07.04.2017 Views

SAMPLE-field-guide-to-fake-news

SAMPLE-field-guide-to-fake-news

SAMPLE-field-guide-to-fake-news

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

A COLLECTION OF RECIPES<br />

FOR THOSE WHO LOVE TO COOK<br />

WITH DIGITAL METHODS<br />

COMPILED BY<br />

Liliana Bounegru<br />

Jonathan Gray<br />

Tommaso Venturini<br />

Michele Mauri<br />

1


A FIELD GUIDE TO FAKE NEWS<br />

Compiled by Liliana Bounegru, Jonathan Gray, Tommaso Venturini<br />

and Michele Mauri<br />

This <strong>guide</strong> explores the use of digital methods <strong>to</strong> trace the<br />

production, circulation and reception of <strong>fake</strong> <strong>news</strong> online. It is a<br />

project of the Public Data Lab with support from First Draft.<br />

The Public Data Lab (publicdatalab.org) is an interdisciplinary<br />

network seeking <strong>to</strong> facilitate research, democratic engagement and<br />

public debate around the future of the data society.<br />

First Draft (firstdraft<strong>news</strong>.com) is dedicated <strong>to</strong> improving skills and<br />

standards in the reporting and sharing of information that emerges<br />

online.<br />

This is a <strong>SAMPLE</strong> of A Field Guide<br />

<strong>to</strong> Fake News released <strong>to</strong> coincide<br />

with the International Journalism<br />

Festival in Perugia in April 2017.<br />

If you’d like <strong>to</strong> be notified when<br />

the full version is released, please<br />

visit: <strong>fake</strong><strong>news</strong>.publicdatalab.org<br />

For further information you can<br />

email contact@publicdatalab.org<br />

© 2017 Public Data Lab.<br />

The <strong>guide</strong> is released under the Creative Commons Attribution<br />

License (creativecommons.org/licenses/by/3.0/). This means you<br />

can freely share, adapt and draw on this work as long as you give<br />

credit, as per the terms of the license.<br />

If you reproduce or draw on material from this <strong>guide</strong>, we’d be<br />

grateful if you could credit and link back as per the following<br />

statement:<br />

“This article draws on A Field Guide <strong>to</strong> Fake News, a<br />

collaboration of the Public Data Lab and First Draft. For further<br />

details see: http://<strong>fake</strong><strong>news</strong>.publicdatalab.org”<br />

A COLLECTION OF RECIPES<br />

FOR THOSE WHO LOVE TO COOK<br />

WITH DIGITAL METHODS<br />

The copyright for the research that this <strong>guide</strong> draws upon remains<br />

with its respective contribu<strong>to</strong>rs.<br />

This project would not have been possible without the contributions<br />

of researchers from the following institutions:<br />

Centre for Journalism Studies, University of Ghent<br />

Citizen Data Lab, Amsterdam University of Applied Sciences<br />

Density Design, Politecnico di Milano<br />

Digital Methods Initiative, University of Amsterdam<br />

Institut National de Recherche en Informatique et en<br />

Au<strong>to</strong>matique (INRIA)<br />

Institute for Policy Research, University of Bath<br />

Labora<strong>to</strong>ire d’Étude des Sciences et des Techniques (STS-Lab),<br />

Université de Lausanne<br />

Labora<strong>to</strong>ire Interdisciplinaire Sciences Innovations Sociétés<br />

(LISIS), Université Paris-Est Marne-la-Vallée<br />

Médialab, Sciences Po, Paris<br />

Techno-Anthropology Lab, Aalborg University Copenhagen<br />

University of Siegen<br />

Published by Public Data Lab.<br />

Sample released in April 2017.


CONTENTS<br />

Contribu<strong>to</strong>rs<br />

Co-investiga<strong>to</strong>rs<br />

Liliana Bounegru, University of Ghent, University of Groningen<br />

Jonathan Gray, University of Bath<br />

Tommaso Venturini, Univ Lyon, Inria, ENS de Lyon, CNRS, UCB<br />

Michele Mauri, Density Design, Politecnico di Milano<br />

DensityDesign Researchers<br />

Agata Brilli, Politecnico di Milano<br />

Ángeles Briones, Politecnico di Milano<br />

Carlo De Gaetano, Politecnico di Milano<br />

Gabriele Colombo, Politecnico di Milano<br />

Mariasilvia Poltronieri, Politecnico di Milano<br />

Michele Invernizzi, Politecnico di Milano<br />

Michele Mauri, Politecnico di Milano<br />

Paolo Ciuccarelli, Politecnico di Milano<br />

Tommaso Elli, Politecnico di Milano<br />

Facilita<strong>to</strong>rs<br />

Alex Gekker, University of Amsterdam<br />

Anders Munk, Aalborg University<br />

Bilel Benbouzid, University of Paris-Est Marne-la-Vallée<br />

Erik Borra, University of Amsterdam<br />

Esther Weltevrede, University of Amsterdam<br />

Jonathan Gray, University of Bath<br />

Jorn Preuss, University of Siegen<br />

Liliana Bounegru, University of Ghent, University of Groningen<br />

Marc Tuters, University of Amsterdam<br />

Mathieu Jacomy, Sciences Po<br />

Natalia Sánchez, University of Amsterdam<br />

Nicolas Baya-Laffite, University of Lausanne<br />

Richard Rogers, University of Amsterdam<br />

Sabine Niederer, Hogeschool van Amsterdam<br />

Tommaso Venturini, Univ Lyon, Inria, ENS de Lyon, CNRS, UCB<br />

Analysts<br />

Agata Troost, University of Utrecht<br />

Anders Grundtvig, Aalborg University<br />

Anna Keuchenius, University of Amsterdam<br />

An<strong>to</strong>nio Martella, DataActive<br />

Asbjørn Fleinert Mathiasen, Aalborg University<br />

Asger Gehrt Olesen, Aalborg University<br />

Carlo Santagiustina, University of Amsterdam<br />

Charlotte Leclercq, University of Amsterdam<br />

Daniel Bach, Aalborg University<br />

Daniela Demarchi, University of Amsterdam<br />

Ecesu Erol, University of Amsterdam<br />

Emil Grosby Jørgensen, Aalborg University<br />

Emillie Vic<strong>to</strong>ria de Keulenaar, University of Amsterdam<br />

Joep Voorn, University of Amsterdam<br />

Jorn Preuss, University of Siegen<br />

Kaspar Beelen, University of Amsterdam<br />

Katerina Gladkova, University of Amsterdam<br />

Lieke Kersten, University of Amsterdam<br />

Lisanne Blomberg, University of Amsterdam<br />

Manon van Hoek, University of Amsterdam<br />

Maria Hayat, University of Amsterdam<br />

Marlene Scherf, University of Amsterdam<br />

Michel Blonk, University of Amsterdam<br />

Mintsje de Witte, University of Amsterdam<br />

Mischa Benjamin Szpirt, Aalborg University<br />

Pieter Vliegenthart, University of Amsterdam<br />

Rina Tsubaki, European Journalism Centre<br />

Ronja Ingeborg Lofstad, Aalborg University<br />

Sal Hagen, University of Amsterdam<br />

Stefani Mans, University of Amsterdam<br />

Stefanie Voortman, University of Amsterdam<br />

Talia Castellanos Usigli, University of Amsterdam<br />

Zoë Versteegen, University of Amsterdam<br />

Research and Edi<strong>to</strong>rial Assistant<br />

Daniela Demarchi, University of Amsterdam<br />

This is a sample of<br />

A Field Guide <strong>to</strong><br />

Fake News, which<br />

will contain further<br />

chapters on how <strong>to</strong><br />

map and investigate<br />

<strong>fake</strong> <strong>news</strong>, online<br />

misinformation and<br />

other associated<br />

phenomena –<br />

including recipes on<br />

memes, bots, trolls,<br />

propaganda and factchecking<br />

initiatives.<br />

7 Introduction<br />

CHAPTER 1<br />

15 MAPPING FAKE NEWS HOTSPOTS ON<br />

FACEBOOK<br />

18 1. What publics does <strong>fake</strong> <strong>news</strong> animate on facebook?<br />

38 2. How may the trajec<strong>to</strong>ry of a <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry be<br />

traced on facebook?<br />

50 3. Do fact-checking initiatives reach the publics of<br />

<strong>fake</strong> <strong>news</strong> on facebook?<br />

CHAPTER 2<br />

59 TRACING THE CIRCULATION OF FAKE NEWS<br />

ON THE WEB<br />

62 1. Where do <strong>fake</strong> <strong>news</strong> originate? By what sites are<br />

they first retransmitted?<br />

76 2. Which are the most visible sources related <strong>to</strong> a <strong>fake</strong><br />

s<strong>to</strong>ry? When and by whom are they mentioned?<br />

CHAPTER 3<br />

93 USING TRACKER SIGNATURES TO MAP THE<br />

TECHNO-COMMERCIAL UNDERPINNINGS OF<br />

FAKE NEWS SITES<br />

95 1. Do <strong>fake</strong> <strong>news</strong> sites use different kinds of trackers<br />

from mainstream media sites?<br />

101 2. How can <strong>fake</strong> <strong>news</strong> and mainstream media sites be<br />

profiled based on their tracker usage?<br />

107 3. How do tracker ecologies on <strong>fake</strong> <strong>news</strong> sites change<br />

over time?<br />

113 4. Which other websites share the same tracker IDs as<br />

<strong>fake</strong> <strong>news</strong> sites?<br />

119 5. Do trackers associated with hyper-partisan and<br />

misinformation sites vary across language spheres?


A FIELD GUIDE TO<br />

FAKE NEWS<br />

INTRODUCTION<br />

[1] See, for example Robert<br />

Darn<strong>to</strong>n, “The True His<strong>to</strong>ry<br />

of Fake News”, The New<br />

York Review of Books,<br />

February 2017, available at:<br />

http://www.nybooks.com/<br />

daily/2017/02/13/the-truehis<strong>to</strong>ry-of-<strong>fake</strong>-<strong>news</strong>/<br />

[2] See, for example, “Sky<br />

Views: Facebook's <strong>fake</strong><br />

<strong>news</strong> threatens democracy”,<br />

Sky News: http://<br />

<strong>news</strong>.sky.com/s<strong>to</strong>ry/skyviews-democracy-burnsas-facebook-lets-<strong>fake</strong><strong>news</strong>-thrive-10652711<br />

What is “<strong>fake</strong> <strong>news</strong>”? And what can be done about it?<br />

Depending on who you ask, <strong>fake</strong> <strong>news</strong> is said <strong>to</strong> represent<br />

a step-change in information warfare; an emerging form of<br />

cynical profiteering; an engine for energising “alt-right” and<br />

other digitally mediated grassroots political mobilisations<br />

around the world; a partisan battle cry for a new liberal<br />

“ministry of truth”; an unwanted byproduct of the online<br />

platforms which organise our digital societies; or a canary<br />

call signalling a collapse of consensus around established<br />

institutions and processes of knowledge production,<br />

heralding a new “post-truth” era in politics and public life.<br />

According <strong>to</strong> some commenta<strong>to</strong>rs <strong>fake</strong> <strong>news</strong> is just old wine<br />

in new bottles – and similar misinformation phenomena<br />

have existed for at least as long as the printing press and<br />

other communication technologies through which they<br />

circulate. [1] Others suggest that new online platforms<br />

accelerate and “supercharge” their circulation in a way<br />

which introduces hither<strong>to</strong> unprecedented challenges and<br />

dynamics. [2] Proposed responses include new media literacy,<br />

educational and fact-checking initiatives; new laws, policies<br />

and fines for technology companies who fail <strong>to</strong> remove<br />

[3] See, John Naugh<strong>to</strong>n,<br />

“Facebook and Twitter<br />

could pay the price<br />

for hate speech”, The<br />

Guardian: https://www.<br />

theguardian.com/commentisfree/2017/mar/19/<br />

john-naugh<strong>to</strong>n-germanyfine-social-media-sitesfacebook-twitter-hatespeech<br />

[4] See, for example, Full Fact,<br />

“The State of Au<strong>to</strong>mated<br />

Factchecking”, https://<br />

fullfact.org/au<strong>to</strong>mated<br />

and http://www.512tech.<br />

com/technology/austinstartup-wants-end-<strong>fake</strong><strong>news</strong>-and-<strong>fake</strong>-everything-else-the-internet/<br />

EcchWFgrl4PQmjPvmkzycJ/<br />

offending content; [3] and a host of new startups and technical<br />

fixes – from authenticated content <strong>to</strong> au<strong>to</strong>mated factchecking<br />

projects. [4]<br />

Across these different kinds of responses, observers<br />

agree that <strong>fake</strong> <strong>news</strong> cannot be straightforwardly defined.<br />

“Fake <strong>news</strong>” comes in many different shades. This need<br />

not be taken as proof of the futility of investigating this<br />

phenomenon. On the contrary: accepting that there is no<br />

easy way <strong>to</strong> demarcate between “<strong>fake</strong>” from “non-<strong>fake</strong>”<br />

across all cases opens interesting research opportunities. It<br />

is precisely because its forms and contents are designed <strong>to</strong><br />

mimic those of mainstream media – and precisely because<br />

it travels through similar circuits – that <strong>fake</strong> <strong>news</strong> offers us<br />

the occasion <strong>to</strong> study not just the strategies and formats of<br />

<strong>fake</strong>ness, but the politics and composition of the media and<br />

information environments of the digital age more generally.<br />

This <strong>guide</strong> aims <strong>to</strong> enrich public debate and catalyse<br />

collective inquiry around this rapidly evolving and highly<br />

contested issue – by suggesting different ways in which it<br />

can be empirically studied, mapped and investigated online.<br />

INTRODUCTION 7


Ultimately our hope is not just <strong>to</strong> provide better accounts of<br />

the issue of <strong>fake</strong> <strong>news</strong> and phenomena associated with it,<br />

but also <strong>to</strong> contribute <strong>to</strong> more substantive forms of public<br />

engagement around it. We hope this <strong>guide</strong> will contribute<br />

<strong>to</strong> facilitating broader public debate and involvement<br />

around processes of reshaping platforms and policies, laws<br />

and infrastructures, technologies and standards that are<br />

implicated in the circulation of <strong>fake</strong> <strong>news</strong> phenomenon.<br />

This includes remaining attentive <strong>to</strong> possible unintended<br />

consequences of these different responses, as well as other<br />

interests and concerns.<br />

The <strong>guide</strong> explores the notion that <strong>fake</strong> <strong>news</strong> is not just<br />

another type of content that circulates online, but that it is<br />

precisely the character of this online circulation and reception<br />

that makes something in<strong>to</strong> <strong>fake</strong> <strong>news</strong>. In this sense <strong>fake</strong><br />

<strong>news</strong> may be considered not just in terms of the form or<br />

content of the message, but also in terms of the mediating<br />

infrastructures, platforms as and participa<strong>to</strong>ry cultures which<br />

facilitate its circulation. In this sense, the significance of <strong>fake</strong><br />

<strong>news</strong> cannot be fully unders<strong>to</strong>od apart from its circulation online.<br />

It is the register of this circulation that also enables us <strong>to</strong><br />

trace how material that starts its life as niche satire can be<br />

repackaged as hyper-partisan clickbait <strong>to</strong> generate advertising<br />

money and then continue life as an illustration of dangerous<br />

political misinformation.<br />

As a consequence this <strong>field</strong> <strong>guide</strong> ecourages a shift from<br />

focusing on the formal content of <strong>fake</strong> <strong>news</strong> in isolation <strong>to</strong><br />

understanding the contexts in which it circulates online.<br />

This shift points <strong>to</strong> the limits of a “deficit model” approach<br />

– which might imply that <strong>fake</strong> <strong>news</strong> thrives only because<br />

of a deficit of factual information. In the <strong>guide</strong> we suggest<br />

new ways of mapping and responding <strong>to</strong> <strong>fake</strong> <strong>news</strong> beyond<br />

identifying and fact-checking suspect claims – including<br />

“thicker” accounts of circulation as a way <strong>to</strong> develop a richer<br />

understanding of how <strong>fake</strong> <strong>news</strong> moves and mobilises<br />

people, more nuanced accounts of <strong>fake</strong>ness and responses<br />

which are better attuned <strong>to</strong> the phenomenon.<br />

[5] For a recent overview see<br />

Shannon Mattern’s “Cloud<br />

and Field”, Places Journal:<br />

https://placesjournal.org/<br />

article/cloud-and-<strong>field</strong>/<br />

While online and platform metrics often serve <strong>to</strong> take<br />

measure of engagement by means of what Theodore Porter<br />

calls “thin descriptions” – i.e. aggregated quantities such<br />

as <strong>to</strong>tal likes, shares, posts – we suggest different ways of<br />

exploring how different publics engage with and ascribe<br />

meaning <strong>to</strong> <strong>fake</strong> <strong>news</strong> and how this moves and mobilises<br />

different ac<strong>to</strong>rs in the process. In doing so while we start<br />

our inquiry with <strong>fake</strong> <strong>news</strong>, we end up surfacing a wide<br />

range of grassroots political, media and participa<strong>to</strong>ry<br />

cultures online and the social and political issues around<br />

which they assemble. Some of these may challenge and<br />

prompt a rethinking of our ideas of the forms and formats of<br />

grassroots political action online.<br />

We have adopted the metaphor of the “<strong>field</strong> <strong>guide</strong>” in the<br />

tradition of a number of recent <strong>guide</strong>s which transpose the<br />

language and imagery of mapping places, flora and fauna<br />

on<strong>to</strong> the cloud, digital infrastructures and life online. [5]<br />

However this metaphor stands in need of some qualification.<br />

Many classical natural his<strong>to</strong>rical “<strong>field</strong> <strong>guide</strong>s” aspire <strong>to</strong><br />

provide systematic taxonomies of natural phenomena by<br />

taking them out of their contexts in order <strong>to</strong> abstract and<br />

compare their features. By contrast with our <strong>guide</strong> we aim<br />

<strong>to</strong> do precisely the opposite – not <strong>to</strong> decontextualise, but <strong>to</strong><br />

recontextualise <strong>fake</strong> <strong>news</strong> phenomena by suggesting ways<br />

<strong>to</strong> follow them “in the wild”: as they travel across the web,<br />

search engines, digital platforms, fact-checking initiatives<br />

and <strong>news</strong> websites.<br />

We do not set out <strong>to</strong> provide a definitive single set of<br />

watercolour portraits, ana<strong>to</strong>mical illustrations, car<strong>to</strong>graphic<br />

charts, satellite imagery or infrastructural diagrams of the<br />

phenomenon in question – or even lists of characteristic<br />

features which may be used for the purposes of identification.<br />

Instead we illustrate a range of methods and procedures<br />

which readers may use in order <strong>to</strong> explore <strong>fake</strong> <strong>news</strong><br />

phenomena online for themselves. As part of this process we<br />

wish <strong>to</strong> extend the reper<strong>to</strong>ire of mapping practices which are<br />

publicly available <strong>to</strong> make sense of <strong>fake</strong> <strong>news</strong> online and in<br />

8 A FIELD GUIDE TO FAKE NEWS INTRODUCTION<br />

9


[6] See Philip E. Agre, “Toward<br />

a Critical Technical<br />

Practice: Lessons Learned<br />

in Trying <strong>to</strong> Reform AI”, in<br />

Geof Bowker, Les Gasser,<br />

Leigh Star, and Bill Turner<br />

(eds), Bridging the Great<br />

Divide: Social Science,<br />

Technical Systems, and Cooperative<br />

Work, Mahwah,<br />

NJ: Erlbaum, 1997. Available<br />

at: http://polaris.gseis.<br />

ucla.edu/pagre/critical.<br />

html<br />

[7] Shannon Mattern, “Cloud<br />

and Field”, Places Journal:<br />

https://placesjournal.org/<br />

article/cloud-and-<strong>field</strong>/<br />

[8] See, for example, the work<br />

of Pamela Smith and her<br />

colleagues on the “Making<br />

and Knowing” project<br />

at Columbia University:<br />

http://recipes.hypotheses.<br />

org/7430 and http://www.<br />

makingandknowing.org/<br />

this sense the graphics that we provide can be unders<strong>to</strong>od as<br />

temporary placeholders <strong>to</strong> encourage further exploration.<br />

We also draw attention <strong>to</strong> different ways of examining how<br />

things are categorised and labelled as <strong>fake</strong> <strong>news</strong> and the politics<br />

of these practices of classification. In this sense we hope <strong>to</strong><br />

cultivate what has been called “critical technical practice” [6]<br />

– which in this case would include reflection on the use of<br />

digital methods and digital data and how these not only serve<br />

<strong>to</strong> designate phenomena which can be straightforwardly and<br />

independently picked out, but how these very methods may<br />

also be involved in the process of articulating what <strong>fake</strong> <strong>news</strong><br />

is. As Shannon Mattern puts it, in undertaking <strong>to</strong> investigate<br />

<strong>fake</strong> <strong>news</strong> online we should be aware of “the shadows cast<br />

by our presence as explorers in the <strong>field</strong>”. [7] And rather than<br />

producing maps for the sake of producing maps, we should<br />

consider what maps do, who and what they are for and<br />

the effects that they produce as social, cultural and political<br />

devices.<br />

Insofar as we focus on providing procedures for inquiry<br />

rather than pictures of the phenomena, this <strong>guide</strong> may<br />

also be considered a kind of “recipe book”. Recent research<br />

suggests that there is an interesting relation between the<br />

documentation of recipes and the emergence of procedural<br />

knowledge in the early modern period – such that practices<br />

of writing down processes for cooking and craft are<br />

entangled with the his<strong>to</strong>ry of the emergence of scientific<br />

method. [8] Over the past few decades the metaphors of the<br />

“recipe” and the “cookbook” have also become popular in<br />

relation <strong>to</strong> software programming. In our <strong>guide</strong>, we illustrate<br />

different approaches <strong>to</strong> mapping and investigating <strong>fake</strong> <strong>news</strong><br />

online through a series of methodological “recipes”. As with<br />

many cookery books, our aim is not just <strong>to</strong> support readers in<br />

following the specific recipes that we present, but rather <strong>to</strong><br />

use these recipes <strong>to</strong> illustrate a certain approach <strong>to</strong> cooking –<br />

with the hope that readers are inspired <strong>to</strong> adapt, modify and<br />

venture beyond them. We also include a number of “serving<br />

suggestions” about how they may be put <strong>to</strong> work<br />

[9] See, Noortje Marres, “The<br />

Redistribution of Methods:<br />

On Intervention in Digital<br />

Social Research, Broadly<br />

Conceived”, The Sociological<br />

Review, June 2012;<br />

and Mike Ananny and Kate<br />

Crawford, “Seeing Without<br />

Knowing: Limitations of the<br />

Transparency Ideal and its<br />

Application <strong>to</strong> Algorithmic<br />

Accountability”, New Media<br />

and Society, December<br />

2016.<br />

[10] M. S. Abildgaard, A.<br />

Birkbak, L. Bounegru, J.<br />

Gray, M. Jacomy, A. K.<br />

Madsen, A. K. Munk (forthcoming)<br />

“Fake News: Seven<br />

Lines of Inquiry”.<br />

We hope that the recipes in this <strong>guide</strong> will enrich<br />

investigations of <strong>fake</strong> <strong>news</strong> in a way which has affinities with<br />

a common narrative approach in mystery fiction – namely the<br />

scenario that in pursuit of solving an apparently simple crime,<br />

the plot thickens, the cast grows, the questions multiply and<br />

there are unexpected twists or changes of perspective. By<br />

following the production, circulation and responses <strong>to</strong> <strong>fake</strong><br />

<strong>news</strong> online – we may end up being drawn in<strong>to</strong> things that<br />

we do not set out <strong>to</strong> investigate: whether the media strategies<br />

of <strong>fake</strong> <strong>news</strong> publishers, propagandists, trolls or bots; the<br />

commercial and technical architectures of online content;<br />

the politics and dynamics of viral content; and how social<br />

life adapts, evolves and innovates in response <strong>to</strong> some of the<br />

world’s biggest online platforms and websites. In this sense,<br />

it will be clear that <strong>fake</strong> <strong>news</strong> involves more than a few rogue<br />

producers or state conspiracies – and raises important and<br />

difficult questions about the role of digital technologies in<br />

society and how we mutually shape and are shaped by them.<br />

In Edgar Allan Poe’s classic mystery s<strong>to</strong>ry “The Purloined<br />

Letter”, the prefect of police – “G” – and his colleagues<br />

search for a letter said <strong>to</strong> contain scandalous information<br />

behind wallpaper, under carpets, in the legs of furniture<br />

and in cushions, only <strong>to</strong> eventually find the letter “hiding<br />

in plain sight”. In a similar vein, we may consider the<br />

algorithmically mediated circulation of <strong>fake</strong> <strong>news</strong> on digital<br />

platforms in terms of what Noortje Marres and Carolin<br />

Gerlitz characterise as “distributed accomplishment” or what<br />

Mike Ananny and Kate Crawford describe as “relational<br />

achievement”. [9] This entails a shift from “seeing in” systems<br />

as a kind of looking “under the hood”, <strong>to</strong> “seeing across”<br />

a diverse range elements which are implicated in the<br />

patterning of collective life online.<br />

Many of the researchers who have contributed <strong>to</strong> the <strong>guide</strong><br />

share a background in a <strong>field</strong> called Science and Technology<br />

Studies (STS). Some of the lines of inquiry pursued in the<br />

<strong>guide</strong> are informed by a forthcoming paper exploring what<br />

STS can bring <strong>to</strong> the study of <strong>fake</strong> <strong>news</strong>. [10] The recipes are<br />

