Content uploaded by Adrian Rauchfleisch
Author content
All content in this area was uploaded by Adrian Rauchfleisch on Apr 06, 2020
Content may be subject to copyright.
Studies in Communication Sciences 18.1 (2018), pp. 69–85
https://doi.org/10.24434/j.scoms.2018.01.005
© 2018, the a uthors. This w ork is licensed under the “Creative Co mmons Attribution –
NonCo mmercial – NoDer ivatives 4. 0 International” licen se (CC BY-NC- ND 4.0).
Digital astroturfing in politics: Definition, typology,
and countermeasures
Marko Kovic, Zurich Institute of Public Affairs Research (ZIPAR), Switzerland*
Adrian Rauchfleisch, National Taiwan University and Zurich Institute of Public Affairs Research
( ZIPAR), Switzerland
Marc Sele, Zurich Institute of Public Affairs Research (ZIPAR), Switzerland
Christian Caspar, Zurich Institute of Public Affairs Research (ZIPAR), Switzerland
*Corresponding author: markokovic@gmail.com
Abstract
In recent years, several instances of political actors who created fake grassroots activity on the Internet
have been uncovered. We propose to call such fake online grassroots activity digital astroturfing, and we
define it as a form of manufactured, deceptive and strategic top-down activity on the Internet initiated by
political actors that mimics bottom-up activity by autonomous individuals. The goal of this paper is to lay
out a conceptual map of the phenomenon of digital astroturfing in politics. To that end, we introduce, first,
a typology of digital astroturfing according to three dimensions (target, actor type, goals), and, second, the
concept of digital astroturfing repertoires, the possible combinations of tools, venues and actions used in
digital astroturfing efforts. Furthermore, we explore possible restrictive and incentivizing countermeasures
against digital astroturfing. Finally, we discuss prospects for future research: Even though empirical re-
search on digital astroturfing is difficult, it is neither impossible nor futile.
Keywords
political communication, Internet, public opinion, astroturfing, bots
1 Introduction
Not long after the advent of the World
Wide Web, there were high hopes that this
new form of access to the Internet might
prove a boon for democracy. Thanks to the
World Wide Web, citizens could gain ac-
cess to politically relevant information far
easier than ever before. Even more than
that: The World Wide Web seemed to car-
ry the potential to make the elusive ideal
of deliberative democracy a reality (or, at
least, to make instances of deliberative
democratic participation easier) (Buch-
stein, 1997; Dahlberg, 2001; Rheingold,
1993). The positive force of the Internet,
the hope went, would make democratic
participation easier than ever and thus it
would promote democratic values, even in
non-democratic countries.
The rapid progress both of Internet
penetration and of Internet-related tech-
nologies has created circumstances in
which the democratic online revolution as
envisioned in the World Wide Web’s early
days could, in principle, take place. Access
to the Internet is increasingly common
even in relatively poor countries, and both
the means for accessing the Internet as
well as the ways in which users can access
political information and engage in politi-
cal discourse have proliferated.
Of course, the World Wide Web has
not produced the democratic utopia that
some might initially have hoped for. But
Internet has certainly had an impact on
politics and, more specifically, on political
discourse. Citizens do indeed seek politi-
cal information online and they do engage
in new forms of digital political discourse.
Unfortunately, the relative ease of engag-
ing in political discourse on the Internet
makes it also relatively easy to undermine,
or at least negatively affect, that very dis-
course. One perfidious way in which this
can happen: Political actors can create
online activity that seems like authentic
70 Kovic et al. / Studies in Communication Sciences 18.1 (2018), pp. 69–85
activity by regular citizens, when it is, in
reality, anything but.
Online users who are not what they
pretend to in order to ruse real people are
the opposite of what is hoped for in the
ideal of deliberative democracy. Fake on-
line users are not interested in participat-
ing in rational political discourse, but rath-
er in deliberately and actively skewing the
discourse according to the goals they are
pursuing. This is not an entirely new tool.
So-called astroturfing, fake grassroots
campaigns and organizations that are
created and backed by political actors or
businesses, have a long tradition (Lyon &
Maxwell, 2004; Walker & Rea, 2014). Astro-
turfing on the Internet is more problemat-
ic than the traditional kind: Digital astro-
turfing is cheaper, has a greater scope, and
is potentially much more effective than
regular astroturfing.
1.1 A new era of political astroturfing
In the wake of the 2016 U.S. presidential
election, the National Intelligence Coun-
cil at the Office of the Director of National
Intelligence released a declassified version
of an assessment of Russian state-spon-
sored activities during the election (Na-
tional Intelligence Council, 2017). In that
report, the National Intelligence Council
describes the Russian interference in the
2016 election as a continuation of co-
vert influence campaigns that Russia and
the Soviet Union were conducting in the
past. However, the report also mentions
the activities of the St. Petersburg-based
“Internet Research Agency” as part of the
Russian influence campaign. The Internet
Research Agency is a state-sponsored on-
line astroturfing organization (Chen, 2015;
MacFarquhar, 2018) specialized in creating
and maintaining sock puppets, fake online
personae that are mimicking regular users
(Bu, Xia, & Wang, 2013). The astroturfing
campaign conducted by the Internet Re-
search Agency was great in scope. In early
2018, Twitter notified over 1.4 million of its
US-based users that they had interacted
with sock puppet accounts likely affiliated
with the Internet Research Agency (Twitter
PublicPolicy, 2018). But the campaign was
also sophisticated, as evidenced by the
persuasive power of some of sock puppets,
such as the influential, yet fictitious “Jenna
Abrams” (Colins & Cox, 2017). The astro-
turfing campaign of 2016 was so pervasive
that it was proposed that the US create a
military unit in order to defend from such
campaigns (Hart & Klink, 2017).
Digital astroturfing is not limited to
U.S. elections. Instances of digital astro-
turfing have been documented in many
countries (Wooley, 2016). Astroturfing ef-
forts can be conducted by outside political
actors, such as in the 2017 French election
(Ferrara, 2017) or in the 2016 Brexit vote in
the United Kingdom (Bastos and Mercea,
2017 ; Llewellyn et al., 2018), or they can
be campaigns of political actors within
their own country, such as the systematic
digital astroturfing strategy of China (King
et al., 2017). Digital astroturfing is also not
limited to manually curated sock puppets.
For example, many digital astroturfing ef-
forts are conducted with automated bots,
whereby computer software impersonates
humans by acting and reacting in as “nat-
ural” ways as possible (Bessi & Ferrara,
2016; Shao et al., 2017; Wooley, 2016).
1.2 The goal of this paper
The phenomenon of digital astroturfing
is of obvious importance. Digital astro-
turfing represents an inherently deceptive
political activity that is detrimental to the
democratic discourse and to democratic
decision-making. In view of this problem,
the goal of this present paper is threefold.
First, we want to define digital astro-
turfing in a generalized, universally ap-
plicable manner. Digital astroturfing is a
complex political activity. Second, we pro-
pose a typology of different forms of digital
astroturfing. That typology is based on the
type of actor that is engaging in the astro-
turfing effort; the persuasion target that
is to be affected by the astroturfing effort,
and the goal that is pursued by the astro-
turfing effort. Third, we discuss possible
countermeasures to either prevent digital
astroturfing or to lessen its impact.
Kovic et al. / Studies in Communication Sciences 18.1 (2018), pp. 69–85 71
2 A generalized definition of digital
astroturfing in politics
In the previous section, we have provi-
sionally introduced digital astroturfing as
a type of political online activity that pre-
tends to be the activity of individual, regu-
lar online users, when it is in fact activity
created by political actors. Such a provi-
sional definition already describes the na-
ture of digital astroturfing to some degree,
but it is derived from an intuitive inductive
process of describing what different cases
of digital astroturfing have in common. A
more precise deductive definition is need-
ed in order to describe digital astroturfing
in a generalized way, independent from
specific singular occurrences of digital
astroturfing. To that end, we propose the
following definition:
Digital astroturfing is a form of manu-
factured, deceptive and strategic top-down
activity on the Internet initiated by politi-
cal actors that mimics bottom-up activity
by autonomous individuals.
