Content uploaded by Hamid Akın Ünver
Author content
All content in this area was uploaded by Hamid Akın Ünver on Apr 02, 2019
Content may be subject to copyright.
Russian Digital Media and
Information Ecosystem in Turkey
H. Akın Ünver | EDAM, Oxford CTGA & Kadir Has University
Cyber Governance and Digital Democracy 2019/1
March 2019
1
Cyber Governance and Digital Democracy 2019/1
Russian Digital Media and
Information Ecosystem in Turkey
H. Akın Ünver | EDAM, Oxford CTGA & Kadir Has Üniversitesi
In recent years, Russian digital information operations,
including disinformation, fake news, and election meddling
have assumed prominence in international news and
scholarly research outlets. A simple Google Trends query
shows us that ‘fake news’ as a term enters into global
mainstream lexicon starting with October 2016, peaking
in the immediate aftermath of the US Presidential Election
in November. Since then, disinformation has been largely
synonymous with Russian digital information operations in
the West, and a number of empirical research projects have
begun focusing on the impact of information warfare on
elections and political behavior.
Russian media ecosystem in Western democracies,
including information and dis-information dynamics,
are quite well-documented1. This focus owes largely to
increased awareness of election meddling, fake news and
digital spoilers such as trolls and bots that often have real-
life effects. In addition to other digital contestation types,
including cyber warfare, Russian information operations are
not conned to the country’s ofcial communication policy.
These strategies are part of the Russian military doctrine,
most relevant of which has been the 2010 Military Doctrine
of the Russian Federation, which sought to “escalate to de-
escalate”2 tensions encompassing the country’s western
borders. To achieve this, the document advised ‘hybrid
war’, which is an umbrella term to dene untraceable and
largely non-violent tools and methods that complement
conventional military efforts. The 2010 doctrine was further
bolstered by the 2013 Gerasimov Doctrine, which, among
other things, diagnosed the “blurring the lines between
the states of war and peace”, adding that “wars are no
longer declared and having begun, proceed according
to an unfamiliar template”3. Hybrid war is not a Russian
invention, nor is Russia the rst state to use non-military
measures to complement military efforts. Rather, the 2010
doctrine was an acknowledgment of the term ‘hybrid war’,
ofcially coined rst by the USCENTCOM in its analysis of
the 2006 Israel-Hezbollah War4. The use of conventional
and unconventional tactics, coupled with the new advances
Hunt Allcott and Matthew Gentzkow, “Social Media and Fake News in the 2016 Election,” Working Paper (National Bureau of Economic Research, January 2017), https://
doi.org/10.3386/w23089; Meital Balmas, “When Fake News Becomes Real: Combined Exposure to Multiple News Sources and Political Attitudes of Inefcacy, Alienation,
and Cynicism,” Communication Research 41, no. 3 (April 1, 2014): 430–54, https://doi.org/10.1177/0093650212453600; Michael C. Dorf and Sidney Tarrow, “Stings
and Scams: ‘Fake News,’ the First Amendment, and the New Activist Journalism,” SSRN Scholarly Paper (Rochester, NY: Social Science Research Network, January 26,
2017), https://papers.ssrn.com/abstract=2906444; David M. J. Lazer et al., “The Science of Fake News,” Science 359, no. 6380 (March 9, 2018): 1094–96, https://doi.
org/10.1126/science.aao2998.
“Text of Newly-Approved Russian Military Doctrine,” Carnegie Endowment for International Peace, February 5, 2010,
https://carnegieendowment.org/2010/02/05/text-of-newly-approved-russian-military-doctrine-pub-40266.
Henry Foy, “Valery Gerasimov, the General with a Doctrine for Russia,” Financial Times, September 15, 2017,
https://www.ft.com/content/7e14a438-989b-11e7-a652-cde3f882dd7b.
Matt M. Matthews, “We Were Caught Unprepared: The 2006 Hezbollah-Israeli War” (Fort Leavenworth, Kansas: US Army Combined Arms Center, 2007),
https://www.armyupress.army.mil/Portals/7/combat-studies-institute/csi-books/we-were-caught-unprepared.pdf.
1
2
3
4
This research has been made possible by funding obtained from the US-based Chrest Foundation
for the project “Turkey and Russia: Context and Prospects”.
INTRODUCTION
2
Cyber Governance and Digital Democracy 2019/1
in communication technology, widened the battleeld and
forced the sides to ght along a broad spectrum of ideas,
images and appearances, all oating in digital space. In
the words of Timothy Thomas, Russian high command was
deeply inuenced by the 2006 USCENTCOM report and that
analysis had led Moscow’s transition into a new thinking in
terms of how to merge new communication technologies
with strategic thinking: “a real cognitive war underway in the
ether and media for the hearts and minds of its citizens at
home and abroad”5.
In many ways, the Internet has become a force domain,
just like land, sea and air. In January 2019, the world has
attained 51% Internet penetration, meaning more than half
of the world is now online and digitally interconnected6.
Foreseeing an inevitable mass global interconnectivity, most
major countries have already set up long-term strategies in
place to situate themselves into a more favorable strategic
position in the digital domain. For the rest of the state actors,
there have been two real wake-up calls to adapt to the
digital medium. The rst was the Arab Spring movement that
rocked the MENA capitals through 2010-12 and the second
was the Occupy-inspired or related movements that did
the same in the West7. Both movements demonstrated the
disruptive capacity of social media platforms to circumvent
and bypass state surveillance and repression. It is during this
period that social media has begun to transform. Instagram
was launched in October 2010, following Facebook’s
politically important geotag function via ‘Places’ app in
August 2010. Facebook bought Instagram in April 2012 and
WhatsApp in February 2014, turning itself into the biggest
heavyweight in social media. In tandem, Twitter emerged
as a more important political communication alternative to
Facebook, as the Arab Spring and Occupy movements used
primarily Twitter to organize and disseminate messages8.
The publicly visible 140-character platform architecture
of Twitter, combined with its fast media upload system,
rendered it the primary venue for critical information ows
during emergencies, protests and civil wars. The Syrian
Civil War, conict in Ukraine and war against ISIS have all
substantially contributed to the rise of Twitter as the primary
emergency-related social media platform9.
Both the NATO Bi-Strategic Capstone Concept10 and the
2010 Russian Military Doctrine11 document have underlined
the threat of an adversary “with the ability to simultaneously
employ conventional and non-conventional means
adaptively in pursuit of their objectives”. These means
were fairly identical in both NATO and Russian military
documents: nuclear proliferation, terrorism, cybercrime
and cyberwar, organized crime and its role in drugs, arms
and human trafcking, migration, ethnic and religious
conicts, population conicts due to resource scarcity and
globalization. Both documents also emphasized the digital
medium as an emerging frontier of political contestation.
NATO followed-up with a 2011 ‘Countering Hybrid Threats’
experiment to develop a unied alliance strategy against
disinformation and media manipulation efforts. This
was ultimately abandoned due to uneven interest and
commitment by the constituent countries12. Compared
to NATO, however, Moscow was quicker to embrace the
uncertainty of the new information revolution, the hybrid
nature of social media and how its intricate twists and turns
could be deployed to support what would later be dened
as the ‘sub-threshold warfare strategy’.
Timothy Thomas, “Russia’s 21st Century Information War: Working to Undermine and Destabilize Populations” (Riga: NATO STRATCOM, 2015),
https://www.stratcomcoe.org/timothy-thomas-russias-21st-century-information-war-working-undermine-and-destabilize-populations.
Abdi Latif Dahir, “Half the World’s Population Used the Internet in 2018 - ITU — Quartz Africa,” Quartz, December 11, 2018,
https://qz.com/africa/1490997/more-than-half-of-worlds-population-using-the-internet-in-2018/.
Philip N. Howard et al., “Opening Closed Regimes: What Was the Role of Social Media During the Arab Spring?,” SSRN Scholarly Paper (Rochester, NY: Social Science
Research Network, 2011), https://papers.ssrn.com/abstract=2595096.
Alexandra Segerberg and W. Lance Bennett, “Social Media and the Organization of Collective Action: Using Twitter to Explore the Ecologies of Two Climate Change Protests,”
The Communication Review 14, no. 3 (July 1, 2011): 197–215, https://doi.org/10.1080/10714421.2011.597250; W. Lance Bennett and Alexandra Segerberg, “Digital Media
and the Personalization of Collective Action,” Information, Communication & Society 14, no. 6 (September 1, 2011): 770–99, https://doi.org/10.1080/1369118X.2011.579141.
Markus Rohde et al., “Out of Syria: Mobile Media in Use at the Time of Civil War,” International Journal of Human–Computer Interaction 32, no. 7 (July 2, 2016): 515–31,
https://doi.org/10.1080/10447318.2016.1177300.
For full text, see: http://www.act.nato.int/images/stories/events/2010/20100826_bi-sc_cht.pdf
For full text English translation, see: https://carnegieendowment.org/les/2010russia_military_doctrine.pdf
Michael Aaronson et al., “NATO Countering the Hybrid Threat,” PRISM 2, no. 4 (2011): 111–24.
5
6
7
8
9
10
11
12
3
Cyber Governance and Digital Democracy 2019/1
Sub-threshold warfare strategy is a sub-strand within the
wider umbrella term of ‘hybrid warfare’, which seeks to
conduct confrontational and combative operations without
triggering the NATO Article 5 obligations or a direct military
retaliation by a NATO country13. This strategy builds upon the
late-Soviet strategy of ‘active measures’, which deployed a
combination of informatics and political framing mechanisms
to divert, distract and mislead institutions and agencies in
Western countries14. As outlined by former KGB Director of
Foreign Counterespionage Oleg Kalugin, ‘active measures’
worked by creating several layers of separation between
the perpetrating agency or gures, rendering the operation
virtually untraceable back to Moscow15. Following decades
of iterations, ‘active measures strategy’ has evolved into its
modern form - sub-threshold warfare – which denes the
sum of non-violent and obstructionist tactics of Russia’s
hybrid warfare operations within NATO countries.
Russia did not invent the sub-threshold warfare, however.
It is the Russian response to the American ‘offset strategy’,
which seeks to alter the balance of power in an unfavorable
standoff through creating a new standoff in a more favorable
contestation area16. The rst American offset strategy was
proclaimed during the early 1950s to deter the Soviet Union
through nuclear means, without spending excessively on
conventional forces. The second offset strategy was during
the 1975-89 period when the United States attempted to
pursue technological deterrence against the Warsaw Pact
to mask NATO’s comparative conventional disadvantage
in Eastern Europe. Finally, the third US offset strategy,
which Russia is currently challenging in direct terms, was
announced in 2014 to bolster US capabilities against anti-
access, area-denial (A2-AD) systems developed by Russia
and China17. This implied bolstering US cyber surveillance,
intelligence, digital media and stealth platforms to preserve
its informatics upper hand in Eastern Europe, especially
along the Russian border. From Russia’s point of view, such
US-origin measures targeted ethnic and religious fault lines in
former Soviet countries, to uproot pro-Russian governments
and leaders from power18. In the same vein, Russia’s sub-
threshold strategy is a mirror image of US offset strategies.
By using digital media and informatics tools, Russia seeks
to offset NATO’s technological and military strength, driving
wedges within and around NATO countries without triggering
their conventional defense mechanisms19.
Alexander Lanoszka, “Russian Hybrid Warfare and Extended Deterrence in Eastern Europe,” International Affairs 92, no. 1 (January 1, 2016): 175–95,
https://doi.org/10.1111/1468-2346.12509.
Steve Abrams, “Beyond Propaganda: Soviet Active Measures in Putin’s Russia,” Connections 15, no. 1 (2016): 5–31.
Evan Osnos, David Remnick, and Joshua Yaffa, “Trump, Putin, and the New Cold War,” February 24, 2017,
https://www.newyorker.com/magazine/2017/03/06/trump-putin-and-the-new-cold-war.
Daniel Fiott, “Europe and the Pentagon’s Third Offset Strategy,” The RUSI Journal 161, no. 1 (January 2, 2016): 26–31, https://doi.org/10.1080/03071847.2016.1152118.
Luis Simón, “The ‘Third’ US Offset Strategy and Europe’s ‘Anti-Access’ Challenge,” Journal of Strategic Studies 39, no. 3 (April 15, 2016): 417–45,
https://doi.org/10.1080/01402390.2016.1163260.
Daniel Fiott, “A Revolution Too Far? US Defence Innovation, Europe and NATO’s Military-Technological Gap,” Journal of Strategic Studies 40, no. 3 (April 16, 2017): 417–37,
https://doi.org/10.1080/01402390.2016.1176565.
Bettina Renz, “Russia and ‘hybrid Warfare,’” Contemporary Politics 22, no. 3 (July 2, 2016): 283–300, https://doi.org/10.1080/13569775.2016.1201316; Sascha Dov
Bachmann and Hakan Gunneriusson, “Russia’s Hybrid Warfare in the East: The Integral Nature of the Information Sphere,” SSRN Scholarly Paper (Rochester, NY: Social
Science Research Network, October 7, 2015), https://papers.ssrn.com/abstract=2670527.
13
14
15
16
17
18
19
4
Cyber Governance and Digital Democracy 2019/1
The highest-prole accusations of Russian meddling in the
West came after the 2016 US General Election. The US
Department of Homeland Security (DHS) and the Ofce of
the Director of National Intelligence (ODNI) have stated that
Russian President Vladimir Putin had personally ordered a
large-scale and high-level ‘inuence campaign’ to increase
the chances of Donald Trump’s victory in the election20. This
was done, according to US security agencies, through a
Russian military intelligence (GRU) led effort in hacking the
Democratic National Committee servers, as well as John
Podesta’s – the director of the Hillary Clinton campaign21 -
account. Later in January 2017, Director of the US National
Intelligence James Clapper asserted in a testimony that
Russia was also involved in a coordinated and state-led
‘fake news campaign’, disseminated across web-based
news and social media platforms. All of this led to the
well-known ‘Mueller investigation’, conducted by the US
Department of Justice Special Counsel Robert Mueller since
May 2017, exploring the extent to which Russia and pro-
Russian networks have been involved in the 2016 election22.
A major sub-thread of the investigation concerns the extent
of Russian digital media operations in the United States that
go beyond more direct attacks such as hacking.
In mid-December 2018, two major empirical studies were
submitted to the US Senate Intelligence Committee that
explored the measurable impact of Russian disinformation
operations in the US elections23. One of these reports,
conducted by the Oxford Internet Institute’s (OII)
Computational Propaganda Project, outlines the extent
to which the Russian Internet Research Agency (IRA) has
used targeted disinformation, bots and trolls to divide the US
public opinion into politically polarized interest groups for
targeted manipulation24. Specically, the report outlines how
Russian efforts pinpointed and exacerbated existing social,
racial and religious tensions and fears among the American
right-wing voters, and fed both inaccurate and fabricated
content to channel those tensions and grievances into pro-
Trump electoral behavior. According to the report, not only
did this targeted disinformation effort contribute substantially
to Donald Trump’s victory, but also continued to bolster his
digital popularity during contested decision phases of his
Presidency.
What is interesting from the researcher’s point of view, is that
Russian disinformation ecosystem in the US is extremely
easy to spot and map-out, as the OII report demonstrates
how 99% of all engagement related to the pro-Russian
and Russian content (likes, shares, retweets, comments)
originated from only 20 accounts on Twitter and Facebook,
all controlled by the IRA, containing account names such as
“Being Patriotic,” “Heart of Texas,” “Blacktivist” and “Army
of Jesus”. An overwhelming majority of these accounts
share links from explicitly Russian news websites such
as Russia Today, Sputnik and RIA Novosti, making web
domain tracking one of the most common analytical tools
to identify a pro-Russian network in a large data cluster.
In other words, Russia hasn’t spent much effort in trying to
conceal its digital inuence operations in the United States,
and most of these accounts still exist in the American
information ecosystem with different names, continuing to
shape opinion within the far-right information networks. In
analytical terms, all of these factors render the US one of the
easiest cases to study Russian disinformation, as inuence
networks and content are still very much ‘out there’ and
can be extracted through a very small sample. Even more
interesting from the researcher’s point of view, Russian
disinformation operations have become even more brazen,
direct and identiable after they were spotted by the Mueller
investigation, rendering their identication and network far
Major Cases of Disinformation and Countermeasures in the West
Karen Yourish and Troy Griggs, “8 U.S. Intelligence Groups Blame Russia for Meddling, but Trump Keeps Clouding the Picture,” The New York Times, July 16, 2018, sec. U.S.,
https://www.nytimes.com/interactive/2018/07/16/us/elections/russian-interference-statements-comments.html,
https://www.nytimes.com/interactive/2018/07/16/us/elections/russian-interference-statements-comments.html.
David E. Sanger and Charlie Savage, “U.S. Says Russia Directed Hacks to Inuence Elections,” The New York Times, December 21, 2017, sec. U.S.,
https://www.nytimes.com/2016/10/08/us/politics/us-formally-accuses-russia-of-stealing-dnc-emails.html.
Jason Breslow, “All The Criminal Charges To Emerge So Far From Robert Mueller’s Investigation,” NPR.org, December 9, 2018,
https://www.npr.org/2018/12/09/643444815/all-the-criminal-charges-to-emerge-so-far-from-robert-muellers-investigation.
Scott Shane, “Five Takeaways From New Reports on Russia’s Social Media Operations,” The New York Times, December 18, 2018, sec. U.S.,
https://www.nytimes.com/2018/12/17/us/politics/takeaways-russia-social-media-operations.html.
Philip N. Howard et al., “The IRA and Political Polarization in the United States, 2012-2018,” Computational Propaganda Research Project (Oxford, UK: Oxford Internet
Institute, December 2018), https://comprop.oii.ox.ac.uk/research/ira-political-polarization/.
20
21
22
23
24
5
Cyber Governance and Digital Democracy 2019/1
easier to map out25. As this research report will show, this
is quite 180-degrees the opposite of what we observed in
Turkey, where pro-Russian information network is elusive,
withdrawn and largely dormant, leading to excruciating
research difculties in mapping out the true extent of similar
information operations.
