Quote:National polls conducted since Monday’s presidential debate have shown Hillary Clinton ahead of Donald Trump by an average of about 4 percentage points — a meaningful improvement from her position before the debate, when she led by just 1 or 2 points.
http://fivethirtyeight.com/features/election-...ng-states/
Now, it’s becoming clearer that battleground state polls are moving toward Clinton as well. These include the first results since the debate from high-quality, live-caller telephone polls; the numbers we’d been getting earlier this week were all from online or automated polls.
Here’s what I wrote on Thursday about what we might expect to see in swing state polls, assuming that Clinton led Trump by 3 to 5 percentage points nationally, as national polls seem to show:
• A 4- to 8-point lead in Pennsylvania, New Hampshire, Colorado, Virginia, Wisconsin and Michigan, which have been slightly bluer than the national average this cycle.
• Somewhere between a tie and a 4-point Clinton lead in Florida and North Carolina, which have been slightly redder than the national average.
• A roughly tied race in Ohio and Iowa, which have been significantly redder than the national average.
So, what data have we gotten since then?
• In New Hampshire, a MassINC Polling Group poll (live caller) had Clinton up by 7 points.
• In Michigan, a Glengariff Group poll (live caller) had Clinton up by 7 points.
• In Florida, a Mason-Dixon poll (live caller) had Clinton up by 4 points.
• Also in Florida, an Opinion Savvy poll (conducted by automated telephone and online) had the race essentially tied (Clinton up by 0.3 points).
• And finally, in Nevada, a Suffolk University poll (live caller) had Clinton ahead by 6 points.
As you can see, these results are pretty much exactly what we’d expect with a Clinton lead of 3 to 5 percentage points nationally.
In fact, they’re mostly toward the high end of the range, which means that her lead over Trump nationally could eventually turn out to be more like 5 points than 3 points as more data comes in.
The most impressive result for Clinton is probably the Suffolk poll of Nevada. I didn’t establish a benchmark for Nevada in Thursday’s write-up because there’s been a divergence between polls and demographics there all cycle, with polls showing it as a Trump-leaning state while demographics imply it should remain Democratic.
But her 6-point lead in the Suffolk poll — the largest lead she’s had in any live-caller poll in Nevada all year — is the sort of number our model was expecting to see there all along. As a caution, Suffolk’s sample sizes are on the smaller side (500 people) so we’ll need to see more data from the Silver State.
Still, the polls have told a pretty consistent story overall. Among the 11 swing state polls conducted since the debate, Clinton has led in all 11.
You may notice that I’ve focused on the top line numbers (“Clinton’s up by 4”) instead of trend lines (“she’s gained 2 points”) in these last couple of updates, because with trend lines there’s more to keep track of.
The period from Sept. 11 through the Sept. 26 debate was one of Clinton’s worst polling stretches of the year, for example, so a new survey from a pollster that last tested the race in that period will probably show Clinton gaining ground since then.
But if a pollster had last surveyed a state in early August, when she was up by 7 or 8 percentage points nationally, you’d still expect Clinton to lose ground since then.
Our models keeps track of all this stuff, of course, although they may not yet have Clinton’s debate bounce fully priced in. Her chances of winning have risen to 67 percent in our polls-only model and 64 percent in polls-plus.
But our hyper-aggressive now-cast has Clinton’s popular vote lead at 4.1 percentage points, as compared with 3.1 points in the polls-only model. Since the now-cast doesn’t need as much data to show a big change, the gap implies that Clinton has some further room to grow in polls-plus and polls-only if we get more polls confirming the results we’ve seen over the past couple days.
1