In this definition, digital astroturfing
as a form of political activity consists of
five necessary conditions: It takes place on
the Internet, it is initiated by political ac-
tors, it is manufactured, it is deceptive and
it is strategic. Digital astroturfing is manu-
factured in that it is not an honest expres-
sion of autonomous individual opinions,
but rather activity created to mimic hon-
est expression of autonomous individual
opinions. It is deceptive, because the goal
of digital astroturfing is to trick the target,
usually the public at large (including in-
dividuals and small groups), into believ-
ing that the manufactured activity is real.
It is strategic, because the political actors
who engage in digital astroturfing pursue
certain goals in doing so. This definition
is relatively simple, but in simplicity, we
believe, lies analytical clarity. So far, only
few authors have attempted to define dig-
ital astroturfing. The most relevant such
attempt is a recent conceptual study by
Zhang, Carpenter, and Ko (2013). In that
study, the authors define online astro-
turfing (the authors use the term online
rather than digital astroturfing) as follows:
“Therefore, we define online astroturfing
as the dissemination of deceptive opinions
by imposters posing as autonomous indi-
viduals on the Internet with the intention
of promoting a specific agenda” (p. 3).
This definition is similar to the one
we introduce in three respects. First, we
have adopted the authors’ very apt no-
tion of “autonomous individuals” into our
own definition because it describes very
precisely what kind of deception is taking
place. We expand the descriptive notion
of autonomous individuals by contrasting
bottom-up and top-down activity; auton-
omous individuals engage in bottom-up
activity, and digital astroturfing is top-
down activity pretending to be bottom-up
activity. Second, and this is nontrivial, the
authors define digital astroturfing as an
activity that is taking place on the Inter-
net. And third, they describe digital astro-
turfing as a deceptive activity. We agree
with those three aspects as necessary
conditions, but the definition lacks accu-
racy and generalizability. The definition
is inaccurate because it implies that the
relevant actors are the imposters posing
as autonomous individuals. However, the
truly relevant actors are not the imposters
themselves, but rather the political actors
initiating the astroturfing efforts. The im-
posters, such as, for example, the employ-
ees of the Internet Research Institute in St.
Petersburg, are simply performing tasks
on behalf of political actors; they are the
means to an end. It is possible that in some
cases, the actors who initiate the digital
astroturfing in pursuit of some goals will
also actively carry out the digital astroturf-
ing effort themselves. For example, a pol-
itician might himself create a number of
fake Twitter accounts with which he would
then follow his own, real account in order
to feign popularity of his account. But even
in such a scenario, the distinction between
who initiates the digital astroturfing effort
is clearly separate from how the digital as-
troturfing effort is actually performed. The
origin of digital astroturfing lies with the
political actors who pursue certain goals
by engaging digital astroturfing, and the
means of actually carrying out the digital
astroturfing effort can be, and in many in-
72 Kovic et al. / Studies in Communication Sciences 18.1 (2018), pp. 69–85
stances probably are, separate from those
political actors.
Furthermore, the above definition by
Zhang et al. (2013) is not fully generaliz-
able because it does not describe a general
principle, but rather one empirical scenar-
io of digital astroturfing: Cases in which
individuals create fake online personæ,
the so-called sock puppets, in order to
communicate certain points of view. While
such a digital astroturfing scenario might
be a very important one, it is not the only
possible one as we will later point out in
our typology.
2.1 The difference between regular
astroturfing and digital astroturfing
Ours is not the first study to tackle the
phenomenon of digital astroturfing. How-
ever, in the previous section, we only dis-
cuss one definition of digital astroturfing
from one recent study. We are not willing-
ly disregarding other definitions of digi-
tal astroturfing, but the status quo of the
conceptual work on digital astroturfing is
rather limited: Most studies that look at
digital astroturfing do not operate with
an explicit definition of digital astroturf-
ing, but they tend to extend the concept of
regular astroturfing into the online realm
instead. For example, Mackie (2009) de-
scribes astroturfing on the Internet in the
following terms: “Astroturfing is the use of
paid agents to create falsely the impres-
sion of popular sentiment (the grass roots
are fake, thus the term astroturf, which
is artificial grass)” (p. 32). Even though
the author goes on to argue that and why
the Internet is vulnerable to astroturfing,
he does not propose a separate, explicit
definition of astroturfing that is occurring
on the Internet. This implies that astro-
turfing is astroturfing, no matter where it
takes place. Furthermore, Mackie (2009)
is only considering paid agents to be in-
dicative of astroturfing. We believe that
unpaid agents can be involved in digital
astroturfing as well, as in the case of sym-
pathizers who become involved in digital
astroturfing because they honestly sup-
port a cause; we delve into this issue in the
sections below. In a similar vein as Mackie
(2009), Boulay (2012) explicitly addresses
the potentials of digital astroturfing, but
she operates with the definition of regular
astroturfing: « L’astroturfing est une straté-
gie de communication dont la source est
occultée et qui prétend, à tort, être d’ori-
gine citoyenne » (p. 201).
Is a conceptual differentiation be-
tween regular and digital astroturfing, as
indirectly proposed by our proposal of an
explicit definition of digital astroturfing,
reasonable given the fact that several au-
thors analyse digital astroturfing, but only
apply a definition of regular astroturfing?
We believe it is, because digital astroturf-
ing conceptually clearly differs from regu-
lar astroturfing. In our definition of digital
astroturfing, we identify five necessary
conditions: The activities take place on
the Internet, they are initiated by political
actors, and they are manufactured, decep-
tive and strategic. If digital astroturfing is
to be understood simply as the extension
of astroturfing onto the Internet, then four
of these five criteria have to apply to reg-
ular astroturfing; obviously, regular astro-
turfing does not have to fulfil the condition
of taking place on the Internet. However,
regular astroturfing is not adequately de-
scribed by the remaining four necessary
conditions. In regular astroturfing, some
political actors are present, and those ac-
tors act strategically in that they pursue
goals. But those actors are not necessari-
ly always initiating the activities, they are
not necessarily directly manufacturing
the activities, and they are not necessari-
ly deceitful about their involvement in the
activities. Vergani (2014) makes this point
clear by introducing the concept of grass-
roots orchestra, a situation where genuine
grassroots activity takes place, but over
time, political actors get involved in that
genuine activity and they influence, if not
usurp the grassroots activity in order to
steer it in a direction beneficial to them-
selves. A recent prominent example of a
grassroots orchestra is the Tea Party move-
ment in the United States: It seems that the
Tea Party was an amalgamation of genuine
grassroots activity with coordinating and
mobilizing influence by political actors
(Formisano, 2012).
Kovic et al. / Studies in Communication Sciences 18.1 (2018), pp. 69–85 73
The distinction between regular and
digital astroturfing, in summary, lies in
the fact that digital astroturfing is a clearly
demarcated dichotomous phenomenon,
while regular astroturfing is not; in regu-
lar astroturfing, varying degrees of genu-
ine grassroots components can be pres-
ent. Of course, this conceptual difference
between digital astroturfing and regular
astroturfing does not preclude the possi-
bility that there are instances of astroturf
campaigns where regular and digital astro-
turfing are combined in order to achieve as
great an impact as possible (Walker, 2014).
2.2 A typology of digital astroturfing
With our generalized definition of digital
astroturfing, it is possible to reflect upon
what digital astroturfing can empirically
look like. Because of the clandestine na-
ture of digital astroturfing, it could be very
misleading to base a typology of digital
astroturfing solely on the exposed cases of
digital astroturfing, since there is no way
of knowing whether those exposed cases,
such as the briefly mentioned ones in the
introduction, cover all possible types of
digital astroturfing. For that reason, we opt
for a theoretico-deductive typology of dig-
ital astroturfing: It is not possible to induc-
tively gather information on all instances
of digital astroturfing, but it is possible to
define a set of dimensions that will plau-
sibly exhaust the possible types of digital
astroturfing.
