Polls ahead of the November 3 midterm elections point to a Democratic “blue wave,” but the scale of that victory remains uncertain. Democrats lead Republicans by eight points in generic congressional preference, a margin comparable to 2018, when they regained the House with a gain of 41 seats.
The Senate picture is more complicated. Democratic candidates are ahead in Texas, Ohio, Iowa and Alaska—states won comfortably by Donald Trump in 2024—but their leads are narrow and fall within the margin of error. A small polling error could therefore produce very different results, ranging from Democratic control of both chambers to a victory limited to the House.
The uncertainty echoes the 2016 presidential election, when polls substantially overstated Hillary Clinton’s prospects and Trump won Pennsylvania, Michigan and Wisconsin. Polling organizations later acknowledged that they had failed to capture the views of some Trump voters.
Charles Franklin, a pollster who leads a polling institute at Marquette University in Wisconsin, says polling errors tend not to be distributed randomly. If surveys underestimate Republicans in one Senate race, they may do so in other states as well. A shared bias can consequently affect several contests at once.
Recent election cycles offer examples of this pattern. Polls overstated Democratic support in the 2006 and 2014 midterms, while they overstated Republican support in 2010 and 2018. In presidential elections, surveys underestimated Barack Obama in 2012 and Trump in 2016, 2020 and 2024.
Franklin links these patterns to national forces such as presidential approval, the economy and attitudes toward the war in Iran. Those factors can push contests across the country in the same direction. This year, Trump’s popularity is described as being at historic lows amid concerns over inflation, gasoline prices and economic performance. Franklin also cautions that polls may be attracting too many Democrats eager to express their opposition to the president.
The House and Senate are exposed to polling shifts in different ways. In the House, the Democratic advantage in the generic vote is broad and consistent enough that Republicans would need a larger polling error than any seen in the past nine elections to retain their current majority.
Franklin estimates that a Democratic gain of fewer than 15 seats would be disappointing, while a gain of 15 to 20 seats would be plausible. A gain of 20 to 25 seats would represent a strong result, and 30 or more would amount to a major Democratic night.
Republicans begin with an advantage in the Senate. Of the 65 senators whose seats are not being contested this year, 31 are Republicans, and Vice President J. D. Vance would break a 50-50 tie. Several competitive races are separated by fewer than three points, meaning an error similar to those recorded in 2014 or 2020 could allow Republicans to retain control.
If polls have overstated Democratic support, particularly in traditionally Republican states such as Texas and Kansas, Republicans could hold the Senate. If they have underestimated Democrats, the party could win the chamber with 52 or 53 seats. Both outcomes remain plausible, according to Franklin.
A Democratic performance substantially above the polls could turn the anticipated wave into a broader victory, putting districts won by Trump by more than 10 points into play and threatening Republican strongholds such as Kansas, South Carolina and Mississippi. Trump’s decision to campaign in those states reflects unexpectedly weak polling results there.
Franklin warns that these scenarios should not be treated as predictions. Polling errors can vary from state to state, and surveys do not fully account for the individual weight of incumbents, challengers or candidate appeal.
Rather than focusing only on whether a candidate leads by one or two points, Franklin recommends examining how widely polls differ within the same state. When surveys point in opposite directions, the clearest conclusion is that the race is highly uncertain. His advice before election day is simple: voters should expect uncertainty.
Sources: El País Internacional.