Click in the button to see annotations by other readers and to add your own.

figure politics-social-media-raised-fist.png
Politics and social media

Daniel Gayo-Avello

Get it on PDF format
Table of Contents
List of Figures


Every one is son of his time; so philosophy also is its time apprehended in thoughts.
It is just as foolish to fancy that any philosophy can transcend its present world,
as that an individual could leap out of his time or jump over Rhodes.
G.W.F. Hegel (translated by S.W. Dyde)
What has been will be again,
what has been done will be done again;
there is nothing new under the sun.
Ecclesiastes 1:9


At first sight, politics and social media could appear to be a rather concrete and well-delimited topic. However, that’s only true if you are reading these words not long after their publication—let’s say in 5 to 8 years. If that’s the case, it is not unlikely that you have heard of Twitter or Facebook, and you may even had got an account (a ‘profile’, in the parlance of the age), while Google+, MySpace, Orkut and Second Life are probably long forgotten services. At the same time, blogs, wikis and email are probably still alive and kicking, no matter the fate that Blogger, Wordpress, Wikipedia and Gmail may have faced. You probably remember president Obama and you may have heard of his historical campaign of 2008, but you may have no clue about what Big Bird, horses, bayonets and binders full of women have to do with the US Presidential Elections of 2012. It is also not unlikely that ‘social media’ as a buzzword had faded into oblivion, and a different one had been coined to describe something ‘entirely new’ that is, below the surface, ‘more of the same’.
Does this mean that this book has a ‘shelf life’? Should you discharge it after, say, 2025? I hope not. And I hope that, not because of my ‘precognoscency’, but because of my aim to broaden that ‘social media and politics’ topic by looking both backwards and sideways. Looking backwards, because I’ll revisit the many different ways in which people have been interacting, engaging in politics, and organizing themselves long before so-called ‘social media’ and ‘Web 2.0’. And looking sideways, because there are much more in politics than just electoral campaigns and voting.
Thus, if you are using any form of Computer-Mediated Communication (CMC) to learn of other people’s ideas on governance, express yours, and maybe influence others to organize actions in order to move forward your point of view, then this book can be of your interest. You will find than your elders were not that different from you, and that except for computers (and maybe smart watches, or in your case ‘carbon sockets’ behind your ears—Gibson, 1984) you (and we before you) have been facing many of the problems that Solon and Pericles tackled with more than 2,500 years ago.
That said, this book is a product of its time and, therefore, you won’t find references beyond 2015; moreover, you will read plenty about cyberspace, forums, Usenet, email, blogs, Facebook, MySpace, Twitter, or YouTube.

What do I talk about when I talk about politics?

Although it can be implied from its contents I think it is much better to explicitly state it: this book is focused in political uses of social media, not in politics affecting social media [A]  [A] Politics affecting social media will be covered only to the extent that they affect political use of social media (see section 2 of chapter 6)., or politics within social media (cf. Margolis and Resnick, 2000: pp. 8-21).
More concretely, this book deals with politics as commonly understood in liberal democracies. Because of that, attention is paid to individual political participation [B]  [B] Chapter 1 Political participation.; the way in which different political actors such as parties, elected officials, legislatures or nonprofit organizations use social media [C]  [C] Chapter 2 Political actors.; the way in which social media can be used to ascertain different kinds of public opinion [D]  [D] Chapter 3 Public opinion.; or the role that social media plays during elections [E]  [E] Chapter 4 Elections..
There are two main reasons to focus on liberal democracy: First, virtually all of the technologies under the umbrella term ‘social media’ have been devised in democratic countries and, indeed, it has been argued that such tools can act as democratic catalyzers in authoritarian regimes [F]  [F] Chapter 5 Contentious politics.. Second, as Churchill said, “democracy is the worst form of Government except all those other forms that have been tried from time to time” or, as it has been rather convincingly argued, liberal democracy is the most satisfying governance system [G]  [G] “Life in a liberal democracy is potentially the road to great material abundance, but it also shows us the way to the completely nonmaterial end of recognition of our freedom. The liberal democratic state values us at our own sense of self-worth. Thus both the desiring and thymotic parts of our souls find satisfaction.” (Fukuyama 1992: p. 200) and, quite likely, the definite one (Fukuyama, 1992).
This by no means implies that liberal democracies are flawless: virtually all of them suffer serious problems and people worldwide have been asking for changes to improve the current state of their respective democracies. Inevitably, social media has also played a role in such struggles [H]  [H] Chapter 5 Contentious politics. and some voices claim that such media can help to revitalize democratic culture in the future, or even change it entirely.
On another hand, the point of view of Fukuyama on liberal democracy is not free of criticism, particularly by assuming that liberal democracies are inextricably linked to economic liberalism (Derrida, 1994: p. 72). Indeed, a fair amount of discussion has been devoted to the question of whether privately-owned social media can help democracy or if they, conversely, may be harmful [I]  [I] “Media socialia. Cui bono?” in chapter 3 Public opinion..
There are still additional caveats to Fukuyama’s thesis that liberal democracy will eventually flourish worldwide, and some of them are provided, rather surprisingly, by Fukuyama himself (2014), albeit an older and maybe wiser one. He identifies two worrisome issues: To start with, some authoritarian regimes did not simply disappear during the 1990s but, instead, they somehow transmuted by incorporating liberal economic elements without changing the political sustrate. Then, some democratic countries have been increasingly incorporating authoritarian elements [J]  [J] It must be noted that Fukuyama (2014) is still hopeful on his thesis as he says: “[D]espite the short-term ebb and flow of world politics, the power of the democratic ideal remains immense. [...] Even as we raise questions about how soon everyone will get there, we should have no doubt as to what kind of society lies at the end of History.”. Needless to say, social media are being used—albeit with more or less difficulties—in such countries and, thus, a review of their political implications under authoritarian regimes is de rigueur [K]  [K] Chapter 6 Social media and authoritarianism..
Finally, there are two threats to democracy that lie in opposite ends of the political spectrum.
At one end we have terrorism which, at the moment of this writing, has got as its most worrisome example jihadism. However, the truth is that the purported goals or credo of terrorist groups are of little interest because all of them share a great deal of common features, including their employment of social media. Besides, it does not matter how dreadful terror acts may seem, they are “a form of of political communication, intended to send a message to a particular constituency” (McNair, 2011); hence, this book would be incomplete if it did not include terrorism when discussing contentious politics and social media [L]  [L] “Social media and terrorism” in chapter 5 Contentious politics..
The other threat to democracy is pseudo-activity which Žižek (2008: p. 183) describes as:
“The urge to ‘be active’, to ‘participate’, to mask the nothingness of what goes on. [...] Those in power often prefer even a ‘critical’ participation, a dialogue, to silence - just to engage us in ‘dialogue’, to make sure our ominous passivity is broken.”
The problem of pseudo-activity regarding politics in social media is covered in this piece from a number of perspectives: under its most conventional form of ‘slacktivism’ or ‘clicktivism’ [M]  [M] “Beginning with ‘S’, political participation—Slacktivism?” in chapter 1 Political participation., but also when describing the nature of communicative capitalism and its impact when interpreting social media as a potential deliberative realm [N]  [N] “Social media and the public sphere” in chapter 3 Public opinion., or when reviewing the subtleties of censorship under networked authoritarianism [O]  [O] Chapter 6 Social media and authoritarianism..
So, in short, my coverage of politics encompasses the different aspects of democratic life; the behavior of the different actors operating under, for, and against democracy—including slacktivists, contenders, and terrorists; and the way in which democracy can eventually succeed over authoritarianism.

What do I talk about when I talk about social media?

It is probably no exaggeration to say that there are as many definitions of ‘social media’ as researchers and practitioners working in the field. However, as I have done with ‘politics’ I want to devote some time to explicitly state which kind of ‘social media’ I am discussing in this book. Hence, I will start with two definitions which are almost ten years apart: the first one is the first entry about ‘social media’ in Wikipedia—a social media itself:
“Social Media is the term used to describe media which are formed mainly by the public as a group, in a social way, rather than media produced by journalists, editors and media conglomerates.” (Benvie, 2006)
The second one is the definition provided by Hogan and Melville (2015):
“Social media represent a set of communication practices that can typically be described as ‘many-to-many.’ In contrast to broadcast media, consumers are typically also producers. In contrast to in-person communication, audiences are often ambiguous or underspecified.”
Both definitions share the concept of ‘user generated content’ but, apart from that, they are quite different and, in all probability, you are already making some addenda to both. Indeed, it seems that every definition of social media is somewhat incomplete without being accompanied by ostensive definitions.
For instance, the second version of Wikipedia’s definition for social media (also by Benvie) included as examples: “blogs, Podcasts, Vlogs, Wikis, MySpace, YouTube, Second Life, Digg, Memeorandum,, Reddit, Flickr, Tailrank and Newsvine” [P]  [P] A problem with ostensive definitions of social media, or for that matter of any technologically-based concept, is that they age badly. At the moment of this writing, Second Life, and MySpace are virtually dead; and, in all honesty, I had not heard of Memeorandum, Tailrank and Newsvine until preparing this piece. I wonder how well (or bad) will age the definition including Wikipedia, social network sites, Facebook, Twitter, Instagram and Reddit..
The more recent piece by Hogan and Melville mentions “collaborative encyclopedias such as Wikipedia, social network sites (SNSs) Facebook and Twitter, photo-sharing sites Instagram, and social news site Reddit.” They also point out that “technologies such as instant messengers, and to some even e-mail are considered social media.”
Thanks to those lists of examples a second feature of social media, one that is usually implicitly assumed, is made much more notorious: social media requires a communication network. In fact, without such a network (internet at the moment of this writing and phone lines before) we cannot properly talk about social media.
Still, there is a crucial point that has been barely hinted: the social part of social media. Going on with prior definitions it is not until the fifth revision of Wikipedia’s entry [Q]  [Q] That revision was contributed by user Mmanuel barely 8 hours after the creation of the entry. It is available at: that some clues on the meaning of ‘social’ are provided:
“people create and share with each other [opinions, insights, experiences and perspectives]”
On the other hand, Hogan and Melville argue that social media are characterized by ‘many-to-many’ communication (in contrast with one-to-many, i.e. broadcasting, or many-to-one, i.e. voting or petitions), and ‘non addressed’ or underspecified audiences (e.g. “all Facebook friends, Twitter followers, readers of a bulletin board, and so forth”).
So, in short, the social part of social media is determined by the particular relations that are established between its users when creating and consuming contents. Everyone can be a creator with their own audience and, at the same time, being part of the audiences of a variety of other users.
Clearly, many of the tools developed in the early XXI century during and after the so-called ‘Web 2.0 revolution’ fit such definitions of social media. Some of them, particularly ‘social networking sites’, are at the moment of this writing considered as the most prominent example of social media and, in some cursory reviews, the only one.
Nevertheless, different services that were available long before ‘social media’ also fit prior definitions. I am not talking about precursors or prior art, I am talking about systems and tools that were (and some of them still are) social media but were known by different names and, somehow, have been eclipsed by the fanciness of ‘social media’ as a buzzword.
For instance, Dahlberg (2015) suggests 1995 as the moment when Web-based digital media that were social media in all except for the name reached a peak, but he also mentions a number of applications predating the Web that deserve the consideration of social media: “bulletin board systems (e.g., Usenet and FidoNet), synchronous online chat (e.g., Internet Relay Chat), multi-user real-time virtual worlds, and, [...] [the] e-mail list”.
Indeed, he is not alone in the vindication of social media before ‘Social Media’; in this regard, the words by Nancy Baym (2015) are not only doing that, but providing a rationale for the coinage of a new term to describe as anew something that was already available:
“Any medium that allows people to make meaning together is social. There is nothing more ‘social’ about ‘social media’ than there is about postcards, landline telephones, television shows, newspapers, books, or cuneiform. There are distinctive qualities to what we call ‘social media’, but being social is not among them. Long before ‘social media,’ the Internet was used to do what Facebook’s mission statement promises. If the words ‘social’ and ‘media’ don’t describe anything distinctive, what cultural work does the term ‘social media’ do?
It obscures the unpleasant truth that ‘social media’ is the takeover of the social by the corporate. ‘Social media’ happened when companies figured out how to harness what people were already doing, make (some of) it a bit easier, call it ‘content,’ and funnel our practices into their revenue streams. The term ‘social media’ puts the focus on what people do through platforms rather than critical issues of ownership, rights, and power.”
It must be noted that this perspective is not particular of Baym but it is not commonplace [R]  [R] Jodi Dean is maybe one of the most prominent scholars working from such critical perspective (e.g. Dean, 2005). Her views have heavily influenced my approach to chapter 4.. However, I think that the fact that a fair amount of political actions are being conducted within privately-owned spaces should be a matter of concern for anyone with a real interest in politics. Actually, the fact that social media is a set of different ‘walled wardens’ can be the most definitory issue of the platforms used during the early XXI century.
But let’s go back to social media before ‘Social media’, which were those services? The truth is that the ecosystem was pretty varied: before the advent of the Web bulletin board systems (BBS), Usenet, Internet Relay Chat, and listservs were very popular (mainly during the late 1970s and 1980s). With the expansion of the Web (early 1990s) services mirroring most of the features of pre-Web platforms appeared, in addition to new tools such as wikis, blogs, or social bookmarking and tagging tools (mid-late 1990s). Eventually, social networking sites appeared (late 1990s and early 2000s) and ‘social media’ was coined [S]  [S] For an interesting discussion about who coined the term ‘social media’ see Bercovici (2010)..
It must be noted, however, that although older services have lost popularity they have not disappeared and, at the moment of this writing, most of them are available. Indeed, the current social media ecosystem can be seen as a number of layers accumulated along the last decades, each of them with its own age of glory and its idiosyncratic buzzword.
Some of such buzzwords are mostly interchangeable with ‘social media’ while others exhibit certain amount of overlap, or refer to some feature or subset of current social media. The most popular, in approximate chronological order, have been: Computer Conferencing, Computer-Mediated Communication, cyberspace, the Net, the Web, and Web 2.0 [T]  [T] For a first hand recollection of the transmutation of 1970s’ Computer Conferencing into 1990s’ Hypermedia please see Turoff (1989)..
So, in short, I use ‘social media’ to describe any communication tool that allows users to consume, share and create multimedia contents which are addressed to unspecified audiences, in a potentially many-to-many fashion. Although not a prerequisite, the most popular social media platforms during the early XXI are privately-owned, and virtually all of them incorporate to some extent social-networking features (such as user profiles and lists of ‘friends’).
Such a broad interpretation clearly covers social-networking sites (SNS) such as Facebook or Twitter, but also the blogosphere, listservs, chatrooms, IRC, or Usenet to name just a few. Hence, although I will mainly cover recent research performed on SNS I will also rely on literature dealing with the intersection of politics and this much broader interpretation of social media.

Why the intersection of politics and social media is crucial?

The whole point of this book is that interaction between social media and politics is not simply important, but crucial. However, to support such a strong claim I need to digress a little.
As I have suggested in the prologue, a cursory approach to the topic could imply a mere interest in social networking sites and elections. If that was the case, a literature review could trace maybe a decade worth of material that would be mainly focused in the United States, and would probably have a heavy emphasis in president Obama’s campaigns. As such, it could be considered interesting, even fashionable, but not universally appealing or generalizable beyond the realms of Facebook or Twitter, or the idiosyncrasies of the USA.
However, such an approach would have been shortsighted, and by focusing in case studies we would risk missing the big picture. A broader interpretation of both politics and social media allows us to go beyond single case studies to, firstly, find deeper long-term implications of the use of computer-mediated communications for political action, and, secondly, avoid repeating prior mistakes (and ubris).
At the moment of this writing, it seems that some social media researchers and practitioners are somewhat intoxicated by what I call ‘social media exceptionalism’. That is, social media—normally reduced to SNS—is considered something entirely new which is to revolutionize every aspect of our lives, including of course politics.
Such exceptionalism is normally received with skepticism and, hence, a dialectical fight eventually arise [U]  [U] Academic debates that tend to be pseudo-activity in the already mentioned sense of Žižek.. A broader interpretation of social media helps to better contextualize this kind of claims and debates, and tone down the social media exceptionalism discourse. Actually, the debate about the purportedly beneficial impact of computer-mediated communications (i.e. social media) in politics, and its corresponding replica can be traced back at least to the mid-1980s.
For instance, Winner (1986: pp. 98-99, 105) said about cyber-optimists:
“In countless books, magazine articles, and media specials some intrepid soul steps forth to proclaim ‘the revolution.’ Often it is called simply ‘the computer revolution’ [...] Other popular variants include the ‘information revolution,’ ‘microelectronics revolution,’ and ‘network revolution.’

Computer scientist J. C. R. Licklider of the Massachusetts Institute of Technology is one advocate especially hopeful about a revitalization of the democratic process. [...] ‘The information revolution,’ he exclaims, ‘is bringing with it a key that may open the door to a new era of involvement and participation. The key is the self-motivating exhilaration that accompanies truly effective interaction with information through a good console through a good network to a good computer.’ It is, in short, a democracy of machines.
Taken as a whole, beliefs of this kind constitute what I would call mythinformation: the almost religious conviction that a widespread adoption of computers and communications systems along with easy access to electronic information will automatically produce a better world for human living.”
Other scholars made similar criticism while focusing in particular instances of social media. For instance, while discussing Usenet newsgroups and email listservs, Streck (1998) said that as any other human product they are subject to human limitations and, hence, in spite of the promises of cyber-optimists, cyberspace would not produce equality or diversity but “a world much like our present one: a heterogeneous place of community, isolation, prejudice, love, hate, intelligence, stupidity, culture, commerce, respect, disregard and just about everything else that makes life at once worth living and almost more than we can stand.”
I guess that this sort of criticism sounds familiar to you but applied to social networking sites; hence, you will neither find novelty in the arguments supporting their purported democratizing power. For instance, Rheingold (1993: p. 133) said about bulletin board systems:
“If a BBS (computer Bulletin Board System) isn’t a democratizing technology, there is no such thing. For less than the cost of a shotgun, a BBS turns an ordinary person anywhere in the world into a publisher, an eyewitness reporter, an advocate, an organizer, a student or teacher, and potential participant in a worldwide citizen-to-citizen conversation.”
And he said the following about being informed in real time about ongoing events (Rheingold 1993: p. 283):
“Information and disinformation about breaking events are pretty raw on the Net. That’s the point. [...] With the Net, during times of crisis, you can get more information, of extremely varying quality, than you can get from conventional media. [...] None of the evidence for political uses of the Net thus far presented is earthshaking in terms of how much power it has now to influence events. But the somewhat different roles of the Net in Tiananmen Square, the Soviet coup, and the Gulf War, represent harbingers of political upheavals to come.”
You could simply replace ‘BBS’ or ‘the Net’ by ‘Twitter’, ‘Facebook’, ‘YouTube’ or virtually any other social media service and you would not notice that the argument is not contemporary but more than 20 years old. Replace ‘Tiananmen Square’, ‘Soviet coup’, ‘Gulf War’ with ‘Arab Spring’, ‘Euromaidan’, and ‘War on Terror’, respectively, and you are not reading about old-fashioned cyberspace but about the huge potential of Twitter as a news source.
Even the threats to the presumed power of social media have been warned a number of times. For instance, Strangelove (1994) said about email, censorship and control:
“E-mail, as a metaphor for networked, global, uncensored communication, is already under attack by the state (through the Clipper chip legislation--an attempt to provide wiretap capability for all electronic communication). E-mail, Internet-based communication, is clearly potentially subversive as it allows bi-directional, unfiltered, uncensored mass communication.”
Rheingold (1993: pp. 298-299) also mentions the chances of exploiting computer-mediated communication “as means of surveillance, control, and disinformation” and provides an additional sort of criticism that perfectly applies to contemporary social media: its commoditization. Indeed he made some poignant but accurate predictions about our contemporary ‘free-model’ in social media (Rheingold, 1993: pp. 312-313):
“You won’t need a dictatorship from above to spy on your neighbors and have them spy on you. Instead, you’ll sell pieces of each other’s individuality to one another. [...] The most insidious attack on our rights to a reasonable degree of privacy might come not from a political dictatorship but from the marketplace. [...] These professional privacy brokers have begun to realize that a significant portion of the population would freely allow someone else to collect and use and even sell personal information, in return for payment or subsidies [V]  [V] It is quite of dramatic that in exchange for all of the private information (including our opinions and thoughts) that we are currently giving to SNS we just obtain the right to keep on using those same SNS..”
You may be wondering what is the point of this detour to the late XX century. Simple: I want to stress that contemporary social media and politics is no more about Twitter and the @POTUS account [W]  [W] Obama (2015). than 1990s’ cyberspace and politics was about president Clinton and vice president Gore announcing they would be reading email [X]  [X] See Rheinghold (1993: pp. 86-87).. If we keep looking at the proverbial finger (i.e. the fashionable case studies about Twitter, Facebook, and so on) we will miss the Moon: the ongoing work that has been conducted during the last three decades to understand how to exploit a powerful set of tools to involve citizens worldwide in actual democratic political action, and to avoid the two dire threats menacing such high hopes: authoritarian control and market commoditization. Those two simple ideas are the justification for this book.

Which is my position and which are my goals with this book?

Along this piece I will provide (at least that is my aim) a balanced review of the literature, but that does not imply that I will act as a dispassionate observer. I have a strong point of view regarding the interaction of social media (whatever its name) and politics, and my position cannot be detached from my analysis of what has been done in the field and what will be likely done in the future. I think that my position will be pretty obvious while reading the book but, as I have done with both politics and social media, I think you deserve to have it explicitly stated.
I am mildly hopeful about the potential of social media to help non conventional or contentious political participation; however, social media is not a panacea for social movements and it can make them grow anarchically and in unproductive ways [Y]  [Y] Chapter 5 Contentious politics.. However, I am hopeful that new and better platforms will appear in the future particularly tailored for this kind of unconventional participation.
I am highly skeptical about the potential of social media to improve the quality of conventional politics under democratic regimes; at least as long as social media equates to privately-owned walled gardens whose aim is to extract maximum profit from personal intercommunication [Z]  [Z] “Media socialia. Cui bono?” in chapter 3 Public opinion., and most users engage in slacktivism [A]  [A] “Beginning with ‘S’, political participation—Slacktivism?” in chapter 1 Political participation.. As with unconventional politics I hope that new platforms will improve these prospects.
I am a non believer on the feasibility of social media to catalyze democratic changes and make authoritarian regimes stumble. I think that a large part of the discourse on the so-called ‘social media revolutions’ has got a white-savior flavor that is profoundly distasteful and condescending with local populations [B]  [B] “The Iranian protests of 2009 and the Arab Spring” in chapter 5 Contentious politics.. Moreover, networked authoritarianism is the best counterexample to the presumed benefits of social media to inspire democracy [C]  [C] “Social media under networked authoritarianism” in chapter 6 Social media and authoritarianism..
I am extremely worried about the growing movements to exploit social media data in the War on Terror. Because of its own nature, lone wolf attacks are virtually impossible to avoid [D]  [D] “Social media and terrorism” in chapter 5 Contentious politics., and trying to automatically do so will devoid of meaning most of the liberties that define liberal democracies. Indeed, as long as social media are privately-owned platforms legally bound by potentially unjust laws the chances of criminalization of innocent citizens will increase, and the prospects of opposition groups under authoritarian regimes will weaken.
In short, although the interaction between social media and politics is usually described as a new ‘agora’ such an analogy is misleading. On one hand the classic agora is reduced to a marketplace which was not its only role in the classic world. On another hand, while current social media are commercial in nature—like a marketplace—they are not truly public. Actually, a better metaphor for social media and politics would be shouting and throwing pamphlets in a mall while trying to not disturb the clientele and avoiding the security guards.
Thus, my overall position is liberal regarding social and individual issues, but not so much regarding commercial ones. I somewhat distrust privately owned platforms not because they are malign, but because their interests are not aligned with those of citizens.
From my point of view, to be a valid political realm social media requires of free, open sourced, distributed, and decentralized tools which rely on strong encryption and complete anonymity (see the concluding remarks of the book).
Certainly, such an approach could render social media platforms as virtually useless for traditional actors such as political parties, lobbies, and politicians. However, that would not be a bug, but a feature: a problem of current social media is that it has been colonized by elites and transformed a presumed many-to-many conversation in a new form of broadcasting [E]  [E] “Mainstream political parties” and “Citizens. This is government” (both in chapter 2 Political actors), and “Social media in electoral campaigns” (in chapter 4 Elections)..
Moreover, gauging representative and trustful public opinion would be increasingly difficult if not impossible at all. Yet, it is a cheap price if it means that authoritarian regimes cannot easily monitor dissent groups.
It is also true that terrorist organizations would take advantage of anonymity and encryption but, after all, they are already exploiting the current tools to their advantage, and, in this regard, I think that the words by Benjamin Franklin perfectly apply: “Those who would give up essential Liberty, to purchase a little temporary Safety, deserve neither Liberty nor Safety.”
To end with, which are my goals when writing this book? What am I trying to convey?
I will be clear, I am trying to persuade you that the complacency of the Panglossian scenario (Barber 1998) in social media is dangerous: social media on itself will not make us freer—particularly if controlled by privately-owned corporations [F]  [F] For a thorough discussion on the perils of social media controlled by a few private companies or driven by market interests see, for instance, McChesney (1996), Dean (2005), Cammaerts (2008), Zittrain (2008), or Thornton (2011), . Indeed, such a believe may pave the way for what Barber labeled as ‘Pandora scenario’: the application of technology for control and repression.
In order to persuade you I will try to make you think about social media, not only in its current state but in its past and, more importantly, in its future state. I will try to make you think about politics and democracy beyond the fanfare of electoral campaigns and voting. I will try to help you develop an informed opinion on the current state of social media and politics, and make you consider if such an state of affairs is the best possible one or if there is room for improvement.
Hopefully, I will be able to convince you that a Jeffersonian scenario [G]  [G] Barber (1998) described the Jeffersonian scenario as a situation where “technologies of information and communications can be nurturing to democracy, … challenge passivity, … enhance information equality, ... overcome sectarianism and prejudice, and they can facilitate participation in deliberative political processes.”, no matter how improbable, is feasible, the most desirable outcome, and something all of us should fight for using our means at hand.

1 Political participation

“You might not be interested in politics, but politics is interested in you.”
Robert D. Holsworth

1.1 Introduction

Verba and Nie (1972: p. 1) said the following about democracy and participation:
“If democracy is interpreted as rule by the people, then the question of who participates in political decisions becomes the question of the nature of democracy in society. Where few take part in decisions there is little democracy; the more participation there is in decisions, the more democracy there is. Such a definition of democracy is crude [...] yet it may get at the heart of the matter, since all other institutions associated with democracy can be related to the general question of who participates or is able to participate in political life.”
Political participation is the core of democracy and, as such, it fulfills a number of crucial functions: it allows the citizenry to communicate with their representatives; it affects, in turn, the behaviors of those in response to citizens’ demands; and, finally, it is a source of civic satisfaction “with the government, and [...] with one’s own role” (Verba and Nie, 1972: p. 5).
Little political participation erodes the legitimacy of the actions of the government because a large part of citizens were not involved in the decisions leading to them; it makes the government actions detached of the people’s demands—or, at least, detached from the demands of some parts of society; and this, in turn, negatively affects the satisfaction of people [H]  [H] This detachment of citizenry from their representatives has been argued as one of the main reasons for the Indignados and other similar movements; see chapter 5 for a review of the application of social media to articulate their discontent..
Indeed, we have been witnessing this ‘democratic disenchantment’ and a constant decline in political participation for a number of years—some such as Putnam (2000) say that for a number of decades—and it is one of the major paradoxes in modern democracy: while democratic ideals are generally praised, the majority of democratic regimes are subject to constant scrutiny and strong criticism by their own citizens, who show increasing levels of distrust and detachment from political institutions (Rosanvallon 2008). Thus, it is hardly surprising that political participation—or, much better, its lack of—is subject of major research by political scientists.
It must be noted however, that although no political scientist denies this worrisome situation, some of them (e.g., Rosanvallon, 2008: p. 18, or Hay, 2013: p. 23) argue that indicators such as increasing abstention are not necessarily a signal of apathy, but maybe of a change in citizens’ behavior towards more unconventional [I]  [I] The use of social media for unconventional political participation is discussed in chapter 5. forms of participation.
In this regard, Rosanvallon says that “voting is the most visible and institutionalized expression of citizenry, the symbol of political participation and civic equality” but it is not the only mode of political participation which is a much more complex issue. In a similar vein Hay (2013) suggests that the only difference between those arguing about the lack of participation, and those arguing a change in modes of participation is their conception of what ‘political participation’ is. Therefore, what is political participation?
Given that I started quoting Verba and Nie, it is de rigeur to include their definition: “those activities by private citizens that are more or less directly aimed at influencing the selection of governmental personnel and/or the actions they take” (Verba and Nie, 1972: p. 2).
Of course, theirs is not the only definition, and other scholars—including Verba and his colleagues—have proposed successive refinements to encompass an increasingly wide spectrum of political actions. Among such definitions, those by Milbrath (1981) and Verba et al. (1995) have been commonly employed in social media research.
Milbrath defined political participation as “those actions of private citizens by which they seek to influence or to support government and politics”, while Verba et al. described voluntary political participation as any “activity that has the intent or effect of influencing government action—either directly by affecting the making or implementation of public policy or indirectly by influencing the selection of people who make those policies”.
This issue, however, is far from settled and, indeed, depending on the considerations about what constitutes or not a mode of political participation very different readings can been made from the same situation. In this regard, the interested reader should consult the recent work of van Deth (2014) who provides an operational definition that can be straightforwardly applied to determine if a behavior (offline or online) is or not political participation.
Nevertheless, for the purpose of this chapter it is enough to know that, according to van Deth, political participation is (1) an action; (2) conducted by particular citizens; (3) voluntarily [J]  [J] The requirement for the actions to be voluntary does not only exclude forced actions but those made by paid staff.; and (4) dealing with politics, the government or the state in a broad sense of those terms.
A non exhaustive list of online and offline actions that fit one or more of prior definitions, and that are mentioned in one or more of the papers covered in this chapter, are provided in Table 1 as an illustration. Please note that some of those actions (e.g., ‘visiting political websites’, usually considered the online equivalent to the traditional action of ‘being informed about politics’) were explicitly excluded by van Deth who did not consider them as activities. Yet, he acknowledged that other authors label such activities as ‘latent forms of political participation’. It is important to emphasize that this is not a simple terminological subtlety: I will discuss later the impact that latent aspects of social media engagement can have in offline political actions such as voting.
Table 1.1 Specimens of typical modes of political participation in offline and online settings.
Offline political participation Online political participation
being informed about politics
contacting elected officials (e.g., in person, by phone or by letter)
engaging in political discussion
writing letters to newspapers
sending support or protest messages to political leaders
wearing/displaying badges, stickers or T-shirts with political messages
attending meetings or rallies
persuading others to vote
joining and supporting a party
campaigning for a party or candidate
making financial contributions to a party or candidate or participating in fund-raising
running for office
working with others on local problems
distributing flyers with political messages
giving speeches
signing a petition
deliberately buying (or boycotting) certain products for political reasons
refusing to obey unjust laws
participating in a legal demonstration or strike
writing political messages on walls
participating in an illegal demonstrationrioting
looking for political information on the web (including videos) visiting gubernamental or public administration websites
visiting websites with political contents or run by political organizations
signing up for a newsletter, subscribing to an RSS feed
signing up as a friend or following a political organization, an elected official or a candidate on a SNS
downloading materials related to a political organization such as screensavers or wallpapers
customizing a webpage or social media profile to display new political or campaign information
contacting a political official or candidate (e.g., via e-mail)
participating in an online Q&A session with a political official
political dialogue/discussion (e.g., chat rooms, email lists, SNS, etc.)
reacting online to a message or article on the Internet (e.g., adding a comment, liking or favoriting it).
posting a message in a blog or microblog expressing a political opinion
creating or uploading videos with a political message (e.g., to YouTube)
starting or joining a political group on a SNS
signing up to volunteer for a campaign
donating online
talking with others about the campaign
trying to influence others (e.g., via e-mail, e-postcards or sharing/forwarding political content by third-parties)
joining a political organization online
working with others in a virtual community to deal with local problems
signing a petition online
participating in online polls
organizing an internet-based boycott or protest
Table 1.
Once clarified what political participation is and its importance, it is unsurprising that increasing participation should be actively pursued in democracy, and that a good amount of research has been conducted on the role that social media [K]  [K] In this chapter and for the rest of the book I am using ‘social media’ in the broad sense introduced in the Foreword, and not as a synonym for social networking sites. can play to achieve that goal.
In this regard, two broad lines of research have been explored: (1) Whether social media decreases, increases, or does not affect political participation at large; and (2) whether social media is able to mobilize previously unengaged (e.g., the youth) or disenfranchised people (e.g., the poor and the less-educated).
Depending on the findings—which, in all honesty, are varied to the point of contradiction, the interaction between social media and politics regarding participation can be depicted in a number of ways [L]  [L] I am using the labels in a rather free manner, the reader may find the same denominations in the literature with slightly different interpretations.:
  1. The demobilization scenario [M]  [M] Please note that I am using here ‘demobilization scenario’ as the opposite end of the ‘mobilization scenario’; it has nothing to do with the demobilization hypothesis that states that recurring to negative campaigning tactics and advertising demobilize the electorate (Bishop & Hillygus, 2013: p. 209). which posits that social media has a demobilizing effect on people and, thus, it decreases political participation.
  2. The mobilization scenario, whose advocates claim that social media is able to move people to take part in politics, especially those who are uninterested in traditional participation, or the disenfranchised.
  3. The normalization scenario, which claims that social media just appeals to those who are able to participate in politics in traditional ways. Thus, social media is seen as a complement for offline modes of political participation.
  4. The reinforcement scenario, which can be seen as a perverse hybrid of mobilization and normalization. Its proponents argue that politically engaged citizens can get additional advantage over unengaged and disenfranchised people because of using social media.
  5. The commodification scenario, which can be considered as a degeneration of normalization; here social media is used as just another mass media and, thus, people’s political participation is basically reduced to information consumption.
  6. The null scenario, which states that social media does not increase political participation but just the opposite: politically engaged people tend to use social media to keep on with their political activities. It must be noted, that most of the time the difference between proponents of normalization and null scenarios does not lay on the results but on their interpretation.
Moreover, it is not unlikely that features from most of these scenarios are taking place simultaneously, but affecting different groups of people and with varied intensity depending on many factors, including idiosyncrasies of each electoral or inter-electoral period. Moreover, we cannot rule out the possibility of a mutual interaction between online and offline actions; that is, citizens participating in traditional ways going online and vice versa, users engaging in politics on social media and then moving offline.
In the following sections I will review a selection of relevant literature regarding each of these scenarios. I close the chapter with the broad picture depicted by the interaction between those contending views, in addition to some proposals for further research in the field.

