Headed into the last week of the campaign, we conducted surveys in Virginia, Minnesota, and Maryland. Each of these states is outside the seven key swing states, and each has a Senate race, with varying degrees of competitiveness among the three. We also felt that each had been under-surveyed by other pollsters in the runup to November 5th.
Our findings
In Maryland, Democrats seem to have little reason to fear. While Republican Senate nominee Larry Hogan manages to peel off a significant number of Kamala Harris voters, the Democratic lean of the state puts this seat just out of reach for Republicans.
President: 56% Harris-33% Trump
Senate: 48% Alsobrooks-40% Hogan
Minnesota is home to Democratic vice presidential nominee Tim Walz, and Democrats do seem to have gotten a small boost from his presence on the ticket—Harris runs much closer to Biden 2020 in Minnesota than in either of the other two states surveyed. In the US Senate race, Amy Klobuchar looks poised to glide to victory, thanks to her bipartisan appeal.
President: 48% Harris-43% Trump
Senate: 52% Klobuchar-35% White
Virginia has the most surprising result of the three states we surveyed. While there have been few surveys of the state, those we have seen have mostly shown Harris performing close to Biden’s 2020 numbers and Tim Kaine running significantly ahead of Harris.
However, our survey tells a different story, with Harris up by .7% (45.2%-44.5%), a statistical tie, and Kaine not leading his Republican opponent Hung Cao by much more (46.4%-44.9%).
This is a good time to remind any Democrats who broke out into a cold sweat reading this that every survey has a margin of error, and ours would still allow for the possibility that Harris is up by the 6 points that the Washington Post poll pegged her lead at earlier this month.
It is also worth noting that the 2020 recalled vote in our survey is only Biden +5.5% in a state he won by 10% four years ago. We decided not to weight by recalled vote, for all the reasons Nate Cohn and others have laid out, and because of our own understanding of the dynamics at play here. But if the electorate looks like 2020, this survey likely underestimates Harris and Kaine by a few points.
In spite of our concerns, we feel it is important to publish outlier surveys like this one. All pollsters get results like these, which fall a few points outside the expected results or the average of other pollsters, and there is always pressure to suppress or play down such results, leading to herding. Given our confidence in our methodology, we see no reason to hide or attempt to manipulate the results. The numbers are what they are.
About The Survey
These surveys were conducted using a combination of live calls and SMS-to-web surveys to a random sample of voters from L2. The sample was weighted to match the age, ethnicity, gender, and education of registered voters in these states. We then narrowed the samples to only likely voters based on their responses to questions about the election and their own likelihood to vote. For Virginia, the likely voter sample has a margin of error of +/-4.3%; for Minnesota, it is 4.24%; and for Maryland, it is 4.34%. For more information, call Brannon Miller at 601.402.3281.
