The polling industry’s biggest problem? Not speaking the language of the American voter
Far too many public pollsters ask questions in the language of politicians rather than the language of everyday Americans.
In recent weeks, I’ve noted that the election forecasting industry has had problems in every election this century. I’ve also pointed out that a big part of the problem is analysis bias and a reliance on outdated polling techniques.
I’m glad that many people are working hard to find ways to improve election polling and forecasting. But I fear that the obsession with this one issue is leading many to ignore the much larger problem associated with the public polling world.
From my perspective, election polling is the least important and least interesting part of our work. The more important task is lifting up the voice of the voiceless so they can be heard in the halls of power. Unfortunately, Election 2020 showed once again that the political elites have no idea how large segments of the American public view the world. In particular, there has been a failure in official Washington to understand rural voters, those without a college degree and minorities.
Unfortunately, polling mistakes are a big part of that particular problem. That’s because most public pollsters frame questions and issues in ways that make sense to politicians but not the public. A simple example of this problem is polling on whether Roe v. Wade should be overturned. That polling questions pops up repeatedly during confirmation hearings of Supreme Court Justices and in exit polls.
It’s certainly a fair question. The problem is that most voters — 56% — don’t know what overturning that ruling would mean from a policy perspective. If most voters don’t understand the policy implications of overturning Roe v. Wade, a simple question about whether it should be overturned is meaningless (and misleading). A better approach would be to focus on how the public thinks about abortion and related issues. Such efforts show public attitudes are more conflicted and nuanced than hardliners on either side of the debate want to admit.
This is a fairly common problem. Far too many public pollsters ask questions in the language of official Washington rather than the language of everyday Americans. They use terms that have a specific meaning to those caught up in the political dialogue but have an entirely different meaning in the popular culture. To paraphrase George Bernard Shaw, the political elites and America are two nations separated by a common language.
Far too many public pollsters ask questions in the language of official Washington rather than the language of everyday Americans.
The best current example of this phenomenon is the word “socialism.” My latest polling on the topic shows about 30% of voters have at least a somewhat favorable opinion of the term. To many in official Washington, this sounds like support for the policies of Sen. Bernie Sanders and Congresswoman Alexandria Ocasio-Cortez.
But that’s far from the case. In fact, most of those who say they favor socialism don’t think it has anything to do with higher taxes or a more powerful government. Few who like the term think it has anything to do with economics (and those who say they like socialism overwhelmingly favor free markets).
Those who think of socialism as it has been historically understood are strongly opposed. That includes many Hispanic and Latino voters who turned away from Democrats in 2020 because of the party’s flirtation with socialism.
What’s really dangerous about this is that political elites don’t recognize the translation error. Support for socialist policies is not growing nearly as much as progressives hope or conservatives fear. In reality, only about 10% of voters favor socialism as it has been historically understood.
One of the best examples of how the language barrier creates real world problems can be found on the issue of immigration. The elite misunderstanding of how Americans view this issue is staggering. And it’s a major reason that Donald Trump was able to win the White House in 2016.
For example, a poll released earlier this year was presented this way: “By a more than two-to-one margin, Americans see immigration into the U.S. as mostly a good thing for the country.” A subsequent HuffPost analysis said the poll “finds that the American public is — and remains — broadly pro-immigrant. As the Trump campaign and Republicans continue to run hard on xenophobia, the polling offers a reminder that the same anti-immigrant strategy that failed from 2017 through 2019 may again backfire with the American public in 2020.”
The real-world results in both 2016 and 2020 contradict this understanding of the issue.
The problem starts with the polling question that was asked (and others that were not asked). Simply asking if immigration is a good thing ignores the way most Americans view the issue. Eight out of 10 consistently say that legal immigration is good for America. That’s much higher than the 54% found in the HuffPost survey. At the same time, 8 out of 10 voters consistently say that illegal immigration is bad for the country.
When you put the results together, you find that a majority of voters see a fundamental distinction between legal and illegal immigration. One is good, the other is bad. So, if you are part of a majority that sees a big difference between legal and illegal immigration, how do you answer a question that does not acknowledge that key distinction?
Donald Trump understood the difference. So did the vast majority of voters who want to put a stop to illegal immigration. Such policies are seen by most as common sense, not xenophobia. The political elite’s misunderstanding causes real policy harm. Not only that, the polling industry suffers when it describes public opinion in ways that make no sense to the public at large.
The same thing happens on issue after issue. This flawed use of polling is contributing to the dysfunction of our political system.
So, while it’s good that many pollsters and analysts are looking for ways to improve election forecasting, the bigger challenge is for pollsters to learn the language of the American people. Rather than trying to figure out which team’s political talking points garners more support; pollsters should use their tool to help the political elites hear what the American people are really saying. That starts with asking questions in the language of everyday Americans.
Scott Rasmussen is an American political analyst and digital media entrepreneur. He is the author of “The Sun is Still Rising: Politics Has Failed But America Will Not.”