Polling is officially unreliable. Are pollsters really just getting it wrong (every time) or is it a disinformation campaign?

The Straw Poll booth at CPAC. 2011. (photo: Gage Skidmore)

Polls don’t just fall out the sky. They don’t grow on trees. The don’t spring up fully formed from the earth.

All polls come from somewhere.

Someone, most often a group of someones, will requisition a poll on a certain topic or topics, hire the requisite staff or polling company, set the parameters of the project and get to work. Maybe they want to take the public’s temperature on a number of social and policy questions for the purposes of a political campaign; maybe they want to measure the saturation of an ad campaign.

Pollsters and the people who hire them don’t do their work purely as a service to humanity; they do it for a purpose, usually, at the bottom of everything, that purpose is money.

Political campaigns, like ad campaigns, are very expensive. Pushing a message that falls flat with likely voters doesn’t make for a successful campaign. Losing campaigns, as many politicians have learned to their great cost, are just as expensive as winning ones, sometimes more. It makes sense to find out how voters feel and tailor messages that appeal to them.

Small businesses often want to know how you heard about their products for the same reason. Advertising is expensive: They want to know what worked and what didn’t so they can do more of one and less of the other.

So why are the polls so wrong?

Pollsters have escaped scrutiny over the past few elections, during the courses of which their predictions have fallen spectacularly to pieces. At first, it was downright Donald Trump-level shocking. By 2018, less so. By 2020, and well into 2021, it is understood that polls aren’t what they used to be.

Why not?

Most types of data collection and data analysis have gotten better, more accurate over the years; not less. Advertisers are so good at micro-targeted marketing, it has gotten retailers like Target in trouble for sending moms-to-be mailers for baby products before they’ve even told their families about the pregnancy.

So why is polling becoming so unreliable?

There are two schools of thought on this answer, which is good because there are two types of people in the world; those who hold some group or other up to a standard of perfection and incorruptibility and realists who know the bitter truth.

Woe betide every cockeyed optimist who meets their heroes. Pity those who find their rightful place in the world and “their people” at last. What a huge disappointment it is to find- in an animal rescue group, or professional writers circle, or homelessness outreach network, among missionaries and humanitarian aid workers- that imperfection, corruptibility and rudeness exists everywhere.

That doesn’t mean there aren’t good people engaged in important work; it just means that no group, however well-meaning, laudable or accredited, should be exempted from scrutiny, given a free pass of unquestioned and unquestionable virtue.

Religious organizations, children’s charities, gymnastics empires, non-profit hospitals, scientists, data analysts, number crunchers; no group or organization is incorruptible. Money and power aren’t minor forces that don’t influence humanity very much.

So when it comes to the question of why polls have become totally unreliable, either you think the people who commission and conduct polls are guileless data scientists getting to the bottom of important social and political questions out of curiosity and a purely altruistic desire for the betterment of humankind, or you know the truth: Polls lie.

Sometimes they lie accidentally; sometimes they can be made to lie. A policy question about taxes or clean energy, phrased one way can illicit a “No” from a respondent, phrased another way it can illicit a “Yes” from that same person.

Pollsters and the people who hire them are not oblivious to this fact. In fact, tweaking questions to try to solicit a “Yes” or a “No”, a favorable or unfavorable response, is part of why politicians and media companies hire pollsters to begin with.

They’re tweaking the message until they get something that resonates with their targeted audience. You don’t hire polling companies to help you understand the voter; you hire a polling company to help you win.

Polls are always funded by some organization or another. That organization almost always has a goal and it isn’t education, anthropology or the study of human civilization in 21st century.

Media companies that solicit polls have an audience to please. The more eyes on their newscasts, the more clicks on their online content, the more advertisers will pay. Media companies don’t exist to tell the public the truth or to educate people about current events.

“The Prophet exists to sell itself,” as J.K. Rowling’s Rita Skeeter so succinctly put it. Let’s not be silly and forget that.

For the largely-liberal audience of MSNBC, for example, would a poll showing NJ Gov. Phil Murphy ahead by a double digit margin in New Jersey do better than, let’s say, a poll breaking the unwelcome news that Murphy was in real trouble?

All the polls, for months, have proclaimed Gov. Phil Murphy the front-runner by a double-digit margin. Democrats expected him to win handily. That Murphy came to within roughly a thousand votes of losing is a staggering blow to the polling industry.

Murphy has been projected as the winner of the New Jersey’s governor’s race, so alls well that ends well. Except, no; there is a major problem with these polls which keep turning out to be not only wrong, but are only ever wrong in one direction:

One that favors Democrats to win.

If polls over the last two months in New Jersey had shown Jack Ciattarelli in a dead-heat with Phil Murphy, how might that have impacted the election?

Would Ciattarelli’s supporters have been more mobilized to get voters to the polls? Would donors have been a little more generous with a campaign that had such a good chance of winning?

The answer to both questions is, of course, yes.

Viewed this way, it is difficult not to start seeing this ongoing pattern of bad polling as a sneaky voter suppression tactic.

If someone were to outfit an ice cream truck with a loudspeaker playing the message, “No need to get out and vote, Democrats; the polls say your candidate has no chance of winning, no chance at all,” in neighborhoods across America, the person operating it would probably be arrested and charged with disenfranchising people and depriving them of their constitutionally-protected right to vote.

Conversely, these polls showing Murphy with such a commanding lead could have persuaded a not-insignificant number of NJ Democrats that it wasn’t worth risking COVID-19 to get out and vote.

For the barrage of bad polls the U.S. electorate has endured over the past few years, pollsters and media companies deserve a little more scrutiny. Polling is broken, and even the best pollsters admit it, along with the twin confession that they have no idea how to fix it.

One proposed way is to increase the spread. A poll with a margin of error of 15 points either way could be safely taken with a grain of salt, as the New Jersey race polls in particular should have been.

Pollsters who object to this method argue that a 15 point error margin makes a poll practically worthless. Yes, it does. It would be nice if pollsters noticed and admitted that fact right up front.

News and media consumers should know what they are getting with polls- news or entertainment?

If a poll over-samples Democrats and is likely to heavily favor the Democratic candidate and progressive policy proposals, it should have a disclaimer and disclose that fact. These polls should not continue to be presented as good-faith estimates of how things are really likely to go on election day.

Polling is at risk of becoming known by over half the U.S. population as misinformation at best and disinformation at worst.

(contributing writer, Brooke Bell)