10 A FIELD GUIDE TO FAKE NEWS INTRODUCTION<br />

11


[11] See, for example, Richard<br />

Rogers, Digital Methods,<br />

Cambridge, MA: MIT Press,<br />

2013.<br />

[12] See, for example, Noortje<br />

Marres, Material Participation:<br />

Technology, the<br />

Environment and Everyday<br />

Publics, London: Palgrave<br />

Macmillan, 2012.<br />

[13] See Tommaso Venturini,<br />

Anders Munk and Axel<br />

Meunier, “Data-Sprint: A<br />

Public Approach <strong>to</strong> Digital<br />

Research” in C. Lury, P.<br />

Clough, M. Michael, R.<br />

Fensham, S. Lammes, A.<br />

Last, & E. Uprichard (Eds.)<br />

Interdisciplinary Research<br />

Methods, London: Routledge,<br />

2017.<br />

[14] Jonathan Gray, Liliana<br />

Bounegru and Lucy Chambers<br />

(Eds.) The Data Journalism<br />

Handbook, Sebas<strong>to</strong>pol,<br />

CA: O’Reilly Media,<br />

2012, available at: http://<br />

datajournalismhandbook.<br />

org/<br />

[15] Upon being invited <strong>to</strong><br />

become Direc<strong>to</strong>r of the<br />

Argentine National Library<br />

at a moment which<br />

coincided with the deterioration<br />

of his eyesight,<br />

Borges famously wrote: “No<br />

one should read self-pity<br />

or reproach / in<strong>to</strong> this<br />

statement of the majesty<br />

/ of God; who with such<br />

splendid irony / Granted<br />

me books and night at one<br />

<strong>to</strong>uch”. See J. L. Borges,<br />

“Poem of the Gifts” in<br />

Selected Poems: Volume<br />

2, London: Penguin Books,<br />

2000, p. 95.<br />

also informed by a “digital methods” research approach that<br />

has developed through an engagement with this <strong>field</strong> and<br />

which many of us have contributed <strong>to</strong> through our teaching<br />

and research. [11] We also draw upon our <strong>field</strong>’s interest<br />

in public engagement and participation around digital<br />

technologies and data infrastructures. [12] As such our focus is<br />

less on advancing particular legal or technical fixes, than on<br />

facilitating processes of public engagement and democratic<br />

deliberation – including provoking curiosity about different<br />

ways of seeing the issue and imagination about the different<br />

ways in which we might respond.<br />

The material in this <strong>guide</strong> has been produced through<br />

a series of “data sprints” and research workshops in<br />

Amsterdam, Copenhagen and Milan, hosted by members<br />

of the Public Data Lab. The “data sprint” is a short form<br />

working format that has emerged at the intersection<br />

between Science and Technology Studies and New Media<br />

Studies, drawing on approaches associated with open-source<br />

software development, open data and civic hacking in order<br />

<strong>to</strong> convene a range of ac<strong>to</strong>rs <strong>to</strong> collaborate around the coproduction<br />

of data and research projects – including between<br />

<strong>field</strong>s of practice with different outlooks. [13]<br />

Two of us have a background in data journalism, having<br />

co-edited The Data Journalism Handbook and undertaken<br />

various initiatives in this <strong>field</strong>. [14] This <strong>guide</strong> builds on a<br />

long-standing interest in supporting productive encounters<br />

between data journalists and digital researchers. While <strong>fake</strong><br />

<strong>news</strong> seems like a remarkably ripe area for experimentation<br />

between these two <strong>field</strong>s, just as the writer Jorge Luis Borges<br />

lamented being granted “books and night at one <strong>to</strong>uch”, [15]<br />

it is not without a sense of irony that we note that as public<br />

attention around this issue grows, <strong>fake</strong> <strong>news</strong> websites are<br />

beginning <strong>to</strong> vanish – leading <strong>to</strong> proposals for a “<strong>fake</strong> <strong>news</strong><br />

archive” amongst our contributing researchers. Happily the<br />

approaches and analytical techniques in this <strong>guide</strong> may be<br />

used <strong>to</strong> inform collaborations between data journalists and<br />

digital researchers around the study of other contentious<br />

issues and controversies as they unfold on digital media, as<br />

well as of the mediating capacities of platforms, algorithms<br />

and infrastructures which shape life online.<br />

The data sprint format has also helped us <strong>to</strong> catalyse new<br />

experimentation and empirical work in a comparatively<br />

short period of time – a distinct advantage given the<br />

pace of developments around <strong>fake</strong> <strong>news</strong>. For this we<br />

are immensely grateful <strong>to</strong> researchers, graduates and<br />

students at Density Design (Politecnico di Milano, Italy),<br />

the Digital Methods Initiative (University of Amsterdam,<br />

Netherlands), the European Journalism Centre, the<br />

Labora<strong>to</strong>ire Interdisciplinaire Sciences Innovations Sociétés<br />

(Université Paris-Est, France), the médialab (Sciences Po,<br />

Paris, France), the Media of Cooperation research group<br />

(University of Siegen, Germany), the STS-Lab (University<br />

of Lausanne, Switzerland), the Techno-Anthropology Lab<br />

(Aalborg University Copenhagen, Denmark) and the Utrecht<br />

University – without whose energy, creativity and dedication<br />

this project would not have been possible.<br />

Jonathan Gray, Liliana Bounegru, Tommaso Venturini<br />

London, March 2017<br />

12 A FIELD GUIDE TO FAKE NEWS INTRODUCTION<br />

13


Chapter #1<br />

MAPPING<br />

FAKE NEWS HOTSPOTS<br />

ON FACEBOOK<br />

What publics does <strong>fake</strong> <strong>news</strong><br />

animate on Facebook?<br />

How may the trajec<strong>to</strong>ry of a<br />

<strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry be traced on<br />

Facebook?<br />

Do fact-checking initiatives<br />

reach the publics of <strong>fake</strong> <strong>news</strong><br />

on Facebook?<br />

14 A FIELD GUIDE TO FAKE NEWS


Introduction - This section provides a set<br />

of recipes for tracing the circulation of <strong>fake</strong><br />

<strong>news</strong> on Facebook. The focus is on circulation<br />

because false and misleading knowledge claims<br />

are not born “<strong>fake</strong> <strong>news</strong>.” To become <strong>fake</strong><br />

<strong>news</strong> they need <strong>to</strong> mobilise a large number of<br />

publics – including witnesses, allies, likes and<br />

shares, as well as opponents <strong>to</strong> contest, flag<br />

and debunk them. Facebook’s architecture<br />

poses challenges <strong>to</strong> the study of circulation<br />

of content due <strong>to</strong> the nature of its access and<br />

permissions system. Hence we focus on tracing<br />

the publics of <strong>fake</strong> <strong>news</strong> through its most<br />

publicly accessible entities: pages and groups,<br />

which may be considered <strong>to</strong> constitute already<br />

assembled publics.<br />

Around the 2016 US presidential elections<br />

commenta<strong>to</strong>rs have noted the emergence of<br />

a Facebook-native, hyper-partisan “political<br />

media machine” that was highly effective in<br />

gathering large numbers of followers and<br />

[1] See, John Herrman, “Inside<br />

Facebook’s (Totally<br />

Insane, Unintentionally<br />

Gigantic, Hyperpartisan)<br />

Political-Media Machine”,<br />

Agust 2016, The New<br />

York Times: https://www.<br />

nytimes.com/2016/08/28/<br />

magazine/inside-facebooks-<strong>to</strong>tally-insane-unintentionally-gigantic-hyperpartisan-political-media-machine.html<br />

generating engagement [1] . This<br />

<strong>fake</strong> <strong>news</strong> dissemination machine<br />

and responses <strong>to</strong> it, is what the<br />

recipes in this section enable <strong>to</strong><br />

explore. The first two recipes<br />

focus on mapping the publics<br />

that are energised by <strong>fake</strong> <strong>news</strong> on Facebook, as<br />

well as the trajec<strong>to</strong>ries through which <strong>fake</strong> <strong>news</strong><br />

s<strong>to</strong>ries travel on Facebook. The third recipe<br />

provides an approach <strong>to</strong> address the effectiveness<br />

of fact-checking initiatives in reaching the<br />

publics of <strong>fake</strong> <strong>news</strong> on Facebook. Through these<br />

recipes we aim <strong>to</strong> gesture <strong>to</strong>wards different ways<br />

of providing “thicker” accounts of circulation and<br />

engagement around <strong>fake</strong> <strong>news</strong> on social media<br />

beyond the “thin descriptions” of aggregated<br />

counts and metrics.<br />

16 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

17


CHAPTER 1 → RECIPE 1<br />

WHAT PUBLICS<br />

DOES FAKE NEWS<br />

ANIMATE ON FACEBOOK?<br />

START<br />

list of 22 URLs of political <strong>fake</strong><br />

<strong>news</strong> s<strong>to</strong>ries<br />

Source: BuzzFeed News<br />

input URLs in<br />

CrowdTangle<br />

output data<br />

> Fake <strong>news</strong> s<strong>to</strong>ry URLs<br />

> Facebook pages and groups<br />

that share the URLs<br />

> Number of interactions<br />

per each page or group<br />

> Date of sharing<br />

of the s<strong>to</strong>ry<br />

identify time intervals with<br />

highest frequency of publication<br />

of <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries<br />

BEFORE STARTING<br />

This recipe takes as a starting point a list of<br />

<strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries. There are different ways of<br />

obtaining these lists – including starting with<br />

existing lists as well as creating your own. To<br />

illustrate this recipe we use an already existing<br />

list of 22 <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries about various political<br />

issues pertaining <strong>to</strong> the 2016 presidential<br />

elections in the US that generated most<br />

engagement on Facebook. These were identified<br />

by BuzzFeed News.<br />

import data in<br />

identify key related<br />

events with<br />

Google News<br />

RAWGraphs<br />

visualise<br />

visualise<br />

a<br />

WHICH MEDIA AND POLITICAL<br />

EVENTS ARE SUCCESSFUL IN<br />

SETTING THE FAKE NEWS<br />

AGENDA?<br />

b<br />

WHICH FACEBOOK PAGES<br />

AND GROUPS SHARE FAKE<br />

NEWS?<br />

The recipe comprises of four steps. We start by<br />

identifying the themes that are exploited in our<br />

set of s<strong>to</strong>ries as well as the key political events<br />

which they edi<strong>to</strong>rialise (a). Next we identify<br />

the most prominent Facebook pages and groups<br />

that share these s<strong>to</strong>ries (b). We also explore<br />

whether certain publics have preferred s<strong>to</strong>ry<br />

themes (c) and profile the publics that are<br />

energised by <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries about the US<br />

elections (d)?<br />

import data in<br />

Gephi<br />

manually categorise<br />

facebook pages<br />

visualise<br />

visualise<br />

c<br />

DO FACEBOOK PUBLICS HAVE<br />

PREFERRED STORY THEMES?<br />

d<br />

WHAT KINDS OF PUBLICS ARE<br />

ENERGISED BY FAKE NEWS?<br />

18 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

19


START<br />

list of 22 URLs of political <strong>fake</strong><br />

<strong>news</strong> s<strong>to</strong>ries<br />

Source: BuzzFeed News<br />

input URLs in<br />

CrowdTangle<br />

identify time intervals with<br />

highest frequency of publication<br />

of <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries<br />

identify key related<br />

events with<br />

Google News<br />

visualise<br />

output data<br />

> Fake <strong>news</strong> s<strong>to</strong>ry URLs<br />

> Facebook pages and groups<br />

that share the URLs<br />

> Number of interactions<br />

per each page or group<br />

> Date of sharing<br />

of the s<strong>to</strong>ry<br />

a<br />

WHICH MEDIA AND POLITICAL<br />

EVENTS ARE SUCCESSFUL IN<br />

SETTING THE FAKE NEWS<br />

AGENDA?<br />

CHAPTER 1 → RECIPE 1<br />

a. EXAMINE THE THEMES EXPLOITED<br />

IN FAKE NEWS STORIES AND<br />

IDENTIFY THE EVENTS WHICH THEY<br />

EDITORIALISE<br />

This analysis may be done by qualitatively<br />

examining the content of each article and<br />

identifying key political or media events<br />

related <strong>to</strong> the issues exploited in the articles,<br />

which occurred around the publication date<br />

of each s<strong>to</strong>ry. This is done <strong>to</strong> enable a better<br />

understanding of the issues that animate the<br />

publics that circulate <strong>fake</strong> <strong>news</strong>.<br />

◊<br />

◊<br />

◊<br />

If the content of the <strong>fake</strong> <strong>news</strong> article is no<br />

longer available on its original URL you<br />

may use the Internet Archive’s → WayBack<br />

Machine <strong>to</strong> check whether an archived<br />

version of the URL is available.<br />

To identify key events occurring around<br />

the dates of publication of the s<strong>to</strong>ries which<br />

are related <strong>to</strong> the themes exploited in the<br />

s<strong>to</strong>ries you may use a <strong>news</strong> aggrega<strong>to</strong>r such<br />

as → Google News Search as well as <strong>news</strong><br />

article archives.<br />

An annotated timeline of s<strong>to</strong>ries and<br />

relevant events occurring around the same<br />

dates might provide a starting point for<br />

reflection about the relationship between<br />

political and media events and <strong>fake</strong> <strong>news</strong><br />

s<strong>to</strong>ries.<br />

20 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

21


WHICH MEDIA AND<br />

POLITICAL EVENTS ARE<br />

SUCCESSFUL IN SETTING<br />

THE FAKE NEWS AGENDA?<br />

Timeline of best performing <strong>fake</strong><br />

<strong>news</strong> s<strong>to</strong>ries about the US elections<br />

on Facebook in 2016 and events they<br />

edi<strong>to</strong>rialise. Successful <strong>fake</strong> <strong>news</strong><br />

s<strong>to</strong>ries appear <strong>to</strong> exploit populist<br />

themes such as anti-establishment<br />

sentiment, nationalist and antiimmigration<br />

sentiment as well as<br />

perceived or projected weaknesses of<br />

political candidates such as misogyny<br />

and corruption. A number of events at<br />

the end of July, mid-Oc<strong>to</strong>ber and early<br />

November are successful in setting the<br />

<strong>fake</strong> <strong>news</strong> “agenda.”<br />

8 March<br />

REACT365.COM<br />

6 July<br />

CHRISTIAN TIMES<br />

Obama passed law for grandparents<br />

<strong>to</strong> get all their grandchildren every<br />

weekend<br />

N/A<br />

11 March<br />

BREAKING Romanian Hacker With<br />

Access To Clin<strong>to</strong>n Emails Found<br />

Dead In Jail Cell<br />

ANTI-CLINTON<br />

23 July<br />

HEAVIERMETAL.NET<br />

7 September<br />

BURRARDSTREETJOURNAL.COM<br />

President Obama Confirms He Will<br />

Refuse To Leave Office If Trump Is<br />

Elected<br />

PRESIDENTIAL RACE<br />

18 September<br />

RAGE AGAINST THE MACHINE To<br />

Reunite And Release Anti Donald<br />

Trump Album<br />

ANTI-TRUMP<br />

3 Oc<strong>to</strong>ber<br />

ABCNEWS.COM.CO<br />

Ac<strong>to</strong>r Bill Murray Announces 2016<br />

Presidential Run<br />

PACIFISM<br />

BIZSTANDARDNEWS.COM<br />

ANTI-DEMOCRAT<br />

POLITICOPS.COM<br />

WORLDNEWSDAILYREPORT.COM<br />

Mike Pence: “Sarah Palin Is My Role<br />

Model For Beautiful, Smart American<br />

Women” - Newslo<br />

GENDER DISCRIMINATION<br />

26 September<br />

Graham Says Christians Must<br />

Support Trump or Face Death Camps<br />

KYPO6.COM<br />

Donald Trump accepts nomination<br />

23 Jul<br />

Trump attacks "corrupt establishment",<br />

states Clin<strong>to</strong>n should be "locked up" 13 Oct<br />

Hillary Clin<strong>to</strong>n accepts nomination 28 Jul Clin<strong>to</strong>n Revelations (So Far)<br />

–Breitbart<br />

and media is in "war against him"<br />

Wikileaks reveals Democratic Party<br />

has a bias against Bernie Sanders<br />

25 Jul<br />

–Vox<br />

The Most Explosive WikiLeaks<br />

25 July<br />

11 Oc<strong>to</strong>ber<br />

ISIS Leader Calls for American<br />

Muslim Voters <strong>to</strong> Support Hillary<br />

Clin<strong>to</strong>n<br />

ANTI-MUSLIM SENTIMENT<br />

ENDING THE FED<br />

Pope Francis Shocks World,<br />

Endorses Donald Trump for<br />

President, Releases Statement<br />

CORRUPTION OF THE POLITICAL ESTABLISHMENT<br />

Pope Francis Shocks World,<br />

Endorses Hillary Clin<strong>to</strong>n for<br />

President, Releases Statement<br />

ANTI-TRUMP<br />

3 November<br />

EMPIREHERALD.COM<br />

African Billionaire Will Give $1 Million To<br />

Anyone Who Wants To Leave America if<br />

Donald Trump is Elected President<br />

DISCRIMINATION AGAINST MINORITIES<br />

14 Oc<strong>to</strong>ber<br />

WORLDNEWSDAILYREPORT.COM<br />

RUPAUL CLAIMS TRUMP TOUCHED<br />

HIM INAPPROPRIATELY IN THE<br />

1990S<br />

MISOGYNY<br />

29 July<br />

BURRARDSTREETJOURNAL.COM<br />

Trump Claims America Should Never<br />

Have Given Canada Its Independence<br />

NATIONALIST SENTIMENT<br />

Trump Says Sexual Assault Allegations<br />

are "Made-Up S<strong>to</strong>ries"<br />

–IB Times<br />

Michelle Obama denounces<br />

Trump's comment on women<br />

–NY Times<br />

THELASTLINEOFDEFENSE.ORG<br />

14 Oct<br />

5 November<br />

Van Full Of Illegals Shows Up To Vote<br />

Clin<strong>to</strong>n At SIX Polling Places, Still<br />

Think Voter Fraud Is A Myth?<br />

VOTER FRAUD, ANTI-IMMIGRATION SENTIMENT<br />

15 Oc<strong>to</strong>ber<br />

USANEWSFLASH.COM<br />

THELASTLINEOFDEFENSE.ORG<br />

FBI Agent Suspected in Hillary Email<br />

Leaks Found Dead in Apparent<br />

Murder-Suicide<br />

CORRUPTION OF THE POLITICAL ESTABLISHMENT<br />

5 November<br />

WHOA! Hillary Caught On Hot Mic<br />

Trashing Beyonce’ With RACIAL SLURS!<br />

ANTI-CLINTON<br />

17 Oc<strong>to</strong>ber<br />

CONSERVATIVESTATE.COM<br />

Hillary Clin<strong>to</strong>n In 2013: “I Would Like To See<br />

People Like Donald Trump Run For Office;<br />

They’re Honest And Can’t Be Bought”<br />

PRO-ENTREPRENEURIALIST SENTIMENT<br />

5 November<br />

DENVER GUARDIAN<br />

FBI Agent Suspected in Hillary Email<br />

Leaks Found Dead in Apparent<br />

Murder-Suicide<br />

CORRUPTION OF THE POLITICAL ESTABLISHMENT<br />

Beyonce, Jay Z & Clin<strong>to</strong>n share a stage<br />

–NY Times<br />

No charge <strong>to</strong> Hillary Cin<strong>to</strong>n and no further investigation over email scandal<br />

–FBI Direc<strong>to</strong>r Comey<br />

Arizona Anti-Immigrant Sheriff <strong>to</strong> Deploy Deputies at Polling Places<br />

–The Guardian<br />

Donald Trump continues <strong>to</strong> accuse Democrats of voter fraud.<br />

On Saturday 5th November, Trump claimed the late opening of<br />

a voting site in a Latino neighorhood of Las Vegas due<br />

<strong>to</strong> long lines pointed <strong>to</strong> a "rigged system."<br />

–Democracy Now<br />

11 December<br />

ABCNEWS.COM.CO<br />

Obama Signs Executive Order<br />

Banning The Pledge Of Allegiance In<br />

Schools Nationwide<br />

NATIONALIST SENTIMENT<br />

11 November<br />

TMZHIPHOP.COM<br />

Trump Offering Free One-Way<br />

Tickets <strong>to</strong> Africa & Mexico for Those<br />

Who Wanna Leave America<br />

ANTI-IMMIGRANT SENTIMENT<br />

5 Nov<br />

6 Nov<br />

7 Nov<br />

ABCNEWS.COM.CO<br />

Obama Signs Executive Order Declaring<br />

Investigation In<strong>to</strong> Election Results;<br />

Revote Planned For Dec. 19th<br />

CORRUPTION OF POLITICAL ESTABLISHMENT<br />

22 November<br />

ABCNEWS.COM.CO<br />

Donald Trump Protester Speaks Out: I<br />

Was Paid $3,500 To Protest Trumps Rally<br />

ANTI-CLINTON<br />

January February March April May<br />

June July August September Oc<strong>to</strong>ber November December<br />

12 December<br />

Fake <strong>news</strong> s<strong>to</strong>ries Events<br />

Ninth Woman Says Trump<br />

Inappropriately Groped or Kissed Her<br />

–The Guardian<br />

15 Oct<br />

Arizona Anti-Immigrant Sheriff<br />

<strong>to</strong> Deploy Deputies at Polling Places<br />

–The Guardian<br />

11 Nov<br />

22 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

23


CHAPTER 1 → RECIPE 1<br />

START<br />

list of 22 URLs of political <strong>fake</strong><br />

<strong>news</strong> s<strong>to</strong>ries<br />

Source: BuzzFeed News<br />

input URLs in<br />

CrowdTangle<br />

output data<br />

> Fake <strong>news</strong> s<strong>to</strong>ry URLs<br />

> Facebook pages and groups<br />

that share the URLs<br />

> Number of interactions<br />

per each page or group<br />

> Date of sharing<br />

of the s<strong>to</strong>ry<br />

b. IDENTIFY THE FACEBOOK PAGES AND<br />

GROUPS THAT SHARE THESE STORIES<br />

This may be done with a social media<br />

moni<strong>to</strong>ring <strong>to</strong>ol such as the browser extension<br />

of → CrowdTangle. The number of followers<br />

per page or group as well as the number of<br />

interactions per posts should be recorded in a<br />

spreadsheet alongside the names of pages and<br />

groups that share <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries.<br />

◊<br />

Please note that a <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry may be<br />

reposted on a number of different websites.<br />

For this reason a methodological decision<br />

needs <strong>to</strong> be taken from the outset as <strong>to</strong><br />

whether only the pages and groups that<br />

share the original URL of the s<strong>to</strong>ry will be<br />

recorded or whether all pages and groups<br />

that share all versions of the <strong>fake</strong> <strong>news</strong><br />

s<strong>to</strong>ry will be collected.<br />

import data in<br />

RAWGraphs<br />

visualise<br />

b<br />

WHICH FACEBOOK PAGES<br />

AND GROUPS SHARE FAKE<br />

NEWS?<br />

◊<br />

◊<br />

◊<br />

You may want <strong>to</strong> take note of the pages or<br />

groups which shared the highest number<br />

of <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries as well as the <strong>to</strong>tal<br />

number of interactions generated by each<br />

group or page.<br />

If you use → CrowdTangle please note<br />

that for Facebook the <strong>to</strong>ol returns the <strong>to</strong>p<br />

500 most important posts from verified<br />

pages as well as from pages with more than<br />

750.000 followers. [1]<br />

You may use a “circle packing” visualisation<br />

<strong>to</strong> represent the pages and groups that<br />

share <strong>fake</strong> <strong>news</strong> items as well as the<br />

number of s<strong>to</strong>ries which they share and<br />

the number of interactions which they<br />

generate. You may use → RAWGraphs for<br />

this operation.<br />

[1]<br />

See, CrowdTangle’s<br />

“Frequently<br />

Asked Questions”,<br />

available<br />

at: https://apps.<br />

crowdtangle.<br />

com/chrome-extension/faq<br />

24 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

25


WHICH FACEBOOK PAGES<br />

AND GROUPS PROMOTED<br />

THE HIGHEST NUMBER OF<br />

FAKE NEWS? WHICH ONES<br />

CREATED THE HIGHEST<br />

ENGAGEMENT?<br />

Facebook pages and groups that<br />

share <strong>fake</strong> <strong>news</strong> items, sized<br />

according <strong>to</strong> the number of items<br />

they share and coloured according<br />

<strong>to</strong> their number of followers. Each<br />

page can share the same item more<br />

than once. The pages and groups that<br />

share the highest number of s<strong>to</strong>ries<br />

are primarily pro-Trump supporters<br />

and anti-Hillary groups. The page<br />

that generates the highest number of<br />

interactions with <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries is<br />

the fan page dedicated <strong>to</strong> republican<br />

TV commenta<strong>to</strong>r, Jeanine Pirro.<br />

AMOUNT OF INTERACTIONS<br />

1 30<br />

0<br />

FAKE NEWS POSTED<br />

1<br />

30<br />

73,732<br />

The Australian Tea Party<br />

The Burrard Street Journal<br />

The Church<br />

of PUNK<br />

The Empire Herald<br />

F@CK YOU<br />

HILLARY CLINTONDemocratic<br />

WE HATE YOU! News<br />

and Events<br />

Deplorables for TRUMP<br />

FU Trump<br />

One Nation Under God<br />

Females<br />

for TRUMP Citizens<br />

For Trump<br />

Judge Jeanine<br />

Pirro has Fans<br />

Don P<br />

The Original Wake<br />

Up America<br />

TRUMP OF CHAMPIONS<br />

Newslo<br />

Indies 4 Trump<br />

PATRIOT<br />

AMERICANS<br />

FOR DONALD TRUMP<br />

The Resistance:<br />

The Last Line<br />

Of Defense<br />

Patriots for a<br />

Free Republic -<br />

Trump Squad<br />

TRUMP FORCE ONE<br />

Chonda Pierce<br />

Freedom Don't<br />

Come Free<br />

& Devildoc<br />

The Trump<br />

American Party<br />

Women SCOUTS for TRUMP (c)<br />

THE TRUMPS STORM GROUP<br />

TRUMP - SPEAK OUT<br />

AGAINST ISLAMIZATION<br />

OF AMERICA<br />

All Things<br />

#trumpvic<strong>to</strong>ry<br />

#neverhillary<br />

Friends Who<br />

Like Donald J. Trump<br />

Politicked<br />

We Endorse<br />

Donald J. Trump<br />

Washing<strong>to</strong>n State for<br />

THE TRUMPERS!!! Donald Trump 2016<br />

NO Hillary for 2016<br />

DONALD TRUMP<br />

COMMANDER IN CHIEF !!!!<br />

Anonymous The<br />

World is Ours<br />

1 Million Strong For<br />

Hillary Clin<strong>to</strong>n in 2016<br />

Hillary Clin<strong>to</strong>n<br />

vote blue<br />

Being Liberal<br />

Means Being<br />

a Hypocrite 2 Colorado<br />

for Donald<br />

Trump 2016<br />

NEVER<br />

HILLARY<br />

CLINTON<br />

Hillary Clin<strong>to</strong>n in 2020<br />

D.L. Hughley<br />

Conservative<br />

Veterans<br />

for America<br />

Donald Trump<br />

Donald J. Trump Donald Trump<br />

Commander<br />

for PRESIDENT America's<br />

in Chief 2020<br />

Facebook Group President (c)<br />

Tomorrow's<br />

News Today.<br />

NEVER HILLARY<br />

Donald Trump for President!<br />

Donald Trump for<br />

President Connecticut<br />

Donald Trump<br />

Criminalize VS. Hillary<br />

Conservatism Clin<strong>to</strong>n<br />

Donald Trump<br />

The Political Movement<br />

Sicilians For<br />

Donald Trump<br />

Donald<br />

Trump<br />

President<br />

Donald Trump For President<br />

Hillary Clin<strong>to</strong>n Revolution<br />

HILLARY IN<br />

2016 & 2020<br />

REPUBLICANS - CONSERVATIVES -<br />

TEA PARTY PATRIOTS-<br />

RIGHT WING AMERICANS<br />

Trump America<br />

WOMEN4TRUMP<br />

Trump<br />

Train<br />

Make America<br />

Great Again<br />

MY VOTE'S FOR<br />

TRUMP<br />

The Deplorables<br />

Trump USA's CEO<br />

Veterans for Donald Trump<br />

Rock Feed<br />

VETERANS FOR PRESIDENT<br />

DONALD J TRUMP 2016<br />

Resist Donald J. Trumpf<br />

Republican<br />

Patriotic<br />

Party<br />

Reclaim<br />

America<br />

UNITE TO FIGHT<br />

BRING AMERICA<br />

BACK !!! TRUMP 2016<br />

TrumpNation<br />

VOTE TRUMP ONLY -<br />

THE AMERICAN PARTY RISING<br />

26 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

27


START<br />

list of 22 URLs of political <strong>fake</strong><br />

<strong>news</strong> s<strong>to</strong>ries<br />

Source: BuzzFeed News<br />

input URLs in<br />

CrowdTangle<br />

output data<br />

> Fake <strong>news</strong> s<strong>to</strong>ry URLs<br />

> Facebook pages and groups<br />

that share the URLs<br />

> Number of interactions<br />

per each page or group<br />

> Date of sharing<br />

of the s<strong>to</strong>ry<br />

CHAPTER 1 → RECIPE 1<br />

c. IDENTIFY WHETHER FACEBOOK<br />

PUBLICS HAVE PREFERRED STORY<br />

THEMES<br />

To explore whether particular s<strong>to</strong>ry themes<br />

assemble publics and <strong>to</strong> qualitatively profile<br />

those publics based on the s<strong>to</strong>ries that animate<br />

them you may conduct a network analysis of<br />

Facebook pages and groups connected by the<br />

s<strong>to</strong>ries which they share.<br />

◊<br />

Starting from the dataset extracted with<br />

→ CrowdTangle’s browser extension, you<br />

may create a network file where each time<br />

a Facebook group or page posts a <strong>fake</strong> <strong>news</strong><br />

