Mitt Romney now leads President Obama in the race for Florida's 29 electoral votes, a significant reversal from earlier this spring, according to a new Quinnipiac University poll released early on Wednesday.
Romney leads Obama overall, 47 percent to 41 percent, the poll shows. Two percent of voters prefer another candidate, 7 percent are undecided, and 3 percent said they would not vote. Three weeks ago, in another Quinnipiac poll, the two candidates were statistically tied, with Romney at 44 percent and Obama at 43 percent.
But less than two months ago, in late March, Obama led Romney, 49 percent to 42 percent; since that time, Obama's numbers have suffered attrition nearly across the board. In late March, he led Romney among women, 52 percent to 38 percent, but that lead is now just a single point, 45 percent to 44 percent. He trailed Romney by only 2 points among men in March, but he now trails by 13 points in the new survey.
Romney leads Obama by 8 percentage points among independent voters, 44 percent to 36 percent. In late March, Obama led among this crucial bloc by 10 points.
Obama's deficit among white voters has also grown, from a lag of 13 percentage points in March to 23 points today. The shift was particularly driven by white voters without a college degree; among this group, Obama now trails Romney by 27 percentage points, 57 percent to 30 percent.
Hispanics have also trended toward Romney in the two surveys. The two candidates now run virtually even among Hispanic voters, with Obama at 42 percent and Romney 40 percent.
Wealthier voters also are breaking for Romney, the poll shows. The former Massachusetts governor leads Obama among voters whose annual household incomes exceed $100,000, 59 percent to 35 percent.
A majority of all voters, 52 percent, disapprove of the job Obama is doing as president, while only 44 percent approve. An equal majority think Obama does not deserve to be reelected. And just 45 percent of voters have a favorable opinion of Obama, compared with 50 percent who have an unfavorable view. In the March survey, Obama's favorable rating was right side up, at 51 percent to 44 percent.
Romney's image ratings, on the other hand, are improving. Forty-four percent have a favorable impression of him, compared with only 35 percent who view him unfavorably. That is the largest favorable-to-unfavorable spread Romney has recorded in Quinnipiac's polls since early January, before the Republican primary campaign began in earnest in the Sunshine State. Voters also continue to see Romney as the candidate who would do a better job on the economy, 50 percent to 40 percent.
Half of voters oppose same-sex marriage, according to the poll, while just 40 percent support it. But the issue has little salience for both sides. Forty-nine percent of voters say the issue will not be important to their vote for president; just 22 percent say it will be "extremely" or "very important."
A quarter of voters say Obama's support for same-sex marriage makes them less likely to support him, while only 11 percent say it makes them more likely. But a wide majority, 63 percent, say it makes no difference.
For Romney, the issue is unlikely to move the needle significantly toward his candidacy. Asked whether his opposition to same-sex marriage makes them more or less likely to vote for him, 23 percent say they are more likely to support him, compared with 19 percent who are more likely to oppose him. But when asked about his opposition to same-sex civil unions, voters tilt against Romney: 15 percent say it makes them more likely to vote for him, but 23 percent say it makes them less likely. On both questions, majorities say it does not make a difference.
Quinnipiac also expanded the horse-race question to include Sen. Marco Rubio, R-Fla., as Romney's running mate. In that matchup, the Romney-Rubio ticket led Obama and Vice President Joe Biden, 49 percent to 41 percent. Overall, 54 percent of Florida voters approve of the job Rubio is doing as senator, the highest Quinnipiac has recorded for the freshman.
The Quinnipiac poll was conducted May 15-21, surveying 1,722 registered voters. The margin of error is +/- 2.4 percentage points.