1.2 The demobilization scenario

I include this one just for the sake of completeness because at the moment of this writing there is abundant evidence disproving its most general form, and many of the arguments that have been used to support it lay on heavy-biased readings of the literature.
Its underlying idea is simple—and includes a touch of luddism: Internet alienates people, increasingly isolates persons from their social environment and, hence, it demobilizes them. A less dramatic version lies on the concept of ‘time displacement’: users devoting time to Internet (or social media) are dedicating less time to other activities, such as political participation.
One of the authors most commonly cited in this regard is Putnam; whose work and thinking has been somewhat reduced to ‘television killed civic America’. Certainly, Putnam (1995) said: “Many possible answers have been suggested for this puzzle [the erosion of social capital]: [...] Television, the electronic revolution, and other technological changes.” However, while many have interpreted the ‘electronic revolution’ as a reference to Internet-mediated communications [N]  [N] In contrast to recent literature that is very explicit regarding the platforms under study (e.g., Twitter, Facebook, etc), literature predating the Web 2.0 explosion is more vague and tend to discuss ‘the Web’, ‘the Internet’, ‘cyberspace’, and so forth. For the purpose of this book I have focused on research involving any kind of interaction among users; however, I will keep the denomination originally used by the respective authors. the truth is that Putnam did not simply mention it.
Ideas in that work were later expanded (see Putnam, 2000) and, again, a number of fragments were available for misquotation, and to subsequently blame the computer-mediated communication for the declining in civic and political participation; for instance:
“[U]nlike those who rely on newspapers, radio, and television for news, those few technologically proficient Americans who rely primarily on the Internet for news are actually less likely than their fellow citizens to be civically involved.”
or this other one:
“The absence of any correlation between Internet usage and civic engagement could mean that the Internet attracts reclusive nerds and energizes them, but it could also mean that the Net disproportionately attracts civic dynamos and sedates them.”
However, Putnam did not blame the Internet; although he neither depicted it as the last hope for democracy:
“The timing of the Internet explosion means that it cannot possibly be causally linked to the crumbling of social connectedness described in previous chapters. [...] By the time that the Internet reached 10 percent of American adults in 1996, the nationwide decline in social connectedness and civic engagement had been under way for at least a quarter of a century. Whatever the future implications of the Internet, social intercourse over the last several decades of the twentieth century was not simply displaced from physical space to cyberspace. The Internet may be part of the solution to our civic problem, or it may exacerbate it, but the cyberrevolution was not the cause.”
Other frequently cited work is those by Kraut et al. (1998) who found that although “the Internet was used extensively for communication [...] greater use of the Internet was associated with declines in participants’ communication with family members in the household, declines in the size of their social circle, and increases in their depression and loneliness”. Interestingly, Kraut et al. (2001) [O]  [O] It is also interesting that the study that basically refutes such glooming effects of the Internet receives just 40% of the cites of the original report. found that such effects had later ‘dissipated’ in participants from their original study, and findings in a second study were exactly the opposite: “more use of the Internet was associated with positive outcomes over a broad range of dependent variables measuring social involvement and psychological well-being”.
Another work—this one reporting mixed results from a social media perspective—was conducted by Shah et al. (2001). To start with they criticized reports that depicted Internet use as overall negative (e.g., Kraut et al., 1998), and provided a much more nuanced picture where information exchange had a positive impact in social capital, and social recreation had a negative one. I consider their results mixed because while they found email exchanging (a kind of social media) as positive they also found that MUDs and chat rooms (which are also social media) had negative effects.
In spite of this, more recent studies have mostly discredit the demobilization scenario, at least, as a universal consequence of Internet or social media use. For instance, Jennings and Zeitner (2003) claim that the “pessimistic view that Internet use would somehow lead to a decline in civic engagement is clearly not warranted”, neither the study by DiMaggio et al. (2004) supported the argument of Internet leading to passivity, and Boulianne (2009) found “little evidence to support the argument that Internet use is contributing to civic decline.”

1.3 The mobilization scenario

This scenario is at the antipodes of the previous one. Its advocates claim that Internet (and social media) have a positive impact in political participation affecting every group in society, but especially those that are unengaged or disenfranchised. It must be noted that there seems to exist a real chance of mobilization only affecting those citizens that are already taking part in politics; such a ‘reinforcement scenario’ is not covered here but in another section.
Regarding the mobilization potential of Internet and social media of the population at large, there are a number of promising reports.
Tolbert and McNeal (2003) found that those looking for political information online during the US 1996 Presidential Elections were more likely to vote, even after controlling for a number of factors such as SES, partisanship, and other media consumption; the same effect was found in the elections held in 2000 although not for the 1998 midterm elections. Besides, other political actions such as discussing politics with others, giving money or volunteering for a candidate were also more likely for those being informed online. Similar results were reported by Kenski and Stroud (2006) but they noted that the impact was not large.
Nisbet and Scheufele (2004) conducted a cross-sectional analysis of data regarding the 2000 US Presidential Elections and the impact that Internet could exert on political participation and knowledge. They found a significant, although modest, impact that was increased when the individuals had also discussed politics with friends and family; it must be noted that they agreed that, indeed, it could also be that Internet use could increase offline political discussion and, in turn, political participation. Interestingly, a later work by Hardy and Scheufele (2006) found that online chat rooms had the same mediating impact of face-to-face political discussion to increase political participation.
Stanley and Weare (2004) reported an interesting case study where a Federal Agency created a web-based discussion forum to ask for public input on a strategic campaign; the forum was run in parallel to a docket and it received much more comments, from a more varied range of constituencies, and covering a broader set of topics. It is a quite limited experience but it clearly shows that, at least for concrete issues, social media systems are able to mobilize the citizenry.
Nevertheless, when discussing the mobilizing possibilities of social media it is far more common to suggest that it can help to engage citizens that are uninvolved or uninterested in politics (e.g., the youth or the women), or simply disenfranchised. Purportedly, the low barriers to entry, the relative unimportance of physical location, and even the possibilities for anonymity or pseudonymity would make especially easier for those people to participate.
As with the demobilization scenario, most of the renderings of this optimistic position have got a few favorite references to drop. One of them is Rheingold (1993), the other one is Varley (1991); a typical quotation from the former could be the following:
“The political significance of CMC lies in its capacity to challenge the existing political hierarchy’s monopoly on powerful communications media, and perhaps thus revitalize citizen-based democracy.” (Rheingold, 1993: p.14)
As for Varley, she described a computer-mediated communication system established in Santa Monica, California during the late 1980s to engage citizens in discussions with other fellow citizens and public officials, about different topics of interest for the community. A common quotation reads as follows:
“For instance, through PEN [Public Electronic Network], a group of residents—including three or four homeless—formed an on-line political organization that lobbied successfully for new city services for the homeless.” (Varley, 1991: p. 44)
Without additional context it seems that both authors are rather optimistic and, besides, Varley is describing some exciting results regarding the participation of the disenfranchised. Unfortunately, both of them are usually misquoted and, thus, unfairly criticized—particularly, Rheingold.
Actually, Varley provides plenty of evidence that the system was far from perfect. For instance, she quotes PEN users describing ‘flame wars’, harassment to women, debates dominated by a minority of users, or the lack of meaningful participation by politicians [P]  [P] Such unglamorous description is perfectly appropriate for Usenet, listservs or Twitter; indeed, all of those problems affect any social media service when used for political participation and debate.. Taking into account all of this—particularly the abuse against women—it seems far stretched to exhibit PEN as support for the positive impact of social media for political mobilization.

1.3.1 Mobilizing the youth

Anyway, the truth is that there exists plenty of literature arguing about the possibilities of mobilization of unengaged or disenfranchised citizens thanks to social media. However, few hard facts support the general feasibility of this idea; at most, there are cautious remarks about its impact to mobilize young adults, and a number of seemingly promising findings.
For instance, Norris (1999) carefully states that the normalization scenario seems to dominate political participation in the net except for the younger generation; thus, she ventured to suggest that “this may provide some grounds for the mobilization thesis”.
In a similar vein Delli Carpini (2000) argues that Internet can increase and enhance the participation of young adults that are already engaged, or interested but not engaged; he also suggests that it may help “to increase the motivation of currently disinterested and unengaged young adults”, although he points out that such a possibility is less clear.
Ward et al. (2003) [Q]  [Q] An additional contribution of Ward et al. (2003) is their prediction about Internet increasing (1) the awareness of citizenry that would, in turn, expect a more responsive government, and (2) the chances of ad hoc protests organized without traditional intermediaries such as parties, trade unions or lobbies. This is particularly interesting because most researchers have focused on finding relation between online activities and conventional offline participation while obviating the, at the moment of this writing, much more common unconventional modes of participation. draw similar conclusions arguing that the Internet “may bring some new individuals and groups into the political process—notably younger people, many of whom have grown up with the Internet as part of their daily lives.” They also noted that, despite the opportunities for interaction and networking, users tend to perform information gathering.
A number of researchers have conducted cross-sectional studies to find support for these hypotheses, especially regarding young adults, whose disengagement is a matter of concern among political scientists. In this regard, Kroh and Neiss (2009) found a positive effect of Internet use in the mobilization of people below 30; Bakker and de Vreese (2011) found that different CMC systems such as email, forums and SNS positively affects political participation, both online and offline; Oser et al. (2013) found that young adults are much more engaged online than the rest of the population and, moreover, so-called ‘online activists’ are “also involved in offline participation”, and Xenos et al. (2014) claimed that that “social media are positively related to political engagement” of young adults.

1.3.2 Beginning with ‘S’, political participation—Slacktivism?

Regarding social networking sites, Valenzuela et al. (2009) analyzed the relation between Facebook use by young adults, particularly Facebook groups, and online and offline political participation finding a positive, albeit, weak correlation. Very similar results were also reported by Bode (2012).
A similar study was conducted by Vitak et al. (2011) achieving mixed results: They found that Facebook activity, particularly if it was used for political purposes, was correlated with both online and offline political participation; however, they also found that the preferred modes of participation were superficial and with a lower degree of commitment, i.e. ‘slacktivism’ (Morozov, 2009; Carr, 2012) [R]  [R] It must be noted that Zuckerman (2014) took up the cudgels for so-called slacktivism: He argues that they are “thin engagement” and, thus, users just need to “show up” because “someone else [...] has done the thinking and concluded that what’s needed to persuade or to make a point is mass participation”..
That conclusion is consistent with prior findings by Baumgartner and Morris (2009) who studied the relation between consuming news through SNS (MySpace and Facebook) and political participation. They found that young adults tend to favour “news that shares [their] preexisting point of view” [S]  [S] At the moment of this writing it is a matter of concern if recommendation algorithms used by social media sites are creating a filter bubble or just reinforcing a situation predating social media use. Given the moment at which Baumgartner and Morris conducted their study it seems that algorithms are not to blame. The reader should consult section 2 of chapter 4 for more details on this issue..; that such kind of information does not improve their political knowledge; and it increases online political activity but not offline political actions, including voting intention.
A similar position is hold by Schlozman et al. (2010): they found that social networking sites such as Facebook were actually mobilizing young persons into political online actions but they also argue that “many forms of political engagement on these venues [e.g., friending a candidate] do not fall squarely under the rubric of a definition of political participation” in the sense of aiming to influence the actions of the government.

1.3.3 Political participation without political knowledge?

Even more worrisome that slacktivism is the apparent lack of impact that social media use has got in the political knowledge [T]  [T] Delli Carpini and Keeter (1997: pp. 10-11) defined political knowledge “as the range of factual information about politics that is stored in the long-term memory.” It is different from “attitudes, beliefs, and opinions”, and “cognitions that are incorrect or that are not subject to reasonable tests of correctness” do not qualify as political knowledge. of young adults. Such a conclusion by Vitak et al. has also been offered by Conroy et al. (2012), and similar findings have been reported by researchers analyzing different social media environments.
For instance, Eveland & Dylko (2012) analyzed the impact of political blog reading during the US 2004 Presidential Elections and found that it was “unrelated to political knowledge”.
Östman (2012) studied the relation between the involvement of young users with user-generated content (namely, the production and consumption of contents in MySpace, YouTube, and blogs) and political participation and knowledge. While UGC involvement can increase political participation it does not seem to affect political knowledge.
In a similar vein, Park (2013) reports that self-identified ‘opinion leaders’ in Twitter are more politically involved, more verbal, and they have got more followers than non leaders but, surprisingly, they do not consume more news than them. In other words, young politically involved Twitter users are not better informed than the audiences following them.

1.3.4 Mobilization of other groups

Regarding other unengaged groups apart of youth, Krueger (2002) found an intriguing negative relation between online political activity and family income: apparently, those individuals with lower incomes (and, thus, a lower SES—Socioeconomic Status) were much more engaged online in comparison to those with higher incomes. Hence Krueger ventures:
“[T]he anonymous nature of the medium allows those from lower status groups to feel more empowered in an online environment compared to analogous offline interactions [and] by drawing on a different set of resources, the same individuals may not be disadvantaged online, thereby potentially expanding the scope of those participating in politics”.
Similar claims were made by Gibson et al. (2005) who found that women and people from poor backgrounds were “equally likely to engage in online participation in general as men and higher social status individuals, once existing levels of political involvement and experience on the Internet are taken into account.”
There is an important caveat to such encouraging findings: Krueger himself points out that Internet access heavily depends on income and, thus, “without equal access, the medium will continue to advantage those types of people already engaged in politics”; moreover, other researchers found just the opposite: “the effect of the Internet is, contrary to the expectations, stronger in citizens with high income” (Kroh and Neiss, 2009).
In addition to that, there are also some daunting findings made by Hoffman (2012) in relation to the online and offline political behavior of those less educated. To start with, she distinguished between ‘political participation’ and ‘political communication’, two sets of actions which are usually conflated in virtually all of the studies. As examples of political participation she suggested ‘contributing money online’, or ‘starting or joining a political group on a SNS’. On the other hand, tweeting, discussing with others about politics using email or text messaging, or posting political material in any sort of social medium, are examples of political communication.
Taking into account such a distinction and performing a cross-sectional study, she made two surprising findings: (1) online political participation predicts voting but online political communication does not; and (2) online political communication is negatively related to education. In other words, less educated users discuss about politics more than other users, but such discussions do not have an impact on voter turnout.

1.3.5 ‘Subliminal’ mobilization?

All of the studies mentioned so far lay on the assumption that users are mobilized because of the more or less ‘explicit’ interactions they have with other users on social media (e.g., they discuss with others, they receive and forward political information, etc.) However, it is also possible that mobilization occurs because of users incurring—although barely noticing it—in latent political participation.
To the best of my knowledge, one of the first authors noting this possibility was Gustafsson (2012) who said regarding politically unengaged Facebook users:
“[P]assive users were also affected by recruitment attempts, by things they saw that their friends did, links their friends posted, and so on. It is time that the distinction between manifest and latent participation is brought into the research field.”
Barely five months after the publication of his work, another publication by an independent team revealed that some kind of ‘subliminal’ mobilization seemed feasible.
Bond et al. (2012) describe a randomized controlled trial where two groups of Facebook users received a message to encourage them to vote in the US 2010 midterm elections, and that also included an ‘I Voted’ button. In one group (1% of the eligible population) the message was purely informational: it just reported the number of friends that had clicked the button, while in the other group (98% of the population) it also showed a random selection of faces of the user’s friends who had already voted. The control group (the remaining 1%) did not receive any message.
The researchers found that both messages were able to mobilize users to click the button, but the so-called ‘social message’—the one with the friends’ faces—spurred more mobilization, it spread farther across the user’s social network, and it had an actual impact on voter turnout [U]  [U] On an individual basis the effect was tiny but given the size of the treated population the overall effect was quite important: Bond et al. estimated that the “social message increased turnout directly by about 60,000 voters and indirectly through social contagion by another 280,000 voters”.. And let’s not forget that the only difference between both messages were the faces.
This research raises additional questions. First, it opens an ethical and legal debate about the implications for social media sites—a debate which is totally out of the scope of this chapter; secondly, it makes one wonder about the myriad of subtle details imbued in social media interactions that may impact others’ political actions.

1.4 The normalization scenario

The normalization of political participation in social media has been approached from two slightly different perspectives. The first one suggests that Internet and social media will be just another channel that will supplement but not replace, nor expand, the possibilities available through other channels. Therefore, people will eventually use the available medium which is more appropriate for the action at hand, according to their interests and skills.
Wellman et al. (2001) can be considered a representative of this position, but similar claims about social media supplementing (while not replacing) offline political actions were made by Shah et al. (2005), or Gibson & Cantijoch (2013).
The other approach to normalization is not so focused on the media but on its users; particularly on their demographic traits, and whether they are different or—much more likely—similar to those of citizens traditionally interested in politics. Actually, this normalization perspective claim that political users of social media will be extremely similar in terms of socioeconomic status to politically engaged citizens, although advocates of this version of ‘social media normalization’ admit that they may be younger.
A number of researchers have supported this position. For instance, Norris (1999) notes “the appeal of the net for the more affluent and more educated”; while Krueger (2006) found that “those with the characteristics that predict conventional mobilization continue to hold a mobilization advantage in the new technological environment; only the mechanism by which they gain an advantage changes.”
This second interpretation of normalization is extremely important because, as Krueger points, it could be that political participation through social media could bring larger inequalities.

1.5 The reinforcement scenario

If social media was just another realm for the engaged citizens to participate in politics then it would simply keep the status quo; i.e., unengaged and disenfranchised citizens would not participate neither offline nor online, while engaged and enfranchised citizens would use any of those options.
However, if social media actually provided an advantage—no matter how small, then such an advantage would be played by those that are already advantaged over those who do not even participate in traditional ways. Hence, depending on the degree of that presumed advantage, social media would deepen the inequalities already present in our democratic systems.
Norris (2001: p.238) is one of the first scholars warning about the risk of social media reinforcing only those persons already interested in politics; later, a number of researchers have provided factual support for such an hypothesis by means of surveys, longitudinal, and cross-sectional studies; for instance: Jennings and Zeitner (2003), Weber et al. (2003), DiMaggio et al. (2004), Di Gennaro and Dutton (2006), Kroh and Neiss (2009), Schlozman et al. (2010), or Oser et al. (2013).
The commodification scenario
As with the reinforcement scenario, the commodification of social media regarding political participation is more a possible consequence of its normalization than an scenario on its own. In fact, one of the first scholars depicting such an outcome—Resnick (1997)— was actually describing it as ‘normalization of cyberspace’.
His position was simple: both utopians and their idea of cyberspace empowering people, and dystopians fearing cyberspace would imply control and repression were wrong. He argued that Web access was transforming cyberspace from a text-based realm with limited access into another mass medium appealing “to the economic, social, and political forces that had previously ignored it.” In this regard he said:
“Cyberspace has not become the locus of a new politics that spills out of the computer screen and revitalizes citizenry and democracy. If anything, ordinary politics in all its complexity and vitality has invaded and captured Cyberspace.”
In other words, the Web made cyberspace universally appealing and, thus, it was to mostly mirror the same virtues and evils of the offline world, including of course political activity. Resnick’s position contains a number of interesting and pessimistic arguments that, although unsupported by data at the time of his writing, have been proven rather far-sighted.
He claims, for instance, that newsgroups and listservs may have been a realm of passionate, individualistic, and egalitarian discussion, but organized groups such as political parties would seldomly reply to posts in such environments because “these groups are not seen as the main vehicles for influencing a mass public.”
Resnick depicts the Web, or better political websites, as the favoured way to reach “the public which is ripe for persuasion and manipulation [...] [the] free-floating agglomeration of Web surfers and searchers.” That is to say, Resnick considers that the deliberation possibilities of cyberspace are irrelevant because it would be transformed in another broadcasting medium; and it would be so because of “the harsh reality of masses of bored and indifferent citizens”.
Resnick’s harsh prospects about this consequence of the normalization of cyberspace were not unique; in fact, slightly earlier than him McChesney (1996) said:
“Aside from the question of access, bulletin boards, and the information highway more generally, do not have the power to produce political culture when it does not exist in the society at large. Given the dominant patterns of global capitalism, it is far more likely that the Internet and the new technologies will adapt themselves to the existing political culture rather than create a new one. Thus, it seems a great stretch to think the Internet will politicize people; it may just as well keep them depoliticized. The New York Times cites Wired magazine approvingly for helping turn «mild-mannered computer nerds into a super-desirable consumer market,» not into political activists”
Hence, an even more worrisome consequence of normalization than greater inequality is this undesirable transformation of social media, where its potential for many-to-many communication is reduced to an unending variety of one-to-many channels offered by organized groups, corporations, and celebrities.
Needless to say, McChesney and Resnick were not the only scholars that noted such commodification as a factor that can make it eventually unfeasible as a meaningful political tool: see, for instance, Rheingold (1993: pp. 301-306), Buchstein (1997), Thornton (2001), or Dahlberg (2005).
Given that at the moment of this writing virtually all major social media platforms are privately owned, the commodification of social media and its potential impact for political participation is an extremely relevant issue, albeit still an open one.

1.6 The null scenario

Strictly speaking, the null hypothesis should refer to the lack of any impact on political participation because of social media use. However, the truth is that an extremely large number of studies have found such an impact—actually, a significant positive correlation between interactive online activities and political actions. Therefore, those criticizing the argument about the positive impact of social media in political participation disagree in how significant such impact is, or in the direction of causality.
In short, critics claim that the impact of social media is weak, or that it is not social media the cause for political participation but just the opposite: politically engaged people use social media as another realm for political actions, and that is the reason for the observed online behaviors.
An additional problem is that most research must rely on cross-sectional studies. Such a limitation is almost unavoidable and, indeed, researchers have used the best available data which is not suitable for longitudinal studies—which are able to find differences before and after the treatment, in this case being engaged with social media.
Nevertheless, a few analysis of this second type have been conducted and, on those bases, some argue that the hype and hope about political participation being increased by social media should be tone down a little.
For instance, Jennings and Zeitner (2003) relied on survey data collected during the years 1965, 1973, 1982, and 1997 for a 1965 sample of high school seniors. Such data contained many details about political participation before—1982—and after—1997—the Internet; and, in addition to that, the data corresponding to 1997 also included surveys of part of the offspring of the original panel. Moreover, the social media types covered in the questionnaires were pretty varied: Usenet news, listservs, chat rooms, WWW and other computer services.
Some of the conclusions of their study were not especially surprising such as finding that the younger generation was using Internet much more than its parents, and that its political use of the medium was even higher. Others were rather interesting such as their suggestion that it is not Internet use what drives political participation but the opposite; in this regard they say: “other factors, most especially previous levels of civic engagement, are generating this relationship between political involvement and political use of the Internet.”
Kroh and Neiss (2009) worked on both cross-sectional data and a natural experiment based on the unequal deployment of home access to broadband Internet (both in Germany). Their analysis of such sources of data led them to conclude that the correlation between Internet use and political participation found in cross-correlation studies is mostly due to self-selection, i.e. citizens who are politically active tend to use Internet more than disinterested persons. Their longitudinal study supports that conclusion and they suggest “that only a small fraction of the correlation is attributable to a causal Internet effect”. This work is mostly consistent with the findings of Jennings and Zeitner (2003), although it does not exclude the causal impact of Internet: it only cuts down its impact.
Bimber and Copeland (2013) also performed a longitudinal study with data from the American National Elections Studies covering the 1996-2008 period in the United States obtaining mixed results. They found no significant correlation between Internet use and political participation in the period 1996-2004, and correlation between such use and a few political activities in the period 2000-2008, particularly ‘trying to persuade others’. They posit that the impact of Internet use on that particular kind of action, combined with the fact that such a correlation was strengthening across the entire period, could be due to “the emergence of social media during the mid-2000s, which are especially conducive to political discussion”.

1.7 Conclusions

Given the abundance of literature about social media and political participation you may be wondering if it has not been subject to any meta analysis. Fortunately, such a meta analysis exists; unfortunately, it depicts a complicated relationship between social media and political participation, to say the least.
Shelley Boulianne (2009) analyzed 38 different studies about Internet use and traditional political participation broadly covering the period 1995-2005, and she found no evidence supporting the demobilization hypothesis. Instead, she found evidence of a positive correlation between Internet use and political participation that seemed to increase across time. She was cautious however, noting that the effect was weak, the increase non monotonic, and the direction of causality unclear. All of this is mostly consistent with the findings of Krohn and Neiss (2009), and not entirely inconsistent with the findings of Bimber and Copeland (2013).
Going beyond that comprehensive survey, when putting side by side the research conducted during the last decades it is poignant how few things can be said about this topic with enough confidence, namely:
Social media is not diminishing political participation; besides, it could help to engage those citizens who are not interested or disenfranchised, particularly if it is used not just for information gathering but to discuss political matters.
However, such an scenario has a number of requirements that are improbable by just relying on market forces (McChesney, 1996). Moreover, while the opportunities that social media offer to disenfranchised and unengaged to participate in politics are important, they are even more so for those engaged and enfranchised.
In other words, social media on its own is unlikely to mobilize previously unengaged people in enough numbers while it is very likely to maintain the status quo, and not unlikely to reinforce it—i.e. widening the participation gap between those engaged and enfranchised, and those unengaged or disenfranchised. There are, however, a few concrete exceptions to that overall normalization-reinforcement scenario.
On one hand, young adults seem to be actually moved by social media to participate in politics; the main caveats are that their political knowledge is poor [V]  [V] It is a matter of reflection which situation is worse for democracy: lack of participation or participation of uninformed citizens?, and that there is a very real risk of slacktivism being their major mode of participation.
On another hand, while disenfranchised people are not those who have the most to gain from social media, parties and social movements targeted at their needs can benefit from social media use (Norris, 2001: p. 238) and, in turn, give them a voice.
Commodification of social media is, however, the most worrisome problem because it can effectively degrade the most important feature of social media, namely, its potential for many-to-many communication.
All of these phenomena are taking place simultaneously, affecting different groups at different moments, and with different intensity.
A final conclusion is that although research is abundant, it has been conducted in such a fragmented way that makes extremely difficult to paint a meaningful picture.
To start with, most of the research conducted up to date makes direct comparisons almost impossible: there is little overlap between the modes of action that are considered political participation in each study (both online and offline); there are important differences in the populations and subpopulations subject to study; and some research was conducted in relation to major elections while other was conducted without any major election in the horizon.
On top of that, the literature has mainly focused in the United States and, thus, it is difficult to determine if the findings can be generalized to other countries—particularly, new democracies.
Finally, most of the research relies on cross-sectional studies and, thus, the direction of the causality link is subject to discussion: i.e. is social media use causing political participation or is the intent to participate, such as deciding a vote, the cause for social media use? With the evidence at hand the most we can say is that if social media is really causing political participation the impact is very weak.
To sum up, further research is needed on this topic but it has to:
1. Determine a minimum range of actions (both online and offline) that define political participation and that, in the case of social media, are stable enough in time, and broadly generalizable across platforms. Those actions should cover both conventional and contentious modes. Researchers should adhere to those ostensive definitions of political participation and not change them without justification.
2. Pay more attention to other countries in addition to the United States, particularly to new democracies and poor democracies.
3. Clearly separate studies conducted in electoral and in non electoral periods.
4. Clearly target different groups of interest such as young adults, minorities, disenfranchised citizens, etc.
5. Aim to conduct longitudinal studies in addition to cross-sectional ones to better determine both the direction of causality and its impact.
6. Aim to develop measures of intervention to shrink the participation gap between engaged and unengaged people or, in other words, try to make the reinforcement scenario less likely and increase the chances of mobilization.
Needless to say, none of those requirements is easy, and, on top of that, the last one could be a matter of ethical and legal concerns [W]  [W] Let’s remember the study by Bond et al. (2012) on manipulating users’ Facebook walls to increase voter turnout.. However, given the crucial importance of political participation and the pervasiveness of social media use, citizenry is entitled to a more substantial answer from the research community than ‘it’s complicated’.

2 Political actors

“Any person or organization depends ultimately on public approval, and is therefore faced with the problem of engineering the public’s consent to a program or goal. We expect our elected government officials to try to engineer our consent for the measures they propose. [...] The engineering of consent is the very essence of the democratic process, the freedom to persuade and suggest.”
(Bernays, 1947: p.14)

2.1 Introduction

Previous chapter rested on the premise that the more political participation of the citizenry the better the democracy. Thus, it discussed a number of modes of participation, as well as the impact of social media on them. The underlying idea is that, no matter the concrete ways of taking part in politics, individual citizens can and must influence the government decisions. In this regard political knowledge is fundamental—although it was not discussed at depth—for citizens’ actions being informed. It was also implied that the definite political action of citizenry is voting, and that votes eventually drive government.
Yet, I did not fully explore what ‘voting’ is. In that regard, I must briefly recall that democracy can be implemented in two rather different ways: representative and direct [X]  [X] The interested reader should consult Haskell (1959: pp. 85-120). It must be noted that advocates of direct democracy regularly suggest that CMC can make direct democracy eventually feasible; however, that is not the point of this chapter: social media and direct democracy are briefly reviewed in chapter 7.. Virtually all modern liberal democracies are of the representative kind, although some instruments of direct democracy—such as referenda—can be employed more or less frequently. Therefore, and simply put, in its most usual sense ‘voting’ refers to citizens choosing their representatives for government.
Needless to say, to make a choice, a pool of alternatives must be available; and, broadly speaking, that is the main roles of political parties. Indeed, despite their flaws, they are usually considered the foundations of modern representative democracies [Y]  [Y] Aldrich (1995: p. 3) suggested that Schattschneider’ statement about democracy being unthinkable save in terms of parties should be paraphrased into “democracy is unworkable save in terms of parties”..
Political parties, however, are not the only political actors playing a role in modern democracy. Like many other concepts appearing in this book, ‘political actor’ has been defined by a number of authors; however, I think that the perspective of McNair (2011: pp. 3-14) is the one that best fits the goals of this piece: he focuses on the communication goals of political actors, and the way in which media—including social media—affects the interplay between them and the audience (i.e. the citizenry).
Indeed, McNair (pp. xiv) endorses the ideas of Walter Lippmann (1921: pp. 135-136) about ‘persuasion’ and ‘manufacture of consent’ to create ‘public opinion’:
“That the manufacture of consent is capable of great refinements no one, I think, denies. The process by which public opinions arise is certainly no less intricate than it has appeared in these pages, and the opportunities for manipulation open to anyone who understands the process are plain enough.
The creation of consent is not a new art. It is a very old one which was supposed to have died out with the appearance of democracy. But it has not died out. It has, in fact, improved enormously in technic, because it is now based on analysis rather than on rule of thumb. And so, as a result of psychological research, coupled with the modern means of communication, the practice of democracy has turned a corner. A revolution is taking place, infinitely more significant than any shifting of economic power.
Within the life of the generation now in control of affairs, persuasion has become a self-conscious art and a regular organ of popular government. None of us begins to understand the consequences, but it is no daring prophecy to say that the knowledge of how to create consent will alter every political calculation and modify every political premise. Under the impact of propaganda, not necessarily in the sinister meaning of the word alone, the old constants of our thinking have become variables. It is no longer possible, for example, to believe in the original dogma of democracy; that the knowledge needed for the management of human affairs comes up spontaneously from the human heart.”
This premise can be grossly simplified as follows: public opinion drives government decisions, mainly through voting but also by means of pressure groups, demonstrations, or even rioting; therefore, those with a saying in government policies either try to attain power themselves, or persuade the public to influence government by proxy. Moreover, government will also try to make its decisions more palatable by persuading the citizenry.
In this regard, McNair (p. 5) provides the definition of political actors which is used along this book:
“[T]hose individuals who aspire, through organisational and institutional means, to influence the decision-making process. They may seek to do this by attaining institutional political power, in government or constituent assemblies, through which preferred policies can be implemented. If in opposition their objectives will be to obstruct existing power-holders, and have them replaced by alternatives.”
He also provides a list of political actors in addition to political parties, namely: public organizations (e.g., trade unions, consumers’ associations, NGOs, corporate lobby groups), pressure groups (e.g., the Occupy movement), terrorist organizations, and governments.
It must be noted that this chapter only focuses on those actors that are more or less institutionalized and that tend to engage in conventional politics: parties, government, and public organizations. Loosely organized political actors which, besides, tend to rely on contentious politics are reviewed in chapter 5.

2.2 Political parties

Political parties in democracy [Z]  [Z] Although it is out of the scope of this chapter, a comprehensive definition of ‘political party’ should also consider those working under non-democratic regimes (cf. Sartori, 1976: pp. 52-57). are organized groups of like-minded individuals that (1) pursue goals reflecting their underlying ideology, (2) try to persuade the citizenry about the correctness of such goals and their corresponding policies, (3) periodically put both to the test of elections, and (4) adhere to the constitutional rules governing the ways to attain, exercise, and leave power (McNair, 2011: p. 5). The Democratic Party in the US, the Liberal Democratic Party in Japan, or the Christian Democratic Union in Germany are typical examples of political parties.
Before discussing how social media has been incorporated into activities of political parties we must take into account that such activities are inevitably intertwined with both the electoral system, and the legislatures and offices that parties aim to control (Aldrich, 2006). However, I have tried to separate those aspects and, thus, social media use by elected officials and representatives is covered in a different section of this chapter, while social media and elections is devoted a chapter of its own.
By organizing the literature that way, it is immediately obvious the relative importance that political parties and politicians give to social media: quite much while campaigning, not so much after elections, and mainly nothing when neither in campaign nor in office [A]  [A] Some anecdotal evidence to support this argument. As of June 2015, Twitter’s website devoted to best practices for different sectors ( provides the following scenarios for ‘government’: ‘government & elections’, ‘elected officials’, ‘candidates’, and ‘agencies’. The equivalent site for Facebook ( depicts government and politics as another industry which goal is winning elections, inter electoral periods are just another opportunity to be prepared for the next election..