s<strong>to</strong>ry a link is established between that page<br />

or group and that s<strong>to</strong>ry.<br />

import data in<br />

Gephi<br />

visualise<br />

c<br />

DO FACEBOOK PUBLICS HAVE<br />

PREFERRED STORY THEMES?<br />

◊<br />

◊<br />

You may use → Table2Net <strong>to</strong> convert your<br />

CSV (comma-separated values) file in<strong>to</strong> a<br />

network file and → Gephi <strong>to</strong> explore the<br />

network. A force-directed layout algorithm<br />

such as ForceAtlas2 [2] can help you<br />

visualise the outcomes.<br />

Identify which s<strong>to</strong>ries are most successful<br />

in energising publics as well as whether<br />

publics have preferred s<strong>to</strong>ry themes.<br />

[2]<br />

See, Mathieu<br />

Jacomy, Tommaso<br />

Venturini, Sebastien<br />

Heymann and<br />

Mathieu Bastian,<br />

“ForceAtlas2, a<br />

Continuous Graph<br />

Layout Algorithm<br />

for Handy Network<br />

Visualization Designed<br />

for the Gephi<br />

Software”, June<br />

2014, PLoS ONE:<br />

http://journals.<br />

plos.org/plosone/<br />

article?id=10.1371/<br />

journal.<br />

pone.0098679<br />

28 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

29


DO FACEBOOK PUBLICS<br />

HAVE PREFERRED<br />

STORY THEMES?<br />

K<br />

Network of Facebook pages and groups<br />

connected by the <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries<br />

which they share. Notable is the core of<br />

the network which consists of a series<br />

of pages and groups associated with<br />

Trump supporters which are animated<br />

by anti-Hillary s<strong>to</strong>ries.<br />

FAKE<br />

NEWS URL<br />

0<br />

AMOUNT OF<br />

INTERACTIONS<br />

N° OF PAGES SHARING<br />

A NEWS ITEM<br />

1<br />

FACEBOOK<br />

PAGE OR<br />

GROUP<br />

0<br />

125<br />

98,032<br />

AMOUNT OF<br />

INTERACTIONS<br />

65,649<br />

H<br />

J<br />

I<br />

T<br />

Reclaim America<br />

America's Veterans<br />

are Loved<br />

M<br />

One Nation<br />

Under God<br />

D<br />

Hillary Clin<strong>to</strong>n<br />

Revolution<br />

Q<br />

C<br />

NO Hillary<br />

for 2016<br />

P<br />

The Trump American Party<br />

Donald Trump<br />

Commander<br />

in Chief 2020<br />

VOTE TRUMP ONLY<br />

THEAMERICAN<br />

PARTY RISING<br />

The Burrard<br />

Street Journal<br />

Colorado for Donald<br />

Trump 2016<br />

G<br />

F@CK YOU HILLARY<br />

CLINTON WE HATE YOU!<br />

A<br />

D.L. Hughley<br />

TrumpNation<br />

N<br />

The Resistance: The Last<br />

Line Of Defense<br />

F<br />

R<br />

S<strong>to</strong>p the Mosque<br />

in Bendigo<br />

TRUMP - SPEAK OUT AGAINST<br />

ISLAMIZATION OF AMERICA<br />

E<br />

L<br />

TRUMP<br />

FORCE ONE<br />

Make America<br />

Great Again<br />

Donald Trump<br />

For President<br />

B<br />

O<br />

Donald Trump<br />

for President<br />

S<br />

U<br />

Fake News Headlines<br />

A FBI Agent Suspected in Hillary Email<br />

Leaks Found Dead in Apparent<br />

Murder-Suicide<br />

B Hillary Clin<strong>to</strong>n In 2013: “I Would Like<br />

To See People Like Donald Trump Run<br />

For Office; They’re Honest And Can’t<br />

Be Bought”<br />

C ISIS Leader Calls for American Muslim<br />

Voters <strong>to</strong> Support Hillary Clin<strong>to</strong>n<br />

D Donald Trump Protester Speaks Out:<br />

"I Was Paid $3,500 To Protest<br />

Trump's Rally"<br />

E Obama Signs Executive Order<br />

Declaring Investigation In<strong>to</strong> Election<br />

Results; Revote Planned For Dec. 19th<br />

F WHOA! Hillary Caught On Hot Mic<br />

Trashing Beyonce’ With RACIAL<br />

SLURS!<br />

G Van Full Of Illegals Shows Up To Vote<br />

Clin<strong>to</strong>n At SIX Polling Places, Still<br />

Think Voter Fraud Is A Myth?<br />

H RAGE AGAINST THE MACHINE To<br />

Reunite And Release Anti Donald<br />

Trump Album<br />

I Obama Signs Executive Order Banning<br />

The Pledge Of Allegiance In Schools<br />

Nationwide<br />

J BREAKING Romanian Hacker With<br />

Access To Clin<strong>to</strong>n Emails Found Dead<br />

In Jail Cell<br />

K Ac<strong>to</strong>r Bill Murray Announces 2016<br />

Presidential Run<br />

L Pope Francis Shocks World, Endorses<br />

Hillary Clin<strong>to</strong>n for President, Releases<br />

Statement<br />

M Pope Francis Shocks World, Endorses<br />

Donald Trump for President, Releases<br />

Statement<br />

N Trump Claims America Should Never<br />

Have Given Canada Its Independence<br />

O Mike Pence: “Sarah Palin Is My Role<br />

Model For Beautiful, Smart American<br />

Women” - Newslo<br />

P RUPAUL CLAIMS TRUMP TOUCHED<br />

HIM INAPPROPRIATELY IN THE<br />

1990S<br />

Q President Obama Confirms He Will<br />

Refuse To Leave Office If Trump Is<br />

Elected<br />

R Graham Says Christians Must Support<br />

Trump or Face Death Camps<br />

S African Billionaire Will Give $1 Million<br />

To Anyone Who Wants To Leave<br />

America if Donald Trump is Elected<br />

President<br />

T Trump Offering Free One-Way Tickets<br />

<strong>to</strong> Africa &amp; Mexico for Those<br />

Who Wanna Leave America<br />

U Obama passed law for grandparents <strong>to</strong><br />

get all their grandchildren every<br />

weekend<br />

30 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

31


START<br />

list of 22 URLs of political <strong>fake</strong><br />

<strong>news</strong> s<strong>to</strong>ries<br />

Source: BuzzFeed News<br />

input URLs in<br />

CrowdTangle<br />

output data<br />

> Fake <strong>news</strong> s<strong>to</strong>ry URLs<br />

> Facebook pages and groups<br />

that share the URLs<br />

> Number of interactions<br />

per each page or group<br />

> Date of sharing<br />

of the s<strong>to</strong>ry<br />

CHAPTER 1 → RECIPE 1<br />

d. PROFILE THE PUBLICS ANIMATED BY<br />

FAKE NEWS<br />

This may be done by conducting a qualitative<br />

analysis of all Facebook pages that share <strong>fake</strong><br />

<strong>news</strong> items based on self-descriptions available<br />

on their “About” pages.<br />

◊<br />

◊<br />

◊<br />

You may take an “emergent coding”<br />

approach <strong>to</strong> identify the themes that<br />

emerge from the description of pages.<br />

You may take note of a more generic<br />

category (e.g. “grass-roots activism”) as<br />

well as a more specific one (e.g. “antiestablishment”).<br />

Sum up the amount of followers across all<br />

pages belonging <strong>to</strong> the same category.<br />

A treemap visualisation may be used <strong>to</strong><br />

represent the weight and hierarchy of each<br />

category. You may use → RAWGraphs for<br />

this operation.<br />

manually categorise<br />

facebook pages<br />

visualise<br />

d<br />

WHAT KINDS OF PUBLICS ARE<br />

ENERGISED BY FAKE NEWS?<br />

32 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

33


WHAT KINDS OF<br />

PUBLICS ARE ENERGISED<br />

BY FAKE NEWS?<br />

Types of Facebook publics animated<br />

by <strong>fake</strong> <strong>news</strong>, according <strong>to</strong> a manual<br />

classification of pages that share <strong>fake</strong><br />

<strong>news</strong> items. Notable are grassroots<br />

activists for a variety of issues,<br />

political candidate loyalists as well as<br />

entertainers.<br />

Truckers<br />

Italian<br />

Left<br />

Pro Latino<br />

Against Gun<br />

Violence<br />

WORKERS<br />

Small<br />

Bussiness<br />

OTHER<br />

Climate<br />

Skeptics<br />

SOCIAL CHANGE<br />

Human<br />

rights<br />

Steelworkers<br />

Page<br />

Dedicated<br />

<strong>to</strong> the<br />

Philipines<br />

Progessives<br />

News<br />

Critical<br />

News<br />

News<br />

Trending<br />

News<br />

Liberal<br />

<strong>news</strong><br />

Non-PC News<br />

NEWS<br />

News<br />

Aggrega<strong>to</strong>r<br />

Liberal<br />

News<br />

Convervative<br />

News<br />

News / Republican<br />

News News<br />

Aggrega<strong>to</strong>rs<br />

Alternative<br />

News<br />

GRASSROOTS ACTIVISTS/ CITIZEN VIGILANTES<br />

For Free<br />

Society<br />

For Liberty<br />

and Property<br />

Againt<br />

Gun<br />

Owners<br />

Against<br />

Mainstream<br />

Media<br />

Against Gun<br />

Control<br />

Advocates<br />

Against<br />

Corruption<br />

Againt<br />

Ted Cruz<br />

Pro<br />

Bernie<br />

White<br />

Lives<br />

Matter<br />

Anti-conservatism<br />

Anonymous<br />

For Against<br />

Disaster Islam<br />

Preparness<br />

For<br />

Information<br />

Freedom<br />

Against Trump<br />

Against<br />

Financial<br />

Terrorism<br />

US Patriots<br />

Pro Latino<br />

Against<br />

Republicans<br />

ENTERTAINERS<br />

PUBLIC FIGURE<br />

Artist<br />

Ac<strong>to</strong>r<br />

Journalist<br />

Porn<br />

Actress<br />

Republican<br />

Politician<br />

EARLY CAMPAIGNERS<br />

Pro-Trump<br />

Early 2020<br />

Bernie<br />

Campaigners<br />

Early 2020 Early 2020<br />

Trump Hillary<br />

Campaigners Campaigners<br />

LOYALISTS<br />

Hillary Loyalists<br />

Trump Loyalists<br />

Bernie Loyalists<br />

DISCUSSION SPACE<br />

Political<br />

Commentary<br />

Anti<br />

establishment<br />

Politcal<br />

Commentary<br />

Anti-Democrat<br />

Punk Tat<strong>to</strong>o<br />

BUSINESS<br />

Fashion<br />

COMMUNITY<br />

MUSIC/ENTERTAIMENT<br />

HIERARCY<br />

CATEGORY<br />

Sub-category<br />

Bussiness<br />

Dating app<br />

Motivational<br />

Speaker<br />

Spritual<br />

Community<br />

Progessive<br />

Community<br />

Survivalist<br />

Democrat Community<br />

Republican Community<br />

Pro Military<br />

FACEBOOK PAGES<br />

PER CATEGORY<br />

OCCUPY MOVEMENT<br />

Pro-Black<br />

Religious<br />

1<br />

2<br />

3<br />

WOMEN-FOCUSED<br />

CLICKBAIT<br />

34 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

35


3 November<br />

EMPIREHERALD.COM<br />

F<br />

IONS<br />

8,032<br />

F<br />

IONS<br />

H<br />

J<br />

African Billionaire Will Give $1 Million To<br />

Anyone Who Wants To Leave America if<br />

Donald Trump is Elected President<br />

Women SCOUTS for TRUMP (c)<br />

DISCRIMINATION AGAINST MINORITIES<br />

3 Oc<strong>to</strong>ber<br />

ABCNEWS.COM.CO<br />

5 November<br />

We Endorse<br />

THELASTLINEOFDEFENSE.ORG<br />

Donald J. Trump<br />

Ac<strong>to</strong>r Bill Murray Announces 2016<br />

Washing<strong>to</strong>n State for<br />

Presidential Run<br />

Van Full Of Illegals Shows Up To Vote<br />

THE TRUMPERS!!! Donald Trump 2016<br />

PACIFISM<br />

11 Oc<strong>to</strong>ber<br />

Clin<strong>to</strong>n At SIX Polling Places, Still<br />

Think Voter Fraud Is A Myth?<br />

11 December<br />

WORLDNEWSDAILYREPORT.COM<br />

THE TRUMPS STORM GROUP<br />

VOTER FRAUD, ANTI-IMMIGRATION SENTIMENT<br />

ABCNEWS.COM.CO<br />

ISIS Leader Calls for American<br />

Muslim Voters <strong>to</strong> Support Hillary<br />

Obama Signs Executive Order<br />

5 November<br />

Clin<strong>to</strong>n<br />

Banning The Pledge Of Allegiance In<br />

WOMEN4TRUMP<br />

ANTI-MUSLIM SENTIMENT<br />

THELASTLINEOFDEFENSE.ORG<br />

Schools Nationwide<br />

WHOA! Hillary Caught On Hot Mic<br />

NATIONALIST SENTIMENT<br />

NO Hillary for 2016<br />

TRUMP - SPEAK OUT<br />

14 Oc<strong>to</strong>ber<br />

Trashing Beyonce’ With RACIAL SLURS!<br />

7 September<br />

12 December<br />

AGAINST ISLAMIZATION<br />

WORLDNEWSDAILYREPORT.COM<br />

ANTI-CLINTON<br />

OF AMERICA<br />

BURRARDSTREETJOURNAL.COM<br />

5 November<br />

ABCNEWS.COM.CO<br />

RUPAUL CLAIMS TRUMP TOUCHED<br />

NEVER<br />

President Obama Confirms He Will<br />

HIM INAPPROPRIATELY IN THE<br />

DENVER GUARDIAN<br />

Obama Signs Executive Order Declaring<br />

Sicilians For<br />

HILLARY Donald Trump<br />

Refuse To Leave Office If Trump Is<br />

1990S<br />

FBI Agent Suspected in Hillary Email<br />

Investigation In<strong>to</strong> Election Results;<br />

CLINTON<br />

Elected<br />

MISOGYNY<br />

Leaks Found Dead in Apparent<br />

Revote Planned For TRUMP Dec. 19th FORCE ONE<br />

PRESIDENTIAL RACE<br />

CORRUPTION OF POLITICAL ESTABLISHMENT<br />

18 September<br />

15 Oc<strong>to</strong>ber<br />

Veterans for Donald Trump<br />

Murder-Suicide<br />

TRUMP OF CHAMPIONS<br />

Hillary Clin<strong>to</strong>n NEVER HILLARY<br />

POLITICOPS.COM<br />

USANEWSFLASH.COM<br />

CORRUPTION OF THE POLITICAL ESTABLISHMENT<br />

vote blue<br />

Mike Pence: “Sarah Palin Is My Role<br />

FBI Agent Suspected in Hillary Email<br />

11 November<br />

Model For Beautiful, Smart American<br />

Leaks Found Dead in Apparent<br />

Hillary Clin<strong>to</strong>n in 2020<br />

Women” - Newslo<br />

Murder-Suicide<br />

TMZHIPHOP.COM<br />

K<br />

Newslo<br />

6 July<br />

GENDER DISCRIMINATION<br />

CORRUPTION OF THE POLITICAL ESTABLISHMENT<br />

Trump Offering Free One-Way<br />

CHRISTIAN TIMES<br />

26 September<br />

Tickets <strong>to</strong> Africa & Mexico for Those<br />

ENDING THE FED<br />

17 Oc<strong>to</strong>ber<br />

BREAKING Romanian Hacker With<br />

Who Wanna Leave America<br />

DONALD TRUMP<br />

Hillary Clin<strong>to</strong>n Revolution<br />

CONSERVATIVESTATE.COM<br />

ANTI-IMMIGRANT SENTIMENT<br />

COMMANDER IN CHIEF !!!!<br />

VOTE TRUMP ONLY -<br />

Access To Clin<strong>to</strong>n Emails Found<br />

Pope Francis Shocks World,<br />

The Australian Tea Party<br />

Donald Trump for President!<br />

THE AMERICAN PARTY RISING<br />

Dead In Jail Cell<br />

Endorses Donald Trump for<br />

Hillary Clin<strong>to</strong>n In 2013: “I Would Like To See<br />

22 November<br />

ANTI-CLINTON<br />

President, Releases Statement<br />

People Like Donald Trump Run For Office;<br />

F@CK YOU<br />

T<br />

23 July<br />

CORRUPTION OF THE POLITICAL ESTABLISHMENT<br />

ABCNEWS.COM.CO<br />

HILLARY CLINTON<br />

D.L. Hughley<br />

They’re Honest And Can’t Be Bought”<br />

Democratic<br />

BIZSTANDARDNEWS.COM<br />

PRO-ENTREPRENEURIALIST SENTIMENTThe Burrard Street Journal<br />

WE HATE YOU!<br />

Donald Trump Protester Speaks Out: News I<br />

Rock Feed<br />

and Events<br />

Donald Trump for<br />

8 March<br />

Graham Says Christians Must<br />

Was Paid $3,500 To Protest Trumps Rally<br />

President Connecticut<br />

REACT365.COM<br />

ANTI-CLINTON<br />

Support Trump or Face Death Camps<br />

Anonymous The<br />

VETERANS FOR PRESIDENT<br />

Obama passed law for grandparents<br />

ANTI-DEMOCRAT<br />

25 July<br />

World is Ours<br />

Donald Trump<br />

DONALD J TRUMP 2016<br />

Indies 4 Trump<br />

Criminalize VS. Hillary<br />

<strong>to</strong> get all their grandchildren every<br />

KYPO6.COM<br />

N The Church<br />

All Things<br />

Conservatism Fake Clin<strong>to</strong>n News Headlines<br />

weekend<br />

of PUNK<br />

Pope Francis Shocks World,<br />

#trumpvic<strong>to</strong>ry<br />

Resist Donald J. Trumpf<br />

N/A<br />

11 March<br />

Endorses Hillary Clin<strong>to</strong>n for<br />

#neverhillary<br />

Donald Trump<br />

A FBI Agent Suspected in Hillary Email<br />

The Political Movement<br />

President, Releases Statement<br />

Deplorables for TRUMP<br />

1 Million Strong For<br />

HEAVIERMETAL.NET<br />

ANTI-TRUMP<br />

29 July<br />

Leaks Found Dead in Apparent<br />

Hillary Clin<strong>to</strong>n in 2016 Conservative<br />

RAGE AGAINST THE MACHINE To<br />

BURRARDSTREETJOURNAL.COM<br />

Veterans Murder-Suicide<br />

Donald<br />

Chonda Pierce<br />

Reunite And Release Anti Donald<br />

for America Trump<br />

Trump Claims America Should Never<br />

The Empire Herald<br />

Trump Album<br />

FU Trump<br />

President<br />

Have Given Canada Its Independence<br />

B Hillary Clin<strong>to</strong>n In 2013: “I Would Like<br />

ANTI-TRUMP<br />

NATIONALIST SENTIMENT<br />

One Nation Under God L<br />

To See People Like Donald Trump Run<br />

Females<br />

Republican<br />

AMOUNT OF INTERACTIONS<br />

For Office; They’re Honest<br />

for TRUMP Citizens<br />

Patriotic And Can’t<br />

D.L. Hughley<br />

For Trump<br />

Being Liberal<br />

Make America Party<br />

January February March April 1 May<br />

June 30 July August September Oc<strong>to</strong>ber November December<br />

Means Being<br />

Be Bought” Great Again<br />

Judge Jeanine<br />

Donald Trump For President<br />

a Hypocrite 2<br />

Pirro has Fans<br />

Colorado<br />

UNITE TO FIGHT<br />

Don P<br />

for Donald C ISIS Leader Calls for American Muslim<br />

BRING AMERICA<br />

0<br />

73,732<br />

Trump 2016<br />

Voters <strong>to</strong> Support Hillary Clin<strong>to</strong>n BACK !!! TRUMP 2016<br />

23 Jul<br />

Trump attacks "corrupt establishment",<br />

Donald Trump accepts nomination<br />

states Clin<strong>to</strong>n should be "locked up" 13 Oct<br />

Beyonce, Jay Z & Clin<strong>to</strong>n share a stage 5 Nov<br />

Donald Trump<br />

P<br />

The Original Wake<br />

and media is in "war against him"<br />

–NY Times<br />

Freedom Don't Donald J. Trump Donald Trump<br />

Commander D Donald Trump MY VOTE'S Protester FOR Speaks Out:<br />

PATRIOT<br />

Wikileaks reveals Democratic Party 25 Jul<br />

FAKE NEWS POSTED<br />

Up America<br />

–Vox<br />

Come Free<br />

for PRESIDENT America's<br />

in Chief 2020 HILLARY IN TRUMP Reclaim<br />

AMERICANS<br />

has a bias against Bernie Sanders<br />

O<br />

"I Was Paid<br />

Facebook Group<br />

The Most Explosive WikiLeaks<br />

No charge <strong>to</strong> Hillary Cin<strong>to</strong>n and no further investigation over email scandal<br />

& Devildoc<br />

President (c)<br />

2016 & 2020 The $3,500 Deplorables To America Protest<br />

FOR DONALD TRUMP<br />

6 Nov<br />

Hillary Clin<strong>to</strong>n accepts nomination 28 Jul Clin<strong>to</strong>n Revelations (So Far)<br />

–FBI Direc<strong>to</strong>r Comey<br />

Trump's Rally"<br />

1<br />

30<br />

Friends Who<br />

–Breitbart<br />

Arizona Anti-Immigrant Sheriff <strong>to</strong> Deploy Deputies at Polling Places The Resistance:<br />

Like Donald J. Trump<br />

E Obama Signs Executive Order<br />

–The Guardian The Last Line<br />

Trump Says Sexual Assault Allegations<br />

Of Defense<br />

14 Oct<br />

Patriots for a<br />

Declaring Investigation In<strong>to</strong> Election<br />

are "Made-Up S<strong>to</strong>ries"<br />

TrumpNation<br />

–IB Times<br />

Donald Trump continues <strong>to</strong> accuse Democrats of voter fraud.<br />

Free Republic -<br />

REPUBLICANS - CONSERVATIVES Results; -Revote Planned For Dec. 19th<br />

On Saturday 5th November, Trump claimed the late opening of 7 Nov<br />

Trump Squad<br />

TEA PARTY PATRIOTS-<br />

Michelle Obama denounces The Burrard<br />

R<br />

a voting site in a Latino neighorhood of Las Vegas due<br />

RIGHT WING AMERICANS<br />

Trump's comment on women Street Journal<br />

<strong>to</strong> long lines pointed <strong>to</strong> a "rigged system."<br />

Politicked<br />

F WHOA! Hillary Caught On Hot Mic<br />

–NY Times<br />

–Democracy Now<br />

WORKERS<br />

NEWS<br />

GRASSROOTS ACTIVISTS/ CITIZEN VIGILANTES<br />

PUBLIC FIGURE<br />

Trashing<br />

Trump America Trump Trump Beyonce’ USA's CEOWith RACIAL<br />

Truckers The Small Trump Steelworkers<br />

Ninth Woman Says Trump<br />

Arizona Anti-Immigrant Sheriff<br />

Tomorrow's News Trending Liberal Convervative For SLURS! Free Train Againt Pro For Against Against Pro Latino<br />

Artist Porn<br />

15 Oct<br />

11 Nov<br />

American Party<br />

Inappropriately Groped or Kissed Her<br />

<strong>to</strong> Deploy Deputies at Polling Places<br />

Bussiness<br />

News Today. News News News<br />

Society Ted Cruz Bernie Disaster Islam Financial<br />

Actress<br />

–The Guardian<br />

–The Guardian<br />

G Van Full Of Illegals Preparness Shows Up To Terrorism Vote<br />

Clin<strong>to</strong>n At SIX Polling Places, Still<br />

Ac<strong>to</strong>r<br />

Republican<br />

Hillary Clin<strong>to</strong>n<br />

OTHER<br />

Think Voter Fraud Is A Myth?<br />

For Liberty White For<br />

Against Trump Against<br />

Politician<br />

Revolution<br />

Italian Climate Page<br />

Critical Liberal News / Republican H and RAGE PropertyAGAINST Lives Information THE MACHINE To Republicans<br />