Given the inuence and success of Russian information
operations in the United States, most analysts have
turned to other NATO countries to see whether Russian-
afliated individuals or networks are involved in elections or
contested political episodes. Russian information footprint
is more visible in some European countries than others
and came in two major waves26. The rst wave, 1991-
2004, focused primarily on former Soviet states and Cold
War frontier countries to promote pro-Moscow political
candidates and shift the public debate into a form more
palatable to Kremlin27. As demonstrated in Way (2015),
where the anti-Russian candidate is a democrat, Russia
promoted more autocratic messaging and information
in digital media channels, whereas if the anti-Russian
candidate was authoritarian, Russian messaging promoted
pluralism, change, and democracy28. The rst wave also
had a generally low level of success. The second wave of
Russian information operations began in 2014, right after the
United States declared its Third Offset Strategy in November
2014, and continues until today. The second wave directly
targeted core NATO countries with two priorities in mind;
a) retaliate against NATO-led ethnic and religious division
strategies in and around Russia, by dividing Western
nations electorally and socially through information warfare,
without triggering Article #5, and b) to ride the wave of rising
far-right and left-wing populism to maximize the effect of
polarization operations29. Compared to the rst wave, the
second wave has been far more successful in terms of its
goal of incapacitating NATO countries’ collective defense
mechanisms and strategic coherence. Beginning with
2014, Russia has specically targeted political parties and
movements that contributed to the polarization in European
countries and also those that posed a more existential
criticism of the political system, rather than individual
political parties30.
In France, Kremlin has partnered with Front National and
has veriably conducted a range of digital warfare attempts,
including the hacking and leakage of Emmanuel Macron’s
campaign team data31. In the UK, pro-Russian accounts
were heavily involved in the Brexit referendum through cyber-
attacks, digital disinformation campaigns and targeted
political advertisements to steer the direction of the vote in
favor of the ‘Leave’ campaign32. In Germany, Russia has
been involved with phishing attacks against political parties
and campaigns that are pro-EU, including a 2015 hacking
of the German Bundestag, stealing 16 gigabytes of emails
(although these emails weren’t leaked)33. Similar digital
disinformation, phishing and campaign hacking cases are
Jane Mayer, “How Russia Helped Swing the Election for Trump,” The New Yorker, September 24, 2018,
https://www.newyorker.com/magazine/2018/10/01/how-russia-helped-to-swing-the-election-for-trump.
Naja Bentzen, “Foreign Inuence Operations in the EU - Think Tank” (Brussels: European Parliament, 2018),
http://www.europarl.europa.eu/thinktank/en/document.html?reference=EPRS_BRI%282018%29625123.
Charlotte Wagnsson and Maria Hellman, “Normative Power Europe Caving In? EU under Pressure of Russian Information Warfare,” JCMS: Journal of Common Market
Studies 56, no. 5 (2018): 1161–77, https://doi.org/10.1111/jcms.12726.
Lucan A. Way, “The Limits of Autocracy Promotion: The Case of Russia in the ‘near Abroad,’” European Journal of Political Research 54, no. 4 (2015): 691–706,
https://doi.org/10.1111/1475-6765.12092.
Peter Pomerantsev, “Authoritarianism Goes Global (II): The Kremlin’s Information War,” Journal of Democracy 26, no. 4 (October 19, 2015): 40–50,
https://doi.org/10.1353/jod.2015.0074.
Alina Polyakova, “Strange Bedfellows: Putin and Europe’s Far Right,” World Affairs 177, no. 3 (2014): 36–40.
Alex Hern, “Macron Hackers Linked to Russian-Afliated Group behind US Attack,” The Guardian, May 8, 2017, sec. World news,
https://www.theguardian.com/world/2017/may/08/macron-hackers-linked-to-russian-afliated-group-behind-us-attack.
Patrick Wintour, “Russian Bid to Inuence Brexit Vote Detailed in New US Senate Report,” The Guardian, January 10, 2018, sec. World news,
https://www.theguardian.com/world/2018/jan/10/russian-inuence-brexit-vote-detailed-us-senate-report.
Paul Carrel and Andrea Shalal, “Germany Says Its Government Computers Secure after ‘Isolated’ Hack,” Reuters, February 28, 2018,
https://www.reuters.com/article/us-germany-cyber-russia-idUSKCN1GC2HZ.
25
26
27
28
29
30
31
32
33
6
Cyber Governance and Digital Democracy 2019/1
also documented in Italy34, Netherlands35 and Sweden36.
By December 2018, the European Union has declared
a belated ‘War Against Disinformation’ in preparation for
renewed Russian meddling efforts in a series of upcoming
elections across Europe in 2019 (including the European
Parliament election)37. Some of the proposed Europe-wide
efforts are the establishment of a rapid warning system to
recognize, isolate and remove fake or manipulated digital
content, establishment of a new ‘digital contract’ with the
main social media platforms Twitter, Facebook, YouTube
and Instagram to ‘get serious’ about tackling disinformation
during key events and set up a European fact-checking
network of local veriers to spot disinformation attempts
in real-time. At the national level, countries have begun
formulating largely converging strategies to combat external
information operations.
France mobilized its intelligence agencies in the run-up to
its 2017 elections. The National Cybersecurity Agency of
France (ANSSI) has produced a cybersecurity handbook,
including a beginner’s introduction to DDoS (Distributed
Denial of Service) attacks, with follow-up briengs for all
political parties38. According to Carnegie Endowment, Marine
Le Pen’s Front National was the only party to be absent
from all of these briengs39. This strategy worked, because
2017 French elections became one of the best-documented
cases of failed Russian election meddling. Expecting a
hack, Emmanuel Macron’s campaign team hired an IT team
specializing in digital disinformation and generated a digital
le storage system that is designed to feed Russia its own
medicine: deliberately fabricated false campaign documents
to mitigate the value of ‘real’ documents that Russia may
have extracted through its hacks. Macron team also hired
three IT lawyers, each tasked with handling different aspects
of disinformation campaigns during the election. Le Monde
has published a list of news websites before the election day,
ranking them according to their reliability40. Also, 30 media
outlets in France partnered with Google to build a networked
fact-checking initiative called CrossCheck. All of this meant
that Russian meddling in French efforts failed, because
France launched a truly national, trans-partisan, inclusive
and heavily institutionalized framework to minimize the
damage caused by Russian digital information operations. It
also used a hybrid strategy of both autonomous, citizen-led
fact-checking efforts and strategically deployed deliberate
disinformation against Russian hackers.
Because the United Kingdom had already suffered
from Russian disinformation attempts during the Brexit
Referendum, its current strategy is structured to learn
from those mistakes retrospectively. Like France, Britain’s
National Cyber Security Center (NCSC) is bringing British
political parties together to brief them about the potential
vulnerabilities of digital information systems and how to
secure their networks against phishing attacks41. NCSC
also published technical primers for politicians on more
advanced topics in cybersecurity, disinformation and
digital leaks42. Because voting in Britain is complex due
to the absence of electronic voting and the responsibility
of organizing voting at the district level, Britain is relatively
more immune to direct digital election meddling. Rather,
Britain is more exposed to digital disinformation and opinion
Stephanie Kirchgaessner, “Russia Suspected over Hacking Attack on Italian Foreign Ministry,” The Guardian, February 10, 2017, sec. World news,
https://www.theguardian.com/world/2017/feb/10/russia-suspected-over-hacking-attack-on-italian-foreign-ministry.
Patrick Wintour and Andrew Roth, “Russia Summons Dutch Ambassador over Hacking Revelations,” The Guardian, October 8, 2018, sec. World news,
https://www.theguardian.com/world/2018/oct/08/russia-summons-dutch-ambassador-over-hacking-revelations.
Erik Brattberg and Tim Maurer, “How Sweden Is Preparing for Russia to Hack Its Election,” May 31, 2018, sec. World, https://www.bbc.com/news/world-44070469.
Daniel Boffey, “EU Raises Funds to Fight ‘Disinformation War’ with Russia,” The Guardian, December 5, 2018, sec. World news,
https://www.theguardian.com/world/2018/dec/05/eu-disinformation-war-russia-fake-news.
Heather A. Conley and Jean-Baptiste Jeangène Wilmer, “Successfully Countering Russian Electoral Interference,” CSIS Briefs (Washington DC: Center for Strategic and
International Studies, June 21, 2018), https://www.csis.org/analysis/successfully-countering-russian-electoral-interference.
Erik Brattberg and Tim Maurer, “Russian Election Interference: Europe’s Counter to Fake News and Cyber Attacks” (Washington DC: Carnegie Endowment for International
Peace, May 23, 2018), https://carnegieendowment.org/2018/05/23/russian-election-interference-europe-s-counter-to-fake-news-and-cyber-attacks-pub-76435.
Laura Daniels, “How Russia Hacked the French Election,” POLITICO, April 23, 2017, https://www.politico.eu/article/france-election-2017-russia-hacked-cyberattacks/.
Oscar Williams, “Russia Is Targeting UK Infrastructure through Supply Chains, NCSC Warns,” New Statesman, April 6, 2018,
https://tech.newstatesman.com/business/russia-uk-critical-infrastructure.
“UK Political Parties Warned of Russian Hacking Threat: Report,” Reuters, March 12, 2017,
https://www.reuters.com/article/us-britain-russia-cybercrime-idUSKBN16J0OE.
34
35
36
37
38
39
40
41
42
7
Cyber Governance and Digital Democracy 2019/1
manipulation attempts, than a direct hack. Thomas Rid’s
US Senate hearing demonstrated that both hacking and
disinformation operations from Russia and China during
the Brexit vote favored no particular candidate, but rather
sought to exacerbate existing divisions and polarization
over the refugee problem, immigration and the political
power balance between London and Brussels43. Like the
US, Russian-origin disinformation in the UK has been easy
to spot and map-out. University of Edinburgh researchers
have discovered that 3000 types of unique content can be
traced to the Russian Internet Research Agency, with around
150,000 unique accounts created by pro-Russian networks
to post specically on the Brexit referendum44. These
accounts were generally involved in anti-NATO and anti-EU
content dissemination and focused on the British far-right
audience through nationalist and isolationist content.
Germany on the other hand, has been targeted from
a multitude of vulnerabilities, including its Bundestag
network, Ministry of Finance digital accounts, Ministry of
Foreign Affairs records and the Christian Democratic Union
(CDU) party infrastructure. Like France, Germany has
been primarily targeted by the APT28 hacker team, a GRU
cyber-extension, as well as carefully curated and targeted
disinformation attempts45. A number of such disinformation
attempts involving the immigrants and immigrant-related
violence have become popular among German far-right
digital circles, perhaps the best-known case being the
‘Lisa’ story46. Even after the exposure of the ‘Lisa story’
as a disinformation case, Russian-origin disinformation
campaigns on immigration and integration still remain
popular among the German far-right47. To defend against
such attempts, political sides in Germany have entered
into a trans-partisan agreement before the September
2017 election to refrain from exploiting each other’s’ leaked
political data and signicantly limit the use of Twitter bots to
boost the spread and engagement of their online political
messages48. Facebook contributed to the training of the
political parties in securing digital infrastructure systems,
as well as how to deal with digital disinformation as fast
as possible, with the help of the voters and national fact-
checking initiatives49. German domestic intelligence agency
BfV and Federal Ofce for Information Security, BSI, both took
an active part in training political parties against a number of
vulnerabilities50. BSI even offered its cyber defense services
to all parties51. Like France, Germany also set up a large
network of citizen-led, autonomous fact-checking networks
to widen its disinformation defense capabilities.
Thomas Rid, “Disinformation: A Primer in Russian Active Measures and Inuence Campaigns,” Pub. L. No. 33017, § Select Committee on Intelligence (2017).
Robert Booth et al., “Russia Used Hundreds of Fake Accounts to Tweet about Brexit, Data Shows,” The Guardian, November 14, 2017, sec. World news,
https://www.theguardian.com/world/2017/nov/14/how-400-russia-run-fake-accounts-posted-bogus-brexit-tweets.
Carrel and Shalal, “Germany Says Its Government Computers Secure after ‘Isolated’ Hack.”
“The ‘Lisa Case’: Germany as a Target of Russian Disinformation,” NATO, 2016,
http://www.nato.int/docu/review/2016/Also-in-2016/lisa-case-germany-target-russian-disinformation/EN/index.htm.
Constanze Stelzenmüller, “The Impact of Russian Interference on Germany’s 2017 Elections,” Brookings (blog), June 28, 2017,
https://www.brookings.edu/testimonies/the-impact-of-russian-interference-on-germanys-2017-elections/.
Michael Schwirtz, “German Election Mystery: Why No Russian Meddling?,” The New York Times, January 20, 2018, sec. World,
https://www.nytimes.com/2017/09/21/world/europe/german-election-russia.html.
Sara Germano, “Facebook, Germany to Collaborate Against Election Interference,” Wall Street Journal, January 20, 2019, sec. Business,
https://www.wsj.com/articles/facebook-germany-to-collaborate-against-election-interference-11548004995.
Sumi Somaskanda, “The Cyber Threat To Germany’s Elections Is Very Real,” The Atlantic, September 20, 2017,
https://www.theatlantic.com/international/archive/2017/09/germany-merkel-putin-elections-cyber-hacking/540162/.
“German Cyber Defense Blends Military and Commerce,” Deutsche Welle, 09 2018,
https://www.dw.com/en/german-cyber-defense-blends-military-and-commerce/a-45636325.
43
44
45
46
47
48
49
50
51
8
Cyber Governance and Digital Democracy 2019/1
With such direct and easily identiable Russian meddling in
some of the most powerful nations of the West, researchers
have recently begun exploring how Russia-origin
disinformation efforts shape the political debate in the rest of
the world. Turkey is a natural case study. After all, Turkey has
been a Cold War buffer country, lies right at the intersection
of Western and Eastern security ecosystems and is adjacent
to three major civil wars – Iraq, Syria and Ukraine. It has long
been a strategically important country and lies adjacent to
some of the most problematic politically contested regions
of the Balkans, Caucasus, Middle East and North Africa.
According to International Telecommunication Union (ITU),
around 70% of the population has access to the Internet52
and according to the UK-based media analytics company
WeAreSocial, Turkey is one of the top countries in terms
of social media usage53. Despite restrictions, Turkey is still
one of the most active countries in terms of discussion and
dissemination of political information online54, and ranks
among the most active countries in terms of using social
media for political communication purposes55. Yet, it is also
one of the most vulnerable countries to disinformation, bot
usage and cyber-attacks56. Therefore, if there is an ideal
country to study the impact of disinformation on politics,
Turkey comes very close to that denition.
Yet, Turkey is also a difcult country to study in terms of
disinformation, because it is already plagued by high-levels
of fake news contamination57. The overall poor state of the
information environment in the country renders disinformation
a norm, not an exception, which makes it harder to isolate the
researched anomaly within a wider pool of other anomalies.
There is also the critical question of causality. Following the
popularized cases in the US, UK, France and Germany, more
countries have begun reporting cases of disinformation, even
when such cases have no measurable effect on any political
outcome58. This availability bias is plaguing the eld with an
enormous volume of contextually irrelevant cases of fake
news that don’t spread beyond a very small network and/or
have no political or social implication59. More importantly, an
excessive focus on Russian disinformation generate myopia
that overlooks the agency of domestic players in making their
countries vulnerable to disinformation in general60. Today, it
is possible to locate Russian disinformation in a large number
of countries, although Russia is by no means the only player
that weaponizes digital information or systematically hacks
international political actors. This creates an availability bias
in the eld, which distorts the extent to which Russia matters
in digital space and substantially downplays the agency of
the pre-existing political and media actors and institutions,
analytically reducing them into non-entities.
This is a problematic way to approach disinformation,
because its conceptual root – propaganda – is certainly not
new to political communication. States have been deploying
propaganda for centuries through the most recent and
widespread media outlets they could access in any given
era. Digital media is simply another step in the long history
of propaganda and political diversion, spread in the past
Charting the Pro-Russian Information Ecosystem in Turkey
“Measuring the Information Society Report 2018,” International Telecommunications Union, 2018, https://www.itu.int/en/ITU-D/Statistics/Pages/publications/misr2018.aspx.
“Digital in 2018: World’s Internet Users Pass the 4 Billion Mark,” We Are Social, January 30, 2018, https://wearesocial.com/blog/2018/01/global-digital-report-2018.
Erkan Saka, “Social Media in Turkey as a Space for Political Battles: AKTrolls and Other Politically Motivated Trolling,” Middle East Critique 27, no. 2 (April 3, 2018): 161–77,
https://doi.org/10.1080/19436149.2018.1439271.
“Reuters Institute Digital News Report” (Oxford, UK: Reuters Institute, Oxford University, 2018), http://www.digitalnewsreport.org/.
Barçın Yinanç, “Poor Media Literacy ‘making Turks Vulnerable to Fake News’ - Turkey News,” Hürriyet Daily News, December 10, 2018,
http://www.hurriyetdailynews.com/poor-media-literacy-making-turks-vulnerable-to-fake-news-139582.
Mark Lowen, “Hunting for Truth in a Land of Conspiracy,” BBC News, November 15, 2018, sec. Europe, https://www.bbc.com/news/world-europe-46137139.
Herman Wasserman, “Fake News from Africa: Panics, Politics and Paradigms,” Journalism, December 17, 2017, 1464884917746861,
https://doi.org/10.1177/1464884917746861; Marco Visentin, Gabriele Pizzi, and Marco Pichierri, “Fake News, Real Problems for Brands: The Impact of Content Truthfulness
and Source Credibility on Consumers’ Behavioral Intentions toward the Advertised Brands,” Journal of Interactive Marketing 45 (February 1, 2019): 99–112,
https://doi.org/10.1016/j.intmar.2018.09.001; Edson C. Tandoc Jr, Zheng Wei Lim, and Richard Ling, “Dening ‘Fake News,’” Digital Journalism 6, no. 2 (February 7, 2018):
137–53, https://doi.org/10.1080/21670811.2017.1360143.
Dominic Spohr, “Fake News and Ideological Polarization: Filter Bubbles and Selective Exposure on Social Media,” Business Information Review 34, no. 3 (September 1,
2017): 150–60, https://doi.org/10.1177/0266382117722446; Tim Groeling, “Media Bias by the Numbers: Challenges and Opportunities in the Empirical Study of Partisan
News,” Annual Review of Political Science 16, no. 1 (2013): 129–51, https://doi.org/10.1146/annurev-polisci-040811-115123.
Samantha Bradshaw and Philip N. Howard, “The Global Organization of Social Media Disinformation Campaigns,” Journal of International Affairs 71, no. 1.5 (2018): 23–32.
52
53
54
55
56
57
58
59
60
9
Cyber Governance and Digital Democracy 2019/1
through word of mouth, sealed envelopes, telegram, radio,
television and motion picture. Success of propaganda
through all of these periods relied primarily on two factors:
the perpetrator’s understanding of the media system, and
the perpetrator’s successful diagnosis of the social divisions
in a target country. Propaganda therefore, always depends
on how well its wielder understands the social impact of
the delivery mechanism, be it telegram, motion picture or
Twitter, and how well it understands what makes its audience
tick. From this perspective, a propagandist’s success relies
largely on the pre-existing information environment and the
extent of grievances within the audience.
