It feels strange calling Monmouth’s latest a bad poll for Trump when their last poll of Florida had him down nine points in the four-way race. He’s cut Hillary’s lead nearly in half! But that last poll was taken in mid-August, during the doldrums of his summer polling. He spent the first half of this month rebounding, to the point where he actually led Clinton by 0.9 points in RCP’s poll average of Florida when today began. The last six surveys there look like this: Tie, Trump by one, Trump by four, Clinton by two, Trump by three, Clinton by one. The race is a toss-up, as Florida traditionally is. Now here comes Monmouth to pee in the punchbowl with a five-point lead for Hillary, suggesting that the last two weeks of the race might not be a “new normal” in the race but rather a blip. Pair this with that questionable topline in Pennsylvania from Muhlenberg showing Trump still down eight points there and Trumpers have new reason to wonder if the race is as newly competitive as they think. It’s more competitive, but is it really competitive?

Needless to say, given what a heavy lift Pennsylvania is, Florida is an absolute must-win for Trump.

Among Florida voters likely to participate in November’s presidential election, 46% currently support Clinton and 41% back Trump. Another 6% intend to vote for Libertarian Gary Johnson and 1% support Green Party candidate Jill Stein, with 5% who are undecided. This compares to Monmouth’s August poll which had Clinton at 48%, Trump 39%, Johnson 6%, and Stein 1%.

Clinton has an overwhelming advantage among Hispanic, black and Asian voters who make up about one-third of the electorate, garnering 69% of this group’s vote to 16% for Trump. Her lead among non-white voters was a similar 69% to 19% last month. Trump’s 53% to 35% advantage among white voters is also about where it was last month (51% to 37%).

How’s she doing it? By consolidating Democratic support, especially vis-a-vis Trump’s troubles with Republicans:

Hillary pulls 94 percent among Dems while Trump tops out at 86 percent among his own party. That surprised me given all the hype lately about Clinton bleeding votes from young left-leaning adults to Gary Johnson. You’d expect to see some artifact of that in her polling with Democrats specifically, but nope. You do see it, though, in Monmouth’s split by age: Clinton actually leads Trump by a greater margin among voters aged 50 or older (six points) than she does among those 18-49 (four points). The reason is Johnson, who’s pulling just three percent in the first group but nine percent in the second, although both Clinton and Trump poll several points worse among the 18-49 crowd, suggesting that Johnson is hurting both of them. Clinton can live with that.

Monmouth is a highly rated pollster (A+ on the FiveThirtyEight scorecard), but if you’re looking for reason to doubt the numbers here, they’ve got Patrick Murphy trailing Marco Rubio in Florida’s Senate race by just two points, 47/45. Murphy has hit 45 percent just once this year and that came at the end of July. Rubio started the day leading him by an average of 6.4 points, which helps explain why Democrats have already seemingly given up on the race. If Monmouth sees Murphy as stronger than everyone else does, maybe they’re expecting a bluer electorate this fall than most other pollsters are, which would help explain why Hillary’s leading comfortably here. Relatedly, yesterday’s splashy NYT/Siena poll of Florida found Clinton clinging to a one-point lead, which is much more in line with other recent polling. I highly, highly recommend reading this NYT analysis of the Siena poll to understand how different pollsters can arrive at different results even when they’re armed with the exact same data. The assumptions you make about demographics and who’s likeliest to turn out can produce everything from a one-point lead to Trump to a four-point lead for Clinton.

There’s one result in Monmouth worth flagging. They asked people whether they’d heard about Trump’s recent Birther walkback and, if they did, whether they thought he was sincere or retreating on it out of simple political interest. Result: 24 percent said sincere, 54 percent said politics. Monmouth then asked if Trump’s walkback made people more or less likely to vote for him. Hmmmm:

That’s not necessarily a disastrous result. If all of the Dems and indies who are giving thumbs down were already #NeverTrump then Trump didn’t lose anything. If all of the Republicans who are giving thumbs up are those squishy college-educated whites whom Trump is trying to claw back from Hillary then he might have done himself some good. Seems pretty clear, though, that most of the public sees the Birther retreat for what it was, i.e. a naked pander designed to make Trump seem more “mainstream.” And it may be that Trump inadvertently helped Clinton simply by handing the media a change of subject after Hillary’s rough “deplorables”/fainting stretch. See this David Drucker post for more on that. “Right when he was starting to show some discipline,” said one GOP strategist, “right when he was starting to gain some momentum and right when he was about to make this race a referendum on Hillary Clinton, Trump goes and finds a way to squander all of it with a self-inflicted wound.” If other polls this week show Clinton’s lead stabilizing, that might help explain way.