Left Skeptics Dedicated<br />

News <strong>news</strong> News News<br />

Matter Freedom<br />

<strong>to</strong> the<br />

Aggrega<strong>to</strong>rs Reunite And Release Anti Donald<br />

Colorado for Donald<br />

Againt<br />

Philipines<br />

Trump Album<br />

Gun<br />

M<br />

Trump 2016<br />

S<br />

Owners<br />

Reclaim America<br />

The Resistance: The Last<br />

I Obama Signs Executive Order Banning<br />

Journalist<br />

Line Of Defense<br />

SOCIAL CHANGE<br />

News News Alternative<br />

Against The Pledge Of Allegiance In Schools<br />

ENTERTAINERS<br />

Pro Latino Human Progessives Aggrega<strong>to</strong>r News<br />

Mainstream Anti-conservatism<br />

US Patriots<br />

rights<br />

Media Nationwide<br />

Against Gun<br />

One Nation<br />

J BREAKING Romanian Hacker With<br />

Control<br />

Under God<br />

Against Gun<br />

Advocates Access To Anonymous Clin<strong>to</strong>n Emails Found Dead<br />

Violence<br />

America's Veterans<br />

C<br />

Against In Jail Cell<br />

Corruption<br />

are Loved<br />

E<br />

Non-PC News<br />

K Ac<strong>to</strong>r Bill Murray Announces 2016<br />

EARLY CAMPAIGNERS<br />

Pro-Trump Presidential Run<br />

Donald Trump<br />

U<br />

Commander<br />

Early 2020 Early 2020 Early 2020<br />

L Pope Francis Shocks World, Endorses<br />

Bernie Trump Hillary<br />

in Chief 2020<br />

Hillary Clin<strong>to</strong>n for President, Releases<br />

I<br />

Campaigners Campaigners Campaigners<br />

LOYALISTS<br />

Q<br />

Statement<br />

F<br />

G<br />

Hillary Loyalists<br />

Trump Loyalists<br />

M Pope Francis Shocks World, Endorses<br />

S<strong>to</strong>p the Mosque<br />

Donald Trump for President, Releases<br />

F@CK YOU HILLARY<br />

Bernie Loyalists<br />

in Bendigo<br />

Statement<br />

CLINTON WE HATE YOU!<br />

DISCUSSION SPACE<br />

N Trump Claims America Should Never<br />

Political Anti<br />

D<br />

Have Given Canada Its Independence<br />

Donald Trump Commentary establishment<br />

For President<br />

O Mike Pence: “Sarah Palin Is My Role<br />

NO Hillary<br />

TrumpNation<br />

Politcal<br />

Model For Beautiful, Smart American<br />

for 2016<br />

Commentary<br />

Women” - Newslo<br />

Anti-Democrat<br />

P RUPAUL CLAIMS TRUMP TOUCHED<br />

HIM INAPPROPRIATELY IN THE<br />

B<br />

1990S<br />

BUSINESS<br />

MUSIC/ENTERTAIMENT<br />

Q President Obama Confirms He Will<br />

Punk Tat<strong>to</strong>o Fashion<br />

COMMUNITY<br />

Refuse To Leave Office If Trump Is<br />

Spritual<br />

Democrat Community<br />

Pro Military<br />

HIERARCY<br />

Elected<br />

Community<br />

Bussiness Motivational<br />

The Trump American Party<br />

TRUMP<br />

Speaker<br />

R Graham Says Christians Must Support<br />

CATEGORY<br />

Progessive<br />

FORCE ONE<br />

Community<br />

Trump or Face Death Camps<br />

A<br />

Dating app<br />

VOTE TRUMP ONLY<br />

Sub-category<br />

Make America<br />

Survivalist<br />

S African Billionaire Will Give $1 Million<br />

Republican Community<br />

THEAMERICAN<br />

Great Again<br />

To Anyone Who Wants To Leave<br />

PARTY RISING<br />

Donald Trump<br />

America if Donald Trump is Elected<br />

FACEBOOK PAGES<br />

for President<br />

President<br />

Pro-Black<br />

PER CATEGORY<br />

TRUMP - SPEAK OUT AGAINST<br />

T Trump Offering Free One-Way Tickets<br />

Religious<br />

ISLAMIZATION OF AMERICA<br />

OCCUPY MOVEMENT<br />

<strong>to</strong> Africa &amp; Mexico for Those<br />

Who Wanna Leave America<br />

Fake <strong>news</strong> s<strong>to</strong>ries Events<br />

CHAPTER 1 → RECIPE 1<br />

SERVING SUGGESTIONS<br />

This recipe may be used <strong>to</strong> better<br />

understand the publics that are<br />

animated by <strong>fake</strong> <strong>news</strong> and the meaning<br />

making activities that they engage in<br />

around <strong>fake</strong> <strong>news</strong>, i.e. how they enroll<br />

<strong>fake</strong> <strong>news</strong> in the service of their own<br />

issue work. This approach may inform<br />

a thicker description of the impact of<br />

<strong>fake</strong> <strong>news</strong> that moves away from its<br />

viral character (the single engagement<br />

number or metric) <strong>to</strong> understanding who<br />

it mobilises and how.<br />

5,649<br />

1<br />

2<br />

3<br />

WOMEN-FOCUSED<br />

U Obama passed law for grandparents <strong>to</strong><br />

get all their grandchildren every<br />

weekend<br />

CLICKBAIT<br />

36 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

37


CHAPTER 1 → RECIPE 2<br />

START<br />

select a <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry<br />

you want <strong>to</strong> trace<br />

HOW MAY THE TRAJECTORY<br />

OF A FAKE NEWS STORY<br />

BE TRACED ON FACEBOOK?<br />

design queries <strong>to</strong> identify<br />

URLs where s<strong>to</strong>ry is published<br />

Google Search<br />

compile list of URLs<br />

where s<strong>to</strong>ry is published<br />

input URL in<br />

CrowdTangle<br />

input each URL in<br />

CrowdTangle<br />

BEFORE STARTING<br />

For this recipe it is recommended that a <strong>fake</strong><br />

<strong>news</strong> s<strong>to</strong>ry is taken as a starting point and the<br />

URL or URLs or on which it is published are<br />

identified. To illustrate this recipe we have<br />

selected as case studies two prominent s<strong>to</strong>ries<br />

about the 2016 US presidential elections, namely<br />

"Trump Offering Free One-Way Tickets <strong>to</strong><br />

Africa & Mexico for Those Who Wanna Leave<br />

America," a s<strong>to</strong>ry that exploits anti-immigrant<br />

sentiment and "Rage Against the Machine<br />

<strong>to</strong> Reunite and Release Anti Donald Trump<br />

Album," which exploits anti-Trump sentiment.<br />

This recipe comes in two flavours. In step one<br />

you will learn <strong>to</strong> trace Facebook pages and<br />

groups in which the original s<strong>to</strong>ry URL is posted<br />

and plot them on a timeline (a). In step two<br />

this analysis will be extended <strong>to</strong> all URLs on<br />

which a s<strong>to</strong>ry has been republished (b).<br />

output data<br />

input data <strong>to</strong><br />

RAWGraphs<br />

a<br />

visualise<br />

> Fake <strong>news</strong> s<strong>to</strong>ry URLs<br />

> Facebook pages and<br />

groups<br />

that share the URLs<br />

> Number of followers<br />

per page or group<br />

> Date of sharing<br />

of the s<strong>to</strong>ry<br />

HOW DOES THE STORY<br />

"RAGE AGAINST THE<br />

MACHINE TO REUNITE<br />

AND RELEASE ANTI<br />

DONALD TRUMP ALBUM"<br />

TRAVEL ON FACEBOOK?<br />

output data for each URL<br />

input data <strong>to</strong><br />

RAWGraphs<br />

visualise<br />

> Fake <strong>news</strong> s<strong>to</strong>ry URLs<br />

> Facebook pages and<br />

groups<br />

that share the URLs<br />

> Number of followers<br />

per page or group<br />

> Date of sharing<br />

of the s<strong>to</strong>ry<br />

b<br />

HOW DOES THE STORY<br />

"TRUMP OFFERING FREE<br />

ONE-WAY TICKETS TO AFRICA<br />

& MEXICO FOR THOSE WHO<br />

WANNA LEAVE AMERICA" AND<br />

ITS DEBUNKED VERSIONS<br />

TRAVEL ON FACEBOOK?<br />

38 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

39


input URL in<br />

CrowdTangle<br />

output data<br />

input data <strong>to</strong><br />

RAWGraphs<br />

a<br />

visualise<br />

START<br />

> Fake <strong>news</strong> s<strong>to</strong>ry URLs<br />

> Facebook pages and<br />

groups<br />

that share the URLs<br />

> Number of followers<br />

per page or group<br />

> Date of sharing<br />

of the s<strong>to</strong>ry<br />

select a <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry<br />

you want <strong>to</strong> trace<br />

HOW DOES THE STORY<br />

"RAGE AGAINST THE<br />

MACHINE TO REUNITE<br />

AND RELEASE ANTI<br />

DONALD TRUMP ALBUM"<br />

TRAVEL ON FACEBOOK?<br />

CHAPTER 1 → RECIPE 2<br />

a. IDENTIFY FACEBOOK PAGES AND<br />

GROUPS THAT SHARE A FAKE NEWS<br />

STORY VIA THE ORIGINAL URL<br />

Facebook pages and groups that share a <strong>fake</strong><br />

<strong>news</strong> s<strong>to</strong>ry may be detected through a social<br />

media moni<strong>to</strong>ring <strong>to</strong>ol such as → CrowdTangle’s<br />

browser extension.<br />

◊<br />

◊<br />

◊<br />

The names of these pages and groups, their<br />

followers’ count, the interactions that they<br />

generate as well as the date of sharing of the<br />

s<strong>to</strong>ry may be recorded in a spreadsheet per<br />

s<strong>to</strong>ry URL.<br />

To explore the temporal dynamics of<br />

the circulation of the <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry<br />

on Facebook, you may plot its trajec<strong>to</strong>ry<br />

across pages and groups on a timeline.<br />

→ RAWGraphs can be used <strong>to</strong> create the<br />

base layer of the visualisation. Note which<br />

publics engage with the s<strong>to</strong>ry as well as<br />

whether the moment of debunking a s<strong>to</strong>ry<br />

affects its circulation.<br />

To take the analysis one step further, a<br />

qualitative analysis of how <strong>fake</strong> <strong>news</strong> is<br />

enrolled by each of these pages <strong>to</strong> support<br />

issue work may be undertaken. This may<br />

be done by examining the context in which<br />

the s<strong>to</strong>ries are shared, i.e. whether they<br />

are shared uncritically or called out as <strong>fake</strong><br />

as well as how they are framed in relation<br />

<strong>to</strong> the issues represented by the pages that<br />

share them. It is <strong>to</strong> be noted that such<br />

analysis might at times be difficult due <strong>to</strong> the<br />

fact that for the most prominent <strong>fake</strong> <strong>news</strong><br />

s<strong>to</strong>ries Facebook posts that share them are<br />

often no longer available via the Facebook<br />

interface and API.<br />

40 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

41


HOW DOES THE STORY<br />

"RAGE AGAINST THE<br />

MACHINE TO REUNITE<br />

AND RELEASE ANTI<br />

DONALD TRUMP<br />

ALBUM" TRAVEL ON<br />

FACEBOOK?<br />

Pubished the 11th of March 2016 on heaviermetal.net<br />

RAGE AGAINST THE MACHINE<br />

To Reunite And Release Anti Donald Trump Album<br />

The Church of PUNK<br />

Justice Through Music<br />

Rock Feed<br />

Rock.It Boy Entertainment<br />

Indie DaaDee<br />

Magic 89.3<br />

98 Rock<br />

Alabama Citizens for Bernie Sanders 2020<br />

Trajec<strong>to</strong>ry of “Rage Against the<br />

Machine <strong>to</strong> Reunite and Release<br />

Anti Donald Trump Album" s<strong>to</strong>ry on<br />

Facebook pages and groups. The s<strong>to</strong>ry<br />

circulates best between March and June<br />

2016 as satire amongst English language<br />

music and entertainment groups. It<br />

is revived in November after the US<br />

elections, when it is also picked up by<br />

Italian music and political pages.<br />

Anonymous The World is Ours<br />

This page is very offensive<br />

Bernie Sanders for President 2020<br />

Occupy Orange County CA<br />

Wunderkynd<br />

Ballin' On A Budget<br />

No Music No Life.<br />

RAGE AGAINST THE MACHINE CHRISTMAS NO.1 2009<br />

Circo Voador<br />

FM 949<br />

Bernie Sanders - Support from abroad<br />

Boogat<br />

FM 949<br />

107.9 The Mix<br />

Occupy London<br />

Occupy movement<br />

United Steelworkers Activists for Political Revolution<br />

United People for Latinos in Film TV and Theater<br />

Missoula for our Green Party Revolution<br />

STORIES SHARED ON FB<br />

Post<br />

Page where is shared<br />

PAKISTAN RISING<br />

Our Revolution Continues- Activist Resources<br />

Dj Mike Style!!!<br />

Heavier Metal<br />

METAL ARMADA<br />

Rock Feed<br />

The s<strong>to</strong>ry reappears on Facebook<br />

at the 11th of November, shared by<br />

the page “Apartment Khunpa”<br />

Interactions<br />

Heavier Metal<br />

Music and/or Radio<br />

Activism<br />

Rock Feed<br />

No sharing of the s<strong>to</strong>ry URL on Facebook from late May <strong>to</strong> early November<br />

Apartment Khunpa<br />

The Guitar Mag<br />

Politics<br />

Banda Bassotti<br />

Others<br />

Marcello Belotti - Sinistra Italiana EU<br />

POSTS INTERACTIONS<br />

0 1588<br />

“Rock Feed” is sharing the s<strong>to</strong>ry<br />

both at the earliest as well as<br />

at the end of the first wave of posting.<br />

ROCK Mundial<br />

The Church of PUNK<br />

Radioactivo 98.5<br />

Los Angeles Punk Museum<br />

POSTS ON A SAME PAGE<br />

Veterans Against Rep. Ignorance<br />

FU Trump<br />

Occupy Orange County CA<br />

11/03/2016 APRIL<br />

MAY JUNE JULY AUGUST SEPTEMBER OCTOBER NOVEMBER DECEMBER<br />

42 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

43


START<br />

select a <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry<br />

you want <strong>to</strong> trace<br />

design queries <strong>to</strong> identify<br />

URLs where s<strong>to</strong>ry is published<br />

Google Search<br />

CHAPTER 1 → RECIPE 2<br />

compile list of URLs<br />

where s<strong>to</strong>ry is published<br />

input each URL in<br />

CrowdTangle<br />

b. IDENTIFY FACEBOOK PAGES AND<br />

GROUPS THAT SHARE ALL INSTANCES<br />

OF A FAKE NEWS STORY<br />

As <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries may be republished by a<br />

number of sources, the previous analysis may<br />

be enriched by tracing the circulation not only<br />

of the original URL on which the chosen s<strong>to</strong>ry<br />

is posted but all instances of s<strong>to</strong>ry republication<br />

across a number of different sites.<br />

output data for each URL<br />

input data <strong>to</strong><br />

RAWGraphs<br />

b<br />

visualise<br />

> Fake <strong>news</strong> s<strong>to</strong>ry URLs<br />

> Facebook pages and<br />

groups<br />

that share the URLs<br />

> Number of followers<br />

per page or group<br />

> Date of sharing<br />

of the s<strong>to</strong>ry<br />

◊<br />

◊<br />

To identify the websites which republish<br />

a s<strong>to</strong>ry as well as those which debunk it,<br />

you may query the title of the <strong>fake</strong> <strong>news</strong> in<br />

a search engine of choice (e.g. → Google<br />

Web Search) using a research browser [1]<br />

and extract the URLs corresponding <strong>to</strong><br />

instances of republication and debunking<br />

of the s<strong>to</strong>ry from the returned list of<br />

results.<br />

Query the resulting URLs in a social<br />

moni<strong>to</strong>ring <strong>to</strong>ol (such as → CrowdTangle)<br />

<strong>to</strong> get the list of Facebook groups and pages<br />

that share the URLs corresponding <strong>to</strong> both<br />

the <strong>fake</strong> s<strong>to</strong>ry and its debunked versions.<br />

[1]<br />

See instructions<br />

on how <strong>to</strong> set up<br />

a research browser<br />

in this video<br />

tu<strong>to</strong>rial: https://<br />

www.youtube.com/<br />

watch?v=bj65Xr9GkJM<br />

HOW DOES THE STORY<br />

"TRUMP OFFERING FREE<br />

ONE-WAY TICKETS TO AFRICA<br />

& MEXICO FOR THOSE WHO<br />

WANNA LEAVE AMERICA" AND<br />

ITS DEBUNKED VERSIONS<br />

TRAVEL ON FACEBOOK?<br />

◊<br />

You may plot these pages on a timeline <strong>to</strong><br />

see whether different <strong>fake</strong> <strong>news</strong> sources<br />

spawn different s<strong>to</strong>ry trajec<strong>to</strong>ries on<br />

Facebook and whether debunked versions<br />

are being acknowledged.<br />

44 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

45


HOW DOES THE<br />

STORY "TRUMP<br />

OFFERING FREE ONE-<br />

WAY TICKETS TO<br />

AFRICA & MEXICO FOR<br />

THOSE WHO WANNA<br />

LEAVE AMERICA"<br />

AND ITS DEBUNKED<br />

VERSIONS TRAVEL ON<br />

FACEBOOK?<br />

Timeline of "Trump Offering<br />

Free One-Way Tickets <strong>to</strong> Africa<br />

& Mexico for Those Who Wanna<br />

Leave America” s<strong>to</strong>ry and its sites<br />

of publication on the web and<br />

Facebook. The s<strong>to</strong>ry is republished<br />

without critical context on multiple<br />

clickbait sites in the week following<br />

its original publication. This gives the<br />

s<strong>to</strong>ry multiple lives on Facebook. Its<br />

FAKE STORY INSTANCES<br />

sharing on a <strong>fake</strong> <strong>news</strong> site animates<br />

political<br />

Earliest<br />

publics<br />

version<br />

while<br />

(original)<br />

its sharing on<br />

clickbait Reposted sites Version sees the s<strong>to</strong>ry being<br />

recycled as clickbait by viral pages.<br />

Debunked version<br />

The publics sparked in<strong>to</strong> being by the<br />

<strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry and the debunked<br />

POSTS ON FACEBOOK<br />

version thereof do not overlap.<br />

Post Page that shares it<br />

Published the 11th of November 2016 on viralmugshot.com<br />

Trump Offering Free One-Way Tickets <strong>to</strong> Africa & Mexico<br />

for Those Who Wanna Leave America<br />

VIRALMUGSHOT.COM<br />

MYHIPHOP.ORG<br />

Daily Media Buzz<br />

YEPSEE.COM<br />

Amazing and Crazy videos<br />

Crazy Accidents 18+<br />

Damn<br />

Embarrassing Party Pho<strong>to</strong>s<br />

EpicFail.com<br />

I post disturbing things.<br />

New York Meta<br />

The Craziest fight videos<br />

THUGIFY.COM<br />

ThugLifeClips<br />

DAILYNEWSPOSTS.INFO<br />

DONALD TRUMP<br />

COMMANDER IN CHIEF !!!!<br />

NO Hillary for 2016<br />

TRUMP - SPEAK OUT AGAINST ISLAMIZATION<br />

OF AMERICA - FIM DA COLONIZAÇÃO ISLÃ<br />

Trump USA's CEO<br />

TrumpNation<br />

VOTE TRUMP ONLY - THE AMERICAN PARTY RISING<br />

ATRENDING.NET<br />

Exposing The Truth √<br />

QATARDAY.COM<br />

Qatar Day<br />

The first instance of the s<strong>to</strong>ry appears on viralmugshot.com in early November, but CrowdTangle only indicates instances of Facebook sharing in late February<br />

The first repost has been<br />

published by myhiphop.org on<br />

the 11th of November<br />

The s<strong>to</strong>ry is published<br />

on Yepsee.com and has<br />

been shared on eight<br />

different Facebook<br />

pages/groups during the<br />

same day<br />

NYMAG.COM/SELECTALL<br />

New York Magazine<br />

Select All<br />

VANITYFAIR.COM<br />

Pakistan Rising<br />

vanityfairmagazine<br />

VFHive<br />

THEVERGE.COM<br />

TechKnowledgeIt<br />

verge<br />

NEWSER.COM<br />

<strong>news</strong>er<br />

BUZZFEED.COM<br />

BuzzFeed News<br />

BuzzFeed Politics<br />

BuzzFeed UK News<br />

BuzzFeed World<br />

Isaac Saul<br />

StyleHaul<br />

The European Journalism Observa<strong>to</strong>ry | EJO<br />

BuzzFeed<br />

Coloradans For Bernie<br />

Ethical Journalism Network<br />

CIVILIZED.LIFE<br />

civilized.life<br />

INQUISITR.COM<br />

theinquisitr<br />

The first debunk has been<br />

published by nymag.com<br />

on the 30th of December<br />

Prince Sam<br />

TheCelebtricity<br />

Don P<br />

Hillary Clin<strong>to</strong>n The People's Choice<br />

FAKE STORY INSTANCES<br />

Earliest version (original)<br />

Reposted Version<br />

Debunked version<br />

POSTS ON FACEBOOK<br />

Post Page that shares it<br />

PAKISTAN RISING<br />

Interactions<br />

Fake S<strong>to</strong>ry<br />

Debunk<br />

POSTS INTERACTIONS<br />

0 1588<br />

POSTS ON A SAME PAGE<br />

PAKISTAN RISING<br />

Interactions<br />

Fake S<strong>to</strong>ry<br />

Debunk<br />

POSTS INTERACTIONS<br />

0 1588<br />

POSTS ON A SAME PAGE<br />

DIYHILFE.COM<br />

Mobbing Hilfe<br />

HOPEFORNIGERIAONLINE.COM<br />

Hope for Nigeria<br />

LUXURYANDGLAMOR.COM<br />

Daily Video<br />

POLITICO.COM<br />

JimHeathTV<br />

LionPublishers<br />

mnsucollegedemocrats<br />

modernliberals<br />

PolPreposterous<br />

No sharing on Facebook from mid November <strong>to</strong> late December<br />

politico<br />

pg/politicomag<br />

politicomediadesk<br />

PENamerican<br />

DavidWilliamBateman<br />

2017<br />

NOV 13 NOV 20 NOV 27 DEC 04 DEC 11 DEC 18 DEC 25 JAN 01<br />

INEWS.CO.UK<br />

theipaper<br />

TRUEAMERICANS.ME<br />

President Donald J. Trump<br />

TRUMP - SPEAK OUT AGAINST [...]<br />

EDUBLOG.SCHOLASTIC.COM<br />

EmmaClarkLibrary<br />

NOWTORONTO.COM<br />

nowmagazine<br />

JAN 08 JAN 15 JAN 22 JAN 29 FEB 05 FEB 12 FEB 19 FEB 26 MAR 05 MAR 12<br />

46 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

47


The s<strong>to</strong>ry reappears on Facebook on<br />

the 11th of November, shared by the<br />

page “Apartment Khunpa<br />

CHAPTER 1 → RECIPE 2<br />

Published the 11th of November 2016 on viralmugshot.com<br />

Trump Offering Free One-Way Tickets <strong>to</strong> Africa & Mexico<br />

for Those Who Wanna Leave America<br />

Apartment Khunpa<br />

The Guitar Mag<br />

Banda Bassotti<br />

DAILYNEWSPOSTS.INFO<br />

Marcello Belotti - Sinistra Italiana EU<br />

DONALD TRUMP<br />

ROCK Mundial<br />

FAKE STORY INSTANCES<br />

Debunked version<br />

The Church of PUNK<br />

Post<br />

Earliest version (original)<br />

Reposted Version<br />

POSTS ON FACEBOOK<br />

PAKISTAN RISING<br />

Interactions<br />

Radioactivo 98.5<br />

POSTS INTERACTIONS<br />

Hope for Nigeria<br />

Los Angeles Punk Museum<br />

0 1588<br />

Page that shares it<br />

Fake S<strong>to</strong>ry<br />

Debunk<br />

POSTS ON A SAME PAGE<br />

VIRALMUGSHOT.COM<br />

MYHIPHOP.ORG<br />

Daily Media Buzz<br />

YEPSEE.COM<br />

Amazing and Crazy videos<br />

Crazy Accidents 18+<br />

Damn<br />

Embarrassing Party Pho<strong>to</strong>s<br />

EpicFail.com<br />

I post disturbing things.<br />

New York Meta<br />

The Craziest fight videos<br />

THUGIFY.COM<br />

ThugLifeClips<br />

COMMANDER IN CHIEF !!!!<br />

NO Hillary for 2016<br />

TRUMP - SPEAK OUT AGAINST ISLAMIZATION<br />

OF AMERICA - FIM DA COLONIZAÇÃO ISLÃ<br />

Trump USA's CEO<br />

TrumpNation<br />

VOTE TRUMP ONLY - THE AMERICAN PARTY RISING<br />

ATRENDING.NET<br />

Exposing The Truth √<br />

QATARDAY.COM<br />

Qatar Day<br />

DIYHILFE.COM<br />

Mobbing Hilfe<br />

HOPEFORNIGERIAONLINE.COM<br />

LUXURYANDGLAMOR.COM<br />

Daily Video<br />

Veterans Against Rep. Ignorance<br />

The first instance of the s<strong>to</strong>ry appears on viralmugshot.com in early November, but CrowdTangle only indicates instances of Facebook sharing in late February<br />

