Crunching Data, Hidden Voters: How Big Data is Shaping Our Politics

Story by Andrea Perng. Cover illustration by Carolyn Duan.

With Election Day fast approaching, both Hillary Clinton’s and Donald Trump’s campaign teams are rushing to convince as many people as possible to vote for their candidate. Gone are the days of polls and ringing up home phones to disgruntled citizens. Instead, campaigns are turning to a more powerful tool to persuade young voters: social media. And to do that, they focus on something that would tell much more than simple percentages on a poll: analytics.

To obtain insight on how to appeal to certain group, campaigns have increasingly started to use something called big data, which is essentially a large amount of data collected on interested voter groups, from as general as voting history and political contributions to something as granular as what videos they watched YouTube, what Facebook posts and pages they liked, and who they follow and retweet on Twitter. Campaigns then use this information to target ads and decide how much they need to invest in engaging a population increasingly establishing its presence on social media, especially to millennials.

According to research by Ipsos MORI, a third of young people think that the use of social media might influence their political vote. Freshman Sreya Kumar agrees.

“Social media has been one of the biggest factors in the election,” Kumar said. “There’s a lot of things involving emails and the Internet.” She believes that the candidates’ use of social media would affect her opinion on the presidential candidates.

But perhaps most telling about how big data is going to swing the election one way or another is the presidential candidates’ approach to social media.

Donald Trump, the Republican candidate, is known for being a big, bold firebrand, saying and doing outrageous things and getting lots of attention for it. “Crooked Hillary Clinton likes to talk about the things she will do but she has been there for 30 years – why didn’t she do them?” reads one of his Tweets. Perhaps it’s this bluntness that brought the Republican candidate 12 million followers on Twitter and 10 million likes on Facebook.

Hillary Clinton, the Democratic candidate, uses Twitter to provide a parallel to Trump’s bombasticity—most of her posts are relatively calm, and occasionally she drops a few memorable lines, as was the case of the great “Delete your account” debacle of June 2016.

Despite being millions behind Trump in follower count, Clinton’s team has at their side their own big data foundation, which is about “building as big a list as humanly possible,” Mitch Stewart, senior strategist for Clinton’s campaign, told Bloomberg in April 2015. Stewart hopes that this list of people who are enthusiastic about Clinton’s campaign will allow the campaign “to set the framework or the foundation for an eventual national effort” and “provide the campaign with an additional head start.”

To utilize this data, Clinton’s campaign applies something that doesn’t cross many minds daily: micro-targeting. Microtargeting uses the digital trail you leave on a certain website, sometimes in the form of small bits of information called cookies and sometimes through its data feed API, to gear advertisements towards you. You see it in action when your Facebook feed is 90% The Onion while your friend’s is almost completely made of AJ+. You see it when you get advertisements on a page completely unrelated to the website you’re on. Micro-targeting has been used in marketing for years, and it’s become increasingly common in politics.

Clinton, however, isn’t the first one to use big data. Usage of big data was made big when Obama’s campaign started collecting data from voter files and consumer databases. Based on the data they assembled, Obama’s campaign team assigned potential voters scores of how strongly they would support Obama. Elan Kriegel, formerly part of the Obama campaign and now managing analytics in Clinton’s campaign, told Pacific Standard in 2014 that out of the four-point margin that Obama won by back in 2012, up to “two points” could be attributed to the Obama campaign’s use of analytics and micro-targeting.

Dan Wagner, formerly the chief analytics officer for the 2012 Obama campaign and now head of Civis Analytics, estimates that the campaign had 15 terabytes in data files from potential voters, from which the campaign sorted through which Facebook pages were liked, events attended, volunteer contacts and where the money they donated went. Hillary Clinton’s campaign is doing the same thing: using big data, Clinton’s statistics and analytics team determines what they should buy so their message gets to as many receptive potential voters as possible.

So does big data actually give Clinton an edge over her opponent?

With Trump shoving aside the possibility of data analytics as “overrated” back in May, members of the GOP are starting to get worried about falling hopelessly behind the modern day culture of fast-paced data-driven campaigns. “If you’re not prepared for it, you can’t catch up,” former digital director for Mitt Romney’s 2012 campaign Zac Moffatt said in an interview with POLITICO. “You can’t have a baby in three months, that’s just the reality of life.”

Junior Joshua Citijaya, who uses Facebook regularly, also believes that targeted social media posts would clearly influence the opinions of undecided voters.

“If I hadn’t already made up my mind through things I found out on my own, I probably would’ve been influenced by [the presidential candidates’ social media activity] for sure,” Citijaya said.