We propose three dimensions that
encompass the different possible types
of digital astroturfing: The political actors
who engage in digital astroturfing, the tar-
get of the digital astroturfing effort, and
the goal of digital astroturfing.
The first dimension of digital astro-
turfing types are the political actors who
engage in digital astroturfing. It is reason-
able use the category of the political ac-
tors that engage in digital astroturfing as a
starting point for a typology of digital as-
troturfing – after all, political actors are the
conditio sine qua non of digital astroturf-
ing; political actors are the initiating force
behind every digital astroturfing effort. We
distinguish between five categories of po-
litical actors: governments, political par-
ties, interest groups, individual politicians,
and individual citizens. The first two actor
categories, governments and political par-
ties, are self-explanatory. The third actor
category, interest groups, perhaps less so.
We use the term interest group to mean a
group that pursues political goals, but is
neither part of the government nor a po-
litical party. This can include loose groups
of like-minded people, nonprofit organi-
zations, businesses, and so forth. Interest
groups are thus groups that pursue polit-
ical interests, and to that end, they inter-
act with the political system. Sometimes,
interest group activity will be in the form
of discrete, non-public direct interaction
with members of government or parlia-
ment; i. e., in the form of lobbying. But in
other situations, interest groups are advo-
cating their position publicly. The nature
of interest groups can be very varied; typi-
cal examples of interest groups in Western
countries are corporations as employer
interest groups and unions as employee
interest groups (Walker, 1983). In the con-
text of interest groups as political actors,
it is important to keep in mind that digi-
tal astroturfing is concerned with political
goals. There is a prominent phenomenon
of non-political, strictly commercial qua-
si-astroturfing that some businesses en-
gage in: The practice of faking online con-
sumer reviews to make one’s products or
services appear more popular than they
actually are or to make competitors’ prod-
ucts or services appear less popular than
they actually are (Yoo & Gretzel, 2009).
This type of deceptive online activity by
commercial interest groups is not digital
astroturfing, because the nature of the ac-
tivity that the actors pursue is not political.
If, however, a business is pursuing political
goals and is engaging in digital astroturfing
to that end, then we subsume that business
in the interest group category of initiating
political actors. The fourth category of
political actors are individual politicians.
Individual politicians are, for the most
part, members of groups such as political
parties, but there are situations in which
individual politicians act on their own
in terms of digital astroturfing. We men-
tioned one such scenario in the introduc-
74 Kovic et al. / Studies in Communication Sciences 18.1 (2018), pp. 69–85
tion: A politician who, as part of his elec-
tion campaign, paid for Twitter followers.
The second dimension of digital as-
troturfing types is the persuasion target
of digital astroturfing. We differentiate
among two general persuasion targets,
the public and specific political actors. It
is quite obvious to imagine the general
public as one target of digital astroturfing,
since most known cases of digital astro-
turfing are cases of public digital astroturf-
ing activities. Of course, political actors
can also perceive public digital astroturf-
ing and be affected by it. But another cate-
gory of digital astroturfing is possible, one
that is specifically aimed at political actors
and is not public.
The third dimension of the typology
of digital astroturfing are the goals the po-
litical actors pursue by engaging in digital
astroturfing. There are, of course, a myriad
specific goals that specific political actors
can pursue in specific political contexts.
The idea of this third dimension of the dig-
ital astroturfing typology is not to describe
every possible situational configuration,
but rather to meaningfully divide the spec-
trum of the goals of digital astroturfing into
a small number of categories. We identify
two such categories of goals: support for
or opposition to policy, and support for
or opposition to political actors. We pro-
pose only two categories, since the goal
of digital astroturfing is to communicate
a valence either towards policy or towards
political actors, and being positive or neg-
ative is simply the property of the valence.
As for the objects of the valence, political
actors who engage in digital astroturfing
will either want to influence public opin-
ion or specific political actors with regard
to certain policy issues, either to support
their own position on those issues, or to
oppose the position of other political ac-
tors. The other broad category of goals
is support for or opposition to political
actors. When political actors pursue this
goal, they will either do so in order to feign
public support for themselves, or to make
other political actors appear unpopular.
In public discussions of digital astro-
turfing, instances of digital astroturfing are
sometimes described as trolling. To equate
digital astroturfing with trolling in such a
manner is a grave conceptual error. We
have defined digital astroturfing as man-
ufactured, deceptive and strategic activity
on the Internet initiated by political actors.
Of these five necessary conditions, trolling
only satisfies two: Trolling takes place on
the Internet, and there is a strategic com-
ponent to it. However, trolling is not initi-
ated on behalf of political actors, it is not
manufactured and it is only somewhat de-
ceptive.
Trolling is best understood as mali-
cious, disruptive or disinhibited (Suler,
2004) online behavior by individuals who
engage in the activity of trolling of their
own individual volition (Hardaker, 2010).
Trolling often also appears as a social,
coordinated activity (MacKinnon & Zuck-
erman, 2012). Still, conceptually, digital
astroturfing is completely separate from
trolling, and mixing these concepts should
be avoided, not least because talking about
digital astroturfing in terms of trolling trivi-
alizes digital astroturfing: Digital astroturf-
ing is a deceptive effort aimed at the public
and launched by political actors. However,
there is one potential connection between
trolling and digital astroturfing: Cases of
digital astroturfing in which the deceptive
activity takes the form of trolling. In such
cases, the trolling that is taking place is not
authentic, but manufactured and strate-
gically used by political actors. We believe
this conceptual distinction is important.
Not every activity that is widely described
as trolling is in fact trolling.
However, it is possible that political
actors engage in digital astroturfing and
the honest bottom-up activity they mim-
ic is trolling. This behavior can still be de-
scribed in terms of our typology, because
a deceptive and strategic use of trolling
is meant to disrupt some public debate
and expression of opinion that a political
actor objects to, and those public debates
and expressions of opinion will, naturally,
refer to some policy or to some political
actors. Fake trolling thus has the goal of
indirectly stymieing expressions of sup-
port for or criticism against policy or for
political actors. For example, the Chinese
government regularly tries to distract the
Kovic et al. / Studies in Communication Sciences 18.1 (2018), pp. 69–85 75
public and change the subject of critical
discussions online with fabricated user
comments (King et al., 2017).
Combining the three dimensions of
the typology leads to the sixteen different
types of digital astroturfing, as reported
in Table 1, provide a useful framework for
analyzing individual cases of digital astro-
turfing as well as for guiding expectations
about digital astroturfing efforts in gen-
eral. Digital astroturfing can occur, as we
argue in the introduction, in very different
political contexts, and as a consequence,
it can be challenging to analyze separate
cases. With our typology, we provide, in
essence, a useful heuristic – a way to think
about digital astroturfing.
2.3 Digital astroturfing repertoires
Our typology consists of sixteen digital as-
troturfing scenarios that describe which
actors pursue what kind of goal with their
digital astroturfing efforts. However, the
different scenarios do not automatically
imply what specific measures the politi-
cal actors take in order to carry out their
digital astroturfing efforts. These specif-
ic efforts consist of three elements: The
specific digital astroturfing tools used, the
specific venues where these tools are ap-
plied, and the specific actions that are tak-
en with those tools in those venues.
In social movement research, the
concept of protest repertoires is used to
describe which tools social movements
use in which contexts (Della Porta, 2013;
Tarrow, 2011). It is useful to use an analo-
gous concept for digital astroturfing: The
concept of digital astroturfing repertoires.
Digital astroturfing repertoires cannot be
defined universally, because the tools and
venues available for digital astroturfing are
very much bound by time and space and
are likely to change – just like the protest
repertoires of social movements are bound
by time and space and likely to change over
time (Biggs, 2013). Also, describing digital
astroturfing repertoires is ultimately an
inductive task of continued observation of
digital astroturfing cases. Even though we
cannot specify a definitive and final list of
context specific digital astroturfing reper-
toires, there are some typical repertoires,
we believe, that encompass a large portion
of contemporary digital astroturfing rep-
ertoires.
Digital astroturfing repertoires can be
thought of, as mentioned above, as combi-
nations of tools, venues and actions. Some
typical digital astroturfing tools are sock
puppets, click farms, and sympathizers.