The first repost has been<br />

published by myhiphop.org on<br />

the 11th of November<br />

The s<strong>to</strong>ry is published<br />

on Yepsee.com and has<br />

been shared on eight<br />

different Facebook<br />

pages/groups during the<br />

same day<br />

NYMAG.COM/SELECTALL<br />

New York Magazine<br />

Select All<br />

VANITYFAIR.COM<br />

Pakistan Rising<br />

vanityfairmagazine<br />

VFHive<br />

POLITICO.COM<br />

JimHeathTV<br />

LionPublishers<br />

mnsucollegedemocrats<br />

modernliberals<br />

PolPreposterous<br />

No sharing on Facebook from mid November <strong>to</strong> late December<br />

politico<br />

pg/politicomag<br />

politicomediadesk<br />

PENamerican<br />

<strong>news</strong>er<br />

DavidWilliamBateman<br />

civilized.life<br />

2017<br />

NOV 13 NOV 20 NOV 27 DEC 04 DEC 11 DEC 18 DEC 25 JAN 01<br />

THEVERGE.COM<br />

TechKnowledgeIt<br />

verge<br />

NEWSER.COM<br />

BUZZFEED.COM<br />

BuzzFeed News<br />

BuzzFeed Politics<br />

BuzzFeed UK News<br />

BuzzFeed World<br />

Isaac Saul<br />

StyleHaul<br />

The European Journalism Observa<strong>to</strong>ry | EJO<br />

BuzzFeed<br />

Coloradans For Bernie<br />

Ethical Journalism Network<br />

CIVILIZED.LIFE<br />

INQUISITR.COM<br />

theinquisitr<br />

The first debunk has been<br />

published by nymag.com<br />

on the 30th of December<br />

INEWS.CO.UK<br />

theipaper<br />

Prince Sam<br />

TheCelebtricity<br />

Don P<br />

Hillary Clin<strong>to</strong>n The People's Choice<br />

TRUEAMERICANS.ME<br />

President Donald J. Trump<br />

TRUMP - SPEAK OUT AGAINST [...]<br />

EDUBLOG.SCHOLASTIC.COM<br />

EmmaClarkLibrary<br />

nowmagazine<br />

JAN 08 JAN 15 JAN 22 JAN 29 FEB 05 FEB 12 FEB 19 FEB 26 MAR 05 MAR 12<br />

NOWTORONTO.COM<br />

SERVING SUGGESTIONS<br />

This recipe may be used <strong>to</strong><br />

understand the trajec<strong>to</strong>ry of a<br />

<strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry on Facebook, the<br />

different phases of its life cycle<br />

as well as key moments and<br />

intermediaries associated with its<br />

dissemination.<br />

FU Trump<br />

Occupy Orange County CA<br />

OBER NOVEMBER DECEMBER<br />

48 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

49


CHAPTER 1 → RECIPE 3<br />

DO FACT-CHECKING<br />

INITIATIVES REACH THE<br />

PUBLICS OF FAKE NEWS<br />

ON FACEBOOK?<br />

START<br />

List of 22 URLs of political<br />

<strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries<br />

Source: BuzzFeed News<br />

query<br />

“Fake News Title 1” + Fake<br />

“Fake News Title 2” + Fake<br />

“. . .” in<br />

Google Web Search<br />

Retain the <strong>to</strong>p ranked URL<br />

of a debunk per <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry<br />

. . .<br />

merge all URLs in a single list<br />

+<br />

BEFORE STARTING<br />

This recipe takes as a starting point a list of<br />

<strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries. There are different ways<br />

of obtaining these lists – including starting<br />

with existing lists as well as creating your<br />

own. To illustrate this recipe we use an<br />

already existing list of 22 <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries<br />

about various political issues pertaining <strong>to</strong><br />

the 2016 presidential elections in the US that<br />

generated most engagement on Facebook.<br />

These were identified by BuzzFeed News.<br />

There are two steps <strong>to</strong> this recipe. The first<br />

is <strong>to</strong> identify URLs that circulate corrections<br />

or “debunking web pages” for each <strong>fake</strong><br />

<strong>news</strong> s<strong>to</strong>ry. The second is <strong>to</strong> explore how<br />

Facebook groups engage with both <strong>fake</strong> <strong>news</strong><br />

s<strong>to</strong>ries and their corresponding debunking<br />

web pages (b).<br />

input URLs in<br />

CrowdTangle<br />

output data<br />

> URLs of <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry<br />

or debunked version thereof<br />

> Facebook pages and groups<br />

that share either the <strong>fake</strong><br />

<strong>news</strong> s<strong>to</strong>ry or its correction<br />

> number of followers<br />

per page or group<br />

input data <strong>to</strong><br />

RAWGraphs<br />

visualise<br />

ARE DEBUNKED VERSIONS<br />

OF FAKE NEWS STORIES<br />

ACKNOWLEDGED BY THE<br />

PUBLICS OF FAKE NEWS?<br />

50 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

51


START<br />

List of 22 URLs of political<br />

<strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries<br />

Source: BuzzFeed News<br />

query<br />

“Fake News Title 1” + Fake<br />

“Fake News Title 2” + Fake<br />

“. . .” in<br />

Google Web Search<br />

Retain the <strong>to</strong>p ranked URL<br />

of a debunk per <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry<br />

. . .<br />

CHAPTER 1 → RECIPE 3<br />

a. IDENTIFY WEB PAGES WHICH AIM TO<br />

DEBUNK FAKE NEWS STORIES<br />

To identify prominent debunking web pages<br />

for a given <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry you may use the<br />

→ Google Web Search engine. Alternatively,<br />

you may also query fact-checking sites for<br />

keywords describing a <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry.<br />

◊<br />

◊<br />

◊<br />

In order <strong>to</strong> find corrections of <strong>fake</strong> <strong>news</strong><br />

articles queries need <strong>to</strong> be designed for<br />

each <strong>fake</strong> <strong>news</strong> item in your list. One<br />

strategy would be <strong>to</strong> use the title of the<br />

s<strong>to</strong>ry in quotation marks followed by the<br />

word “<strong>fake</strong>” (e.g. “Trump Offering Free<br />

One-Way Tickets <strong>to</strong> Africa & Mexico for<br />

Those Who Wanna Leave America”, <strong>fake</strong>).<br />

You may use the search engine ranking<br />

as an indication of salience of correction<br />

and select the highest ranked URLs<br />

corresponding <strong>to</strong> a corrected version of the<br />

<strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry in question.<br />

The result of this step is a list of URLs<br />

containing the most highly ranked<br />

debunking web pages per <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry.<br />

52 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

53


START<br />

List of 22 URLs of political<br />

<strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries<br />

Source: BuzzFeed News<br />

merge all URLs in a single list<br />

+<br />

input URLs in<br />

CrowdTangle<br />

query<br />

“Fake News Title 1” + Fake<br />

“Fake News Title 2” + Fake<br />

“. . .” in<br />

Google Web Search<br />

Retain the <strong>to</strong>p ranked URL<br />

of a debunk per <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry<br />

output data<br />

> URLs of <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry<br />

or debunked version thereof<br />

> Facebook pages and groups<br />

that share either the <strong>fake</strong><br />

<strong>news</strong> s<strong>to</strong>ry or its correction<br />

> number of followers<br />

per page or group<br />

. . .<br />

CHAPTER 1 → RECIPE 3<br />

b. MAP THE OVERLAP BETWEEN THE<br />

PUBLICS OF FAKE NEWS STORIES AND<br />

WEB PAGES WHICH AIM TO DEBUNK<br />

THEM<br />

Facebook pages and groups that share both <strong>fake</strong><br />

<strong>news</strong> s<strong>to</strong>ries as well as web pages which aim <strong>to</strong><br />

debunk them may be detected through a social<br />

media moni<strong>to</strong>ring <strong>to</strong>ol such as → CrowdTangle’s<br />

browser extension.<br />

◊<br />

◊<br />

To explore whether the debunking web<br />

pages are acknowledged by the publics<br />

which share the <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries, identify<br />

whether there is an overlap between the<br />

Facebook pages and groups that share <strong>fake</strong><br />

<strong>news</strong> s<strong>to</strong>ries and those debunking web<br />

pages issued in responses.<br />

This may be illustrated by means of a<br />

“circle packing” visualisation. You may use<br />

→ RawGraph for this operation.<br />

input data <strong>to</strong><br />

RAWGraphs<br />

visualise<br />

ARE DEBUNKED VERSIONS<br />

OF FAKE NEWS STORIES<br />

ACKNOWLEDGED BY THE<br />

PUBLICS OF FAKE NEWS?<br />

54 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

55


ARE DEBUNKING<br />

WEB PAGES<br />

ACKNOWLEDGED<br />

BY THE PUBLICS OF<br />

FAKE NEWS?<br />

Fake <strong>news</strong> pages and debunking<br />

web pages have different publics<br />

on Facebook. Only six of the pages<br />

that share <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries have<br />

acknowledged web pages which aim<br />

<strong>to</strong> debunk them. While Google looks<br />

<strong>to</strong> prioritise debunking web pages, on<br />

Facebook it is <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries that<br />

circulate better. While both progressive<br />

and conservative pages share <strong>fake</strong><br />

<strong>news</strong> s<strong>to</strong>ries it is primarily progressive<br />

Facebook pages and those pertaining<br />

<strong>to</strong> journalists and fact-checking<br />

initiatives that share web pages which<br />

aim <strong>to</strong> debunk <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries.<br />

Rage Against the Machine To Reunite<br />

And Release Anti Donald Trump Album<br />

Pubished the 11th of March 2016<br />

on heaviermetal.net<br />

Rupaul claims Trump <strong>to</strong>uched<br />

him inappropriately in the 1990s<br />

Pubished the 14th of Oc<strong>to</strong>ber 2016<br />

on world<strong>news</strong>dailyreport.com<br />

D.L. Hughley<br />

Ac<strong>to</strong>r Bill Murray Announces<br />

2016 Presidential Run<br />

Pubished the 3th of Oc<strong>to</strong>ber 2016<br />

on www.abc<strong>news</strong>.com.co Pubished the 18th of September 2016<br />

on www.politicops.com<br />

African Billionaire Will Give $1<br />

Million To Anyone Who Wants<br />

To Leave America if Donald Trump<br />

is Elected President<br />

Pubished the 3th of November 2016<br />

on www.empireherald.com<br />

Mike Pence: “Sarah Palin Is<br />

My Role Model For Beautiful,<br />

Smart American Women”<br />

BREAKING Romanian Hacker<br />

With Access To Clin<strong>to</strong>n Emails<br />

Found Dead In Jail Cell<br />

Pubished the 6th of July 2016<br />

on Christian Times<br />

Bestweb<strong>to</strong>ols<br />

Anonymous Portugal<br />

Internacional<br />

President Obama Confirms<br />

He Will Refuse To Leave Office<br />

If Trump Is Elected<br />

Pubished the 7th of September 2016<br />

on www.burrardstreetjournal.com<br />

ISIS Leader Calls for American<br />

Muslim Voters <strong>to</strong> Support<br />

Hillary Clin<strong>to</strong>n<br />

Pubished the 11th of Oc<strong>to</strong>ber 2016<br />

on world<strong>news</strong>dailyreport.com<br />

Trump Offering Free One-Way<br />

Tickets <strong>to</strong> Africa; Mexico for Those<br />

Who Wanna Leave America<br />

Pubished the 11th of November 2016<br />

on www.tmzhiphop.com<br />

Obama passed law for grandparents<br />

<strong>to</strong> get all their grandchildren<br />

every weekend<br />

Pubished the 8th of March 2016<br />

on www.react365.com<br />

WHOA! Hillary Caught On<br />

Hot Mic Trashing Beyonce’<br />

With RACIAL SLURS!<br />

Pubished the 5th of November 2016<br />

on www.thelastlineofdefense.org<br />

Hillary Clin<strong>to</strong>n In 2013: “I Would Like<br />

To See People Like Donald Trump Run<br />

For Office; They’re Honest And Can’t<br />

Be Bought”<br />

Pubished the 17th of Oc<strong>to</strong>ber 2016<br />

on conservativestate.com<br />

Graham Says Christians<br />

Must Support Trump or Face<br />

Death Camps<br />

Pubished the 23th of July 2016<br />

on www.bizstandard<strong>news</strong>.com<br />

Trump Claims America<br />

Should Never Have Given<br />

Canada Its Independence<br />

Pubished the 29th of July 2016<br />

on www.burrardstreetjournal.com<br />

Van Full Of Illegals Shows Up To<br />

Vote Clin<strong>to</strong>n At SIX Polling Places,<br />

Still Think Voter Fraud Is A Myth?<br />

Pubished the 5th of November 2016<br />

on www.thelastlineofdefense.org<br />

FBI Agent Suspected in Hillary<br />

Email Leaks Found Dead<br />

in Apparent Murder-Suicide<br />

Pubished the 15th of Oc<strong>to</strong>ber 2016<br />

on www.usa<strong>news</strong>flash.com<br />

Page or group which shares both<br />

<strong>fake</strong> s<strong>to</strong>ry and its correction<br />

Fake<br />

Debunk<br />

Obama Signs Executive Order<br />

Declaring Investigation In<strong>to</strong> Election<br />

Results; Revote Planned For Dec. 19th<br />

Pubished the 12th of December 2016<br />

on www.abc<strong>news</strong>.com.co<br />

Pope Francis Shocks World,<br />

Endorses Hillary Clin<strong>to</strong>n for<br />

President, Releases Statement<br />

Pubished the 25th of July 2016<br />

on www.kypo6.com<br />

Pope Francis Shocks World,<br />

Endorses Donald Trump for<br />

President, Releases Statement<br />

Pubished the 26th of September 2016<br />

on Ending The Fed<br />

Obama Signs Executive Order<br />

Banning The Pledge Of Allegiance<br />

In Schools Nationwide<br />

Pubished the 11th of December 2016<br />

on www.abc<strong>news</strong>.com.co<br />

Donald Trump Protester<br />

Speaks Out: I Was Paid $3,500<br />

To Protest Trump's Rally<br />

Pubished the 22th of November 2016<br />

on www.abc<strong>news</strong>.com.co<br />

POST INTERACTIONS<br />

Hillary Clin<strong>to</strong>n<br />

for President in 2016<br />

Criminalize<br />

Conservatism<br />

100<br />

41271<br />

Connecticut<br />

Progressives<br />

PolitiFact<br />

Read the damn <strong>news</strong><br />

56 A FIELD GUIDE TO FAKE NEWS CHAPTER 1: MAPPING FAKE NEWS HOTSPOTS ON FACEBOOK<br />

57


Chapter #2<br />

1.3.A<br />

Rage Against the Machine To Reunite<br />

And Release Anti Donald Trump Album<br />

Pubished the 11th of March 2016<br />

on heaviermetal.net<br />

Ac<strong>to</strong>r Bill Murray Announces<br />

Mike Pence: “Sarah Palin Is<br />

2016 Presidential Run<br />

My Role Model For Beautiful,<br />

Pubished the 3th of Oc<strong>to</strong>ber 2016<br />

Smart American Women”<br />

on www.abc<strong>news</strong>.com.co Pubished the 18th of September 2016<br />

on www.politicops.com<br />

CHAPTER 1 → RECIPE 3<br />

President Obama Confirms<br />

He Will Refuse To Leave Office<br />

If Trump Is Elected<br />

Pubished the 7th of September 2016<br />

on www.burrardstreetjournal.com<br />

Trump Offering Free One-Way<br />

Tickets <strong>to</strong> Africa; Mexico for Those<br />

Who Wanna Leave America<br />

Pubished the 11th of November 2016<br />

on www.tmzhiphop.com<br />

Obama passed law for grandparents<br />

<strong>to</strong> get all their grandchildren<br />

every weekend<br />

Pubished the 8th of March 2016<br />

on www.react365.com<br />

TRACING THE<br />

CIRCULATION OF FAKE<br />

NEWS ON THE WEB<br />

Rupaul claims Trump <strong>to</strong>uched<br />

him inappropriately in the 1990s<br />

Pubished the 14th of Oc<strong>to</strong>ber 2016<br />

on world<strong>news</strong>dailyreport.com<br />

African Billionaire Will Give $1<br />

Million To Anyone Who Wants<br />

To Leave America if Donald Trump<br />

is Elected President<br />

Pubished the 3th of November 2016<br />

on www.empireherald.com<br />

BREAKING Romanian Hacker<br />

With Access To Clin<strong>to</strong>n Emails<br />

Found Dead In Jail Cell<br />

Pubished the 6th of July 2016<br />

on Christian Times<br />

ISIS Leader Calls for American<br />

Muslim Voters <strong>to</strong> Support<br />

Hillary Clin<strong>to</strong>n<br />

Pubished the 11th of Oc<strong>to</strong>ber 2016<br />

on world<strong>news</strong>dailyreport.com<br />

WHOA! Hillary Caught On<br />

Hot Mic Trashing Beyonce’<br />

With RACIAL SLURS!<br />

Pubished the 5th of November 2016<br />

on www.thelastlineofdefense.org<br />

Bestweb<strong>to</strong>ols<br />

Anonymous Portugal<br />

Internacional<br />

D.L. Hughley<br />

Hillary Clin<strong>to</strong>n In 2013: “I Would Like<br />

To See People Like Donald Trump Run<br />

For Office; They’re Honest And Can’t<br />

Be Bought”<br />

Pubished the 17th of Oc<strong>to</strong>ber 2016<br />

on conservativestate.com<br />

Graham Says Christians<br />

Must Support Trump or Face<br />

Death Camps<br />

Pubished the 23th of July 2016<br />

on www.bizstandard<strong>news</strong>.com<br />

Trump Claims America<br />

Should Never Have Given<br />

Canada Its Independence<br />

Pubished the 29th of July 2016<br />

on www.burrardstreetjournal.com<br />

Van Full Of Illegals Shows Up To<br />

Vote Clin<strong>to</strong>n At SIX Polling Places,<br />

Still Think Voter Fraud Is A Myth?<br />

Pubished the 5th of November 2016<br />

on www.thelastlineofdefense.org<br />

FBI Agent Suspected in Hillary<br />

Email Leaks Found Dead<br />

in Apparent Murder-Suicide<br />

Pubished the 15th of Oc<strong>to</strong>ber 2016<br />

on www.usa<strong>news</strong>flash.com<br />

Where do <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries originate?<br />

By which sites are they first disseminated?<br />

Page or group which shares both<br />

<strong>fake</strong> s<strong>to</strong>ry and its correction<br />

Fake<br />

Debunk<br />

POST INTERACTIONS<br />

100 41271<br />

Obama Signs Executive Order<br />

Declaring Investigation In<strong>to</strong> Election<br />

Results; Revote Planned For Dec. 19th<br />

Pubished the 12th of December 2016<br />

on www.abc<strong>news</strong>.com.co<br />

Hillary Clin<strong>to</strong>n<br />

for President in 2016<br />

Connecticut<br />

Progressives<br />

Pope Francis Shocks World,<br />

Pope Francis Shocks World,<br />

Endorses Hillary Clin<strong>to</strong>n for<br />

Endorses Donald Trump for<br />

President, Releases Statement<br />

President, Releases Statement<br />

Pubished the 25th of July 2016<br />

Pubished the 26th of September 2016<br />

on www.kypo6.com<br />

on Ending The Fed<br />

Read the damn <strong>news</strong><br />

Obama Signs Executive Order<br />

Banning The Pledge Of Allegiance<br />

In Schools Nationwide<br />

Pubished the 11th of December 2016<br />

on www.abc<strong>news</strong>.com.co<br />

Donald Trump Protester<br />

Speaks Out: I Was Paid $3,500<br />

To Protest Trump's Rally<br />

Pubished the 22th of November 2016<br />

on www.abc<strong>news</strong>.com.co<br />

Criminalize<br />

Conservatism<br />

PolitiFact<br />

Which are the most visible sources<br />

related <strong>to</strong> a <strong>fake</strong> s<strong>to</strong>ry? When and by<br />

whom are they mentioned?<br />

SERVING SUGGESTIONS<br />

This recipe may be used as<br />

one way <strong>to</strong> assess the impact<br />

of attempts <strong>to</strong> debunk <strong>fake</strong><br />

<strong>news</strong> by examining whether<br />

debunking responses <strong>to</strong> <strong>fake</strong><br />

<strong>news</strong> are acknowledged on<br />

the platform that generates<br />

most engagement with <strong>fake</strong><br />

<strong>news</strong>, Facebook, and by the<br />

particular publics which share<br />

and engage with <strong>fake</strong> <strong>news</strong>.<br />

58 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

59


Introduction - Fake <strong>news</strong> are not just “false<br />

<strong>news</strong>”. They are interesting not so much<br />

because their content or form are different<br />

from that of “authentic <strong>news</strong>”, but because they<br />

travel as much as (and sometimes more than)<br />

mainstream <strong>news</strong>. If a blog claims that Pope<br />

Francis endorses Donald Trump, it's just a lie. If<br />

the s<strong>to</strong>ry is picked up by dozens of other blogs,<br />

retransmitted by hundreds of websites, crossposted<br />

over thousands of social media accounts<br />

and read by hundreds of thousands, then it<br />

becomes <strong>fake</strong> <strong>news</strong>.<br />

The following recipes investigate the<br />

circulation of <strong>fake</strong> <strong>news</strong> for two reasons. Firstly,<br />

from a political point of view many have<br />

expressed disappointment that techniques and<br />

tactics commonly used <strong>to</strong> tackle <strong>fake</strong> <strong>news</strong> have<br />

not lived up <strong>to</strong> expectations. Fact-checking and<br />

debunking, in particular, often do not succeed<br />

in preventing the circulation of hoaxes and<br />

rumors. On the contrary: they can inadvertently<br />

contribute <strong>to</strong> making them even more visible<br />

on the web. A better understanding of how <strong>fake</strong><br />

<strong>news</strong> travels online can help <strong>to</strong> inform responses<br />

that are more attuned <strong>to</strong> the phenomena. From<br />

a methodological point of view, as there is<br />

no “on<strong>to</strong>logical” difference between <strong>fake</strong> and<br />

authentic <strong>news</strong>, studying <strong>fake</strong> <strong>news</strong> circulation<br />

can help us understand more about how other<br />

kinds of <strong>news</strong> travels.<br />

This recipe comes in two flavours. Firstly,<br />

we propose a manual variant which can be<br />

executed without the need for any particular <strong>to</strong>ol<br />

or technical knowledge. This variant is easy <strong>to</strong><br />

execute but also time consuming. It is based on<br />

a search for web pages referring <strong>to</strong> <strong>fake</strong> <strong>news</strong><br />

s<strong>to</strong>ries and on the manual identification of the<br />

dates of their publication and the sources that<br />

they cite. Secondly, we propose a semi-au<strong>to</strong>mated<br />

version which allows this approach <strong>to</strong> be scaled<br />

<strong>to</strong> more pages, but demands more technical skills<br />

and may require more manual verification.<br />

60 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

61


CHAPTER 2 → RECIPE 1<br />

START<br />

open Chrome browser<br />

in incogni<strong>to</strong> mode<br />

WHERE DO FAKE<br />

NEWS ORIGINATE?<br />

BY WHAT SITES ARE THEY<br />

FIRST RETRANSMITTED?<br />

query<br />

[ Pope AND endors* AND (Trump OR Clin<strong>to</strong>n)<br />

AND NOT (hoax OR “<strong>fake</strong> <strong>news</strong>” OR lie OR debunk) ]<br />

Google Web Search<br />

compile a list with<br />

the <strong>to</strong>p resulting URLs<br />

a<br />

input URLs in a spreadsheet<br />

> The page advocating<br />

or debunking the <strong>fake</strong><br />

<strong>news</strong> item<br />

> Rank in the search<br />

engine results<br />

> the date of publication<br />

of each of the page<br />

BEFORE STARTING<br />

For this recipe you will need <strong>to</strong> choose a <strong>fake</strong><br />

<strong>news</strong> s<strong>to</strong>ry whose circulation you’d like <strong>to</strong><br />

trace. The more distinctive your s<strong>to</strong>ry is, the<br />

easier it will be <strong>to</strong> follow its circulation. In<br />

the example, we focus on the “Pope Endorses<br />

Trump” s<strong>to</strong>ry that was widely circulated<br />

around the US Elections.<br />

record all sources cited in the appearances<br />

of your s<strong>to</strong>ry, noting down:<br />

+<br />

+<br />

+<br />

extract network of instances<br />

and references with<br />

Table2Net<br />

> if they are cited as evidence<br />

or counter-evidence<br />

> if they are cited through<br />

hyperlink, textual reference<br />

or copying/pasting of its<br />

content<br />

Make sure you visit both the results of<br />

your initial search and all the sources<br />

cited by those results<br />

import data in<br />

Gephi<br />

visualise<br />

b<br />

HOW DO THE<br />

OCCURRENCES OF THE<br />

“POPE ENDORSES<br />

TRUMP” STORY CITES<br />

EACH OTHER?<br />

Plot URLs on a timeline with<br />

Graph Recipes<br />

visualise<br />

c<br />

WHAT IS THE LIFE OF<br />

THE “POPE ENDORSES<br />

TRUMP” STORY<br />

ACCORDING TO PAGES<br />

IN SEARCH ENGINE<br />

RESULTS?<br />

62 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

63


START<br />

open Chrome browser<br />

in incogni<strong>to</strong> mode<br />

query<br />

[ Pope AND endors* AND (Trump OR Clin<strong>to</strong>n)<br />

AND NOT (hoax OR “<strong>fake</strong> <strong>news</strong>” OR lie OR debunk) ]<br />

Google Web Search<br />

compile a list with<br />

the <strong>to</strong>p resulting URLs<br />

a<br />

input URLs in a spreadsheet<br />

> The page advocating<br />

or debunking the <strong>fake</strong><br />

<strong>news</strong> item<br />

> Rank in the search<br />

engine results<br />

> the date of publication<br />

of each of the page<br />

CHAPTER 2 → RECIPE 1<br />

a IDENTIFY AND ANALYSE THE<br />

OCCURRENCES OF YOUR STORY IN SEARCH<br />

ENGINE RESULTS<br />

Successful <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries always appear on<br />

several web pages. In the first step of this recipe,<br />

you will identify and collect information about these<br />

occurrences.<br />

◊<br />

◊<br />

◊<br />

◊<br />

◊<br />

Identify the occurrences of your s<strong>to</strong>ry by<br />

querying one or more search engines (we used<br />

→ Google Web Search). Since <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries<br />

evolve while circulating, consider keywords that<br />

may capture different variants of the s<strong>to</strong>ry.<br />

Rely on search engines <strong>to</strong> rank results by<br />

relevance and concentrate on the first results<br />

(under the working assumption that they are the<br />

ones that circulated the most).<br />

To avoid "filter bubbles" and personalised<br />

results, consider using a dedicated research<br />

browser. [1]<br />

Be aware that search engines may give more<br />

visibility <strong>to</strong> debunkers than <strong>to</strong> original sources<br />

– and be careful not <strong>to</strong> overestimate the<br />

circulation of debunkers based on this. Also<br />

bear in mind that with this approach you see<br />

the phenomena “through the eyes” of the search<br />

engine that you selected – which will become a<br />

part of your s<strong>to</strong>ry or research.<br />

Record all relevant metadata for each relevant<br />

search result. You can collect as many variables<br />

as you like, but make sure <strong>to</strong> characterise how<br />

the page refers <strong>to</strong> the s<strong>to</strong>ry (e.g. as a reliable<br />

<strong>news</strong> source, or as a problematic claim <strong>to</strong> be<br />

debunked), as well as some indica<strong>to</strong>r of its<br />

“visibility” (e.g. the rank in the search results).<br />

[1]<br />

See instructions<br />

on how <strong>to</strong> set up a<br />

research browser<br />

in this video tu<strong>to</strong>rial:<br />

https://www.<br />

youtube.com/<br />

watch?v=bj65Xr9GkJM<br />

64 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

65


START<br />

open Chrome browser<br />

in incogni<strong>to</strong> mode<br />

query<br />

[ Pope AND endors* AND (Trump OR Clin<strong>to</strong>n)<br />

AND NOT (hoax OR “<strong>fake</strong> <strong>news</strong>” OR lie OR debunk) ]<br />

