It is hard to think of a time before public opinion polling in politics, but try to imagine: the year is 1932, and George Gallup decides to use his skills from his work in market research and advertising to do some polling for his mother, who is running to be Iowa’s next Secretary of State. Shortly after, he founds the American Institute of Public Opinion Polling, which correctly predicts Roosevelt’s victory in 1936 using a sample of 50,000 respondents, contradicting and publicly embarrassing Literary Digest, which predicted a Roosevelt defeat from sampling over two million of its own readers and registered voters.
The stunning prediction, the first of its kind in politics, caused public intrigue. Coinciding with the rise of the social sciences, research was conducted into different methodologies and random sampling (which Literary Digest failed to do) and the revelation of a new and lucrative field led to the emergence of The Gallup Organization and several other polling companies.
It’s hard to overstate the way public opinion polling has changed politics, both during campaigns or once leaders are elected. Near instantaneous polls are fairly reliable ways to capture the collective mood of constituencies beyond just election days that occur every two, four, or six years. While it is true that voting is the constitutional check we have to hold our representatives accountable, strong public opinions can either grant, or deny, leaders different degrees of political capital. During the Democratic House sit-in this past June, lawmakers repeatedly cited the statistic that approximately 90 percent of Americans supported the gun control measures that have been introduced by the Senate. Shortly after his election, Franklin Delano Roosevelt (FDR) – known for his Fireside Chats and for being the first modern president highly popular in his use of media – was the first to use public opinion polling. Roosevelt consulted Princeton professor Hadley Cantril during U.S. intervention in World War II, although hesitantly and secretly.
Today, polling is anything but, especially during this election season – whether results and subsequent analyses making headlines, or internal polling done by campaigns. Recent polls have caused pause, showing that the presidential race between Clinton and Trump may be leveling, with Clinton down to a lead between one and four percentage points. This is certainly reflective of a tough weekend for Mrs. Clinton, and it is difficult at times to discern what degree fluctuations are part of the typical ups and downs of a dramatic campaign cycle, or actually indicative of a greater trend.
While Trump proudly claims he is not like other politicians, he is arguably more captivated by the polls than any – routinely mentioning his lead in them in the Republican primary debates like a fifth grader believing he had literally won the “popularity contest.” If Mr. Trump wins the most important poll, that which occurs in voting booths across America on Election Day, he is going to have a hard time understanding that being popular does not make one a good president, just as being a good president rarely makes one popular.
The Upshot most recently gave Hillary Clinton a 73 percent chance of winning the presidency, which they equate to saying her chance of losing has the same probability that an NFL kicker misses a 45-yard field goal. To continue with the sports analogy, the good news is that anytime between now and Election Day, Donald Trump will likely incur a series of penalties that will push Clinton up into better field goal range. However, this will require that the referees, the voters, call him on it and do not let an already rough game continue to get out of hand.
While we don’t know exactly how the polls will change until November, what is known for certain is that constant exposure to polls can lead to voter apathy – with parallels currently being drawn to the polling done in the weeks leading up to the historic Brexit vote, which many argue made “Remain voters” feel overconfident and stay at home, and underrepresented “Leave voters” in sampling. While it is informative to watch the polls and stay up to date, it is important to keep the canddiates flaws and shortcomings in mind. Most fundamentally, we must remember that polls and forecasts are not deterministic; voter turnout will determine the outcome of this election just as it has in every election, before and after 1936.