In light of the growing criticism in the disinformation literature
against studies that merely state that ‘disinformation exists’
or focus solely on Russian disinformation diffusion without
any political and media environment context, this report
aims to go one step further. Rather than only looking at
whether there is Russian disinformation in Turkey (there
is), this study seeks to explore whether Russian information
operations in Turkey matter, and have any inuence on
Turkey’s wider information landscape. Do they inuence
any mainstream conversation and have a measurable
effect such as polarization and/or shifting election results,
or do they merely exist in isolation, without any substantial
engagement and relevance?
To do that, this study dissects 3 of the most important events
in recent Turkish-Russian bilateral relations (downing of the
SU24 jet, assassination of the Russian ambassador and the
S400 negotiations) and arguably 2 of the most important
domestic events in recent Turkish politics (2016 coup attempt
and 24 June 2018 general elections) and aims to measure
the impact of Russian-origin information campaigns against
the wider Turkish information network in these most critical
episodes. This method yields a more accurate perspective
to evaluate the severity of Russian information efforts in
Turkey compared to isolating low-engagement or marginally
relevant content types that have no measurable effect on the
general information network.
Mainstream pro-Russian information ecosystem in Turkey is
fairly straightforward. Russia’s primary mainstream Turkish
language media outlet is Sputnik-Türkiye, which was one
of the rst foreign language branches of the agency after
it replaced all previous Russian foreign-language services
in November 2014. Around the same time, Rusya’nın Sesi –
the primary pro-Russian news radio in Turkey – was named
RSFM and continued under the aegis of the Sputnik News
Agency. Aydınlık is a well-known pro-Russian outlet, that
regularly reports events directly related to Turkish-Russian
relations with an emphasis on the Russian view. There is
also a pro-Turkish-Russian relations website called TurkRus.
com, run by journalist Suat Taşpınar, where news and
opinion on Turkish-Russian relations are shared frequently.
Outside Sputnik-Türkiye, RSFM and Aydınlık however,
what constitutes a pro-Russian outlet in Turkey is a highly
contested and often a context-specic designation. Being
considered ‘pro’ any foreign country, be it Russia, United
States or otherwise, is generally considered libelous in
Turkey. To that end, being called pro-Russian, like being
called pro-American, is usually an external allegation to
defame an actor and is also usually denied by the target(s).
This complicates an analyst’s job, especially when the study
is empirical, because the outlining the ‘pro-Russian media
environment’ becomes a moving target. Regardless, and
as demonstrated in this research report, all shades of the
Turkish media spectrum have published news reports and
analyses that could be considered as ‘pro-Russian’ or ‘close
to Putin’s position’, under different contexts and content. In
other words, the mainstream media in Turkey measurably
shifts into a pro-Russian narrative on issues directly related
to Turkish-Russian relations.
Prominent studies61 that explore Russian disinformation in
the US or EU have so far identied ‘pro-Russian’ accounts in
three ways: a) URL/domain root tracking to see if they lead to
Sputnik, RT or other Russian-language media or website link,
b) location information of the account(s) involved (most US
and UK disinformation studies traced pro-Russian networks
based on their stated physical location within Russian
territory), and c) dominant language of the text shared by
the account (whether it is mostly in Russian). A combination
of these methods usually yield a fairly reliable network of
Russian inuence actors in an information ecosystem. The
rst method can be applied to the Turkish case, as pro-
Savvas Zannettou et al., “Disinformation Warfare: Understanding State-Sponsored Trolls on Twitter and Their Inuence on the Web,” arXiv:1801.09288 [Cs], January 28, 2018,
http://arxiv.org/abs/1801.09288; Michael Jensen, “Russian Trolls and Fake News: Information or Identity Logics?,” Journal of International Affairs 71, no. 1.5 (2018): 115–24;
Yevgeniy Golovchenko, Mareike Hartmann, and Rebecca Adler-Nissen, “State, Media and Civil Society in the Information Warfare over Ukraine: Citizen Curators of Digital
Disinformation,” International Affairs 94, no. 5 (September 1, 2018): 975–94, https://doi.org/10.1093/ia/iiy148; Denis Stukal et al., “Detecting Bots on Russian Political Twitter,”
Big Data 5, no. 4 (December 1, 2017): 310–24, https://doi.org/10.1089/big.2017.0038; Emilio Ferrara et al., “The Rise of Social Bots,” Communications ACM 59, no. 7 (June
2016): 96–104, https://doi.org/10.1145/2818717.
61
10
Cyber Governance and Digital Democracy 2019/1
Russian opinion actors in Turkey do share Sputnik or Russia
Today news frequently. The other two however, are not very
reliable in the Turkish context, because they can be easily
faked. Like the well-known attribution problem in cyber-
attacks (i.e. who really conducted the attack), location and
language-based attribution of information operations can
be masked, leading to false attribution. Furthermore, in our
study, less than .0001% of the content (14 tweets only out
of 183 million) was sent by accounts that contains location
information in Russia, or had Russian-language text in their
content data.
Further complications arise from separating a pro-Russian
media outlet from an anti-Western media outlet. A range of
Turkish outlets have been designated as ‘pro-Russian’ due
to their anti-NATO and anti-EU reporting bias, although some
of these outlets sometimes reject this designation, dening
themselves as ‘nationalist’. From an analyst’s perspective it
becomes further difcult to separate NATO, EU and Russia
news reports of a nationalist or an Islamist outlet, given both
forms of reporting are heavily against Western institutions
and report Turkish-NATO and Turkish-EU cooperation in
overwhelmingly critical terms. This doesn’t necessarily make
them pro-Russian, as most of those outlets are also often
critical of Moscow. Even further complicating the picture,
and as demonstrated empirically in this research report,
in the last 3 years, most mainstream, high-circulation pro-
government and opposition news outlets alike have begun,
at different times, reporting content that is aligned with
Russia’s position. As the mainstream media environment
shifts gradually into a more pro-Russian tone, identifying pro-
Russian outlets and networks become even harder. The only
exceptions, of course are Sputnik-Türkiye, RSFM, because
they are explicitly Russian-owned and funded. The closest
domestic political outlet that comes closest to being an
indigenous pro-Russian player is Aydınlık, given its frequent
explicit support for Russian policy. However, in light of the
new data presented in this study, it is more accurate to track
‘pro-Russian content’, rather than an outright ‘pro-Russian
outlet’, as the former can be traced across all shades of the
Turkish media environment, whereas the latter becomes a
subjective designation, open to debate.
From a methodological standpoint, all of this makes a truly
objective computational, large-volume study on Russian
information operations in Turkey tricky. Instead of following
the best-known mainstream studies that focus on Russian
disinformation in the West through focusing on follower
networks and most commonly shared news domains, we
had to go a step further. In this study, we use the follower
network and web URL tracing approaches, while adding
a third dimension: sentiment. We have trained our topic
modelling algorithm in Turkish political text to identify
positive and negative sentiment digital content related to
events and topics on Russian-Turkish relations. This three-
layered strategy was necessary, as Russian information
domain is not nearly as explicit, straightforward and brazen
as we observe in other Western cases.
Data source. The data used in this study is generated via
the Twitter streaming API. Our research group has built a
crawler that live-scrapes all tweets and their metadata since
24 November 2015 (SU-24 downing incident) that contain
the n-grams ‘Rus_’, ‘Putin_’ and ‘Moskova_’. These n-grams
are derived from the keyword analysis tool KWFinder and
topic-wise allow us to extract 99.998% of all Twitter content
related to Turkish-Russian relations in Turkish-language.
We don’t call our approach ‘stemming’, which is mostly
understood as an acronym for n-grams. Stemming works
better with inecting languages like English, whereas has
a low reliability in agglutinative languages like Turkish. In
the latter, n-gram approach yielded a more reliable result
compared to ‘stemming’. We primarily use Twitter in this
study, with periodic robustness checks on Facebook. In
comparison to our scraping effort on public accounts on
Facebook, which yielded a total of 1,163,856 words over 3
years, Twitter gave us 486,052,996 words in total through the
same time period, rendering Twitter a far better venue for this
type of research. In our study on Turkish elections, we have
pre-designated 6 of the most-shared proven disinformation
cases that emerged during the election period. To explore
disinformation during the failed 2016 coup attempt, we
similarly analyzed 5 of the most prominent cases of fake
news. These disinformation types have been exposed by
Teyit.Org, a major Turkish fact-checking initiative.
Case selection. Our cases are selected due to their
signicant digital popularity compared to other cases in
Turkish-Russian relations, and their political impact on
bilateral relations. 24 June elections were picked, as these
were arguably the most important political election in
Methodology
11
Cyber Governance and Digital Democracy 2019/1
Turkey, with the higher stakes and acute rivalry compared
to previous elections. Most importantly, 24 June elections
inaugurated the new political system in Turkey, and has
a regime-shift component. The importance of 24 June
elections is evidenced by their popularity on social media.
Selected Cases Percentage of Dirty Data Clean Data
S400 negotiations
(5 benchmarks – longitudinal) 3.937 44,394,129
2016 Coup attempt 21.593 27,459,214
Assassination of the Ambassador 11.353 18,667,492
SU24 downing incident 9.825 13,560,108
Discarded Cases Percentage of Dirty Data Clean Data
Russia’s annexation of Crimea (Mar 2014) 15.024 1,901,403
TurkStream Negotiations + Signing (longitudinal) 2.079 871,031
White Helmets (longitudinal) 46.492 139,059
Akkuyu Nuclear Power Plant (longitudinal) 1.291 84,493
Selahattin Demirtaş visit to Moscow (Dec. 2015) 11.938 37,381
Table 1 - Selected and discarded cases based on the proportion of cleaned data and post-cleaning data size
Selected Robustness Check Percentage of Dirty Data Clean Data
24 June 2018 Elections (Presidential + General) 17.884 79,823,491
Discarded Robustness Check Alternatives Percentage of Dirty Data Clean Data
2017 referendum 21.091 24,043,032
1 November 2015 Elections (General) 14.726 18,129,491
June 2015 Elections (General) 11.958 14,939,251
2014 Presidential elections 8.083 7,249,219
2014 local elections 6.179 3,140,402
Table 2 - Selected robustness check case and discarded alternatives based on the proportion of cleaned data
and post-cleaning data size
12
Cyber Governance and Digital Democracy 2019/1
Data cleaning. In this study, we clear out a type of bot-driven
content commonly observed in Turkey: gibberish, randomly-
selected words that don’t contain any meaning. This type
of automated content is easy to automatically detect and
remove because The First Letters Of All Words In Such
Tweets Are Capitalized. Often, bots that have nothing to
do with Turkish-Russian relations (such as sports-related
or advertiser bots) are programmed to auto-select n-grams
based on their popularity, with no relation to any political
incident. This results in quite a large volume of meaningless
data that skews the results of Turkey-based disinformation
research, and thus, have to be cleaned. We don’t clean out
tweets that are bot-driven, but have a meaningful sentence
structure. The total number of tweets we study in this report
is 183,904,434 post-cleaning. Percentages of dropped
tweets during the cleaning phase in each case are listed in
Tables 1 and 2.
Bot designation. We use a bot-detection algorithm that auto-
identies suspected bot accounts by employing four of the
most established methods in the methodological literature:
a) friend-to-follower ratio (Wang et. al., 201062), number of
tweets (Howard and Kollanyi, 201663), account creation
date (Jones, 201764) and text duplicates (Thieltges et. al.
201865). Only if an account posts content that meets one of
these criteria, we do a second robustness check through a)
the use of URL shorteners (a main indicator of automation,
because URL shorteners, like trib.al, bit.ly or tinyurl.com,
track trafc to a link), b) its tweet history of using more than
3 languages (bot accounts post automated messages in
an average of 5-6 different languages to push a particular
narrative in multiple linguistic and time-zone domains), c)
follower-to-like ratio (an account that contains too many
posts with disproportionate likes and retweets compared to
its follower count, there is a high likelihood that it is a bot).
Automated Content/Sentiment Analysis. For automated text
analysis we use ReadMe: Software for Automated Content
Analysis66, CEM: Coarsened Exact Matching Software and
WhatIF: Software for Evaluating Counterfactuals, all of which
we trained with Turkish-language ofcial document, media
text and news website comments text specically related
to Turkish-Russian relations and Turkish foreign policy over
the course of 6 months. Through random selection semi-
supervision tests, this heavy focus on texts related specically
Turkish-Russian political relations has allowed us to reach
98% reliability in detecting Turkish-language sentiment
(including sarcasm) correctly67. In addition to specialized
learning through focused text, we believe that this high level
of reliability also owes to the linguistic properties of Turkish
as an agglutinative language as opposed to English as an
inecting language.
Latent Dirichlet Allocation: LDA is a topic model type,
which is a statistical text-mining method that auto-discovers
relevant topic clusters in a large body of text. Its relevance
algorithm is driven by Dirichlet distributions that are built as
‘topic-per-document’ and ‘words-per-topic’ classication.
Our pre-processing routine includes tokenization, removing
words fewer than 3 words, removing stopwords, lemmatizing
words and reducing n-grams to their root form. In this study,
we use the LDA approach dened by McCallum (et. al.
200768).
Alex Hai Wang, “Detecting Spam Bots in Online Social Networking Sites: A Machine Learning Approach,” in Data and Applications Security and Privacy XXIV, ed. Sara Foresti
and Sushil Jajodia, Lecture Notes in Computer Science (Springer Berlin Heidelberg, 2010), 335–42.
Philip N. Howard and Bence Kollanyi, “Bots, #Strongerin, and #Brexit: Computational Propaganda During the UK-EU Referendum,” SSRN Scholarly Paper (Rochester, NY:
Social Science Research Network, June 20, 2016), https://papers.ssrn.com/abstract=2798311.
Marc Owen Jones, “Hacking, Bots and Information Wars in the Qatar Spat,” Washington Post, June 7, 2017,
https://www.washingtonpost.com/news/monkey-cage/wp/2017/06/07/hacking-bots-and-information-wars-in-the-qatar-spat/.
Andree Thieltges et al., “Effects of Social Bots in the Iran-Debate on Twitter,” arXiv:1805.10105 [Cs], May 25, 2018, http://arxiv.org/abs/1805.10105.
62
63
64
65
Daniel J. Hopkins and Gary King, “A Method of Automated Nonparametric Content Analysis for Social Science,” American Journal of Political Science 54, no. 1 (2010): 229–47,
https://doi.org/10.1111/j.1540-5907.2009.00428.x.
Peng Liu et al., “Sarcasm Detection in Social Media Based on Imbalanced Classication,” in Web-Age Information Management, ed. Feifei Li et al., Lecture Notes in Computer
Science (Springer International Publishing, 2014), 459–71.
A. McCallum, X. Wei, and X. Wang, “Topical N-Grams: Phrase and Topic Discovery, with an Application to Information Retrieval,” in Seventh IEEE International Conference
on Data Mining (ICDM 2007)(ICDM), 2007, 697–702, https://doi.org/10.1109/ICDM.2007.86.
66
67
68
13
Cyber Governance and Digital Democracy 2019/1
Figure 1 - N-gram text discovery and topic model approach used in this research report, as described in McCallum et. al.
(2007): “The graphical model presentation of this model is shown in Figure 1(c). Its generative process can be described as
follows: 1. draw Discrete distributions φz from a Dirichlet prior β for each topic z; 2. draw Bernoulli distributions ψzw from a Beta
prior γ for each topic z and each word w; 3. draw Discrete distributions σzw from a Dirichlet prior δ for each topic z and each
word w; for each document d, draw a Discrete distribution θ(d) from a Dirichlet prior α; then for each word wi
(d) in document d:
(a) draw xi
(d) from Bernoulli ψz (d) i−1w (d) i−1 ; (b) draw zi
(d) from Discrete θ(d) ; and (c) draw wi
(d) from Discrete σz (d) i w (d) i−1 if xi
(d) = 1;
else draw wi
(d) from Discrete φzi
(d)
.”
14
Cyber Governance and Digital Democracy 2019/1
The 24 November 2015 downing of the Russian SU-24 in Syria
is a major ashpoint between Turkey and Russia and offers
a good case study for longitudinal digital misinformation
research. Predicting that the incident would change Turkish-
Russian relations to a great extent, our research cluster has
begun scraping social media data soon after the news broke
out. Initially, we scraped Twitter, Facebook and web pages,
but given the vast analytical size of the former, we decided
to go with Twitter only. Ultimately, the decision to begin
scraping relevant data that date became an auspicious
one, as we are still continuing with the same corpus and
keywords that have been building up since then.
The Russian Sukhoi-24 was shot down by a Turkish F-16
after the Russian aircraft violated the Turkish airspace by
about 2.19 kilometers. Two Russian pilots had ejected,
but the main pilot was shot down and killed by a band of
Syrian rebels on the ground, while Russian troops ultimately
rescued the second pilot. Russia contested Turkey’s
decision to shoot down the SU-24, asserting at the highest
level that the craft remained within Syrian territory and never
strayed into Turkish airspace. Turkey, on the other hand,
released a number of audio recordings, including the radio
communications between the Turkish air base and the F-16.
In addition, Ankara argued that the Russian SU-24 was
unidentied (meaning its electronic radar identication was
marked as ‘unknown’ craft) and the airbase command thought
the jets belonged to the Syrian government. The contested
narratives between the two sides continued for months
and could only be diffused following intense diplomatic
cushioning. The event was also highly internationalized as
all NATO countries, including the United States, entered the
fray through diplomatic de-escalation moves.
Initial hours of the incident on social media are marked
mostly by objective statements, with a substantial presence
of retweets from main news outlets, rather than organic
new content. The severity of the crisis was generating an
unexpectedly cautious response from the digital audiences
on both sides; rather than making assertive statements,
content focused mainly on the technicalities: how far
the SU24 penetrated Turkish air space, what were the
engagement rules and the exact location of the shooting
down. It is after Day-1 that the rst cases of sentiment-
relevant content start to emerge. Our group fed the collected
tweets into both ReadME and SentiStrength – two popular
text analysis tools. Ultimately, we decided to go ahead with
semi-supervised ReadME as it provided greater reliability in
Turkish-language political text. Eventually, our algorithm has
discovered two main topic clusters, that correspond to two
narrative ecosystems: one pro-Russian and one pro-Turkish:
Narrative-A: the topic cluster of tweets that blame Russia for
violating Turkish airspace,
Narrative-B: the topic cluster of tweets that blame Turkey for
shooting down the jet outside of Turkish airspace.