Sock puppets are, as we describe in the in-
troduction, fake online personae (Bu et al.,
2013) that can be used for a variety of pur-
poses. Click farms are a relatively recent
variant of sock puppets tied to the rise of
social media. Click farms provide services
that, essentially, boil down to faux social
media activity, such as fake followers on
Twitter or fake likes on Facebook (Clark,
2015). Both sock puppets and click farms
can contain various degrees of automa-
Table 1: Typology of digital astroturfing
Initiating political actor Persuasion target Goal: Support for or opposition to:
Policy Political actors
Government The Public 1 2
Political party 3 4
Individual politician 5 6
Interest group 7 8
Government Political actors 9 10
Political party 11 12
Individual politician 13 14
Interest group 15 16
Note. Each number represents one digital astroturfing scenario as a combination of three dimensions. Reading example: The digital astroturf-
ing type with the number 7 describes a digital astroturfing scenario in which an interest group has the goal of influencing the public opinion
on some policy issue.
76 Kovic et al. / Studies in Communication Sciences 18.1 (2018), pp. 69–85
tion through bots, programs that operate
automatically on behalf of agents (Geer,
2005). For example, fake user profiles used
in click farms are usually created manual-
ly, but the accounts are subsequently op-
erated automatically. A lot of the current
research in the area of digital astroturfing
is focused on bots. Bots are certainly im-
portant, not least quantitatively, since
creating and deploying bots is fairly easy
and cheap. However, bots are still, for the
most part, not able to persuasively mim-
ic humans (Stieglitz et al., 2017), limiting
their potential persuasive power. Bots as
automated sock puppets or as automat-
ed click farms are certainly important,
but manually curated and operated sock
puppets and click farm profiles still mat-
ter. The idea of sympathizers as digital as-
troturfing tools might, prima facie, seem
a misclassification. After all, sympathiz-
ers are real people who honestly hold the
opinions they put forward. The question
whether sympathizers as digital astroturf-
ing tools honestly believe in what they do
is conceptually irrelevant. Political actors
are still orchestrating, i. e., manufacturing
some activity that is supposed to look like
non-manufactured, spontaneous activity;
that activity is deceptive (the target of the
astroturfing attempt has no knowledge of
its manufactured origins); and the activi-
ty is strategic from the point of view of the
political actor. Whether sympathizers as
the tools of digital astroturfing sympathize
with the political actors on whose behalf
they act has no conceptual impact on the
nature of the activity at hand. Howard
(2006) describes such a case in which sym-
pathizers were mobilized in an orches-
trated attempt without even knowing that
they were instrumentalized by a political
actor. Finally, paid supporters are similar
to sympathizers: They both engage in ac-
tivities on behalf of political actors without
openly declaring so, and they both do so
with their true identities rather than with
sock puppets. However, paid supporters
are not (only) acting out of personal con-
viction, but out of pecuniary interests –
they are being paid for their activities.
Some typical venues of digital astro-
turfing are social media, websites, com-
ment sections (predominantly those of
news websites), and emails. We under-
stand social media as platforms where
users, be they individuals, groups or or-
ganizations, can connect with other users
(Boyd & Ellison, 2007) and where the us-
ers themselves are the primary creators of
content (Kaplan & Haenlein, 2010). Typical
examples of social media are services such
as Facebook, Twitter, Instagram, Snapchat,
and so forth. We also consider online pe-
titions to be a form of social media; after
all, when signing online petitions, users
connect with each other and by doing so
create a certain type of content. Websites
are separate from social media in that the
users who curate websites do not simply
create profiles on social media sites, but
rather create separately hosted, indepen-
dent content. Comment sections are the
sections on websites that allow website
visitors to leave written comments below
content on the website, and sometimes,
to like or dislike existing comments. In the
context of digital astroturfing, comment
sections of news websites are potentially
one of the most relevant venues. In prin-
ciple, comment sections could be thought
of as just another form of social media, but
we treat them as a separate venue. Users
are the primary creators of content on so-
cial media, but not in comment sections
of news websites – in comment sections,
users only react to content created by jour-
nalists. The final venue of digital astroturf-
ing are direct messages. Describing direct
messages (such as emails, online contact
forms, private messages on social media,
and so forth) as venues is perhaps some-
what confusing, because direct messages
aren’t public places on the Internet. But
what we mean by venue is not a public lo-
cation, but rather the place where the digi-
tal astroturfing tools are applied. However,
direct messages are a somewhat special
case of digital astroturfing, because direct
messages can be targeted at specific polit-
ical actors, without the general public tak-
ing notice.
Finally, there are only two digital as-
troturfing actions: Actively creating con-
tent, or passively signaling (dis-)approv-
al. The difference between active content
Kovic et al. / Studies in Communication Sciences 18.1 (2018), pp. 69–85 77
creation and passive signaling of (dis-)ap-
proval can be exemplified with comment
sections on websites. Writing a comment
constitutes an active creation of content,
whereas liking or disliking is merely a pas-
sive signaling of approval or disapproval.
The different tools, the different venues
and the different actions of digital astro-
turfing are summarized in Table 2.
A digital astroturfing repertoire may con-
sist of any combination of tools, venues
and actions from Table 2. For example,
the combination of sock puppets with
comment sections (of news websites) and
creating content is the digital astroturfing
repertoire used by the Russian Internet
Research Agency described in the intro-
duction. The combination of click farms
with social media and the signaling of ap-
proval is the digital astroturfing repertoire
used by the Swiss politician mentioned in
the introduction who bought fake twitter
followers. A repertoire consisting of sym-
pathizers combined with direct messages
and creating content is a digital astroturf-
ing repertoire sometimes applied by polit-
ical candidates running for office who en-
courage sympathizers to send pre-written
emails to friends and to newspaper editors
(Klotz, 2007).
3 Countermeasures
Digital astroturfing is a clandestine activ-
ity, and as such, it is difficult to devise de-
fense mechanisms against it. While there
is probably no way to completely prevent
digital astroturfing from occurring, there
are ways to limit the probability of digital
astroturfing occurring or, at least, of dig-
ital astroturfing being effective in some
venues. More specifically, we identify
two broad strategies of countermeasures:
Restrictive countermeasures and incen-
tivizing countermeasures. Restrictive
countermeasures are measures that limit
the spectrum of possible online activities,
and this limitation of online activities in
general also limits digital astroturfing ac-
tivities. Incentivizing countermeasures, in
contrast, are measures that are conducive
to honest forms of online activity. Incen-
tivizing countermeasures do not neces-
sarily prevent digital astroturfing, but by
rewarding more openly honest activity,
incentivizing countermeasures make dig-
ital astroturfing less prominent and thus,
potentially, less impactful.
Before we delve into both types of
countermeasures in more detail, it is in
order to justify why exactly we propose
countermeasures in the first place – the
fact that we actively discuss countermea-
sures implies that we regard digital astro-
turfing to be a problem.
There are two major reasons why
digital astroturfing is an important and
problematic phenomenon, a normative
one and an empirical one. Normatively,
digital astroturfing is challenging because
the act of engaging in digital astroturf-
ing, essentially, amounts to subterfuge,
and subterfuge is nothing more than a
form of lying (Grover, 1993). Empirically,
digital astroturfing almost certainly has
some real-world impact. While it is, for
obvious reasons, very difficult to quantify
the effects of digital astroturfing on pub-
lic opinion or on political actors, it is not
far-fetched to assume that some effects
are very likely to exist. If, for example, a
user comment in the comment section of
a news website is written by a sock pup-
pet, and some individual reads that com-
ment without realizing that it was written
by a sock puppet, then the minimal effect
of digital astroturfing has happened –
someone was deceived by the digital as-
troturfing effort. The probability of such
a minimal effect of digital astroturfing is
very high; if the amount of people who get
deceived by digital astroturfing efforts is
greater than zero, then the minimal effect
Table 2: Tools, venues and actions of digital
astroturfing repertoires
Tools Venues Actions
sock puppets social media creating content
click farms websites signalling (dis-)
approval
sympathizers comment sections
paid supporters Direct messages
78 Kovic et al. / Studies in Communication Sciences 18.1 (2018), pp. 69–85
exists, and consequently, digital astroturf-
ing has an empirical impact. What is not
clear, however, is the full extent of the em-
pirical impacts of digital astroturfing. For
example, there is some evidence that us-
er-generated comments on news articles
can, in general, affect the perception of
the content of those news articles (Ander-
son, Brossard, Scheufele, Xenos, & Ladwig,
2014; Lee, 2012). It is entirely possible that
astroturfed comments can produce simi-
lar effects.