2.2.1 Non-mainstream political parties

One of the most obvious promises of social media for the political party system is the feasibility of bootstrapping new parties online. Phillips (1995) was one of the first suggesting such a possibility and, although not directly related to social media, his depiction still deserves attention:
“As technology gives Americans greater options in entertainment, media and shopping, pressures are rising for more say in politics. The choice between Republicans and Democrats, the two megaparties created 140 to 170 years ago in the Industrial Revolution, is increasingly unsatisfactory. Some 53% to 58% of Americans want a new third party.”
Similar remarks were also made by Norris (2001: p. 238) but the truth is that evidence did not seem to support that kind of optimism (Gibson et al. 2000; or Margolis and Resnick, 2000: pp. 53-66). Actually (1) while in 1994 most political party websites in the US belonged to minor parties, by 1996 the majority were affiliated with the Democratic or the Republican party; (2) minor parties pioneered the use of Usenet and listservs but once the Web arrived they were not able to reach the same content quality and updating frequency of major parties; and (3) minor parties achieved better visibility on the Web than on traditional media but, still, they were behind major parties, particularly during elections.
This prospects have changed rather recently, in Europe at least. A relatively large number of emerging political parties (e.g., the Five Star Movement in Italy, Syriza in Greece, or Podemos in Spain) are heavily using social media and gaining ground on traditional parties. However, it would be quite audacious to say that social media is the driving force behind them [B]  [B] sIt is of interest, however, that many of those new parties can be considered somewhat successors of social movements which took great advantage of social media (e.g., the Five Star Movement originating in the V-Day, or Podemos in the 15M movement). Such kind of movements is explored in chapter 5.. Still, it is also unquestioned that social media is playing an important role, and there is an incipient literature on the topic—particularly regarding the Five Star Movement (M5S) which I am using as an illustrative example [C]  [C] The reader may wonder why I am not using other examples such as the Pirate Parties of Sweden or Germany. I have two main reasons: (1) bluntly put, given the number of representatives they have been able to obtain during its decade-long life, it is clear that they are not considered as a viable option by the electorate, (2) they are too focused in a limited number of issues and, in that regard, they resemble more pressure groups rather than parties. For more information on this parties the reader should consult Burkart (2014)..
Bartlett et al. (2013) say the following about the way in which M5S started:
“Beppe Grillo is one of the first politicians to have embraced this transformational change, and by using social media [...] his movement has grown from practically nothing to a major political force in Italy in the space of three years. The creation of the Movimento 5 Stelle [...] was announced by Grillo, then a popular Italian comedian and blogger, in a blogpost dated 9 September 2009. […] Grillo has been able to transform this online following and support [one million Facebook fans, and 700,556 followers on Twitter] into real world political impact.”
Yet, we must take into account that Grillo did not simply decide to found a political party; when he made such a movement he already had abundant experience engaging people in massive demonstrations such as the V-Days [D]  [D] See the section “Seattle, Philippines, Spain and Italy: Triggering demonstrations” in chapter 5 Contentious politics. and, thus, he had reasonable expectations about the outcome of such calling.
Anyway, Bartlett et al. also insist that social media is not the only factor in the success of M5S, and they point out that it has got a wide array of offline actions which are equally important—e.g., weekly meetings across local communities, or the many rallies all over the country where Beppe Grillo takes part. Very similar views are exposed by Bordignon and Ceccarini (2013) who provide an extremely detailed review of the history of M5S. They also give an additional reason that Blatter et al. do not mention, and that was later fully explored by Miconi (2014)—Grillo’s charisma [E]  [E] As a matter of fact, in the original statutes of the party it was established that the “headquarters” was Beppe Grillo’s blog.. Indeed, Miconi makes a suggestive proposal: that the success of M5S is not actually an example of the power of the Web, but an example of power in the Web.
If such interpretations are correct it may be that Beppe Grillo—as Obama earlier—is not necessarily a prime example of the potential of social media in politics but a brilliant exception [F]  [F] Gibson et al. (2000) analyzed how Italian political parties were using Internet and the Web; they found that, “during election campaigns it appears that major parties have the capacity to outstrip their rivals, even in cyberspace. The smaller, less well-represented parties do not appear to be capitalizing on the new medium to communicate their message.” If non-mainstream Italian parties were underperforming during the early 2000s despite of the Web, it seems far-stretched to claim that blogs were the reason for the success of M5S—particularly when campaign blogs have been consistently underused by parties (see chapter 4 Elections).. At this point, however, we have no other alternative but to wait for events to unfold, and only then check whether social media can be a determinant for emerging parties or not. Yet, we must also take into consideration that it seems to be particularly appropriate for populist parties (Bartlett, 2014):
“It is distributed, non-hierarchical and democratic. [...] an alternative to the mainstream media, which many supporters of populist parties strongly distrust. [...] not controlled by the elites: The content is generated by us—the honest, hard-working, ordinary citizens—exactly those people who the populists are defending. [...] It [Internet] is the only media source populists have more faith in than their compatriots. The short acerbic nature of populist messages works well too. [...] Humour, outspokenness, pithy put downs and catchy slogans: These are the DNA of cyber culture.”

2.2.2 Mainstream political parties

As I said above, mainstream political parties tend to pay attention to social media mainly during electoral periods, and literature on that issue is covered in chapter 3. It is also true, however, that since the mid-1970s—at least regarding the US, to a lesser extent in other countries—‘permanent campaigning’ [G]  [G] Simply put, ’permanent campaigning’ is an strategy where governing acts are exploited for the purpose of sustaining popularity, purportedly to increase prospects of reelection (Blumenthal, 1982—cited by Heclo, 2000: p. 2). has been an increasingly common way of doing politics. Needless to say, this has had an impact in the way parties and politicians use social media.
A number of researchers have studied the application of social media by mainstream political parties when not campaigning. The main research question they aimed to answer is whether social media is a useful tool for intra party democracy, and if it can be used to engage the citizenry at large. Unfortunately, the short answer is ‘no’, in part because of social media features, in part because of political parties themselves.
Mainstream parties do not seem to be interested in increasing internal democracy or forwarding feedback from their rank-and-file members to their elites; indeed, many of them are actually reducing the opportunities for interaction. On another hand, mainstream political parties seem to try to exert certain level of control on social media, and they tend to use it as another broadcasting top-down medium. Finally, political parties seldom interact with the citizenry and, in fact, they have much to lose in such interactions. Evidence on such issues has been mainly collected through surveys and interviews with those responsible of ICT or communication within political parties.
For instance, Gibson and Ward (1998) found that UK parties recognized “the interactive opportunities that the Internet offers, [but] they consider its capacity for the downward dissemination of information to be of greater importance.” A later study by the same authors, (Gibson and Ward, 1999) focused on the application of intranets [H]  [H] Actually, Gibson and Ward used the term ICCS (Internal Computerised Communication System); in this particular case I have chosen to replace the original one because it seems an unnecessary neologism that was briefly used by political scientists in the UK. by parties, obtained similar findings: intranets were not being used to collect feedback, extending participation, or empowering the party members; they were mainly used to spread information in a top-down manner.
Gibson and Ward (1999) also explored how intra-party websites were being used. They considered websites belonging to local branches, elected officials (such as members of parliament and members of the European parliament—MEPs), and pressure groups orbiting the party. The similarities between those websites and current social networking political profiles, Facebook groups and pages are obvious and, thus, there are some useful readings here:
Most of that is still valid today, particularly in social networking sites: subnational profiles are less common and they publish content less frequently; representatives try to depict an image of constant activity in front of their constituents, and—although not necessarily at the request of the national party—it is pretty common for politicians in social networking sites to remove material [I]  [I] Deleted tweets are particularly common among politicians and, although according to Twitter’s Developer Agreement they should be removed, there are services such as Politwoops (originally from the Netherlands) that inform about such tweets (see Fig. 3.2). Shortly after finished writing this chapter Twitter decided to “to pull the plug on Politwoops” effectively killing the service (Gates, 2015)..
Gibson and Ward (1999) also noted that despite the centralized and lack of editorial control on the internet, national parties (particularly if conservative) tried to exert a certain amount of oversight on local websites; similar findings were also reported by Nixon and Johansson (1999) and it would be surprising that such a trend had disappeared in SNS.
Conversation, a commonly cited feature of social media, is extremely uncommon, not very productive and, actually, political parties tend to avoid it. For instance, Nixon and Johansson (1999) said about chat rooms:
“[they] are often ‘ordinary members’ exchanging ideas between themselves, and not a bottom up flow informing the policy makers. The leadership/party officials do engage in discussion via the net, but this tends to be on special occasions that are time limited and, generally, have the leader responding to carefully pre-selected points or questions. [...] The chat room will often have a moderator who will monitor the discussion and has the power to eject those asking inappropriate or difficult questions. Even when a party decides to ‘moderate’ a chat session and to use the data collected as an element in its decision making [...] the costs of administration and response are high and people may not get the level of contact that they desire at the time that they wish.”
figure FnTSM-Politics and Social Media-Political Actors and Political Participation-Fig-1.png
Figure 2.1 President Barack Obama replies to a question asked by a Twitter user using the hashtag #askpotus on May 2015. A number of users try to participate in the exchange by replying the President’s answer. The thread is interactive but cannot be considered any sort of meaningful conversation.
figure FnTSM-Politics and Social Media-Political Actors and Political Participation-Fig-2.png
Figure 2.2 An assorted selection of deleted tweets by Republican presidential candidate Ted Cruz as originally collected and shown by Politwoops. By allowing social media users removal of historical content and enforcing such measures onto third parties, SNS are effectively allowing political actors to rewrite their own history at will.
All of this still applies to SNS (e.g., Small, 2010) and it has even been exacerbated. For instance, it is virtually impossible for any staff to tackle with all the mentions a political party receives; when Q&A sessions are organized the number of mentions can be extremely large, only a few will be replied, and not necessarily to the satisfaction of the citizen or the audience (see Fig. 3.1). Moreover, although Twitter is not a chat room, troublesome accounts can be blocked, or reported for suspension. Similar issues affect other SNS and social media services and they are, indeed, deeply imbued in the nature of the medium.
On top of that, SNS, chat rooms and forums are ripe for trolling and abuse and, hence, they can easily backfire. For instance, Pedersen and Saglie (2005) report that some Norwegian parties had to close their forums because they were being used by right-wing extremists; Jackson and Lilleker (2009) describe how some political Facebook walls were ‘defaced’ without the political party removing the comments—revealing that for that party Facebook was above all posturing.
So, in short, political parties and their elites are not actually engaging in conversations with rank-and-file party members or with the electorate; instead, they create SNS profiles and use other social media instruments to “give the impression of interactivity-as-process while actually offering interactivity-as-product” (Jackson and Lilleker, 2009). This is problematic because, as they tend to ignore most of the non written norms of social media, political parties risk being not well accepted by part of the users.

2.3 Government

Under this epigraphe I am conflating an exceedingly large number of entities that can be classified under one of the following broad categories: executives, legislatures, subnational governments (including their own executive and legislative branches), and the so-called ‘machinery of government’ from the cabinet down to agencies and units, going through the different departments or ministries.
It is out of the scope of this chapter to provide detailed definitions for each of them; moreover, it would be of little value because, when using social media, all of such entities seek the same basic goals: to publicize their work, to increase transparency, and to engage citizens in both governance and policy-making (adapted from Katz et al., 2013: pp. 16-17).
Indeed, the first two goals are closely related, and to attain them the government approaches citizenry in a different manner than to attain the last one. Klang and Nolin (2011) labeled the first approach as the ‘transparent citizenship strategy’—which “aims to further the traditional values of freedom of information access”—and the second one as the ‘interactive citizenship strategy’—which “emphasizes social media as an instrument for generating feedback”.
Needless to say, there is no sharp line dividing them, and the same medium can be employed under both strategies. Moreover, activities that purportedly belong to the interactive citizenship strategy (e.g., tweeting from a politician’s account) can be used solely for transparency purposes which, in turn, can be in fact permanent campaign tactics.
Anyway, this section is organized in two different subsections: The first one covers literature about how the government has been using social media to inform the public, not as much as collective bodies but mainly as individual officials. The second one surveys research about the different ways in which social media has been used to engage citizenry to take part in governance and policy-making.

2.3.1 Citizens. This is government

It goes without saying that government in modern democracies is not monolithic and different and independent branches exist; typically but not necessarily those branches are three: executive, legislative and judiciary. It is quite interesting—and somewhat worrisome—that when referring to ‘the government’ some people means just the executive, or at most, the executive and the legislative; of course, it is not different in social media [J]  [J] It must be noted, however, that the implications of members of the judiciary using social media and interacting with members of the executive or the legislative, and with other political and economical actors would be troublesome to say the least. As Janoski-Haehlen (2011) put it: “Judges should be aware of the repercussions of dishonoring the judicial system and should try to avoid doing so at all costs. That is not to say that judges cannot participate in social media sites in their personal capacity, but they must be cautious in what they post, share, comment, and tweet on social media sites.” To the best of my knowledge the judiciary is not employing social media beyond public relations and press releases and, hence, it is not playing any role as a political actor within social media..
Margolis and Resnick (2000: pp. 79-97) argued that the executive (no matter national or subnational) was much better represented in the Web than the legislative, giving the citizenry the impression that it is the key player within the government. Though, they stated that such a situation would not persist for a long time, and that legislatures would eventually provide information and services on the Web in similar ways to the executive. However, they also noted that it was not unlikely that most of such information would focus on the legislators as individuals instead of the legislature as a collective organ. They argued that such “personal” websites would “be used largely as a communications device to protect and to enhance the advantages of incumbents.”
Their concerns in this regard were well-founded: although virtually all legislatures across the world have got a website (Leston-Bandeira, 2007) they are not heavily using social media, legislators are the ones who do it, and they use social media as a permanent campaign tool. Needless to say, members of the executive are not different. However, branches of the executive, and particularly certain agencies are using social media in interesting ways as I will discuss in the next section.
Taking all of that into account, it does not make sense to differentiate politicians as belonging to the executive or the legislative, to national or subnational governments. Instead, it will be much useful—and illustrative—to explore their use of different types of social media along time; that way we may discern a stable pattern of use. Thus, I will cover email, websites, blogs, and social networking sites. Use of email

I already mentioned in the Introductory chapter the announcement by the Clinton administration about their use of email to “bring the Presidency and this Administration closer and make it more accessible to the people.” Actually, Bill Clinton did not use email but, still, citizens were sending their messages to the president. On this, Owen and Davis (2008) said:
“The Internet staff in the Clinton administration was inundated, and printed e-mail messages piled up on the floor. By 2003, the White House was receiving approximately 15,000 e-mails a day to which the only response was an acknowledgment that the message was received”
The acknowledgement message in the website was even more disheartening: “There will be no further electronic response. If appropriate, a reply will be sent to you via U.S. mail.”
The relation of George W. Bush administration with email was not different and, indeed, the situation has not changed: politicians and agencies of any type and condition receive too much email, they read too little, and they reply virtually none (Taylor and Kent, 2004; Williamson, 2009). Under the light of these facts it is quite shocking that email is usually praised by politicians as the most convenient way to interact with citizenry (e.g., Kent, 2004), and email addresses are prominently shown at different governmental websites across the world. Use of websites

Properly speaking, websites (at least as conceived during the 1990s) are not social media; Margolis and Resnick (2000: pp. 4-5) described the differences between both mediums extremely well:
“If human conversation is the real-world paradigm of a newsgroup, a presentation is the paradigm of a Web site. [...] Unlike conversations, presentations value talents and demand expertise. They are fundamentally inegalitarian. [...] An audience ordinarily is not expected to participate in the presentation. They are there to be entertained, informed, inspired, awed, manipulated, energized, and the like. [...] The Web site is typical of the new Net, just as the newsgroup or listserv is typical of the old Net.”
This means that if I strictly adhered to my own depiction of social media I should skip one decade—from the early-1990s to the early 2000s—and describe it as ‘the Social Media Winter’. However, some words are required on websites for two main reasons: First, politicians invested huge amounts of effort (and money, mainly public) into government websites because they, purportedly, would be able to inform and engage people. Second, although minority, certain amount of interaction was available in websites under the guise of forums, and later blogs.
So, in which ways did politicians employ websites?
To answer that question a good amount of research has been conducted, mainly on the websites of representatives. Among the most commonly studied legislatures we can find: the British House of Commons [K]  [K] e.g., Ward and Lusoli, 2005; Norton, 2007; or Vicente-Merino, 2008.; the US House of Representatives [L]  [L] e.g., Zittel, 2003; or Taylor and Kent, 2004.; the Swedish Riksdag [M]  [M] e.g., Zittel, 2003; or Vicente-Merino, 2008.; the German Bundestag [N]  [N] e.g., Zittel, 2003.; the US Senate [O]  [O] e.g., Taylor and Kent, 2004.; the Portuguese Assembleia da República [P]  [P] e.g., Vicente-Merino, 2008.; or the EU Parliament [Q]  [Q] e.g., Vicente-Merino, 2008. That research has usually relied on content analysis of the websites, and interviews with the representatives; its findings are not encouraging.
First, as Margolis and Resnick said, most websites were purely informational [R]  [R] e.g., Taylor and Kent, 2004; Ward and Lusoli, 2005; Norton, 2007; or Vicente-Merino, 2008., or as Zittel (2003) succinctly described them: “digital brochures”. Despite this, legislators claimed that the Web was useful not only to keep informed their constituents and the media, but also to receive feedback from their constituents and give them timely responses for their answers.
Unfortunately, although representatives were talking about dialogue, their websites implied that they “believe that providing information (one-way communication) is ‘dialogue’” (Taylor and Kent, 2004), and only a few ones were genuinely trying to engage with their constituents (Ward and Lusoli, 2005).
Moreover, an important number of representatives provided a website just for posturing (Ward and Lusoli, 2005), or considered it an obligation due to their position (Vicente-Merino, 2008); many of them were based on templates provided by their parties, and after the elections many representatives tended to neglect them (Ward and Lusoli, 2005).
When deeper analyses of the contents were conducted it was found that websites were used “to bolster the position of [one’s own] party in the political system” treating the citizens as a passive audience (Norton, 2007). In other words, as Margolis and Resnick had feared, purportedly legislative websites were being used as a permanent campaign tool for the advantage of the incumbents. Use of blogs

Maybe the first politician with a blog was Howard Dean (Meeks, 2003) who launched his Blog for America (see Fig. 3.3) during his short campaign in the 2004 Democratic primaries. However, its blog was not only authored by Dean. Thus, the privilege of being the first politician that blogged about his activity is for Tom Watson—a Labour member of the parliament when he started blogging. He considered a blog was the best way to inform about his job, express his ideas, rebut claims made against him directly, and also to engage with citizens. In his own words:
“It’s wonderful to get feedback and emails from readers, and nice comments from other bloggers. [...] At the moment I believe I’m the first Westminster MP to have a blog, but I believe it will be the main method of communication for politicians within ten years.”
It is true that other representatives and politicians started blogs after Watson but they are not the norm; after all, as Watson said, “it’s a political risk”.
Quite interestingly, despite being American the first politicians with blogs—the aforementioned Howard Dean, but also George W. Bush with his campaign blog—it seems that blogging politicians are a “British thing”. Hence, most reports about politicians’ blogs have dealt with those by UK representatives. The main research question being whether blogs could fill the gap between constituents and legislators’ websites.
On this issue Auty (2005) found that blogs provided a more personal approach to the representative, and were not as partisan as websites (similar findings were also reported by Jackson, 2008); they allowed the legislator to inform their constituents about their activities, and receive feedback from users through the comments. Unfortunately, comments were seldom replied and no proper dialogue was conducted; indeed, top-down communication was the norm.
Similar findings were reported by Jackson (2008) and Williamson (2009); moreover, both noted that blogs were not very popular among politicians because they were time-consuming and provided little value.
Nevertheless, those representatives that blogged regularly were growing some sort of community, and when comments were available the blog helped to start conversations not only with the legislator but among constituents (Jackson, 2008). On another hand, such communities seemed to be different from the constituency of the representative and, thus, the legislators could be devoting time and resources addressing topics with which their constituents were not actually concerned.
It must also be noted that members of parliament—particularly if they engaged in novel ways of interaction with citizenry—were able to reach certain prominence in the media and, in turn, attract readers to their blogs. For lesser known politicians, such as those at subnational levels, blogging was probably not worthing the value. In this regard, Wright (2009) found that blogs by UK councillors had got a scarce readership and even fewer comments (which were seldom replied, it must be said), and, still, some councillors admitted they were self-censoring themselves to avoid conflicts with their political party.
figure FnTSM-Politics and Social Media-Political Actors and Political Participation-Fig-3.png
Figure 2.3 Howard Dean’s blog when running as Democratic presidential candidate in 2003. Use of social networking sites

Services such as MySpace, Facebook or Twitter have got a number of features that differentiate them from websites and blogs, and make them simple and attractive for many people, including politicians. To name a few: SNS were the first time that hosting one’s online presence using a free service was not a stigma but a sign of modernity [S]  [S] Try to imagine the public’s reaction if Bill Clinton had decided to host the White House site in GeoCities—provided it existed at the date.; the stress about having something to say was virtually gone because in SNS verbosity is the real problem; finally, audiences are there just waiting for you to join—if you are a politician—so they can follow you. Moreover, the ease for interacting with other users, uploading multimedia content, and spreading any kind of material were also of interest.
figure FnTSM-Politics and Social Media-Political Actors and Political Participation-Fig-4.png
Figure 2.4 Screenshot of the Twitter account of President George W. Bush in late 2007. The President did not tweet himself but the tweets were mirrored from The White House RSS feed.
Because of this, SNS have become another realm where politicians ‘must be’; however, unlike websites or blogs, they are simple enough for anyone being able to single-handedly manage it. For instance, at the moment of this writing, about 75 % of the UK MPs, 99% of US representatives, and all US senators have got a Twitter account [T]  [T] As a matter of curiosity, one of the first politicians with an associated SNS profile, concretely on Twitter, was George W. Bush under the name of @TheWhiteHouse (see Fig. 3.4). However, he was not tweeting and, indeed, the profile was just a mirror for the RSS feed provided on the White House website.; and in other countries the situation is comparable, with substantial numbers of politicians having a profile on Twitter, a Facebook page, a YouTube channel, or all of them.
Thus, the question is no more whether politicians are using SNS but for what are they using them. Fortunately, there is some research on the topic although it is mostly focused in UK and US politicians. According to Williamson (2009) there are four reasons for politicians to use social networking tools, namely: “organisational, activism (issues based), campaigning (for re-election) and just to be there!” However, the principal is campaigning: representatives use their SNS profiles “to keep constituents informed and the MP on their radar”.
Jackson and Lilleker (2009, 2011) reached similar findings when studying the use of SNS by UK representatives: they use SNS mainly for self promotion, there is little interaction with constituents and, instead, “they interact with personal friends and local party members”. Except for a few legislators, SNS are mainly used as a broadcasting medium. Still, Jackson and Lilleker argue that Twitter seems to be different from other prior tools such as websites and blogs, and it seems to be “attracting those not previously considered to be at the forefront of Internet usage”.
Representatives in the US Congress are not different: Golbeck et al. (2010) found that legislators who use Twitter employ it for self-promotion by informing of their activities, and publishing links to news pieces about themselves, or to their own blogs. Like UK MPs, US representatives who were interacting with the public were a minority and, yet, Golbeck et al. noted that there could be problems of scale when Twitter reached larger audiences—actually, at this moment there is such a problem. Finally, Golbeck et al. also argued that Twitter use in the US Congress has little to do with transparency and everything with permanent campaigning.
In addition to the research above mentioned, this self-promotion and permanent campaigning behavior has been consistently observed in different countries and institutions: e.g., the Norwegian Storting (Sæbø, 2011), majors of Turkish cities (Sobaci and Karkin, 2013), or members of the European Parliament (Lilleker and Koc-Michalska, 2013).
So, in short, social media of different kinds have been employed by politicians—both in the executive and the legislative—during the last two decades. The purported aims of using such technologies are publicizing their work and increasing transparency. However, most politicians are focused on publicity over transparency; moreover, when they inform about their duties and activities are not actually giving information to the public but campaigning for reelection.
With regard to citizens’ engagement, it seems that politicians are not particularly interested on it; still, if they were it would be an enormous technical challenge to have meaningful dialogues.

2.3.2 Government. This is the people

Employing social media to engage citizens in policy-making and governance has some ties to e-government, a field with an extensive literature. I am not interested, however, in “classic” approaches to e-government where the user solely interacts with a site. Instead, I am interested in those approaches where the citizen expects—and is expected—to interact with other fellow citizens or with officials and representatives to improve the operation of government, or to contribute to some policy issue.
This form of incorporating social media to e-government could arguably serve different purposes, such as giving people a voice regarding policy making and implementation, co-production of governmental materials and services, and crowdsourcing solutions and innovations (Jaeger et al., 2012).
It must be noted, however, that we have a long way to go to fully understand how to properly integrate citizens feedback in policy making and other government activities (Bertot et al., 2010). Certainly, there exist a number of success cases showing the feasibility of co-production and crowdsourcing in certain domains (see next section); however, it is not easy to generalize those lessons to other domains.
The truth is that at this moment citizens are trying to engage with their governments but that engagement is far from being productive. Hence, I am organizing this section along the following topics: engagement with top level officials; engagement with government agencies or subnational governments; and other kinds of engagement—mainly spontaneous and unsolicited feedback. Engagement with top level officials

To the best of my knowledge, one of the first administrations providing citizens with a scheduled online mechanism to interact with officials was that of George W. Bush. Inaugurated on April 2003, “Ask the White House” [U]  [U] You can visit the archive version as of December 2008 at: (see Fig. 3.5) was an online forum for citizens to submit questions to executive officials. The transcripts of those chat sessions were archived and publicly available and, moreover, citizens were able to suggest the officials they wanted to take part in the sessions; besides, from August 2004 it offered a sharing mechanism through email. The service run without interruption until December 2008 and covered tens of officials, from the Chief of Staff to the Secretary of Homeland Security, going through the First Lady. Even George W.Bush took part in one of the sessions, on January 2008 [V]  [V]
figure FnTSM-Politics and Social Media-Political Actors and Political Participation-Fig-5.png
Figure 2.5 First chat session of the many held by The White House during George W. Bush second term.
figure FnTSM-Politics and Social Media-Political Actors and Political Participation-Fig-6.png
Figure 2.6 President Barack Obama participated in a Twitter Town Hall on 2011.
figure FnTSM-Politics and Social Media-Political Actors and Political Participation-Fig-7.png
Figure 2.7 President Barack Obama took part in a Reddit’s Ask Me Anything (AMA) on 2012.
That format is not very different of Q&A sessions run through Twitter, Reddit AMAs (short for “Ask Me Anything”) or Twitter Town Halls [W]  [W] Receiving questions via Twitter and replying them out loud by streaming.. Indeed, Barack Obama has engaged with citizens under all this formats (see Fig. 3.6 and Fig. 3.7). On July 2011 he conducted a Twitter Town Hall during which he answered 24 questions for about one hour; on August 2012, during the electoral campaign, he took part in a Reddit AMA session [X]  [X] for half an hour and addressed 10 questions; and at the moment of this writing he has held three Twitter Q&A’s: on May 2012 he answered 7 questions on assorted topics; on December 2012 he addressed 8 questions about taxes; and on May 2015 12 questions on climate change.
Other politicians, including some holding office, have taken part in similar experiences although none of them have reached so much media coverage as Obama’s actions. From that point of view, this kind of meetings are attractive because they may improve the reputation of the official organizing them; but, given that, one must wonder if they are not just another permanent campaign tactic. On another hand there certainly are risks such as troublesome questions, but we cannot forget that little is left to chance and that these meetings are “choreographed and controlled by staff” (Katz et al., 2013: p. 54). Anyway, despite their attractive, it is dubious that these formats can exert any meaningful impact in policy making (Katz et al., 2013: p. 71) and, from that point of view, they are far from being a success. Engagement with subnational governments or with government agencies

One of the first online systems to allow citizens directly engage with a local government and other citizens was the Public Electronic Network (PEN) established in Santa Monica during the late 1980s (Varley, 1991)—and that has been already described in chapter 1.
PEN was a forum where users were able to discuss topics of broad interest for the community. Purportedly it should have helped to bootstrap new policies and, indeed, there were some success cases. However, PEN suffered from the same shortcomings of any online forum: threads drifting towards off topic issues, flame wars, verbal violence, and conversation dominated by a few vocal users. Moreover, elected officials eventually abandoned PEN and, thus, the conversation held within was simply ignored by the government.
If we compare the situation in the early 1990s with the 2010s we find that many local governments have embraced some kinds of social media but stay away from forum-like services. For instance, Mossberger et al. (2013) found that in 2011 more than 85% of the 75 largest US cities used Facebook and Twitter [Y]  [Y] These findings are similar to those of Oliveira and Welch (2013). They worked on data from a survey conducted in 2012 among officials from 500 cities in the US, including all 184 cities with a population over 100,000. They found that Facebook and Twitter were the most common social media tools: 92% and 78%, respectively., but just 2.7% offered discussion boards, and only 8% held virtual town hall meetings. Interestingly, almost 75% provided a form in their website to send comments, but less than 23% had got blogs for elected officials.
This data can be read as if local governments don’t mind if citizens send comments, but want to avoid those comments to be too easy to exchange among fellow citizens. In other words, vertical communication, even bottom-top communication, is welcome, horizontal communication not so much. If such an interpretation is correct then these cities can be employing social media to increase transparency and to collect feedback [Z]  [Z] Oliveira and Welch (2013) found that while 91% of the cities used social media to disseminate information, only 45% used it to receive feedback on service quality, and 66% to engage with citizens. but they are failing to promote scenarios for crowdsourcing and co-production. Under such circumstances it is also dubious if uncoordinated feedback provided by citizens is of any utility to improve local policies.
There are, however, government departments and agencies which are incorporating social media at a higher pace. In the US this situation has been undoubtedly accelerated by the Open Government Initiative [A]  [A], but similar projects at different stages of development are taking place in other countries. The literature on this topic is extremely abundant so I will just cover a few selected pieces with a focus on the interaction between agencies and citizens in social media.
According to Mergel (2013) government departments and agencies seem to lack any formal guidance when adopting social media; they mostly act by mimicking the behavior of other governmental actors and learn by experimentation. This means that their approaches to social media are extremely varied although there exist three main tactics: push, pull and networking.
Agencies choosing a push approach are using social media as a broadcasting channel, do not interact with citizens nor accept comments or feedback. They fundamentally distrust social media.
Under the pull approach “agencies are actively trying to encourage their audiences to co-create and share content in different formats with them” (Mergel 2013). A good example of this kind of action was the #GlobalSelfie [B]  [B] initiative of NASA:
“On Earth Day 2014, NASA asked people all around the world a simple question – ‘Where are you on Earth Right Now?’
We asked people to answer the question on social media, with a selfie. The goal was to use each picture as a pixel in the creation of a ‘Global Selfie’ – a mosaic image that would look like Earth appeared from space on Earth Day.”
Undoubtedly such campaigns engage people in the co-creation of materials, but they are a far cry from engaging citizens in government decisions of any kind. They may also help to raise awareness about a given topic but, on another hand, they could simply be government-sponsored slacktivism.
Finally, the networking approach is a mixture of push and pull that also accepts unsolicited comments, feedback and provide opportunities for sharing government materials in social media. Mergel argues that this approach can be empowering for users.
Unfortunately the truth is that, up to now, there is no evidence that any single government agency has incorporated feedback by citizens in any meaningful way—provided they are accepting it, which may not be the case. Moreover, the available metrics about social media participation are useless to understand actual public engagement (Smead, 2013).
It must also be noted than when agencies in the US refer to social media they are not generally discussing “classic” social media such as discussion boards, wikis, or listservs, but commercial platforms such as Twitter (157 accounts), Facebook (103 accounts), or YouTube (97 accounts), but also Scribd, Vimeo or Slideshare [C]  [C] Data obtained from Access: June 7th, 2015..
Blogs are the only exception with 28% of a sample of agencies having a blog (Snead, 2013); however, about 40% of the agencies with a blog do not allow comments or do not publish users comments [D]  [D] One prominent blog without comments is that of the White House.. Given that a blog without comments can hardly be considered a social medium, only 17% of the agencies are properly using blogs. Anyway, user participation in blogs is much lower than for the rest of social media (Snead, 2013).
Finally, although the research here reviewed is focused in the US, the situation in other countries is not very different: governmental approach to social media is commonly based on social networking sites, and broadcasting tactics are majority. Other kinds of engagement

In the two prior sections I have mentioned a number of times users feedback, and whether it is or not incorporated in any meaningful way by government. I have also mentioned that feedback is welcome in theory, but actually not all government officials and agencies are encouraging users to provide it, and some even disable it. I have also shown that when officials engage in Q&A sessions, or agencies apply pull tactics they are actually driving the topics and the kind of feedback they are interested in—e.g., Obama’s Q&A on climate change, or NASA’s #Globalselfie.
What I have not discussed is the nature of such feedback which, needless to say, is far from monolithic. Indeed, most of the feedback that users are willing to provide is spontaneous, not necessarily aligned with the topics of interest for the government, and highly unstructured. In this subsection I will cover two important aspects regarding this kind of feedback: the real value of such feedback for the government, and how the different governments have provided tools to try to “discipline” it.
To start with, we must confront the stark reality that for governments not all voices and opinions can be considered equally or, at least, they are not of the same quality.
For instance, Shulman (2009)—who studied mass email campaigns targeted at federal agencies during rule making processes—says that “a case can be made against mass e-mails citing overwhelming evidence of low-quality, redundant, and generally insubstantial commenting by the public.” He argues that the public has been mislead into thinking that quantity of comments will influence the decision and, thus, they flood agencies and politicians with comments. Instead, public information processes aim “to elicit new information and identify issues not already raised”. He finally adds that, by flooding with low quality and duplicated comments the government, the activists are making harder for the high quality comments to be detected [E]  [E] There is a good example showing how massive comments can drive public dialogue although not influencing decisions: the marijuana question during Obama’s Twitter Town Hall. On that meeting he addressed a question about legalizing marijuana which is not among the topics most pressing for public opinion. The thing is that questions were not issued live, but previously crowdsourced—i.e. people voted on the most popular questions—and because of that “the marijuana legalization folks shanghaied the forum and voted their question to the top” (Jeff Howe quoted in Katz et al., 2013: pp. 54-55). Obama eventually answered the question with a joke; so, the morale is that social media can be manipulated by certain groups to force government officials to pay attention to non pressing issues; they, in turn, can simply ignore or ridicule them; meanwhile, pressing issues may very well remain ignored..
Farina et al. (2012) made similar remarks on Web 2.0 mediated participation in rulemaking:
“[N]ot all citizens’ preferences about policy outcomes are created equal. [...] Unlike electoral democracy [...] the legitimacy of rulemaking derives from a formally transparent process of reasoned deliberation. The types of preferences expressed in mass comments may be good enough for electoral democracy, but they are not good enough for rulemaking.”
They also make some harsh critiques about requesting public feedback which is not going to be valued nor incorporated in the decision processes. They qualify that as “political showmanship and not open government.” Indeed, they provide a number of recommendations to implement valid systems to engage citizens in rulemaking. Though, they also state that such systems require first to determine which kinds of participation will be valued, and therefore to understand that the kind of participation which is pervasive in social media may not be the most appropriate one.
It must be noted that although no government has implemented Rulemaking 2.0 systems such as those depicted by Farina et al. they have offered some tools to try to “discipline” the feedback citizens provide. Among the most common we can find e-petitions systems. Such systems allow members of the public to submit petitions to the government; such petitions must be signed and an email address is required. Normally, those e-petitions are open to the public to sign them and, if they reach a predefined number of signatures before a deadline, they are addressed by the government.
One of the earliest examples of such a system is the one launched by the Scottish Parliament on june 1999; other systems are those by the German Bundestag (launched on 2005), the UK Prime Minister (launched on 2006), or “We The People” in the White House (launched on 2011).
Although e-petitions are unsolicited feedback as tweets or emails targeted at politicians or agencies, they tend to be much more structured and better argued than those other kinds of feedback. The reason for that is simple: no one in the government is going to handle a petition unless it has got enough signatures and, thus, the petitioner must be able to persuade a large enough group of citizens about the value of their petition. Those systems allow users to search for petitions or filter by issues, but they also have lists with the most popular (i.e. those with most signatures) and the most recently submitted petitions.
It must be acknowledge that e-petitions are a step forward because they can be used to “put topics on the public agenda” (Jungherr and Jürgens, 2010) and, moreover, they can avoid flooding government staff with thousands of similar emails or tweets.
However, they are still ripe for manipulation and even trolling. A much celebrated example of the later was a petition to “Secure resources and funding, and begin construction of a Death Star by 2016” submitted on November 2012 to the White House. Eventually, the petition reached enough signatures to be addressed [F]  [F] the answer was used not only to argue the unfeasibility of the idea (the pun) but to exhibit a number of actions by NASA, DARPA, and the White House (self promotion).
This example shows that every system can be gamed, and that the government can play an advantage on it for the sake of its reputation. Moreover, there are a number of worrisome facts about e-petitioners and their behavior.
Jungherr and Jürgens (2010) analyzed data collected from the e-petitions system of the German Bundestag and found that a large part of users are ‘activism consumers’; that is, they sign many petitions across many topics. If we discount ‘single issue stakeholders’—users interested in at most two petitions and on a single topic—activism consumers amount for 99% of “active users” of the system. The remaining 1% of them are ‘new lobbyists’ and ‘hit and run activists’; the only difference among them is that while the former use the system for extended periods of time, ‘hit and run activists’ use it for a short time and then they abandon it.
An orthogonal perspective was provided by Lindner and Riehm (2011) who surveyed both traditional petitioners and e-petitioners using the Bundestag system. They found that it was not attracting unengaged or disenfranchised users but, instead, exacerbating the already existing biases.
So, in short, e-petitions systems fail to increase participation, most users are interested in just one single issue and lose interest in quite a short time. The remaining few active users are dominated by ‘activism consumers’ that are willing to sign virtually every kind of petition. Finally, such systems can be gamed and, hence, the most popular e-petitions are not necessarily meaningful or deal with pressing issues for the public opinion.

