Kamala Harris and Donald Trump Photo: Shutterstock/Gage Skidmore
In the run-up to the US presidential election on 5 November, most polls predict a close race between the candidates, with Donald Trump having an advantage in key states. However, two new polls point to a possible victory for Democrat Kamala Harris, even in states previously dominated by Republicans.
According to a Selzer poll for the Des Moines Register/Mediacom released over the weekend, Harris may unexpectedly win Iowa. The survey, conducted among 808 likely voters between 28 and 31 October (margin of error 3.4%), shows Harris ahead of Trump by three percentage points, 47% to 44%. For context, in the September Selzer poll, Trump was 4% ahead of Harris, and in June – by as much as 18%.
Independent voters, women, and older Americans have a big impact on the outcome in the state. Among women aged 65 and older, Harris is supported by 63% to 28% for Trump, and among men in the same category, she has a slight advantage – 47% to 45%.
Why is Iowa important?
Iowa was not considered a battleground in this election, as Trump won the state in 2016 and 2020. However, a likely victory for Harris here could point to broader trends that could tip the scales in her favour in the national race.
How reliable is this survey?
Selzer’s polls are considered to be among the most accurate in the US. In particular, in 2008, Selzer was the first to predict the victory of the then little-known Barack Obama in the Iowa primary, where he beat Hillary Clinton. According to Forbes, a reputable statistician Nate Silver rates Selzer as the second most reliable polling agency in the US after the New York Times/Siena.
However, other studies have not yet confirmed Selzer’s findings. For example, the Emerson College/RealClearDefence poll, also conducted in Iowa, shows Trump ahead of Harris with 53% to 43%.
Trump reacted to the new data by calling the Selzer results inaccurate and stressing that other polls give him an advantage. His campaign also claimed that this poll, as well as a recent New York Times poll (where Harris is ahead of Trump in four of the seven key states), are allegedly intended to demoralise Trump supporters, The Guardian reports.
An innovative survey from the UK
Another important study was conducted by the British company Focaldata. Over the past month, it surveyed more than 31,000 American voters using the MRR methodology – a regression analysis with post-stratification, popular in Britain after its success in predicting Theresa May’s loss of the parliamentary majority in 2017, Politico writes.
According to Focaldata, Harris has a high chance of winning in Michigan (+5%), Nevada (+2%), Pennsylvania and Wisconsin. Trump is likely to win Georgia and North Carolina, while the outcome in Arizona remains uncertain. Harris’s advantage is explained, in particular, by support from older white women, a traditionally conservative group of voters. At the same time, Trump has strengthened his position among younger male voters of African-American and Latino descent.
Should you trust Focaldata’s data?
Although the data from Focaldata reinforces the Democrats’ position, analyst James Kanagasuriam is cautious about its reliability, noting that Harris’ slight advantage does not guarantee the results on Election Day.
Although the MRR method has gained popularity in British politics, it is not without its drawbacks. Expert Joe Twyman, a former YouGov analyst, explains that the success of polls depends on how well the questions are asked and how accurately the demographic data is reflected. Therefore, if the methodology is inaccurate, no amount of sophisticated analytics can compensate for potential errors.
Thus, although the polls show a strong competition, the elections remain unpredictable, and any outcome could be a surprise for both camps.