Google Web Search<br />

compile a list with<br />

the <strong>to</strong>p resulting URLs<br />

record all sources cited in the appearances<br />

of your s<strong>to</strong>ry, noting down:<br />

+<br />

+<br />

+<br />

extract network of instances<br />

and references with<br />

Table2Net<br />

> if they are cited as evidence<br />

or counter-evidence<br />

> if they are cited through<br />

hyperlink, textual reference<br />

or copying/pasting of its<br />

content<br />

Make sure you visit both the results of<br />

your initial search and all the sources<br />

cited by those results<br />

import data in<br />

Gephi<br />

a<br />

input URLs in a spreadsheet<br />

> The page advocating<br />

or debunking the <strong>fake</strong><br />

<strong>news</strong> item<br />

> Rank in the search<br />

engine results<br />

> the date of publication<br />

of each of the page<br />

visualise<br />

b<br />

HOW DO THE<br />

OCCURRENCES OF THE<br />

“POPE ENDORSES<br />

TRUMP” STORY CITES<br />

EACH OTHER?<br />

CHAPTER 2 → RECIPE 1<br />

b. EXTRACT THE NETWORK OF<br />

REFERENCES AROUND YOUR STORIES<br />

Fake <strong>news</strong> is supported or opposed through<br />

a network of references: websites that share<br />

rumours cite other pages <strong>to</strong> support their<br />

claims, while debunking initiatives flag <strong>to</strong>xic<br />

websites or refer <strong>to</strong> sources denying them.<br />

In this step, we will trace the network of<br />

references of a specific <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry.<br />

◊<br />

◊<br />

◊<br />

◊<br />

Record all sources cited in the occurrences<br />

of your s<strong>to</strong>ry. For each, note down if the<br />

source is cited as evidence or counterevidence<br />

and if it is cited through a<br />

hyperlink (e.g. http://snopes.com), a<br />

textual reference (e.g. “the website WTOE<br />

5 News”) or copying and pasting its<br />

content.<br />

Make sure you visit not only all the results<br />

of your initial search, but also all the<br />

sources cited by those results.<br />

Extract the networks of occurrences and<br />

references (you can use → Table2Net).<br />

Visualise the network (using → Gephi, for<br />

instance), applying a force-directed layout;<br />

sizing the nodes according <strong>to</strong> the number<br />

of citations they receive; and colouring the<br />

nodes according <strong>to</strong> how they report the<br />

s<strong>to</strong>ry (advocating or debunking).<br />

66 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

67


HOW DO THE<br />

OCCURRENCES OF<br />

THE “POPE ENDORSES<br />

TRUMP” STORY CITES<br />

EACH OTHER?<br />

europe.<strong>news</strong>week.com/<strong>fake</strong>-...<br />

rednationrising<br />

reddit.com/r/thedonald/co...<br />

trend.network/2016/09/28/...<br />

complex<br />

uk.complex.com/life/2016/...<br />

mercury<strong>news</strong>.com/2016/12/1...<br />

Network of cross-references between<br />

the pages mentioning the “Pope<br />

Endorses Trump” s<strong>to</strong>ry. In the image<br />

the nodes represent the different pages<br />

on which the <strong>fake</strong> s<strong>to</strong>ry appears. The<br />

comparison of the colour of the nodes<br />

(which indicates whether the page cites<br />

or debunks the s<strong>to</strong>ry) and their size<br />

(which indicates the number of citations<br />

received by the page) reveals the greater<br />

visibility of the debunking and neutral<br />

pages compared <strong>to</strong> websites that spread<br />

the <strong>fake</strong> s<strong>to</strong>ry as authentic<br />

NUMBER OF MENTIONS<br />

theglobeandmail.com/<strong>news</strong>/...<br />

thejournal.ie/<strong>fake</strong>-<strong>news</strong>-f...<br />

buzzfeed<br />

endingthefed<br />

thebassbarn.com/forum/12-...<br />

godlikeproductions.com/fo...<br />

cnbc.com/2016/12/30/read-...<br />

uk.businessinsider.com/fa...<br />

nymag.com/selectall/2016/...<br />

quora<br />

theguardian.com/commentis...<br />

enh<br />

nytimes.com/2016/11/13/op...<br />

raptureforums.com/forums/...<br />

snopes<br />

nytimes.com/2016/11/13/op...<br />

niemanlab.org/2016/11/the...<br />

morning<strong>news</strong>usa.com/pope-f...<br />

latimes<br />

catholic<strong>news</strong><br />

w<strong>to</strong>e5<br />

dinarvets.com/forums/inde...<br />

erumors<br />

buzzfeed_strangest<br />

sco<strong>news</strong>.co.uk/<strong>news</strong>/52118/...<br />

independent.co.uk/<strong>news</strong>/pe...<br />

<strong>news</strong>tatesman.com/world/20...<br />

hoaxslayer<br />

truthorfiction.com/pope-f...<br />

hackinghumanity.com/uncat...<br />

earnthenecklace.com/fact-...<br />

justplainpolitics.com/sho...<br />

thats<strong>fake</strong>.com/pope-franci...<br />

msn<br />

groups.google.com/forum/#...<br />

linkedin.com/pulse/pope-f...<br />

gv<strong>news</strong>.com/opinion/factch...<br />

<strong>news</strong>breakshere.com/pope-f...<br />

wp<br />

commonwealmagazine.org/bl...<br />

southend<strong>news</strong>network.com/n...<br />

heavy.com/<strong>news</strong>/2016/11/fa...<br />

buzzfeed_partisan<br />

fox<strong>news</strong>.com/world/2016/12...<br />

guardian<br />

vox.com/new-money/2016/11...<br />

buzzfeed_macedonia<br />

dailydot.com/unclick/face...<br />

1<br />

18<br />

reuters.com/article/us-us...<br />

Fake News<br />

Debunk<br />

Neutral<br />

slate.com/blogs/future_te...<br />

w<strong>to</strong>e5_about<br />

factcheck.org/2016/10/did...<br />

ap<strong>news</strong>.com/86073aedf0c245...<br />

wgal.com/article/did-the-...<br />

cbs<strong>news</strong>.com/<strong>news</strong>/pope-fra...<br />

wgal<br />

msn.com/en-us/<strong>news</strong>/crime/...<br />

kypo6<br />

youtube.com/watch?v=8pzfw...<br />

nbc<strong>news</strong>.com/<strong>news</strong>/us-<strong>news</strong>/...<br />

cathola.com/2017/02/19/po...<br />

guardian_pope<br />

buzzfeed_viral<strong>fake</strong><br />

thedailybeast.com/article...<br />

bigs<strong>to</strong>ry.ap.org/article/5...<br />

68 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

69


START<br />

open Chrome browser<br />

in incogni<strong>to</strong> mode<br />

query<br />

[ Pope AND endors* AND (Trump OR Clin<strong>to</strong>n)<br />

AND NOT (hoax OR “<strong>fake</strong> <strong>news</strong>” OR lie OR debunk) ]<br />

Google Web Search<br />

compile a list with<br />

the <strong>to</strong>p resulting URLs<br />

record all sources cited in the appearances<br />

of your s<strong>to</strong>ry, noting down:<br />

+<br />

+<br />

+<br />

> if they are cited as evidence<br />

or counter-evidence<br />

> if they are cited through<br />

hyperlink, textual reference<br />

or copying/pasting of its<br />

content<br />

a<br />

input URLs in a spreadsheet<br />

> The page advocating<br />

or debunking the <strong>fake</strong><br />

<strong>news</strong> item<br />

> Rank in the search<br />

engine results<br />

> the date of publication<br />

of each of the page<br />

CHAPTER 2 → RECIPE 1<br />

c. VISUALIZE THE FAKE NEWS INSTANCES<br />

OVER TIME<br />

The network extracted in the previous step<br />

can help you understanding not only who cited<br />

whom, but also how and in which direction your<br />

<strong>fake</strong> <strong>news</strong> travelled. To reveal the circulation use<br />

the dates that you collected in the first step of this<br />

recipe.<br />

◊<br />

Arrange the network of instances extracted<br />

in the previous step chronologically. You can<br />

use different visual styles <strong>to</strong> represent the<br />

different kind of citations (<strong>to</strong> do so, we used<br />

a cus<strong>to</strong>m script of → Graph Recipes <strong>to</strong>ols.<br />

Make sure you visit both the results of<br />

your initial search and all the sources<br />

cited by those results<br />

extract network of instances<br />

and references with<br />

Table2Net<br />

import data in<br />

Gephi<br />

visualise<br />

b<br />

HOW DO THE<br />

OCCURRENCES OF THE<br />

“POPE ENDORSES<br />

TRUMP” STORY CITES<br />

EACH OTHER?<br />

Plot URLs on a timeline with<br />

Graph Recipes<br />

visualise<br />

c<br />

WHAT IS THE LIFE OF<br />

THE “POPE ENDORSES<br />

TRUMP” STORY<br />

ACCORDING TO PAGES<br />

IN SEARCH ENGINE<br />

RESULTS?<br />

70 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

71


WHAT IS THE LIFE OF<br />

THE “POPE ENDORSES<br />

TRUMP” STORY<br />

ACCORDING TO PAGES<br />

IN SEARCH ENGINE<br />

RESULTS?<br />

Chronological network of the crossreferences<br />

between the pages<br />

mentioning the “Pope Endorses<br />

Trump” s<strong>to</strong>ry. In this image it is not<br />

the nodes, but their connections that<br />

are highlighted: their colour indicates<br />

whether a source is cited as evidence<br />

or counter-evidence of the s<strong>to</strong>ry and<br />

the style the type of citation. The<br />

high presence of orange dotted lines<br />

(counter-evidence textual citations)<br />

shows how debunking initiatives tend <strong>to</strong><br />

mention original sources but not link <strong>to</strong><br />

them. This technique is used <strong>to</strong> flag <strong>fake</strong><br />

websites without increasing their online<br />

visibility by explicitly linking <strong>to</strong> them.<br />

Buzzfeed<br />

Theguardian<br />

Truthorfiction<br />

Washing<strong>to</strong>npost<br />

Thedailybeast<br />

Commonwealmagazine<br />

Catholic-<strong>news</strong><br />

W<strong>to</strong>e5<strong>news</strong><br />

Every<strong>news</strong>here<br />

Kypo6<br />

Thejournal<br />

Vox<br />

Dinarvets<br />

Godlikeproductions<br />

Truthorfiction<br />

Linkedin<br />

Groups<br />

Earnthenecklace<br />

Justplainpolitics<br />

Newsbreakshere<br />

Thebassbarn<br />

Snopes<br />

Nymag<br />

Niemanlab<br />

Nytimes<br />

Nytimes<br />

Uk<br />

Morning<strong>news</strong>usa<br />

Theguardian<br />

Hoax-slayer<br />

Buzzfeed<br />

Thats<strong>fake</strong><br />

Gv<strong>news</strong> Independent<br />

Newstatesman<br />

Sco<strong>news</strong><br />

Raptureforums<br />

Msn<br />

Youtube<br />

Trend<br />

Reddit<br />

Twitter<br />

Europe<br />

Endingthefed<br />

Quora<br />

Latimes<br />

Wgal<br />

Buzzfeed<br />

Heavy<br />

Msn<br />

Hackinghumanity<br />

Cnbc<br />

Factcheck<br />

Slate<br />

CITED<br />

SOURCE<br />

source in original list<br />

source identified through<br />

the analysis<br />

SOURCE<br />

Buzzfeed<br />

Complex<br />

Buzzfeed<br />

Dailydot<br />

Uk<br />

Bigs<strong>to</strong>ry<br />

TYPE OF CITATION<br />

counter-evidence<br />

cited as hyperlink<br />

counter-evidence<br />

cited as text<br />

evidence cited<br />

as hyperlink<br />

evidence cited<br />

as text<br />

evidence cited<br />

as copy-paste<br />

No sharing of the s<strong>to</strong>ry URL from March <strong>to</strong> July<br />

Nbc<strong>news</strong><br />

Ap<strong>news</strong><br />

Fox<strong>news</strong><br />

Theglobeandmail<br />

Cbs<strong>news</strong><br />

Southend<strong>news</strong>network<br />

Mercury<strong>news</strong><br />

Reuters<br />

Theguardian<br />

Cathola<br />

GOOGLE RANK<br />

1 100<br />

WITHOUT DATE<br />

JANUARY<br />

FEBRUARY MARCH APRIL MAY JUNE JULY AUGUST SEPTEMBER OCTOBER NOVEMBER DECEMBER JANUARY FEBRUARY<br />

2016<br />

2017<br />

72 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

73


mercury<strong>news</strong>.com/2016/12/1...<br />

endingthefed<br />

commonwealmagazine.org/bl...<br />

12-...<br />

godlikeproductions.com/fo...<br />

bc.com/2016/12/30/read-...<br />

msn<br />

nsider.com/fa...<br />

groups.google.com/forum/#...<br />

quora<br />

catholic<strong>news</strong><br />

latimes<br />

southend<strong>news</strong>network.com/n...<br />

heavy.com/<strong>news</strong>/2016/11/fa...<br />

buzzfeed_partisan<br />

CHAPTER 2 → RECIPE 1<br />

Buzzfeed<br />

Theguardian<br />

thenecklace.com/fact-...<br />

Truthorfiction<br />

pope-f...<br />

tplainpolitics.com/sho...<br />

Washing<strong>to</strong>npost<br />

Thedailybeast<br />

Commonwealmagazine<br />

linkedin.com/pulse/pope-f...<br />

w<strong>to</strong>e5<br />

gv<strong>news</strong>.com/opinion/factch...<br />

<strong>news</strong>breakshere.com/pope-f...<br />

Catholic-<strong>news</strong><br />

Every<strong>news</strong>here<br />

Truthorfiction<br />

W<strong>to</strong>e5<strong>news</strong><br />

dinarvets.com/forums/inde...<br />

st<br />

sco<strong>news</strong>.co.uk/<strong>news</strong>/52118/...<br />

independent.co.uk/<strong>news</strong>/pe...<br />

<strong>news</strong>tatesman.com/world/20...<br />

oaxslayer<br />

truthorfiction.com/pope-f...<br />

thats<strong>fake</strong>.com/pope-franci...<br />

wp<br />

erumors<br />

Kypo6<br />

Raptureforums<br />

Dinarvets<br />

Theguardian<br />

vox.com/new-money/2016/11...<br />

Hoax-slayer<br />

Buzzfeed<br />

Thats<strong>fake</strong><br />

Gv<strong>news</strong> Independent<br />

Newstatesman<br />

Sco<strong>news</strong><br />

Cnbc<br />

Endingthefed<br />

Msn<br />

Youtube<br />

Reddit<br />

Twitter<br />

buzzfeed_macedonia<br />

Europe<br />

Buzzfeed<br />

Factcheck<br />

Godlikeproductions<br />

Groups<br />

Earnthenecklace<br />

Justplainpolitics<br />

Newsbreakshere<br />

Thebassbarn<br />

Snopes<br />

Nymag<br />

Niemanlab<br />

Nytimes<br />

Nytimes<br />

Uk<br />

Morning<strong>news</strong>usa<br />

Trend<br />

Quora<br />

Linkedin<br />

Msn<br />

fox<strong>news</strong>.com/world/2016/12...<br />

Slate<br />

Vox<br />

guardian<br />

Heavy<br />

Thejournal<br />

Latimes<br />

dailydot.com/unclick/face...<br />

Wgal<br />

Hackinghumanity<br />

SERVING SUGGESTIONS<br />

This recipe may be used <strong>to</strong> repurpose data<br />

obtained through search engine results in<br />

order <strong>to</strong> identify and follow the different<br />

occurrences of a given <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry<br />

as it is cited and referenced by different<br />

online sources, as well as <strong>to</strong> retrace its<br />

circulation over time. You will also be able<br />

<strong>to</strong> see when the debunking activities <strong>to</strong>ok<br />

place, who promoted them and what effect<br />

this had on the circulation of <strong>fake</strong> <strong>news</strong><br />

s<strong>to</strong>ries and web pages.<br />

nbc<strong>news</strong>.com/<strong>news</strong>/us-<strong>news</strong>/...<br />

Buzzfeed Dailydot<br />

No sharing of the s<strong>to</strong>ry URL from March <strong>to</strong> July<br />

cathola.com/2017/02/19/po...<br />

guardian_pope<br />

Complex<br />

Buzzfeed<br />

Uk<br />

Bigs<strong>to</strong>ry<br />

Nbc<strong>news</strong><br />

Ap<strong>news</strong><br />

Fox<strong>news</strong><br />

Theglobeandmail<br />

Cbs<strong>news</strong><br />

om/article/did-the-...<br />

wgal<br />

en-us/<strong>news</strong>/crime/...<br />

6<br />

WITHOUT DATE<br />

2016<br />

JANUARY<br />

outube.com/watch?v=8pzfw...<br />

buzzfeed_viral<strong>fake</strong><br />

bigs<strong>to</strong>ry.ap.org/article/5...<br />

thedailybeast.com/article...<br />

Southend<strong>news</strong>network<br />

Mercury<strong>news</strong><br />

FEBRUARY MARCH APRIL MAY JUNE JULY AUGUST SEPTEMBER OCTOBER NOVEMBER DECEMBER JANUARY FEBRUARY<br />

2017<br />

Reuters<br />

Theguardian<br />

Cathola<br />

74 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

75


START<br />

CHAPTER 2 → RECIPE 2<br />

WHICH ARE THE MOST<br />

VISIBLE SOURCES<br />

RELATED TO A FAKE STORY?<br />

WHEN AND BY WHOM<br />

ARE THEY MENTIONED?<br />

find a query identifying<br />

your <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry<br />

identify the occurrences<br />

of your s<strong>to</strong>ry in<br />

Radarly<br />

export all data<br />

> Full text of each<br />

occurrence<br />

> Date of publication<br />

> Metadata<br />

compile a list of potential sources<br />

crawl the websites with<br />

Hyphe<br />

websites collected<br />

and curated<br />

by the project<br />

Décodex by Le Monde<br />

BEFORE STARTING<br />

This recipe enables a scaling up of the approach<br />

presented in the previous recipe, but requires a<br />

bit more technical knowledge, as well as some<br />

bigger datasets. In particular, you will need <strong>to</strong><br />

have access <strong>to</strong>:<br />

◊<br />

◊<br />

A web archive (we used → Radarly by<br />

Linkfluence).<br />

A list of all the possible web sources in<br />

which your chosen <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry may<br />

have appeared (we used the list curated by<br />

→ Le Monde Décodex).<br />

b<br />

WHICH ARE THE<br />

SOURCES CITED IN THE<br />

OCCURRENCES OF THE<br />

FAKE NEWS STORY?<br />

search mentions<br />

of potential sources<br />

in the instances of your s<strong>to</strong>ry<br />

Csv Rinse Repeat<br />

export a graph file<br />

Project the instances<br />

of your s<strong>to</strong>ry<br />

on the basemap<br />

produce time-sliced<br />

networks using<br />

Graph Recipes<br />

export a graph file<br />

it represents<br />

the general<br />

connectivity<br />

among potential<br />

sources<br />

create a basemap<br />

network with<br />

76 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

77<br />

visualise<br />

c<br />

HOW MANY OCCURRENCES<br />

OF THE FAKE NEWS ARE<br />

PUBLISHED IN EACH PERIOD<br />

OF TIME AND WHAT<br />

SOURCES DO THEY CITE?<br />

visualise<br />

Gephi<br />

a<br />

visualise<br />

WHAT ARE THE<br />

MAIN SPHERES<br />

IN THE FRENCH<br />

MEDIA SYSTEM?


START<br />

compile a list of potential sources<br />

crawl the websites with<br />

Hyphe<br />

export a graph file<br />

it represents<br />

the general<br />

connectivity<br />

among potential<br />

sources<br />

create a basemap<br />

network with<br />

Gephi<br />

websites collected<br />

and curated<br />

by the project<br />

Décodex by Le Monde<br />

visualise<br />

CHAPTER 2 → RECIPE 2<br />

a DEFINE A BASE MAP OF NEWS<br />

PROVIDERS<br />

Identify (or compile) a list of all the possible<br />

Web sources in which you think your <strong>fake</strong><br />

<strong>news</strong> item might have appeared (try <strong>to</strong> be as<br />

exhaustive as possible). You can use one of the<br />

many lists of <strong>fake</strong> <strong>news</strong> websites maintained by<br />

debunking initiatives and combine it with a list<br />

of mainstream media outlets.<br />

◊<br />

◊<br />

◊<br />

Identify how the sources in your list are<br />

associated with each other through web<br />

crawling and hyperlink analysis. We used<br />

→ Hyphe for this.<br />

Visualise the resulting network and apply a<br />

force-directed layout algorithm <strong>to</strong> identify<br />

clusters of sources. You can use → Gephi<br />

for this task.<br />

Manually highlight and name the clusters.<br />

a<br />

WHAT ARE THE<br />

MAIN SPHERES<br />

IN THE FRENCH<br />

MEDIA SYSTEM?<br />

78 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

79


WHAT ARE THE MAIN<br />

SPHERES IN THE FRENCH<br />

MEDIA SYSTEM?<br />

Network analysis of the media<br />

sources active in French public<br />

debate. The image shows the <strong>news</strong><br />

sources listed by the Décodex project<br />

by Le Monde and the hyperlinks<br />

connecting them. A force-directed<br />

layout has been applied <strong>to</strong> reveal the<br />

main clusters of websites and their<br />

respective associations and positions.<br />

LOCAL MEDIA<br />

EAST FRANCE MEDIA<br />

L'Est républicain<br />

Le Progrès<br />

Version Femina<br />

L'Alsace-Le Pays<br />

GROUPE CENTRE-FRANCE<br />

L'Yonne républicaine<br />

Canal+<br />

Wanted Pedo<br />

GOSSIP MEDIA<br />

Al Kanz<br />

Paris Match<br />

Dailymotion<br />

La République du Centre<br />

Facebook<br />

SATIRICAL WEBSITES<br />

Twitter<br />

SOCIAL MEDIA AND OTHERS<br />

YouTube<br />

Snapchat<br />

Instagram<br />

Wikihow<br />

Tumblr<br />

Vice<br />

Mashable<br />

Reddit<br />

The Washing<strong>to</strong>n Post<br />

Bloomberg<br />

The Guardian The New York Times<br />

BuzzFeed<br />

The Independent<br />

INTERNATIONAL MEDIA<br />

US MEDIA<br />

UK MEDIA<br />

Change.org<br />

The Daily Telegraph<br />

Ouest-France<br />

Europe 1<br />

France Inter<br />

RTL<br />

Les Echos<br />

Russia Today<br />

Conscience du peuple<br />

RUSSIAN MEDIA<br />

Mes Opinions<br />

AFP<br />

Debunkers de hoax<br />

Wikipédia<br />

Les Chroniques de Rorschach<br />

Link between<br />

providers<br />

BFMTV<br />

Le Figaro<br />

Slate<br />

L'Express<br />

Le Monde<br />

Lablogalupus<br />

alalumieredunouveaumonde<br />

Blog du"Plan C", par Etienne Chouard S<strong>to</strong>p Mensonges<br />

Franceinfo<br />

Libération<br />

Le6emetiroir<br />

EXTREME-RIGHT Le Parisien-Aujourd’hui MEDIAen France<br />

Le Point<br />

Huffing<strong>to</strong>n Post<br />

Arrêt sur images Les Crises<br />

Reopen911<br />

Révolution Vibra<strong>to</strong>ire<br />

Novopress<br />

Marianne<br />

Les Mou<strong>to</strong>ns enragés<br />

Contre-Info<br />

L'Obs<br />

Cercle des Volontaires<br />

IlFat<strong>to</strong>Quotidiano.fr<br />

Les Brins d'Herbes<br />

Agoravox<br />

Egalité et Réconciliation<br />

Provider<br />

NUMBER OF CONNECTIONS<br />

MAINSTREAM FRENCH MEDIA<br />

CONSPIRACY THEORISTS<br />

80 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

81


START<br />

find a query identifying<br />

your <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry<br />

identify the occurrences<br />

of your s<strong>to</strong>ry in<br />

Radarly<br />

export all data<br />

> Full text of each<br />

occurrence<br />

> Date of publication<br />

> Metadata<br />

WHICH ARE THE<br />

SOURCES CITED IN THE<br />

OCCURRENCES OF THE<br />

FAKE NEWS STORY?<br />

search mentions<br />

of potential sources<br />

in the instances of your s<strong>to</strong>ry<br />

b<br />

Csv Rinse Repeat<br />

export a graph file<br />

visualise<br />

Project the instances<br />

of your s<strong>to</strong>ry<br />

on the basemap<br />

compile a list of potential sources<br />

crawl the websites with<br />

Hyphe<br />

export a graph file<br />

it represents<br />

the general<br />

connectivity<br />

among potential<br />

sources<br />

create a basemap<br />

network with<br />

Gephi<br />

websites collected<br />

and curated<br />

by the project<br />

Décodex by Le Monde<br />

CHAPTER 2 → RECIPE 2<br />

b. HIGHLIGHT THE OCCURRENCES OF YOUR<br />

STORY ON THE BASE MAP<br />

In this step we will explore how <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ries<br />

are associated with different sources. This is<br />

interesting, as while a <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry might –<br />

for example – start out its life as a piece of satire,<br />

as it travels it can become more prominently<br />

associated with non-satirical sources. Here we<br />

will identify which of the sources in the base map<br />

of the French media system are mentioned in the<br />

pages in which your <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry occurs.<br />

◊<br />

◊<br />

◊<br />

◊<br />

Create a query that identifies the <strong>fake</strong> <strong>news</strong><br />

s<strong>to</strong>ry that you want <strong>to</strong> trace. Use keywords<br />

specifically associated with your s<strong>to</strong>ry and<br />

the s<strong>to</strong>p-words <strong>to</strong> exclude "false positives".<br />

