The 2016 Election Is The Most Data Dependent Ever

With the campaigns turning increasingly partisan, data gives a clear picture


Before 2012, Nate Silver’s role was ‘Stats Guy’ at the New York Times. He was a baseball writer, certainly not somebody that you would point to and say they were a political expert. However, after President Obama won in 2008 and 2012, he became one of the most respected political pollsters in the world.

This came about because of the accuracy of his predictions - 49 of 50 states correct in 2008 and 50 of 50 correct in 2012. This is phenomenal given that he was essentially a baseball fan who knew how to analyze baseball data. However, the site created after his impressive results,, has since become one of the most important political sites in the world.

Fivethirtyeight looks at the data involved in elections and political movements, then explains it through clever data visualizations and clear, concise editorials. This election season has seen the site rise to within the top 900 in the world, averaging millions of visits a day.

There is one key reason behind this massive increase in traffic. This has been an election where truth and honesty seem to have been largely ignored. It isn’t necessarily in what is being said by the candidates (although according to politifact one has lied in 52% of their major claims compared to the other’s 12%) but the way that many people are interpreting what is happening.

Unfortunately, traditional polls today are not as accurate as they once were, simply because of the way many are conducted. Take the LA Times poll conducted the week after the first debates for instance, which gave Donald Trump a 4 point lead. This was conducted online, meaning that anybody could vote in it as many times as they liked, which is one of the explanations behind it being the only one showing Trump ahead, whereas a consensus of scientific national polls shows the opposite, with Clinton holding a significant lead.

What Nate Silver’s team have done so well with their site is to amalgamate this data in a way far more thorough way than a simple poll. Rather than discussing overall trends (Clinton is up by 6% etc), they look it at a state level and then make a prediction of who has the greater chance of winning rather than traditional percentage points. This works because there are certain states who are more or less guaranteed for republicans (such as Texas and Oklahoma) and others nailed on Democrat (such as California and Maryland) and it is in specific swing states where the election is generally won or lost.

Fivethirtyeight has managed to stay relatively impartial in this highly poisonous and partisan campaign, instead concentrating on the numbers and largely ignoring the controversies or comments of candidates. This, combined with the accuracy they have had in the past two campaigns has seen them become one of the most respected sources on US elections. They have been named ‘Data Journalism Website of the Year’ for 2016 by the Global Editors Network, ‘Best Political Blog’ from the International Academy of Digital Arts and Sciences in 2013 and Silver was named amongst the powerful people on Earth in 2010.

This success is indicative of the shift towards fact and data driven insights that many non-partisan voters want to see, rather than the clear bias employed by media outlets like Fox and the New York Times. This media bias is then bought across to social media, where it becomes even stronger, until a slightly biased media slant becomes a full blown polarizing issue. Data that has been well explained and clearly outlined, like Fivethirtyeight, is giving voters an accurate picture of the numbers that they could never find on social media and to some extend even traditional media.

If I type ‘Trump polls’ on Twitter, I am told that he is up by 3 or 4 points. If I type in ‘Clinton polls’ I am told she is up by 6. When I look at Fivethirtyeight I can see that Clinton actually has a 78.8% chance of winning when all scientific polls are aggregated. Data being used in this way is a good way to prevent people seeing what they want to see and instead shows them what they need to.

Vision small

Read next:

Big Data Forecasting In Pharma