The race to win the Minnesota House of Representatives majority between DFL and Republican House candidates is likely to be decided by razor-thin margins.

We see evidence of that in our latest KSTP/SurveyUSA poll in which we asked Minnesotans if they generally plan to vote for a Republican or Democratic candidate. According to the poll conducted May 8-11, 45% of respondents say they will likely vote for a Republican and 43% say Democrat. Another 4% say they prefer other candidates and 8% are undecided.

In our first poll of the year in January, Democrats had a three-point lead in the generic ballot question.

“There is leakage amongst independents and suburbanites away from the DFL. If the Republicans are ever going to take the state House, they need to do well with those two groups, and right now, they’re making progress,” said Carleton College political analyst Steven Schier.

Our poll shows Republicans have a 10-point lead among independents, 40% to 30%, and a three-point lead in the suburbs, 46% to 43%. There is also a sizeable gender gap, with Republicans leading 50% to 37% among men and Democrats leading 49% to 41% among women.

The poll also shows distinct differences between urban Minnesota and greater Minnesota. Democrats lead 56% to 34% in urban areas and Republicans lead 53% to 32% in rural areas.

The poll includes 39% of respondents identifying themselves as Democrats, 35% as Republicans and 22% as independents.

Schier says the final week of the session could impact the elections in November because several issues are in play that impact many Minnesotans, including the ride-share pay issue — a bill in the legislature that could cause Uber and Lyft to stop doing business in Minnesota if it becomes law.

“If Uber and Lyft leave Minnesota that is a huge issue on Election Day,” said Schier.

SurveyUSA interviewed 825 adults from the state of Minnesota May 8-10. Of the adults, 710 were identified as being registered to vote; of the registered voters, 625 were determined by SurveyUSA to be likely to vote in the November general election and were asked the substantive questions that followed. This research was conducted online, using a 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 recall the 2020 presidential vote.

You can see the full survey results below: