Voters cast their votes during the US presidential election in Elyria, Ohio, on November 8, 2016. REUTERS/Aaron Josefczyk
The polling community has taken a lot of heat right now, given the failure of forecasters and data journalists to predict the Trump triumph in the 2016 elections. Politico, USA Today, the Guardian, and many other headlines asked how the polls could have been so wrong. CNN’s Jake Tapper said the polling and projection industry would be put out of business and the Huffington Post pronounced “The Death of Polling.”
It will take a few more weeks to sort out what exactly went wrong based on actual voting data, but people are offering some explanations based on exit polls. These reasons include some analysts’ underestimation of late deciders breaking for Trump in key states, effects of non-response bias especially if non-responders were inclined toward Trump, low voter turnout, larger than usual state polling errors, reluctance to admit voting for Donald Trump (shy Trumpers), and likely voter models that failed, among other possible factors in inaccurate projections.
My survey team at the Chicago Council Survey is lucky. We aren’t tasked with predicting the outcome of elections. To borrow a phrase from A Game of Thrones, I thank the old gods and the new for this. We monitor trends in American attitudes toward foreign policy issues – whether everyday Americans support international engagement, a military presence abroad, and participation in NATO and other international organizations. We measure the extent of concern across the country over climate change, immigration, and conflicts overseas. We don’t focus on two to three percentage point differences. It must be a nerve-wracking enterprise for my fellow survey researchers who have to do so.
But the failings of 2016 predictions aside, the national surveys (versus some state polls) did show a close race and generally were not off the mark—in fact, they look to be closer to the final mark than national polls were in 2012. Moreover, survey research still plays an important role in our democracy beyond predicting elections. Giving every person an equal chance to voice their opinions on subjects often reserved for the experts - such as foreign policy - is a rare equalizer in politics. This is perhaps even more important now than ever before. While Donald Trump won the election in the end, let us not forget that the popular choice was Hillary Clinton. And the 2016 Chicago Council Survey shows that a majority of Americans sided more with Hillary Clinton’s take on US international engagement – not with Donald Trump’s vision.
For example, people might assume that since Donald Trump won the presidency, most Americans support building a wall with Mexico and deporting undocumented immigrants given his positions. But in fact, our surveys and other polls have consistently found Americans are either deeply divided about or opposed to building a border wall with Mexico, depending on the survey. And when it comes to the eleven million unauthorized immigrants living in the US, a majority of Americans favor offering them a pathway to US citizenship.
Moreover, while Trump was particularly critical of international trade and globalization, Americans think globalization has been mostly good and generally support free trade (though they say it has been bad for jobs). And though Donald Trump has criticized NATO and alliance partners, Americans want to maintain the US commitment to the alliance and support maintaining US military bases overseas.
In this way, public opinion polls can help to counter loud, vocal minorities that are pushing an unpopular position – and even a future president – by exposing the desires of the majority. Surveys can also alert policymakers about the limits to public support. Nate Silver (of FiveThirtyEight fame) expects that Donald Trump will be “very obsessed” with his polling numbers and approval ratings and that “if he detects public pushback against his agenda, he might abandon things pretty quickly.” In other words, public opinion can help shape policy decisions at various points in the process.
In today’s hyper-partisan media environment, some may misinterpret intensely ideological commentaries expressed through internet and digital media as stances that speak for the public. Here accurately-designed surveys can give voice to people in the middle, including those who are less partisan and less committed to a particular position. George Gallup argued back 1957 that without scientifically-based representative probability surveys, policymakers would be guided “only by letters to congressmen, the lobbying of pressure groups, and the reports of political henchmen.”
So yes, the polling community is on alert and is working to improve methods and try to overcome the challenges posed by super low response rates in the United States. The 2016 elections made the challenges very clear (and painful). Survey consumers need to be aware of the shortcomings and pitfalls of surveys and differentiate between well-designed and poorly-designed surveys. But we should not lose faith in all public opinion surveys. They serve an important purpose in keeping policymakers accountable to and in touch with the views on Main Street. Surveying the views of all Americans tries to ensure that those with quieter voices are considered alongside interest groups, lobbyists, activists, or the one percent.