3.1 Restrictive countermeasures, or:
Limiting online activity
The most sweeping restrictive counter-
measure against digital astroturfing is be
to outlaw it. However, an outright crimi-
nalization of digital astroturfing is, at the
very least, hardly enforceable (digital as-
troturfing is, after all, clandestine). Also,
digital astroturfing activities that do not
use sock puppets but individuals who en-
gage in digital astroturfing activities with
their real identities, would present a dif-
ficult legal situation: While the activity is
clearly digital astroturfing (a political actor
manufactures deceptive activity in order
pursue his own goals), the individuals who
represent the digital astroturfing tools can
easily claim to honestly have the opinions
that they express through their activities.
Outlawing digital astroturfing thus seems
very unrealistic. Furthermore, such action
would potentially harm freedom of speech
and thus, from a consequentialist point of
view, do more harm than good (e. g., Deib-
ert et al., 2008).
More realistic means of restrictive
countermeasures are countermeasures
applied at the different digital astroturfing
venues as summarized in Table 2. But not
all of the venues are suited for restrictive
measures. Specifically, digital astroturfing
can neither be meaningfully restricted
when it comes to creating websites (it is
trivially easy to create a website anony-
mously) nor when it comes to sending di-
rect messages, such as emails, to political
actors. But for the other two venues, re-
strictive countermeasures can be imple-
mented.
The first of those venues is social me-
dia. One potential restrictive measure for
social media is the detection of inauthen-
tic user accounts. There already exists con-
siderable research in the area of automat-
ed detection of fake social media accounts,
usually in the context of spam accounts
that post links to products and services
or to other unrelated online content
(Mukherjee, Liu, & Glance, 2012; Zheng,
Zeng, Chen, Yu, & Rong, 2015). While there
is some research that extends the question
of automated detection of fake profiles
to digital astroturfing (Chen et al., 2011),
that research is still scarce. Another pos-
sible restrictive countermeasure for social
media platforms is the implementation of
a mandatory real-name policy for users.
Implementing a mandatory real-name
policy might sound drastic, but it has been
the official policy of one of the largest so-
cial media platforms, Facebook, for years
(Wauters, Donoso, & Lievens, 2014; Giles,
2011). However, even though using one’s
real name is the default policy at Facebook,
not every user has to or is able to provide
credentials for their identity. Facebook is
detecting potential pseudonyms algorith-
mically and, apparently, selectively, which
has led to instances of users accounts be-
ing blocked despite being honest expres-
sions of real identities (Dijck, 2013; Lingel
& Golub, 2015). Mandatory real-name
policies for social media also exist beyond
Western social media platforms, notably
in China. A prominent step towards ban-
ning anonymity on Chinese social media
has been taken in 2011, when a manda-
tory real-name policy for so-called mi-
croblogging sites that are headquartered
in Beijing, such as Sina Weibo, had been
introduced. Under that policy, users have
to register with their real identities, but are
allowed to use nicknames on their public
profiles. However, the policy has not yet
been thoroughly enforced (Jiang, 2016).
The second digital astroturfing venue
where restrictive countermeasures can be
implemented with a high probability of
the desired impact are comment sections
on news websites. By and large, comment
sections of news websites are already gov-
erned by some degree of restrictive mea-
Kovic et al. / Studies in Communication Sciences 18.1 (2018), pp. 69–85 79
sures in that most news websites have
some moderation policies in place (Can-
ter, 2013; Hille & Bakker, 2014). Comment
moderation policies can be made more
restrictive in order to prevent astroturfed
comments. One restriction measure is to
ban anonymous comments and enforce a
real-name policy, similar to the social me-
dia examples cited above. From a logical
point of view, banning anonymous com-
ments will always prevent sock puppets
from commenting, which means that this
measure is definitively effective, but only
if it is systematically enforced by means of
an explicit verification process.
Enforcing a real-name policy is not a
viable restrictive countermeasure against
all digital astroturfing tools listed in
Table 2: Paid supporters and sympathiz-
ers will still be able to be active on social
media and to comment on news websites,
since they do so with their real identity.
Another possible restrictive measure
is not aimed at the tools used for digital
astroturfing, but at the specific actions of
digital astroturfing: Disabling the possi-
bility of signaling approval or disapprov-
al for existing comments through liking
or disliking them. From a logical point of
view, disabling the possibility of liking or
disliking comments will always prevent
this form of astroturfing. A third and most
far-reaching restrictive measure with re-
gard to comments is to completely disable
comments. A number of prominent news
websites have opted for disabling their
comment section in recent years (Finley,
2015), but not primarily because of per-
ceived inauthentic commenters. Disabling
comments will, from a logical point of
view, always prevent all instances of astro-
turfed comments and, as a logical conse-
quence, it will always prevent all instances
of astroturfed “liking” or “disliking” of ex-
isting comments as well. This means that
completely disabling comments is the
most effective restrictive measure against
astroturfed comments, since it always pre-
vents 100% of astroturfing instances.
In summary, there are ways to imple-
ment restrictive countermeasures against
digital astroturfing which are very likely
to lead to the desired outcome, which is
reducing or completely preventing in-
stances of digital astroturfing. However,
implementing such restrictive measures
can also have some downsides, and we
believe that the benefit of reducing digital
astroturfing by restrictive measures does
not necessarily outweigh the downsides.
When it comes to the automatic detection
of unauthentic user profiles on social me-
dia platforms, the potential downside are
primarily false positives: When automatic
means of detecting and deleting inauthen-
tic user accounts are implemented, there
is a chance that a portion of the accounts
thus detected will actually be authentic
ones. However, this downside of false pos-
itives can be actively controlled and dialed
down. How many false positives there are
is to a substantial degree a trade off with
the desired sensitivity of the automatic de-
tection algorithms. If a maximum sensitiv-
ity is desired, meaning that a perfect rate
of true positives is the goal, then the false
positive rate will, probably, be at its high-
est. But if a lower sensitivity is accepted,
meaning that not all true positives will be
detected, then the false positive rate can
be lowered.
The downsides of restrictive counter-
measures in comments sections on news
websites are less open to adjustment than
the downsides of automatic detection of
inauthentic user accounts on social media.
The first such measure we discuss above
is enforcement of a real-name policy in
the comment sections. Perhaps it is not
self-evident why non-anonymity should
have downsides at all – if one has an hon-
est opinion to publicly share, why not do
so using their true identity? The problem
with that notion is that in many non-dem-
ocratic countries, voicing one’s opinion
publicly can have dire consequences. A
prominent example of such a country is
China, where online users criticize the
government under the guise of anonymity
because such criticism is essentially illegal
(Qiang, 2011). Limiting the possibilities of
anonymous online commenting in such
political contexts will either lead to less
expression of opinion, or to an increase in
sanctions because those who speak out are
easier to be tracked down. Another down-
80 Kovic et al. / Studies in Communication Sciences 18.1 (2018), pp. 69–85
side of doing away with anonymous com-
menting is commercial in nature. Remov-
ing anonymous commenting reduces the
total comment volume (Fredheim, Moore,
& Naughton, 2015), but only a small part of
those anonymous comments is likely to be
astroturfed. This means that websites will
lose a lot of user-generated content and,
consequently, website traffic. This can ul-
timately lead to loss in revenue, which is
a relevant factor for commercial operators
of social media and of news websites.