2.3.3 Other applications of social media by governments

Up to now I have covered social media use by government to reach the citizens, and to help citizens to reach the government; in other words, to increase government transparency and citizens’ participation. Although this may be the most visible and common application, it is not the only one. In this regard, Osimo (2008) describes a number of domains where social media (Web 2.0 in his report) can have a substantial impact on government activities; namely: service provision, law enforcement, regulation, cross-agency collaboration and knowledge management.
Although governmental activities, they cannot be considered politics and, thus, it is out of the scope of this book to cover them. Hence, I highly encourage the interested reader to consult Osimo’s report, and the survey by Magro (2012). Still, I will provide a couple of examples of e-government applications of social media for illustration purposes.
The first one is related to law enforcement which, although is a core government competence, can exploit social media to engage citizens to help on its efforts (Osimo, 2008: p. 36). A good example of this are ‘#tweetredada’ actions organized by Spanish national police corps [G]  [G] At the moment of this writing @policia is, with 1.61 million followers, the most followed law enforcement account on Twitter. A rather uncommon feature is that the account has no followings and, thus, there is a running joke: “The police follows no one”. (@policia). Using Twitter and YouTube, Spanish police encourages citizens to anonymously report on drug traffic and sale. Such cyber raids are quite effective and have resulted in a number of detentions. In addition to those reports, @policia has organized other Twitter raids and is frequently informed about other crimes such as phishing, child pornography, or online harassment. Other examples of social media use by police departments can be found in Heverin and Zach (2010), Crump (2011), or Meijer and Thaens (2013).
Another area where social media has been suggested as potentially useful is inter-agency collaboration. Purportedly, social media can break up the different information silos within government and “support horizontal and vertical information-sharing needs” (Mergel, 2011). In this regard, there are two subfields where such kind of collaboration is essential and where some success stories of social media use are available: disaster response and security.
Social media use for disaster response seems to have promising prospects—particularly when using crowdsourcing tools and mining data from services such as Twitter. Here social media would help government agencies to perform two tasks: (1) to obtain information on the ground, and (2) to coordinate their efforts. The interested reader should consult Goolsby (2010) to learn more about this topic.
Regarding the application of social media to help agencies to share information, Osimo provides as a success case Intellipedia—“a wiki platform managed by the CIA, which enables the direct collaboration between the analysts of the 14 US Intelligence agencies.”

2.4 Public organizations

McNair (2011: pp. 7-8) defined public organizations as “non-party organizations with political objectives” and cited as examples “trade unions, consumers’ associations, NGOs, and corporate lobby groups”. Unlike parties—where ideology is the main common feature of their members—public organizations are defined by a relatively narrow goal which makes beneficial for their members to join together (e.g., workforce issues, environmental concerns, or just the intent to influence policy-makers for the advantage of a given concrete industry). Unlike pressure groups they are usually well-organized, and tend to have access to both policy-makers and mass media. Finally, unlike terrorist organizations they abide to the law and avoid violence. Needless to say, public organizations are extremely varied and very different one from another. Hence, for the sake of brevity I will just provide a few examples on social media use by unions and Non-Governmental Organizations.

2.4.1 Unions

Unions are “[c]ollective organizations of workers whose purpose is to substitute a collective bargain for separate individual bargaining and thereby maintain and improve the standard of living of their members” (Burnham, 2003). Among their means to achieve their ends strike and picketing are “absolutely essential ingredients” (Robertson, 2004: p. 481), but we cannot forget that unions in modern democracies are not only players in the labor arena but also in the political one (Streeck and Hassel, 2003)—usually with ties to some left or center-left party.
As with other organizations, during the late 1990s and early 2000s the prospects of using Internet-based communications (i.e. social media) within unions seemed extremely promising for their goals (Shostak, 1999 and 2002; Lee, 2000; Lucore, 2002; or Barab, 2007). Among the different media unions were starting to use or planning to use we could find websites, email, listservs, chat rooms, bulletin boards, web forums, or blogs. Purportedly, such tools would help unions to engage new members, to allow commonly underrepresented members to take part in virtual meetings, and to improve internal democracy at large. They could also help to better organize strikes and pickets, even they could be realms for “virtual strikes” [H]  [H] Blodgett and Tapia (2011) describe how IBM employees in Italy conducted a strike in Second Life. The strikers reached their goals but, arguably, the success of the strike was due to being covered by traditional media because of their novelty..
In reality, websites, newsletters, and even Web 2.0 tools have been mostly used for broadcasting (Shostak, 2002: pp. 140-141; Ward and Lusoli, 2003; Schradie, 2015); email have been used for outreach purposes and although it is true that massive email campaigns have been orchestrated by unions (Newman, 2005) we have already seen that such actions are not always beneficial; finally, discussion systems (e.g., chat rooms, bulletin boards, and web forums, but also blogs and blog comments) have faced distrust (Ward and Lusoli, 2003) and moderation attempts by unions’ leaders (Barab, 2007), and even shutdowns when the discussion was not considered appropriate (Lee, 2000).
However, the truth is that there are two ways of understanding unionism and, thus, two ways of employing social media for its purposes (Schradie, 2015).
If an organization is focused on ‘lobbyist unionism’ it is very likely that it will use social media mostly for broadcasting information and reaching members and supporters; it can find social media useful to conduct lobbying or to organize pressure campaigns, but classic strikes, much less protests and other kinds of contentious politics [I]  [I] Godard (2011) discusses a number of hypotheses for the decline in strike activity including that employees are using other forms of conflict, or that conflict is now more pervasive, not unique of labor relations and it is “reflected in broader political unrest [...] as workers attribute perceived injustice to the behaviour of national and even global political and economic elites rather than individual employers.” I do not dare to posit which hypothesis is the right one but it is clear that contentious politics are increasingly common and that traditional unions are not at the forefront of current protest movements., are unlikely.
On another hand, a union can be contentious with employers and government and, in that case, one may think that it would exploit social media in similar ways to other social movements. However, some recent research has put that idea into question: Schradie (2015) analyzed the behavior of two unions in North Carolina—one a traditional lobbyist union, the other a contentious one—and found that the contentious union was not exploiting social media. Its members used the most appropriate tool to reach people, and social media (including email) was not always the best option. Moreover, a common concern among the members of that union was that social media tools could “atomize” a grassroot movement; one of them said on this:
“I have problems with folks overrelying so much on technology. Everything is through tweeting and Twittering and that kind of stuff, and for me, that helps really accomplish one of the goals of our class enemy and the people’s enemy: the atomization of folks. It leads to a certain amount of fragmentation. Even though people can quickly see struggles [...] but you still got to have some sense of a coherent development of strategy and strategic thinking and folks being able to at least collaborate with some sense of strategic objectives to maximize the impact of fighting back.”
This problem is not unique of unions, but pervasive in any social movement employing social media tools, thus, I will discuss it in depth in chapter 5. Anyway, without further research, it is difficult to ascertain if the behavior of this particular union is shared by other unions across the world or if it is due to their contentious approach. What it seems clear is that traditional unions are not fully exploiting the interaction features of social media, and they exhibit a number of commonalities with political parties on that behalf.

2.4.2 Non-Governmental Organizations

Robertson (2004: p. 351) provides the following definition for Non-Governmental Organizations:
“NGOs are private bodies [...] operating on a ‘not for profit’ basis to provide wide-ranging benefits for individuals or societies. [...] [P]art of their activity will involve bringing public pressure on governments and international organizations to adopt their preferred policy. They are, however, much more than pressure groups because they take it upon themselves to achieve ends rather than merely try to influence the governmental provision of goods.”
To achieve their goals NGOs need funds—obtained from donors; staff—mainly volunteers; and outreach to persuade the public, and to get the attention of both the media and the government. As expected, the opportunities that social media could bring to NGOs have been praised a number of times (e.g., Johnson, 1999; Spencer, 2002; or Briones et al., 2011). Purportedly, social media would help NGOs to find new donors, recruit volunteers and coordinate them on the field, organize campaigns, reach the media and politicians, and engage in a dialogue with the general public.
In this regard, some reports show that many NGOs representatives claim to be doing that (e.g., Briones et al., 2011; or Obar et al., 2012). However, when analyzing the content they publish and their overall behavior in social media a different picture appears.
To start with, many NGOs representatives consider that information dissemination is the main use of Internet (Kenix, 2008) and, indeed, a number of reports have found broadcasting as the major purpose of websites (Kang and Norton, 2004; or Reber and Kim, 2006), Facebook (Waters et al., 2009; or Saxton and Waters, 2014), and Twitter (Waters and Jamal, 2011; Lovejoy and Saxton, 2012; or Guo and Saxton, 2013).
Moreover, the lack of resources, particularly staff, greatly affects NGOs responsiveness in social media (Spencer, 2002; Kenix, 2008; or Briones et al., 2011). This does not mean that NGOs are not interacting with potential donors, volunteers, media and the public, but it is a very minor activity.
In this regard, Lovejoy and Saxton (2012) found that only 25.8% of tweets by NGOs were community-building tweets and, of those, just a third were replies—i.e. conversations; calls to action were even scarcer (15.6%) with a negligible presence of requests for donations (3.1%), calls for volunteers (0.8%), or lobbying activity (0.6%). Similar results were reported by Guo and Saxton (2013): 19.7% of tweets were community-building ones, and 11.6% calls to action. The situation on Facebook was not very different with almost half of the messages being of informational nature (Saxton and Waters, 2014).
In short, the majority of messages by NGOs in SNS are one-to-many communication; messages to weave a community are substantial but actual conversations are minority—at least on Twitter; finally, important calls to action such as asking for donations, volunteers and lobbying efforts are almost negligible.
Indeed, Guo and Saxton (2013) depict the behavior of NGOs on Twitter (but it likely applies to other SNS) as a pyramid with most of the messages devoted to “reaching out people”, an important part used to “keep the flame alive”, and a minority for “stepping up to action”. All of this casts some doubts on the actual application by NGOs of social media beyond mere top-down diffusion of information.
Finally, we must take into account that some dynamics operating in social media—particularly in SNS—may be especially negative for NGOs in the long term despite their apparent immediate “success”. Most of such dynamics lay on the idea that only what can be easily measured is important and, hence, NGOs can try to “go viral” [J]  [J] One unfortunate example is the Kony 2012 video and campaign by Invisible Children whose purpose was to put pressure on the international community to get Joseph Kony arrested by the end of 2012 (cf. Waldorf, 2012). Arguably, some resolutions were put forward in the US Congress on that regard; however, they were little more than empty words, and no more resolutions on the issue were passed beyond 2012. Moreover, the public lost interest in the cause scarcely a month after the video was released. At the moment of this writing the video has been watched more than 100 million times in YouTube and still receives on the order of 10,000 weekly visits. Meanwhile, Kony is still free and his terrorist group active.. This can eventually increase clicktivism or slacktivism. Such a phenomenon is negative not only for NGOs but for any collective movement and, hence, it is devoted a whole section in chapter 5.

2.5 Conclusions

In this chapter I have reviewed how political parties, the government (i.e. representatives, elected officials, and agencies), unions and NGOs have been using social media to interact with citizens.
Purportedly, social media should have allowed all of them to engage in meaningful dialogue with their potential voters, constituents, members, donors, and volunteers, in addition to the media and the public at large.
Moreover, social media should have increased internal democracy of political parties and unions, allowing rank-and-file members to provide feedback to their elites; in a similar fashion, the government under its different branches should be able to incorporate the opinion and knowledge of the people when making laws and rules.
Finally, social media should help minor and fringe parties to reach the citizens; it could even facilitate the creation of new ones; and public organization could use social media to better organized their advocacy and lobbying campaigns.
In short, social media should have been a powerful tool for both political actors and the citizens interacting with them.
However, we cannot forget the definition of political actor by McNair (2011: p. 5): organizations that aim to influence the decision-making process either directly by attaining power, or indirectly by means of the public opinion. Eventually, political actors need to persuade the public either to influence on the government, or—if in government—to accept the decisions that they have taken.
Dialogue is not needed to persuade the public; what is more, it is undesirable. Hence, political actors tend to broadcast information in a top-down fashion, and social media is used that way most of the time. Meanwhile, worthwhile interactions with simple citizens are rare: if they are focused (e.g., a Twitter Town Hall or a Q&A session) then the number of individuals involved is extremely small; on the contrary, if they are open to anyone (e.g., a request for comments about a new law) then they are massive and mostly disregarded. We must seriously consider if this ‘illusionary interactivity’ (Owen and Davis, 2008) is the best that social media can achieve, and if it is the best our democracies deserve.

3 Public opinion

“Public sentiment is everything.”
Abraham Lincoln
“Public opinion? There is no such thing.
There is only published opinion.”
Winston Churchill,
in Hochhuth, Rolf. Soldiers: An Obituary for Geneva. Grove Press, 1968.

3.1 Introduction

The second edition of the Oxford Concise Dictionary of Politics (McLean, 2003) describes public opinion as follows:
“First used in its obvious, literal sense in eighteenth-century political thought. Began to acquire a more precise meaning (without losing its general one) with the advent of scientific opinion polling in the 1930s. General statements about public opinion, which often turn out to be the opinion of two taxi-drivers and ten consecutive passers-by, should be treated with caution.”
Such a brief definition conveys three related but very different interpretations.
On its most basic and unscientific form, public opinion is equated to an assorted collection of beliefs randomly gathered from the public—what in journalism is commonly known as “man on the street” or “vox pop”. An obvious weaknesses of this interpretation is its unrepresentativeness of the whole population; that and other problems were addressed during the early 20th century giving birth to the polling industry.
Under that second approach, public opinion refers to verbalizations about concrete issues which are important to many people; such verbalizations are produced by many individuals who are aware of others reacting to the same issues, but are not necessarily physically together (Allport, 1937). In other words, the pollster approach to public opinion focuses on situations and societal groups which are well defined and, thus, amenable to quantification.
On that regard, Allport (1937) briefly commented on two theories that he acknowledged could be valid “in the realm of possible abstract truth” but that he claimed are unfeasible for scientifically discovering public opinion. They were the “group-product theory” and the “eulogistic theory” which, simply put, argued that public opinion was some sort of product which emerged by means of discussion within a given group, and that was different from individual opinions and consensus opinion, and, indeed, “superior in character” to both.
This higher interpretation of public opinion is actually the oldest one—dating back late-XVII and early-XVIII centuries (cf. Noelle-Neumann, 1979)—and it has exerted a strong influence in the implementation of liberal democracies. Under this interpretation, public opinion is the evolving outcome of the different interactions and discussions amongst free—and informed—citizens. The realm holding those dynamics is the so-called ‘public sphere’ (Habermas, 1991).
It is rather obvious the impedance mismatch between the pollster approach and the public sphere interpretation. On one hand, opinion pollsters find the public sphere too abstract to be useful. On another hand, advocates of theoretical approaches when exploring public opinion have strongly criticized polls. For instance, Papacharissi (2010: p. 14) argued that because of polls “citizens are not called upon to deliberate, but merely to report agreement or disagreement with certain questions”; Bourdieu (1979) went further in asserting that polls’ goal is not to determine public opinion but, instead, to create “the idea that a unanimous public opinion exists in order to legitimate a policy”.
Be that as it may, both perspectives have been taken when studying the relation of social media with politics. To start with, it is a matter of debate whether social media is or is not host to democratic debate. Moreover, and leaving aside that question, social media contents have been mined to try to distill public opinion in thesense used by pollsters. The first matter is pertinent because social media has become a realm where many of us spend an important part of our lives and, thus, it is crucial to know if it is healthy (or unhealthy) for our political life. The second one is also important because, feasible or not, it is being done.
Thus, this chapter will cover both topics. I start trying to determine if social media fits the requirements of the Habermasian public sphere; then I discuss alternative interpretations to the public sphere; after that I cover two important issues that affect all approaches to public opinion, namely, polarization and agenda setting; finally, I close the chapter discussing the most common approaches to distill public opinion from social media, and their limitations.

3.2 Social media and the public sphere

When discussing public opinion and democracy it is de rigueur to pay a visit to Habermas’ public sphere (Habermas, 1991). Simply put, it is a realm that facilitates the public’s participation and deliberation on key issues of common interest and which, in turn, produces public opinion to both shape policies and hold representatives accountable. Yet, according to Habermas, the liberal public sphere that appeared in the XIX century was transformed during the XX century by private interests and elites, and constructive debate was largely replaced by relatively sterile discussion managed through mass media.
Needless to say, this theory is not free of criticism (cf. Calhoun, 1992); however, despite its stylized idealization of the liberal public sphere [K]  [K] Habermas worked on the bourgeois public sphere which excluded most of the population and that, by current standards, was far from democratic. , and the available arguments against the decadence of the public sphere in the 20th century [L]  [L] The 1960s Civil Rights Movement in the US and second-wave feminism are usually used as counterexamples of the purported decadence of the public sphere in the 20th century., the concept is still powerful and pervasive. Thus, one may wonder if social media can contribute to the public sphere and, in that case, to what extent.
In this regard, Kellner (2000) extended Habermas’ ideas to include cyberspace in the public sphere. However, he warns that
“[C]omputer technologies [should] be used to serve the interests of the people and not corporate elites [...] to inform and enlighten individuals rather than to manipulate them [...] to articulate their own experiences and interests, and to promote democratic debate and diversity, allowing a full range of voices and ideas to become part of the cyberdemocracy of the future.”
Hence, there are some important questions to answer before considering social media part of the public sphere and producing public opinion; namely: (1) Does social media really serve the people? (2) Is it used to inform or to manipulate? (3) Does it promote debate and diversity, or partisanship and polarization? In this section I will shed some light on these questions.

3.2.1Media socialia. Cui bono?

In its early stages social media services were extremely decentralized. BBS, Usenet, mailing lists, web forums, or blogs were deployed in myriad of servers. Some of them were privately owned, while others belonged to public institutions; however, there were so many of them that talking about the ownership of, for instance, Usenet or the blogosphere was simply ludicrous.
All of this has changed and social media has been colonized by market forces (Cammaerts, 2008). BBS and Usenet are virtually defunct; most blogs in the world are hosted in a few platforms (e.g., Blogger, Wordpress, or Tumblr); mailing lists and forums were overcome by private services such as Yahoo! Groups or Google Groups; and finally, most social networking sites are privately owned.
Today it is obvious that social media belongs to the market. Indeed, some social media companies are publicly traded and they have got enormous capitalizations—on the order of tens of billions of dollars. This could be due to a new dot-com bubble but, still, it clearly shows that social media has raised huge expectations of monetization. Given that virtually all of the services are offered for free (or at negligible prices), it seems obvious that social media owners will protect their users only to the extent it does not affect their investors—nor the laws in the different countries where they are operating.
There is nothing wrong [M]  [M] Actually, nothing illegal; it is a matter of debate if some of those companies’ behaviors are ethical (cf. Geiger, 2015). with that business model, of course. However, it would be extremely ingenuous to consider that current services have been developed in the interest of people; particularly, to allow them free and unrestricted political action. I am not denying that such activities are taking place in privately owned social media; however, those were not its goals, quite to the contrary, they are just positive externalities of current social media.
As a matter of fact, modern social media services are products of communicative capitalism (Dean, 2010) and their main objective is to “capture their users in intensive and extensive networks of enjoyment, production, and surveillance”. The actual message, its author or its audience are largely irrelevant, provided they contribute to the global stream of communication and to the profit of the company owning the service.
Because of this, it has been claimed that current social media is actually depoliticizing (Dean, 2005), and it acts as a fetish for some persons feeling they are “doing something” (Žižek 2008: p. 183), or connected with pressing issues, or just venting their frustration. In this sense, it would not be farfetched to qualify social media as the new “opium of the people”.
Even more, social media has been appropriated by different elites (Cammaerts, 2008). Politicians, parties, and news outlets have opened social media profiles, they have engaged (more or less successfully) in social media dynamics, and they are trying to set the political agenda also within social media.
For instance, leaders’ debates are not only televised but live-tweeted by the audience (Diakopoulos and Shamma, 2010): the so-called ‘viewertariat’ (Anstead and O’Loughlin, 2011). Certainly, such viewertariat is engaged in a discussion but not in a reasonable debate of any sort. They engage, at most, in a guided debate, driven by the agenda established by the media and the political actors. This is even clearer when social profiles from elite users engage in those social media “debates” to rebut criticisms from regular citizens (Elmer, 2012).
It is also quite revealing that elite users (politicians, journalists and pundits) have occupied central positions in political social networks, they interact mostly among them, they only rarely engage with regular users (Ausserhofer and Maireder, 2013), they publish contents with an overt agenda-setting aim (Parmelee, 2013), and still their contents are favored by most users over those from non-elite users (Hawthorne et al., 2013).
In this scenario it is rather clear that social media—at least as it is now—does not serve the people’s interests and, thus, it does not fulfil the first of Kellner’s criteria to consider it a public sphere.

3.2.2 Information, misinformation and disinformation in social media

Many users rely on social media to get political information, and for some of them it is their main source (e.g., Mitchell et al., 2015). Thus, it would be crucial that such information was factual and trustworthy for citizens taking informed decisions. If such information was misleading it would weaken even more the chances of social media to constitute a valid public sphere. Hence, there are two related but different questions in this regard: (1) Do users trust information in social media? And (2) is that information actually trustworthy or do manipulation attempts exist?
The answer to the first question is rather nuanced. To start with, global users’ trust has been steadily growing from 2011, and it is now close to that of traditional media. While in 2011 only 8% of users trusted information published in social media (Edelman Trust, 2012), in 2014 45% of them trusted it versus 58% who trusted traditional media (Edelman Trust, 2014).
However, there are stark differences between users in emerging countries and those in developed ones: 58% of the users in emerging countries trust social media versus 65% trusting traditional ones; in developed countries only 26% of the users trust social media and 51% trust traditional outlets.
This means that while users in developed countries would be somewhat sensitive to misleading information in social media, users in emerging countries (where democratic struggles are much more important) would be much more sensitive to that kind of manipulation. This takes us to the second question: can we really trust information in social media?
First of all, as with any user generated content, there exists a substantial amount of false information, not only regarding politics but on any current issue. Indeed, the problem is pressing enough to have sparked a good number of methods to automatically determine the credibility of social posts (e.g., Castillo et al., 2011; Kang et al., 2012; Gupta et al., 2012; Kim and Ahmad, 2013; or Abbasi and Liu, 2013). Unsurprisingly, most of them are supervised techniques and, thus, a prior labelling phase is required for training. From such labelling we have some anecdotal evidence on the prevalence of false information in newsworthy tweets: 8.6% of them are almost certainly false, 31.8% are likely false, 18.6% are uncertain, and just 41% are likely true (Castillo et al., 2013).
Needless to say, the fact that an important amount of tweets (or any other social content) is false would not necessarily imply an underlying intention to manipulate the opinion of the users. It could be that what is spread is misinformation but not disinformation. Rather unfortunately, there is ample evidence of different attempts to spread the latter; particularly with regard to political issues.
For instance, in 2006 the progressive blog instigated a link bombing campaign against Republican incumbents running for congress; the goal was altering search results to bring to the public’s attention negative stories about those politicians (Metaxas and Mustafaraj, 2009). In 2008, a smear campaign against Massachusetts’ Democratic candidate for the senate, Martha Coakley, was orchestrated using Twitter (Mustafaraj and Metaxas, 2010). In addition to smear campaigns, astroturfing in social media is also pretty common, and a number of methods have been developed to detect it early (e.g., Ratkiewicz et al., 2011; or Lee et al. 2011).
As a side note, it is interest to point out that one of those projects (Truthy) was itself subject to a smear campaign from mid to late-2014 [N]  [N] 92 under the false claims of being a governmental tool to restrict free speech of those opposing Obama—see Figure 3.1 for an assorted selection of defaming tweets.
Furthermore, although actual users are eventually drawn into these kinds of campaigns and contribute their own “content”, there is evidence that they are initially arranged using automated and semi-automated methods (Chu et al., 2010), socialbot networks (Boshmaf et al., 2011), and “online persona management services” (Giles, 2011). For instance, in 2012 it was claimed that the campaign of Enrique Peña Nieto, the PRI’s candidate for the Mexican Presidential Elections, was using these tactics for astroturfing (Orcutt, 2012).
Finally, although not aiming to manipulate the available information, spammers are a frequent source of noise in political topics. For instance, one common tactic in Twitter, trend hijacking, consists of producing tweets containing hashtags from trending topics. This kind of contents is neither disinformation nor misinformation, just pollution; however, it affects the signal to noise ratio and can severely disrupt the experience of users trying to follow a topic. Indeed, the 2009 Iranian election protests were one early example of this problem, when the hashtag #iranelection was subject to spam by trend hijacking (e.g., Grier et al., 2012; or Benevenuto et al., 2010).
Hence, given the limited users’ trust in social media, and the problems due to astroturfing, smear campaigns and spam, it seems clear that social media also fail to fulfil the second Kellner’s criterion to consider cyberspace a public sphere: “to inform and enlighten individuals”.
figure FnTSM-Politics and Social Media-Public Opinion and Elections-Fig-1.png
Figure 3.1 On August 2014 University of Indiana’s project Truthy was covered in Fox News in a misleading way. That, in turn, spearheaded a number of misinformed attacks by some Twitter users onto the project and the researchers behind it.

3.2.3 Is democratic deliberation taking place in social media?

Self-expression of political views (Schlozman et al., 2010; or Vitak et al., 2011) is pretty common in social media, with about one third of the users having posted messages regarding their political ideas, or redistributing political contents online (Rainie et al., 2012). However, this self-expression does not produce higher levels of political debate (Schlozman et al., 2010). Moreover, politically-centered discussions in social media tend to be of relatively poor quality.
In this regard it seems appropriate to quote Dahlgren (2005) who said that
“[D]ialogue is preferable to violence, and good dialogue is preferable to poor dialogue, but with the referent of the Habermasian ideal speech situation, demanding criteria are placed on the nature of political discussion”.
Shortly put, proper democratic deliberation assumes citizens are equal participants, opposing points of view are welcome, and the main goal of the discussion it to reach “rationally motivated consensus”. Little of this is common in social media and, indeed, monopolization of the debate by a minority, impoliteness, flame wars, trolls, and harassment have been common since its early days—e.g., Varley, 1991; Schneider, 1996; Buchstein, 1997; Davis, 1999: pp. 157-161; O’Sullivan and Flanagin, 2003; or Hardaker, 2010.
However, all of that is not enough to discard the existence of democratic deliberation in social media. To support such a strong argument, I will focus on two key aspects: exposure to diversity, and debate quality. Regarding the former, I will check whether social media users are prone to interact with users with opposing views, and what is the outcome of such interactions. Regarding the latter, I will review research on the kind of topics covered by users in online political debates, and the way in which they are handled.
A good amount of research about political homophily [O]  [O] The trend of individuals to interact more frequently with like-minded people. and polarization in social media exists, and, given its importance, a whole section is devoted to the topic elsewhere; therefore here, I just provide a brief summary of the matter. To start with, social media is usually depicted as a polarized—or at least fragmented—realm; however, reality is more complex: (1) social media users are, to some extent, exposed to ideas opposing theirs, even if they are partisans; (2) they may engage in discussions with other users with opposing ideas but they prefer to avoid it, especially if they are partisans; (3) when they encounter those opposing ideas they do not actively spread them within their own partisan network of contacts; and (4) political homophily is a strong force shaping the connections in SNS. Hence, although social media users are not totally isolated in echo chambers, they are far from being widely exposed to diverse points of view or engaging with people holding them, and even further of spreading ideas different of their own.
Regarding the quality of online political discussions the available reports are not encouraging. To start with, I have already mentioned a number of issues that hinder quality conversations, such as impoliteness, trolling, flaming and harassment.
In addition, users posting about politics tend to rely on their own authority, and do not usually provide external evidence. This was found by Davis (1999: p. 161) on Usenet but similar behavior has been reported about Twitter users (Park, 2013).
Moreover, amusement is a major driving force for users of commercial social media services and sometimes this precludes from any reasonable deliberation. For instance, Hess (2009) says about YouTube:
“Much of the content found on YouTube is dismissive of serious discussion and an examination of the most popular uses of YouTube indicates that entertainment and play are much more fitting comprehensions of the website. Often, serious discussion is juxtaposed with crude humor and user arguments in the form of flaming.”
This tilt of social media towards humorous contents has made of faux pas and witty remarks a key aspect of political “discussion”. It is somewhat vexing that Big Bird and “horses and bayonets” became the “debate-defining moments” (Freelon and Karpf, 2015) of the first US presidential debate in 2012, and that they “may have influenced public opinion and media coverage”. Indeed, that kind of reactions are common to live-tweeted debates; for instance, during the debate between Angela Merkel and her challenger in 2013 one of the most attention grabbing topics was her necklace! (Trilling, 2014). Still, humorousness does not always implies the trivialization of the topic at hand, as many times it is used as a soft protest device (‘t Hart, 2007).
But even ignoring buffoonery and poorly supported argumentations and harassment and flaming and trolling and impoliteness the fact is that the political agenda in social media is quite different from both the mass media and the public agendas (e.g., Delwiche, 2005; or Ausserhofer and Maireder, 2013). Social media favors discussion on political topics related to technology and media, and topics with shorter cycles tend to overcome long-running topics which are usually the most important ones (e.g., the financial crisis or corruption scandals).
In regard to the quality of the conversation, if we pay attention to the dynamics of social media chatter (such as in Twitter) we find that they are better defined as “blinkered deliberation” (Goodin, 2000):
“[They] are not so much ’talking to one another’ as they are ’posting notices for all to read’. Others might (or might not) take note of them, and reply. But insofar as they reply in similarly public fashions, they too are essentially just posting other notices for all to note (or not), in turn. What we find in the public sphere, in short, is not so much ’public deliberation’ as ’deliberation in public.”
Finally, even in those cases in which actual political discussions are held in social media—i.e., civilized and meaningful ones—not all of them would serve democracy. As Schudson (1997) brightly put it, “[c]onversation provides no magic solution to problems of democracy” because proper democratic conversations must be governed according to norms and pursue an objective. In this regard, conversations about politics on social media “resemble the political discussions that take place among friends in pubs or coffeehouses” (Papacharissi, 2004) and, as of today, no proper way to handle norm-governed debates has been incorporated to social media services at large.
Hence, if we consider all of the above broad issues, namely, the lack of diversity in the interactions due to homophily and partisanship, the low quality of the discussions’ contents, their biases, and the blinkered deliberative style, it seems that social media at large also fails to meet the third and last requirement that Kellmer asked for cyberspace being an online public sphere.
There is some hope however, at least regarding the chance of stimulating democratic deliberation. Price and Cappella (2002) described an extremely interesting exercise where CMC was used to hold democratic discussions. Such discussions took part in small groups (about 15 persons) with a stable membership over time; the sessions were pre-scheduled and moderated, and the moderator was also in charge of “dropping” questions to the group to stimulate discussion.
All of that is certainly possible with most social media services but it does not emerge spontaneously and it requires a prior careful organization. In other words, democratic deliberation must be nurtured in social media; in this regard, moderation tools such as those of Reddit and previously Slashdot could be a good starting point (cf. Poor, 2005) but they should be carefully considered (Wright, 2006). Moreover, another sensitive issue is the selection of the topics to deliberate about (e.g., Katz et al., 2013: pp. 54-55). New social media services should try to incorporate those and other deliberation features “by-design” (e.g., Wright and Street, 2007).
Arguably, promoting social media services for the sake of democratic debate may seem far-fetched; however, it has not only been proposed before (e.g., Iosifidis, 2011) but it has even been tried (e.g., Dahlberg, 2001c; or Jensen, 2003).

