In the weeks leading up to and directly following the November 5th election, pollsters have again taken center stage in national politics. Polls create attractive news: they provide pseudo-certainty for voters, political pundits, and candidates in the face of an uncertain future. The “data-driven” nature of polling can add an air of legitimacy to reporting, which pollsters work hard to maintain. George Gallup, founder of one of the earliest and longest-running US public polls, once said, “Most laymen see no difference between forecasting an election and picking the winner of a horse race. In due time these people will be educated to the difference.” Gallup’s assertion of legitimacy, though, ignores a startling fact: since their conception, US political polls have been consistently inaccurate.
Modern polling first emerged in 1932 through simple mail-in surveys. These early polls, carried out by Literary Digest Magazine, seemed to predict outcomes with such great accuracy that one paper joked we should “quit holding elections and accept the Digest poll as final.” In the wake of these successes, many outlets created their polls, each with their own — often secret — methodology for predicting voter behavior. Issues with polling became apparent shortly after. In subsequent years, varying polls asserted with confidence that Franklin Roosevelt, Harry Truman, and Dwight D. Eisenhower would all lose their Presidential bids by considerable margins. In response to each of these failures, pollsters adjusted their methodologies, but high poll error rates have persisted.
Two notable elections in recent history highlight stark issues with polls. In the 1980 Presidential race, polls predicted a close contest between Republican nominee Ronald Reagan and the incumbent, Jimmy Carter. Contrary to these polls, Reagan secured a landslide victory, winning 44 states. Similarly, in 2016, many respected political polls indicated that Hillary Clinton would have a decisive win in the Presidential election. However, Donald Trump defied expectations, clinching a narrow electoral majority to claim victory.
Polling is, of course, not always wrong. Nate Silver, a well-respected statistician and pollster, predicted nearly every state in the 2008 and 2012 Presidential elections correctly. Junkets also saw broad success in their polls for George W Bush’s Presidential runs. Political analysts and journalists argue that polling is a useful tool for understanding the electorate. Over the last decades, campaigns have also relied on public and internal polls to guide their platform and messaging. This often unreliable tool, though, comes at the cost of millions of dollars in spending for journalists and candidates alike.
Reliance on polling — both from the press and campaigns — creates not just a financial suck, though. The elevation of the poll as a rigorous, scientific endeavor has led candidates and the press to shape the narratives of an election entirely around polling numbers. This creates a few glaring issues. Candidates are now driven solely by statistics on the desires of potential voters. Instead of running on a coherent political vision, candidates are driven to maximize their numbers by adopting the most popular political positions — at least, according to their flawed internal polling. Under the pressure of polling, candidates are limited in the issues they can focus on. If an emerging issue has no representation in polling numbers, why raise it? If polling shows swing state voters care primarily about the economy, why not concede on other issues like education, infrastructure, or health to attract voters from an opposing party?
Journalists often elevate polls to lend an air of scientific authority to their election reporting. Viewing polls, though, can create the perception that the election is a foregone conclusion or heavily favoring one outcome, which influences voter behavior. Polls also act as a tool for “sane washing,” presenting candidates’ positions as equally valid or partisan, even when one stance—such as denying climate change—lacks factual or scientific merit. Additionally, polls can oversimplify political landscapes, fostering false dichotomies between candidates who may have similar policy goals but differ in tone or rhetoric, especially on issues like foreign policy.
Between creating incentives for non-substantive policy visions that change with the latest polling, and encouraging horse-race style reporting that ignores nuances and complexities in a given election, applying polling beyond its limits has contributed to a vapid and unenergized political environment.
