Commentary

The Polls are Wrong. Here’s Why It Matters.


The day before the most recent presidential election, I spent hours in my common room skimming over polls and trying to map out every possible outcome of the election. It was stressful to say the least. I remember how early estimates predicted around an 8 percent lead for the Democrats, according to 270towin. However, after votes were cast, the lead turned out to be much less significant; at only 4.5 percent, the disparity between the polls and reality left many people confused.

The 2016 election shows just how great of an extent the polling system is flawed. The vast majority of the polls seemed to predict a Democratic landslide win over the Republicans. A month before the November election, “The New York Times” estimated that Hillary Clinton had a 91 percent chance of winning. In the end, however, the Republican nominee was victorious, with 306 electoral college votes, flabbergasting political pundits and everyday Americans alike.

How did this happen? Was it just a product of inaccuracy, or was there simply a silent majority among us? These discrepancies between actual results and polling data seem to be due to a lack of consistency between surveys.

Polls are all asking different things, and all have different ways of asking. One of the most blatant examples, according to Nate Cohn, a domestic correspondent for “The New York Times,” is in the definition of “Hispanic,” which seems to change from poll to poll. As a result, the public is confused about what polls actually mean, and folks will interpret statistics incorrectly. This is a recurrent issue in polling methodology that has affected many significant elections in the past, and without change, it threatens to impact elections in the future.

This fault in the polling system prevents American democracy from reaching its true potential. Incorrect polling directly affects voter turnout, and in turn, significantly impacts real election results. In 2016, my parents, who voted in every election since they turned 18, chose not to vote. Looking back on the election three months after that, I asked them why they had withdrawn from their civic duty, and they responded, “Well, everyone thought Clinton would win by a landslide, so why go out and vote if it’s already decided?”

When someone is predicted to have a wide lead, their supporters ease up on turnout, assuming that their individual vote will not be impactful. If you are playing soccer and it looks like you’ll win by 15 goals, most people on the winning team would be less driven to score more goals, as each additional one becomes less impactful. Of course, if everyone on the winning team were to sit out, that 15 point lead might disappear. In the past couple of years, early polls sucked Democrats into the illusion that they were 15 goals ahead, and that their victory was therefore secure. However, this oversight allowed Republicans to steal the victory in the end.

Current polling systems are clearly flawed, and as a result, they have drastically impacted American politics. Given that most people acknowledge that the system is flawed, why not fix it before it causes even more damage? Polls aren’t accurate, and they shouldn’t be the model everyone looks to when trying to calculate election results. You will never be able to predict the outcome of any election down to the tee, so at least until polls are sufficiently reformed, we need to stop relying so heavily upon them to predict the outcome of elections and make our voting decisions.