We have our first two polls since last week’s presidential debate: one national poll and one poll of Pennsylvania.
Combined, they’re a bit of a puzzle.
In the national poll, Kamala Harris and Donald J. Trump are tied among likely voters, 47 percent to 47 percent — a slight gain for Ms. Harris since our last national survey, taken immediately before the debate.
At the same time, Ms. Harris had a four-point lead in a New York Times/Philadelphia Inquirer/Siena College poll of Pennsylvania, 50 percent to 46 percent.
Before getting into the head-scratching details, let’s start with the big picture:
Not much change since the debate. Despite a strong debate performance, Vice President Harris did not gain much ground compared with our last polls of the nation and Pennsylvania. The poll is full of evidence that our respondents thought she did well in the debate — and that Mr. Trump did poorly — but it hasn’t made a big difference, at least for now and at least in our polling.
Pennsylvania, Pennsylvania, Pennsylvania. Ms. Harris may not have gained much, but her campaign will surely be happy with the result in Pennsylvania. The national result, on the other hand, is quite favorable for Mr. Trump (that’s the head-scratching part we’re about to examine). But our elections are decided by the Electoral College, and no state figures more prominently in the electoral math than Pennsylvania.
Now let’s consider our puzzle: a clear lead for Ms. Harris in Pennsylvania, but a tie nationwide? This is unexpected. Four years ago, President Biden won the national vote by 4.5 percentage points, but won Pennsylvania by just 1.2 points. Similarly, our poll averages have shown Ms. Harris doing better nationwide than in Pennsylvania. This poll is nearly the opposite.
Usually, I’d say that this is probably just statistical noise — the inevitable variation in poll results inherent to random sampling. And it might well be, as we shall see. But I think it’s hard to assume that this is simply noise, for two reasons:
It’s what we’ve shown before. It’s easy enough to dismiss any single poll result as a statistical fluke. But we’ve now found similar results in our last two polls of the nation and Pennsylvania.
This is becoming a trend among high-quality pollsters. Yes, our poll average shows Ms. Harris doing better nationally than in Pennsylvania, but if you focus only on higher-quality polls (which we call “select pollsters” in our table), the story is a bit different. Over the last month, a lot of these polls show Ms. Harris doing relatively poorly nationwide, but doing well in the Northern battleground states.
A note on “select” pollsters: To be considered select on our poll averages page, pollsters must meet two of three criteria: a track record of outperforming other pollsters; a transparent methodology; the use of a method that has a chance to reach most or all potential voters. This isn’t a perfect approach (it omits some pretty good polls, and it includes some that aren’t great), but it includes most of the heavy hitters in the industry and it weeds out much of the junk.
When you focus on these higher-quality polls, you get a surprising picture: There are a lot of good polls for Mr. Trump nationally, and a lot of polls showing Ms. Harris doing relatively well in the Northern battleground states like Pennsylvania.
If you just take a simple average of all of the “select” polls fielded since Aug. 1, you get Ms. Harris ahead by three points in Wisconsin and Michigan, and ahead by two in Pennsylvania and nationwide. The Times/Siena poll doesn’t look quite so surprising against that backdrop.
This pattern has been there for a while, but I assumed it was partly or mostly because of timing. Many top national polls were taken soon after Ms. Harris announced her candidacy or just before the debate, while many high-quality state polls were fielded in between. This suggested that the state polls might have caught Ms. Harris at the peak of her post-announcement surge, while the national polls caught her just before and after a political sugar high. Today’s poll makes this interpretation more complicated. And looking back, this was never that simple of an explanation.
The lower-quality polls have generally shown the opposite relationship: better for Ms. Harris nationwide, but better for Mr. Trump in the key Northern states. You may wonder why we include lower-quality pollsters in our averages at all, but many of them do have value, and in some cases they have even been more accurate than polls we typically think of as higher quality. They receive less weight in our average, but they do receive some — enough to sometimes cancel out the “select pollsters,” given their greater numbers.
What’s clear is that recent results from higher-quality polls are very different from those of the last presidential election. If true, it would suggest that Mr. Trump’s advantage in the Electoral College, relative to the popular vote, has declined significantly since 2020.
This wouldn’t come from nowhere: Almost exactly one year ago, I wrote that there were signs that Mr. Trump’s Electoral College advantage edge was fading, including in the 2022 midterm elections. In fact, today’s poll result is reminiscent of our polling ahead of the midterms, which found Republicans leading nationally but Democrats running strong in Pennsylvania and other battlegrounds. It was hard to believe given recent history — I didn’t believe it, and neither did others pollsters I spoke with — but it turned out to be right.
We will re-examine the case for a fading Trump Electoral College advantage soon, including a dive into the geographic distribution of his strength in Times/Siena national surveys over the last year.
Post-debate mirage?
The period after a debate is always fraught for pollsters. One reason: The supporters of the consensus winner of the debate might become especially energized and likelier to respond to a poll.
While the poll didn’t show much shift, there were signs that Democrats were likelier to respond than usual, which may be cause for caution. Overall, white Democrats were 20 percent likelier to respond than white Republicans. (We’re limiting the comparison to white respondents to isolate the effect of partisanship, not race.) And, indeed, Ms. Harris’s strength in the poll was concentrated among white college graduates — the group you might expect to be newly energized after a debate.
There were also a few signs the more committed Pennsylvania Democratic voters were especially likely to respond on the first night or two of polling, immediately after the debate. The proportion of voters who had previously contributed to Democratic campaigns, based on Federal Election Commission records on the L2 voter file, was unusually high over the first two nights — something that was not seen in the national survey. This returned to normal by the end, but there were other signs that Pennsylvanians took unusual interest in the debate: The debate was held in Pennsylvania; its media markets had the highest ratings for the debate anywhere in the country; and Ms. Harris campaigned in the state afterward.
It’s hard to say whether any of this affected the result of the survey, but it suggests the debate at least temporarily influenced who wanted to take a poll. It’s also a reminder that the campaign plays out very differently in the key battleground states.
Russian doll methodology
We did something unusual with this poll: We conducted a poll inside a poll inside a poll.
In partnership with The Philadelphia Inquirer, this national survey was fielded and weighted as three separate polls: a poll of the city of Philadelphia, a poll of the rest of Pennsylvania, and a poll of the other 49 states and the District of Columbia. The three polls were combined (and balanced in proportion to the share of the population).
If you’re wondering, the poll of Philadelphia found Ms. Harris ahead 79-16 (compared with 81-17 for Mr. Biden in 2020).
We have a few other fun national survey designs coming over the next month.
The post Harris Ahead in Pennsylvania and Tied Nationally? Unpacking an Unexpected Result. appeared first on New York Times.