TAMPA, Fla. -- Mitt Romney may have a bigger problem than his low likability and favorability ratings. Even among voters who support him, many say he will lose to President Barack Obama.
Pollsters say that's a serious handicap that deserves more notice than it's receiving. Republican image-makers, focused on trying to make Romney seem warmer, might be wiser to try to make him more plausible -- as a commander in chief and as a politician who can defeat a seasoned campaigner like Obama.
Americans like winners, pollsters say. The "who do you THINK will win" question has a strong -- but certainly not infallible -- track record of signaling where an election is going.
The numbers are startling. Poll after poll shows Obama and Romney roughly even when people are asked the "horse race" question: who do you support for president? But Romney trails Obama badly when the same people are asked: Who do you believe will win?
This by no means assures Obama of victory. And there's plenty of time for the numbers to shift as confidence in Romney's chances grows. For now, however, veteran Republican pollsters say Romney's team should be concerned.
"If I could know only one thing in a campaign's home stretch, that's what I'd want to know," GOP pollster Mike McKenna said of the likely winner question. "It's a barometer of the real mood out there."
"That question gives a sense of the enthusiasm, turnout, last-minute volunteers and fundraising" that are likely to materialize, McKenna said.
Another GOP pollster, Steve Lombardo, agrees. "It reflects where the race is," he said of Obama's advantage on the who's-likely-to-win question.
"As much as people want to believe the race is tied," Lombardo said, Obama has an edge, at least for now. People who support a certain candidate, he said, "are less likely to go to the polls if they don't think their candidate can win."
Lombardo and McKenna say they expect Romney to narrow the "who will win" gap with Obama, and this week's nominating convention should help.
Lombardo said Romney needs to shrink his disadvantage to single digits.
He has a ways to go.
A recent AP-GfK poll found Romney and Obama essentially tied on the question of who they would vote for if the election were held today. But 57 percent of voters say Obama will win in November, while 33 percent think he will lose.
A May poll by Gallup found 56 percent of Americans predicting an Obama win, and 36 percent picking Romney.
Pollsters say no pre-election question is a foolproof predictor of an election's outcome. Despite popular perceptions, the ubiquitous who-would-win-today question is not meant to predict Election Day results, but rather is "a snapshot in time," as pollsters say.
The less-frequently asked "who do you think will win" question is somewhat different, said Andrew Kohut, veteran pollster for the Pew Research Center. "Its record for predicting is pretty good," he said, even if the reasons are not entirely clear.
Polling history supports his view. A mid-August 2008 poll for CBS News and The New York Times found that 48 percent of Americans thought Obama would win the election, and 34 percent named John McCain. Obama won, 53 percent to 46.
In August 2004, the Pew Research Center found that 44 percent thought George W. Bush would win, and 37 percent named John Kerry. Bush won the November election with 51 percent of the vote.
The 2000 election supports arguments both for and against the "who will win" question's ability to predict presidential results. A CBS News poll found that 47 percent of Americans expected Bush to win, compared to 33 percent for Al Gore. Gore won the popular vote that fall, but Bush won the Electoral College vote, and thus the White House.
In August polls in 1992 and 1996, Americans correctly predicted Bill Clinton's wins over incumbent George H.W. Bush and challenger Bob Dole, respectively. In August 1988, a poll similarly found confidence in Bush's eventual win over Michael Dukakis.
"Any indicator that has been as reliable as this, for whatever reason, has to be taken seriously," Kohut said.
It's impossible to read voters' minds. But the combination of low faith in Romney's ability to defeat Obama and his relatively low favorability ratings suggests he has yet to persuade most Americans he is cut from presidential timber.
There's ample time to remedy that. And, of course, the weak economy might sink Obama without Romney having to raise his popularity.
Romney's camp undoubtedly wants to use the convention's remaining hours -- especially his Thursday night acceptance speech -- to burnish his likability and plausibility. There's work to do.
Most registered voters viewed Romney unfavorably when the convention opened, a recent Washington Post poll found. He is the first party nominee with higher unfavorable than favorable ratings at this stage since Walter Mondale, the 1984 Democratic choice.
In the same poll, 61 percent said Obama was more "friendly" and "likable" than Romney; 27 percent saw it the other way around.
Several Republican strategists said they are not terribly worried about Americans' predictions of the Nov. 6 winner.
"That data fails to measure the intensity that voters bring to the table," which favors Republicans, said veteran consultant Terry Holt. "Romney still has to close the sale, but he's still got plenty of time."
Holt added: "I'd be more concerned if it were two or three days before the election."
Millions of Americans will tune in Thursday night for Romney's acceptance speech. Aside from the three October debates with Obama, it will probably mark his best chance to start closing the plausibility gap.