — mainly tend to reinforce preconceived notions rather than change them; most of each candidates' supporters say it's their guy who won.
Typical was our first debate-reaction poll in 2000: Seventy-nine percent of Al Gore supporters said Gore won; 70 percent of George W. Bush's supporters said Bush won. And 93 percent of each candidate's supporters stuck with their pre-debate vote preference. Among debate watchers, Gore's support was 45 percent before the debate and 45 percent after it; Bush's went from 48 percent to 49 percent.
Still, while the debates rarely prompt much change, post-debate evaluations can. In 1992, immediately after the first debate, 24 percent of debate watchers said Perot had won. By the very next night, however, that perception had grown to 37 percent among people who either had watched it, or heard or read about it. And as noted, Perot's support did advance, from 6 percent before the debates to 17 percent after them.
The post-debate polls we've been describing are full-population surveys that capture the state of the overall race. We'll have one of those soon enough. First, though, we will have an immediate debate-reaction poll tonight.
This is a different kind of animal. Debate-reaction polls do not measure the race overall, among all registered or likely voters. They're done only among registered voters who watched the debate. This group does not necessarily look like the broader voting public; in 2000, for instance, Republicans tuned in disproportionately.
If the horse race among debate watchers does not look like recent likely voter horse races, do not be surprised.
Debate-watcher polls are "panel" surveys. In the ABC News/Washington Post poll we reported Monday, we asked people if they'd be watching the debate, and if yes, whether we could call back after and ask what they thought. That is what we'll be doing tonight: Around 125 interviewers jumping on the phones, calling back debate watchers from our random national sample, and asking three questions: Did you watch, who won, whom do you support? In 20 or 25 minutes, we should have the results.
One caution: Different debate-reaction polls can get different answers, because they ask different questions. Our ABC News poll asks: "Who, in your opinion, won the debate?" The answer can be Kerry, Bush, or "tie."
Others go at it differently. Some ask, "Who do you think did the better job — (this guy) or (that guy)?" Doing the better job is different from winning, and naming the guys steers people away from calling it a tie. Others ask who did the better job, but without naming the participants. Still others might not allow "tie" as an option at all. All polls should be evaluated on the basis of what they asked.
Lastly, someone somewhere is sure to run an Internet click-in "poll." Run screaming.
In 2000, both CNN and our very own ABCNews.com ran online click-ins on the debate. Earlier that day Jim Nicholson, then-chairman of the Republican National Committee, sent around a mass e-mail telling his people to go to CNN.com and ABCNews.com, "log on tonight: Vote after the debate and make your voice heard!"
Voilà: The ABCNews.com click-in had Bush winning the debate, 58 percent to 41 percent. Our actual poll had it 42 percent to 39 percent, Gore-Bush.
More ABC News polls can be found in the ABC News Poll Vault.