With Super Tuesday just around the corner, we know one thing about this election: it’s the most riveting in years. But why? What makes it so different from everything that came before?
The obvious answer is the surprising line-up of final contenders. A year ago, the idea that a Donald Trump or a Bernie Sanders could have a serious shot at the White House was unthinkable, especially to the “experts” working off twentieth-century playbooks. Now here we are, and the real question is how the dynamics of public opinion — and influence, in particular — have changed to make this new kind of politics possible.
What is influence? Voters have long formed opinions about candidates in two basic ways: directly, by watching and listening to them through debates, interviews and other campaign events carried on radio and TV, as well as political advertising; and indirectly, by consuming news stories, opinions and other election information offered up by journalists and commentators in the public sphere, along with family, friends and co-workers.
Just a few decades ago, it was easy to say who were the key influencers: the powerful political and media players — both organizations and individuals — who orchestrated and largely defined the election conversation.
What has changed? A few things. Thanks to the digital revolution, the old behemoths of political influence, the two major parties and the traditional media, have lost their former dominance. The parties are still hugely powerful and will wind up nominating two final candidates, one of whom it’s very likely will be the next president. But each party is now contending with a magnetic renegade they didn’t see coming.
This occurred largely through the media, but not just the traditional media. There’s no question the old outlets still exert major influence. When the Pew Research Center recently asked American adults to name the election information source they find most helpful, the number one answer was cable news. Yet only 24% gave that answer. The rest cited sources spread widely among ten other categories of information. Tied for second place after cable were an odd couple: local TV — a longtime influencer in communities across the country — and social media, which have been steadily gaining influence over the last several elections.
In short, the old influence hierarchy has been shattered, replaced by a new mosaic of influence in which social media play a growing role.
So who’s influencing Election 2016? For the last year, our research group, the Laboratory for Social Machines, which is part of the MIT Media Lab, has been capturing and saving all the election news published by a basket of influential media outlets. We also have access to the entire Twitter archive, plus the full “firehose” of about 500 million new tweets each day. We have been crawling this data, pulling the election-related tweets and classifying them in various ways. By applying algorithms to both data sets, journalism and Twitter, we were able to identify the election influencers who loom large at the intersection of news and social media. (For more details about how we did the analysis, see the FAQ at the end.)
Below are the top 150 influencers, as calculated by presence and impact in the election news coverage and the Twitter conversation since last August. In this social media age, individuals have voices on the same platforms as huge organizations, so we have combined them all in a single list.
Some things to note:
- Trump took the number-one influencer spot by scoring highest in all four of the metrics used to construct the list, two for news influence and two for Twitter influence. Further confirmation that he is the master of both domains.
- After Trump, the top of the list is unsurprisingly dominated by other presidential candidates, both those still in the race and some who have exited. However, non-candidates also appear in the upper ranks. For instance, Daniel Scavino Jr., Trump’s social media director, is influencer number eighteen, beating out eight former presidential candidates including Martin O’Malley and Rick Santorum.
- Fox News comes in at number twelve (just ahead of Bill Clinton) as the most influential news organization. And Fox’s Megyn Kelly, famous for sparring with Trump, is the most influential journalist.
- Just because you’re an influencer doesn’t mean your views are carrying the day. Jeb Bush is ranked number 4 even though he just withdrew from the race. That’s because over the nearly seven-month period covered by our data, he played a very prominent role in the conversation. The popular fascination with his withdrawal confirms this.
- Some influencers aren’t active participants in the conversation. Russian premier Vladimir Putin and ex-president Ronald Reagan, who’s deceased, made the list by figuring in the election news coverage (both have been mentioned often in presidential debates). Reagan is also one of a handful of people on the list who are shown without a Twitter handle, because they are not on Twitter (accounts about these people don’t count).
- This is by no means a definitive list of election influencers, but rather an attempt to illuminate an important piece of the new influence landscape. Big-money campaign donors didn’t appear often enough in our data to make the list, though arguably they have more election influence than anyone. No newspaper or magazine journalists made it, though many produce highly influential work. Reason: When their work is passed around online via links, the bylines tend not to be mentioned, so their individual influence wasn’t captured. However, The New York Times, The Washington Post and Time all made the list as organizations.
