(KSTP) — Democrat Kamala Harris has an eight-point lead in our final KSTP/SurveyUSA poll of the 2024 presidential race. Harris leads Republican Donald Trump 51% to 43%, with three percent undecided and two percent preferring another candidate. The poll has a credibility interval, or margin of error, of +/- 4%.

Harris led by six points in our survey in late September and has led every survey since she entered the race in late July and named Minnesota Gov. Tim Walz her running mate in August.

“This is a snapshot,” Carleton College political analyst Steven Schier said after reviewing the results. “We’ll see next Tuesday what the actual electorate looks like.”

Schier says he was surprised so few respondents identified themselves as “independents” in the poll, a group where Trump holds a one-point edge. Forty-two percent identified as Democrats and 35% as Republicans.

“Usually about a third of Minnesotans say they’re independent, less than 20% say they are in this survey,” Schier says.

Schier also notes an anomaly in the gender demographic breakdown where Harris and Trump are now tied at 48% among men and Harris leads by 16 points among women. Last month, Trump led men by 12 points and Harris led women by 18.

“What we see is a collapse of support for Trump among men of Minnesota in this survey,” Schier says. “A huge shift in just a few weeks which seems unusual. Likewise, Harris’ decline amongst women seems curious in a few weeks.”

One part of the poll remained true to form. Harris leads by 33 in urban areas of the state and by 13 in suburban areas. Trump leads by 23 in rural areas.

“You have very tight races in all the swing states and nationally and I think what you’ll get is probably a narrow Harris victory in Minnesota,” said Schier. “I don’t expect an eight-point margin.”

SurveyUSA interviewed 900 adults from the state of Minnesota 10/24/24 through 10/28/24. Of the adults, 801 were identified as being registered to vote; of the registered voters, 728 were determined by SurveyUSA to be likely to vote in the November general election and were asked the substantive questions which follow. This research was conducted online, using nonprobability sample of online adult panelists chosen randomly by Lucid Holdings LLC of New Orleans. The combined pool of survey respondents was weighted to US Census ACS targets for gender, age, race, education, and home ownership, and to recalled 2020 presidential vote.