The big unknown behind Democrats' polling leads

Poll after poll spells trouble for GOP candidates — in purple states and deep-red ones alike — but while Democrats see a blue wave coming, some Republicans see a polling mirage.Why it matters: Pollsters worry they're flying partly blind. Even a modest polling error could mean the difference between Republicans retaining control of Congress and President Trump facing Democratic majorities in both chambers.The big picture: Polling Trump voters has long been notoriously difficult. The polls underestimated Trump's support in 2016, 2020 and 2024. Midterm polling has been more accurate, though some still underestimated Republican support in 2022, notes Scott Keeter, senior survey adviser at Pew Research Center.Catch up quick: With less than a month until the midterms, the polls paint a bleak picture for Republicans.The House: Democrats lead Republicans 49.8%-41.9% in the RealClearPolling generic congressional ballot average. Cook Political Report projects Democrats will gain nine to 19 seats, enough to win a narrow majority.The Senate: Cook now sees a potential Democratic net gain of two to six seats. Democrats need four to take control, and Cook just moved the Kansas Senate race into its toss-up column.Trump: The president's approval rating stands at 37.6% in Cook's polling tracker. A recent Economist/YouGov poll found 35% approving and 60% disapproving of his performance.State of play: Research suggests Republicans have become somewhat less likely than Democrats to participate in telephone surveys, and response rates overall have plummeted since the late 1990s, per Pew."We definitely have more people answering surveys who voted for [former Vice President Kamala] Harris than who voted for President Trump," says Brent Buchanan, a Republican pollster and founder of Cygnal. "But if you're a good pollster, you still get the other people."Pollsters can compensate by giving greater weight to Trump voters or adjusting samples to better reflect the Republican electorate, which skews male, white and working class.Advanced statistical modeling tools can also take a poor sample and make it look like the true electorate, though results may be inaccurate for subgroups like young people or Hispanic voters, Pew's Keeter says.But that has risks: "The adjustment assumes that the Republicans who respond to a poll are similar to the Republicans who don't respond to the poll," Rice University political science assistant professor Matthew Tyler tells Axios.Tyler said weighting GOP respondents more heavily can backfire if those willing to answer polls are also more willing to vote Democratic than the Republicans who never respond. That could inflate the apparent size of a Democratic wave.Between the lines: Predicting who will actually vote can dramatically change the results.YouGov chief scientist Douglas Rivers found Democratic leads ranging from six to 16 points depending on how likely voters were identified in a September survey — and called the 16-point margin "almost