- The analysis surfaced numerous people who have appeared now and then in election news and social chatter, but it was hard to gauge their impact without data. Now we have it. Examples: Ivanka Trump, Chelsea Clinton, #BlackLivesMatter activist and Baltimore mayoral candidate DeRay Mckesson, Cher and Pope Francis.
- The “Other” category includes many whose influence is entirely Twitter-centric, including political activists of both right and left. It also includes some people and organizations that made it through humor and parody, such as The Onion, @LOLGOP and @weknowwhatsbest, a conservative parody of the White House press secretary.
How does influence look by category? Below is a chart showing how it stacks up, followed by break-out lists for all the categories.
It’s important to remember that the technologies disrupting the public sphere are still evolving, so influence will remain a moving target. We hope that this effort to quantify it in a new way (see below for an explanation of our method) will inspire others to try their own approaches.
Technology aside, influence is by nature always shifting — that’s part of what makes it interesting. By the fall when the election is nearing, the list might look quite different. We’ll let you know!
William Powers is a research scientist at the Laboratory for Social Machines and author of the New York Times bestseller Hamlet’s BlackBerry.
Eric Chu, a graduate student at the Laboratory for Social Machines, developed the analytics and visualizations for this post. Lisa Conn, a student at the MIT Sloan School of Management, also made significant contributions to the project.
Photo credits: PBS, Independent Journal
(Prepared with Eric Chu, Soroush Vosoughi and Prashanth Vijayaraghavan, all of the Laboratory for Social Machines)
How exactly did you measure influence?
We wanted to identify election influencers from both news and social media. Some people and organizations figure prominently in the news, but not so much in social media, while for others it’s the reverse. So we started with two “gates” for becoming a candidate for the list, the news gate and the Twitter gate.
On the news side, you had to be mentioned at least 25 times in the election news coverage we’ve captured since last August 1, or at least three times per month during that same period. We settled on these numbers after testing different thresholds to ensure we got a healthy number of prospective influencers. About 1,200 met these criteria and made it through the news gate.
On the Twitter side, we developed algorithms that allow us to identify and save only tweets that are about the election and also compile a list of those who authored them. So we began by isolating all of those election commenters, many millions of people and organizations. We then multiplied the number of election tweets from each person or organization by the number of people following them. So someone who tweeted frequently about the race but had a relatively small following could potentially make it in, as could a celebrity who only tweeted once but reached millions with that tweet. The 10,000 people and organizations with the highest scores made it through the Twitter gate.
We then took everyone who made the first cut (news or Twitter) and rated them for four different influence metrics: 1) total mentions in election news, 2) election news centrality, 3) total Twitter mentions in election tweets, 4) total number of retweets on election tweets. We weighted these equally at 25% each. The 150 top scorers made the list.
What is centrality?
Centrality is a kind of network analysis. It’s a way of measuring the degree to which a person or other entity within a network is connected to others in the same network. For instance, the PageRank algorithm that drives Google measures the centrality of web pages.
Identifying the most influential people within a human network is a classic use of centrality. In our analysis, news centrality is the degree to which each of our influencer candidates appeared in election news coverage along with other influential people or organizations.
How do retweets show influence?
To be a true influencer, it’s not enough just to make a lot of noise about the election. There has to be evidence that other people are listening to this person or organization, that their voice is rippling, having impact. Retweets are an indicator of that.
Is this analysis part of a larger project?
Yes, our group has built an election analytics machine called The Electome. Funded in part by a grant from the Knight Foundation, the project aims to further public understanding of how politics and news are changing in the digital age, and to show how campaign journalism can move beyond the traditional “horse race” coverage. We have several partners with whom we’ll be collaborating on projects. We’ve already done one analysis with The Washington Post and are planning more. We’re also publishing pieces on our own, including two (here and here) by our colleague Andrew Heyward about how we’re tracking what we call the Horse Race of Ideas.
Some on the list are not exactly household names. Did you check the identities and credentials of the people behind them?
We did not investigate the backgrounds of the influencers. Some of the Twitter accounts are clearly not under real names. Some have content we found offensive. We didn’t eliminate anyone on these bases. This is not a list of “good” influencers or a seal of approval for anyone’s views. Whatever they’ve been saying about the election, if their voice is being heard widely according to our metrics, they were deemed influential and made the list.