3.3 If not deliberation, what is happening on social media?

Given that social media does not fulfil any of the aforementioned criteria for being a public sphere, it is clear that is not currently producing public opinion in Habermasian terms (cf. Dahlberg, 2001a; or Iosifidis, 2011). However, as I have already said, Habermas’ conception of the public sphere is not itself free of criticism and, moreover, there are alternative approaches to democratic debate under which social media discussion may be more appropriately and positively interpreted. Hence, in this section I will cover other theories such as the marketplace of ideas, agonistic pluralism, and non-deliberative discourse.

3.3.1 Social media as a marketplace of ideas

The so-called “marketplace of ideas” is a metaphor commonly used to defend free speech that can also be used to depict public opinion dynamics. It was introduced by Justice Oliver W. Holmes (1919) with these words:
“But when men have realized that time has upset many fighting faiths, they may come to believe even more than they believe the very foundations of their own conduct that the ultimate good desired is better reached by free trade in ideas—that the best test of truth is the power of the thought to get itself accepted in the competition of the market, and that truth is the only ground upon which their wishes safely can be carried out.”
Simply put, citizens produce ideas that compete among one another to be adopted (i.e., consumed) by other citizens; “good” ideas are prefered to “bad” ideas and, eventually, the best ideas prevail in society. Under this approach there is no need for norm-governed rational deliberation among peers [P]  [P] However, the marketplace of ideas does not discard debate and discussion. and, indeed, many ideas can be accepted or rejected without anything close to a conversation.
From this perspective, it seems rather plausible to consider political expression in social media as a marketplace of ideas; indeed, different social media services have been argued to hold such a marketplace: for instance, bulletin board systems (Schlachter, 1993), the blogosphere (Coleman, 2005), or Web 2.0 at large (Tapscott, 2009: pp. 259-263).
If we consider social media as a marketplace for ideas it can be argued that issues such as spam, trolling, flaming and harassment would be “handled” by the market; moreover, the best political ideas would be increasingly common and, somehow, they would eventually permeate to the public opinion (albeit not in a Habermasian sense). However, there are a number of reasons why this can work poorly.
To start with, the marketplace metaphor assumes that all ideas are available for citizens to judge and compare, so they can each individually decide which are the best for them. Unfortunately, that is an inaccurate depiction of social media; not because free speech is being hampered, but because of users’ decisions. I have already discussed homophily and partisanship as dire issues for the public sphere, but they are also a major problem for the marketplace of ideas; indeed, it could be argued that users in social media are not operating in a free market but under different ‘ideological monopolies’.
Moreover, as with any other depiction of markets, the marketplace of ideas assumes that participants are rational which, sometimes, seems a too optimistic assumption (Ingber, 1984). I am not staunchly defending the thesis by Bauerlein (2008) that digital media—including social media—is stupefying, but we must acknowledge that serious political ideas are less “popular” for the marketplace than jokes and memes [Q]  [Q] Remember 2012’s Presidential campaign memes again: Big Bird, “horses and bayonets”, and “binders full of women”.. Hence, as the public sphere, the marketplace of ideas if also negatively affected by the jocularity of social media.
Finally, the idea of the marketplace of ideas has some serious flaws. For instance, Ingber (1984) says:
“[I]deas that support an entrenched power structure or ideology are most likely to gain acceptance within our current market. Conversely, those ideas that threaten such structures or ideologies are largely ignored in the marketplace.”
Gordon (1997) further elaborated the bias caused by power or majorities:
“We do not all enter the marketplace with the same buying power. The market responds to those individuals with the most buying power individually, or to those groups who, collectively, exert power in the marketplace. The goods that we see surviving or prevailing in the marketplace will be those demanded either by the most numerous or the most wealthy buyers, since aggregate demand is simply the sum of individual demand. What are the implications if we stick with our analogy? In the marketplace of ideas, the ideas that survive or prevail will be those espoused either by the most powerful or the most numerous in the society.”
Finally, Ingber also argues that the marketplace of ideas does not aim to find the “best” ideas because that depends on different groups within society; actually
“[T]he marketplace serves as a forum where cultural groups with differing needs, interests, and experiences battle to defend or establish their disparate senses of what is ‘true’ or ‘best.’ Official adoption and support of one group’s position, allegedly due to its success in the marketplace, merely enhances through legal mechanisms the stature of that group’s subculture; it does not represent a universal acceptance of that group’s perspective.”
In short, it may be that social media is a marketplace of political ideas but, in that case (1) polarization and partisanship are preventing many users to access to a large part of the available ideas; (2) the humorousness tendency of social media may give misplaced importance to anecdotal events over pressing issues; (3) the elites may bias the market; and (4) most popular ideas—not necessarily the best ones—will prevail.
Because of this, social media can provide only a skewed version of public opinion and, at most, a version that would not be different from the one depicted by mass media and elites. Moreover, as it was said in the chapter devoted to political participation, minorities and dissidents are not being empowered by social media; they certainly achieve a wider outreach than in mass media or public spaces but they are far from being leveled with elites.

3.3.2 Social media as agonistic pluralism

Habermas’ concept of the public sphere has been criticized from different angles. On one hand, the public sphere is said to be a normative model; that is, it does not explain how democracy or public opinion work, but it is an idealized way in which they could work. Advocates of the public sphere argue that such a model is feasible, and that it has not been fully implemented because of practical issues [R]  [R] Both the public sphere and the marketplace of ideas fail to consider that citizens may not be interested in politics, that civility may not be the norm when discussing, and that popular topics and ideas may not be actually important for a large part of the society, or merely produced for amusement. I have shown evidence that such issues are pervasive and, thus, any normative approach to public opinion is likely to fail when facing reality.. Indeed, the concept is popular when discussing social media—and previously the Web and “cyberspace”—because, purportedly, the available new tools could help to overcome such problems and, thus, achieve that ideal public sphere in practice.
On another hand, it has been argued that by focusing on the bourgeois public sphere, large parts of society are excluded from consideration; namely, women, the working class, and minorities at large. Indeed, by including those very different groups when discussing the public sphere, it is increasingly clear that one Common Good or one Public Interest satisfying all of the groups does not exist and, thus, it is unrealistic to try to find a consensus opinion. This critique is similar to the one made about the marketplace of ideas being in fact a battleground for different groups in society.
Finally, some scholars have argued that the public sphere is not only unfeasible because of practical issues but because of being based on flawed assumptions. According to this approach, conflict in society is unavoidable, and there is no way that deliberation can put an end to it. This so-called ‘agonistic pluralism’ (Mouffe, 1999) captures to some extent both the criticism about the normative nature of the public sphere, and the omission of groups of society with conflictual relations among them. In this regard, Mouffe says:
“[C]ompromises are possible [...] But they should be seen as temporary respites in an ongoing confrontation. Hence, the importance of distinguishing between two types of political relations: one of antagonism between enemies, and one of agonism between adversaries. We could say that the aim of democratic politics is to transform an ‘antagonism’ into an ‘agonism.’ [...] Contrary to the model of ‘deliberative democracy,’ the model of ‘agonistic pluralism’ [...] asserts that the prime task of democratic politics is not to eliminate passions nor to relegate them to the private sphere in order to render rational consensus possible, but to mobilise those passions towards the promotion of democratic designs. Far from jeopardizing democracy, agonistic confrontation is in fact its very condition of existence.”
It must be noted that, although agonistic pluralism seems to be a good depiction of both modern democracies and the nature of political discussion in social media, it also implies that there is no public opinion but, actually, a diversity of public opinions. On another hand, given the polarization of political social media it is difficult to ascertain if we are witnessing agonistic publics interacting, or isolated public spheres with occasional skirmishes among them. Moreover, not every conflictual political behavior in social media should be considered as agonism and a democratic contribution—e.g., flaming, trolling, harassment, and hate speech do not deserve such a consideration. Finally, from this perspective it is also clear that different groups would try to set the agenda to gain allies for their interests which, again, is similar to some criticism made to the marketplace of ideas.

3.3.3 Non-deliberative democratic communication

Most approaches to public opinion and political debate on social media have assumed that deliberation is crucial—neither the marketplace of ideas nor agonistic pluralism discard deliberation; however, there exist other communication practices apart of deliberation that are taking place in online political discussions and that have been neglected in research. From that perspective, it is futile to try to determine whether social media consists of a single deliberative sphere, a set of isolated public spheres, or a set of at times agonistic, at times antagonistic publics sharing a common space.
In this regard the framework proposed by Freelon (2010)—and heavily influenced by Dahlberg (2001b)—is very interesting. He proposes the existence of three different approaches by users to online democratic discourse, namely: the liberal individualist, the communitarian, and the deliberative.
Liberal individualists emphasize individual self-expression and, hence, communication is fundamentally one-way with occasional rebuttals of other users’ opinions. Those adept to the communitarian model focus on the technological features to reinforce or create cohesive communities with clear boundaries. Finally, the deliberative model could somewhat correspond to the kind of debate that Habermas considered a requirement for the public sphere but not necessarily—I have already argued that not every political discussion is a deliberation.
Arguably, the co-existence of those models of communication would produce the different portrayals conveyed by the literature, such as the observed polarization and agonistic publics.
In a later work Freelon (2015) focused his analysis on two different platforms: Twitter and online newspapers comment systems. He found that none of the platforms perfectly fit one single model. However, Twitter (concretely hashtag-mediated discussions) seemed to be specially well-suited for communitarian discourse while comments to online news seemed to facilitate both liberal individualistic and deliberative approaches.
It must be noted that the direction of causality is undetermined and, thus, we cannot assert that each platform attracts a certain type of communicators or that platforms are causing certain varieties of communication. This question is important because if we are interested in promoting political discussion in social media we must pay attention to the way in which different affordances may facilitate certain behaviors while deterring others.
Public opinion in social media is complicated
So, social media does not correspond to the stylized public sphere of Habermas; instead, it holds a large number of mostly independent public spheres engaged in a complex and evolving game of antagonism, agonism and alliances among them. Such a game implies communication of widely varied quality: from blinkered deliberation to the scarce civil and rational debate, going through the more common exchange of poorly informed viewpoints, peppered with incivility.
Given that such communication acts are held in public spaces they are constantly providing ideas to the marketplace to be judged. However, those participating in the market (both active members of the different spheres and non-involved users) do not chose the prevailing ideas on a pure rational basis, but as the result of a complex combination of their personal interests and biases with characteristics of the ideas, such as adherence to the status quo or popularity.
Finally, we must acknowledge that all of those actions—even those considered as communitarian—are individualistic, albeit networked in Wellman’s sense (cf. Rainie and Wellman, 2012); as Papacharissi (2010: p. 21) said:
“Via the affordances of technological environments, individuals fraternize from the privacy of their own spheres, practicing a form of networked yet privé sociality that is formulated within a private social sphere.”
Such individualism may profoundly affect collective actions mediated through social media in at least two ways: Firstly, with regard to their ability to produce large numbers of high quality ideas; secondly, with regard to their ability to market them effectively—i.e., eventually making those ideas prevailing. All of this will be covered in chapter 5.

3.4 Ideological diversity, partisanship and polarization

Although I have partially covered this issue when discussing the public sphere, polarization is much more than a problem when interpreting social media from a theoretical perspective; it has got an impact without regards to the underpinnings of any interpretation and, moreover, it does not only affect social media but public opinion outside social media.
The third edition of the concise Oxford dictionary of politics (McLean and McMillan, 2009) defines polarization as:
“Any general move of political actors from centrist to extreme political positions. Some factors that may lead to polarization include: ethnic or religious violence and counter-violence; political leaders taking up ‘expressive’ positions expressing ideology rather than ‘instrumental’ positions aiming to win the next election; and changes in the electoral system such that it becomes more profitable to woo one’s core supporters than aim for the median voter.”
When we say that social media is polarized we are not only referring to political actors but to common users who adopt partisan stances, and are increasingly more militant and antagonistic towards supporters of other parties or causes. This situation is problematic for a number of reasons: (1) It discourages rational-critical deliberation and, thus, it impairs the public sphere. (2) It disrupts the operation of the marketplace of ideas because citizens do not have access to all the available information. (3) It is a step back in a plural society because antagonism is entrenched instead of cherishing agonism. (4) Users with discourse forms other than liberal individualism are dissuaded to participate, or they may withdraw to those spaces where they are not attacked—i.e., their own partisan sphere; in the former case they are effectively silenced, in the later their debate is inane because they are preaching among the converted. Finally, (5) polarization is self-sustained because each group is dominated by the most vocal—and sometimes extreme— users (Mustafaraj et al., 2011).
It has been argued that polarization in social media has been worsened after information filtering has been applied to automatically curate the timelines of users. Indeed, such a possibility was hypothesized as early as 2001 by DiMaggio et al. but it has been to some extent rebutted quite recently (Bakshy et al., 2015).
The truth is that the underlying cause of polarization in social media is ‘selective exposure’ and homophily—i.e., the tendency of users to consume contents that do not confront their point of view, and interact with like-minded individuals. Such behaviors have been observed way earlier than recommendation algorithms were being used; for instance, in Usenet (e.g., Wilhelm, 1998; or Davis, 1999: ch. 7); chat rooms and message boards (e.g., Wojcieszak and Mutz, 2009 [S]  [S] Wojcieszak and Mutz also found that user may occasionally engage in political discussion with users with opposing ideas in forums that are not presumed to be political. They argued that such kind of casual engagements are hopeful in the face of polarization. I must disagree, despite such discussions among people in disagreement, we cannot rely on serendipity alone to avoid polarization in social media.); political websites (e.g., Sunstein, 2007: pp. 54-55), the political blogosphere (e.g., Adamic et al., 2005; Sunstein, 2007: pp. 145-149; Hargittai et al., 2008; or Elmer et al., 2009); and social networking sites (e.g., Ancu et al., 2009; Himelboim et al., 2013; or Gruzd and Roy, 2014). On top of that, even when users consume an algorithmically curated timeline—e.g., Facebook—their ideological biases are much stronger than those introduced by the filter (Bakshy et al., 2015).
Yet, it must be noted that such studies found that users highly prefer contents and other users fitting their ideology, not that they are avoiding opposing ideas at all cost—e.g., Kobayashi and Ikeda, 2009; or Johnson et al., 2010. Indeed, there exist somewhat optimistic reports which claim that the potential for polarization has been overstated, and that a non-negligible amount of online political discussion is being held among individuals with different opinions. It is also true that none of those researchers claims that such kind of discussion is the most common behaviour and, moreover, many of them relied on surveys conducted among the users who may have depicted a too glossy picture of their behavior.
In this regard we can cite Stromer-Galley (2003) who surveyed users of Usenet, chat rooms and forums; Brundidge and Rice (2009) who focused on chat rooms and email, and also used surveys; Kim (2011) who argued that consumers of political information in SNS are exposed to diverse points of view, even those who are partisans; or Hayes et al. (2015) who reported similar results for Facebook.
When considering both bodies of literature a convoluted scenario is depicted where users do not avoid opposing points of view but, at the same time, they prefer to interact with like-minded people and non-challenging ideological contents. This is even clearer if we pay attention to the behavior of users when facing different points of view. For instance, Valenzuela et al. (2012) found that “discussion agreement was positively related to participation online while disagreement was negatively related with it”; that is, when social media users find opposing views they try to avoid discussion. This has been supported by additional studies (e.g., Rainie and Smith, 2012; or Barberá, 2015), and similar results have been found when analyzing blogs (Elmer et al., 2009), or news-sharing (An et al., 2013).
This setting becomes even more complicated when we incorporate sharing behaviors [T]  [T] Tufekci (2014) describes a number of practices—such as subtweeting, screen captures and hate-linking—that imply interaction among users but are virtually invisible when collecting data. She argues that by performing a standard data collection we would obtain “a polarized map of groups not talking to each other, whereas the reality is a polarized situation in which contentious groups are engaging each other but without the conventional means that make such conversations visible to algorithms and researchers.”. Thus, Conover et al. (2011a) found that mention networks in Twitter (i.e., interactions among users) are not particularly polarized, but retweet networks are. In other words, Twitter users may interact with others holding opposing views but they do not spread those diverse points of view within their own network. This “gatekeeping” role has also been found in the German political Twittersphere (Jürgens et al., 2011).
Furthermore, it seems that online social networks are to some extent woven and unwoven because of homophily (Rainie and Smith, 2012). According to that report, 18% of users of SNS had broken some social connection because of political disagreements, while 16% of the users had established a new social link “because that person shared the user’s political views”. These results seem compatible with those reported by Hayes et al. (2015) who found that about one fifth of users managed their exposure to opposing political ideas by unfriending or muting friends.
So, in short, social media users are not isolated in echo chambers; however, exposure to cross-cutting ideological contents and users is neither wanted nor valued; indeed, exposure is mostly serendipitous, and when challenging contents are found they are seldom discussed or spread to their own enclaves.
Nevertheless, in the same way that deliberation could be nurtured “by design”, it seems that polarity could be handled in an algorithmic fashion. In this regard there is a recent work by Graells-Garrido (2015) showing how to find intermediate topics between two opposing camps. Certainly, there is much work to do on this matter but it seems a promising line of research to fight both ‘echo chambers’ and ‘filter bubbles’.

3.5 Agenda setting

Dearing and Rogers (1996: pp. 1-2) defined agenda-setting as “an ongoing competition among issue proponents to gain the attention of media professionals, the public, and policy elites.” This definition does not clearly state who are the proponents neither what issues are; however, under the most common interpretation, proponents of issues are political actors of every kind—i.e., parties, governments, representatives, officials, public organizations and pressure groups—and issues are “demand[s] that the members of a political system are prepared to deal with as a significant item for discussion through the recognized channels in the system.” (Easton, 1957). It must also be noted that there is not just one agenda but three: those of the media, the public and the politicians, all of them influencing to each other (cf. Dearing and Rogers).
This implies that different actors have got diverse demands from the political system, and they use the means at their hand to first transform their demands into political issues and, then, try to get them satisfied. Some actors may attain power and, thus, they will try to set both the media and public agenda to first persuade the public in order to get votes to reach power; and then, to convince the public that their policies are adequate. Those actors that cannot attain power need to set the public agenda, the media agenda, or both to, eventually, set the policy agenda. Finally, the media and policy agendas are influenced by the public agenda substantiated through public opinion—via opinion polls, letters to the press, public meetings, strikes and demonstrations, and, obviously, through the ballot box (Stanyer, 2005).
Taking this into consideration we could think whether social media constitutes a different agenda or, to the contrary, it is just another facet of the public agenda—i.e., public opinion. Certainly, politicians, mass media outlets and journalist are present in social media but they are a minority, and the immense majority of users are regular citizens. Besides, social media has been commonly approach as a proxy for public opinion both in theoretical (e.g., the public sphere) and practical terms (e.g., a complement or a replacement for polling surveys).
Those evidences lead me to argue that social media should be consider another manifestation of the public agenda; indeed, this would explain the attempts of different elites to appropriate social media (Cammaerts, 2008): they would be pursuing agenda building goals. Yet, setting the public agenda on social media is not just a prerogative of elites, and pressure groups of different kinds are also actively engaged in such processes.
Furthermore, agenda setting processes in social media are crucial no matter which theoretical framework is employed to interpret the public opinion it holds. Thus, understanding how elites are setting the agenda in social media would help us to explain why it is far from being an ideal public sphere nor a perfect egalitarian marketplace of ideas. On another hand, if we consider social media from an agonistic perspective it is even more clear that agenda setting is crucial for each of the different ‘agonic’ groups. Finally, if we just focus in the different discourse styles—namely, liberal individualist, communitarian, and deliberative—they are still agenda setting processes which determine the topics to discuss about, no matter the mode of discussion.
So, in which ways have been different actors trying to set the agenda in, and by means of, social media?
To start with, we must acknowledge that in its original depiction agenda-setting referred to the ability of the press to shape political reality by means of the selection and distribution of topics, and the amount of information devoted to each one (McCombs and Shaw, 1972). However, if the press is able to set the agenda it should be considered a political actor, right? The truth is that this question is somewhat open to debate; while some consider press and mass media as political actors (e.g., McNair, 2011: ch. 5), others are not that sure (e.g., Page, 1996; or Patterson, 1997). Indeed, Page made on this issue the following reflection:
“The concept of ‘political actor,’ [...] implies observable action that is purposive [...] and sufficiently unified so that it makes sense to speak of a single actor. A critical question, therefore, concerns whether—or to what extent—media outlets do in fact use their publications and broadcasts in a purposive and unified fashion to pursue policy objectives.”
Hence, although it is undeniable that virtually all mass media outlets have noticeable biases, and a good amount of them are clearly partisan, it is exceedingly difficult to discuss in this book mass media as a whole, in a similar fashion to that applied to political parties, politicians or pressure groups. Because of that, mass media was not covered in chapter 2. However, the role—if any—of mass media setting the agenda in social media is pertinent and it will be covered here. In addition to that, I will also review how politicians have been using social media to try to set the mass media agenda; and also the bidirectional interaction between social media (particularly blogs) and mass media.

3.5.1 Mass media setting the public agenda in social media

The effect of mass media in the topics discussed in social media has been explored a number of times. For instance, Roberts et al. (2002) analyzed whether mass media set the agenda of bulletin boards. To that end they collected information about four different issues (immigration, health care, taxes and abortion) and they found that discussion in three of them positively correlated with online news, lagging between 1 and 7 days. It must be noted, however, that lag times varied among issues and not all online media were able to set the agenda for all of the issues.
Similar results were reported by Lee et al. (2005) for the first level of agenda-setting, but they also found that discussion observed in bulletin boards influenced mass media at a second level. That is, there seemed to be an interaction between mass media and social media where discussion topics were broadly set by the press, and then they were somewhat qualified and reframed following the reaction of the public as seen in social media. Needless to say, this does not necessarily implied that mass media was directly influenced by social media; it could be that public opinion perceived by the press in other ways was similar to the opinion observed in social media.
Finally, there are two recent and contradictory reports on mass media coverage of politicians and the attention they received on social media. On one hand, Hong and Nadler (2012) found a positive relation between mass media reporting about a politician and they being mentioned on Twitter; according to their report, “a 10% increase in the number of media mentions for a politician is associated with a 4 to 6% increase in the number of Twitter mentions.” On another hand, Murthy and Peto (2015) found no correlation between sentiment about primaries candidates in press and frequency of mentions or sentiment in tweets about them. This topic is highly related with electoral forecasting by means of social media traces and, thus, it will be covered in depth in the next chapter.

3.5.2 Politicians using social media to set the mass media and public agendas

Another kind of actor trying to set the agenda—in this case that of the press—are politicians. However, it must be noted that—except for little-known politicians—they do not consider social media as their main communication channel (Skovsgaard and Van Dalen,2013). Actually, virtually all of them—especially those that are well known—tend to rely on social media as a leverage to get press coverage; however, by using social media they also get the advantage of setting the agenda, particularly on the second level.
This kind of “journalists friendly” approach mainly consists of offering ‘information subsidies’—e.g., press releases, pictures, videos, lists of position on issues—and has been put to practice since politicians started to use personal websites (e.g., Lipinski & Neddenriep, 2004). Still, it has greatly expanded with current social media tools, particularly Twitter (e.g., Broersma and Graham, 2012; Parmelee, 2013; or Hemphill et al., 2013).
Actually, current social media has got a number of advantages over traditional websites; as Broersma and Graham (2012) put it:
“[Social media services] are relatively cheap, engaging, and easy-to-control tools to reach out to voters without the mediation of traditional news media and journalism, while on the other hand they offer possibilities to set the news agenda and get their message included in the mass media virtually untouched by journalists.”
Similarly, Parmelee (2013) commented:
“Tweets [...] have the opportunity to influence the public directly as well as through media coverage. [...] Political tweets inject issues and views into the public domain that journalists may feel the need to cover, either to appear current with their readership or for fear that competing news organizations will scoop them.”
In other words, social media in general and Twitter in particular are being actively used by politicians to try to set both the mass media and the public agendas by choosing the issues and their framing. Whether they are being successful or not is another matter that would require further research.

3.5.3 Social media setting the mass media agenda—and vice versa

Given the hype surrounding the democratizing possibilities offered by cyberspace, the Web and social media, it was a matter of time that some tool would be claimed to provide citizens with the power to directly set the media and policy agendas—beyond other channels, that is. Such an instrument was the blog. Among the advocates of the power of blogs to set the agenda we can find Drezner and Farrell (2004) who said:
“Under specific circumstances—when key weblogs focus on a new or neglected issue—blogs can act as a focal point for the mainstream media and exert formidable agenda-setting power. [...] The blogosphere also acts as a barometer for whether a story would or should receive greater coverage by the mainstream media. The more blogs that discuss a particular issue, the more likely that the blogosphere will set the agenda for future news coverage."
It must be noted that such claims were very early toned down by Murley and Roberts (2005); they argued that political blogs were not influencing the media agenda but, instead, reframing the issues raised by the mass media: “Weblogs [...] often tell their publics how to think about items that the mainstream media say we should think about.” Cornfield et al. (2005) reported very similar findings.
Later, it was found that the relation among blogs and mass media was much more complex and bidirectional; that is, neither mass-media was always influencing blogs nor the opposite. Wallsten (2007a) found that there were issues which when covered by the press effectively received coverage by blogs, but there were also issues that did not have that effect, and even issues which exhibited a negative correlation between mass media and blog coverage. However, he also noted that for a number of issues there was a reinforcement cycle between mass media and blogs.
Arguably, such kind of behavior could suggest inter-media agenda-setting and, thus, blogs—at least those studied by Wallsten—would be perceived by mass media as similar to other news outlets. Indeed, this hypothesis was suggested by Sweetser et al. (2008) to describe the mutual interaction between mass media and candidate blogs during the 2004 US Presidential Election. Similar results about the bidirectional interaction between mass media and blogs were reported by Messner and Garrison (2011), and Neuman et al. (2014).
Moreover, not all blogs are equal and, in fact, most of them followed the agenda set by a minority of them (Haas, 2006); indeed, agenda-setter blogs are just a handful—“the elite of the elite” (Nahon et al., 2011)—and most of them “do not socially construct frames, nor do they set the political agenda.”
On this light, it may be argued that agenda-setters in the blogosphere exhibit more commonalities with mass media outlets that with common blogs. Taking this into account we may wonder if such kind of blogs can set the policy agenda as mass media; in this regard Wallsten (2007b) analyzed the influence of blogs in political elites and found that such influence certainly exists, but it is short and highly dependent on the issue at hand. Hence, although blogs may have some impact on setting the policy agenda it is fairly limited.
So, in short, despite claims of blogs influencing mass media the truth is that only a tiny minority of them have that power, and the rest just follow them and the press; blogs are not able to influence the policy agenda at large, and it seems unlikely that other social media services have such a power. Mass media are very likely to have got a larger agenda-setting power in social media, but it is not general for every conceivable issue. Finally, politicians are actively trying to use social media such as Twitter to set both the public and media agendas.

3.6 Gauging public opinion from social media

As mentioned in the Introduction, ‘public opinion’ describes two related but widely different matters. First, it may refer to the resulting product of collective deliberation about matters of interest for the population at large. Second, it may apply to the findings of surveys concerning a set of topics, and administered to representative samples of the public.
I have already argued that social media does not yet produce the first kind of public opinion. Though, it has been claimed that public opinion under its second meaning can be distilled from the deluge of opinionated messages published in social media. This seems somewhat plausible because, after all, many of those messages fit the definition by Allport (1937) of public opinion: they are verbalizations on concrete issues of importance to many people, which are produced by many individuals who are aware of others reacting to the same issues, but are not physically together.
Furthermore, and as I will show below, there is a growing body of research on this topic with rather promising reports. Indeed, the question has already left the academic realm to enter that of the polling practitioners. In September 2012 the council of the American Association for Public Opinion Research commissioned a report (Murphy et al., 2014) with two main points of interest: the uses of social media as a platform to help in the polling process, and the exploitation of social media as a data source. The report raises more questions than it answers but it also sketches a number of lines of research for the near term.
My goal for this section is similar in spirit to that report and, thus, I will also cover the most relevant works conducted up to date, and the many limitations and challenges affecting them. Furthermore, I also dare to suggest a number of promising lines of research that could eventually help to alleviate some of those limitations.

3.6.1 Common approaches

We can distinguish a number of approaches to the problem of gauging public opinion from social media (Anstead and O’Loughlin, 2012).
The simplest method is the so-called “vox pop” which “illustrates” the public’s opinion using as anecdotal evidence assorted selections of messages (see, for instance, Figure XXX). This is pretty common in journalism although it cannot be considered public opinion in any way.
Quantitative illustration is a slightly more advanced approach. It basically consists of first defining the topic of interest on the basis of keywords, hashtags, and, sometimes, users; and then, collecting different social media metrics for such topic. For instance, we can report how many tweets belong to a hashtag, and it has been rather common to report the number of friends or followers a candidate has got in Facebook, or Twitter, respectively.
Two different refinements have been proposed to that basic method. The first approach relies on applying sentiment analysis methods to the contents (e.g., blog posts, statuses or tweets), and then producing a time series representing the mood of the population. Once data has been collected during enough time, it is possible to detect abnormal or unexpected peaks and, consequently, mine the contents to try to ascertain the topic that has produced such peaks (e.g., Lansdall-Welfare et al., 2012).
Methods following this approach have been applied, for instance, to blogs (Balog et al., 2006; or Demartini et al., 2011) or Usenet (González-Bailón et al., 2012), and they have shown that online public’s mood reacts to mass emergencies, terror attacks, or major political decisions. This approach has also been quite popular to exploit Twitter data as akin to pre-electoral polls in order to forecast elections; however, results in this case have been mixed (Gayo-Avello, 2013).
A serious problem of that kind of approach is that it tries to only employ social media data and, thus, results can be quite inconsistent. That is, neither the peaks location nor their heights always fit the reaction of the public depicted by surveys: in some cases social media seems to overreact, and in others to underreact.
To solve this, it is possible to correlate time series gathered from social media traces with others obtained from actual public opinion polls. Such an approach has been used, for instance, to nowcast consumer confidence (e.g., O’Connor et al., 2010; or Vosen and Schmidt, 2012), presidential approval ratings (O’Connor et al., 2010); or to forecast electoral results by training on pre-electoral polls (Tsakalidis et al., 2015).
Finally, we have so-called ‘semantic polling’, that is, “mining and natural language reading of textual data such as Tweets to draw conclusions about public opinion” (Anstead and O’Loughlin, 2012). This method could be considered the closest to open-ended surveys since it could, at least in theory, determine the public agenda. Such an approach has been used, for instance, to detect political issues discussed in the Russian blogosphere (Koltsova and Koltcov, 2013). Taking into account the abundant prior research in topic detection and tracking on data streams (e.g., AlSumait et al., 2009; or Sayyadi et al., 2009) this seems a fruitful line of research. The main problems are the aforementioned biases in social media towards topics that are humorous, fast pace or tech related. In other words, we risk taking ‘trending topics’ as public’s opinion.