Narrative-A was disseminated quite centrally by pro-
government news outlets, closely supported by government
members’ accounts, ultimately reaching hegemonic status
within the Turkish Twitter ecosystem by the end of Day-
1. This narrative orbited around legalistic arguments and
emphasis on international law over airspace violations, and
asserted that the decision was correct. Narrative-B on the
other hand primarily relied on Sputnik Türkiye, Aydınlık, and
RT-domain content, yet received less support in the Turkish
Twitter ecosystem. Narrative-B followed Russian defense
establishment view that the SU24 was shot down outside
Turkish airspace, while it was still cruising within Syrian
territory. Furthermore, Narrative-B contained early radar
screen images disseminated by the Russian high command
to support the claim that the jet was shot down within Syrian
airspace. However, Narrative-A quickly monopolized the
frame and the narrative in Turkish information ecosystem,
and Narrative-B lingered on with diminishing popularity and
eventually got marginalized in the ecosystem by Day-8.
Case-1: SU24 Downing Incident
15
Cyber Governance and Digital Democracy 2019/1
Narrative-A Top Keywords Occurrence Narrative-B Top Keywords Occurrence
Uça_ 4,180,868 Uça_ 1,628,426
Rus_ 4,012,673 Rus_ 1,574,634
Savaş 3,963,145 Suriye 1,506,418
Sınır_ 3,461,369 SU24 1,469,765
İhlal 3,405,951 Dışı_ 1,369,748
Havasaha_ 3,208,413 Havasaha_ 1,267,932
Türk_ 2,794,486 Gir_ 1,134,662
Suriye 2,408,018 Devriye 1,068,484
Taraf_ 2,097,666 Sınırın_ 1,097,255
Savunma 1,468,234 F-16 1,031,320
Düşür_ 1,168,989 Savunma 947,824
Alçak_ 896,725 Moskov_ 864,768
Haber_ 711,266 Düşür_ 845,250
Sputnik_TR 702,431 Düştü 764,318
İçerisinde 626,482 RT 745,326
Güneyinde 554,936 Sputnik_TR 741,255
RIA 524,864 Dışarı_ 645,262
F-16 464,629 Kuzey_ 468,598
Table 3 - Most frequently occurring n-grams in LDA-designated Narrative-A and Narrative-B content and engagement clusters
Table 3 - Top domains included in tweets from LDA-designated Narrative-A and Narrative-B*
Narrative-A Domain Diffusion Share % Narrative-B Domain Diffusion Share %
Sabah.com.tr 17.593 tr.sputniknews.com 88.931
Yenisafak.com.tr 16.394 aydinlik.com.tr 6.394
Hurriyet.com.tr 16.014 Rt.com 4.675
HaberTurk.com.tr 13.931
Milliyet.com.tr 11.293
aa.com.tr 9.429
Cnnturk.com.tr 6.291
Ntv.com.tr 4.485
Ahaber.com.tr 4.570
Domain popularity indicates how much users prefer to share content that contains these URLs. It doesn’t mean that these accounts publish more content. It is a sign of how
frequently that domain is referred to, not how much that domain publishes on a given topic. This is an important disclaimer that is also relevant for other domain popularity tables
in this report.
*
16
Cyber Governance and Digital Democracy 2019/1
Figure 2 - 24-hour time-series analysis of LDA-designated Narrative-A and Narrative-B engagement metrics,
sorted by organic/bot and authentic/retweet designation. Count value in 000s.
Figure 3 - 7-day time-series analysis of LDA-designated Narrative-A and Narrative-B engagement metrics,
sorted by organic/bot and authentic/retweet designation. Count value in 000s.
17
Cyber Governance and Digital Democracy 2019/1
After Day-8 (2 December 2015) however, a new pro-
Russian line emerges. This line [Narrative-C] mounted a
centralized challenge against the Turkish government for
aiding and supporting the Islamic State (ISIS), specically
by purchasing crude oil from the areas controlled by
the group via tankers69. This new line was also heavily
disseminated through Sputnik-Türkiye and a number of
Turkish-language anonymous accounts and became
more successful in terms of its spread and staying power,
compared to Russia’s early narrative that the SU-24 was
shot down outside Turkish airspace. In all of our cases, this
is the most successful, high-impact and lasting Russian
information operation, and arguably, Russia’s success
with this narrative has put Turkey on the defensive from a
digital standpoint. This new pro-Russian criticism of Turkey
suggests heavy centralized message control owing to the
Russian Defense Ministry statement on 2 December 2015,
that it had satellite images offering proof that Turkish tankers
were involved in oil smuggling from ISIS-controlled territory70.
Beginning with January 2016, this narrative substantially
proliferated in the Turkish Twitter ecosystem, and becomes
a more dominant line of argument, compared to Turkey’s
narrative-A (that Russian jet was shot down within Turkish
airspace). Furthermore, the Russian line was frequently
retweeted and shared by Western media outlets as well71,
perhaps as an expression of frustration towards Turkey in
general, or its Syria policy in particular. Regardless, pro-
Russian information operations on Turkey-ISIS link went
beyond Turkish information ecosystem and became truly
global, with substantial engagement in all NATO countries.
In terms of diplomatic messaging and signaling, Russia has
demonstrated to Ankara that it could easily draw a wedge
between Turkey and its Western allies and create a very
large rift within NATO if it tried. Arguably, the success of
Narrative-C has been one of the digital drivers, along with
more immediate physical drivers related to Syria, of Turkey’s
growing distancing from NATO and its slide into the Russian
orbit. The main evidence on this can be found with the other
cases studied in this report: after Narrative-B, Turkey never
ofcially challenged a Russian digital narrative on social
media in a signicant way.
“Russia Presents Proof of Turkey’s Role in ISIS Oil Trade,” RT International, December 2, 2015, https://www.rt.com/news/324263-russia-brieng-isis-funding/.
Ben Taub, “The ISIS Oil Trade, from the Ground Up,” December 4, 2015, https://www.newyorker.com/news/news-desk/the-isis-oil-trade-from-the-ground-up.
Tom Brooks-Pollock, “Russia Releases ‘Proof’ Turkey Is Smuggling Isis Oil over Its Border,” The Independent, December 2, 2015,
http://www.independent.co.uk/news/world/europe/russia-releases-proof-turkey-is-smuggling-isis-oil-over-its-border-a6757651.html;
Maria Tsvetkova and Lidia Kelly, “Russia Says It Has Proof Turkey Involved in Islamic State Oil Trade,” Reuters, December 2, 2015,
https://www.reuters.com/article/us-mideast-crisis-russia-turkey-idUSKBN0TL19S20151202; Greg Botelho, “Russia, Turkey Trade Charges: Who Bought Oil from ISIS?,”
CNN, December 2, 2015, https://edition.cnn.com/2015/12/02/europe/syria-turkey-russia-warplane-tensions/index.html.
69
70
71
Narrative-C Occurrence Narrative-D Occurrence
Türk_ 6,216,826 FET_ 4,136,193
IŞİD 5,120,631 Rus 3,849,679
Petrol 4,597,623 Cemaa_ 3,426,716
Yasadışı 2,236,641 Uçağı_ 3,201,329
Tanker 1,619,374 NATO 2,797,634
Suriye 1,396,417 Talimat_ 2,643,418
Erdoğan 946,674 Tarafı_ 2,325,824
Militan_ 863,219 ABD 1,946,264
Ticareti_ 634,546 Güle_ 1,245,357
Görüntü_ 422,279 Pilot_ 904,526
Sınır_ 316,748 Moskov_ 843,145
Haftada 201,367 Gizli 751,422
Bakanl_ 102,267 Emir 526,214
Table 5 - Most frequently occurring n-grams in Narrative-C and Narrative-D content and engagement cluster
18
Cyber Governance and Digital Democracy 2019/1
The ‘ISIS oil’ information strategy has been demonstrably
successful from the Russian perspective. Although the
information ecosystem in Turkey still discussed and
debated that Ankara was correct in deciding to shoot down
the Russian jet, the pro-Russian ecosystem has managed to
widen the information battlefront and distract substantially
from Ankara’s main argument. Through 2 December 2015
till 6 August 2016 meeting between Presidents Erdoğan and
Putin (around 8 months), the oil smuggling narrative has
dominated the online agenda on Turkey-Russia relations
and reframed the whole episode into Turkey’s ‘oil trade with
ISIS’. Nothing Ankara did, either diplomatically or public
relations-wise was enough to counter Russian narrative. As
mentioned earlier, Russia was able to hack into Turkey’s
relations with NATO through this narrative as NATO country
media outlets became some of the most prominent central
hubs in the dissemination of this narrative. Even today, ‘ISIS
oil’ narrative is shared widely through US-based far-right
conspiracy theory outlets like Breitbart72 and InfoWars73. This
narrative ended quite sharply, however, following the rst-
ever meeting between Presidents Recep Tayyip Erdoğan
and Vladimir Putin in St. Petersburg on 9 August 2016. Even
before that, the failed coup attempt of 15 July 2016 seems
to have made a mark on reducing Russia’s weight behind
Narrative-C.
The Turkish side has responded to Russia’s Narrative-C, by
both retaining its insistence on Narrative-A, and diverting
gradually into a new narrative. Narrative-D asserted that the
jet shooting was ordered by a military chain of command loyal
to the exiled cleric Fethullah Gülen74. It is difcult to assess
why this shift happened by purely computational tools.
However, it is likely that this shift was a product of Turkey’s
realization that it didn’t have much leverage against Moscow
in strategic terms and had to de-escalate the situation, even
though Turkish version of events ultimately proved to be
correct and the allegations of Turkey’s ‘oil purchases from
ISIS’ have been played down by the US State Department
itself75. Regardless, Russia’s ‘ISIS oil’ information campaign
had become too hard to challenge for Turkey, given its
domestic and international diffusion rate, necessitating a
diversion, rather than a direct challenge. According to the
new Narrative-D, the decision to shoot down the Russian
jet did not follow the regular chain of command but was
deliberately put into play by a clandestine pro-Gülen and/
or NATO military network. This line of reasoning shifts into
two different but not mutually exclusive sub-narratives: that
a) the order was given by a pro-Gulen commander, b) the
decision was made autonomously by a pro-Gulen pilot, and
c) decision was made by a much higher ranking general
Narrative-C Domain Diffusion Share % Narrative-D Domain Diffusion Share %
tr.sputniknews.com 47.493 Sabah.com.tr 22.193
Cumhuriyet.com.tr 12.328 Yenisafak.com.tr 20.594
Odatv.com 11.387 Ahaber.com.tr 17.603
Aydinlik.com.tr 9.395 Sozcu.com.tr 13.491
Evrensel.net 7.203 Haberler.com 10.839
Diken.com.tr 6.382 Aksam.com.tr 9.394
Miscellaneous blogs/websites 5.812 Ahaber.com.tr 5.886
Table 6 - Top domains within Narrative-C and Narrative-D content and engagement clusters
Edwin Mora, “Report: Turkey, Syria Helped Keep Islamic State Alive by Buying Their Oil,” Breitbart, July 3, 2018,
https://www.breitbart.com/national-security/2018/07/03/report-turkey-syria-helped-keep-islamic-state-alive-by-buying-their-oil/.
Michael Snyder, “Obama Knows That Turkey Is Buying Oil From ISIS And He Isn’t Doing Anything To Stop It,” InfoWars (blog), November 28, 2015,
https://www.infowars.com/obama-knows-that-turkey-is-buying-oil-from-isis-and-he-isnt-doing-anything-to-stop-it/.
“Rus Savaş Uçağının Fetö Tarafıdan Düşürüldüğü İddiası,” Milliyet, October 7, 2017,
http://www.milliyet.com.tr/rus-savas-ucaginin-feto-taradan-dusuruldugu-sivas-yerelhaber-2323116/.
“Oil Smuggled into Turkey Not Enough to Be Protable: U.S. Ofcial,” Reuters, December 4, 2015, https://www.reuters.com/article/us-syria-oil-usa-idUSKBN0TN2P920151204.
72
73
74
75
19
Cyber Governance and Digital Democracy 2019/1
operating with a hidden, pro-NATO agenda76. According
to line-C, the decision to shoot down the jet was a NATO
clandestine instigation, set in motion to create signicant
escalation in Turkey-Russia relations as Russia had recently
entered the Syrian war. This line of reasoning suggested
that a major diplomatic incident with Turkey would be a
deterrent for Russia and that’s why clandestine pro-Gülen
and/or rogue pro-NATO cohort within the Turkish command
structure sought to create an articial crisis to cause military
escalation between Ankara and Moscow. All three lines
were nearly evenly distributed across the pro-government
accounts, without any particular narrative becoming more
popular than others. All of these different sub-narratives
were intended to divert further from Russia’s Narrative-C
(ISIS oil) and tried to seek a digital narrative consensus
that both sides could agree on; at least on paper. After the
failed coup attempt in July 2016, Narrative-D became the
mainstream narrative in Turkish-language Twitter.
Although Turkey prevailed over Russia in the rst round
(Days 1-8) over the framing of the jet downing incident,
Russia’s distraction strategy of widening the information
battleeld in round-2 (2 December 2015 - 6 August 2016)
has been more successful. This is a perfect example of
how Russia adapted to and surpassed the inherent logic
of the US ‘offset strategy’. The extent of this success is
evidenced both by the engagement and spread statistics in
Figure-4, but also by the fact that although Turkey resumed
its narrative-A, it has also shifted to an alternative narrative
that the decision to shoot down was made outside regular
decision-making channels. This may have had several
goals in mind, from exporting the blame away from recently
altered engagement rules against Syria, or giving Russia a
backdoor reconciliation option, or a signal of de-escalation,
or all of them. This shift of narratives has contributed to the
de-escalation of the crisis, as Sputnik-Türkiye, Aydınlık, and
OdaTV, along with regular pro-government newspapers
like Yeni Safak, Sabah and Takvim too, have both gradually
begun reporting on a possible ‘Gülenist meddling’. By
the end of August 2016, Russia had abandoned its line
furthering ‘Turkey-ISIS oil’ narrative, and Turkish line had
shifted from ‘we were right’, into ‘the decision was made by
conspirators’.
Fuad Safarov, “Rus uzman, Rus uçağını düşürme emrini kimin verdiğini açıkladı,” Sputnik Türkiye, 12 2016,
https://tr.sputniknews.com/rusya/201612041026128807-rus-uzman-gulen-ucak/; Elvan Alkaya, “Elvan Alkaya: FETÖ bağlantılı cinayetler ve şaibeli davalar (2),”
Yeni Şafak, July 26, 2016, https://www.yenisafak.com/yazarlar/elvanalkaya/feto-baglantili-cinayetler-ve-aibeli-davalar-2-2030700;
Mehmet Y. Yılmaz, “Davutoğlu bıçağın sırtında,” Hürriyet, July 19, 2016, http://www.hurriyet.com.tr/yazarlar/mehmet-y-yilmaz/davutoglu-bicagin-sirtinda-40154956.
76
Figure 4 - Time-series engagement metrics of four main LDA-designated narrative clusters (December 2015 – October 2016)
20
Cyber Governance and Digital Democracy 2019/1
The jet downing incident has been the rst and the last time that
Moscow exed its digital hybrid war muscles in dealing with
Ankara. Most certainly with the addition of more immediate
and ‘real’ factors related to Syria, growing Russian military
capabilities in the Black Sea and widening rift between
Turkey and NATO, Moscow’s information demonstration has
made a lasting mark on the Turkish information ecosystem.
This was also the rst, and the last time we have observed
a direct digital information confrontation between Russia
and Turkey and also the only time that we can conclusively
assert from data that Russian information operations had a
signicant inuence on Turkey’s information ecosystem in a
way that had a measurable effect on policy.
Disinformation during the coup attempt and in following
days was rife. The uncertainty of the rst few hours of the
coup attempt produced a number of contextually important
and potentially dangerous cases of disinformation that
served to distract, confuse and mobilize people. In the days
following the night of 15-16 July, the type of disinformation
attempts changed, replacing mobilization-oriented content
with cases of political disinformation that is harder to fact-
check and verify.
1. ResulK vs. F-16 - By far, most widespread of these
factually untrue types of popular content has been the man
who allegedly tried to jump on a rogue F-16 as it was diving
down into the Kızılay square in Ankara77. As unbelievable as it
sounds in hindsight, the story of Resul K., the man who ‘tried
to jump on an F-16 and wanted to smash its windows with
a crowbar’, became the widest-spread false information on
social media (Facebook, Twitter and Instagram combined).
This disinformation type, featuring a photoshopped image
of the Kızılay square, emerged within a cluster not afliated
either with the government, or the opposition, but based on
our study, was generated and spread through the ‘wisdom
of the crowds’, with no observable central node controlling
the spread. However, after a day after the introduction of this
particular disinformation case on social media, mainstream
pro-government and nationalist outlets have picked up on
this disinformation and began to share it on social media.
Days later however, the Milliyet78 newspaper reported that
Resul Kaptancı was indeed killed during his attempt to attack
the rogue troops taking over the General Staff headquarters
in Ankara, but the F-16 story was wrong. Yet, digital content
related to Resul Kaptancı was overwhelmingly shared as
part of the disinformation related to the attempt to jump on
an F-16 and remained that way. This disinformation content
is still popular on Youtube and popular culture websites in
Turkey.
Case-2: July 2016 Coup Attempt
“Şeytan taşladı,” Takvim, 08 2016, https://www.takvim.com.tr/guncel/2016/08/16/seytan-tasladi.
“15 Temmuz Şehidi Resul Kaptancı’nın Ailesi İdam İstiyor,” Milliyet, March 5, 2017,
http://www.milliyet.com.tr/15-temmuz-sehidi-resul-kaptanci-nin-ankara-yerelhaber-1884645/.
77
78
21
Cyber Governance and Digital Democracy 2019/1
2. Responsibility Waiver – The second most popular type of
disinformation was in fact, a meta-disinformation. It sought
to protect people against disinformation, but it was itself
disinformation. It featured a fabricated message allegedly
sent by, or on behalf of the Gendarmerie Headquarters
calling on everyone on social media to post a legal waiver
on their account pages, waiving their legal responsibility ‘in
case other malicious actors hacked into their accounts and
posted disinformation or anti-government messages without
their knowledge’. In the immediate post-coup environment,
where disinformation and confusion were rife, this case of
disinformation added fuel to the re and led hundreds of
thousands of Turks to post this disclaimer on their accounts.
Another version of this disinformation type cited the Prime
Ministry as the source. Regardless, neither the Turkish
Penal Code nor the country’s IT law contain any clause or
requisite related to such waivers and the text message was
fact-checked as essentially meaningless. This particular
disinformation type is traceable back to the opposition
news network and popular public gures, but later took on
a non-political tint, shared by a large portion of the Turkish
social media users. As demonstrated in Figure-6, this
disinformation type was fact-checked within the rst hour of
its emergence, but still lingered on as one of the widest-
shared disinformation types in the aftermath of the coup
attempt.
