Home >Opinion >Columns >Pollsters cannot seem to read the US voter’s mind right

The US had more than 230,000 deaths due to covid-19. More than 10 million people were infected with the virus. The economy is not in a good shape at all. With such a disastrous performance on the health front, it should have been easy to write off President Donald Trump’s chances in the US presidential election. So it is not surprising that most pollsters predicted that Donald Trump will lose. But to the surprise of most pollsters, Trump seems to have a big chance of re-election. As this goes to press, there is no clear winner, but the race is far closer than many expected. Why do pollsters find it so difficult to read voters’ minds?

In 2016, while most pollsters predicted that Hillary Clinton would win, one pollster, The Trafalgar Group, called that election for Trump. The same pollster was one of the very few who predicted a Trump win again in 2020. What makes its research methodology so different from other polling agencies? According to Robert Cahaly of The Trafalgar Group, the belief that guides its research methodology is that people are not honest with pollsters. This is because of a “social desirability bias".

A social desirability bias is the tendency to under-report socially undesirable attitudes and behaviours, and to over-report what are considered desirable. It is a type of response bias in which people tend to answer questions in accordance with how their answers will be viewed by others, instead of replying truthfully. Respondents may give you answers that are more politically correct. The economist Timur Kuran, author of the book, Private Truths, Public Lies: The Social Consequences of Preference Falsification, says this behaviour tendency to falsify one’s real opinion under perceived social pressure is very common in everyday life. Preference falsification is far more easier in stating one’s voting preference because, unlike many other decisions which can be verified later, one’s real voting behaviour will always remain anonymous.

During an election, voters may experience the “hovering pencil effect". Although they feel one candidate may be the best, there is almost always some attribute of an optional candidate that makes voters think twice. In The Elephant in the Brain: Hidden Motives in Everyday Life, its authors Kevin Simpler and Robin Hanson remind us that in many a decision we take in life, there is a hidden, mostly non-conscious motive that really influences one’s final decision. Many a time, that motive will be a socially-sensitive one, and it might not be considered politically correct to reveal in public.

In America’s 2016 polls, this elephant-in-the-brain issue was probably gender. It was the first time that a major US political party had fielded a woman candidate. Voters had to decide whether they really wanted a woman as their president. Very few people like to publicly admit a gender bias. So, when pollsters asked for their preference, they gave the politically correct answer. But in the anonymity of polling booths, they may have felt free to express their hidden biases. Hillary Clinton lost a US election that produced its widest gender gap ever among voters picking a president.

This election, one might think that the issues that determined voting behaviour were Trump’s pandemic management, the state of America’s economy and each candidate’s tax policies. But the likely elephant-in-the-brain issue was race: Who supports my race? Yet again, it is not politically right to admit a strong racial bias in public. But deep within, many are racially biased. Several incidents leading unto this election stoked racial tensions. Pollsters failed to pick it up, or fix their data appropriately to account for it, but this crucial factor influenced who Americans voted for.

This had happened long ago, too. During California’s 1982 election for its governor, an African-American called Tom Bradley was leading in all pre-poll surveys. But when the results came in, he lost. Social scientists realized for the first time then that it happened because some Caucasian voters who intended to vote for a candidate of their own race would nonetheless tell pollsters that they were either undecided or likely to vote otherwise. This was called “the Bradley effect". This year’s White House election has also demonstrated facets of the same effect.

There is yet another reason why what they say in pre-poll surveys differs from how they behave on voting day. They are often not aware of why they really prefer one candidate over another. Given our own rationalizing selves, we want to believe that we vote for a candidate based on rational matters like economic policies, administrative capabilities, and so on. But the truth is that decisions to vote for a candidate are based on reasons one might not even be aware of. As Herbert W. Simons noted in his book Persuasion in Society, there was a significant change in the preference of women voters for Al Gore in the 2000 US presidential election after his 4.8-second lip-to-lip kiss with his wife at the Democratic Party convention. It is very unlikely that many voters could even know that this incident actually influenced their final voting behaviour.

The difficulty that most pollsters seem to have had in predicting the winner of this year’s US presidential election is yet further proof that human behaviour is too complex a phenomenon to decipher. There are many non-conscious factors that influence any human decision. Uncovering the real reasons, the elephant-in-the-brain issues, that go into any human decision is almost always a very tough task.

Biju Dominic is the chief evangelist, Fractal Analytics and chairman, FinalMile Consulting

Subscribe to Mint Newsletters
* Enter a valid email
* Thank you for subscribing to our newsletter.

Never miss a story! Stay connected and informed with Mint. Download our App Now!!

Edit Profile
My ReadsRedeem a Gift CardLogout