3.6.2 Limitations

The limitations affecting the feasibility of mining public opinion from social media can be classified into four broad groups; namely, demographic biases, self-selection biases, adversarial behaviors, and limitations in opinion mining and big data methods.
Demographic biases are due to the fact that not everybody uses social media, and different groups are over- or underrepresented. Because of that, social media data cannot be considered a random and representative sample of the population and, thus, public opinion derived from such data is always skewed.
Self-selection biases occur with certain individuals or groups decide not to express their opinion; this may be an habitual behavior of some users, it may depend on the political issues under discussion, or it may be caused by the dynamics of the ongoing debate—i.e., vocal users may create the impression onto other users that a minority opinion is majority and, hence, instill a spiral of silence. Anyway, self-selection biases introduce additional skewness on top of demographic biases.
Adversarial behaviors are those where individuals or groups try to manipulate the overall perception of public opinion in social media. They cover a broad diversity of actions, and they commonly include the spreading of misinformation and disinformation.
Finally, even though social media was a representative sample of the population, all users were equally likely of expressing their opinions, and those opinions actually conformed to reality, the truth is that there exist additional limitations. On one hand, political texts are still difficult to handle by automated opinion mining techniques because of irony, sarcasm, double entendres and framing. On another hand, cursory applications of big data methods fail to acknowledge not only demographic and self-selection biases, but they also implicitly make flawed assumptions like equating online and offline behaviors, assuming that such behaviors are persistent along time, or ignoring the biases due to the affordances of the platform whose data is used as input.
All of these limitations are covered in detail in the following subsections. Demographic biases

An obvious problem of gauging opinion from social media is that not everybody uses it, and that the percentage of users widely varies across different demographic groups. For instance, in the United States—which has got a high Internet penetration rate—we find that while 95% of adults younger than 33 use Internet, only 30% of those older than 74 use it. If we pay attention to different social media services the gap is even wider: 83% vs 16% for SNS, 80% vs 20% for watching online videos, or 43% vs 15% for reading blogs (Zickuhr, 2010). Furthermore, use of different social media services is not homogeneous across groups, and the user bases of different platforms present very different mixes of gender, racial, ethnic, socioeconomic status, and even ideological backgrounds (Duggan et al., 2015; and Rainie et al., 2012).
All of this means that data collected from social media services is not representative of the population as a whole. In general, social media users are younger, more urban, and more educated than the population at large. Of course, this demographic bias has been warned about time and time again (e.g., boyd and Crawford, 2012; Hargittai, 2015; Nagler and Tucker, 2015); yet, social media data is still relied on as a kind of panacea, and the size of data is considered redeemer of its underlying biases. Needless to say, that is ingenuous and dangerous.
Fortunately, demographic bias is somewhat solvable: we can always correct the data according to the actual ratio of each demographic strata within the population (e.g., Gayo-Avello, 2001b; or Nagler and Tucker, 2015). Certainly, the amount of users’ demographic metadata varies widely from one service to another—it is close to nonexistent in Usenet, chat rooms or web forums; scarce in some SNS such as Twitter, and extremely rich in others such as Facebook. Yet, if there exists a certain amount of users who provide demographic metadata it is possible (1) to train models to perform content-based user profiling, or (2) to exploit homophilous relations to label other users in their neighborhood. Indeed, it has been repeatedly shown that relevant traits for public opinion analysis, such as gender [U]  [U] Only male and female: Rao et al. (2010), Burger et al. (2011), Gayo-Avello (2011a), Al Zamal et al. (2012), Bergsma et al. (2013)., age [V]  [V] Usually age intervals: Rao et al. (2010), Gayo-Avello (2011a), Al Zamal et al. (2012), Sloan et al. (2015)., political leaning [W]  [W] Rao et al. (2010), Conover et al. (2011b), Gayo-Avello (2011a), Golbeck and Hansen (2011), Pennacchiotti & Popescu (2011a, 2011b), Al Zamal et al. (2012), Barberá (2012, 2015), Boutet et al. (2012), Cohen & Ruths (2013), Volkova et al. (2014), Bond and Messing (2015)., race or ethnicity [X]  [X] Gayo-Avello (2011a), Pennacchiotti & Popescu (2011a, 2011b), Bergsma et al. (2013), religious affiliation [Y]  [Y] Gayo-Avello (2011a)., sexual orientation [Z]  [Z] Ibid., and even social class [A]  [A] Sloan et al. (2015)., can be ascertained with reasonable accuracy for unlabeled social media users. Another feature which is crucial and is not readily available for all users is location; on this regard Jurgens et al. (2015) have written an interesting survey about location inference for Twitter users. Self-selection biases

Another source of bias, even more worrisome than demographics, is self-selection bias which results in users expressing their opinions on some issues but not in others, and with widely different intensity.
Mustafaraj et al. (2011) were one of the first reporting this issue; they found that there are many occasions when most content in social media is the product of small but very vocal groups while a large majority of users remains mostly silent. That would not be a problem provided both groups were similar, but they found that their contents are significantly different and, thus, vocal minorities are able to skew the depiction of public opinion in social media.
Furthermore, vocal users tend to be those with firmer—and more extreme—opinions (Rainie et al., 2012) and they also tend to self-identify as ‘opinion leaders’ (Park, 2013). Those users with a more moderate stance are less likely to express their opinions, and there are even users who although interested in politics “do not let it show in their digital presence” (Gustafsson, 2012).
Besides, such behaviors are constantly changing according to the different political issues: on this regard Mitchell and Hitlin (2013) found that Twitter reaction was sometimes more liberal and others more conservative than results from polling surveys. The most simple explanation for this is that so-called liberal and conservative blocs are far from monolithic and, thus, liberal and conservative individuals react in different manners to different topics and, sometimes, they remain silent on issues even having an opinion.
For obvious reasons, drawing any inference from a sample that is not only unrepresentative and biased, but it is constantly changing is extremely challenging. Yet, there are some initial steps to estimate the size of silent majorities (Venkataraman et al., 2012) and to reduce the impact of self-selection bias in data (Zagheni and Weber, 2015).
Finally, silent majorities are not only problematic because of the the unpredictable biases they introduce when gauging public opinion from social media; they are also a source of concern because they produce a spiral of silence [B]  [B] The spiral of silence introduced by Noelle-Neumann (1974) states that people tend to not voice their opinions if they think they are minority and the opposite, voice them when they think they are majority. (Hampton et al., 2014) which, in turn, may increase polarization within the different groups. Social media is an adversarial scenario

I am using ‘adversarial’ here in a similar sense to its use in the field of Adversarial Information Retrieval; i.e., to describe an environment where (1) some actors try to manipulate the information available within or the dynamics taking place on it to obtain an undeserved benefit; while (2) other actors are aware of the existence of such malicious behaviors and try to detect them and—if possible—undo the harm. Common users are not normally fully aware of the outcome of the manipulations except for the most obvious ones.
Needless to say, abuse in social media does not only affect public opinion mining and, in fact, spam is one of the most common forms of abuse; however, for the purpose of this section we are not interested in every kind of malicious behavior but just on those aimed to manipulate public’s perception, and on their detection. Hence, I will briefly sketch three main lines of research: detection of political manipulation in social media [C]  [C] Virtually all of the research on these topics has been conducted on Twitter. Of course that does not imply that astroturfing, automated accounts and misleading information do not occur in other SNS; it simply confirms that researchers have focused in the most convenient platform while neglecting others., detection of automated activity, and identification of mis- and disinformation.
Concerning the first issue, Metaxas and Mustafaraj have conducted extensive research, particularly in relation to electoral campaigns. They have explored the use of link bombing techniques by bloggers to boost negative results in search engines (2009), and they were the first to observe the use of Twitter to orchestrate smear campaigns—the so-called ‘Twitter bombs’ (2010). Their findings suggest that such behaviors are relatively common; after all “[i]n times of political elections, the stakes are high, and advocates may try to support their cause by active manipulation of social media.” (2012). Their research spearheaded later developments such as Truthy (Ratkiewicz et al., 2011a, 2011b), a system to detect astroturfing campaigns at early stages. Another system similar to Truthy but independently developed was described by Lee et al. (2011).
To be successful, that kind of campaigns requires hundreds of accounts acting together; hence, another fruitful line of research has studied how automated and semi-automated activity is organized and conducted in SNS. Here, I would like to highlight the works by Chu et al. (2010), Boshmaf et al. (2011), and Wang et al. (2012). They have covered, respectively, the detection of bots and semi-automated accounts; the vulnerabilities of SNS to massive infiltration of such kind of accounts; and the inner working of astroturfing systems.
Although initially operated by a limited number of ‘puppeteers’, astroturfing and smear campaigns may eventually get enough traction to draw actual individuals. Moreover, smear campaigns may also start because of a piece of misleading information or a rumor, and spread widely. Thus, a third important line of work is devoted to better understand the spread of misleading information—not necessarily disinformation—and its early detection. Such line of research can be traced back to the work of Castillo et al. (2011), although Qazvinian et al. (2011) also developed a system to detect rumors on Twitter. Interesting recent research on this area has been conducted by Metaxas et al. (2015) and Mocanu et al. (2015) Limitations of opinion mining applied to political texts

Virtually all of the works on public opinion mining in social media automatically exploit user generated contents. Certainly, some have relied on mere counts of specific keywords but the majority applied more complex techniques, and plenty of them one form or another of sentiment analysis, involving in many cases machine learning and natural language processing methods.
Liu (2015: p. 18) provides a structured definition for opinions that has been assumed (either implicitly or explicitly) by virtually every research in the field:
“An opinion is a quadruple (g, s, h, t) where g is the sentiment target, s is the sentiment of the opinion about the target g, h is the opinion holder (the person or organization who holds the opinion), and t is the time when the opinion is expressed.”
Hence, at the core of opinion mining we found the problem of assigning a polarity (usually in the spectrum negative-positive) to a string of free text. A number of methods have been devised up to now, and the reader should consult the books by Pang & Lee (2008), and Liu (2015). Nevertheless, a high level survey of such techniques is needed.
Probably, the most simple approaches are lexicon-based (or dictionary) methods. Simply put, a weighted list of words is required; then unseen texts are scored according to those weights and, hence, the polarity of those texts is determined. Needless to say, producing such a weighted lexicon is not a trivial task and, because of that, most researchers tend to use a limited set of lexicons which are somewhat de facto standards: ANEW (Bradley & Lang, 1999), the lexicon compiled by Wilson et al. (2005), or Sentiwordnet (Esuli & Sebastiani, 2006).
Predefined lexicons, however, exhibit a number of problems, maybe the most obvious one is that polarity is context dependent; as Turney (2002) put it:
“For example, the adjective ‘unpredictable’ may have a negative orientation in an automotive review, in a phrase such as ‘unpredictable steering’, but it could have a positive orientation in a movie review, in a phrase such as ‘unpredictable plot’.”
The implications of this are simple: when applying lexicons not tuned for a given domain (such as political opinion) results are subpar. Actually, Metaxas et al. (2011) found that some common lexicons are not a competitive option for political texts: they are only slightly better than random classifiers, they are unable to handle misinformation or propaganda and, moreover, there is no correlation between predicted sentiment and the political orientation of the users. Similar results have been reported by González-Bailón and Paltoglou (2015).
To solve those issues different approaches have been proposed such as producing context-dependent lexicons, or not using single words but n-grams (usually bigrams). The works by Turney (2002), Turney and Littman (2003), or Taboada et al. (2011) are representative examples. Another extremely common approach is to apply machine learning methods; however, they still struggle with some unique features of political texts.
For instance, Irony is especially common in political opinions; thus, while overtly negative material is easy to detect, statements labeled as positive tend to be to a substantial extent ironic comments that are not positive at all—up to 35% (Carvalho et al., 2009). Furthermore, a large part of political opinions which appear to be neutral in sentiment actually have got a clear polarity implied through their word election (Yu et al., 2008). Indeed, sentiment analysis on its own is not enough to fully analyze political texts and, thus, it is needed to determine the ideological leaning of their authors.
In short, opinion mining methods are routinely used to gauge public opinion from social media data but they face a number of important challenges; the truth is that automated political opinion mining is much more difficult than general opinion mining, and the accuracy of current methods is still fairly limited—about 60% (Bakliwal et al., 2013). General limitations of big data approaches

Virtually all approaches to mine public opinion from social media data rely on big data methods. Such methods are not inherently wrong but many of their underlying assumptions when applied to social media data have been shown to be flawed. This does not mean that we cannot build useful services exploiting such data (cf. Lin, 2015), but we must be extremely cautious when making claims about the society we are purportedly studying. This is particularly true when reporting the alleged overall public opinion on any political issue.
Some of those problematic assumptions have been already discussed such as demographic and participation biases, or adversarial behaviors. However, there are additional issues which are more subtle and that would still hold even if social media data was a representative and trustworthy sample. They are the assumptions that online behavior equates offline behavior, and that such behavior is stable in time (Huberty, 2015).
In the first case we are assuming that users’ activities and opinions in social media mirror their offline counterparts. If social media data was actually representative of the whole population—which is not—and online traces closely followed offline behaviors then we would be able to observe population at painstaking detail. However, there are no reasons to think that what users do and say online is what they do and say offline. Indeed, there are circumstantial evidence that both kinds of behavior are different and, moreover, that differences shift along time in unpredictable ways. As an example, Mitchell and Hitlin (2013) found not only that “[t]he reaction on Twitter to major political events and policy decisions often differs a great deal from public opinion as measured by surveys” but that such a reaction is inconsistent; that is, sometimes Twitter reactions is more conservative and others more liberal.
In the second case we assume that users’ behavior is stable for a long time and, thus, once an automatic method has been developed and fine-tuned it should work indefinitely. Again, it seems a strong assumption to make, and indirect evidence suggests that systems exploiting social media data may degrade in capricious ways. A paramount example of this is how Google Flu Trends has ended up overestimating flu incidence (Lazer et al., 2014); or the UK general election of 2015 [D]  [D] It must be acknowledge that neither polls worked for that election but that just emphasize my point: people’s behaviours are changing and unpredictable. being the umpteenth case of Twitter-based electoral forecasting methods not working very well (cf. Jackson and Thorsen, 2015).
Finally, nowadays a huge part of methods purportedly based on social media data just exploit Twitter data. However, this choice is not due to Twitter being the largest SNS—which it is not—but because of convenience: Twitter data is relatively easy to collect in comparison to data from other sources. Though, by focusing on one single service all of the above mentioned issues—demographic and participation biases, mismatch between online and offline behaviours, changing behaviors along time—are exacerbated.
Furthermore, additional problems are introduced because of the idiosyncrasies of the platform. For instance, given that it is easy—and free—to collect Twitter data in realtime while access to historical data requires a fee, “researchers are much more likely to focus on something in the present or immediate past” (boyd and Crawford, 2012). Another example: most researchers rely on the free Streaming API which is a subsample of the comprehensive (and expensive) Firehose access; however, it has been shown that data from the streaming data is skewed and, thus, not totally reliable (Morstatter et al., 2013; and Wang et al., 2015).
Of course, this does not mean that we should abandon social media data when trying to gauge public opinion, quite to the contrary. However, we must acknowledge the biases and issues pervading such data, check whether the strong assumptions we are making are plausible or not, and if not provide the means to correct the biases as best as we can.
Actually, the solution to many of those problems may lay in finding high quality small data within the flood of biased and poor quality big data. In this regard, I find the work by Diaz et al. (2014) very interesting: they compellingly argue that social media data could be more appropriately treated as a panel—by focusing in a selection of representative users—than as a survey. I think that their work opens a line of research that deserves further exploration.

3.7 Conclusions

In this chapter I have explored the relation between social media and public opinion from two very different angles. First, I considered whether social media is a realm where public opinion as the ‘voice of the people’ on issues of common interest can emerge. To answer that question I reviewed the theory of the public sphere, and checked a number of requisites for social media being such a thing. Shortly put, social media at its current state is not a public sphere because it does not attempt to serve people, misleading information abounds, and no proper deliberation is held within.
Social media neither fits the idealized marketplace of ideas, and it is much better described by agonistic pluralism. Unfortunately, polarization is a worrisome concern and it risks to degrade agonism (which is positive for democracy) into antagonism (which is harmful). Thus, I suggested that, in order to make of social media a realm for democratic debate, features to nurture deliberation should be incorporated by design to new services.
After that, I devoted some time to describe how social media is being used for agenda-setting purposes by politicians and mass media, and I went through the complex bidirectional interaction between social media and mass media in terms of agenda building.
Finally, I concluded the chapter analyzing the use of social media as a data source to ascertain public opinion in a pollster sense. I described common approaches and the main issues of concern, such as demographic and self-selection bias, but also the many problems due to the adversarial nature of social media, or the limitations due to opinion mining and big data approaches. Yet, for each of those challenges a number of promising lines of research were suggested.
To sum up, in both cases, to nurture the emergence of the ’voice of the people’ or to gauge the pulse of society, social media offers an exciting road ahead but it will not be absent of difficulties.

4 Elections

“Online campaigning is about as newsworthy as the use of electricity or modern plumbing in electioneering.”
Steven Clift
“Honesty. That’s the thing in the theater today. Honesty …
and just as soon as I can learn to fake that, I’ll have it made.”
Anonymous actor circa 1962
“The predictive power of Twitter regarding elections has been greatly exaggerated.”
Daniel Gayo-Avello

4.1 Introduction

Robertson (2004) defines ‘election’ in these terms:
“[A] method of choosing among candidates for some post or office, [...] elections have become the only fully respectable method for selecting political leaders and governors throughout the world.”
Elections are fundamental at democracies, and in prior chapters I have covered important topics related to them and their interaction with social media; namely, political participation—which includes voting—and political parties—which provide the candidates running for election. In this chapter I will review the different ways in which social media has been employed during elections, such as their use by electioneers to conduct political campaigns, by citizens to make sense of those campaigns, and finally by public opinion practitioners to forecast electoral results.
It is not my aim, however, to provide an in-depth analysis of those matters, especially when some of them have been already covered at large in the literature. Thus, those readers interested in general online campaigning should consult Bimber and Davis (2003), or Foot and Schneider (2006). Electoral forecasting from social media data—mainly, Twitter—has been surveyed by this author (Gayo-Avello, 2013) and more recently by Jungherr (2015).
Instead, in this chapter I will cover research on social media campaigning; I will pay attention to the use of social media by citizens during campaigns given its pervasiveness (Rainie, 2012); and I will provide a top-level but critic view of social media based electoral forecasting. Before proceeding further, I must warn the reader on all of those topics.
First, social media is not a silver bullet for electioneering. The success of Obama’s campaigns cannot be solely attributed to social media use; moreover, other candidates had used social media before him without succeeding, and many others have tried to replicate his tactics to no avail.
Second, although citizens are certainly using social media to discuss electoral campaigns, they are undoubtedly influenced by the media and political agendas, they tend to skip the most dense issues and, in contrast, they focus in the most frivolous aspects of campaigns.
Finally, electoral forecasting using social media data is surrounded by tremendous hype, it is far away from offering any sound methodology, and naïve methods are periodically introduced as the next “polling-killer”.
This does not mean that social media is unimportant to elections. However, electioneers should forget about benchmarking and, instead, focus in the underlying reasons for electoral success such as the traits of the candidate and their manifesto, the strength of their grassroots and sympathizers support, and their appeal to the overall population. They should also take for granted that only a minority of users would use social media during the campaign in constructive and non frivolous ways, and they should plan in advance how to tackle with that scenario. Finally, electoral forecasters should avoid hype and relying on single case successes, and acknowledge that social media based forecasting may very well be unfeasible.

4.2 Social media in electoral campaigns

At the moment of this writing it seems unavoidable for a candidate to use social media during a political campaign. Purportedly, it can help to organize the campaign, raise funds, boost grassroot support, or persuade undecided voters (e.g., Harfoush, 2009; Jaeger et al., 2010; Cogburn and Espinoza-Vasquez, 2011). Social media can also help the candidate to spread their message—even offline—(Karlsen, 2015), to get feedback from the electorate (e.g., Meeks, 2003), to engage with citizens in fruitful discussion (e.g., Enli and Skogerbø, 2013; Ross and Bürger, 2014), and it may even get a minor but still valuable boost in votes (Gibson and McAllister, 2011). Such kind of arguments are usually seasoned with carefully selected anecdotes such as the huge success of Obama’s fundraising (Jaeger et al., 2010; or Cogburn and Espinoza-Vasquez, 2011), or how Ségolène Royal used forums to develop her manifesto (Lilleker and Malagón, 2010).
That line of reasoning, however, suffers from evident flaws. For instance, it tends to forget the many examples where social media candidates lost against “traditional” ones (e.g., Howard Dean or Ségolène Royal), and it also tends to omit the fact that social media candidates usually neglect their followers and avoid interacting with them (Roper, 1997; Stromer-Galley, 2000; Gibson et al., 2003), thus, mainly using social media for broadcasting (D’Alessio, 2000; Gibson et al., 2003).
More importantly, not all of the practices are readily transferable to every conceivable election since they are mainly taken from campaigns in the United States. This is crucial because raising funds, growing an organization, and running personalistic campaigns are characteristic of the US, but not necessarily of other countries and elections [E]  [E] For a comparison of different electoral systems the interested reader should consult Gueorguieva and Simon (2009). Kluver et al. (2007) compiled an interesting set of use cases comparing Web use across different countries during national elections.. For instance, in most European countries candidates can rely on the party machinery to run the campaign and, thus, bootstrapping an organization is unnecessary, while mobilizing party sympathizers and persuading undecided voters is. Moreover, in elections ruled by proportional representation voting systems running individualized campaigns makes little sense [F]  [F] It must be noted, however, that a trend towards more individualized campaigns is evident in many countries, even in those “with party-based electoral systems [...] campaigns increasingly concentrate on the image of the parties’ top candidates” (Plasser & Plasser, 2002: p. 344). Under such circumstances social media is an additional tool for top candidates, and a crucial one for those which aim at re-nomination or want to achieve better list positions in the future (Karlsen, 2011)., while it is unavoidable under plurality voting systems.
Nevertheless, there are four broad practices in online campaigning that are common across countries, no matter the characteristics of their respective electoral systems and political contexts: informing, involving, connecting, and mobilizing (Foot et al., 2007).
“Informing” comprises all the online actions aim to provide “information about the central figures in each electoral web sphere, the substance of their public discourse, and the election process itself.” This includes candidates’ biographies (or a description of the political actor if it is not an individual), their positions on different issues, information about the campaign and the election, plus additional multimedia materials such as speeches.
“Involving” refers to those actions and tools aiming to help users to interact with the campaigners; this includes calendars about future campaign events, the possibility of subscribing to additional sources of information, joining the organization, volunteering for the campaign, or contributing money.
“Connecting” consists of any feature used by the campaigners to bridge users to other political actors; such bridges can be purely cognitive (e.g., comparing the candidate with their opposer) or hyperlinks.
Finally, “mobilizing” refers to those features aiming to transform users into advocates. This includes materials to distribute either online or offline, tools to encourage or help visitors to publicly express their support for the campaign, or to persuade other people—e.g., sharing a party message with a friend by email.
All of these features were identified before the hey-day of social networking sites; however, it is clear that although SNS may be slightly more convenient to conduct some of those actions, they are not providing anything radically different from what blogs, newsgroups, chat rooms, web forums or mailing lists have always delivered to online campaigns.
Moreover, it is also clear that the abovementioned features are valid to any electoral system. If the campaign takes place under a plurality voting system, information would be focused on the candidate, while if the system is a proportional one, more attention would be paid to the party and its manifesto. On another hand, while in some countries mobilization would be focused on voters in others would also include volunteers to run the campaign.
Therefore, social media is not a “secret sauce” that can improve political campaigns, nor a new tool that can replace prior ones. Instead it is a new medium to inform, involve and mobilize voters, and also a realm where voters debate about the campaign and conduct their own supportive actions. Thus, in the following subsections I will cover how different social media tools have been used by campaigners and citizens during electoral periods from those different perspectives.

4.2.1 As used by candidates and parties

Although involving and mobilizing citizens are two of the major practices during campaigning (both offline and online) the truth is that most campaigns are still focused in information and persuasion. Thus, despite the interactive possibilities of social media usually praised by candidates (Enli and Skogerbø, 2013; Ross and Bürger, 2014), the tools are mainly used for broadcasting purposes (D’Alessio, 2000; Gibson et al., 2003). Such a problem was early noted by Bonchek (1996) who said:
“They [candidates] still see Internet users as an audience and are basing their Internet strategy on the delivery of a «message.» This approach may be effective for television, but not for the Internet. The Internet works best when it is used to create virtual communities in which people interact and relate to each other, sharing information, opinions, and experiences.
Instead of building fancy Web sites around pictures and press releases, the candidates should be building virtual communities around issues and ideas. [...] The priority for any candidate is therefore to get people talking about the issues that matter most to his or her campaign.
[...] The candidates should have their volunteers moderate chat rooms, newsgroups, unofficial Web sites, and mailing lists about key elements in the candidate’s platform and the candidate’s positions on current events. [...] Whether the participants in the forums would agree or disagree with the candidate’s position is unimportant. The purpose is to get voters to engage with the issues and proposals that matter to the candidate and to associate the candidate’s name with this engagement.
[...] The lesson that has yet to be learned by the candidates is that the Internet is not television. Television is about informing and entertaining; the Net is about relating and exploring. The candidate able to make the transition will take his or her place alongside Roosevelt and Kennedy as a political visionary.”
In other words, Bonchek was asking candidates to increase the opportunities for citizens being involved in the campaign, and connecting among them. He also makes an additional strong point about creating communities around the different issues the candidate is championing. All of those ideas were later famously exploited by Barack Obama but he was not the first candidate to apply them; moreover, although such an approach may be necessary to win an election, it is clearly not sufficient—as evidenced by a number of unsuccessful campaigns, particularly, the one by Howard Dean in 2004.
Indeed, the traits depicted by Bonchek have been incrementally incorporated into online political campaigns. In this regard, we may consider three crucial milestones: The first one is the US 2000 Presidential Election which witnessed pervasive attempts towards interactivity and multimedia in campaign websites, and also an increasing online involvement of users through donations and volunteering. The second one is the US 2004 Presidential Election with the use of campaign blogs to nurture advocate communities, and initial attempts of on-the-ground organization using meetups. The third milestone is the 2008 presidential campaign of Barack Obama; there, his experience during the 2004 senatorial campaign proved unvaluable to manage on the ground community organization using social media tools.
As of today, different candidates running for different elections have tried to apply the lessons from Obama’s campaigns. Needless to say, it is not an easy task: First, it requires a candidate and a message able to inspire a large part of the population, encourage them to get mobilized and in turn persuade others, and then provide them with a suite of tools that empowers them and, at the same time, gives the campaigners top-down control of all of the actions. The US 2000 Presidential Elections: Increasing interactivity

Lewicki and Ziaukas (2000) described the use of the Web towards the US Presidential elections of 2000 by Republican primary candidates (John McCain and George W. Bush), the Democratic nominee Al Gore, plus alternative candidates Pat Buchanan and Ralph Nader. All of them provided plenty of up to date information and they offered different ways for users to get involved in the campaign and, in turn, mobilize the vote.
Actually, we can discern some interesting advances in all of those campaign websites. For instance, Buchanan and Ralph Nader offered the chance to share some stories by email with other users. The Gore-Lieberman’s website included the so-called “Instant MessageNet”, where users could introduce their AOL, MSN or ICQ handle to then interact with other nearby supporters interested in the same issues (Lewicki and Ziaukas, 2000; Davis et al., 2009); in addition to that, it provided tools to create advocate websites, and even a mobile application for PDAs. McCain’s website offered polls, a mailing list and, in contrast to George W. Bush website, it was clearly aimed at involving and mobilizing support. Actually, McCain’s online campaign managers said:
“We had to use the Internet. We didn’t have a choice. We looked at the Internet as a national staging ground for grassroots activities ... We looked at the Internet as a way to mobilize in those states where we couldn’t afford to put up a bricks and mortar building” (Lewicki and Ziaukas, 2000)
Despite such efforts into mobilizing online support, McCain lost the primaries to George W. Bush. The eventual Bush-Cheney website was much more sophisticated, and it included abundant multimedia material (both audio and video). However, except for a newsletter, it did not offer any chance for voters to interact with the campaign or the candidates.
Still, the experimentation conducted by the different candidates during the 2000 primaries and presidential campaign set the path for future online campaigns; particularly, regarding the involving and mobilization of supporters. In this regard, Lewicki and Ziaukas (2000) argued:
“The successful Presidential Internet candidate of 2004 or 2008 [...] will raise money like McCain, develop warrior tools like Gore, have clean and efficient design like Nader, organize shovelware like Buchanan, have tightly-written, regularly-delivered e-mail like Bush when things get tough, develop tribing-up tools like Gore, and have the e-buttling efficiency of Bush. [And] She or he will report and respond to news on the campaign site.” The US 2004 Presidential Election: Blogs everywhere

Certainly, all of the candidates running for the Democratic Primaries of 2004 applied the lessons learned in the prior electoral cycle, and they incorporated some new features, particularly, blogs. In this regard, Howard Dean is usually considered the pioneer (Meeks, 2003; Davis et al., 2009), although the rest of candidates added a blog to their campaign websites shortly after him.
It must be noted that although labeled as “blogs” they departed a great way from blogger orthodoxy. For instance, the candidates were virtually absent from them and, instead, campaign staffers were the most common contributors. Besides, some candidates websites were providing a space for supporters to create their so-called blogs (e.g. Clark) but it was closer to a forum with separate channels managed by different users.
In addition to blogs, most of the candidates were using other social media paraphernalia such as online chats with campaign staffers (Kerry), online forums (Dean and Kerry), Meetups (Clark, Dean and Kerry), and newsletters (all of the candidates).
Finally, Howard Dean’s website included something that can only be described as a social networking site: DeanLink (see Figure XX). Supporters had got a profile that included their picture, name, location, a link to their homepage, list of personal interests, list of political issues, date of membership, number of supporters recruited, other supporters linked to them, and “badges” granted by other supporters—e.g., “hard working volunteer”, “great organizer” or “enthusiastic Dean supporter”.
Needless to say, the rest of candidates were also nurturing their grassroots movements and encouraging the creation of groups around different issues. For instance, Kerry’s website linked to different “communities” (e.g., African Americans, Students, LGBT, Latinos, etc.); Clark had got an extensive network of grassroots teams, each of them with its own website and linked from the campaign website; and Edwards relied on a volunteer-run grassroots website which also provided a forum.
All of these features earned a number of benefits for both candidates and advocates: they allowed the candidates to keep “in touch with their public and getting feedback”, while they helped “to build community and give the bloggers a sense of belonging” (Meeks, 2003). Moreover, growing up the support base in a peer-to-peer way proved better than top-down approaches (Davis, 2005). Such a way of nurturing communities was also praised by Lessig (2003):
“[W]hen done right, as the Howard Dean campaign apparently is doing, the blog is a tool for building community. The trick is to turn the audience into the speaker. A well-structured blog inspires both reading and writing. And by getting the audience to type, candidates get the audience committed. Engagement replaces reception, which in turn leads to real space action. The life of the Dean campaign on the Internet is not really life on the Internet. It’s the activity in real space that the Internet inspires. [...] Trippi [Dean’s campaign manager] let control of the blogs go and thus was born the first open source presidential campaign. The Dean campaign engages hundreds of blogs without policing who says what when, or who is on-message how much of the time.”
Although highly inspiring, such “open source” approach to content production was not always beneficial. For instance, Meeks (2003) says that posts in Dean’s blog “lack of substance”, and she criticized the opacity about who the bloggers were and how they were granted to post; meanwhile, Kerbel and Bloom (2005) noted that “[a]lmost half of the discussions [in Dean’s blog] were self-references invoking the online community.”
Despite such efforts, Dean’s withdrew from the race. Though, the movement he started has survived to these days and, hence, his case offers one clear lesson: communities must be decentralized to thrive (Lessig, 2003; Kerbel and Bloom, 2005). At the same time Dean’s approach raises two important questions: Is it possible to replicate that kind of experience? Is it advisable for a candidate to run such a campaign?
Regarding the first question, it does not seem an easy task. Neither the blog communities of the other Democratic primaries candidates, nor the eventual Bush-Cheney blog were able to spur the same communitarian feelings. Kerbel and Bloom (2005) suggested two reasons for that: First, the candidate must have a “charismatic appeal” for the voters; then, the campaigners must let some of their control go, and allow their supporters to run part of the campaign on their own. The first factor is difficult to find, and the second one is a tough decision to take because it can backfire at moments.
With regard to the second question, Howard Dean had little to lose when he decided to run a decentralized campaign (Kerbel & Bloom, 2005) but other candidates may find that style of campaigning too uncontrollable. Still, we cannot attribute Dean’s failure to his heavy use of social media but to other reasons; for instance, concerns among Democratic supporters about his electability if eventually nominated, the fact that most primary voters were much more moderate than Dean’s advocates (Hindman, 2005; McSweeney, 2007), or the fact that his advocates, although enthusiastic, were just a minor part of the primaries electorate (Davis et al., 2009).
When compared to the Dean’s website those of major presidential candidates were much less exciting—particularly because they were exceedingly similar (see Figure XXa XXb). Still, there were a number of interesting features, particularly in the Bush-Cheney campaign site.
Both websites hosted a blog but the one by Bush-Cheney was much easier to browse through different topics; Kerry-Edwards provided a forum for supporters to chat among them, but Bush-Cheney offered live chats with campaign staffers and major supporters [G]  [G] The Bush-Cheney campaign chat was identical to the “Ask the White House” chat described in chapter 2 (cf. “Engagement with top level officials”). (Vaccari, 2008; Davis et al., 2009); both offered personalized subsites for different communities although Kerry-Edwards had a much greater variety; finally, both campaigns paid attention to volunteers, but Bush-Cheney website was much more vivid and introduced “gamification” features such as a leader board or profiles of featured volunteers (Vaccari, 2008).
So, in short; the 2004 Democratic primaries and Presidential campaign witnessed the adoption of an important number of social media features such as blogs, forums, chat rooms, social networking sites, user generated content, and even attempts to gamification. Blogs were particularly praised because of their purported capacity to nurture a community of supporters around a candidate; however, not all candidate blogs exhibited such a capacity. Still, virtually all of the candidates thoroughly applied the lessons learned in 2000 (Lewicki and Ziaukas, 2000) and, in that regard, all of the campaigns shared a number of social media features.
figure FnTSM-Politics and Social Media-Public Opinion and Elections-Fig-2.png
Figure 4.1 A screenshot of DeanLink, a social networking site for Howard Dean’s supporters. It provided user profiles, links among users, endorsements and a certain amount of gamification by displaying the number of members recruited by each supporter.
figure FnTSM-Politics and Social Media-Public Opinion and Elections-Fig-3a.png figure FnTSM-Politics and Social Media-Public Opinion and Elections-Fig-3b.png
Figure 4.2 The campaign websites for the Bush-Cheney and Kerry-Edwards tickets. Both are strikingly similar, including the usage of newsletter, blog, volunteer centers, etc. Still, the Bush-Cheney website included additional features such chats with campaign staff or individualizing concrete volunteers in the front-page. The US 2008 Presidential Election: My Barack Obama

Shortly after winning the 2008 elections, Barack Obama was qualified as “the first Internet president”, and it was claimed that his campaign “utilized the Internet and information technology unlike any previous political campaign” and that it would change “how politicians and the public interact” (Greengard, 2009).
Along with the idea of social media fueling Obama’s campaign was that John McCain lagged behind in Internet. That is totally unfair, especially taking into account that in 2000 he was considered an “Internet candidate” because of his important online fund raises (Lewicki and Ziaukas, 2000) and the use of sophisticated features such as web casts (Xenos and Foot, 2008).
So, without belittling Obama’s campaign of 2008, the truth is that it fine tuned methods used in a prior and equally successful campaign that has, however, received much lesser attention: Obama’s 2004 campaign for the Senate (Cogburn and Espinoza-Vasquez, 2011). Without analyzing his senatorial race—contextualized within the 2004 US Presidential Election—we cannot fully understand Obama’s campaigns of 2008 and 2012. Anyway, it was not social media which made Obama won, but the impressive social support he enjoyed—although it is true that it was efficiently managed by means of social media (Talbot, 2005; Vaccari, 2010).
As a matter of fact, many of the tactics applied during the 2008 and 2012 Obama’s campaigns were already present in the final version of his 2004 campaign site: the splash screen to recruit volunteers (Fig. XXa and XXb), the blog, or the tools to organize the grassroots movement (in 2004 they were called “Barack Brigades”). However, those tools were not there from the beginning and they neither were a novelty introduced by Obama’s staff. Indeed, tracing the evolution of Obama’s campaign for the US Senate provides very interesting information.
Indeed, the 2003 version of Obama’s campaign website was basically “brochureware” (see Figure XX). It is true that it provided a newsletter, but they were common since the 2000 elections. Though, the first interactive feature was added quite early: credit card donations that had already been exploited by McCain in 2000; during the rest of 2003 the website introduced a call for volunteers—a form identical to that of other candidates; meetups; and the first calls to action targeted at particular communities—women, asian americans, youth and students—although they only provided a contact email and were common among candidates running for the US Presidential Election.
figure FnTSM-Politics and Social Media-Public Opinion and Elections-Fig-5.png
Figure 4.3 Obama’s campaign website in 2003; it was standard brochureware, none of the features commonly associated with Obama’s use of social media are yet present.
figure FnTSM-Politics and Social Media-Public Opinion and Elections-Fig-4a.png figure FnTSM-Politics and Social Media-Public Opinion and Elections-Fig-4b.png
Figure 4.4 Splash screens used in Obama’s websites for the Senatorial and Presidential campaigns. They differ in aesthetics but functionality is the same in both cases.
During 2004 meetups were intensively used to organize events and to get volunteers to help on each of them—in a similar way to that of Dean. Around mid-2004 a major revamp was conducted introducing the splash screen to recruit volunteers, the Barack Brigades, and the blog—just after they started being used in the US 2004 Presidential campaign.
A fortunate coincidence allows us to discern some features that were a true novelty of Obama’s senatorial campaign. The last version of the website was powered by which was being used by other candidates such as Byron Dorgan (Senate) and Betty McCollum (House). That software provided campaigners with a content management system, and tools to raise funds and register volunteers. In addition to that Obama introduced the blog— although that was taken from the Presidential campaign—and a totally different way of managing volunteers.
Actually, it was his way of organizing the grassroots that made the difference; certainly, we can find similarities between that and Dean’s organization but the truth is that (1) Obama’s campaign was much more efficient, (2) they did their best to translate online support to offline support, and (3) online was far from being their main goal. Indeed, Obama already had plenty of ground support before running for Senate [H]  [H] See for instance Foulkes (2003) who describes how ACORN actively campaigned on the ground for Obama: “Good door-knocking by community residents for even two half days can impact turnout. Good door-knocking by paid and supervised canvassers for two weeks can have dramatic impact. And a combination of the two, especially with experienced community leaders working with the paid canvassers, can make a huge difference.” and he used online tools to empower the campaigners and, in turn, mobilize the supporters.
Given the success of the 2004 campaign it made perfectly sense to apply the same methods in 2008 and again in 2012. Needless to say, the stages—particularly in 2008—were much higher but also were the funds. So, the website was much more professional, virtually every social media platform was targeted—from Facebook or Twitter to AsianAve, BlackPlanet, or MiGente; a SNS was developed: MyBarackObama [I]  [I] That feature is certainly comparable to DeanLink although it is on a different order of magnitude (Talbot, 2008). As Dean’s manager Joe Trippi said “They [Obama campaign] were Apollo 11, and we were the Wright Brothers.”; and the campaign also provided a mobile app [J]  [J] A novelty originally introduced by the Gore-Lieberman campaign in 2000. (Cogburn and Espinoza-Vasquez, 2011). Still, on the ground actions were at the core: volunteering, making phone calls, and knocking doors.
In short, from 2003 to 2008 Obama’s staff perfected a number of tools—that had been originally introduced by other candidates—to do what they did the best: organizing communities on the ground (Dreier, 2008; Ganz, 2009). Probably that is the main lesson when comparing Dean’s and Obama’s campaigns: the candidate must empower volunteers but a top-down control of the whole organization is also needed (Vaccari, 2010).
With regard to the McCain-Palin campaign website it was, unsurprisingly, very similar to that of Obama-Biden. It also had a blog, issue specific groups, a volunteer HQ which included some gamification aspects, and also a SNS of its own—McCainSpace—which was inferior to MyBO (Talbot, 2008). In general, and despite of their attempts to improve the socialization and mobilization features, Obama’s website made easier for his supporters to take action, and his campaign was able to earn much more supporters in social media platforms (Pew Research Center, 2008).
Of course, the eventual result of the 2008 elections was not due to the poor user experience in McCain’s website nor to his smaller support in social media (Cogburn and Espinoza-Vasquez, 2011); as Vaccari (2010) points out:
“The effectiveness of Internet applications depends on contextual factors, such as the personality and message of the candidate and the ability to elicit a strong grassroots response from a large enough portion of the electorate. When these forces are already in place, the new media can efficiently channel such energies toward electoral and political goals; by contrast, when these preconditions are absent, the impact of online tools can be expected to be marginal. One implication of these findings is that a one-size-fits-all approach to e-campaigning by political actors is generally unwarranted. Although in 2008 most candidates adopted a similar set of online tools, the outcomes of their efforts depended on contextual and organizational factors.” Congressional elections in the US

