Two polling stories

There were nine national surveys from “select pollsters” (which means they meet certain criteria for reliability) released this past week, and they couldn’t have shown a closer race. Three showed a tie, including the Times/Siena College poll, and three showed Harris ahead. Her best, an ABC News/Ipsos survey, found her up by 4 points.

On the other hand, three of the polls showed Trump ahead – his best, the Wall Street Journal poll, found him up by 3 points.

If you just take the simple average of those nine polls, you find Harris ahead by less than a point. The maths behind the Times polling average is fancier than that, but it gives you the same answer.

In nearly every case, the national polls this past week showed Trump gaining compared with the last survey by the same firm.

The Project 538 average puts Harris ahead of Trump by more than a point.

In state polling, meanwhile, Marist showed a close race in Georgia, North Carolina and Arizona — ranging from a tied race in Georgia to a 2-point Trump lead in North Carolina. In each of the three, the polls were close to the Times average and similar to the prior Marist polls of the same states.

The latest polls in the northern battlegrounds stayed close as well. Quinnipiac found the race tied in Wisconsin, with Harris ahead by 4 points in Michigan. Importantly, the two results were quite a bit better for her than the last Quinnipiac poll, which showed Trump ahead in both states.

Why has Harris stayed resilient in the battlegrounds?

One possibility is that she hasn’t. There were a lot of high-quality national polls last week from many different pollsters, but there weren’t many state polls – and most of the state polls came from just two firms. We expect many pollsters’ final state polls over the final stretch; perhaps they will belatedly show a shift towards Trump.

Loading

Another possibility is that her resilience is the result of the intense campaign in the battlegrounds. Millions of dollars are being spent on advertisements there as the candidates barnstorm this small group of states. The rest of the country is not experiencing this barrage, and it’s possible that has contributed to the relative stability of the race there.

There’s a final possibility: one set of polls – the national polls or the state polls – could simply be more accurate than the other. Maybe Harris will ultimately win by a fairly comfortable margin nationwide, or, Trump could win easily in the battlegrounds. It certainly wouldn’t be the first time that the battleground state polls underestimated Trump.

What if the polls are wrong?

There is precedent for the state and national polls to err in different directions.

Loading

In 2012, the national polls underestimated Barack Obama, but were right about his clear lead in Ohio – and therefore in the Electoral College.

The story was flipped around in 2016 and in the 2022 midterms. In those two elections, the national polls were pretty good, while the core battleground state polls were off. And interestingly, they were off in different directions. In 2016, the battleground state polls badly underestimated Trump, while the polls of key states for Senate races underestimated Democrats in 2022.

If the polls err as they did in 2020 or 2022, the outcome could look completely different.

With the race so close, it wouldn’t even take anything like a 2020 or 2022 polling error to yield a different outcome. After all, six of the battleground states are within a point. The polls are never exactly right; if they err one way or another by a point or two, as usual, then one side could win by a decisive margin.

New Senate polls

Is the Senate still in play? It’s a long shot, but the answer might still be “yes”, according to Monday’s Times/Siena polls in Texas and Nebraska.

Republicans led both contests, but not by much.

Senator Ted Cruz, the Republican incumbent, led Democrat congressman Colin Allred by 4 points in the Texas Senate race – same as the 4-point lead Cruz held in our poll earlier this month.

In Nebraska, the Republican incumbent, Senator Deb Fischer, led independent candidate Dan Osborn by 2 points in the surprisingly competitive Senate race.

Sweeping these two seats would most likely give Republicans Senate control, and while Republicans are ahead in each poll, neither lead is especially solid.

Why not?

Let’s start with Texas.

The Times/Siena poll found Trump leading Harris by 10 points – one of Trump’s largest leads in a Texas poll this cycle, up from 7 points in the last Times/Siena poll. There are two basic possibilities here: one is that Trump is gaining, as he is nationwide, and this will help get Cruz over the finish line. The other possibility is that the poll is a bit favourable for Republicans, in which case Cruz’s lead could quickly start to look a little dicey.

Loading

Then there’s Nebraska. Needless to say, a race between a Republican and an independent is a lot more uncertain than a traditional partisan contest. Osborn has a run a strong campaign – if you don’t believe me, watch an ad or two.

Importantly, the race isn’t close merely because there are a lot of Republican undecided voters, as I would have guessed before the poll: only 4 per cent of Republican-leaning voters are undecided, compared with 4 per cent of Democratic-leaning voters. It’s not inevitable that the race will break towards Fischer as undecided voters make up their minds.

Osborn also led among all registered voters, even as he trailed with the likely electorate. Nebraska is not a state that’s accustomed to having a close race – at least outside of Nebraska’s Second Congressional District, where Harris held a commanding 12-point lead in our poll and where a potentially pivotal Electoral College vote is at stake.