Figure 5 - Sample content and rst-hour diffusion network of the ResulK vs. F-16 disinformation
22
Cyber Governance and Digital Democracy 2019/1
3. Fake Azerbaijan Incursion – The third most widespread
disinformation type after the coup attempt was an alleged
statement attributed to the President of Azerbaijan, Ilham
Aliyev. According to the attributed quote, President Aliyev
had declared his support for the Turkish people’s resistance
against the coup attempt and added that had the people’s
resistance failed, the Army of Azerbaijan would move into
Turkey to help people quash the coup attempt. When traced
back to its original quote, President Aliyev did declare his
support for the resistance against the coup, but made no
statement that would even remotely resemble ordering his
army moving into Turkey79. Regardless, this disinformation
originally emerges within the ultranationalist social media
groups that have too few followers to make an impact. It
became viral, when the JPEG image of this disinformation
was shared on a popular Twitter account called @
TuhafAmaGercek (2.25 million followers as of writing this
report). This could have been a good candidate for pro-
Russian information operation, given Russia’s role in Turkish-
Azerbaijani relations, but the account @TuhafAmaGercek has
no connection to either Russia or the pro-Russian information
ecosystem in Turkey and is a widely followed popular culture
trivia account.
Figure 6 - Sample content and rst-hour diffusion network of the ‘Responsibility Waiver’ disinformation
“İlham Aliyev’den FETÖ Darbe Girişimine Kınama,” TRT Haber, 07 2016, https://www.trthaber.com/haber/dunya/ilham-aliyevden-feto-darbe-girisimine-kinama-261298.html.
79
23
Cyber Governance and Digital Democracy 2019/1
Figure 7 - Sample content and rst-hour diffusion network of the ‘Fake Azerbaijan incursion’ disinformation
4. Rogue Soldiers – The fourth largest spread disinformation
is a combination of distorted versions of a factually correct
news about the violence against rogue troops following the
failure of the coup attempt. The best-known and documented
case was the surrender of the rogue troops on Istanbul’s
second bridge, during which six soldiers were lynched80.
As one of the tensest ashpoints of the coup attempt, the
2nd Istanbul Bridge had witnessed rogue troops ring on the
crowd killing several civilians. In return, and although loyalist
troops and police were there to oversee the surrender, several
rogue troops faced lynching, and one of those had died
on the scene. This tense episode was portrayed on social
media through a urry of fake images and disinformation,
and can be traced back to OdaTV and Haber.Sol.Org.Tr81.
One of the most prominent cases was the spread of an
image that belongs to an ISIS beheading in Syria, as if it
was the image of the lynching in Istanbul. In addition, the
images of several alive soldiers were disseminated on social
media as the identity of the deceased soldier, adding further
confusion and anger to an already loaded situation. This type
of disinformation had spread through both pro-government
and opposition social networks and the involvement of
accounts that can be traced to Russia is non-existent. This is
one of the most contested disinformation types related to the
coup attempt and the fake versions of the actual incident still
circulate widely on Turkish-language social media platforms.
“6 askere linç,” Hürriyet, 07 2016, http://www.hurriyet.com.tr/gundem/6-askere-linc-40150768.
“‘Başı kesilen asker’ haberine gelen yalanlama ve o haberin hikayesi,” Sözcü, 07 2016,
https://www.sozcu.com.tr/2016/gundem/basi-kesilen-asker-haberine-yalanlama-ve-o-haberin-hikayesi-1319809/.
80
81
24
Cyber Governance and Digital Democracy 2019/1
5. Police HQ Bombing – Another highly potent case of
disinformation has been the video that some of the prominent
media outlets shared as the ‘bombing of the Ankara police
headquarters’82. The headquarters building was indeed
attacked during the coup night, but a doctored video
belonging to 2014 Israeli airstrikes in Gaza became more
widespread compared to the actual event footage. The fact
that such an easily veriable video (Google reverse video
search function easily brings the original Youtube version83)
was shared across prominent media outlets gives a good
idea on how crisis episodes render even trained journalists
into sharing unveried false information. This has been by
far the most problematic type of fake news to verify. The
video is a doctored version of Russia Today’s live coverage
of the 2014 Israeli bombardment and thus, most links shared
on social media contain the ‘rt.com’ domain extension.
However, it is impossible to verify who really doctored the
video and the central nodes in the dissemination of this
video in its very early hours are all Turkish accounts.
Figure 8 - Sample content and rst-hour diffusion network of the ‘Rogue Soldiers’ disinformation
A recording of the video on CNN-Turk can be accessed here: https://twitter.com/t24comtr/status/754384016657289217
For the video link, please visit: https://www.youtube.com/watch?v=8pffdijUI1U
82
83
25
Cyber Governance and Digital Democracy 2019/1
A nal word about the Russian role in disseminating news on
NATO and/or US involvement or negligence/apathy during
the coup attempt. Sputnik-Türkiye and a number of other
pro-Russian accounts in Turkey did push this line during
and after the coup attempt. However, this line was already
a mainstream view in the Turkish information ecosystem in
the immediate aftermath of the coup attempt and dominantly
disseminated by the pro-government news networks. Our
survey found 5 such content types by Sputnik-Türkiye,
but discovered that their reach was below 800 aggregate
engagements. In contrast, however, digital content
originating from mainstream Turkish accounts blaming the
US or NATO had already reached around 100,000 aggregate
engagements by the end of 17 July 2016.
Figure 9 - The original RT video, which was doctored and used as a disinformation footage and
the rst-hour diffusion network of the ‘Ankara bombing’ disinformation
26
Cyber Governance and Digital Democracy 2019/1
The second major crisis after the SU24 incident and months
of diplomatic rebalancing was the assassination of the
Russian Ambassador to Ankara, Andrey Karlov. Already
tense and mutually distrustful from the damage caused
by the SU-24 incident, the assassination of Ambassador
Karlov on 19 December 2016 in Ankara, during a public
event, threatened to hamper reconciliation efforts. The
assassination followed a cautious improvement in Turkish-
Russian relations ongoing since the Erdogan-Putin meeting
in August. The assassination had profound implications,
given the assassin was an off-duty police ofcer, who
sneaked into Ambassador Karlov’s security detail using
his police ID84. The weeks that ensued were marked by
signicant noise in digital information ows, mostly regarding
the allegiances of the attacker, as well as the motivations
that led to the assassination. Turkey’s ofcial line has been
that the attacker was a member of the Fethullah Gülen
network; a point that was personally made by President
Recep Tayyip Erdoğan85. Other alleged allegiances of the
attacker were to ISIS, and Jabhat Fatah al-Sham86 - a jihadi
group active in north-western Syria. Recently in late-2018, a
new theory emerged that the attacker belonged to another
Islamic congregation – an allegation made none other than
by Sputnik-Türkiye itself87.
The sentiment scores of the aggregate data in the rst 3 hours
of the assassination are mixed (positive and negative) and
display signicant confusion. Hostile-designated content are
those that either a) question whether the assassination really
happened or whether it was ‘staged’, and b) expressed
opinion that the assassination was a Russian ‘inside job’
that attempted to further embarrass Turkey by exposing its
internal security decit, or that c) it was a ‘Western instigation’
aimed to disrupt ongoing Turkish-Russian rapprochement.
This confusion owes largely to the fact that there has been
no centrally-coordinated messaging in Turkish information
ecosystem within the rst hour, and all major information
brokers tried to weave a narrative on their own. It is with the
2-hour mark that the Presidency and government accounts
begin to assert a positive narrative to end the confusion in
the digital domain. With the identication of the attacker as
a police ofcer another wave of confusion ensues, trying to
contextualize this security decit. But this urry of negative-
sentiment content also dies down by the 9-hour mark as the
ofcial narrative settles in and denes the act ofcially as a
terrorist attack and declares solidarity with Moscow. Indeed,
by the 12-hour, both organic and bot-driven engagement
begins to revolve around keywords related to terror/
terrorism, and also condolences and sympathies towards
Russia and the Ambassador’s family, revealing that the
Turkish-speaking Twitter ecosystem largely accepted the
government narrative. By the end of Day-1, two additional
negative-sentiment peaks emerge, following the ofcial
denouncement of the attack by President Erdoğan and Prime
Minister Binali Yıldırım. By Day-2, a new inux of negative-
sentiment content is observable, both at the organic and bot-
driven-level, blaming the Gülen network for its involvement
in the assassination. This line of narrative is similar to the
one in the jet downing incident, as three distinct plot lines
emerge: a) the assassin was directed by a rogue police
chief, b) the assassin himself was a rogue operator, and c)
decision was made by a clandestine network operating with
a hidden, pro-NATO agenda.
Case-3: Assassination of the Russian Ambassador in Ankara
“Karlov suikastı iddianamesi mahkemede,” NTV, November 23, 2018, https://www.ntv.com.tr/turkiye/karlov-suikasti-iddianamesi-mahkemede,pG6pIihRcUaQQcbvWqgpEg.
“Erdoğan: Suikastçı FETÖ’ye Mensup,” December 21, 2016, sec. Türkiye, https://www.bbc.com/turkce/haberler-turkiye-38394837.
“Karlov suikastını, eski adı El Nusra olan Fetih el Şam üstlendi,” Sputnik Türkiye, December 21, 2016,
https://tr.sputniknews.com/ortadogu/201612211026428039-karlov-fetih-el-sam/.
“Karlov suikastı sanığı: Menzil tarikatına bağlıyım,” Sputnik Türkiye, January 11, 2019,
https://tr.sputniknews.com/turkiye/201901111037037495-karlov-suikasti-sanigi-menzil-tarikatini-bagliyim/.
84
85
86
87
27
Cyber Governance and Digital Democracy 2019/1
We measure the centrality of actors, sentiment and content in dedicated network clusters based on the following function
(Saxena et. al. 201788)
Where Cmid denotes closeness centrality of the best-connected node in the cluster, n is the aggregate number of nodes, p
is the degree of the logistic curve at median.
Akrati Saxena, Ralucca Gera, and S. R. S. Iyengar, “A Faster Method to Estimate Closeness Centrality Ranking,” ArXiv:1706.02083 [Physics], June 7, 2017,
http://arxiv.org/abs/1706.02083.
88
Positive-Sentiment Top Keywords Occurrence Negative-Sentiment Top Keywords Occurrence
Karlov 12,954,371 Karlov 3,917,394
Rus_ 11,720,845 Büyükelçi_ 3,742,197
Suikast_ 9,384,054 Öldü_ 3,248,492
Büyükelçi_ 7,920,581 Provok_ 2,928,915
Üzü_ 6,183,401 Tehdit_ 2,271,601
Putin_ 5,193,932 Ankara_ 1,293,857
Başsağlı_ 4,910,356 Oyun_ 1,104,381
Dost_ 3,291,603 Suriye_ 958,398
Andrey 2,869,039 Halep_ 769,271
Kın_ 1,958,204 FET_ 481,293
Halkı_ 984,812 NATO 385,926
İşbirli_ 750,386 Amerika_ 204,385
Table 7 - Most frequently occurring n-grams in positive and negative-sentiment content and engagement clusters
Figure 10 - Weighted centrality network of inter-account engagement within the Positive-Sentiment cluster (rst 4 hours)
28
Cyber Governance and Digital Democracy 2019/1
Figure 12 - 24-hour time-series analysis of LDA-designated Positive and Negative-sentiment engagement metrics,
sorted by organic/bot and authentic/retweet designation. Count value in 000s.
Figure 11 - Weighted centrality network of inter-account engagement within the Negative-Sentiment cluster (rst 4 hours)
29
Cyber Governance and Digital Democracy 2019/1
Figure 14 - Weighted centrality measures of the most frequently appearing words in bot-driven negative sentiment tweets
(rst 2 hours)
Figure 13 - Weighted centrality measures of the most frequently appearing words in organic positive sentiment tweets
(rst 2 hours)
30
Cyber Governance and Digital Democracy 2019/1
While the positive-sentiment mention network measures
of the rst 3-4 hours demonstrate convergence, negative-
sentiment mention network is more fragmented. This means
that tweets designated as ‘positive’ by our algorithm indicate
greater centralized control (i.e., government-led framing
attempts), whereas negative sentiment tweets originate
from accounts that are either loosely tied to the government
or the opposition network. This is equally valid for both
organic and bot-driven content. The negative-sentiment
ecosystem on the other hand (again, both organic and bot-
driven) is greatly fragmented in terms of discussed topics
and topic weights. Although positive sentiment score topics
contain mostly organic sentiments of sadness, sorrow and
solidarity between Turkey and Russia, negative sentiment
score topics contain a variety of narratives driven primarily
by hashtags. These are namely a) assassination being a
‘provocation’, b) an ‘external game played on Turkey’, c) a
‘successful revenge’ against Russian military role in Aleppo,
or d) an incident aiming to pressure Turkey internationally.
Also in the rst 3-hours, there are occasional bot-driven
negative sentiment spikes dominated by implications of
US and NATO involvement in the assassination attempt.
Although by the 6th hour these suggestions are eliminated
from the information ecosystem as there have been sporadic
resurfacing of NATO and US involvement allegations within
the same ecosystem that makes such blames through
the Gülen organization. The NATO and US-related blame
eventually disappear from the ecosystem by the 12th hour
as the most frequent negative sentiment mentions become
‘terror/terrorism’, ‘treason/traitor’ and ‘FETÖ’. By the end of
the rst week and the return of Ambassador Karlov’s body
back to Moscow, the information ecosystem settles into a
generally positive equilibrium as content frequency and
substance both reveal mutual understanding, sympathy
and condolences. Through all time frames observed (1,
3, 6 12-hour, 1-day intervals) the pro-Russian information
ecosystem in Turkey has been remarkably silent. Sputnik-
Türkiye and accounts that regularly lie within the Sputnik
network have resorted to very few content shares. Those
that they did share were mainly quotes directly attributed to
the Presidency, government, and security ofcials.
In digital terms, this particular episode in Turkish-Russian
relations has demonstrated substantial restraint on the
part of the pro-Russian information apparatus in Turkey,
given the silence and direct reliance on ofcial sources in
its news. Especially when compared to the aftermath of
the SU24 downing incident, pro-Russian media behavior
in Turkey has been signicantly toned-down, asserting
Moscow’s condence in the way Turkish authorities dealt
with the situation. This suggests high-level coordination
and condence-building at play that would force Moscow
to refrain from the kind of media assault it did with the ‘ISIS
oil’ strategy.
Another important observation is that both positive and
negative sentiment diffusion clusters orbit the same high-
level ofcial sources. Regardless of whether a network
represents negative or positive sentiment content, its
constituent accounts retweet and follow the same key
ministries, security agencies and the Presidential ofce.
This is an interesting nding, since all of the widely-shared
content types overwhelmingly reference ofcial organs, but
interpret the statements of these ofcial organs completely
differently. This nding is especially visible within the rst
hour of the assassination, as negative-sentiment bots and
organic accounts represent ofcial statements in a way
that ts their own agenda. It is only after the rst two hours
and a more explicit assertion of the ofcial view through
Presidency and ministerial accounts that these negative
sentiment clusters switch to a more conciliatory and de-
escalating tone.
31
Cyber Governance and Digital Democracy 2019/1
Figure 15 - 7-day time-series analysis of LDA-designated Positive and Negative-sentiment engagement metrics,
sorted by organic/bot and authentic/retweet designation. Count value in 000s.
Perhaps the most researched aspect of Russian involvement
in digital space has been election meddling in Western
democracies. While Russia’s involvement in Western
elections is relatively well documented, there is virtually
no substantive and empirical research on similar election
interference beyond these countries - and denitely no data-
driven research on Turkey. This is important, since checking
Russian involvement in any election meddling in Turkey
would serve as a good robustness check of our ndings
related to the most signicant crises in relations.
Although a separate, longitudinal analysis on disinformation
in all Turkish election since 2010 is necessary, that would
lie beyond the scope of this project. There is an endless
debate in Turkish political science literature on the question
of which recent election(s) were ‘really’ the most important
for Turkey. Presenting this research in multiple occasions,
we observed an acute disagreement among the attendees,
all offering a different combination of elections as ‘the most
important’. To that end, we decided to go only with the 24
June 2018 elections for two reasons. First, it has the largest
volume of data compared to all other Turkish elections with
nearly 80 million tweets containing solely disinformation-
related content, after data cleaning (the total volume of
election data is much larger). Second, it had the largest
turnout (87%89) among recent Turkish elections since the
1999 general elections and it was the Turkish election with
the largest ever nominal votes cast (51,183,72990), including
the 2017 referendum. To that end, it is a better robustness
check compared to the alternatives.
Case-3: Potential Election Meddling
“Voter Turnout in Turkey Elections Was 87 Percent: State Broadcaster,” Reuters, June 24, 2018, https://www.reuters.com/article/us-turkey-election-turnout-idUSKBN1JK0SP.
“Son dakika: YSK, 2018 kesin seçim sonuçlarını açıkladı,” Hürriyet, July 4, 2018,
http://www.hurriyet.com.tr/gundem/son-dakika-ysk-2018-kesin-secim-sonuclarini-acikladi-40886560.
89
90
32
Cyber Governance and Digital Democracy 2019/1
To trace Russian involvement in Turkish digital disinformation
ecosystem during the election, this study focuses on 6 of
the most popular of such cases (measured by more than
50,000 engagements within the rst day) on or before 24
June. These cases are exposed as disinformation and were
fact-checked by Teyit.org.
1. Allegation that Erdogan called Meral Aksener ‘zilli’
(1,492,493 combined engagement): One of the odder,
yet signicant disinformation types was the allegation that
President Erdoğan had called a major opposition leader
Meral Aksener ‘zilli’91 - a slang word that has a wide spectrum
of meanings, closest in this context being ‘shrewish’.
Originally appearing in a public Facebook page called ‘İzci’,
the disinformation attempt lashed out at President Erdoğan
for his lack of tact during his party’s convention, by referring
to Meral Akşener in slang terms. In reality, there never was
such a statement, evidenced by President Erdoğan’s full
speech, available on Youtube92. The content spread quite
rapidly across opposition-nationalist information ecosystem,
also spreading to the pro-government network within the
same day. Similar to other disinformation cases, although
the correction and fact-checking content appeared on social
media outlets within the a few hours of its dissemination, this
disinformation type took on a life of its own and it was shared
frequently until the election day on 24 June. There is no node
or network in this disinformation ecosystem that could be
traced to pro-Russian sources.
“Erdoğan’ın Meral Akşener hakkında ‘Zilli Meral Kemal’in eteklisi’ dediği iddiası,” teyit.org (blog), May 7, 2018,
https://teyit.org/erdoganin-meral-aksener-hakkinda-zilli-meral-kemalin-eteklisi-dedigi-iddiasi/.