Although they are the most commonly analyzed, Presidential Elections are not the only one benefiting from social media use. Actually, a few researchers have explored how congressional candidates have approached social media as a campaigning tool. From that literature it seems, that Obama’s run for the US Senate was the most interesting one and the rest of candidates have been extremely cautious and traditional.
This is remarkable because one notable difference between presidential candidates and those running for Congress is that the later are expected to respond to their constituents much more frequently and about much more concrete issues. Therefore, interaction with their constituency should be of utter importance for them (see Chapter 2 on this regard), and they could start those good practices before reaching office; i.e., during their campaign.
In spite of this, it has been argued that tools such as “message boards, forums, and live chats [...] allow for less control over the flow of information, require strong logistical capabilities, and may, in fact, be more technologically interesting than politically useful” (Druckman et al., 2007).
Gulati and Williams (2007) reported similar findings, and argued that features to allow content coproduction, and interaction among users and with campaigners were not being highly used during the 2006 campaign for two main reasons: lack of control over the message and lack of resources to manage such tools. They also found that despite the hype surrounding blogs and SNS they were hardly used by candidates: about 40% of 2006 Senate candidates had got a campaign blog or a Facebook profile, and the ratio among House candidates was even lower.
They also argued that SNS cannot substitute a campaign website but they may provide some benefits for the candidate (even when not interacting with supporters): they offer exposure to demographics that are not usually interested in politics, they may be a pool to recruit volunteers, they can serve to build relationships in the long term and, more importantly, they expose supporters to each other and, thus, they can help to foster a community around the candidate.
Still, Gulati and Williams are cautious because they argue that other prior novelties such as blogs or meetups did “not lived up to expectations for their election role”. Sweetser and Lariscy (2008) disagree on this, and they consider that candidates neglecting a “free way to build a mailing list of constituents seems almost uncaring”, particularly when most of those constituents are young voters who are not only difficult to reach but just to find.
Similar results were reported by Lassen and Brown (2011) regarding Twitter use in the US Congress. While they assumed that members facing re-election in 2008 would adopt Twitter to interact with their voters, the truth is that its usage “took off sharply in early 2009 as the 111th Congress convened”; one might even think that many of them started using Twitter just because of Obama.
Indeed, Lassen and Brown mentioned one reason for Congress members usage of Twitter: “the electoral benefits of appearing instep with their constituents not only politically and ideologically, but culturally and technologically as well.” In other words, Twitter in this case—although it can apply to any other social medium— is of interest for candidates not due to its purported usefulness to interact with voters, but because of the electoral credit that it can provide when used as a gadget for posturing. The United Kingdom

As with many other campaigning tools and tactics, social media use has spread from the United States to campaigns in other countries. In this point I will cover some research regarding its application in the United Kingdom while the following one will focus on other countries. The survey will be brief because no real novelties have been introduced, and most candidates are just experimenting with different tools, maybe trying to replicate Obama’s success, maybe just trying to keep pace with their electorate.
During the 2001 General Election candidates faced the same overall criticism that US congressional candidates before: their lack of interactivity and abuse of broadcasting (Ward and Gibson, 2003). Although email, bulletin boards and chat rooms were well known at that moment, they were seldom used: only 8% of the campaign websites provided any kind of interactivity, and only 3 candidates [K]  [K] It must be noted that 44 parties and 3284 candidates run for those elections. engaged in online chats with voters (Ward and Gibson, 2003). In addition to the already mentioned need of candidates of fully controlling their message, Ward and Gibson mentioned two additional reasons for their paucity regarding interaction: lack of public interest and, surprisingly, the scrutiny of the media and political adversaries. They said: “anything controversial or off-message may well be picked up and used against the candidate” and, thus, most of them opted for reduce interaction and provide bland information [L]  [L] A similar cautious approach to Twitter by UK candidates due to journalistic pressure has been argued by Graham et al. (2014)..
Ferguson (2005) described how the 2005 General Election was eagerly anticipated, mainly because of the US 2004 Presidential Elections. However, and despite the hype it had little to offer in terms of online campaigning. Blogs were used by 65 candidates (less than 2% of the total) but most of them did not allow users to post comments, and many were not even written by the candidate. Ferguson labeled such “blogs” as “the ‘latest and safest’ media-friendly tool used by those who wanted to give the impression of being comfortable with getting upclose-and-personal with the public.” Similarly harsh remarks were made by Jackson (2006), this time about party blogs:
“[B]logs were essentially used as one-way communication channels which added colour to party web sites. As a result, such weblogs may have encouraged visitors to return because of some form of voyeurism, but they were not either effective conversational, campaigning, or promotional tools.”
Rather surprisingly, one party, the UK Independence Party (right-wing populist), decided not to offer a blog precisely to avoid receiving feedback that they assumed was going to be negative (Jackson, 2006). Long live democracy!
Still, the 2005 General Election eventually had got an interesting aftereffect from a social media perspective. The day after the election, the Conservative leader announced he was to step down, and in October that year David Cameron was elected as new party leader. One year later, on September 2006, ‘webcameron’ was released (Woodward, 2006).
Initially, the site was supposed to be centered around a video blog which would cast Cameron addressing disaffected voters; however, it very quickly added new features such as blog written by Cameron himself, a guest blog, and an open blog where any registered user could post. Users could vote posts from the open blog, and Cameron was to respond the top 5 posts—later the top 3. Moreover, all posts were open for comments but they were not replied by the staff.
Such use of blogs is certainly laudable but, unfortunate, it lasted too little: on May 2007 Cameron stopped responding user inquiries, and shortly after that, both the guest blog and the open blog disappeared. His blog was renamed into a diary that was open to comments but which were not replied. Later, a web forum was opened for users to debate with each other. From there some issues were selected for Cameron to respond on his diary (every two weeks approximately); still, the web forum was also short lived and it was closed by late 2007 (Chadwick, 2008).
On early 2008, ‘webcameron’ disappeared as a website of its own, and only the videos were preserved in the Conservative Party website—although it was no more a video blog. Apparently, the last ‘webcameron’ video was produced for the General Election of 2010, and in 2013, in a surprising move, the Conservative Party removed a decade of speeches including most of ‘webcameron’ (Ramesh and Hern, 2013).
Given the relatively short life of ‘webcameron’ (little more than a year) and the number of changes it suffered, it was clearly a tool for experimentation. Taken into account the background in the US some features were not really novel (e.g., the videos and forums, even the blog) while others could seem audacious such as responding to questions raised by the users. However, although far from common, that was not a novelty in the UK; indeed, during the 2001 General Election some candidates responded to cherry picked questions received by email (Ward and Gibson, 2003), and Tony Blair opened his Campaign Diary by replying to ten questions presumably sent by voters (Jackson, 2006).
Taking into account the eventual fate of ‘webcameron’ it seems that the Conservative Party did not greatly value the experience. It is unlikely that it was a matter of resources as in minor parties and, instead, it could be attributed to two main reasons. On one hand, the feedback obtained from the open blog first, and from the forum later, could be too meager and, at the same time, too costly in terms of reputation. Indeed, the Conservative forums were at occasions surrounded by controversy and criticism from the Labour party (e.g., Williams, 2007). On another hand, the project was qualified as hollow and insincere (Chadwick, 2008), and, more importantly, lacking any intention of engaging in a real dialogue with the citizenship or connecting with other actors beyond the party. It was referred as “a one-way diatribe of not-quite policies” (Bell, 2006).
In comparison with the struggles of ‘webcameron’ the 2010 General Election was much less interesting. As expected, commercial SNS such as Facebook or Twitter were used by a substantial amount of candidates (Baxter et al., 2011; Graham et al., 2013). However, the majority of the candidates’ messages were broadcasting (Baxter et al., 2011) and they were not usually focused on policies but on campaign and partisan issues (Graham et al., 2013). Candidates used Twitter as a tactical tool, mainly aimed at setting the media agenda or just getting mass media attention; and when interacting with regular users it was not to discuss politics but as social grooming.
It was also unsurprising that virtually all of the parties released action centers similar to MyBO (Lilleker and Jackson, 2010). The Labour party had Membersnet, the Conservative party myconservatives, and the Liberal Democrats offered LibDemAct. It must be noted, however, that although the sites were modeled after Obama’s SNS some of them expanded pre-existing party intranets predating Obama’s campaign. Indeed, Labour’s Membersnet was released on early 2007 (as Labour Members Net - MpURL) and it already allowed sympathizers to create their own blogs, interact with others in forums, and look for campaigns to volunteer.
Rather shockingly, the party that most faithfully tried to replicate Obama’s online tactics was the British National Party (far-right). It offered a full-fledge SNS which was fairly active, all of the party messages were open for comments, and abundant multimedia material—particularly video—was available. Still, it seems that the party at large does not greatly value online campaigning; after all, the Activists’ & Organisers’ Handbook released by the BNP devotes just one page to Internet and outpours distrust on both the medium and its users. Moreover, the campaign website was a one-person effort and a few days before the elections the webmaster closed it (Taylor and Muir, 2010).
In all likelihood, this may be the most powerful image illustrating the vacuity of trusting social media as a campaigning panacea: a far-right homophobic racist party used the same tools as Obama and they still were not able to win the hearts and minds of Britons. But it’s no different anywhere…

I have devoted much attention to the cases of the US and the UK for two main reasons. First, most of the electioneering applications of social media were originally developed in the US. Second, the United Kingdom is, after the US, one of the countries whose elections have been more thoroughly covered in the literature. Still, the use of social media for political campaigning is very similar in other countries; candidates and parties try to incorporate the latest tools while, at the same time, tightly controlling the message.
Indeed, the adoption of innovations closely follows the trace set by the US. Thus, brochureware was commonplace in campaigns during the early 2000s (e.g., Schweitzer, 2005). Blogs were rather enthusiastically adopted after 2004—e.g., the Danish general election of 2005 (Klastrup and Pedersen, 2006), or the German federal election of 2005 (Albrecht et al., 2007)—but they were launched just before the campaigns, comments were not allowed or not replied when allowed, and most campaign blogs were abandoned after the elections. Finally, SNS such as Facebook or Twitter are used but, again, mainly for broadcasting with scarce interactions with common users and, in those cases, mostly with party members or sympathizers (e.g., in Spain—Aragón et al., 2013; in Switzerland—Klinger, 2013; or in the Netherlands—Graham et al., 2014).
What politicians seem to value the most are the chances that social media offers for inexpensive broadcasting. Actually, Gulati and Williams (2010) found that congressional candidates find YouTube much more attractive than Facebook, and Klotz (2010) argue that it is being used by “established political participants” to broadcast traditional advertisements.
Using social media to look “authentic”
We cannot close the topic of social media use by campaigning politicians without briefly mentioning the issue of authenticity. Lilleker (2006: pp. 39-41) defined it as:
“The perception of political actors as being ‘real people’, intrinsically a part of the community they represent, rather than being detached and part of an elite.”
Simply put, voters demand of candidates to be like them and, moreover, they are the voters who settle that matter, not the politicians. Nevertheless, although “difficult to construct by the actor” it is not impossible to craft authenticity; however, it “must be believed to be real and not fake”. In this regard, appearing in non-political contexts or showing emotions are common tactics. As of today, social media is also playing a role in the raid for authenticity.
For instance, some politicians has assumed that by writing a blog, or using Twitter or Facebook they may look like regular folk (e.g., Weiss, 2003; Ferguson, 2005; Lawson-Borders and Kirk, 2005; Klastrup and Pedersen, 2006; Baxter et al., 2011; or Ross and Bürger, 2014). Also, the abovementioned webcameron site clearly fell under such kind of attempts but it was spotted as staged, manufactured and, thus, unauthentic (Bell, 2006; Chadwick, 2008).
To avoid such pitfalls Gilpin et al. (2010) have proposed a framework to evaluate the authenticity of any digitally mediated interaction with the public; unfortunately, it is out of the scope of this chapter to fully discuss it. Enough to say that it considers four dimensions—namely, identity, authority, transparency and engagement—and it has been applied to analyze the social media presence of four congressional candidates running for the 2012 elections (Grow and Ward, 2013). From that analysis Grow and Ward made a number of recommendations to increase a candidate’s “authenticity”: e.g., providing a complete social profile, using all of the available features in the platform, sharing information about hobbies and interests (including pictures), and actively interacting with constituents both online and offline.
Such recommendations are in line with best practices outlined by Facebook or Twitter for government and politics (e.g., Twitter, 2014: pp. 56-59) and reveal that, as in many other aspects, social media has been normalized: After all, it has become just another tool in spin doctors’ bags.
Still, one of the most powerful aspects of authenticity—namely, replying to those users addressing a candidate for a conversation [M]  [M] In this regard, Utz (2009) conducted an extremely interesting experiment in Hyves (an extinct Dutch SNS similar to Facebook). She created a profile for a candidate and replied to users comments. Those candidates engaging with potential voters are more positively evaluated than non-repliers, and there are two situations where the positive evaluation is much stronger: right-wing politicians and left-wing voters. Utz argued that left-wing voters value responsive politicians much more than right-wing voters do, and that right-wing candidates benefit more from interaction because, traditionally, they are much more prone to ignore social media users. It must be noted, however, that comments in Facebook-like platforms are much better suited for a dialogue than tweets—again, platform affordances matter.—is extremely difficult to spin without being noted in the long run.

4.2.2 As used by citizens

Although I started discussing the use of social media during elections by candidates, the key actor is the electorate. Looking for information to decide their vote is an increasingly common application of social media but it is not the only one. Indeed, by using social media, voters are able to play much more active roles such as expressing their opinions on the candidates and the campaign, interacting with candidates and also with other voters. However, some words of caution are needed with regard to the findings reported about these activities.
First, all of the caveats about automated content analysis from social media described in the previous chapter apply to the research surveyed in this one. For instance, although users produce plenty of contents during a political campaign, not all of them can be considered as grassroots campaigning; indeed, recent research has found that what could be cursory interpreted as negative campaigning by users was actually mere discontent [N]  [N] It must also be noted that social media comments during campaign are, generally, much more negative than coverage in mass media (Rosenstiel et al., 2012a). (Hosch-Dayican et al., 2014).
Moreover, most of this research has been conducted on Twitter data and this can certainly bias the findings. To start with, trying to obtain a coherent picture from just one social media platform is extremely problematic: Rosenstiel and Jurkowitz (2011) shown that “political discussion on Twitter is measurably different than the one found in the blogosphere”, and Rosenstiel et al. (2012b) revealed that Twitter, Facebook and blogs “offered a different sense of the candidates”. Moreover, mentions of users are normally taken at face value as interactions, even as dialogues; however, as Bruns and Highfield (2013) have shown, users tend to use politicians’ handles not to engage them in a conversation but just to refer to them.
In short, during political campaigns there is probably much less dialogue that it may have been reported, what has been assumed to be user campaigns of some sort may simply be individuals coalescing because of an external event or a trending topic, and any conclusion obtained from a single platform must be faced with healthy skepticism.
Future research should focus on finer grained details (both in terms of contents and behavior) before reaching conclusions about engagement between users, or between users and politicians; after all, if social media is not a proper public sphere out of electoral periods, much less during campaigns when the level of noise is much higher. Besides, in the we should aim to collect data from all sources at hand and not only from the most convenient ones; otherwise, we will not be reporting the campaign but just a highly skewed vision of it.
That said, let us see how users have relied on social media—mainly Twitter but also Facebook—to be informed about political campaigns, express their opinions, and interact among them and with candidates. To keep informed and decide their vote

United States is, without a doubt, the country where Web and social media use during elections has been studied more thoroughly and the longer. There we can paint a broad picture covering the last 15 years. For other countries we have more fragmented data but, still, we can use it to sketch how citizens are using social media during elections.
To start with, social media is mainly used as a source of information and just a minority of users express their opinions or interact with others. For instance, in the US the percentage of Internet users getting news about the campaign online grew up from 33% in 2000 (Kohut and Rainie, 2000), to 52% in 2004 (Rainie et al., 2005), and 60% in 2008 (Smith, 2009). During that period, those discussing politics or expressing their political opinion online were 8% in 2000 (Kohut and Rainie, 2000), 9.5% in 2004 (Rainie et al., 2005), and almost 20% in 2008 (Smith, 2009).
The situation in other countries is not better: In the UK in 2005, 28% of Internet users got online news about the elections and only 1.5% discussed politics (Ward and Lusoli, 2005). In France in 2007, only 8% of the population considered Internet its main source of political news, and just 7% of Internet users expressed their opinions online (Vaccari, 2008).
Moreover, the majority of users do not consider that online information or discussion have got any sensible influence on their voting decisions. In US in 2000, 43% of Internet users made such a claim (Kohut and Rainie, 2000); in UK in 2005, 17.8% of users said that they had made a better informed choice due to online activity, and almost 13% said that it had helped them to make up their mind (Ward and Lusoli, 2005).
More recent studies have disaggregated social networking use from other activities and, except for the younger cohorts, results are similar to those of prior campaigns. In the US in 2010, just 13% of users of SNS expressed their political opinions and only 7% followed or friended a candidate or a political group (Smith, 2011). In 2012, the percentage of SNS users expressing online grew up to 34%, and 20% followed or friended politicians (Rainie et al., 2012). When looking at registered voters following politicians the trend is also positive—from 6% in 2010 to 16% in 2014—but it is still a minor behavior (Smith, 2014).
There are no information about the influence on voting decisions that US users attribute to social media but in other countries it is well below that of mass media. For instance, in Finland in 2011 only 4% of users considered that SNS had got any influence, versus 30% claiming that TV had impacted their decision, and 27% stating the same for newspapers (Strandberg, 2013). There are two recent reports from the UK conducted in relation to the 2015 General Election. According to Geary (2015) 47% of Twitter users aged 18-34 had made up their mind because of information they had consumed in the site. Comparable results were reported by Duffy (2015): 34% of citizens aged 18-24 considered social media had influenced their vote; however, the percentage of adults claiming the same is just 13%, while newspapers had purportedly influenced 20% of adults, and leadership debates 40% of them. Still, the trend is positive because only 2% of adults considered social media had got any influence on their decisions in the 2010 elections.
So, in short. Social media users in Western countries are mainly consumers and not producers of political material during electoral campaigns, and in both cases such actions are not majority. Moreover, mass media is still much more influential than social media and, anyway, I will show that social media contents produced during campaigns are broadly set by the media agenda. To express their opinion

Although users expressing online during campaigns are a minority they provide a real time glimpse into the voters’ opinion—with all the caveats due to the unavoidable biases. After analyzing those contents for a number of campaigns in different countries [O]  [O] The research surveyed in this subsection was conducted in Australia (Bruns et al., 2011), Germany (Trilling, 2015; Jungherr, 2015), Italy (Marchetti and Ceccobelli, 2015), Norway (Kalsnes et al., 2014), Spain (Aragón et al., 2013), Sweden (Larsson and Moe, 2012), the UK (Stanyer, 2006), and the US (Shamma et al., 2009; Diakopoulos and Shamma, 2010; Mustafaraj et al., 2011; Hong and Nadler, 2012; Bekafigo and McBride, 2013; Freelon and Karpf, 2015). a set of commonalities appear.
First, campaign events, leaders debates and the mass media have a strong influence on the topics discussed in the blogosphere (Stanyer, 2006) and Twitter [P]  [P] Shamma et al., 2009; Diakopoulos and Shamma, 2010; Bruns et al., 2011; Larsson and Moe, 2012; Hong and Nadler, 2012; Aragón et al., 2013; Bekafigo and McBride, 2013; Kalsnes et al., 2014; Freelon and Karpf, 2015; Trilling, 2015; Jungherr, 2015; and Marchetti and Ceccobelli, 2015. However, user generated contents in social media do not simply mirror the political or media agendas (Bruns et al., 2011; Jungherr, 2015). Instead, social media users pay little attention to some topics while they overemphasize others, particularly if they are closely aligned with their interests, such as Internet and technology (Bruns et al. ,2011).
In addition to that, social media users tend to discuss more about campaign tactics and strategies, and tangential issues of leaders debates than about substantial political issues (Bruns et al., 2011; Kalsnes et al., 2014; Trilling, 2015; Jungherr, 2015).
Moreover, humor amounts for a substantial volume of the contents users produce (Freelon and Karpf, 2015; Trilling, 2015; Jungherr, 2015). Such humorous framing of campaign moments—especially during leaders debates—can exert a strong influence in the subsequent coverage in mainstream media, particularly if the framing is of memetic nature. Given that online political humor tend to have a negative effect on candidates’ image (Baumgartner, 2007) and that humor is one important tool in the protest repertoire (‘t Hart, 2007), this phenomenon cannot be lightly dismissed.
Except for humorous memes, contents produced by social media users seem to have little impact on the media agenda. Though, it is not uncommon for mass media to closely monitor trending topics to quickly detect shifts in public opinion; hence, from time to time, relatively small numbers of politically active social media users can exert an impact on mass media (Marchetti and Ceccobelli, 2015).
Finally, another common trait of political social media contents if that they are produced by a vocal minority (Mustafaraj et al., 2011). Usually, they are elite users such as top bloggers and journalists, but also strong partisans or highly engaged individuals (Larsson and Moe, 2012; Bekafigo and McBride; 2013). Still, those users are not the only key actors, and other non-political elite users—e.g., celebrities, sport people, show business stars, etc.—can play an important role in framing certain campaign events or in setting the focus on particular issues, given their massive followership (Freelon and Karpf, 2015). To interact with candidates and with other citizens

The interactive behavior of users during political campaigns does not greatly differ from political interactions described in prior chapter; still, the number of users discussing politics obviously increases during elections.
As it has been discussed in previous section, politicians very rarely interact with users but, anyway, some users address them—although not necessarily in dialogic manners. For instance, Sweetser and Lariscy (2008) reported that young users write in candidates’ Facebook walls in friendly terms, usually with positive and supportive messages even when candidates do not reply them. In Twitter, candidates are frequently mentioned but it seems that users are not trying to engage them in a conversation (Bruns and Highfield, 2013) and, indeed, the number of mentions received by candidates is not correlated with candidates’ activity on Twitter but, instead, with the press coverage they receive (Hong and Nadler, 2012).
Facebook candidates’ walls offer an opportunity for supporters to engage among them in discussion, sometimes vehemently (Sweetser and Lariscy, 2008). However, given the way in which Facebook data is usually crawled by researchers, it is difficult to say how pervasive it is such a practice. In Twitter, where data is usually collected by choosing keywords or hashtags, it has been found that only a tiny minority of users actually engage in dialogues: for instance, during a campaign in Sweden just 7% of tweets were replies (Larsson and Moe, 2012).
Moreover, and as was discussed in prior chapter, users prefer to engage with like-minded users; however, it is not uncommon during campaigns that supporters of one candidate or issue “invade” the realm of other group. Depending on the affordances of the platform, such actions may be more or less intrusive.
For instance, Meeks (2003) described how supporters of Howard Dean trolled Kerry’s blog with harassing messages (Miga, 2003) and, thus, Kerry supporters call for doing the same in Dean’s forums. “Hashjacking” is a similar behavior that takes place on Twitter; it consists of co-opting hashtags commonly used by political adversaries to proselytize and agitate (Bode et al., 2015).
Although anecdotal, these examples reveal that interaction among opposing factions during campaigns are very likely antagonistic and not agonistic and, thus, users campaigning does not seem to try to persuade undecided voters—who may be observing the unfolding discussion—by the strength of their arguments but, unfortunately, by their sheer volume.

4.3 Social media and electoral forecasting [Q]  [Q] Please not that this section may sound slightly patronizing, I beg you to forgive me for that and, please, consider that this topic is my particular groundhog day where I am frequently dragged when wearing my referee hat.

In prior chapter I described the different sources of bias affecting social media to ascertain public opinion and, thus, a reader not familiarized with the topic might think that strong claims about its predictive power regarding elections are uncommon. Quite to the contrary, the area is extremely popular, and there is abundant literature exploring the feasibility of using social media data to forecast elections. A vast majority of such papers have used Twitter data but, still, a few have explored other data sources such as Facebook (Williams and Gulati, 2008; Lappas et al., 2010; Giglietto, 2012), web searches (Lui et al., 2011; Granka, 2013; Yasseri and Bright, 2014), Flickr (Jin et al., 2010) and Wikipedia (Yasseri and Bright, 2014).
As usual, Twitter data is used for the sake of convenience and despite its unrepresentativeness of the voting population at large. Moreover, many approaches have relied on extremely simplistic metrics such as the raw number of tweets, likes, followers or friends a candidate accumulates. Needless to say, there are many reasons for eventual winners achieving high values on such metrics (Nagler and Tucker, 2015) and, as with many other political scenarios, social media can be certainly influenced by external events but still does not closely mirror them.
Anyway, and no matter the challenges, the question is pertinent and the literature should have provided an answer so far. Thus, can Twitter forecast elections or not?
The most earnest answer to that is quite amusing: “yes, no, maybe”; indeed, and for the sake of precision [R]  [R] Data used to reach such an “inconclusion” can be found in Table XX., it should be: “yes” according to 42% of the researchers, “no” according to 18%, and “maybe” according to the remaining 40%. Given that only 18% of the authors have claimed a resounding “no”, and “yes” slightly surpasses “maybe”, we could simply agree on something like “research shows that Twitter-based electoral forecasts are outstandingly promising”. The only problem is that research does not work that way; negative results are not noise or outliers, and in a number of cases they are quite strong rebuttals of prior “promising” results.
So, simply put, after 5 years, almost 40 papers, and the work of more than a hundred researchers from 54 institutions in 16 different countries we, don’t, know. Such a situation is unacceptable and there is something extremely wrong in the way we, as a community, are approaching this particular topic. If it only affected electoral forecasting it could be a relatively minor problem; however, I am afraid it reflects major issues in other areas where STEM researchers are approaching social sciences—as, in this case, politics.
Therefore, instead of surveying the literature on the topic I have decided to offer a reflection on how we have reached this situation. First, I provide some context for the problem of Twitter-based electoral forecasting, in particular, its origins and the way it has been developed so far. I then provide some high-level conclusions from the literature, and the underlying reasons to explain the contradictory reports. Finally, I describe some research that, although promising, failed rather dramatically during the recent 2015 UK General Election.
I hope this case study can be useful to understand the challenges affecting not just this particular problem, but any other area of confluence between social media research and political science.

4.3.1 The state of the question

The first work describing an actual attempt to forecast an election is the work by Tumasjan et al. (2010). That team analyzed the 2009 German federal elections and concluded that “the mere number of messages mentioning a party reflects the election result”. In other words, one just needs to collect tweets about candidates and parties for a number of days or weeks, count them, and, voilà, one has forecasted an election. The idea is astonishingly simple and, hence, the paper is extremely popular and has received, at the moment of this writing, 903 citations.
Not as famous but close with 804 citations is the work by O’Connor et al. (2010) which was published in the same venue that the work of Tumasjan et al. (ICWSM 2010). In that paper, time series showing the evolution of sentiment in Twitter are successfully correlated with some indices (such as the Presidential Job Approval) and unsuccessfully with others (such as pre-electoral polls for the 2008 US Presidential elections). That’s curious! Two papers in the same venue achieving contradictory conclusions.
Much less known, with 108 citations, is the work by Jungherr et al. (2011) where the method by Tumasjan et al. is rebutted. Jungherr et al. shown that if tweets for all of the parties running for election had been counted (instead of just those with seats in the parliament) the victory of the Pirate Party would have been predicted. This can seem a minor issue for readers in the U.S. but not taking into account new parties running for election is a big mistake as colleagues from Europe can tell you [S]  [S] Think for instance of Movimento 5 Stelle in Italy or Podemos in Spain (see Chapter 2)..
Those three papers constituted the foundation of a new subfield, and the thorough analysis and criticism from Jungherr et al. to the work by Tumasjan et al. helped to raise the bar of some (but not all) of the research conducted later: claims about electoral forecasts should not be simply supported on some numbers which are close enough to the result of the elections.
After those three papers, plenty of research has been published but, as I have already said, I will not survey it. The interested reader should consult the works by Gayo-Avello (2013) and Jungherr (2015). Instead, I want to provide a 30,000 foot view so if you are unfamiliar you can still gain an accurate impression of the field, if not of the technical details.
To that end, I will provide some hard facts summarizing the current state of the question. To obtain them I conducted an exhaustive literature research to find any kind of refereed manuscript dealing in all or in part with forecasting elections, or voting intention and pre-electoral polls. For each of those manuscripts a comprehensive set of features was obtained:
? Metadata such as the year of publication, the venue where it was published, or the kind of venue (e.g. conference, workshop, journal, etc.)
? The affiliations of the authors (both institution and country).
? The field of science for each author of each paper (e.g. Computer Sciences and Technology, Political Science, Sociology, etc.). This data was used to determine if the team authoring each paper was multidisciplinary (at least one author from STEM fields and another one from Social Sciences) or non multidisciplinary (STEM only, or Social Sciences only).
? Whether the paper cites literature about “traditional” electoral forecasting by political scientists and not only foundational work in Twitter-based electoral prediction.
? Whether the paper explicitly mentions future work to be conducted or not.
? Whether the team authoring the paper has published later work on electoral forecasting studying a different election.
? The election or elections studied in the paper, including the year when they took place, the countries involved, and the presumed language of the tweets collected for the study.
? Whether they collected their own data or used third party data, and the method to collect it in the former case (e.g. Search API, Streaming API, Firehose, etc.)
? Whether the authors mention a third party technique to make the forecast or devised one of their own.
? Whether they use external data (such as pre-electoral polls or census data) or they rely solely on Twitter data.
? And, of course, their response to The Question: Yes, No, Maybe.
In addition to the features of interest a set of criteria were established to conduct the collection and the codification:
1. Surveys were omitted; papers correlating Twitter data with non electoral polls were not considered, and I also excluded papers studying the predictability of elections at large working on aggregated data (e.g., DiGrazia et al., 2013; or Spierings and Jacobs, 2013).
2. Duplicated papers or extended versions of conference papers submitted to journals were not considered further work. To be considered as separated studies, papers should deal with a different election.
3. The possible answers for “The Question” were Yes, No, and Maybe. It must be noted that I did not judge the plausibility of the results reported in the papers; that is, any strong claim of accurate forecasts, or outperformance of pre-electoral polls accounts for “Yes”. Similarly, any strong statement about inaccurate or failed predictions, or underperformance when compared to polls was labeled as “No”. “Maybe” was used for those papers that either reported mixed results, or described their results as “promising” but asked for caution with regard to future elections. I also used this label for papers with contradictory claims: for instance, authors stating that their method achieved results comparable to those of polls, and providing, at the same time, figures that clearly show their results underperformed the polls. Cases where a claim on predictability for national levels is made while forecasts at state or provincial levels are unsuccessful were also labeled as “Maybe”.
4. Fields of science were based on the UNESCO nomenclature but, needless to say, sometimes it is quite difficult to determine such kind of information for an author; in that case they were labeled as “Unknown”. However, only 5% of the authors were eventually labeled this way.
Once the features to obtain and the criteria to be used were clear, I started the recollection of literature. To that end I relied on Google Scholar; it would have not made sense to use Web of Science or Scopus given that most of the literature is published in CS conferences that those repositories do not cover very well. Two different processes were applied to obtain the final list of candidate papers.
On a first phase the papers by Tumasjan et al. (2010), O’Connor et al. (2010) and the survey by Gayo-Avello (2013) were used as seeds. The underlying idea was that any paper on the topic would cite one of them (particularly the two foundational papers) or would be related to them by means of co-citation patterns. Then, those papers with a promising title or snippet were used as seeds for an additional iteration.
On a second phase, the query [T]  [T] Actually the query was not submitted that way for two reasons. First, to achieve maximum recall even under the restriction of 1,000 results from Google Scholar; secondly because of the limitations of the program Publish or Perish used to automate the downloading. That way, individual queries such as “twitter electoral forecasts”, “twitter elections forecasts”, etc. were submitted separately. (twitter OR tweets OR microblogs OR microblogging) AND (electoral OR elections) AND (forecasts OR forecasting OR predicts OR prediction) was used to collect additional bibliographical data. The rationale for that was that papers from outside the computer science community could have been missed by just relying on the prior snowball sample. In this case, no additional iterations were performed.
After completing both phases a list of about 4,500 papers was ready for preprocessing. First of all, works written in languages other than English were removed. The remaining ones were shortlisted (according to their title and abstract) with the goal of keeping just papers on Twitter-based electoral forecasting. That short list contained 115 papers; 59 of them were actually related to the topic and from that list 21 were removed because they were either surveys, duplicated work, or non-peer-reviewed works. Eventually, the list contained 38 papers published between 2010 and April 2015 that were downloaded and analyzed [U]  [U] The papers analyzed were O’Connor et al., 2010; Tumasjan et al., 2010; Bermingham and Smeaton, 2011; Chung and Mustafaraj, 2011; Fonseca, 2011; Gayo-Avello, 2011; Livne et al., 2011; Trumper et al., 2011; Borondo et al., 2012; Chung et al., 2012; He et al., 2012; Jungherr et al., 2012; Sang and Bos, 2012; Shi et al., 2012; Skoric et al., 2012; Soler et al., 2012; Beauchamp, 2013; Fink et al., 2013; Franch, 2013; Gaurav et al., 2013; Huberty, 2013; Jensen and Anstead, 2013; Lampos et al., 2013; Mahmood et al., 2013; Mejova et al., 2013; Pimenta et al., 2013; Sanders and van den Bosch, 2013; Thapen and Ghanem, 2013; Caldarelli et al., 2014; Cameron et al., 2014; Ceron et al., 2014; Mehndiratta et al., 2014; Razzaq et al., 2014; Unankard et al., 2014; Ceron et al., 2015; Khatua et al., 2015; Murthy and Petto, 2015; Tsakalidis et al., 2015a..
The main conclusion from that analysis was already introduced: 42% of the authors claim that elections can be accurately forecasted using Twitter data, 40% describe promising or mixed results, and just 18% claim that such kind of forecasts are unfeasible. How have we reached such contradictory situation?
To start with, there is little incentive to replicate prior results: either one achieves the same conclusion, and hence the paper is not novel; or one achieves a different result, and hence one must justify the reason for positive results if the prior paper achieved negative ones, or simply assume that something was wrong if achieving negative results when the prior paper claimed positive ones. In other words, a no-win situation.
However, by choosing a different election—and possibly a new venue—getting a publication seems much more feasible: if we achieved positive results, we “confirm” prior findings; if we achieved negative results, we “argue” that even more research is needed. In other words, a win-win situation. This is not just a cynical argumentation, we just need to check the data:
With regards to the forecasting methods applied, very few authors replicate prior ones, and when doing that, the most simple ones are chosen: those by Tumasjan et al. (29%) and O’Connor et al. (13%). If we aggregate the results of teams using each of such methods, we find that tweet counting has been reported to correctly forecast just 43% of the elections, while lexicon based sentiment analysis has apparently failed in every single forecast.
Still, papers not replicating those methods are not necessarily devising novel approaches; instead, most authors make variations of volume methods (counting tweets, retweets, users, unique users, etc.) and lexicon-based sentiment analysis.
On top of that, researchers insist on using only Twitter data, despite its biases, and only 18% of the papers relied on external data [Y]  [Y] Papers relying on external data are precisely the most promising ones., with pre-electoral polls (or prior electoral results) being the most popular source of ground truth data—71% of the papers using external data and 13% of the global.
Rather shockingly, taking into account the fruitfulness of the field, virtually none of the teams have further proceeded with their research. Actually, from 2010 to 2013 [Z]  [Z] I used that period given that seems unreasonable for authors publishing in 2014 to have a second paper on the topic in the same year or by mid 2015. 30 different teams published works in this area and only one has published a second paper [A]  [A] Interestingly, that team reported “positive” results in their first paper and “mixed” results in their second one. (i.e., using a different election). This is intriguing because 53% of the teams explicitly discussed future work to be conducted. If we pay attention to the relation between reported results and planning future work we have that 85% of those reporting “positive” results describe further research (although they eventually have not delivered), versus 36% of those reporting “mixed” results, and 17% of those reporting “negative” results. For some reason, teams with a purportedly working solution decide that no further proofs are needed.
Finally, we must consider which kind of researchers are working in this area. 72% of them come from STEM fields; what is more, a staggering 52% of them are computer scientists. However, no matter the technological requirements to work with Twitter’s API and processing the data, the truth is that electoral forecasting is a Political Science topic and, rather unfortunately, political scientists amount for just 10% of the authors. Statisticians, who would be obviously useful, are even rarer: 2.5%.
If we pay attention to the mix of backgrounds of the different teams we find that only 24% of the papers were a multidisciplinary product. Indeed, there are harsh differences in the results reported by each kind of team: While only 9% of STEM teams report negative results, 29% of social science teams, and 33% of multidisciplinary teams report such kind of outcome. This is surprising given that the percentage of negative results in STEM fields is above 20%, and in social sciences it is about 15% (Fanelli, 2010).

