Behind the big stage, where in two weeks a decision will be made about the new US president, public opinion pollsters are fighting hard to ensure that the estimates they have published thus far closely match the outcome.
They are battling to restore their reputation, which suffered setbacks in the 2016 and 2020 election cycles.
Regarding the latter, the American Association of Public Opinion Research described them as "the profession's biggest misfire since 1980," as polls predicted a close outcome, but Ronald Reagan emerged victorious with a landslide victory over the current president, Jimmy Carter.
Pollsters today do not want a repeat of 2016, when most of them gave Hillary Clinton a much higher chance than Donald Trump, nor the last election, when polls at both the national and state level showed a much greater advantage for Joe Biden over Trump than the later election result showed.
Agencies that conduct polls improve their methods from election to election, rely more on modern technologies, and eliminate factors that previously led them to make mistakes.
Despite this, we should still approach the results they publish with caution and not take them for granted. The reality of the election day often denied them.
Tight race and margin of error
The difference in support between Donald Trump and Kamala Harris has long been quite small in most surveys, and this creates one of the biggest challenges for pollsters.
Is the margin of error that the polls publish with their results small enough to show their findings as reliable? If the survey showed that the difference in support for two candidates is only a few per cent, what is the margin of error? Are such results even useful for the public to conclude the true balance of power between the candidates?
In a study on the eve of the 2016 presidential elections, The New York Times, based on a comparison of the results of more than 4,000 surveys between 1998 and 2014 with later election results, determined that the margin of error is not two to three per cent, as is often stated, but twice as high (6 to 7%).
The possibility of a large margin error cannot be ignored today
The pollsters' methodology, particularly sample selection and data collection technology, have improved over time, so we must be fair to them and take this into account.
But the possibility of a large margin error cannot be ignored today. In the balance of power between Kamala Harris and Donald Trump, which is often measured by only a few percentage points, this survey feature could play a crucial role for their reputation after the November 5 election.
Fewer phone and more online surveys
According to the Pew Research Centre, most pollsters have technologically advanced their data collection process and rely less on the once-dominant method of telephone interviews.
In 2000, nearly all 29 pollsters conducted live telephone surveys; today, only four organisations still do so. Despite the increase in the number of national public pollsters, most, including the Pew Research Centre, now rely on online surveys or a combination of online and telephone polling.
The good news is that people with deep knowledge of polling are working hard to fix the problems exposed in 2016 and 2020
This approach significantly reduced and nearly eliminated the possibility of error, which had been a long-standing issue for researchers obtaining results from a survey of landline phone users, typically senior citizens.
"The good news is that people with deep knowledge of polling are working hard to fix the problems exposed in 2016 and 2020, experimenting with more data sources and interview approaches than ever before," stated Pew Research Centre researchers.
Are there Trump's shy voters?
Donald Trump's re-entry into the election presents a major challenge for pollsters. The challenge lies in identifying the quantity of "shy voters" who support the former president and the current Republican candidate.
The primary reason for the insufficient precision of the surveys in 2016 and 2020 was the significant number of "shy voters" supporting Trump. These are individuals who, allegedly, feel embarrassed to inform pollsters that they intend to vote for Trump, yet intentionally provide inaccurate information about their political preferences.
Donald Trump's re-entry into the election presents a major challenge for pollsters
Some analysts—even those from the Democratic camp—rejected this theory as a myth in 2020, using it as an excuse for the mistakes they made organising their polls in the previous two election cycles.
We will see who among the organisers of the survey "hit" and who among them "missed" the final result, regardless of the fact that most experts persistently say that their job is not to predict the future, nor should we expect that from pre-election polls.
However, their daily polls once again shaped the public's expectations based on the results they witnessed in the media. Some of them even base their decisions on the surveys they have seen, a phenomenon known as the bandwagon effect, where they follow the majority's opinion.
This is the group that will be most frustrated with the opinion pollsters if their decision turns out to be wrong.