Identify the occurrences of your s<strong>to</strong>ry,<br />

running your query on the archive that<br />

you have chosen <strong>to</strong> use. For each of the<br />

results, collect the full text and the date of<br />

publication.<br />

Detect mentions of the sources of your base<br />

map, searching for the URLs as well as for<br />

the names of your sources (e.g. sputnik<strong>news</strong>.<br />

com, Sputnik). In our example we used a<br />

cus<strong>to</strong>m script for → CSV Rinse Repeat.<br />

Project the occurrences of your s<strong>to</strong>ry on<strong>to</strong><br />

your base map, by merging them with the<br />

source network and connecting each of them<br />

<strong>to</strong> the sources that they mention. While<br />

keeping the source-nodes fixed, apply a force<br />

directed spatialisation algorithm (you can<br />

do this using → Gephi) <strong>to</strong> move the nodes<br />

representing the <strong>fake</strong> s<strong>to</strong>ry occurrences<br />

closer <strong>to</strong> clusters of the base map that they<br />

cite the most.<br />

82 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

83


WHICH ARE THE<br />

SOURCES CITED IN THE<br />

OCCURRENCES OF THE<br />

FAKE NEWS STORY?<br />

Projection of the <strong>fake</strong> <strong>news</strong><br />

occurrences on the network of media<br />

sources. In this image, the occurrences<br />

of the <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry are positioned on<br />

the base map displayed by the previous<br />

network according <strong>to</strong> the sources they<br />

cite. The tendency <strong>to</strong> refer <strong>to</strong> social<br />

media is visible as well as the relevance<br />

of “Russia Today” and “Sputnik<br />

International” in this particular s<strong>to</strong>ry.<br />

Version Femina<br />

L'Yonne républicaine<br />

La République du Centre<br />

Facebook<br />

Snapchat<br />

Instagram<br />

Wikihow<br />

Tumblr<br />

Vice<br />

Mashable<br />

Reddit<br />

Canal+<br />

Twitter<br />

The Washing<strong>to</strong>n Post<br />

L'Est républicain<br />

Le Progrès<br />

L'Alsace-Le Pays<br />

Wanted Pedo<br />

Al Kanz<br />

Paris Match<br />

Dailymotion<br />

YouTube<br />

Bloomberg<br />

The Guardian The New York Times<br />

BuzzFeed<br />

The Independent<br />

Change.org<br />

The Daily Telegraph<br />

Ouest-France Europe 1<br />

Mes Opinions AFP<br />

France Inter<br />

Les Echos<br />

RTL<br />

Wikipédia<br />

Debunkers de hoax<br />

Les Chroniques de Rorschach<br />

Russia Today<br />

Conscience du peuple<br />

TRACKER POPULARITY<br />

Link between provider<br />

and <strong>fake</strong> <strong>news</strong> article<br />

Link between<br />

providers<br />

Provider<br />

BFMTV<br />

Le Figaro<br />

Slate<br />

L'Express<br />

Le Monde<br />

Lablogalupus<br />

alalumieredunouveaumonde<br />

Blog du"Plan C", par Etienne Chouard S<strong>to</strong>p Mensonges<br />

Franceinfo<br />

Libération<br />

Le6emetiroir<br />

Huffing<strong>to</strong>n Post<br />

Le Parisien-Aujourd’hui en France<br />

Le Point<br />

Reopen911<br />

Révolution Vibra<strong>to</strong>ire<br />

Arrêt sur images<br />

Les Crises<br />

Novopress<br />

Marianne<br />

Les Mou<strong>to</strong>ns enragés<br />

Contre-Info<br />

L'Obs<br />

Cercle des Volontaires<br />

IlFat<strong>to</strong>Quotidiano.fr<br />

Les Brins d'Herbes<br />

Agoravox<br />

Egalité et Réconciliation<br />

Fake News article<br />

NUMBER OF CONNECTIONS<br />

84 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

85


START<br />

find a query identifying<br />

your <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry<br />

identify the occurrences<br />

of your s<strong>to</strong>ry in<br />

Radarly<br />

export all data<br />

> Full text of each<br />

occurrence<br />

> Date of publication<br />

> Metadata<br />

search mentions<br />

of potential sources<br />

in the instances of your s<strong>to</strong>ry<br />

Csv Rinse Repeat<br />

export a graph file<br />

Project the instances<br />

of your s<strong>to</strong>ry<br />

on the basemap<br />

compile a list of potential sources<br />

crawl the websites with<br />

Hyphe<br />

export a graph file<br />

it represents<br />

the general<br />

connectivity<br />

among potential<br />

sources<br />

create a basemap<br />

network with<br />

Gephi<br />

websites collected<br />

and curated<br />

by the project<br />

Décodex by Le Monde<br />

CHAPTER 2 → RECIPE 2<br />

c. VISUALISE THE SPREAD OF YOUR FAKE<br />

STORY ON THE BASE MAP<br />

In this step, you will reveal how the reference<br />

patterns identified in the previous step evolve<br />

over time.<br />

◊<br />

◊<br />

◊<br />

Slice your network of occurrences and<br />

references by month, by week or by day,<br />

according <strong>to</strong> the speed of circulation of<br />

your s<strong>to</strong>ry. In this example we grouped<br />

<strong>news</strong> by month and then zoomed in on<br />

a four day window <strong>to</strong> explore the most<br />

important period of circulation.<br />

While keeping the source base map stable,<br />

visualise the different temporal slices of<br />

<strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry occurrences.<br />

In order <strong>to</strong> make the changes and patterns<br />

more legible, you can represent the <strong>fake</strong><br />

s<strong>to</strong>ry occurrences not as single nodes, but<br />

through a density heatmap (the example<br />

has been produced using a → Graph<br />

Recipes).<br />

c<br />

HOW MANY OCCURRENCES<br />

OF THE FAKE NEWS ARE<br />

PUBLISHED IN EACH PERIOD<br />

OF TIME AND WHAT<br />

SOURCES DO THEY CITE?<br />

produce time-sliced<br />

networks using<br />

Graph Recipes<br />

visualise<br />

86 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

87


HOW MANY<br />

OCCURRENCES OF THE<br />

FAKE NEWS STORY<br />

ARE PUBLISHED IN<br />

EACH PERIOD AND<br />

WHAT SOURCES DO<br />

THEY CITE?<br />

14 MAR 2016 - 11 APR 2016 28 MAR 2016 - 25 APR 2016 11 APR 2016 - 9 MAY 2016 25 APR 2016 - 23 MAY 2016 9 MAY 2016 - 6 JUN 2016 23 MAY 2016 - 20 JUN 2016<br />

Temporal evolution of the <strong>fake</strong> <strong>news</strong><br />

s<strong>to</strong>ry in the whole observed period.<br />

In this image, the occurrences of the<br />

<strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry are divided in slices of<br />

4 weeks (with an overlap of two weeks)<br />

and represented as a density heat map<br />

rather than as individual points. Though<br />

mentions of the s<strong>to</strong>ry have been present<br />

for more than one year, its circulation<br />

appears <strong>to</strong> spike up in February 2017,<br />

when a new strand of the <strong>fake</strong> s<strong>to</strong>ry<br />

is published by the Russian website<br />

“Sputnik International”.<br />

6 JUN 2016 - 4 JUL 2016 20 JUN 2016 - 18 JUL 2016 4 JUL 2016 - 1 AUG 2016 18 JUL 2016 - 15 AUG 2016 1 AUG 2016 - 29 JUL 2016 15 AUG 2017 - 12 SEP 2016<br />

BASEMAP<br />

Social Media<br />

and Others<br />

Local Media International<br />

Media<br />

29 AUG 2016- 26 SEP 2016 12 SEP 2016 - 10 OCT 2016 26 SEP 2016 - 24 OCT 2016 10 OCT 2016 - 7 NOV 2017 24 OCT 2017 - 21 NOV 2017 7 NOV 2016 - 5 MAY 2016<br />

Extreme<br />

Right Media<br />

Conspiracy<br />

Theorists<br />

Mainstream<br />

French Media<br />

Fake <strong>news</strong> density<br />

21 NOV 2017 - 19 DEC 2016 5 DEC 2016 - 2 JAN 2017<br />

19 DEC 2016 - 16 JAN 2017 2 JAN 2017 - 30 JAN 2017 16 JAN 2017 - 13 FEB 2017 30 JAN 2017 - 27 FEB 2017<br />

88 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

89


HOW MANY<br />

OCCURRENCES OF THE<br />

FAKE NEWS STORY<br />

ARE PUBLISHED IN<br />

FEBRUARY 2017 AND<br />

WHAT SOURCES DO<br />

THEY CITE?<br />

Temporal evolution of the <strong>fake</strong><br />

<strong>news</strong> s<strong>to</strong>ry in the February 2017.<br />

This image represents a ‘temporal<br />

zoom’ of the previous one. Here the<br />

occurrences of the <strong>fake</strong> <strong>news</strong> s<strong>to</strong>ry<br />

are break up in slices of 4 days (with<br />

an overlap of two days).<br />

14 MAR 2016 - 11 APR 2016 28 MAR 2016 - 25 APR 2016 11 APR 2016 - 9 MAY 2016 25 APR 2016 - 23 MAY 2016 9 MAY 2016 - 6 JUN 2016 23 MAY 2016 - 20 JUN 2016<br />

1 FEB 2017 - 5 FEB 2017 3 FEB 2017 - 7 FEB 2017 5 FEB 2017 - 9 FEB 2017 7 FEB 2017 - 11 FEB 2017 9 FEB 2017 - 13 FEB 2017<br />

6 JUN 2016 - 4 JUL 2016 20 JUN 2016 - 18 JUL 2016 4 JUL 2016 - 1 AUG 2016 18 JUL 2016 - 15 AUG 2016 1 AUG 2016 - 29 JUL 2016 15 AUG 2017 - 12 SEP 2016<br />

11 FEB 2017 - 15 FEB 2017 13 FEB 2017 - 17 FEB 2017 15 FEB 2017 -19 FEB 2017 17 FEB 2017 - 21 FEB 2017 19 FEB 2017 - 23 FEB 2017<br />

BASEMAP<br />

Social Media<br />

and Others<br />

Local Media International<br />

Media<br />

29 AUG 2016- 26 SEP 2016 12 SEP 2016 - 10 OCT 2016 26 SEP 2016 - 24 OCT 2016 10 OCT 2016 - 7 NOV 2017 24 OCT 2017 - 21 NOV 2017 7 NOV 2016 - 5 MAY 2016<br />

Extreme<br />

Right Media<br />

Conspiracy<br />

Theorists<br />

Mainstream<br />

French Media<br />

21 FEB 2017 - 25 FEB 2017 23 FEB 2017 - 27 FEB 2017 25 FEB 2017 - 1 MAR 2017 27 FEB 2017 - 3 MAR 2017<br />

Fake <strong>news</strong> density<br />

21 NOV 2017 - 19 DEC 2016 5 DEC 2016 - 2 JAN 2017<br />

19 DEC 2016 - 16 JAN 2017 2 JAN 2017 - 30 JAN 2017 16 JAN 2017 - 13 FEB 2017 30 JAN 2017 - 27 FEB 2017<br />

90 A FIELD GUIDE TO FAKE NEWS CHAPTER 2: TRACING THE CIRCULATION OF FAKE NEWS ON THE WEB<br />

91


GROUPE CENTRE-FRANCE<br />

L'Yonne républicaine<br />

GOSSIP MEDIA<br />

La L'Yonne République républicaine du Centre<br />

SATIRICAL WEBSITES<br />

CHAPTER 2 → RECIPE 2<br />

Snapchat<br />

Facebook Instagram<br />

Tumblr<br />

Facebook<br />

Vice<br />

US MEDIA<br />

20 JUN 2016 - 18 JUL 2016 4 JUL 2016 - 1 AUG 2016 18 JUL 2016 - 15 AUG 2016 1 AUG 2016 - 29 JUL 2016 15 AUG 2017 - 12 SEP 2016<br />

Tumblr<br />

Vice<br />

Mashable<br />

Reddit<br />

Version Femina<br />

Mashable<br />

Reddit<br />

Version Femina<br />

Canal+<br />

The Washing<strong>to</strong>n Post<br />

OCAL MEDIA<br />

Twitter<br />

Canal+<br />

The Washing<strong>to</strong>n Post<br />

Twitter<br />

Bloomberg<br />

CE MEDIA<br />

L'Est républicain<br />

The Bloomberg Guardian The New York Times<br />

Wanted Pedo<br />

L'Est républicain<br />

YouTube The Guardian The New York Times<br />

Wanted Pedo<br />

Al Kanz<br />

YouTube<br />

BuzzFeed<br />

Le Progrès<br />

Al Kanz<br />

BuzzFeed<br />

Le Progrès<br />

Paris Match<br />

The Independent<br />

L'Alsace-Le Pays<br />

Paris Match<br />

Dailymotion<br />

The Independent<br />

L'Alsace-Le Pays<br />

Dailymotion<br />

UK MEDIA<br />

1 FEB 2017 - 5 FEB 2017 3 FEB 2017 - 7 FEB 2017 5 FEB 2017 - 9 FEB 2017 7 FEB 2017 - 11 FEB 2017 9 FEB 2017 - 13 FEB 2017<br />

Change.org<br />

The Daily Telegraph<br />

Contre-Info<br />

L'Obs<br />

This recipe may be used <strong>to</strong><br />

Agoravox<br />

Egalité et Réconciliation<br />

identify which websites appear<br />

Egalité et Réconciliation<br />

most often when a <strong>fake</strong> <strong>news</strong><br />

s<strong>to</strong>ry is mentioned in different<br />

settings. These are not necessarily<br />

the original sources of the <strong>fake</strong><br />

<strong>news</strong>, but are often the most<br />

influential media outlets that<br />

contribute <strong>to</strong> its circulation<br />

(whether as a rumour or as a<br />

debunked s<strong>to</strong>ry).<br />

SOCIAL MEDIA AND OTHERS<br />

Instagram<br />

Wikihow<br />

INTERNATIONAL MEDIA<br />

Change.org<br />

The Daily Telegraph<br />

France Inter<br />

Russia Today<br />

Ouest-France Europe 1<br />

Les Echos<br />

Conscience du peuple<br />

France Inter RTL<br />

Russia Today<br />

Ouest-France Europe 1<br />

Les Echos<br />

RUSSIAN MEDIA<br />

Conscience du peuple<br />

RTL<br />

Wikipédia<br />

Mes Opinions<br />

FEB AFP 2017 - 15<br />

Debunkers de hoax<br />

17 FEB<br />

Wikipédia<br />

2017 15 FEB 2017 -19 FEB 2017 17<br />

Les Chroniques de Rorschach<br />

11 FEB 2017 13 FEB 2017 - FEB 2017 - 21 FEB 2017 19 FEB 2017 - 23 FEB 2017<br />

Mes Opinions AFP<br />

Debunkers de hoax<br />

Les Chroniques de Rorschach<br />

Slate<br />

BFMTV<br />

Le Figaro<br />

L'Express<br />

Lablogalupus<br />

alalumieredunouveaumonde<br />

Slate<br />

BFMTV<br />

Le Figaro<br />

Le Monde<br />

L'Express<br />

Lablogalupus<br />

Le Monde<br />

Blog du"Plan C", par<br />

alalumieredunouveaumonde<br />

Etienne Chouard S<strong>to</strong>p Mensonges<br />

Franceinfo<br />

Blog Libération<br />

21 FEB 2017 - 25 FEB 2017 23 FEB 2017 - 27 FEB 2017<br />

du"Plan<br />

25 FEB 2017 - 1 MAR 2017<br />

C", par<br />

27 FEB 2017<br />

Etienne<br />

- 3 MAR 2017<br />

Chouard<br />

LEGENDA<br />

S<strong>to</strong>p Mensonges<br />

Franceinfo<br />

Le6emetiroir<br />

Libération<br />

Huffing<strong>to</strong>n Post<br />

Le Parisien-Aujourd’hui en France<br />

Le6emetiroir<br />

Huffing<strong>to</strong>n Post Le Point<br />

Reopen911<br />

Révolution Vibra<strong>to</strong>ire<br />

Les Crises<br />

EXTREME-RIGHT Le Parisien-Aujourd’hui MEDIAen France<br />

Arrêt sur images<br />

Le Point<br />

Reopen911<br />

Révolution Vibra<strong>to</strong>ire<br />

Novopress<br />

Arrêt sur images Marianne Les Crises<br />

Les Mou<strong>to</strong>ns enragés<br />

Novopress<br />

SERVING SUGGESTIONS Cercle des<br />

Marianne<br />

Les Volontaires Mou<strong>to</strong>ns enragés IlFat<strong>to</strong>Quotidiano.fr<br />

Contre-Info<br />

L'Obs<br />

Cercle des Volontaires<br />

IlFat<strong>to</strong>Quotidiano.fr Les Brins d'Herbes<br />

Agoravox<br />

Les Brins d'Herbes<br />

MAINSTREAM FRENCH MEDIA<br />

La République du Centre<br />

12 SEP 2016 - 10 OCT 2016 26 SEP 2016 - 24 OCT 2016 10 OCT 2016 - 7 NOV 2017 24 OCT 2017 - 21 NOV 2017 7 NOV 2016 - 5 MAY 2016<br />

Snapchat<br />

Wikihow<br />

CONSPIRACY THEORISTS<br />

5 DEC 2016 - 2 JAN 2017 19 DEC 2016 - 16 JAN 2017 2 JAN 2017 - 30 JAN 2017 16 JAN 2017 - 13 FEB 2017 30 JAN 2017 - 27 FEB 2017<br />

Chapter #3<br />

USING TRACKER SIGNATURES<br />

TO MAP THE TECHNO-<br />

COMMERCIAL UNDERPINNINGS<br />

OF FAKE NEWS SITES<br />

Do <strong>fake</strong> <strong>news</strong> sites use<br />

different kinds of trackers from<br />

mainstream media sites?<br />

How can <strong>fake</strong> <strong>news</strong> and<br />

mainstream media sites be profiled<br />

based on their tracker usage?<br />

How do tracker ecologies on <strong>fake</strong><br />

<strong>news</strong> sites change over time?<br />

Which other websites share<br />

the same tracker ids as <strong>fake</strong><br />

<strong>news</strong> sites?<br />

Do trackers associated with<br />

hyper-partisan and misinformation<br />

sites vary across language<br />

spheres?<br />

92 A FIELD GUIDE TO FAKE NEWS CHAPTER 3: USING TRACKER SIGNATURES TO MAP THE TECHNO-COMMERCIAL UNDERPINNINGS OF FAKE NEWS SITES<br />

93


CHAPTER 3 → RECIPE 1<br />

Introduction - Over the past few decades<br />

many responses <strong>to</strong> misinformation have<br />

focused on mapping and debunking claims<br />

made or repeated by politicians, journalists or<br />

other public figures. What are the prospects of<br />

mapping <strong>fake</strong> <strong>news</strong> online not just by looking at<br />

the circulation of claims, but by examining the<br />

technical infrastructures of the websites through<br />

which these claims are published?<br />

Many websites use “trackers” – small<br />

bits of embedded code – in order <strong>to</strong> moni<strong>to</strong>r<br />

engagement, including visi<strong>to</strong>r numbers, visi<strong>to</strong>r<br />

behaviour and the effectiveness of ads. In this<br />

section we look at how data about web trackers<br />

can be repurposed in order <strong>to</strong> investigate the<br />

technical and commercial underpinnings of<br />

websites associated with <strong>fake</strong> <strong>news</strong> and other<br />

misinformation phenomena.<br />

DO FAKE NEWS SITES<br />

USE DIFFERENT KINDS OF<br />

TRACKERS FROM MAINSTREAM<br />

MEDIA SITES?<br />

BEFORE STARTING<br />

For this recipe you will need two lists of<br />

URLs: one list of <strong>fake</strong> <strong>news</strong> URLs and one list<br />

of mainstream media URLs. How these lists<br />

are obtained is a crucial part of the research<br />

process. You can either draw on existing<br />

lists, or create your own (e.g. by compiling a<br />

selection, triangulating from other sources,<br />

or obtaining from different platforms or<br />

media sources). The starting point that you<br />

choose will affect how <strong>to</strong> read and what<br />

you can do with the results. To illustrate<br />

this recipe, we start with a selection of <strong>fake</strong><br />

<strong>news</strong> pages obtained from a list created by<br />

BuzzFeed News (ordered by most engaged<br />

with content according <strong>to</strong> the BuzzSumo<br />

<strong>to</strong>ol), as well as a list of mainstream media<br />

web pages obtained by triangulating lists from<br />

BuzzFeed News and Alexa.<br />

94 A FIELD GUIDE TO FAKE NEWS CHAPTER 3: USING TRACKER SIGNATURES TO MAP THE TECHNO-COMMERCIAL UNDERPINNINGS OF FAKE NEWS SITES<br />

95


START<br />

obtain seed lists<br />

of <strong>fake</strong> <strong>news</strong> URLs<br />

input URLs in<br />

DMI Tracker Tracker<br />

output data <strong>to</strong><br />

> Extract list of trackers<br />

associated with web pages<br />

on each list<br />

> Calculate tracker usage<br />

per site type<br />

import data in<br />

RAWGraphs<br />

visualise<br />

DO MAINSTREAM<br />

MEDIA AND FAKE NEWS<br />

WEBSITES SHARE THE SAME<br />

TRACKER ECOLOGY?<br />

CHAPTER 3 → RECIPE 1<br />

CALCULATE TRACKER USAGE PER<br />

SITE TYPE<br />

From the source code of web pages it is often<br />

possible <strong>to</strong> see which third-party tracking<br />

services are used.<br />

◊<br />

◊<br />

◊<br />

Collect data about trackers associated with<br />

the web pages on each list. You may use the<br />

→ DMI Tracker Tracker <strong>to</strong>ol <strong>to</strong> collect this<br />

information.<br />

Count the usage of each tracker in <strong>fake</strong><br />

<strong>news</strong> websites and in mainstream <strong>news</strong><br />

websites.<br />

You may use a scatter plot <strong>to</strong> visualise<br />

the resulting data. Each circle represents<br />

one tracker coloured by category. On the<br />

horizontal axis, you can show, for example,<br />

the distribution of trackers usage by<br />

mainstream media and <strong>fake</strong> <strong>news</strong> websites.<br />

On the vertical axis, you can indicate the<br />

overall usage of the tracker. We used<br />

the → RAWGraphs <strong>to</strong>ol <strong>to</strong> generate this<br />

visualisation.<br />

96 A FIELD GUIDE TO FAKE NEWS CHAPTER 3: USING TRACKER SIGNATURES TO MAP THE TECHNO-COMMERCIAL UNDERPINNINGS OF FAKE NEWS SITES<br />

97


DO MAINSTREAM MEDIA<br />

AND FAKE NEWS WEBSITES<br />

SHARE THE SAME TRACKER<br />

ECOLOGIES?<br />

Scatterplot representing tracker<br />

usage on a series of <strong>fake</strong> <strong>news</strong> and<br />

mainstream media sites. While <strong>fake</strong><br />

<strong>news</strong> sites and mainstream media<br />

sites share popular tracker services<br />

such as Google Adsense, DoubleClick<br />

and Google Analytics, mainstream<br />

media sites appears more mature and<br />

sophisticated in its use of trackers in<br />

terms of the number and diversity of<br />

trackers that it uses.<br />

300<br />

Google Adsense<br />

DoubleClick<br />

ScoreCard Research Beacon<br />

Google Analytics<br />

250<br />

Facebook Connect<br />

Tracker Popularity<br />

200<br />

150<br />

Google Publisher Tags<br />

ChartBeat<br />

Aggregate Knowledge<br />

AppNexus<br />

Rubicon<br />

TradeDesk<br />

BlueKai<br />

LiveRamp<br />

eXelate<br />

Google Syndication<br />

Quantcast<br />

Twitter But<strong>to</strong>n<br />

DoubleClick<br />

Ad Exchange-Seller<br />

Datalogix<br />

TRACKER POPULARITY<br />

100<br />

Moat<br />

Facebook Cus<strong>to</strong>m Audience<br />

Facebook Social Graph<br />

Facebook Social Plugins<br />

Gravatar<br />

1<br />

307<br />

Omniture (Adobe Analytics)<br />

Outbrain<br />

Wordpress Stats<br />

RevContent<br />

50<br />

PulsePoint<br />

Content.ad<br />

TRACKER CATEGORY<br />

OwnerIQ<br />

ShareThis<br />

Disqus<br />

Ad<br />

Acxiom<br />

Zypmedia<br />

Analytics<br />

20 40 60 80<br />

Tracker<br />

Widget<br />

100%<br />

MAINSTREAM<br />

Tracker Fakeness Index (%)<br />

100%<br />

FAKE NEWS<br />

98 A FIELD GUIDE TO FAKE NEWS CHAPTER 3: USING TRACKER SIGNATURES TO MAP THE TECHNO-COMMERCIAL UNDERPINNINGS OF FAKE NEWS SITES<br />

99


Google Adsense<br />

Google Analytics<br />

250<br />

Facebook Connect<br />

CHAPTER 3 → RECIPE 2<br />

TY<br />

Tracker Popularity<br />

200<br />

150<br />

100<br />

Moat<br />

50<br />

Google Publisher Tags<br />

Aggregate Knowledge<br />

AppNexus<br />

CHAPTER 3 → RECIPE 1<br />

Rubicon<br />

ChartBeat<br />

TradeDesk<br />

Datalogix<br />

Facebook Cus<strong>to</strong>m Audience<br />

Omniture (Adobe Analytics)<br />

Outbrain<br />

BlueKai<br />

LiveRamp<br />

eXelate<br />

Quantcast<br />

Twitter But<strong>to</strong>n<br />

Google Syndication<br />

HOW CAN FAKE NEWS<br />

DoubleClick<br />

AND MAINSTREAM MEDIA SITES<br />

Ad Exchange-Seller<br />

BE PROFILED BASED ON<br />

Facebook Social Plugins<br />

THEIR TRACKER USAGE?<br />

Gravatar<br />

Facebook Social Graph<br />

Wordpress Stats<br />

RevContent<br />

Content.ad<br />

PulsePoint<br />

Y<br />

OwnerIQ<br />

ShareThis<br />

Disqus<br />

Acxiom<br />

Zypmedia<br />

100%<br />

MAINSTREAM<br />

SERVING SUGGESTIONS<br />

This recipe can be used <strong>to</strong> profile<br />

the tracking practices associated<br />

with different kinds of websites –<br />

including which trackers are either<br />

mainly or exclusively associated<br />

with <strong>fake</strong> <strong>news</strong> websites and what<br />

these trackers do – as well as<br />

identifying most commonly used<br />

trackers. It can also be used for<br />

exploring the “long tail” of smaller<br />

and more specialised trackers.<br />

20 40 60 80<br />

Tracker Fakeness Index (%)<br />

BEFORE STARTING<br />

For this recipe you will need two lists of<br />

URLs: one list of <strong>fake</strong> <strong>news</strong> URLs and one list<br />

of mainstream media URLs. How these lists<br />

are obtained is a crucial part of the research<br />

process. You can either draw on existing lists, or<br />

create your own (e.g. by compiling a selection,<br />

triangulating from other sources, or obtaining<br />

from different platforms or media sources). The<br />

starting point that you choose will affect how <strong>to</strong><br />

read and what you can do with the results. To<br />

illustrate this recipe, we start with a selection<br />

of <strong>fake</strong> <strong>news</strong> pages obtained from a list created<br />

by BuzzFeed News (ordered by most engaged<br />

with content according <strong>to</strong> the BuzzSumo <strong>to</strong>ol),<br />

as well as a list of mainstream media web pages<br />

obtained by triangulating lists from BuzzFeed<br />

News and Alexa.<br />

100%<br />

FAKE NEWS<br />

100 A FIELD GUIDE TO FAKE NEWS CHAPTER 3: USING TRACKER SIGNATURES TO MAP THE TECHNO-COMMERCIAL UNDERPINNINGS OF FAKE NEWS SITES<br />

