American democracy relies on polls. Polls give insight to politicians about what the public care about. Polls inspire American voters to turn out if they see that the candidate that they support seem to be behind.

​

Polls also fuel horse-race journalism. They allow political reporters to provide a constant update on how candidates are performing; when one politician is pulling ahead by a few points, and when another is falling behind. Rather than poll results coming out on the first Tuesday of November, presidential poll results come out the moment that primary elections end. They turn elections into sporting events that last weeks and months. Theoretically, they turn election day into a confirmation of what we already know will happen. Yet, somehow they seem to not always get it right.

​

I suggest that pre-election polling is hurting American democracy more than it is helping. These poll results are most often found to be inaccurate and unreliable, yet the American public can simply not get enough. Nate Silver has become an icon in America politics in the weeks leading up to elections- and not just because of his sloppy hair. People decide who to vote for, or if they will vote at all, based on what they hear the “rest of America” is doing. Horse-race journalism fuels intense polarization, and turns voters into “fans” of their favorite “team”.

​

In the 21st century the entire country is conditioned to have every answer to any question at their fingertips. People simply cannot stand to not have all available information in front of them. But when it comes to polling, it seems that any information we are given ahead of election day is simply speculation, even if the public wants to take it as fact.

​

In general, voters are less likely to turn out to the polls if they believe that the election will be a landslide victory, especially if the predicted victory is for the candidate that they support. For instance, Democratic voters in California often feel that their vote does not matter because pre-election polls show that the population of California is overwhelmingly going to choose the Democratic candidate anyway, despite whether they take the time to drive to the polls to vote. It can also work the opposite way.

​

Polling creates a constant pressure to follow the majority. As events arise, like Black Lives Matter protests across the country, it has become commonplace to turn to polls to tell us how that has shifted the vote account. It is important to note that just as preferences change during major events months before the election, surely those preferences may change again. 

​

Many people come to understand polls as a prediction of the future. Polls are meant to represent public opinion at a specific moment in time. When asking someone in September who they are voting for the following November, they can only answer with the knowledge that they have at that very moment. There is no way to predict breaking news or major attitude changes. 

​

As I see it, there is one major factor that will contribute to less accurate polls results in the near future. Early voting, largely through mail-in voting, was used by more Americans in the 2020 elections in unprecedented numbers, and very well may continue to become the dominant form of voting. The earlier someone votes, the less time, and likelihood, they would have to change their opinion based on important news right before the election. As early voting continues to rise in popularity across the nation, it will bring with it even more potential for polls to be riddled with inaccuracies, because people are not voting with all the available information on the date of the election, they're voting with all the available information several days or weeks before the election.

​

Early voters vote with different (and less) information than people who vote on election day. Pollsters historically could rely on people attitudes just days before the election to be generally similar to their attitudes on election day. With the emergence of early voting, voters may have voted up to a month before the election, therefore breaking the trends created by the pre-election polling of their peers.