Full video can be accessed at: https://www.youtube.com/watch?v=FKpQ0irG6aE
91
92
33
Cyber Governance and Digital Democracy 2019/1
Figure 17 - Time-series engagement graph showing the popularity of the fake news(below) and fact-checked (above) content
Figure 16 - Sample content and rst 2-hour diffusion network of the ‘zilli’ disinformation
34
Cyber Governance and Digital Democracy 2019/1
2. Muharrem Ince dance (2,491,385 combined
engagement): The widest-shared disinformation regarding
the main opposition candidate Muharrem Ince was a
photoshopped image of him, dancing in a mosque93. As
absurd as the statement sounds, the doctored image of
a dancing Muharrem Ince overlaid into a mosque interior
became one of the most shared digital content types of the
election. The content can be traced back to a very distinct
pro-government network, in which former Ankara mayor,
Melih Gokcek appears to be the overwhelmingly central
node based on the function we import from Saxena (et.
al. 201794). There is also a distinct bot network involved in
the initial spread. Rather than any pro-Russian ecosystem,
this disinformation type can be traced into a distinctly pro-
government media and opinion network.
“Fotoğrafın Muharrem İnce’nin camide halay çektiğini gösterdiği iddiası,” teyit.org (blog), May 9, 2018,
https://teyit.org/fotogran-muharrem-incenin-camide-halay-cektigini-gosterdigi-iddiasi/.
Saxena, Gera, and Iyengar, “A Faster Method to Estimate Closeness Centrality Ranking.”
93
94
Figure 18 - Time-series engagement with the fake news (rst 1.5 days).
Blue: Engagement through popular (high follower count) accounts, Red: Nominal value of engagement
35
Cyber Governance and Digital Democracy 2019/1
Figure 19 - The diffusion network (overwhelmingly pro-government) of the fake news. The large red dot represents
@06melihgokcek – which assumes disproportionate centrality in the diffusion network.
3. Alleged Election Fraud in Diyarbakır (1,375,482 combined
engagement): Another major disinformation case that
proliferated on the day of the election was the allegation
that a voting site (Mesut Yilmaz primary school building)
in Diyarbakır was visited by three members of the Higher
Electoral Board (YSK)95. These ofcials, according to rumors,
looked suspicious and tried to force ballot box administrators
and observers to agree to fraudulent behavior. Shared by
a local district ofcial from the People’s Democracy Party
(HDP), the image of the YSK members accompanied by
police ofcers circulated far and wide on social media,
creating widespread upheaval in an already-tense city.
After multiple inquiries, it was revealed that the YSK team
was sent to the voting site following an additional envelope
request by the voting site scrutineers and had ultimately
reached an agreement with the HDP ofcials96. Although the
veried version of events was also disseminated on social
media, the misunderstood image and the content became
shared far more across social media venues, warning all
observers and voters to ‘be careful against YSK ofcials’.
This predictably caused substantial problems in a number
of voting sites during regular and scheduled YSK observer
visits.
“Diyarbakır Mesut Yılmaz İlkokulu’nda boş zararla dolaşan kişiler kim?,” teyit.org (blog), June 24, 2018,
https://teyit.org/diyarbakir-mesut-yilmaz-ilkokulunda-bos-zararla-dolasan-kisiler-kim/.
“Emniyet Müdürü: YSK’dan zarf talep etmiş,” HaberTurk, June 24, 2018, https://www.haberturk.com/diyarbakir-da-muhurlu-bos-zararda-yanlis-anlasilma-2029291.
95
96
36
Cyber Governance and Digital Democracy 2019/1
This particular case of disinformation invalidates an important
theory of communication, which in essence, argues that
strengthening local news is the best way to combat fake news
at the national level97. Proponents of this argument posit that
if local journalism is supported and strengthened, on-the-
ground reporters would do a much faster job in validating
false claims before they could reach the national level. In
our particular case however, the opposite happened. Local
news agencies have been the main sources from which
disinformation emerged, spread and picked up by national
news. Although HDP ofcials later claried the situation,
most local news agencies did not share the fact-checked
version of events, creating around 6-8 hours’ time lag until
engagement with this content declined visibly. The weighted
centrality of this disinformation case is clustered around local
news outlets in Diyarbakır. In addition, there are a number
of marginally important anonymous organic accounts that
are unidentiable in terms of their political allegiances, either
from their prole information, or text analysis of their tweets.
These accounts cannot be traced to any pro-Russian
network in Turkey and remain outside of the central pro-
Kurdish news network diffusion mechanics (see Fig. 21).
Damian Radcliffe, “How Local Journalism Can Upend the ‘fake News’ Narrative,” The Conversation, November 27, 2018,
http://theconversation.com/how-local-journalism-can-upend-the-fake-news-narrative-104630.
97
Figure 20 - Sample content types and time-series engagement with the Diyarbakır fake news case
37
Cyber Governance and Digital Democracy 2019/1
Figure 21 - Spread network of the disinformation. Starting from local news sources,
the fake news was ultimately nationalized through Twitter
4. Alleged Erdoğan Rally Mistake (1,371,964 combined
engagement): This case of disinformation alleged that
President Erdoğan, in an election rally in Bursa, mistakenly
addressed the crowd as ‘the people of Sakarya’, confusing
which city he was in98. This disinformation type was shared
frequently with other minor attempts that pointed to the
old age of the President and his slip ups and mistaken
statements, questioning whether he was t to rule the
country. In reality, Erdoğan was citing a folk song, which
contains the words ‘Sakarya’, but did not confuse his
audience99. Originating within the opposition network cluster
more aligned with the Republican People’s Party (CHP)
this was one of the fastest-spreading fake news cases we
encountered (See. Fig.21). The role of bots is minimal in this
case, as the organic opposition network has contributed
heavily to the dissemination of such content. Despite the
fast fact-checking of Teyit.Org, the fact-checked version
of events was marginalized by the vast wave of accounts
involved in the disinformation effort, willingly or unwillingly.
No pro-Russian network is observable in this case.
For a popular example, see: https://twitter.com/avcimucahit/status/1006221787846324225
“Cumhurbaşkanı Erdoğan’ın Bursa mitingindeki kalabalığa ‘Sakarya’ diye seslendiği iddiası,” teyit.org (blog), June 12, 2018,
https://teyit.org/cumhurbaskani-erdoganin-bursa-mitingindeki-kalabaliga-sakarya-diye-seslendigi-iddiasi/.
98
99
38
Cyber Governance and Digital Democracy 2019/1
Figure 22 - Time-series diffusion and engagement patterns of the fake news (blue) and its fact-checked version (red)
Figure 23 - Despite fact-checks (red), the overwhelming majority of the engagement with the fake news (blue)
demonstrates wide reach and dissemination. Politically, the cluster is overwhelmingly made up of opposition politicians,
journalists and celebrities that are politically highly engaged with each other. In other words, the network represents an
overwhelming opposition network.
39
Cyber Governance and Digital Democracy 2019/1
5. Allegation that Ince was in a coffeehouse during the
coup attempt: The allegation that Muharrem Ince, the
main opposition candidate was in a coffeehouse, playing
rummikub during the night of the coup was disseminated
widely by the pro-government conventional news accounts
like Yeni Akit, A Haber and Takvim100. Using a photo from
December 2014, these news outlets began circulating
the disinformation attempt online. This was one of many
similar disinformation styles of the pro-government media
ecosystem blaming senior members of the opposition
parties for not ‘doing enough’ during the coup attempt. The
interesting point with this particular case of disinformation is
that the opposition network is equally active in disseminating
it as the pro-government network. Of further interest is the
fact that it was another pro-government account (@Ankara_
Kusu) that eventually spearheaded the fact-checking effort
in the rst hours of the spread.101 There are no pro-Russian
accounts observable in the early impact network of this
content.
Figure 24 - Muharrem Ince-related disinformation content example and diffusion network
“Darbe gecesi tavla partisindeymiş,” Sabah, June 22, 2018, https://www.sabah.com.tr/gundem/2018/06/22/darbe-gecesi-tavla-partisindeymis.
See: https://twitter.com/ankara_kusu/status/1010272819073179649
100
101
40
Cyber Governance and Digital Democracy 2019/1
6. Disappearing ink: By noon on the election day, Ruhat
Mengi - a Turkish journalist - warned her 62.5k followers
that a friend’s relative had spotted a tampered ballot stamp.
According to the allegation, a ‘special ink’ was distributed
by the government to voting sites that regularly vote against
the ruling AKP and Erdoğan, and was an elaborate plan to
render all votes in opposition districts ‘null’. This message
was retweeted and shared across media platforms within
minutes and became shared by some of the most inuential
journalists and politicians online.
Our analysis shows that the tweet was actually shared much
earlier in the morning (06:26am) by a university student
in Ankara102, and spread across several other platforms
and social chat channels before reaching Ruhat Mengi by
noon. Evidently, there was no veried report in any medium
regarding any ‘disappearing ink’; this allegation was later
fact-checked and corrected. Still, the tweet retained its
potency and got shared online for hours to come. This was
a truly cross-platform disinformation example, as tracing
its course solely through Twitter yields insufcient results.
Several examples on Facebook, Whatsapp, Eksisozluk and
Instagram were spotted by our research team, although
establishing a clear causal cross-platform linkage is very
difcult.
Diffusion patterns reveal that although the fake news originally
emerged within the opposition network, it was later picked up
by a number of highly inuential ultranationalist anonymous
accounts to fact-check and stop its diffusion. After its
fact-checked versions appeared online, pro-government
accounts spread both the fake news and the fact-checked
version in order to mock the opposition network due to the
strangeness of their claim. Although the opposition network
originated the content, it was mostly the pro-government
networks mocking the claim that contributed to the widest
diffusion of this disinformation type. We haven’t encountered
any accounts in this network that could be connected to the
pro-Russian information ecosystem.
Figure 25 - Sample content type engagement metrics of the ‘disappearing ink’ fake news
Tweet link: https://twitter.com/MazharCoban/status/1010786264944128001
102
41
Cyber Governance and Digital Democracy 2019/1
Another critical case study for Russian information ecosystem
in Turkey is the S400 saga. Syrian Civil War had heightened
Turkey’s threat perceptions and made the country’s long-
term dependence on NATO for aerial defense more glaring.
Although Patriot missile batteries were deployed along the
border, operated by American, German, Dutch and Spanish
crews in turns, Turkey viewed the delays in the arrival of these
batteries as highly problematic to its immediate national
security concerns. Eager to become more self-sufcient
and prevent such critical delays in the future, Turkey sought
resident anti-air batteries. Initially reaching out to Chinese
anti-air systems, this move was abandoned due to China’s
reluctance to fulll Turkey’s technology transfer criteria103.
Then, Turkey showed interest in Russian S400 systems,
seeking to secure a better technology transfer and co-
production deal. Ultimately, an agreement was reached in
September 2017, when President Erdoğan stated that the
deal with Russia was a foregone conclusion, leading to the
2.5 billion US Dollar signing in December. Yet, this deal has
no technology transfer or co-production clause as well. The
rst S400s systems were then announced to be delivered
in 2019.
In a search for pro-Russian information ows through this
period, this study has benchmarked ve events:
10 October 2016 when Turkey and Russia declared that
serious Presidential-level negotiations were underway
over S400 sales,
Erdoğan’s 10 March 2017 visit to Moscow to assert
Turkey’s commitment to S400,
29 December 2017 commercial agreement between
the two sides,
3 April 2018 President Erdoğan’s statement on Turkey’s
‘point of no return’ on S400 purchase,
19 August 2018 President Putin’s statement that
deliveries could be made a year earlier than planned.
In these periods, we have measured pro-Russian and
anti-Russian sentiments on Turkish-language Twitter, as
measured by our topic modeling algorithm that we trained
on Turkish political text. It is interesting to observe the
gradual transition of the sentiment scores related to S400s
from mixed (negative and positive) to mostly positive
through these ve cases we observed. The October 2016
Case-4: S400 Procurement Process
Figure 26 - Diffusion network of the fake news (rst 6 hours)
“Turkey Conrms Cancellation of $3.4 Billion Missile Defence...,” Reuters, November 18, 2015,
https://www.reuters.com/article/us-turkey-china-missile-idUSKCN0T61OV20151118.
103
a.
b.
c.
d.
e.
42
Cyber Governance and Digital Democracy 2019/1
declaration was mostly met with mixed views by the Turkish-
language Twitter. Digital opinions shared on the matter
split between whether S400s are ultimately good (positive
sentiment) or bad (negative sentiment) for Turkey. Due
to the technical nature of S400 negotiations, we observe
a general restraint by users to post authentic tweets, and
rather their preference for retweeting reports on the deal
published by familiar media outlets. The even split between
positive and negative sentiments gradually eases down
and positive sentiment scores prevail at an increasingly
greater margin across ve benchmarked events, ultimately
becoming the mainstream consensus in Turkish-language
Twitter. Especially with the August 2018 statement by Putin,
the overall interest in S400s both decline and converge on a
generally positive sentiment.
As shown on Table-8, some of the main words designated
as ‘negative-sentiment’ topic cluster contain interoperability
issues with existing NATO equipment, commitment to NATO
regarding the deployment of Patriot missile batteries, and
to what extent S400 acquisitions will inuence Turkey’s
bid to buy F-35 jets from the United States. Such content
viewed S400s not as diversication, but strategic confusion
as it brought operational problems with potentially real-life
consequences. Positive-sentiment topic cluster words, on
the other hand, contain n-grams that belong to digital content
that question the relevance of NATO, importance of Turkey’s
security autonomy, and technical information about S400s
(such as its range and capabilities) largely interwoven with
Erdoğan’s supportive statements about the missile systems.
Because we look at the longitudinal changes in sentiment
scores, S400 deal is an overall good way to observe the
gradual shift of the Turkish information ecosystem from an
ambivalent view towards Russia, into a mostly pro-Russian
view. The topic clusters show, however, that this pro-Russian
sentiment is less about Turks ‘liking’ Russia, and more about
their views about converging strategic interests in a limited
pool of issue areas.
Long-term sentiment scores also add substantial depth to
our previous ndings. In terms of the domain roots of positive
and negative-sentiment content, we clearly observe that the
pro-Russian opinions have been quite well-integrated into
the pro-government mainstream (Table-9). Because S400s
have largely been framed within the context of strategic
autonomy and greater self-sufciency in anti-air defensive
capabilities, both nationalist anti-government and pro-
government media clusters (that are usually non-converging)
are mostly converging in unison support. Negative sentiment
clusters appear to have converged along the mainstream
center-right and centrist media outlets, although their share
gradually declines as positive-sentiment content becomes
the main expression regarding the S400s.
Figure 27 - - Time-series diffusion of positive and negative sentiment scores towards S400 missiles based on
organic/bot and authentic/retweet clusters. Date range 1 October 2016 – 1 September 2018.
Peaks represent engagement during ve benchmarked events.
43
Cyber Governance and Digital Democracy 2019/1
Negative Sentiment
Top Keywords Occurrence Positive Sentiment
Top Keywords Occurrence
S400 1,382,385 S400 6,492,695
Füze_ 1,294,016 Türkiye 6,329,593
Birlikte_ 1,184,603 Savunma 5,938,149
NATO 1,035,596 Hava 5,394,014
F16 939,503 Karşı 4,205,699
Patriot 929,458 Erdoğan 4,144,923
Entegre 837,814 Bağımsız_ 3,948,391
Savunma 742,559 Yüksek 3,854,553
Altyapı 693,941 Kapasite_ 3,469,319
Kullanılabi_ 491,394 Menzil_ 3,194,966
Türkiye 443,644 Füze 2,847,471
Kongre_ 385,193 Güvenli_ 2,638,383
F35 293,004 Sınır 2,385,993
Negative Sentiment Domain % Share of
Aggregate Tweets Positive Sentiment Domain % Share of
Aggregate Tweets
Hurriyet.com.tr 22.382 Sabah.com.tr 27.294
CnnTurk.com.tr 19.504 Yenisafak.com 21.953
Sozcu.com.tr 17.828 HaberTurk.com.tr 19.144
Cumhuriyet.com.tr 15.382 Ahaber.com.tr 13.048
T24.com.tr 13.449 Star.com.tr 8.032
bbc.com (Turkish) 6.486 OdaTV.com 5.392
DW.com (Turkish) 3.291 Tr.Sputniknews.com 5.019
Table 8 - Most frequently occurring n-grams in positive and negative-sentiment content and engagement clusters
Table 9 - Top domains within positive and negative-sentiment content and engagement clusters
44
Cyber Governance and Digital Democracy 2019/1
This report sought to explore the impact and relevance of
pro-Russian information operations in Turkey. To do this, it
traced pro-Russian information ows on Twitter across some
of the most important events in Turkish-Russian relations and
two of the most important cases for Turkey individually to act
as an analytical robustness check. In doing so, this study
has become one of the largest longitudinal digital information
studies ever conducted. This study has discovered that with
the exception of the Russian allegation that Turkey was
selling oil to ISIS, accounts or content traceable to Russia
with a high degree of certainty have close to zero inuence
on Turkey’s digital information ecosystem. Compared to
Russian disinformation and opinion manipulation efforts in
most of the Western countries that are both easy to map
out and have a visible impact on politics, similar activities
in Turkey are both less explicit, and also overwhelmingly
insignicant. This doesn’t mean Russian disinformation or
information operations don’t exist in Turkey. Rather, this
means that they don’t have any measurable impact on the
wider Turkish-language information ecosystem, whereas
distinctly pro-Russian views are disseminated by mainstream
pro-government and opposition media networks.
This nding is important. Given the scale and directness
of both disinformation and election meddling in Western
democracies, Russian digital media presence in Turkey is
minimal. This begs the obvious question: Why?
The rst answer that this research can provide directly through
the empirical evidence is that pro-Russian opinion is well-
embedded within the existing Turkish information sphere.
In that, pro-Russian sentiments and opinion are already
integrated into the media mainstream without any need
for pro-Russian information operations. This doesn’t mean
that criticism of Russia has completely disappeared from
the ecosystem; rather, in cases most directly relevant and
important for Russia (Syria, Ukraine, natural gas partnership,
defense deals or nuclear energy) the pro-government media
apparatus visibly follows a pro-Russian line. When the pro-
Russian content pressures the government, it is picked by
the opposition network. In contrast, when the pro-Russian
content validates the government’s position, it is picked by
the pro-government network. This bandwagoning behaviour
can be viewed as a by-product of the strategic alignment (or
entrapment, depending on one’s point of view) between the
two countries that could be observed in tangible strategic/
security policy areas.