101


CHAPTER 3 → RECIPE 2<br />

START<br />

obtain seed lists<br />

of <strong>fake</strong> <strong>news</strong> URLs<br />

input URLs in<br />

DMI Tracker Tracker<br />

output data <strong>to</strong><br />

> Extract list of trackers<br />

associated with web pages<br />

on each list<br />

import data in<br />

Gephi<br />

visualise<br />

HOW DO FAKE NEWS SITES<br />

AND MAINSTREAM MEDIA<br />

CLUSTER ACCORDING THEIR<br />

TRACKER USAGE?<br />

GRAPH RELATIONS BETWEEN PAGES<br />

AND TRACKERS<br />

In order <strong>to</strong> explore how different URLs share<br />

the same patterns of tracker usage we can<br />

create a network graph <strong>to</strong> highlight associations<br />

between web pages <strong>to</strong> their corresponding<br />

trackers.<br />

◊<br />

◊<br />

◊<br />

Extract lists of trackers associated with the<br />

initial lists of <strong>fake</strong> <strong>news</strong> and mainstream<br />

media pages. You may use the → DMI<br />

Tracker Tracker <strong>to</strong>ol <strong>to</strong> collect this<br />

information.<br />

Create a network in order <strong>to</strong> show the<br />

tracker usage patterns of the different<br />

web pages. We used → Gephi in order <strong>to</strong><br />

visually explore the network using a force<br />

directed network layout <strong>to</strong> help read the<br />

data.<br />

You can annotate the network graph in<br />

order <strong>to</strong> highlight the cluster of URLs<br />

(e.g. <strong>fake</strong> <strong>news</strong> clusters, or mainstream<br />

media clusters).<br />

102 A FIELD GUIDE TO FAKE NEWS CHAPTER 3: USING TRACKER SIGNATURES TO MAP THE TECHNO-COMMERCIAL UNDERPINNINGS OF FAKE NEWS SITES<br />

103


HOW DO FAKE NEWS<br />

SITES AND MAINSTREAM<br />

MEDIA CLUSTER<br />

ACCORDING THEIR<br />

TRACKER USAGE?<br />

empire<strong>news</strong>.net +<br />

<strong>news</strong>4ktla.com +<br />

straights<strong>to</strong>ned.com<br />

Bipartite network of trackers and<br />

websites that use them. Shared tracker<br />

signatures may be used <strong>to</strong> explore<br />

tracker practices or strategies amongst<br />

a set of websites or <strong>to</strong> detect <strong>fake</strong> <strong>news</strong><br />

“media groups.”<br />

abc<strong>news</strong>.com.co<br />

<strong>news</strong>examiner.net<br />

nationalreport.net<br />

<strong>news</strong>lo.com<br />

<strong>news</strong>biscuit.com<br />

en.mediamass.net<br />

thevalleyreport.com<br />

the<strong>news</strong>nerd.com<br />

satiratribune.com<br />

<strong>news</strong>hub.info<br />

celebtricity.com<br />

empireherald.com<br />

ncscooper.com<br />

burradstreetjournal.com<br />

notallowed<strong>to</strong>.com<br />

baltimoregazzette.com<br />

dailyfinesser.com<br />

huzlers.com<br />

real<strong>news</strong>rightnow.com<br />

wm21<strong>news</strong>.com<br />

stuppid.com<br />

kata33. om<br />

NUMBER OF CONNECTIONS<br />

Fake <strong>news</strong> sites<br />

Mainstream media sites<br />

Landrypost.com<br />

Tracker<br />

104 A FIELD GUIDE TO FAKE NEWS CHAPTER 3: USING TRACKER SIGNATURES TO MAP THE TECHNO-COMMERCIAL UNDERPINNINGS OF FAKE NEWS SITES<br />

105


<strong>news</strong>4ktla.com +<br />

straights<strong>to</strong>ned.com<br />

CHAPTER 3 → RECIPE 2<br />

<strong>news</strong>lo.com<br />

abc<strong>news</strong>.com.co<br />

<strong>news</strong>examiner.net<br />

nationalreport.net<br />

<strong>news</strong>biscuit.com<br />

en.mediamass.net<br />

CHAPTER 3 → RECIPE 3<br />

HOW DO TRACKER ECOLOGIES<br />

ON FAKE NEWS SITES<br />

CHANGE OVER TIME??<br />

thevalleyreport.com<br />

the<strong>news</strong>nerd.com<br />

satiratribune.com<br />

<strong>news</strong>hub.info<br />

celebtricity.com<br />

empireherald.com<br />

ncscooper.com<br />

burradstreetjournal.com<br />

notallowed<strong>to</strong>.com<br />

baltimoregazzette.com<br />

dailyfinesser.com<br />

huzlers.com<br />

SERVING SUGGESTIONS<br />

This recipe can be used <strong>to</strong> explore how a<br />

set of web pages can be grouped based<br />

on their tracker signatures. This provides a<br />

complementary picture <strong>to</strong> lists or metrics<br />

(e.g. of most and least used trackers across<br />

the pages) by facilitating exploration of<br />

relations between trackers and websites.<br />

For example it could be used as a starting<br />

point <strong>to</strong> identify potential <strong>fake</strong> <strong>news</strong><br />

“media groups” for further investigation,<br />

or <strong>to</strong> explore the different web tracking<br />

practices, styles and footprints of <strong>fake</strong><br />

<strong>news</strong> web pages – including comparisons<br />

between pages associated with different<br />

regions, issues or sources.<br />

wm21<strong>news</strong>.com<br />

stuppid.com<br />

kata33. om<br />

Landrypost.com<br />

real<strong>news</strong>rightnow.com<br />

BEFORE STARTING<br />

For this recipe you will need the source code<br />

of the same web page (or set of web pages) at<br />

two different moments in time. You can obtain<br />

saved copies of the same page over time (e.g.<br />

through manually or au<strong>to</strong>matically saving the<br />

source code yourself) or you can use public web<br />

archiving projects such as the Internet Archive’s<br />

→ The Wayback Machine.<br />

106 A FIELD GUIDE TO FAKE NEWS CHAPTER 3: USING TRACKER SIGNATURES TO MAP THE TECHNO-COMMERCIAL UNDERPINNINGS OF FAKE NEWS SITES<br />

107


START<br />

obtain seed lists<br />

of <strong>fake</strong> <strong>news</strong> URLs<br />

input URLs in<br />

The Wayback Machine<br />

merge all URLs in a single list<br />

+<br />

input URLs in<br />

DMI Tracker Tracker<br />

output data <strong>to</strong><br />

import data in<br />

Gephi<br />

visualise<br />

> Extract list of trackers<br />

associated with web pages<br />

on each list<br />

obtain past versions<br />

of the pages<br />

CHAPTER 3 → RECIPE 3<br />

GRAPH RELATIONS BETWEEN PAGES<br />

AND TRACKERS<br />

This recipe can be used <strong>to</strong> identify which<br />

trackers were being used by a given web page<br />

at different moments in time. It might be useful<br />

<strong>to</strong> chart changes in tracking practices – for<br />

example by examining the impact and responses<br />

<strong>to</strong> events like Google and Facebook’s bans of<br />

<strong>fake</strong> <strong>news</strong> providers from their ads programs in<br />

November 2016.<br />

◊<br />

◊<br />

◊<br />

◊<br />

Obtain archived copies of a webpage. You<br />

may use → The Wayback Machine <strong>to</strong> see<br />

how a given page changed over time.<br />

Identify associated trackers with the<br />

current and previous version of the page.<br />

You may use the → DMI Tracker Tracker<br />

<strong>to</strong>ol <strong>to</strong> collect such information.<br />

Identify the trackers which are only present<br />

on the first date, the ones that which are<br />

only present on the second date and the<br />

ones that are shared across both dates.<br />

You can group trackers in<strong>to</strong> three lists,<br />

colouring them accordingly.<br />

HOW DO FAKE NEWS SITES<br />

ADAPT THEIR TRACKER USAGE<br />

IN RESPONSE TO BLACKLISTING<br />

FROM MAJOR AD NETWORKS?<br />

108 A FIELD GUIDE TO FAKE NEWS CHAPTER 3: USING TRACKER SIGNATURES TO MAP THE TECHNO-COMMERCIAL UNDERPINNINGS OF FAKE NEWS SITES<br />

109


HOW DO FAKE NEWS SITES<br />

ADAPT THEIR TRACKER<br />

USAGE IN RESPONSE TO<br />

BLACKLISTING FROM<br />

MAJOR AD NETWORKS?<br />

Tracker ecologies on <strong>fake</strong> <strong>news</strong> sites<br />

before and after blacklisting from<br />

major ad networks. While ad networks<br />

from which <strong>fake</strong> <strong>news</strong> sites have been<br />

blacklisted remain in the source code<br />

of these sites and hence are present<br />

in the graphic even after the moment<br />

of blacklisting, the visualisation also<br />

illustrates new ad networks that <strong>fake</strong><br />

<strong>news</strong> sites have moved <strong>to</strong>. A manual<br />

review of ad services used <strong>to</strong> serve ads<br />

on the website interface may help <strong>to</strong><br />

further refine this analysis and identify<br />

false positives (i.e. tracker services that<br />

are no longer in use but whose code<br />

remains embedded in these sites).<br />

TRACKER POPULARITY<br />

TRACKER<br />

TRACKER<br />

TRACKER<br />

TRACKER CATEGORY<br />

Ad<br />

Analytics<br />

Tracker<br />

Widget<br />

Admarvel<br />

Clicksor<br />

Drawbridge<br />

Facebook Cus<strong>to</strong>m Audience<br />

Google Publisher Tags<br />

Gumgum<br />

Media.net<br />

Sekindo<br />

Doubleverify<br />

Visible Measures<br />

Omniture (Adobe Analytics)<br />

TRACKERS PRESENT<br />

ONLY BEFORE DECEMBER<br />

Doubleclick<br />

Google Analytics<br />

Google Adsense<br />

Doubleclick Ad Exchange-seller<br />

Google Syndication<br />

Scorecard Research Beacon<br />

Gravatar<br />

Wordpress Stats<br />

Facebook Connect<br />

Exelate<br />

Disqus<br />

Brightroll<br />

Adobe Audience Manager<br />

Bluekai<br />

Amazon Associates<br />

Appnexus<br />

Bidswitch<br />

Criteo<br />

Mediamath<br />

Openx<br />

Pubmatic<br />

Tradedesk<br />

Facebook Social Plugins<br />

Taboola<br />

Twitter But<strong>to</strong>n<br />

Adtech<br />

Advertising.com<br />

Index Exchange (Formerly Casale Media)<br />

Pulsepoint<br />

Quantcast<br />

Rubicon<br />

Spoutable<br />

Stickyads<br />

Tapad<br />

Teads<br />

Turn Inc.<br />

Yahoo Ad Exchange<br />

Tubemogul<br />

Krux Digital<br />

Liveramp<br />

Acloudimages<br />

Acuity Ads<br />

Adscale<br />

Eyeview<br />

Revcontent<br />

Smart Adserver<br />

Sovrn (Formerly Lijit Networks)<br />

Twitter Advertising<br />

Zypmedia<br />

At Internet<br />

Twitter Analytics<br />

Aggregate Knowledge<br />

Lotame<br />

Owneriq<br />

Rocket Fuel<br />

Videology<br />

Addthis<br />

Facebook Social Graph<br />

Lockerz Share<br />

Pinterest<br />

Sharethis<br />

Twitter Badge<br />

Typekit By Adobe<br />

TRACKERS<br />

ALWAYS PRESENT<br />

Adform<br />

Infectious Media<br />

Yahoo Ad Manager Plus<br />

Adap.tv<br />

Adroll<br />

Bidswitch<br />

Crimtan<br />

Datalogix<br />

Dataxu<br />

Digilant<br />

Dstillery<br />

Getintent<br />

Improve Digital<br />

Infolinks<br />

Internet Billboard<br />

Smaa<strong>to</strong><br />

Smartclip<br />

Spotxchange<br />

Switch Concepts<br />

Yieldlab<br />

Kxcdn<br />

Beeswax<br />

Bidtheatre<br />

Chango<br />

Do<strong>to</strong>mi<br />

Kixer<br />

Mythings<br />

Netmining<br />

Pagefair<br />

Radiumone<br />

Sumome<br />

Tumblr Dashboard<br />

TRACKERS PRESENT<br />

ONLY AFTER JANUARY<br />

110 A FIELD GUIDE TO FAKE NEWS CHAPTER 3: USING TRACKER SIGNATURES TO MAP THE TECHNO-COMMERCIAL UNDERPINNINGS OF FAKE NEWS SITES<br />

111


CHAPTER 3 → RECIPE 4<br />

Doubleclick<br />

CHAPTER 3 → RECIPE 3<br />

Google Analytics<br />

Google Adsense<br />

Doubleclick Ad Exchange-seller<br />

Google Syndication<br />

Scorecard Research Beacon<br />

Gravatar<br />

Wordpress<br />

SERVING SUGGESTIONS<br />

Stats<br />

Facebook Connect<br />

Given debates and proposals<br />

Exelate about s<strong>to</strong>pping the ad revenue<br />

of <strong>fake</strong> <strong>news</strong>, this recipe may be<br />

Disqus<br />

used <strong>to</strong> understand how <strong>fake</strong><br />

Brightroll <strong>news</strong> websites are adapting <strong>to</strong><br />

Adobe the measures Audience taken Manager by trackers<br />

Bluekai services, technology companies<br />

Amazon and advertisers Associates – as well as how<br />

Appnexus effective these measures are.<br />

Bidswitch For example, it can show which<br />

trackers have been dropped, which<br />

Criteo<br />

remain and which are added at<br />

Mediamath<br />

different moments in time.<br />

Openx<br />

Pubmatic<br />

Tradedesk<br />

Facebook Social Plugins<br />

Taboola<br />

Twitter But<strong>to</strong>n<br />

Adtech<br />

WHICH OTHER<br />

WEBSITES SHARE<br />

THE SAME TRACKER IDS<br />

AS FAKE NEWS SITES?<br />

BEFORE STARTING<br />

Before you start you will need <strong>to</strong> compile<br />

or identify seed lists of <strong>fake</strong> <strong>news</strong> or other<br />

misinformation websites. We illustrate this<br />

recipe by examining which websites use the<br />

same Google Analytics IDs as a list of websites<br />

from the EU Disinformation Review.<br />

112 A FIELD GUIDE TO FAKE NEWS WHICH OTHER WEBSITES SHARE THE SAME TRACKER IDS AS FAKE NEWS PAGES?<br />

113


START<br />

obtain seed lists<br />

of <strong>fake</strong> <strong>news</strong> URLs<br />

CHAPTER 3 → RECIPE 4<br />

obtain the Google Analytics ID<br />

associated <strong>to</strong> each page<br />

input IDs in<br />

Spyonweb<br />

get the list of associated pages<br />

obtain screenshots<br />

of associated pages<br />

categorise websites groups<br />

visualise<br />

WHAT MEDIA GROUP STRATEGIES<br />

CAN BE DETECTED THROUGH<br />

SHARED GOOGLE ANALYTICS IDS?<br />

IDENTIFY WEBSITES WHICH SHARE<br />

TRACKER IDS WITH A SEED LIST OF<br />

PAGES OR SITES<br />

This recipe can be used <strong>to</strong> identify which other<br />

websites share the same tracker IDs as web<br />

pages on a given list.<br />

◊<br />

◊<br />

◊<br />

◊<br />

Extract the Google Analytics ID for<br />

each URL in your starting list. You can<br />

do this manually (e.g. by looking in the<br />

source code for a string in the form “UAxxxxxxx”)<br />

or au<strong>to</strong>matically through web<br />

scraping or other <strong>to</strong>ols (in this example we<br />

wrote a cus<strong>to</strong>m script in order <strong>to</strong> extract<br />

this information from the metadata of the<br />

website).<br />

Obtain a list of pages associated with the<br />

same ID. We used the API of<br />

→ Spyonweb.com <strong>to</strong> get this information.<br />

Take a screenshot of each web page. We<br />

used a script <strong>to</strong> au<strong>to</strong>mate the process of<br />

obtaining screenshots, in order <strong>to</strong> visually<br />

compare the different websites <strong>to</strong> identify<br />

different kinds of media groups.<br />

Place <strong>to</strong>gether screenshots of pages with<br />

the same ID <strong>to</strong> spot differences and<br />

similarities between websites across and<br />

within groups.<br />

114 A FIELD GUIDE TO FAKE NEWS WHICH OTHER WEBSITES SHARE THE SAME TRACKER IDS AS FAKE NEWS PAGES?<br />

115


WHAT MEDIA GROUP<br />

STRATEGIES CAN BE<br />

DETECTED THROUGH<br />

SHARED GOOGLE<br />

ANALYTICS IDS?<br />

rt.com<br />

xryshaygh.com<br />

defense<strong>news</strong>.com newcoldwar.org almanach.cz<br />

19 disinformation s<strong>to</strong>ries 1 disinformation s<strong>to</strong>ry 1 disinformation s<strong>to</strong>ry 7 disinformation s<strong>to</strong>ries 3 disinformation s<strong>to</strong>ries<br />

UA-5773642<br />

UA-4839940<br />

UA-841082<br />

UA-15942468<br />

UA-3004323<br />

A selection of websites which share<br />

the same Google Analytics IDs, based<br />

on seed list from EU Disinformation<br />

Review. This illustrates the diversity of<br />

online settings where claims labelled<br />

as Russian disinformation are shared<br />

– from large media groups such as<br />

Russia Today, <strong>to</strong> themed clusters (e.g.<br />

military or mysticism), and geographical<br />

clusters (e.g. Canadian). One can also<br />

identify distinctive visual styles and<br />

possible shared CMS features amongst<br />

different websites in these clusters,<br />

which may be used as the basis for<br />

further investigations in<strong>to</strong> the media,<br />

publication and communication<br />

strategies of websites associated with<br />

online misinformation.<br />

Media group<br />

(Russia Today)<br />

assange.rt.com<br />

doc.rt.com<br />

Lone<br />

Webmaster<br />

asco.gr<br />

royalparadise.gr<br />

Themed network<br />

(Navy, Airforce)<br />

armedforcesjournal.com<br />

sightlinemediagroup.com<br />

Canadian socialism<br />

and unions<br />

oakvillendp.ca<br />

socialiststudies.com<br />

Themed network<br />

(Mysticism, Liberland)<br />

grand-mystical-lodge.com<br />

malachim.cz<br />

learnrussian.rt.com<br />

discoverthassos.com<br />

militarytimes.com<br />

ndpsocialists.ca<br />

liberlandpress.com<br />

russian.rt.com<br />

thassosinn.gr<br />

airforcetimes.com<br />

ccu-csc.ca<br />

illuminati-journal.com<br />

catalog.rt.com<br />

marinecorpstimes.com<br />

oldm.cz<br />

116 A FIELD GUIDE TO FAKE NEWS WHICH OTHER WEBSITES SHARE THE SAME TRACKER IDS AS FAKE NEWS PAGES?<br />

117


t.com<br />

xryshaygh.com<br />

defense<strong>news</strong>.com newcoldwar.org almanach.cz<br />

19 disinformation s<strong>to</strong>ries 1 disinformation s<strong>to</strong>ry 1 disinformation s<strong>to</strong>ry 7 disinformation s<strong>to</strong>ries 3 disinformation s<strong>to</strong>ries<br />

UA-5773642<br />

UA-4839940<br />

UA-841082<br />

UA-15942468<br />

UA-3004323<br />

CHAPTER 3 → RECIPE 5<br />

Media group<br />

(Russia Today)<br />

assange.rt.com<br />

CHAPTER Lone 3 → RECIPE 4<br />

Webmaster<br />

asco.gr<br />

Themed network<br />

(Navy, Airforce)<br />

armedforcesjournal.com<br />

Canadian socialism Themed network<br />

and unions<br />

ASSOCIATED WITH HYPER-<br />

(Mysticism, Liberland)<br />

oakvillendp.ca<br />

DO TRACKERS<br />

PARTISAN AND MISINFORMATION<br />

SITES VARY ACROSS<br />

LANGUAGE grand-mystical-lodge.com<br />

SPHERES?<br />

doc.rt.com<br />

royalparadise.gr<br />

sightlinemediagroup.com<br />

socialiststudies.com<br />

malachim.cz<br />

SERVING SUGGESTIONS<br />

BEFORE STARTING<br />

learnrussian.rt.com<br />

russian.rt.com<br />

This recipe may be used in the<br />

discoverthassos.com<br />

service of expanding a group of<br />

<strong>fake</strong> <strong>news</strong> web pages – in order<br />

<strong>to</strong> derive lists of other websites<br />

which share the same tracker IDs.<br />

It may also be used <strong>to</strong> provide<br />

context <strong>to</strong> the thassosinn.gr digital strategies<br />

and “media groupings” of <strong>fake</strong><br />

<strong>news</strong> providers<br />

militarytimes.com<br />

airforcetimes.com<br />

ndpsocialists.ca<br />

ccu-csc.ca<br />

For this recipe, you will need lists of <strong>fake</strong> <strong>news</strong>,<br />

liberlandpress.com<br />

hyper-partisan or misinformation sites in<br />

different language spheres in order <strong>to</strong> compare<br />

their trackers and tracking practices. We<br />

illustrate this recipe with reference <strong>to</strong> hyperpartisan,<br />

<strong>fake</strong> <strong>news</strong> and misinformation sites in<br />

illuminati-journal.com<br />

Dutch, English and German language spheres.<br />

catalog.rt.com<br />

marinecorpstimes.com<br />

oldm.cz<br />

118 A FIELD GUIDE TO FAKE NEWS WHICH OTHER WEBSITES SHARE THE SAME TRACKER IDS AS FAKE NEWS PAGES?<br />

119


CHAPTER 3 → RECIPE 5<br />

START<br />

obtain seed lists<br />

of <strong>fake</strong> <strong>news</strong> URLs<br />

input URLs in<br />

DMI Tracker Tracker<br />

output data <strong>to</strong><br />

> Extract list of trackers<br />

associated with web pages<br />

on each list<br />

compare trackers lists with<br />

DMI Triangulation Tool<br />

IDENTIFY TRACKERS PER LANGUAGE<br />

SPHERE<br />

◊<br />

◊<br />

◊<br />

Extract trackers associated with lists of the<br />

web pages for each language sphere. We did<br />

this using the → DMI Tracker Tracker <strong>to</strong>ol.<br />

Identify the trackers which are shared across<br />

and which are unique <strong>to</strong> different languages<br />

spheres within the dataset. We did this using<br />

the → DMI Triangulation <strong>to</strong>ol.<br />

You can illustrate the results using the visual<br />

metaphor of magnets. Each of the three<br />

languages are represented on the corner of a<br />

triangle. The trackers are distributed in the<br />

triangle according <strong>to</strong> their usage: if a tracker<br />

is used by all three languages it will appear in<br />

the middle, if it is used by two languages the<br />

tracker will be placed on the edge between<br />

the two and so on.<br />

visualise<br />

DO MISINFORMATION AND<br />

HYPER-PARTISAN WEBSITES<br />

IN DIFFERENT LANGUAGE<br />

SPHERES HAVE DISTINCT<br />

TRACKER ECOLOGIES?<br />

120 A FIELD GUIDE TO FAKE NEWS WHICH OTHER WEBSITES SHARE THE SAME TRACKER IDS AS FAKE NEWS PAGES?<br />

121


DO MISINFORMATION<br />

AND HYPER-PARTISAN<br />

WEBSITES IN DIFFERENT<br />

LANGUAGE SPHERES<br />

HAVE DISTINCT TRACKER<br />

ECOLOGIES?<br />

veeseo<br />

Mark & Mini<br />

Ligatus<br />

Atlas<br />

DUTCH<br />

AdLantic<br />

Platform161<br />

SkimLinks<br />

LinkShare<br />

Visualisation of tracker ecologies<br />

associated with hyper-partisan or<br />

misinformation sites across three<br />

language spheres. While popular<br />

ad and widget services such as<br />

DoubleClick, Google Adsense and<br />

Facebook Connect are shared across<br />

language spheres, unique services per<br />

language sphere may also be detected.<br />

For example, trackers associated with<br />

the Russian-language focused Mail.<br />

ru Group are only found in the set of<br />

websites associated with the German<br />

language sphere.<br />

Facebook<br />

Connect<br />

Media<br />

Innovation<br />

Group<br />

DoubleClick<br />

Ad Exchange-Seller<br />

Gravatar<br />

DoubleClick<br />

Wordpress<br />

Stats<br />

Google Adsense<br />

Google<br />

Twitter But<strong>to</strong>n Syndication<br />

Google<br />

Analytics<br />

SumoMe ENGLISH<br />

Google<br />

Tag Manager<br />

ChartBeat Content ad<br />

Spoutable<br />

Distroscale<br />

Zedo<br />

First Impression<br />

TRACKER POPULARITY<br />

1<br />

78<br />

TRACKER CATEGORY<br />

Ad<br />

Analytics<br />

Tracker<br />

Widget<br />

INFOnline<br />

Plista<br />

Piwik<br />

Tynt<br />

adNET de<br />

VKontakte Widgets<br />

AMP Platform<br />

M P NewMedia<br />

GERMAN<br />

122 A FIELD GUIDE TO FAKE NEWS WHICH OTHER WEBSITES SHARE THE SAME TRACKER IDS AS FAKE NEWS PAGES?<br />

123


CHAPTER 3 → RECIPE 5<br />

Facebook<br />

Connect<br />

Media<br />

Innovation<br />

Group<br />

DoubleClick<br />

Ad Exchange-Seller<br />

Gravatar<br />

DoubleClick<br />

Wordpress<br />

Stats<br />

Google Adsense<br />

Google<br />

Twitter But<strong>to</strong>n Syndication<br />

Google<br />

Analytics<br />

SumoMe ENGLISH<br />

Google<br />

Tag Manager<br />

ChartBeat Content ad<br />

Spoutable<br />

Distroscale<br />

Zedo<br />

First Impression<br />

s<br />

SERVING SUGGESTIONS<br />

This recipe can be used <strong>to</strong><br />

identify trackers for further<br />

investigation – including<br />

language sphere specific and<br />

cross-language trackers. It<br />

may help <strong>to</strong> provide lines<br />

of inquiry for looking in<strong>to</strong><br />

what is distinctive about the<br />

commercial and technical<br />

underpinnings of <strong>fake</strong> <strong>news</strong> in<br />

different language spheres.<br />

124 A FIELD GUIDE TO FAKE NEWS WHICH OTHER WEBSITES SHARE THE SAME TRACKER IDS AS FAKE NEWS PAGES?<br />

125


126 A FIELD GUIDE TO FAKE NEWS

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!