Polls under fire after failing to predict Trump win despite new methodologies

US polling agencies are under fire after the third presidential election in a row for failing to foresee the emphatic ballot box triumph that will propel Donald Trump back to the White House.

- Robert Tait

Having seriously underestimated Trump's support in the 2016 and 2020 elections, polling agencies trumpeted a recalibrated methodology for 2024 that was meant to more realistically reflect his standing while restoring their own credibility.

Instead, pollsters are now being called on to explain a broad range of surveys that had shown two candidates essentially deadlocked both nationally and in battleground states in a race that was deemed too close to call.

Compounding the embarrassment, many experts in the run-up to polling day predicted a narrow electoral college victory for Kamala Harris, who was foreseen by some as just about achieving a win in a majority of the seven key swing states: Pennsylvania, Michigan, Wisconsin, North Carolina, Georgia, Arizona and Nevada.

In fact, Trump had won five of those states at the time of publication and was leading in Nevada and Arizona, which had yet to be called.

Most notable was a poll published last weekend by the Des Moines Register that showed Harris with a three point lead over Trump in the Republican stronghold state of lowa - supposedly fuelled by widespread outrage among female voters over the restriction of abortion rights.

The poll, carried out by Ann Selzer - an Iowa pollster renowned among her peers for reliability - fed Democratic hopes of a groundswell of support among female voters that could potentially carry over to neighbouring Michigan and Wisconsin.

Selzer continued to vouch for its findings even as Trump's campaign dismissed it as a "fake poll" and "a clear outlier".