The strongest measure with regard
to comments on news websites that we
discuss above is the complete disabling
of comments. Completely removing the
possibility for users to comment is a mea-
sure, we believe, that represents a step
backwards in terms of the development of
the Internet, because it is a measure that
reduces the opportunities for honest citi-
zens to participate in public discourse. Of
course, the Internet is a big place, and any
one user can, for example, easily create a
website to express whatever opinions he
or she may hold. Furthermore, news orga-
nizations obviously have no obligation to
host user-generated content (Jönsson &
Örnebring, 2011) of any sort on their on-
line platforms. But the possibility for cit-
izens to directly react to news content by
commenting on it and by engaging with
other users who comment as well does
seem like an added discursive value which
is lost by disabling comments altogether.
Restrictive countermeasures against
digital astroturfing are possible, but they
carry with them some degree of down-
sides. Because of those downsides, it is
worth exploring alternative strategies of
countermeasures: Such countermeasures
that do not restrict, but instead incentiv-
ize.
3.2 Incentivizing countermeasures, or:
Rewarding honesty
The general idea behind incentivizing
countermeasures against digital astro-
turfing is to provide incentives that moti-
vate users to behave in ways that are more
likely to be honest. Incentivizing honest
behavior does not prevent digital astro-
turfing from occurring, but by rewarding
honest behavior, honest online activity
can become publicly discernible from on-
line activity for which there is no informa-
tion about its authenticity.
One goal of such incentives is to en-
courage user participation under one’s
true identity, without enforcing it. In the
case of social media, encouraging users
to use their true identity consists of two
steps. First, there needs to be an actual
possibility for users to verify their identity
on a given social media platform. Second,
after a user has verified his or her identity,
the information of the authenticity of the
social media profile in question needs to
be publicly visible. Some prominent social
media platforms, such as Twitter and Face-
book, have such a verification program in
place, and verified accounts are awarded
with a publicly visible badge. However,
both Twitter and Facebook do not current-
ly implement their verification scheme
widely for their whole user base, but only
to a relatively small portion of users who
are a “public figure, media company or
brand” (Facebook, 2015). We believe that
users can be incentivized to opt into a ver-
ification process by nudging them towards
it. Users can be presented with advantages
of a verified account, such as the pro-so-
cial norm of transparency and authentici-
ty, but also with the rather playful element
of receiving a badge that is proof of their
verified status. However, such nudging
strategies are, to a certain degree, paternal-
istic in nature (Thaler & Sunstein, 2009).
We think that the paternalistic element of
verification options should be minimized
by explicitly presenting verification as a
choice that brings with it certain benefits.
User verification is also possible for
the venue of comment sections. Besides
identity verification, another type of veri-
fication is possible for comment sections:
Incentives that allow for anonymous com-
menting, but reward a history of good and
consistent commenting behavior. One
such scheme is currently applied by the
New York Times, where some comment-
ers are algorithmically selected for a pub-
licly visible verification badge based on
their commenting history (Sullivan, 2014).
Commenters become eligible for verifica-
Kovic et al. / Studies in Communication Sciences 18.1 (2018), pp. 69–85 81
tion if they have not run afoul of comment
moderators in the past. The main incen-
tive under that scheme is that verified
commenters are able to immediately get
their comments published, without prior
moderation.
In comparison with restrictive
countermeasures, incentivizing ones have
the great advantage that they do not limit
the spectrum of online activities, but rath-
er expand it by offering rewards for behav-
ior that is more likely to be honest. Howev-
er, incentivizing countermeasures against
digital astroturfing are not without down-
sides. One major downside is the tempo-
ral component of incentivization: It takes
time, first, to implement incentives, and
second, for a critical part of a given user
base to react to incentives. For example,
if Twitter and Facebook decided to offer
verification to all of their users, that would
pose a non-trivial technical challenge, and
the widespread adoption of verification
would probably not be immediate but
rather follow adoption rates and patterns
similar those of other technological inno-
vations (Rogers, 2003).
4 Conclusion: Is research on digital
astroturfing feasible?
Laying out a conceptual map of digital
astroturfing is only a means to an end –
a map is meant to be used to explore. We
have described a typology with ideal-types
(Weber, 1922). In the real world, no two
instances of digital astroturfing will be ex-
actly the same, and they might employ any
mix of astroturfing repertoires. Still, the
map that we propose, even though prelim-
inary and likely incomplete, should be of
some use for maneuvering the terrain of
digital astroturfing. But how exactly can
digital astroturfing be explored empirical-
ly? After all, digital astroturfing is a clan-
destine activity, and if it is carried out suc-
cessfully, we do not know that it has taken
place. This makes any kind of research
inherently challenging. For example, re-
search designs that are routinely used in
the study of other forms of political com-
munication can be impossible to imple-
ment in the study of digital astroturfing,
since very basic facts such as who is doing
what in the pursuit of which strategic goals
is, by definition, absent in digital astroturf-
ing. Even though it undoubtedly poses
unique challenges, the empirical study of
digital astroturfing is not futile.
A first step in the study of digital as-
troturfing is the establishment of a plau-
sible conceptual framework of digital
astroturfing. The very goal of our study is
to contribute to this first step. In order to
conduct empirical research, we first need
a sound understanding of how to think
about our object of study. In this sense, we
do not think that research on digital astro-
turfing should be exploratory in nature,
as is sometimes suggested for research on
regular astroturfing (Boulay, 2013).
Some important recent research ef-
forts in the area of digital astroturfing are
focusing on so-called “social bots” (Bas-
tos and Mercea, 2017; Ferrara et al., 2016;
Woolley, 2016;). Social bots represent one
digital astroturfing tool (automated sock-
puppets) used within one venue (social
media). As such, they most certainly war-
rant scientific scrutiny. However, it is im-
portant to note that social bots probably
represent the low hanging fruit of digital
astroturfing. Social bots are a relatively
crude form of digital astroturfing, and be-
cause of their highly automated nature,
they are relatively easy to detect. Further-
more, and perhaps just as importantly, the
less sophisticated a digital astroturfing ef-
fort, the less probable it is that the effort
will have the impact intended by the initi-
ating actor behind the digital astroturfing
effort. Not only could there be no effect,
but there could actually be a negative ef-
fect (from the point of view of the initiat-
ing actor): When people become aware
of persuasion attempts and thus develop
persuasion knowledge, they tend not only
to not be persuaded, but to actually react
negatively (Friestad & Wright, 1994).
There are several empirical research
strategies that can be applied to the study
of digital astroturfing. A very important
one are case studies. To date, most re-
search on cases of digital astroturfing is
done by investigative journalists. Such
82 Kovic et al. / Studies in Communication Sciences 18.1 (2018), pp. 69–85
journalistic work is, of course, very valu-
able, but a more scientifically vigorous
analysis of digital astroturfing cases is still
necessary, not least because of the need
to ground case study research in a sound
conceptual framework. An obvious dis-
advantage of case studies is the fact that
the cases that can be analyzed cannot
be freely selected – only those instances
of digital astroturfing that have been re-
vealed to be digital astroturfing are avail-
able for case study analysis. This a priori
selection could potentially introduce bias,
because observed digital astroturfing is
perhaps different from unobserved digital
astroturfing. However, this potential bias
is not an insurmountable problem. Giv-
en a conceptual foundation that consists,
among other things, of a typology and a
reasonable expectation of digital astro-
turfing repertoires, the potential selection
bias in case study research can be actively
addressed. For example, if case study re-
search consistently failed to identify a cer-
tain type of digital astroturfing that could
indicate that that type of digital astroturf-
ing is very successful at remaining hidden.
Another possible research strategy
consists of surveys and interviews of polit-
ical actors. It might seem naive to suggest
to simply ask political actors whether they
engage in digital astroturfing. However, it
is not at all unheard of for political actors to
disclose sensitive information for scientif-
ic purposes as long as they are guaranteed
anonymity. But talking to political actors
about digital astroturfing does not have
to produce a direct “admission of guilt”
to be useful. For example, the perception
of digital astroturfing by political actors is
valuable data, because it could indicate
how prevalent a phenomenon digital as-
troturfing is. Talking to people who are
potentially involved in or knowledgeable
about digital astroturfing does not have to
be limited to political actors. Communica-
tion professionals in the business of politi-
cal consulting are also willing to talk about
their line of work, given a sufficient layer of
anonymity (e. g., Serazio, 2014).
