The staying power of public opinion polls

When the initial returns of the ballots cast in the US elections showed incumbent president Donald Trump leading over Democratic candidate Joe Biden in the popular vote and in Pennsylvania, Michigan, and Wisconsin, the political analysts in broadcast network CNN’s panel of commentators began to talk of the end of public opinion polling. All the major pollsters had forecast a Biden victory in the popular vote as well as in the three states.

Bokat-Lindell, a staff editor of The New York Times, wrote on Nov. 5, “Less than 24 hours after the first ballot counts began to pour in on election night, the media and political classes had already declared a loser: polling.”

On the same day, W. Joseph Campbell of The Chicago Tribune wrote: “Election polling is facing yet another reckoning following its uneven-at-best performance in this year’s voting. Although the outcome in the 2020 presidential race remained uncertain through Thursday morning, it was evident that polls collectively faltered, overall, in providing Americans with clear indications as to how the election would turn out.

“And that misstep promises to resonate through the field of survey research, which was battered four years ago when Donald Trump carried states such as Wisconsin, Michigan and Pennsylvania, where polls indicated he had almost no chance of winning.”

Margaret Sullivan, noted in her column in The Washington Post: “Polling seems to be irrevocably broken, or at least our understanding of how seriously to take it is.” Jeff Kowalsky of The Boston Globe posted, “America’s pollsters missed the mark again.”

Poll-bashing has been the favorite pastime of many prominent American journalists since the early days of public opinion polls. Walter Lippmann, a legend in the world of journalism, wrote in 1936 that election polls were “a nuisance.”

In 1952, broadcast giant Edward R. Murrow said on CBS Radio that pollsters “have been undone again” after Dwight Eisenhower was elected president in a landslide victory, contrary to the projections of pollsters. Eric Sevareid, the longtime CBS News commentator, confessed to “a secret glee and relief when the polls go wrong.”

Arianna Huffington, founder of Huffington Post, was a poll-basher, too. From the late 1990s into mid-2000s, she conducted a campaign to “get the dominance of polling out of our political life.” In 2004, Jimmy Breslin called pre-election polls “monstrous frauds” in his syndicated column in New York City’s Newsday.

Bashers of political polls went to town when Donald Trump was declared winner in the 2016 election, contrary to the pollsters’ forecast of a Hillary Clinton victory. Actually, Clinton received more votes than Trump as the pollsters projected. But the winner in the US electoral system is not based on who garners the most votes. The votes cast by electors of the Electoral College determine the winner.

In the recently concluded electoral process, the pollsters called it right, with Joe Biden winning over incumbent Trump in the popular vote and in the states of Pennsylvania, Michigan, and Wisconsin although they were a few points off. But American elections are decided by swing states in the Electoral College, not the popular vote, and it is at the state level that polling data missed the mark.

As David Byler, data analyst and political columnist at The Washington Post, wrote two days after the election, “After Donald Trump’s surprise victory in 2016, the 2020 election became a referendum on both Trump’s first term and the pollsters and forecasters who appeared to underestimate his chances four years ago. That first question hasn’t yet been settled. But many were quick to declare a verdict on the second. “But if it was too early to declare a winner Wednesday, it was also too early to declare a major failure of polling. Polling may… err, (but) it’s still better than any conceivable alternative.”

“The apparent results of the presidential election fall within the ‘likely range’ of outcomes projected by polls and forecasts. Those ranges exist for a reason: For all that some readers treat these mathematical analyses as dispatches from the future, that’s an impossible expectation.

“The nature of public opinion research, which uses a sample as a stand-in for a larger entity such as a congressional district, state or the whole nation, means that polls are more shotguns than sniper rifles: They don’t have long-range precision, and they produce a spread of possibilities rather than one pinpoint prediction. Polls underestimated Trump’s support, but not by as much as you might think.”

The sample must be representative of the larger entity it stands for, which the field reporters and statisticians of CNN call “the universe,” which may be a county, congressional district, state, or the entire nation. Pollsters in the US draw their samples from telephone directories. Pollster Pew Research Center estimates only about 3% of households in the US do not have access to any phone.

To ensure that the samples represent all adults who have access to telephone, either landline or mobile, the individuals selected to make up the samples are chosen at random with the use of computers. This is what is called random sampling, a method generally, if not universally, accepted in the physical and social sciences. People who took up advanced studies in those fields will swear to the integrity of random sampling.

As the samples cannot represent the universe exactly, pollsters allow for errors. That is the 3% or 5% error allowed in election projections. In the Philippines, respondents are picked at random from the lists of voters which the Commission on Elections makes accessible to the local pollsters. Here, interviews are done face-to-face or in person. That is why I say election projections, those made by Social Weather Stations and Pulse Asia anyway, are more accurate.

David Lauter, senior Washington correspondent of The Los Angeles Times, filed this report on Nov. 5: “As returns from quick-counting states like Florida and Texas flowed in Tuesday night, pollsters quickly became targets in America’s easily triggered political culture. With Democratic challenger Joe Biden running well behind poll-created expectations in those states, accusations of ‘collapse’ and ‘disaster’ peppered the nation’s pollsters. But as states have counted more results, the picture has shifted: Polls did underestimate Trump’s vote, but by a relatively small amount — less, for example, than they underestimated President Obama’s strength in 2012, when he ran for re-election.”

Besides, the preference of voters changes over time. The changing preference is to a large extent influenced by the events immediately preceding the conduct of the surveys. The survey results reflect the sentiments of the voters at the time the survey is conducted. Pollsters in the US ask respondents the question: “As of today, who would you likely vote as president?” or words to that effect. Respondents’ preferences right after the national conventions may change during the campaign period.

That is why Trump left Walter Reed Hospital over the advice of his doctors to campaign in an effort to change the preference of those who were pro-Biden. He went to industrial cities where he said Biden would lock down the place to control the spread of COVID-19. In oil-producing states, he said Biden would shut down the pumps to protect the environment. In Florida, where there are many immigrants (escapists or refugees from socialist countries like Cuba and Nicaragua) Trump branded Biden an extreme socialist.

Biden countered by going to cities with large populations of African-Americans, like Atlanta and Detroit, where he called Trump a white supremacist. He campaigned hard in Arizona and New Mexico, states with huge populations of people of Mexican descent, reminding them that Trump called Mexicans rapists and murderers.

Such rhetoric could have changed the preferences of many voters, that is why the variance between pollsters’ projections and actual election results. The closer the surveys are conducted to Election Day, the smaller the variance is, as voter’s preference is more firm near Election Day.

Maybe that is the reason why polling has endured all these years despite the acrimonious commentaries of legendary columnists — politicians have more faith in the projections of public opinion pollsters than in the assessment of the political situation made by newspaper columnists.

Oscar P. Lagman, Jr. is a retired corporate executive, business consultant, and management professor. He has been a politicized citizen since his college days in the late 1950s. He was in charge of public opinion surveys of Robot Statistics, the Gallup Poll affiliate, in the early 1960s.