The second answer is that the Turkish information ecosystem
is already so plagued with disinformation and domestically-
generated fake news, that bits and pieces of cases
traceable to Russia end up getting suffocated within the
larger ecosystem. In two highly important cases for Turkey,
the failed 2016 coup attempt and 24 June 2018 elections,
domestic and indigenous disinformation attempts completely
overwhelm the information ecosystem, rendering (dis)
information content traceable to Russia highly insignicant.
This shouldn’t come as a surprise, as Turkey is already one of
the most bot-infected countries in the world and have one of
the lowest resistance to fake digital news. Yet, this evidence
connects to the popular theoretical debate on whether pre-
existing disinformation ecosystem renders further external
disinformation easier or harder to make an impact. The
general consensus in the political communication literature
is that a free press, open media environment and freedom
of expression render disinformation less effective and easier
to spot, as the ‘marketplace of ideas’ is expected to quickly
verify fake information and remove it quickly. This is what we
have seen in France during the 2017 election, for example.
Yet, the evidence presented in this report doesn’t support
this theory; it actually supports the exact opposite.
This leads to a second question: are Russian information
attempts insignicant because of Turkey’s pre-existing
information landscape, or, bluntly put, is Russia not trying?
Based on the sustained damage Russia wreaked with the
‘ISIS oil’ information campaign, it is possible to argue that
in most cases, Russia has so far refrained from exing its
digital informatics muscles. This could be a direct result of
changing strategic prerogatives in Ankara and Moscow,
and the resultant convergence of short-term security
interests. To that end, after Ankara’s gradual strategic shift
that began with Russia’s entry into the Syrian Civil War in
the summer of 2015, and especially following the rst
Erdogan-Putin meeting in August 2016, Russia appears to
have followed a different ‘sub-threshold’ protocol for Turkey
compared to other NATO countries. This is best exemplied
by the general silence of the pro-Russian accounts in Turkey
following what could have become a near-ideal exploitation
point: the assassination of the Russian Ambassador. Yet,
instead of launching an information campaign similar to
those observed in other major NATO countries, pro-Russian
accounts have chosen to share direct quotes from Turkish
ofcials and stuck to the facts. The S400 negotiations are also
indicative of the gradual transition of the Turkish information
CONCLUSION
45
Cyber Governance and Digital Democracy 2019/1
ecosystem from a divided sentiment on Russia to a generally
positive equilibrium. Evidenced by the high shares of pro-
Russian content shared across mainstream Turkish media
outlets, a major Moscow-led effort to establish a separate
large-scale pro-Russian outlet or network in Turkey doesn’t
seem necessary. In other words, there seems to be no logic
in destabilizing an already pro-Russian information network
through further disinformation. Furthermore, Turkey’s native
information ecosystem is already highly contaminated
with regular fake news that surmounting the pre-existing
disinformation hurdles would probably require a great deal
of nancial and human-resource investment on Russia’s
part.
On a closing note, disinformation in Turkey in general and
external information operations in particular (pro-Russian
or otherwise) are gradually becoming better observable
in the ‘dark social media’ (i.e., comments sections of
hidden or restricted pages). Instead of the open and easily
observable medium of Twitter and Facebook, newer forms
of opinion manipulation are emerging in restricted pages on
Facebook, Instagram and Turkey’s own Reddit: EkşiSözluk.
These outlets are the next emerging frontier in disinformation
research; although at their current state, they are not very
signicant or able to inuence the mainstream debate.
Future studies interested in disinformation research in
Turkey could benet from an extended digital ethnography
work on several of these restricted/hidden pages to see
to what extent discussion there inuences the wider
debate. Another potentially important avenue would be
communication apps like WhatsApp, Signal or Telegram,
although these avenues bring their own data availability
challenges. Yet, the most promising line of research seems
to compare Russian information operations in Turkey with
other countries where Russia is expected to meddle in, but
doesn’t. Bulgaria, Hungary, Poland or Romania could offer
important comparative insight.
Furthermore, there is a strong case for looking into Russian
disinformation about Turkey, in other NATO countries. It is
by now common knowledge that Russian disinformation
campaigns are aimed to polarize and distract public opinion
within NATO. It is also known that in almost all NATO
countries, immigration, Turkey’s role in the EU refugee deal
and Turkey’ EU membership bid have all been used to fuel
the rising far-right and contribute substantially to issue-
specic polarization. The most obvious case would be the
impact of the well-known Brexit referendum fake news,
which asserted that Turkey would be joining the EU and also
suggesting that Turkish migrants would come to the UK in
large numbers if the voters chose to remain in the Union104.
Given the success of Russian-origin immigration and refugee
disinformation in the European Union (on Syrian refugees105)
and the United States (on Mexican immigrants106), dividing
NATO further by pushing anti-Turkish agenda in native-
language digital domains in the West would t well with wider
Russian plans. After all, the ‘ISIS oil’ information operation
demonstrated how well Russia could turn NATO against
Turkey when it wants to. Further research on how Turkey-
related disinformation content is shared within European
and US information ecosystem would be a natural next step
to follow up after this study.
Yet, in exploring both ‘dark social media’ or Turkey-related
information operations among NATO countries studies have
to look beyond the availability bias of ‘disinformation exists’
and try to explain whether such attempts matter or have any
impact on either the mainstream information ecosystem. Most
importantly do Turkey-related disinformation efforts generate
any measurable outcome such as mobilization or alter
electoral behaviour like it did during the Brexit referendum
campaign. Most recently, Grinberg et. al. (2019107) article
in Science has demonstrated how, despite an avalanche
of attention and deep-dive into Russian disinformation
operations during the 2016 US Election, disinformation
(both Russian and indigenous) has accounted for only 6%
of aggregate cumulative digital media consumption before,
during and after the vote. Furthermore, it was only less
than 1% of the American digital media audience that was
exposed to 80% of the fake news disseminated around the
Anoosh Chakelian, “Facebook Releases Brexit Campaign Ads for the Fake News Inquiry – but What’s Wrong with Them?,” New Statesman America, July 27, 2018,
https://www.newstatesman.com/politics/media/2018/07/facebook-releases-brexit-campaign-ads-fake-news-inquiry-what-s-wrong-them.
Thomas Brey, “Fake News and Alternative Facts Target EU’s Core,” Deutsche Welle, December 29, 2018,
https://www.dw.com/en/opinion-fake-news-and-alternative-facts-target-eus-core/a-46889022.
Donie O’Sullivan, “Newly Released Facebook Ads Show Russian Trolls Targeted Mexican-Americans after Trump Election,” CNNMoney, May 10, 2018,
https://money.cnn.com/2018/05/10/technology/russian-facebook-ads-targeted-mexican-americans/index.html.
Nir Grinberg et al., “Fake News on Twitter during the 2016 U.S. Presidential Election,” Science 363, no. 6425 (January 25, 2019): 374–78,
https://doi.org/10.1126/science.aau2706.
104
105
106
107
46
Cyber Governance and Digital Democracy 2019/1
election campaign period. That study is a good reminder
of the necessity of contextualizing disinformation within the
wider information sample and establish some degree of
explanatory authority on the tangible and measurable effects
of digital propaganda. The impromptu Twitter debate108
between Gary King and other prominent disinformation
researchers is highly instructive in this context.
Ultimately, it is possible to argue with a great degree of
certainty that Russia has so far did not follow the same kind
of information strategy in Turkish-language digital domain
that it does in well-known cases in the West. This is a result
of the convergence in security/strategic relations between
Ankara and Moscow, and the heavy dose of pre-existing
indigenous disinformation environment in Turkey raising
the barriers of entry for external information meddling. This
state of affairs, of course, is dependent on the continuation
of converged strategic interests between the two sides,
and shouldn’t be interpreted as a structural explanation of
why Russian information operations in Turkey are largely
dormant.
The link to the debate can be accessed here: https://twitter.com/kinggary/status/1079503260657041408
108
47
Cyber Governance and Digital Democracy 2019/1
“6 askere linç.” Hürriyet, 07 2016. http://www.hurriyet.com.tr/gundem/6-askere-linc-40150768.
“15 Temmuz Şehidi Resul Kaptancı’nın Ailesi İdam İstiyor.” Milliyet, March 5, 2017.
http://www.milliyet.com.tr/15-temmuz-sehidi-resul-kaptanci-nin-ankara-yerelhaber-1884645/.
Aaronson, Michael, Sverre Diessen, Yves De Kermabon, Mary Beth Long, and Michael Miklaucic.
“NATO Countering the Hybrid Threat.” PRISM 2, no. 4 (2011): 111–24.
Abrams, Steve. “Beyond Propaganda: Soviet Active Measures in Putin’s Russia.”
Connections 15, no. 1 (2016): 5–31.
Alkaya, Elvan. “Elvan Alkaya: FETÖ bağlantılı cinayetler ve şaibeli davalar (2).” Yeni Şafak, July 26, 2016.
https://www.yenisafak.com/yazarlar/elvanalkaya/feto-baglantili-cinayetler-ve-aibeli-davalar-2-2030700.
Allcott, Hunt, and Matthew Gentzkow. “Social Media and Fake News in the 2016 Election.” Working Paper.
National Bureau of Economic Research, January 2017. https://doi.org/10.3386/w23089.
Bachmann, Sascha Dov, and Hakan Gunneriusson. “Russia’s Hybrid Warfare in the East: The Integral Nature
of the Information Sphere.” SSRN Scholarly Paper. Rochester, NY: Social Science Research Network,
October 7, 2015. https://papers.ssrn.com/abstract=2670527.
Balmas, Meital. “When Fake News Becomes Real: Combined Exposure to Multiple News Sources and Political
Attitudes of Inefcacy, Alienation, and Cynicism.” Communication Research 41, no. 3 (April 1, 2014):
430–54. https://doi.org/10.1177/0093650212453600.
“‘Başı kesilen asker’ haberine gelen yalanlama ve o haberin hikayesi.” Sözcü, 07 2016. https://www.sozcu.com.
tr/2016/gundem/basi-kesilen-asker-haberine-yalanlama-ve-o-haberin-hikayesi-1319809/.
Bennett, W. Lance, and Alexandra Segerberg. “Digital Media and the Personalization of Collective Action.”
Information, Communication & Society 14, no. 6 (September 1, 2011): 770–99.
https://doi.org/10.1080/1369118X.2011.579141.
Bentzen, Naja. “Foreign Inuence Operations in the EU - Think Tank.” Brussels: European Parliament, 2018.
http://www.europarl.europa.eu/thinktank/en/document.html?reference=EPRS_BRI%282018%29625123.
Boffey, Daniel. “EU Raises Funds to Fight ‘disinformation War’ with Russia.” The Guardian, December 5, 2018,
sec. World news.
https://www.theguardian.com/world/2018/dec/05/eu-disinformation-war-russia-fake-news.
Booth, Robert, Matthew Weaver, Alex Hern, Stacee Smith, and Shaun Walker. “Russia Used Hundreds of Fake
Accounts to Tweet about Brexit, Data Shows.” The Guardian, November 14, 2017, sec. World news.
https://www.theguardian.com/world/2017/nov/14/how-400-russia-run-fake-accounts-posted-bogus-
brexit-tweets.
Botelho, Greg. “Russia, Turkey Trade Charges: Who Bought Oil from ISIS?” CNN, December 2, 2015.
https://edition.cnn.com/2015/12/02/europe/syria-turkey-russia-warplane-tensions/index.html.
Bradshaw, Samantha, and Philip N. Howard. “The Global Organization of Social Media Disinformation
Campaigns.” Journal of International Affairs 71, no. 1.5 (2018): 23–32.
Brattberg, Erik, and Tim Maurer. “How Sweden Is Preparing for Russia to Hack Its Election,” May 31, 2018, sec.
World. https://www.bbc.com/news/world-44070469.
REFERENCES
48
Cyber Governance and Digital Democracy 2019/1
———. “Russian Election Interference: Europe’s Counter to Fake News and Cyber Attacks.” Washington DC:
Carnegie Endowment for International Peace, May 23, 2018.
https://carnegieendowment.org/2018/05/23/russian-election-interference-europe-s-counter-to-fake-
news-and-cyber-attacks-pub-76435.
Breslow, Jason. “All The Criminal Charges To Emerge So Far From Robert Mueller’s Investigation.” NPR.Org,
December 9, 2018. https://www.npr.org/2018/12/09/643444815/all-the-criminal-charges-to-emerge-so-
far-from-robert-muellers-investigation.
Brey, Thomas. “Fake News and Alternative Facts Target EU’s Core.” Deutsche Welle, December 29, 2018.
https://www.dw.com/en/opinion-fake-news-and-alternative-facts-target-eus-core/a-46889022.
Brooks-Pollock, Tom. “Russia Releases ‘proof’ Turkey Is Smuggling Isis Oil over Its Border.” The Independent,
December 2, 2015. http://www.independent.co.uk/news/world/europe/russia-releases-proof-turkey-is-
smuggling-isis-oil-over-its-border-a6757651.html.
Carrel, Paul, and Andrea Shalal. “Germany Says Its Government Computers Secure after ‘isolated’ Hack.”
Reuters, February 28, 2018.
https://www.reuters.com/article/us-germany-cyber-russia-idUSKCN1GC2HZ.
Chakelian, Anoosh. “Facebook Releases Brexit Campaign Ads for the Fake News Inquiry – but What’s Wrong
with Them?” New Statesman America, July 27, 2018. https://www.newstatesman.com/politics/
media/2018/07/facebook-releases-brexit-campaign-ads-fake-news-inquiry-what-s-wrong-them.
Conley, Heather A., and Jean-Baptiste Jeangène Wilmer. “Successfully Countering Russian Electoral
Interference.” CSIS Briefs. Washington DC: Center for Strategic and International Studies, June 21,
2018. https://www.csis.org/analysis/successfully-countering-russian-electoral-interference.
“Cumhurbaşkanı Erdoğan’ın Bursa mitingindeki kalabalığa ‘Sakarya’ diye seslendiği iddiası.” teyit.org (blog),
June 12, 2018.
https://teyit.org/cumhurbaskani-erdoganin-bursa-mitingindeki-kalabaliga-sakarya-diye-seslendigi-iddiasi/.
Dahir, Abdi Latif. “Half the World’s Population Used the Internet in 2018 - ITU — Quartz Africa.” Quartz,
December 11, 2018.
https://qz.com/africa/1490997/more-than-half-of-worlds-population-using-the-internet-in-2018/.
Daniels, Laura. “How Russia Hacked the French Election.” POLITICO, April 23, 2017.
https://www.politico.eu/article/france-election-2017-russia-hacked-cyberattacks/.
“Darbe gecesi tavla partisindeymiş.” Sabah, June 22, 2018.
https://www.sabah.com.tr/gundem/2018/06/22/darbe-gecesi-tavla-partisindeymis.
“Digital in 2018: World’s Internet Users Pass the 4 Billion Mark.” We Are Social, January 30, 2018.
https://wearesocial.com/blog/2018/01/global-digital-report-2018.
“Diyarbakır Mesut Yılmaz İlkokulu’nda boş zararla dolaşan kişiler kim?” teyit.org (blog), June 24, 2018.
https://teyit.org/diyarbakir-mesut-yilmaz-ilkokulunda-bos-zararla-dolasan-kisiler-kim/.
Dorf, Michael C., and Sidney Tarrow. “Stings and Scams: ‘Fake News,’ the First Amendment, and the New
Activist Journalism.” SSRN Scholarly Paper. Rochester, NY: Social Science Research Network, January
26, 2017. https://papers.ssrn.com/abstract=2906444.
49
Cyber Governance and Digital Democracy 2019/1
“Emniyet Müdürü: YSK’dan zarf talep etmiş.” HaberTurk, June 24, 2018.
https://www.haberturk.com/diyarbakir-da-muhurlu-bos-zararda-yanlis-anlasilma-2029291.
“Erdoğan: Suikastçı FETÖ’ye Mensup,” December 21, 2016, sec. Türkiye.
https://www.bbc.com/turkce/haberler-turkiye-38394837.
“Erdoğan’ın Meral Akşener hakkında ‘Zilli Meral Kemal’in eteklisi’ dediği iddiası.” teyit.org (blog), May 7, 2018.
https://teyit.org/erdoganin-meral-aksener-hakkinda-zilli-meral-kemalin-eteklisi-dedigi-iddiasi/.
Ferrara, Emilio, Onur Varol, Clayton Davis, Filippo Menczer, and Alessandro Flammini. “The Rise of Social Bots.”
Communications ACM 59, no. 7 (June 2016): 96–104. https://doi.org/10.1145/2818717.
Fiott, Daniel. “A Revolution Too Far? US Defence Innovation, Europe and NATO’s Military-Technological Gap.”
Journal of Strategic Studies 40, no. 3 (April 16, 2017): 417–37.
https://doi.org/10.1080/01402390.2016.1176565.
———. “Europe and the Pentagon’s Third Offset Strategy.” The RUSI Journal 161, no. 1 (January 2, 2016): 26–
31. https://doi.org/10.1080/03071847.2016.1152118.
“Fotoğrafın Muharrem İnce’nin camide halay çektiğini gösterdiği iddiası.” teyit.org (blog), May 9, 2018.
https://teyit.org/fotogran-muharrem-incenin-camide-halay-cektigini-gosterdigi-iddiasi/.
Foy, Henry. “Valery Gerasimov, the General with a Doctrine for Russia.” Financial Times, September 15, 2017.
https://www.ft.com/content/7e14a438-989b-11e7-a652-cde3f882dd7b.
“German Cyber Defense Blends Military and Commerce.” Deutsche Welle. 09 2018.
https://www.dw.com/en/german-cyber-defense-blends-military-and-commerce/a-45636325.
Germano, Sara. “Facebook, Germany to Collaborate Against Election Interference.” Wall Street Journal, January
20, 2019, sec. Business.
https://www.wsj.com/articles/facebook-germany-to-collaborate-against-election-interference-11548004995.
Golovchenko, Yevgeniy, Mareike Hartmann, and Rebecca Adler-Nissen. “State, Media and Civil Society in the
Information Warfare over Ukraine: Citizen Curators of Digital Disinformation.” International Affairs 94, no.
5 (September 1, 2018): 975–94. https://doi.org/10.1093/ia/iiy148.
Grinberg, Nir, Kenneth Joseph, Lisa Friedland, Briony Swire-Thompson, and David Lazer. “Fake News on Twitter
during the 2016 U.S. Presidential Election.” Science 363, no. 6425 (January 25, 2019): 374–78.
https://doi.org/10.1126/science.aau2706.