A third promising research strategy
is the direct collaboration with venues of
digital astroturfing. Two of the venues that
are summarized in Table 2 are suitable for
this, social media services and news orga-
nizations that have comment sections on
their websites. Both social media services
and news organizations should, in princi-
ple, be interested in reducing, or at least in
understanding digital astroturfing on their
respective platforms. The specific nature
of the collaboration can be twofold. First,
the combined effort of researchers and
operators of social media or of news plat-
forms can be focused on the detection of
digital astroturfing. The second possible
kind of collaboration is the implementa-
tion and monitoring of countermeasures
against digital astroturfing. For this kind
of collaboration, greater weight should
be given to the implementation of incen-
tivizing countermeasures than to restric-
tive ones. Restrictive countermeasures
are mostly a technical affair with poten-
tially big downsides, while incentivizing
countermeasures involve not only tech-
nical innovation, but observation of and
interaction with users as well. Restrictive
countermeasures carry the promise of
relatively fast short-term solutions, but in
order to effectively combat digital astro-
turfing in the long term, we believe that
innovations in the form of incentivizing
countermeasures are inevitable.
References
Anderson, A. A., Brossard, D., Scheufele, D. A.,
Xenos, M. A., & Ladwig, P. (2014). The “Nas-
ty Effect”: Online incivility and risk per-
ceptions of emerging technologies. Journal
of Computer-Mediated Communication,
19(3), 373–387. doi:10.1111/jcc4.12 009.
Apfl, S., & Kleiner, S. (2014). Die Netzflüsterer
[The net whisperers]. DATUM, 2014(11).
Retrieved from https://datum.at/die-netz-
fluesterer/.
Bastos, M. T., & Mercea, D. (2017). The Brexit
botnet and user-generated hyperpartisan
news. Social Science Computer Review.
Advanced online publication, 1–18.
doi:10.1177/0894439317734157.
Bessi, A., & Ferrara, E. (2016). Social bots distort
the 2016 U.S. Presidential election online
discussion. First Monday, 21(11). Retrieved
Kovic et al. / Studies in Communication Sciences 18.1 (2018), pp. 69–85 83
from http://firstmonday.org/article/
view/7090/5653.
Biggs, M. (2013). How repertoires evolve: The
diffusion of suicide protest in the twenti-
eth century. Mobilization: An International
Quarterly, 18(4), 407–428. doi:10.17813/
maiq.18.4.njnu779530x55082.
Boulay, S. (2013). Can methodological require-
ments be fulfilled when studying con-
cealed or unethical research objects? The
case of astroturfing. ESSACHESS – Journal
for Communication Studies, 6(2/12),
177–187.
Boulay, S. (2012). Quelle(s) considération(s)
pour l’éthique dans l’usage des technolo-
gies d’information et de communication
en relations publiques? Analyse de cas
d’astroturfing et réflexion critique. Revista
Internacional de Relaciones Públicas, 2(4),
201–220.
Boyd, D. M., & Ellison, N. B. (2007). Social net-
work sites: Definition, history, and schol-
arship. Journal of Computer-Mediated
Communication, 13(1), 210–230.
doi:10.1111/j.1083-6101.2007.00393.x.
Bu, Z., Xia, Z., & Wang, J. (2013). A sock puppet
detection algorithm on virtual spaces.
Knowledge-Based Systems, 37 (January),
366–377. doi:10.1016/j.knosys.2012.08.016
Buchstein, H. (1997). Bytes that bite: The Inter-
net and deliberative democracy. Constel-
lations, 4(2), 248–263. doi:10.1111/1467-
8675.00052.
Canter, L. (2013). The misconception of online
comment threads. Journalism Practice,
7(5), 604–619. doi:10.1080/17512786.2012
.740172.
Chen, A. (June 2, 2015). The agency. The New
York Times. Retrieved from http://www.
nytimes.com/2015/06/07/magazine/
the-agency.html.
Chen, C., Wu, K., Srinivasan, V., & Zhang, X.
(2011). Battling the Internet water army:
Detection of hidden paid posters. arX-
iv:1111.4297 [cs]. arXiv: 1111.4297.
Clark, D. B. (April 21, 2015). The bot bubble:
How click farms have inflated social media
currency. The New Republic. Retrieved
from https://newrepublic.com/arti-
cle/121551/bot-bubble-click-farms-have-
inflated-social-media-currency.
Dahlberg, L. (2001). The Internet and demo-
cratic discourse: Exploring the prospects
of online deliberative forums extending
the public sphere. Information, Com-
munication & Society, 4(4), 615–633.
doi:10.1080/13691180110097030.
Deibert, R., Palfrey, J., Rohozinski, R., & Zittrain,
J. (Eds.) (2008). The information revolu-
tion and global politics. Access denied: The
practice and policy of global Internet filter-
ing. Cambridge, MA: MIT Press.
Della Porta, D. (2013). Repertoires of con-
tention. In D. A. Snow, D. Della Porta, B.
Klandermans, & D. McAdam (Eds.), The
Wiley-Blackwell encyclopedia of social
and political movements. Oxford, UK:
Blackwell. doi:10.1002/9780470674871.
wbespm178.
DePaulo, B. M., Kashy, D. A., Kirkendol, S. E.,
Wyer, M. M., & Epstein, J. A. (1996). Lying
in everyday life. Journal of Personality
and Social Psychology, 70(5), 979–995.
doi:10.1037/0022-3514.70.5.979.
Dijck, J. v. (2013). “You have one identity”:
Performing the self on Facebook and
LinkedIn. Media, Culture & Society, 35(2),
199–215. doi:10.1177/0163443712468605
Erat, S., & Gneezy, U. (2011). White lies.
Management Science, 58(4), 723–733.
doi:10.1287/mnsc.1110.1449.
Facebook (2015). What is a verified page or
profile? Retrieved from https://www.face-
book.com/help/196050490547892.
Ferrara, E. (2017). Disinformation and social
bot operations in the run up to the 2017
French presidential election. First Monday,
22(8). Retrieved from https://arxiv.org/ftp/
arxiv/papers/1707/1707.00086.pdf.
Fielding, N., & Cobain, I. (March 17, 2011). Re-
vealed: US spy operation that manipulates
social media. The Guardian. Retrieved
from http://www.theguardian.com/
technology/2011/mar/17/us-spy-opera-
tion-social-networks.
Finley, K. (October 8, 2015). A brief history
of the end of the comments. WIRED.
Retrieved from http://www.wired.
com/2015/10/brief-history-of-the-de-
mise-of-the-comments-timeline/.
Formisano, R. P. (2012). The Tea Party: A brief
history. Baltimore, MD: Johns Hopkins
University Press.
Fredheim, R., Moore, A., & Naughton, J. (2015).
Anonymity and online commenting: An
empirical study (SSRN Scholarly Paper No.
84 Kovic et al. / Studies in Communication Sciences 18.1 (2018), pp. 69–85
ID 2591299). Social Science Research Net-
work. Rochester, NY.
Geer, D. (2005). Malicious bots threaten net-
work security. Computer, 38(1), 18–20.
doi:10.1109/MC.2005.26.
Giles, J. (2011). Zeroing in on the real you. New
Scientist, 212(2836), 50–53. doi:10.1016/
S0262-4079(11)62669-9.
Grover, S. L. (1993). Lying, deceit, and subter-
fuge: A model of dishonesty in the work-
place. Organization Science, 4(3), 478–495.
doi:10.1287/orsc.4.3.478.
Hardaker, C. (2010). Trolling in asynchronous
computer-mediated communication:
From user discussions to academic defini-
tions. Journal of Politeness Research. Lan-
guage, Behaviour, Culture, 6(2), 215–242.
doi:10.1515/JPLR.2010.011.
