After Labor Day, StimSight Research polled 498 likely voters in New Jersey’s seventh district from 8 to 12 September, with a margin of error of 4.7 points. In fact, Bennett led 47 to 42. The survey was run by Patrick Murray. He spent years directing the Monmouth University poll before it closed, and this was the first public poll of the race that did not come from a campaign.

So the number carries more weight than the internals of any campaign memo. Even so, one survey is one survey.

Then, on 16 September, Sabato’s Crystal Ball shifted the seat from Toss Up to Leans Democratic and called Kean a modest underdog. Meanwhile Cook Political Report has kept it at Toss Up. The two raters are looking at the same district, which Donald Trump carried by a single point in 2024 and Mikie Sherrill carried by two in last year’s race for governor.

Since last year we have argued that a service record helps Democrats in seats like this one. So it is worth saying plainly that this poll gives that argument almost nothing.

Still, the incumbent is the story in the crosstabs. Kean’s job rating in his own district sits at 37 percent approval against 60 percent disapproval. Trump’s, in the same survey, is 44 and 58. An incumbent Republican polling worse than the president in a district the president carried is a problem no challenger biography created.

Some of that is self-inflicted. Because Kean was away from Congress for months this year without a public explanation, voters had time to draw conclusions. He later said he had been receiving treatment for depression. Kyle Kondik of the Crystal Ball read the damage as coming from the silence rather than the treatment, and that reading seems right to us. Getting care is not a political failing. Leaving a swing district to guess for months is a different matter.

In 2022 Kean won this seat by unseating Tom Malinowski. Then he won it again in 2024 by five points, while Trump carried it by one. That is a real record of running ahead of his ticket. This year he is running behind it.

Kyle Kondik described the district the way most analysts do. It is affluent and highly educated, which is the kind of seat Republicans have struggled to hold since 2016. That matters, because it explains the swing without reference to either candidate. Bennett is running in a district that was already moving toward her party.

Among voters who called the cost of household goods an important factor, 39 percent said Bennett was more likely to bring those costs down. By contrast, seventeen percent said Kean. Meanwhile she leads by nine among unaffiliated voters, by nine among self-identified independents, and by fifteen among moderates. Kean’s clearest strength is voters without a college degree, where he leads by sixteen.

Yet nothing in the public release attributes any of that to her time in the Navy. The pollster’s own summary points at a bloc of voters worried about paying bills.

That is the honest version. A Navy veteran is leading a Republican incumbent in a district Trump won, and the measurable reasons are groceries and an absent congressman. Our thesis can claim the candidate. It cannot claim the lead.

For comparison, two other veteran Democrats sit in seats Cook still calls toss-ups: Cait Conley in New York’s seventeenth and JoAnna Mendoza in Arizona’s sixth. We have not seen an independent public poll giving either of them a lead. So this is the strongest survey evidence any candidate on our board has, which is exactly why the caveats below belong with it.

InsiderNJ sponsored the survey. The New Jersey Globe reported that the outlet is owned by a Democratic National Committeeman and his son, who have together given $10,500 to Bennett’s campaign. That does not make Murray’s work partisan, and his firm is the most credible name in New Jersey polling. It does mean the sponsor has a stake in the result, which is a fact readers deserve alongside the number.

Two more cautions belong here. A five-point lead in a poll with a 4.7-point margin of error is a lead you cannot bank. Then there is the state itself, since New Jersey has embarrassed pollsters in three consecutive statewide elections, a point the Globe made in its own write-up of this survey.

None of that makes the poll wrong. Murray called the race a horse race himself, and he was careful about the bloc driving it. Still, a five-point lead from a sponsor with a stake, in a state where the polling has missed, is a reason to wait for the second poll before anyone celebrates.

Here is the case against reading anything about veterans into this race. The seventh district is affluent and highly educated, the kind of place that has moved away from Republicans throughout the Trump era. Sherrill carried it last year. Kean’s approval is under water. A generic Democrat with no biography at all might be leading by a similar margin right now, and we would have no way to tell the difference.

The narrower version of our argument survives that, barely. It says a service record makes a first-time candidate acceptable in a district that does not default to Democrats, which is a claim about the floor rather than the ceiling. Even so, three in ten voters still have no opinion of Bennett. If her floor holds while the attacks land, the record may be part of the reason, and we will not be able to prove it.

What we can do is refuse the easy version. If Bennett wins in November, the first explanations are the national environment and an incumbent who disappeared. Anyone who tells you the uniform did it is selling something.

Meanwhile both campaigns are working the same small pool of voters. Murray identified it plainly: people who are not locked in by party and who worry about paying their bills. Because that bloc decides the seat, the fall here will be fought over grocery prices rather than over either candidate’s biography. So even a Bennett win teaches a narrower lesson than it first appears to.

Three things. Whether a second independent poll finds the same five points, since one survey with this sponsor is not a trend. Whether Cook follows Sabato, because the two raters disagreeing is itself information. And whether the debate changes the share of voters who have no view of Bennett.

Finally, watch the money. Because outside groups decide late in districts this expensive, the reservations placed in October will show which side believes this rating change.

The race is still close enough that ordinary campaign events can move it. What has already moved is the rating, and the reason was the incumbent.