Digital media played a prominent role in the recent US presidential election, with social media platforms channelling previously fringe universes of political culture, rooted in populism and post-truth politics, right into the mainstream of US political discourse. Meanwhile, traditional mechanisms, from polling to mainstream media, failed to adequately capture public sentiment around political events. Are new instruments needed to understand the socio-technical fabric of the post-truth political landscape? And what can digital researchers do to contribute? Liliana Bounegru outlines examples of approaches being developed at the Digital Methods Initiative that hope to assist digital researchers, data journalists, civil society groups and others looking to increase public understanding of these phenomena.
As Washington tradition dictates, inauguration eve saw a multitude of balls and galas where tuxedo and gown-clad Donald Trump supporters, donors, organisers and fundraisers gathered to acknowledge campaign efforts and celebrate victory. And while these events usually celebrate the efforts of those involved in raising campaign contributions, knocking on doors and organising rallies – this year saw the recognition of a different kind of political work to bring the golden-blonde billionaire to the White House. The “DeploraBall”, hosted by the National Press Club, a self-described “meeting of the trolls”, lauded mobilisations to “meme him into the presidency”. Elsewhere, the owner of a website that generated widely circulated fake news about the presidential election claimed that his site helped Trump get elected, with a BuzzFeed News investigation indicating that fake news did indeed generate more engagement than real news on Facebook.
These claims are illustrative of the prominent role that digital media has played in relation to the US presidential elections – both in practice and in the public imagination. Perhaps most distinctive are the ways in which social media platforms have acted as engines to channel previously fringe universes of political culture, rooted in right-wing populism and post-truth politics, into the mainstream of American politics.
In light of a Trump victory that appears to have been successful in harnessing these digital engines (in a way which was notably different in character to Obama’s 2008 election campaign), there is also a perception of traditional mechanisms having failed in capturing public sentiment around political events, from polls and surveys to mainstream media. How can we account for the failure of these instruments to capture the popular appeal that Trump generated? Are new instruments needed to understand the socio-technical fabric of the post-truth political landscape? And what can digital researchers contribute to the public understanding of these phenomena?
Last month, a group of new media researchers, journalism students, graphic designers and programmers assembled at the University of Amsterdam. The group sought to address these questions by developing ways to creatively repurpose digital traces from online platforms in the service of critical social and cultural research. The aim was to shed light on these emerging digitally mediated political cultures, as well as their associated forms of political action. As eyes turn nervously towards upcoming European elections in 2017, can critical digital research around the US elections be leveraged in order to understand emerging forms of information politics – and the mainstreaming of previously fringe right-wing formations in other countries – before the polls open? Below are a few noteworthy examples of approaches from our ongoing research at the Digital Methods Initiative that we hope might assist digital researchers, data journalists, civil society groups and others looking to increase public understanding of these phenomena.
Mapping fake news sites through linking patterns and tracker signatures
While the notion of “fake news” dominates discussions in the wake of the US election results, what counts as fake news and to whom is far from settled. One approach to understanding this issue is to use linking patterns and tracker signatures to profile fake news ecosystems. An analysis of how websites labeled as fake news link amongst themselves may reveal the various spheres of hyper-partisan information and misinformation that comprise fake news spaces in different languages.
Figure 1: Network map of English language websites labeled as “fake news” clustered according to their linking patterns generated with the Issue Crawler tool. Source: https://wiki.digitalmethods.net/Dmi/WinterSchool2017FakeNewsSites.
Similarly, web trackers, such as ads and analytics, may also be used to map spheres of misinformation. An analysis of the nature of trackers present on fake news sites over time may shed light on the economic underpinnings and business models underlying these sites as well as the types of user data which they capture.
Figure 2: The rise of ad trackers (blue nodes) on a set of English language fake news sites between 1999 and 2016. Data collected with the Internet Archive Wayback Machine Link Ripper and the TrackerTracker tool and visualized with Gephi. Source: https://wiki.digitalmethods.net/Dmi/WinterSchool2017FakeNewsSites.
Mapping the infrastructures and tactics of the alt-right online
Another approach would make visible the infrastructures through which alt-right groups organise online and the tactics they deploy. Here one might devise methods to capture and analyse data from popular alt-right spaces (such as the subreddit, The_Donald). Projects may seek to trace how forms of political action such as “Google bombs” are coordinated to manipulate search results in order to favor some candidates and damage others. A glimpse into the techniques of new right-wing formations may be evidenced by comparing sources and platforms appearing in top posts in popular spaces online – such as these examples from The_Donald.
Figure 3: Network map of targeted individuals and phrases to be associated with their names through Google bombs coordinated on /r/The_Donald. Source: https://wiki.digitalmethods.net/Dmi/WinterSchool2017TrumpAltRight
Figure 4: Network of co-occurring domain names in top posts and associated comments on /r/The_Donald. Source: https://wiki.digitalmethods.net/Dmi/WinterSchool2017TrumpAltRightActivism
Understanding filter bubbles beyond Facebook
Other approaches may look to extend the study of filter bubbles beyond political news on Facebook. This was the focus of a Wall Street Journal interactive project that enables readers to navigate the liberal and conservative political news filter bubbles on Facebook side by side. One way to advance such work would be to develop approaches to demarcate election filter bubbles created by Trump and Clinton supporters on Twitter through the use of hashtags. This may serve to illustrate the media sources specific to and shared by different political camps on social media – in order to inform communications and media outreach activities in relation to elections.
Figure 5: Network map of media sources shared on fringe and conventional hashtags used by Trump and Clinton supporters on Twitter. Source: https://wiki.digitalmethods.net/Dmi/WinterSchool2017BeyondTheBubbleInside
Finally, one might look for common ground between these different political camps and bubbles on Facebook, by mapping so-called page like networks; i.e. networks of likes amongst Facebook pages. For example, in the case of the US elections, identifying pages liked by both liberal and conservative Facebook pages used in the Wall Street Journal’s Blue Feed, Red Feed may form the basis of a “Purple Feed”, which would display content from pages that are liked and shared by both groups.
As we gear up for another gruelling round of elections in several European countries, this kind of critical digital research could become a locus for collaboration between different actors interested in advancing public understanding of the socio-technical fabric of “post-truth” politics. For this will surely be needed in order to equip citizens and civil society groups concerned about the ascent of the new right with the means to understand and respond to alliances, practices, information sources and forms of meaning-making that are emerging on the media that surround us.
This blog post draws on research projects conducted at the Digital Methods Winter School 2017 at the University of Amsterdam. Further details can be found on the website of the Digital Methods Initiative.
Note: This article gives the views of the author, and not the position of the LSE Impact Blog, nor of the London School of Economics. Please review our comments policy if you have any concerns on posting a comment below.
About the author
Liliana Bounegru is a researcher specialising in data journalism and digital methods. She is affiliated with several universities in Europe and the US, including the universities of Amsterdam, Groningen and Ghent, Sciences Po, and Columbia University. She is also a data journalism advisor for the European Journalism Centre where she previously founded and led the Data Driven Journalism initiative and co-edited the Data Journalism Handbook.