Hart, S. W., & Klink, M. C. (2017). 1st Troll Battal-
ion: Influencing military and strategic op-
erations through cyber-personas. In 2017
International Conference on Cyber Conflict
(CyCon U.S.) (pp. 97–104). doi:10.1109/
CYCONUS.2017.8167503.
Hille, S. & Bakker, P. (2014). Engaging the social
news user. Journalism Practice, 8(5), 563–
572. doi:10.1080/17512786.2014.899758
Howard, P. N. (2006). New media campaigns
and the managed citizen. Communication,
society and politics. Cambridge: Cam-
bridge University Press.
Jiang, M. (2016). Managing the micro-self: the
governmentality of real name registration
policy in Chinese microblogosphere.
Information, Communication & Soci-
ety, 19(2), 202–220. doi:10.1080/136911
8X.2015.1060723.
Jönsson, A. M., & Örnebring, H. (2011). Us-
er-generated content and the news. Jour-
nalism Practice, 5(2), 127–144. doi:10.1080
/17512786.2010.501155.
Kaplan, A. M., & Haenlein, M. (2010). Users
of the world, unite! The challenges and
opportunities of social media. Business
Horizons, 53(1), 59–68. doi:10.1016/j.bush-
or.2009.09.003.
King, G., Pan, J., & Roberts, M. E. (2017). How
the Chinese government fabricates social
media posts for strategic Distraction, not
engaged argument. American Political Sci-
ence Review, 111(3), 484–501. doi:10.1017/
S0003055417000144.
Klotz, R. J. (2007). Internet campaigning for
grassroots and astroturf support. Social
Science Computer Review, 25(1), 3–12.
doi:10.1177/0894439306289105.
Lee, E.-J. (2012). That’s not the way it is: How
user-generated comments on the news
affect perceived media bias. Journal
of Computer-Mediated Communica-
tion, 18(1), 32–45. doi:10.1111/j.1083-
6101.2012.01597.x.
Lingel, J., & Golub, A. (2015). In face on Face-
book: Brooklyn’s drag community and
sociotechnical practices of online com-
munication. Journal of Computer-Me-
diated Communication, 20(5), 536–553.
doi:10.1111/jcc4.12125.
Llewellyn, C., Cram, L., Favero, A., & Hill, R. L.
(2018). For whom the bell trolls: Troll be-
haviour in the Twitter Brexit debate. ArX-
iv:1801.08754 [Cs]. Retrieved from http://
arxiv.org/abs/1801.08754.
Lyon, T. P., & Maxwell, J. W. (2004). Astroturf:
Interest group lobbying and corporate
strategy. Journal of Economics & Manage-
ment Strategy, 13(4), 561–597. doi:10.1111/
j.1430-9134.2004.00023.x.
MacFarquhar, N. (February 18, 2018). Inside
the Russian troll factory: Zombies and
a breakneck pace. The New York Times.
Retrieved from https://www.nytimes.
com/2018/02/18/world/europe/rus-
sia-troll-factory.html
Mackie, G. (2009). Astroturfing infotopia. Theo-
ria: A Journal of Social and Political Theo-
ry, 56(119), 30–56.
MacKinnon, R., & Zuckerman, E. (2012). Don’t
feed the trolls. Index on Censorship, 41(4),
14–24. doi:10.1177/0306422012467413.
McCornack, S. A. & Levine, T. R. (1990). When
lies are uncovered: Emotional and rela-
tional outcomes of discovered deception.
Communication Monographs, 57(2), 119–
138. doi:10.1080/03637759009376190.
National Intelligence Council. (2017). Assessing
Russian Activities and Intentions in Recent
US Elections (Intelligence Community
Assessment). Washington, DC. Retrieved
from https://www.dni.gov/files/docu-
ments/ICA_2017_01.pdf.
Neuhaus, C. (September 2, 2015). Hans-Peter
Portmanns Twitter-Wunder: 11 000 Freun-
de sollt ihr sein. Neue Zürcher Zeitung.
Kovic et al. / Studies in Communication Sciences 18.1 (2018), pp. 69–85 85
Qiang, X. (2011). The battle for the Chinese In-
ternet. Journal of Democracy, 22(2), 47–61.
doi:10.1353/jod.2011.0020.
Rheingold, H. (1993). The virtual community:
Homesteading on the electronic frontier.
Cambridge, MA: MIT Press.
Rogers, E. M. (2003). Diffusion of innovations.
New York: Free Press.
Serazio, M. (2014). The new media designs of
political consultants: Campaign produc-
tion in a fragmented era. Journal of Com-
munication, 64(4), 743–763. doi:10.1111/
jcom.1 2078.
Shao, C., Ciampaglia, G. L., Varol, O., Flammi-
ni, A., & Menczer, F. (2017). The spread
of misinformation by social bots. ArX-
iv:1707.07592 [Physics]. Retrieved from
http://arxiv.org/abs/1707.07592.
Stieglitz, S., Brachten, F., Berthelé, D., Schlaus,
M., Venetopoulou, C., & Veutgen, D. (2017).
Do social bots (still) act different to hu-
mans? – Comparing metrics of social bots
with those of humans. In G. Meiselwitz
(eds), Social computing and social media.
Human behavior (pp. 379–395). Cham,
Switzerland: Springer. doi:10.1007/978-3-
319-58559-8_30.
Suler, J. (2004). The online disinhibition effect.
CyberPsychology & Behavior, 7(3), 321–326.
doi:10.1089/1094931041291295.
Sullivan, M. (April 28, 2014). How “verified
commenters” earn their status at The
Times, and why. The New York Times.
Retrieved from http://publiceditor.blogs.
nytimes.com/2014/04/28/how-verified-
commenters-earn-their-status-at-the-
times-and-why/.
Tarrow, S. G. (2011). Power in movement: Social
movements and contentious politics. Cam-
bridge, MA: Cambridge University Press.
Thaler, R. H., & Sunstein, C. R. (2009). Nudge:
Improving decisions about health, wealth,
and happiness. New York: Penguin.
Twitter PublicPolicy. (2018). Update on Twitter’s
review of the 2016 U.S. election. Retrieved
from https://blog.twitter.com/official/
en_us/topics/company/2018/2016-elec-
tion-update.html.
Vergani, M. (2014). Rethinking grassroots
campaigners in the digital media: The
“grassroots orchestra” in Italy. Australian
Journal of Political Science, 49(2), 237–251.
doi:10.10 80/10361146.2014.898129
Walker, E. T. (2014). Grassroots for hire: Public
affairs consultants in american democracy.
Cambridge, MA: Cambridge University
Press.
Walker, E. T., & Rea, C. M. (2014). The political
mobilization of firms and industries. An-
nual Review of Sociology, 40(1), 281–304.
doi:10.1146/annurev-soc-071913-043215
Walker, J. L. (1983). The origins and mainte-
nance of interest groups in America. The
American Political Science Review, 77(2),
390–406. doi:10.2307/1958924.
Wauters, E., Donoso, V., & Lievens, E. (2014).
Optimizing transparency for users in
social networking sites. info, 16(6), 8–23.
doi:10.1108/info-06-2014-0026.
Weber, M. (1922). Wirtschaft und Gesellschaft.
Tübingen, Germany: Mohr.
Woolley, S. C. (2016). Automating power: Social
bot interference in global politics. First
Monday, 21(4).
Yoo, K.-H., & Gretzel, U. (2009). Comparison of
deceptive and truthful travel reviews. In
D. W. Höpken, D. U. Gretzel, & D. R. Law
(Eds.), Information and Communication
Technologies in Tourism 2009 (pp. 37–47).
Vienna: Springer. doi:10.1007/978-3-211-
93971-0_4
Zhang, J., Carpenter, D., & Ko, M. (2013). Online
astroturfing: A theoretical perspective.
AMCIS 2013 Proceedings.
Zheng, X., Zeng, Z., Chen, Z., Yu, Y., & Rong,
C. (2015). Detecting spammers on social
networks. Neurocomputing, 159, 27–34.
doi:10.1016/j.neucom.2015.02.047.