Groeling, Tim. “Media Bias by the Numbers: Challenges and Opportunities in the Empirical Study of Partisan
News.” Annual Review of Political Science 16, no. 1 (2013): 129–51.
https://doi.org/10.1146/annurev-polisci-040811-115123.
Hern, Alex. “Macron Hackers Linked to Russian-Afliated Group behind US Attack.” The Guardian, May 8, 2017,
sec. World news. https://www.theguardian.com/world/2017/may/08/macron-hackers-linked-to-russian-
afliated-group-behind-us-attack.
Hopkins, Daniel J., and Gary King. “A Method of Automated Nonparametric Content Analysis for Social
Science.” American Journal of Political Science 54, no. 1 (2010): 229–47.
https://doi.org/10.1111/j.1540-5907.2009.00428.x.
50
Cyber Governance and Digital Democracy 2019/1
Howard, Philip N., Aiden Duffy, Deen Freelon, Muzammil M. Hussain, Will Mari, and Marwa Maziad. “Opening
Closed Regimes: What Was the Role of Social Media During the Arab Spring?” SSRN Scholarly Paper.
Rochester, NY: Social Science Research Network, 2011. https://papers.ssrn.com/abstract=2595096.
Howard, Philip N., Bharath Ganesh, Dimitra Liotsiou, John Kelly, and Camille Francois. “The IRA and Political
Polarization in the United States, 2012-2018.” Computational Propaganda Research Project. Oxford,
UK: Oxford Internet Institute, December 2018.
https://comprop.oii.ox.ac.uk/research/ira-political-polarization/.
Howard, Philip N., and Bence Kollanyi. “Bots, #Strongerin, and #Brexit: Computational Propaganda During the
UK-EU Referendum.” SSRN Scholarly Paper. Rochester, NY: Social Science Research Network, June
20, 2016. https://papers.ssrn.com/abstract=2798311.
“İlham Aliyev’den FETÖ Darbe Girişimine Kınama.” TRT Haber, 07 2016.
https://www.trthaber.com/haber/dunya/ilham-aliyevden-feto-darbe-girisimine-kinama-261298.html.
Jensen, Michael. “Russian Trolls and Fake News: Information or Identity Logics?” Journal of International Affairs
71, no. 1.5 (2018): 115–24.
Jones, Marc Owen. “Hacking, Bots and Information Wars in the Qatar Spat.” Washington Post. June 7, 2017.
https://www.washingtonpost.com/news/monkey-cage/wp/2017/06/07/hacking-bots-and-information-
wars-in-the-qatar-spat/.
“Karlov suikastı iddianamesi mahkemede.” NTV, November 23, 2018. https://www.ntv.com.tr/turkiye/karlov-
suikasti-iddianamesi-mahkemede,pG6pIihRcUaQQcbvWqgpEg.
“Karlov suikastı sanığı: Menzil tarikatına bağlıyım.” Sputnik Türkiye, January 11, 2019.
https://tr.sputniknews.com/turkiye/201901111037037495-karlov-suikasti-sanigi-menzil-tarikatini-bagliyim/.
“Karlov suikastını, eski adı El Nusra olan Fetih el Şam üstlendi.” Sputnik Türkiye, December 21, 2016.
https://tr.sputniknews.com/ortadogu/201612211026428039-karlov-fetih-el-sam/.
Kirchgaessner, Stephanie. “Russia Suspected over Hacking Attack on Italian Foreign Ministry.” The Guardian,
February 10, 2017, sec. World news. https://www.theguardian.com/world/2017/feb/10/russia-suspected-
over-hacking-attack-on-italian-foreign-ministry.
Lanoszka, Alexander. “Russian Hybrid Warfare and Extended Deterrence in Eastern Europe.”
International Affairs 92, no. 1 (January 1, 2016): 175–95. https://doi.org/10.1111/1468-2346.12509.
Lazer, David M. J., Matthew A. Baum, Yochai Benkler, Adam J. Berinsky, Kelly M. Greenhill, Filippo Menczer,
Miriam J. Metzger, et al. “The Science of Fake News.” Science 359, no. 6380 (March 9, 2018): 1094–96.
https://doi.org/10.1126/science.aao2998.
Liu, Peng, Wei Chen, Gaoyan Ou, Tengjiao Wang, Dongqing Yang, and Kai Lei. “Sarcasm Detection in Social
Media Based on Imbalanced Classication.” In Web-Age Information Management, edited by Feifei
Li, Guoliang Li, Seung-won Hwang, Bin Yao, and Zhenjie Zhang, 459–71. Lecture Notes in Computer
Science. Springer International Publishing, 2014.
Lowen, Mark. “Hunting for Truth in a Land of Conspiracy.” BBC News. November 15, 2018, sec. Europe.
https://www.bbc.com/news/world-europe-46137139.
51
Cyber Governance and Digital Democracy 2019/1
Matthews, Matt M. “We Were Caught Unprepared: The 2006 Hezbollah-Israeli War.” Fort Leavenworth, Kansas:
US Army Combined Arms Center, 2007. https://www.armyupress.army.mil/Portals/7/combat-studies-
institute/csi-books/we-were-caught-unprepared.pdf.
Mayer, Jane. “How Russia Helped Swing the Election for Trump.” The New Yorker, September 24, 2018.
https://www.newyorker.com/magazine/2018/10/01/how-russia-helped-to-swing-the-election-for-trump.
McCallum, A., X. Wei, and X. Wang. “Topical N-Grams: Phrase and Topic Discovery, with an Application to
Information Retrieval.” In Seventh IEEE International Conference on Data Mining (ICDM 2007)(ICDM),
697–702, 2007. https://doi.org/10.1109/ICDM.2007.86.
“Measuring the Information Society Report 2018.” International Telecommunications Union, 2018.
https://www.itu.int/en/ITU-D/Statistics/Pages/publications/misr2018.aspx.
Mora, Edwin. “Report: Turkey, Syria Helped Keep Islamic State Alive by Buying Their Oil.” Breitbart, July 3, 2018.
https://www.breitbart.com/national-security/2018/07/03/report-turkey-syria-helped-keep-islamic-state-
alive-by-buying-their-oil/.
“Oil Smuggled into Turkey Not Enough to Be Protable: U.S. Ofcial.” Reuters, December 4, 2015.
https://www.reuters.com/article/us-syria-oil-usa-idUSKBN0TN2P920151204.
Osnos, Evan, David Remnick, and Joshua Yaffa. “Trump, Putin, and the New Cold War,” February 24, 2017.
https://www.newyorker.com/magazine/2017/03/06/trump-putin-and-the-new-cold-war.
O’Sullivan, Donie. “Newly Released Facebook Ads Show Russian Trolls Targeted Mexican-Americans after
Trump Election.” CNNMoney, May 10, 2018. https://money.cnn.com/2018/05/10/technology/russian-
facebook-ads-targeted-mexican-americans/index.html.
Polyakova, Alina. “Strange Bedfellows: Putin and Europe’s Far Right.” World Affairs 177, no. 3 (2014): 36–40.
Pomerantsev, Peter. “Authoritarianism Goes Global (II): The Kremlin’s Information War.” Journal of Democracy
26, no. 4 (October 19, 2015): 40–50. https://doi.org/10.1353/jod.2015.0074.
Radcliffe, Damian. “How Local Journalism Can Upend the ‘fake News’ Narrative.” The Conversation, November
27, 2018. http://theconversation.com/how-local-journalism-can-upend-the-fake-news-narrative-104630.
Renz, Bettina. “Russia and ‘Hybrid Warfare.’” Contemporary Politics 22, no. 3 (July 2, 2016): 283–300.
https://doi.org/10.1080/13569775.2016.1201316.
“Reuters Institute Digital News Report.” Oxford, UK: Reuters Institute, Oxford University, 2018.
http://www.digitalnewsreport.org/.
Rid, Thomas. Disinformation: A Primer in Russian Active Measures and Inuence Campaigns, Pub. L. No.
033017, § Select Committee on Intelligence (2017).
Rohde, Markus, Konstantin Aal, Kaoru Misaki, Dave Randall, Anne Weibert, and Volker Wulf. “Out of Syria:
Mobile Media in Use at the Time of Civil War.” International Journal of Human–Computer Interaction 32,
no. 7 (July 2, 2016): 515–31. https://doi.org/10.1080/10447318.2016.1177300.
“Rus Savaş Uçağının Fetö Tarafıdan Düşürüldüğü İddiası.” Milliyet, October 7, 2017.
http://www.milliyet.com.tr/rus-savas-ucaginin-feto-taradan-dusuruldugu-sivas-yerelhaber-2323116/.
“Russia Presents Proof of Turkey’s Role in ISIS Oil Trade.” RT International. December 2, 2015.
https://www.rt.com/news/324263-russia-brieng-isis-funding/.
52
Cyber Governance and Digital Democracy 2019/1
Safarov, Fuad. “Rus uzman, Rus uçağını düşürme emrini kimin verdiğini açıkladı.” Sputnik Türkiye, 12 2016.
https://tr.sputniknews.com/rusya/201612041026128807-rus-uzman-gulen-ucak/.
Saka, Erkan. “Social Media in Turkey as a Space for Political Battles: AKTrolls and Other Politically Motivated
Trolling.” Middle East Critique 27, no. 2 (April 3, 2018): 161–77.
https://doi.org/10.1080/19436149.2018.1439271.
Sanger, David E., and Charlie Savage. “U.S. Says Russia Directed Hacks to Inuence Elections.” The New York
Times, December 21, 2017, sec. U.S. https://www.nytimes.com/2016/10/08/us/politics/us-formally-
accuses-russia-of-stealing-dnc-emails.html.
Saxena, Akrati, Ralucca Gera, and S. R. S. Iyengar. “A Faster Method to Estimate Closeness Centrality Ranking.”
ArXiv:1706.02083 [Physics], June 7, 2017. http://arxiv.org/abs/1706.02083.
Schwirtz, Michael. “German Election Mystery: Why No Russian Meddling?” The New York Times, January 20,
2018, sec. World. https://www.nytimes.com/2017/09/21/world/europe/german-election-russia.html.
Segerberg, Alexandra, and W. Lance Bennett. “Social Media and the Organization of Collective Action: Using
Twitter to Explore the Ecologies of Two Climate Change Protests.” The Communication Review 14, no. 3
(July 1, 2011): 197–215. https://doi.org/10.1080/10714421.2011.597250.
“Şeytan taşladı.” Takvim, 08 2016. https://www.takvim.com.tr/guncel/2016/08/16/seytan-tasladi.
Shane, Scott. “Five Takeaways From New Reports on Russia’s Social Media Operations.” The New York Times,
December 18, 2018, sec. U.S.
https://www.nytimes.com/2018/12/17/us/politics/takeaways-russia-social-media-operations.html.
Simón, Luis. “The ‘Third’ US Offset Strategy and Europe’s ‘Anti-Access’ Challenge.” Journal of Strategic Studies
39, no. 3 (April 15, 2016): 417–45. https://doi.org/10.1080/01402390.2016.1163260.
Snyder, Michael. “Obama Knows That Turkey Is Buying Oil From ISIS And He Isn’t Doing Anything To Stop It.”
InfoWars (blog), November 28, 2015. https://www.infowars.com/obama-knows-that-turkey-is-buying-oil-
from-isis-and-he-isnt-doing-anything-to-stop-it/.
Somaskanda, Sumi. “The Cyber Threat To Germany’s Elections Is Very Real.” The Atlantic, September 20, 2017.
https://www.theatlantic.com/international/archive/2017/09/germany-merkel-putin-elections-cyber-
hacking/540162/.
“Son dakika: YSK, 2018 kesin seçim sonuçlarını açıkladı.” Hürriyet, July 4, 2018.
http://www.hurriyet.com.tr/gundem/son-dakika-ysk-2018-kesin-secim-sonuclarini-acikladi-40886560.
Spohr, Dominic. “Fake News and Ideological Polarization: Filter Bubbles and Selective Exposure on Social
Media.” Business Information Review 34, no. 3 (September 1, 2017): 150–60.
https://doi.org/10.1177/0266382117722446.
Stelzenmüller, Constanze. “The Impact of Russian Interference on Germany’s 2017 Elections.” Brookings (blog),
June 28, 2017. https://www.brookings.edu/testimonies/the-impact-of-russian-interference-on-germanys-
2017-elections/.
Stukal, Denis, Sergey Sanovich, Richard Bonneau, and Joshua A. Tucker. “Detecting Bots on Russian Political
Twitter.” Big Data 5, no. 4 (December 1, 2017): 310–24.
https://doi.org/10.1089/big.2017.0038.
53
Cyber Governance and Digital Democracy 2019/1
Tandoc Jr, Edson C., Zheng Wei Lim, and Richard Ling. “Dening ‘Fake News.’” Digital Journalism 6, no. 2
(February 7, 2018): 137–53. https://doi.org/10.1080/21670811.2017.1360143.
Taub, Ben. “The ISIS Oil Trade, from the Ground Up,” December 4, 2015.
https://www.newyorker.com/news/news-desk/the-isis-oil-trade-from-the-ground-up.
“Text of Newly-Approved Russian Military Doctrine.” Carnegie Endowment for International Peace, February 5, 2010.
https://carnegieendowment.org/2010/02/05/text-of-newly-approved-russian-military-doctrine-pub-40266.
“The ‘Lisa Case’: Germany as a Target of Russian Disinformation.” NATO, 2016. http://www.nato.int/docu/
review/2016/Also-in-2016/lisa-case-germany-target-russian-disinformation/EN/index.htm.
Thieltges, Andree, Orestis Papakyriakopoulos, Juan Carlos Medina Serrano, and Simon Hegelich. “Effects of
Social Bots in the Iran-Debate on Twitter.” ArXiv:1805.10105 [Cs], May 25, 2018.
http://arxiv.org/abs/1805.10105.
Thomas, Timothy. “Russia’s 21st Century Information War: Working to Undermine and Destabilize Populations.”
Riga: NATO STRATCOM, 2015. https://www.stratcomcoe.org/timothy-thomas-russias-21st-century-
information-war-working-undermine-and-destabilize-populations.
Tsvetkova, Maria, and Lidia Kelly. “Russia Says It Has Proof Turkey Involved in Islamic State Oil Trade.” Reuters,
December 2, 2015.
https://www.reuters.com/article/us-mideast-crisis-russia-turkey-idUSKBN0TL19S20151202.
“Turkey Conrms Cancellation of $3.4 Billion Missile Defence...” Reuters, November 18, 2015.
https://www.reuters.com/article/us-turkey-china-missile-idUSKCN0T61OV20151118.
“UK Political Parties Warned of Russian Hacking Threat: Report.” Reuters, March 12, 2017.
https://www.reuters.com/article/us-britain-russia-cybercrime-idUSKBN16J0OE.
Visentin, Marco, Gabriele Pizzi, and Marco Pichierri. “Fake News, Real Problems for Brands: The Impact of
Content Truthfulness and Source Credibility on Consumers’ Behavioral Intentions toward the Advertised
Brands.” Journal of Interactive Marketing 45 (February 1, 2019): 99–112.
https://doi.org/10.1016/j.intmar.2018.09.001.
“Voter Turnout in Turkey Elections Was 87 Percent: State Broadcaster.” Reuters, June 24, 2018.
https://www.reuters.com/article/us-turkey-election-turnout-idUSKBN1JK0SP.
Wagnsson, Charlotte, and Maria Hellman. “Normative Power Europe Caving In? EU under Pressure of Russian
Information Warfare.” JCMS: Journal of Common Market Studies 56, no. 5 (2018): 1161–77.
https://doi.org/10.1111/jcms.12726.
Wang, Alex Hai. “Detecting Spam Bots in Online Social Networking Sites: A Machine Learning Approach.” In
Data and Applications Security and Privacy XXIV, edited by Sara Foresti and Sushil Jajodia, 335–42.
Lecture Notes in Computer Science. Springer Berlin Heidelberg, 2010.
Wasserman, Herman. “Fake News from Africa: Panics, Politics and Paradigms.” Journalism, December 17,
2017, 1464884917746861. https://doi.org/10.1177/1464884917746861.
Way, Lucan A. “The Limits of Autocracy Promotion: The Case of Russia in the ‘near Abroad.’” European Journal
of Political Research 54, no. 4 (2015): 691–706. https://doi.org/10.1111/1475-6765.12092.
54
Cyber Governance and Digital Democracy 2019/1
Williams, Oscar. “Russia Is Targeting UK Infrastructure through Supply Chains, NCSC Warns.” New Statesman,
April 6, 2018. https://tech.newstatesman.com/business/russia-uk-critical-infrastructure.
Wintour, Patrick. “Russian Bid to Inuence Brexit Vote Detailed in New US Senate Report.” The Guardian,
January 10, 2018, sec. World news. https://www.theguardian.com/world/2018/jan/10/russian-inuence-
brexit-vote-detailed-us-senate-report.
Wintour, Patrick, and Andrew Roth. “Russia Summons Dutch Ambassador over Hacking Revelations.” The
Guardian, October 8, 2018, sec. World news. https://www.theguardian.com/world/2018/oct/08/russia-
summons-dutch-ambassador-over-hacking-revelations.
Yinanç, Barçın. “Poor Media Literacy ‘Making Turks Vulnerable to Fake News’ - Turkey News.” Hürriyet Daily
News. December 10, 2018. http://www.hurriyetdailynews.com/poor-media-literacy-making-turks-
vulnerable-to-fake-news-139582.
Yılmaz, Mehmet Y. “Davutoğlu bıçağın sırtında.” Hürriyet, July 19, 2016.
http://www.hurriyet.com.tr/yazarlar/mehmet-y-yilmaz/davutoglu-bicagin-sirtinda-40154956.
Yourish, Karen, and Troy Griggs. “8 U.S. Intelligence Groups Blame Russia for Meddling, but Trump Keeps
Clouding the Picture.” The New York Times, July 16, 2018, sec. U.S. https://www.nytimes.com/
interactive/2018/07/16/us/elections/russian-interference-statements-comments. html, https://www.
nytimes.com/interactive/2018/07/16/us/elections/russian-interference-statements-comments.html.
Zannettou, Savvas, Tristan Cauleld, Emiliano De Cristofaro, Michael Sirivianos, Gianluca Stringhini, and
Jeremy Blackburn. “Disinformation Warfare: Understanding State-Sponsored Trolls on Twitter and Their
Inuence on the Web.” ArXiv:1801.09288 [Cs], January 28, 2018. http://arxiv.org/abs/1801.09288.
Cyber Governance and Digital Democracy 2019/1
March 2019
H. Akın Ünver | EDAM, Oxford CTGA & Kadir Has University
Cyber Governance and Digital Democracy 2019/1
Russian Digital Media and
Information Ecosystem in Turkey