4.3.2 If it seems too good to be true, it probably is

Please note that I am not suggesting any sort of malpractice in the papers discussed above. There are a number of alternative and plausible factors explaining such a situation.
To start with, it seems clear that some researchers feel overly optimistic when approaching this problem. As soon as they have positive results, they quickly report them without reflecting on the underlying causes for such outcome; after all, their results agree with reality.
Negative results, in contrast, are much more difficult to report. On one hand, there are a myriad reasons that can explain them and, on top of that, since most “forecasting” methods are extremely crude researchers don’t feel confident enough to submit their research for publication. On the other hand, even if submitted, such kind of reports have not been traditionally welcome in conferences or journals. As an example of this I provide two comments by referees that my coauthors and me received in the past:
“[U]nless a ‘negative results’ paper is methodologically impeccable, it is hard for its conclusions to be believed.”
“The main difficulty with this type of counter-argument paper is that it is hard to make it immune to attacks of them form: «you would have done better if you did a different kind of analysis.» (if only you had looked at this other time periods, a broader set of tweets, applied this other sentiment analysis technique, etc.) It’s not clear to me what can be done to relieve this, but by concentrating on the failure of a specific set of techniques it is not obvious how the reader can take this as evidence of failure of the idea in general.”
From such comments a cynic would conclude two things: First, positive results papers do not need to be methodologically impeccable for their conclusions to be believed. Second, one should keep trying until they reaches positive results.
Given such state of affairs, it is very likely that for every “accurately” forecasted election ever reported there are multiple negative reports that have been put inside a drawer. Actually, there are two corollaries by Ioannidis (2005) that perfectly describe this area and which, unfortunately, imply that most of this research is false:
“The greater the flexibility in designs, definitions, outcomes, and analytical modes in a scientific field, the less likely the research findings are to be true.”
“The hotter a scientific field (with more scientific teams involved), the less likely the research findings are to be true.”

4.3.3 Promising recent research—that eventually failed

Despite what I have argued above, there have been a number of promising works; not due to their results (indeed, some of them were even “negative”) but because of their efforts towards systematizing a social media based forecasting methodology. It must be noted that their underlying premise is rather simple but it is still far from common in this particular area.
All of those teams have combined Twitter data with external ground truth data to “train” a model. Such ground truth has been of two main types: electoral results from previous elections or pre-electoral polls. Needless to say, using that kind of information is standard when developing any nowcasting system; however, in the case of Twitter-based electoral forecasting many researchers have purposely avoid it—maybe because it is difficult to claim something is going to kill the polling industry by using their data.
Pre-electoral polling data has been used by Shi et al. (2012), Beauchamp (2013), Franch (2013), Lampos et al. (2013) and Tsakalidis et al. (2015a). Past electoral data has been used by Huberty (2013) and Burnap et al. (2015).
In addition to that, these works include a few additional and interesting features. For instance, Franch (2013) aggregated data from different platforms, namely Facebook, Twitter, YouTube and Google; Tsakalidis et al. (2015a) applied their method to different elections held the same year to test its universality; Burnap et al. (2015) and Tsakalidis et al. (2015a) published their forecasts before the elections—something that seems obvious but which is uncommon in this area.
With regards to results, all of the teams reported promising or positive results; except for Huberty that reported that his method was unable to produce an accurate forecast, and Burnap et al. that published before the elections were held.
In a fortunate turn of events, Tsakalidis et al. produced a second prediction for the UK General Election of 2015 (Tsakalidis et al., 2015b) which was the same that Burnap et al. forecasted. Both attempts failed to accurately forecast the election and, although their mistakes have some commonalities there are also some intriguing differences.
Both forecasted a narrow win of the Labour Party. Burnap et al. projected 306 seats, and Tsakalidis et al. a vote share of 34.05%. The party actually obtained 232 seats with a vote share of 30.4%. In other words, it obtained 24% less of the projected seats and 11% less of the projected vote share.
Conservatives were supposed to lose the election by a small margin. Burnap et al. projected 285 seats, and Tsakalidis et al. a vote share of 33.24%. Tories actually won the election with 330 seats and 36.9% of the votes. That is, an increase of 16% in seats, and of 11% in vote share over the prediction.
Both teams failed when forecasting the outcome of the Liberal Democrats; however, while Tsakalidis et al. overestimated the results by 0.9 percentage points, Burnap et al. predicted the party would win 13 seats more than it actually achieved—an overestimation of 163%.
Burnap et al. also dramatically underestimated the projection for the Scottish National Party—to the point of acknowledging it even before the election—and the UK Independence Party. Curiously, the forecast by Tsakalidis et al. for both parties were perfect.
Conversely, the forecasts by Burnap et al. for the Green Party and Plaid Cymru were exact, but Tsakalidis et al. failed both of them.
Given that the method by Tsakalidis et al. relies on pre-electoral polls it could be argued that the failure of opinion polls in that particular election (Cowling, 2015) was the reason for the wrong forecasts. Still, that would not explain how results for other parties were accurately predicted. Moreover, Burnap et al. did not use polling data but results from the previous election and, thus, they were affected by other sources of error.
Anyway, at the moment of this writing some of the most thoroughly devised methods—even one that had been already proven in multiple elections—are still prone to failure. Worse, we have very little clues about why they behave in such a way; unless, of course, we accept that it may be that we cannot forecast elections in a systematic way using Twitter data.

4.3.4 What about the future?

In short, there exist an interesting and practical question that has been explored for almost 5 years with no conclusive results. Moreover, it may be that the answer to that question is negative; i.e., the methodology we are trying to develop is unfeasible.
That is problematic for two reasons. First, if social media based electoral forecasting is not possible it is exceedingly difficult to support the idea that other forms of public opinion can be mined from the same data source. Second, if such kind of forecasting is possible then it is unacceptable to use methods that are unreliable and operate as black boxes.
Needless to say, it will not be easy to make progress in this line of research, and among the issues hampering it we may find:
  1. The task is much harder than initially assumed: Social media is an adversarial scenario, sophisticated solutions to perform sentiment analysis are required, data is highly biased, the public is far from monolithic, and its self-expressive behavior is poorly understood. This does not only affect electoral forecasting but any form of public opinion mining from social media.
  2. The research community is fragmented: Every new “niche” increases its balkanization, it lacks actual multidisciplinarity, and prior work in psephology in particular and social sciences in general is mostly ignored by researchers without that background.
  3. We are producing small (almost anecdotal) advances but little actual breakthroughs. This does not only affect this particular area but most of social media research (Van Osch and Coursaris, 2015). I attribute this to the fact that the yearly cycle of major conferences and long cycles for journals drive the research calendar.
  4. No matter their importance, negative results are rarely published [B]  [B] This is slowly changing; see, for instance, the #FAIL! The Workshop Series ( and, hence, results are usually “enhanced” so that “promising” results tend to be eventually published, even though promised “future work” is rarely delivered.
  5. Elections and public opinion outside the US, and languages other than English are mostly ignored.
  6. Free data is unrepresentative and complete data is extremely expensive. In both cases it is sensitive and difficult to share for replication analysis. On top of that, even complete data from one single platform may not provide a coherent picture of society.
  7. Tools are not usually shared by researchers but just shallowly described in the papers.
  8. Expensive infrastructure is needed to both store the data and perform the computations.
All of those are very real problems but I am not saying that an international collaboration to build a Social Supercollider (Watts, 2013) is the solution to all our woes. Certainly, it would be nice if we could persuade ourselves that yearly cycles of conferences are a problem, and forget our scholar rivalries and feuds to embrace a huge collaborative effort, so we could secure hundreds of thousands, if not millions, in grants to buy data and cloud computing power.
Unfortunately, such a plan is impractical. As I have said, its outcome is extremely uncertain; after all, it may well be that we cannot exploit social media to predict elections on a general basis—or ascertain public opinion of any kind. If that was the case it would be a really expensive negative result. Moreover, even if elections could be accurately predicted on the basis of user generated content it would be difficult to consider such an outcome a major scientific breakthrough; at least, not taking into account its costs.
What we really need is an intermediate approach; something that could go beyond individual approaches tackling with single case scenarios and proof-of-concept methods, but was still easier to handle and, specially, less expensive than an international collaboration. In this regard, the TREC approach could be particularly well suited as it treasures plenty of experience in comparable initiatives, such as the TREC Microblog track (Ounis et al., 2011).
Ideally, an eventual initiative organized by TREC—or NTCIR or CLEF for that matter—would start by asking social scientist of different branches to sketch the broad lines of research and make the tough questions. Then, panels within the organization would decide on the more feasible tracks—both from a technical perspective and the level of interest among the community. Finally, calls for participation would be issued globally, and teams would work on an independent basis but on the same datasets. This way, year after year the state of the art would be slowly advanced, additional challenges—and dead ends—would be identified.
I myself have lobbied in two different occasions to organize an electoral forecasting data challenge. The first time was shortly after ICWSM 2011, and the second one before PLEAD 2013 (Weber et al., 2013). It goes without saying that I am not the best of the lobbyists but, still, I am hopeful that someone will eventually take up the gauntlet. In my opinion, this is the most feasible approach to conduct social media research with maximum guarantee of validity—and maybe we can once and for all know if Twitter can or cannot forecast elections.

5 Contentious politics

Stéphane Hessel
“If you want to liberate a society, just give them the internet.”
Wael Ghonim
“The revolution will not be tweeted.”
Jon B. Alterman

5.1 Introduction

The chapter devoted to political participation described it as voluntary actions conducted by particular citizens and broadly dealing with politics, the government or the state (van Deth, 2014). Voting, visiting political websites, following a candidate in a SNS, or persuading others to vote, were used as examples. However, those actions illustrate conventional politics and there are others that are unconventional, even contentious, forms of political participation—some were already mentioned in that chapter such as striking, demonstrating or rioting.
In that regard, Tarrow (2013) defined contentious politics as:
“[E]pisodic, public, collective interaction among makers of claims and their objects when: (a) at least one government is a claimant, an object of claims, or a party to the claims, and (b) the claims would, if realized, affect the interests of at least one of the claimants or objects of claims.”
Such a definition comprises not only the few aforementioned examples of unconventional participation but also social movements, strikes, civil wars and democratization processes. Previous chapters did not cover that kind of political actions even though in recent decades they have become commonplace (Thomassen, 2001: p. 181; Norris, 2002; Ortiz et al., 2013) and, on top of that, there are abundant examples of their use of social media [C]  [C] Just between 2009 and 2014 a large number of protests swept the world from Moldova to Hong Kong, passing through Iran, Tunisia, Egypt, Spain, or Ukraine. In all of them social media purportedly played a central role..
Hence, we need to discuss the confluence of social media with contentious politics in two main areas: First, to understand how social movements employ social media, and to what extent the nature of social media may be shaping them. Second, to tackle with the question of whether social media can or cannot boost democracy in authoritarian countries.
Finally, in spite of not being always considered a form of contentious politics, a brief review about terrorism and social media is pertinent in this chapter for a number of reasons: (1) Terror acts are undoubtedly of unconventional nature, (2) they are a recurring tacting by some political actors (Tilly, 2004b); (3) they are a form of political communication (McNair, 2011: pp. 167-172); and, (4) social media is being used by terrorists.
Social movements
Before analyzing how social movements are using social media for contentious politics we must first acknowledge that contention is unavoidable, and then understand the circumstances that may trigger it, and the role that social movements play in contentious politics.
Simply put, politics arises when people disagree about matters of public concern and rely on different mechanisms to settle such disagreements (Allison, 2009). In representative democracies citizens delegate in their representatives, and they can contact and lobby them; in addition to joining, for instance, political parties, unions, or non-governmental organizations. Needless to say, representatives are a subset of citizenry and, at the same time, the government may ask for citizens’ opinion in a number of ways.
Still, there are occasions when part or most of the citizens cannot (or do not want to) use those conventional mechanisms and, instead, they need (or prefer) to directly exert power against the state or the elites by means of disruptive actions (Tarrow, 2011: p. 6). It goes without saying that sporadic contentious behaviors are old as history itself; in contrast, contentious behaviors sustained through social movements are instead a more recent—and successful—invention.
Actually, for contention to appear powerless actors only need of a threat and a political opportunity; however, under those circumstances the confrontation against the more powerful actors will be probably short. Yet, if the resourceless actors can rely on a dense social network, they have developed a collective identity, they are action-oriented, and their claims are legitimate, then the contention can be sustained for longer periods, and in that case the group of actors must be considered a social movement (Della Porta and Diani, 2006: pp. 20-22; Tarrow, 2011: p. 16).
Thus considered, social movements can be traced back to the XIX century although they adopted their current form and most of its tactic repertoire along the XX century (cf. Tilly, 2004a). From the mid 1990s the Internet first and later social media have also been included (Norris, 2002; Bennett, 2003; Tarrow, 2011: pp. 137-139).
The impact of CMC is maybe one of the aspects that makes new social movements increasingly popular among citizens disenchanted with conventional political institutions (Rosanvallon, 2008: p. 18, Hay, 2013: p. 23). CMC offers them a diversity of convenient ways of participating in an action (Postmes and Brunsting, 2002) or “joining” a movement:
“Via the affordances of technological environments, individuals fraternize from the privacy of their own spheres, practicing a form of networked yet privé sociality that is formulated within a private social sphere.” (Papacharissi, 2010: p. 21)
Unfortunately, that convenience is a double-edged sword: On one hand it can greatly increase the outreach of a given social movement but, on the other hand, there is little control on participants (Postmes and Brunsting, 2002) and their level of involvement can be exceedingly low. For instance, slacktivism (Morozov, 2009b) or hashtag activism (Carr, 2012) are inexpensive forms of showing support or joining a campaign, but they suppose virtually null commitment and, moreover, they cannot be seriously considered as contentious actions [D]  [D] Hence, although this kind of manifestations are not conventional—or, much better, they were not conventional some years ago—they are not covered in this chapter but they were instead reviewed in the chapter devoted to political participation.. Yet, slacktivism is just one of the perils for new social movements; it is much more worrisome the way in which Internet and social media may be shaping some social movements and eventually tampering them.

5.2 Collective and connective action

We must distinguish between movements that incorporate Internet and social media into their repertoire from those that exist because of Internet and social media (Van Laer and Van Aelst, 2010). In the first case the movement would still conduct collective actions (mediated or not through social media) while in the second case we would be talking about connective actions (Bennett and Segerberg, 2012). However, before analyzing the features of such connective actions, we have to briefly describe those of collective actions and how they are affected by the incorporation of CMC technologies.
Collective action has been defined as “a group’s application of pooled resources to common ends” (Tilly, 1981) and although it is an obvious trait of social movements, it is not only characteristic of them. Indeed, when referring to the actions of such movements it would be much proper to call them contentious collective actions (Gould, 2005). Anyway, collective endeavors of any sort require some kind of organization—formal or informal—which in the contentious case may even “predate whatever the overt political issue is” (Gould, 2005).
However, when the resources to conduct the actions are inexpensive (e.g., handheld cameras or smartphones) or virtually free (e.g., video-sharing websites, blogging platforms, instant messaging) the costs of coordination, communication and information access are negligible and, thus, the need for a more formal or tight organization are less important (Bimber et al., 2005). Besides, computer mediated communications allow for transnational social movements and coalitions that were unthinkable before the Internet (Norris, 2002; Bennett, 2003).
Such circumstances have allowed for purportedly flat, distributed, and leaderless movements to emerge that have been largely praised (e.g., Castells, 2013). Indeed, most of the so-called Twitter and Facebook “revolutions” that will be covered later in this chapter somewhat respond to that idealized description; however, there are two important caveats to such idealization:
First, many of those allegedly leaderless and horizontal movements have well defined coordinating procedures and participants are assigned different roles. See as an example the description by McPhail and McCarthy (2005) of Black Bloc structure along affinity groups, and its reliance on coordinating technologies to conduct their actions [E]  [E] This should illustrate the fact that there is a huge difference between anarchism and chaos—although they are separated by a thin line., or the more recent comments by Gerbaudo (2012: p. 159) on social movements mediated through SNS:
“[D]espite their repeated claims to leaderlessness, contemporary social movements do have their own ‘choreographers’ and these choreographers are not identical with the ‘dancers’ or participants.”
Second, under its ideal form, such movements suffer of a number of weaknesses that may hamper their eventual success. To start with, heavy decentralization even without using CMC “permit[s]—and indeed encourage[s]—a lack of coordination and continuity” (Tarrow, 2011: pp. 131-132). When CMC is added to the mix, “problems of control, decision making and collective identity” also appear (Bennett, 2003).
Finally, for some of those flat and leaderless movements CMC is defining and thus it may be more appropriate to discuss connective rather than collective actions (Bennett and Segerberg, 2012). While collective actions can exploit CMC without any important effect on their dynamics, connective actions basically emerge from the digital tools they rely on. Moreover, they are individualized (Bennet, 2012) in a sense close to that of “networked individualism” (cf. Rainie and Wellman, 2012).
Thus, connective actions tend to draw from frames that are personally appealing to very different individuals, and easy to share in social media. At the same time, participants in connective actions do not consider themselves members of any group but individuals temporally connecting with others to pursue some concrete goal which is important on a personal basis. All of this means that collective identity [F]  [F] The argument by Bennett and Segerberg (2012) about the lack of collective identity in digitally networked actions has been criticized very recently by Gerbaudo and Treré (2015) and Bakardjieva (2015). Their criticism has not settled the issue but it seems that a much more nuanced conceptualization of collective identity is needed. First, different participants may self-identify with a movement at very different degrees; second, different movements can produce collective identities with bonds of varied strengths. The works by Gerbaudo & Treré and Bakardjieva belong to a special issue entirely devoted to social media and collective identity which I wholeheartedly recommend. Still, among the papers comprising that number, I consider the work by Treré (2015) particularly inspiring and a must read for anyone interested in studying the emergence of collective identity in digital environments. is unlikely to emerge in connective actions, and organizational requirements are scarce—in contrast to collective actions that imply some degree of organization and the eventual emergence of collective identity. Individualistic participation, in addition to dilute collective identity, may be eventually undermining of the movement as a whole (Fenton and Barassi, 2011).
Because of this, connective actions risk to become chaotic and unproductive. This can be avoided in two different (although complementary) ways: First, some kind of formal organization can emerge. If that did not happen, it is still possible that some action-oriented frames eventually resonate with a large enough number of individuals; in that case, and if the network is redundant and dense enough, those frames can spark potentially successful actions.
Of course, both collective and connective dynamics can occur among the same social movement, and because of that digitally networked actions are such a challenge for both scholars and politicians to properly understand. Still, it is undeniable that they are, at the moment of this writing, the most common form of contentious politics and one of the most attractive ways for many individuals to become politically engaged. In the next section I will review a number of such movements devoting some time to the social media technologies they used and the eventual results of their struggles.

5.3 Social media revolutions?

Protesters are quintessential bricoleurs [G]  [G] In the sense of Levi-Strauss (1966: pp. 17-22).: they take the means at hand and add them to their tactics repertoire after having repurposed them (Harlow, 2013). Naturally, the first time a tool plays a visible role during a protest it attracts the public attention. When it is a technology it may attract much more than its fair share, and lead to hyperbolic claims about its importance for the development of the protest.
This occurred with social media—actually SNS—during the cycle of protests of 2009-2014, and now it seems preposterous to talk about Twitter or Facebook revolutions. Not because those platforms did not play a role in the protests, but because their use was somewhat expected, like other technologies such as cell phones—and unlike others such as the telegraph. The 2009 Iranian protests, the Arab Spring or the Occupy movement were no more Twitter revolutions than the Mexican Revolution was a Leica Revolution (Mejías, 2010) or the 1992 Black May in Thailand was a “cellular phone revolution”.
Still, contemporary press labeled as such the events of May 1992 in Bangkok because protesters were using mobile phones—but also fax machines—to spread information and, supposedly, coordinate their actions (Shenon, 1992). It was also the press who quickly claimed Twitter revolutions were unfolding first in Moldova (Barry, 2009) and later in Iran (The Washington Times, 2009). Such approaches are both shortsighted and hyperbolic: They devote too much attention to anecdotal details while, at the same time, they make claims difficult to support, and set expectations almost impossible to fulfill.
Of course, that does not imply that social media is useless for contentious purposes; after all, if that was the case then protesters would not use it. Still, its use is highly varied, the implications of such use change from one protest to another—e.g., police can monitor social media to react to the protests; and its impact on the eventual outcome of the protests is difficult to disaggregate from the rest of actions.
Anyway, a brief chronological survey of different contentious actions and their use of CMC is needed to find common patterns. So, the next section will cover a selection of relevant movements with the exception of democratization uprisings which are devoted its own section.

5.3.1 Chiapas station, direction Hong Kong

Simplistic approaches to contentious uses of social media seem to imply not only that Twitter or Facebook have been instrumental in recent protests, but also that events such as the Arab Spring or the Indignados are somehow “new” forms of contention.
Yet, nonviolent occupation—as seen in Tahrir Square, Puerta del Sol, Zuccotti Park, Gezi Park or the Admiralty—is just one of the many methods compiled by Gene Sharp as early as 1973, and it was a major tactic during the French May 1968 (cf. Feenberg and Freedman, 2010) and well before that.
Thus, we must contextualize those recent movements within a broader collection of events; to that end, I will cover relevant actions taking place during two decades—from 1994 to 2014—and spanning 9 countries. Chiapas: Getting outreach

On January 1, 1994 the North American Free Trade Agreement (NAFTA) signed by Canada, Mexico and the United States came into force. The agreement was opposed by Mexican peasants, particularly indigenous people, for two main reasons: unrestricted imports of US subsidized agricultural products; and large amendments and section repeals in Article 27 of the Constitution of Mexico that affected land property and use by indigenous people (Leyva García, 1992).
In consequence, Indigenous peasants faced NAFTA as a direct threat to their way of living and during 1992 and 1993 massive marches protesting against the agreement were organized, and dutifully ignored by the Mexican government (Castells, 2010: p. 78). Therefore, some peasants changed their strategy and constituted the Zapatista Army of National Liberation (known in Spanish as EZLN).
The same day that NAFTA came into force EZLN went public by issuing a manifesto (EZLN, 1993) and seizing control of a number of populated areas in the Mexican state of Chiapas (Martínez-Torres, 2001). The manifesto was in fact a declaration of war against the Mexican state; they called the population and the army to raise against the de facto dictatorship of PRI (the party that had ruled Mexico for 70 years), and fight for freedom, democracy, and social welfare. The EZLN depicted themselves as true Mexican patriots, and based their call to arms on Article 39 of the Mexican constitution that claims that “[a]ll public power originates in the people [and] [t]he people at all times have the inalienable right to alter or modify their form of government.”
EZLN failed to trigger a revolution and they retreated as soon they were confronted by the Mexican army. Thus, they had to change their strategy again (Martínez-Torres, 2001): they started a communication strategy which heavily relied on the Internet to spread their message undistorted by the Mexican government or the mass media (Castells, 2010: pp. 82-84). That way they were able to earn support within Mexico but particularly abroad (Garrido & Halavais, 2003). That support attracted international attention—especially from foreign investors—to the response that Mexico’s government would give to EZLN, and quickly forced the government to negotiate with them (Martínez-Torres, 2001).
Eventually, some legislative reforms addressing issues affecting indigenous people have been passed and there are a substantial amount of de facto indigenous self-government in certain areas of Chiapas. Moreover, the EZLN has somewhat muted into an anti-globalization movement which still heavily relies on the Internet (Castells, 2010: p. 86).
The interested readers should consult the extensive—although at times hyperbolic—work by Ronfeldt et al. (1998). There, EZLN is described as a paramount example of netwar [H]  [H] Please note that the netwar concept is not focused on technological networks but in the broadest sense of networking, including offline methods in addition to online ones., defined as contentious actions where “the protagonists depend on using network forms of organization, doctrine, strategy, and technology” (Ibid.: p. xi) and that is “acephalous at times, and polycephalous at other times” (Ibid.: p. 13). Seattle, Philippines, Spain and Italy: Triggering demonstrations

EZLN used the Internet to spread their message abroad, earn support and, in turn, leverage that international support and attention to force the Mexican government to negotiate. Other movements have used social media to mobilize people and trigger concrete actions on the ground, instead. Among the most relevant examples we can find the 1999 Seattle protests against the World Trade Organization, the Philippines uprisings of 2001, the March 13 demonstrations in Spain in 2004, or the V-Day Celebration in Italy in 2007.
Seattle was the venue for the 1999 meeting of the World Trade Organization (WTO); the goal of that meeting was to launch a new round of multilateral trade negotiations. The WTO is frequently object of criticism by different organizations and movements on the basis that its actions favor richer countries and corporations, and that their decision making processes are obscure and secretive.
The Seattle meeting was heavily disrupted by tens of thousands of protesters from all around the world. Such impressive protest was the outcome of months of coordination by very different movements and NGO. They necessarily had to resort to the Web given the transnational nature of the action. Interlinked websites (Van Aelst and Walgrave, 2002) were used to spread information about the WTO, its practices and consequences. Such information certainly reached huge numbers of potential activists and helped to weave a dense network among the activists; however, in order to attract them to Seattle, much more was needed. In that regard, different mailing lists were used for coordination purposes—both at national and international levels, allocation of resources and for discussion among the activists (Eagleton-Pierce, 2010).
Two years later, the then president of Philippines, Joseph Estrada, was facing impeachment on charges of corruption. The trial was being broadcasted by radio and television and population was obviously interested. After a close vote, the court decided to ignore a piece of evidence and the trial was aborted. That night anti-Estrada protesters gathered to ask for his resignation, many of them instructed by means of SMS (Tilly, 2004a: p. 95; Castells et al., 2005). Further mobilization messages were shared among protesters during the following days by means of cell phones and, to a lesser extent, computers (Rafael, 2012). Eventually, hundreds of thousands of persons attended the demonstrations and, finally, they were backed by the army, effectively ousting Estrada who was replaced by his vice president.
Given that outcome one could claim victory for the social media triggered uprising. However, despite the irregularities during the trial, the truth is that this event was a soft coup [I]  [I] Joseph Estrada was democratically elected and he was forced to resign not impeached.. Moreover, as Tilly (2004a: p. 96) or Rafael (2012) note, some elites (including the Catholic Church in the Philippines) were overtly calling for the mobilization and, thus, the People Power II movement cannot be considered a grassroot movement. Finally, it is difficult to determine the impact of social media (SMS sharing in this case) in the outcome. The most that can be said is that the volume of messages handled by operators during the uprising were much higher than average (Ibid.), and that they even needed to upscale the infrastructure in the area where the demonstrations took place (Castells et al., 2005).
SMS also played a central role during the March 13 demonstrations in Spain (13M). Those protests were addressed towards the conservative ruling party because of their response to the deadly train bombings that had took place 2 days earlier, and merely 3 days before the general elections. Protesters were asking the conservative party to tell the truth given that many believed that the official version was aimed to secure an electoral victory.
Two days before, four commuter trains were simultaneously bombed during the morning rush hour. Almost 200 people died and 1,800 were wounded. The government quickly claimed that the terrorist group ETA was responsible. Given that the attacks were short before the elections if such theory was confirmed the ruling party would likely secure its victory. Indeed, the government insisted during the following days on ETA’s authorship, despite the terrorist group denied it the day after the attacks, and that the islamist theory was gaining strength among the police.
The day before the elections many citizens already distrust the official version in the mass media and were visiting foreign news outlets looking for information (Castells et al., 2005). The same day, a SMS calling for concentrations in front of the conservative party headquarters was spread, even though demonstrations and political events are unlawful in Spain the day before elections. That evening about 4,000 persons gathered in front ot the party’s offices asking for the truth; the press that was there to cover the elections live broadcasted the concentration. Shortly after that, a group of suspects was arrested and the government had to acknowledge that they were associated to al-Qaeda. The next day, the conservative party lost the elections.
Unlike prior cases it has been possible to prove that this chain of events had a real impact on the electoral results. Montalvo (2011) compared results of postal ballots issued abroad (days before the attacks) with results of ballots issued the day of the elections. He found that turnout increased by 2.7 percentage points and that, if not for the attacks, the conservative party would have won the elections. However, as with the case of Philippines, it is almost impossible to determine the impact of the demonstrations; still, the SMS volume was abnormally high that day (Flesher Fominaya, 2011). Also like the People Power II movement, it has been argued that the original message was not spontaneous but crafted by left-wing activists to trigger the mobilization and hamper with the electoral process (Ibid.). As a matter of fact, Pablo Iglesias, leader of Podemos [J]  [J] Podemos is a political party that has emerged from the breeding ground of 15M, a movement discussed later in this section. has recently claimed that the message was devised by people close to him (Manetto, 2014).
The last case of social media triggered demonstrations is Beppe Grillo’s V-Day [K]  [K] Beppe Grillo and his V-Days were already introduced in chapter 2 Political Actors when discussing the Five Star Movement. which sheds some light on the actual impact of social media. As introduced in chapter 2 Political actors, Grillo is a comedian turned into blogger turned into politician. He founded the Five Star Movement (M5S), a movement and later a party that has enjoyed huge success in recent elections, obtaining representation in both the Italian senate and chamber of deputies, and in the European Parliament.
Before starting M5S Grillo organized several V-Days which were days of action addressed at the government and the political elites. The first one was announced on Grillo’s blog on June 14, 2007 and set for September 8, 2007. The goal was to collect signatures to “demand a referendum for a new law banning convicted criminals from Parliament” (Pepe and di Gennaro, 2009) and it was a huge success: one million people demonstrated in more than 200 cities across Italy, and more than 330,000 signatures were collected, while only 50,000 are needed in Italy to call for a referendum (Ibid.)
Unlike prior cases where it was impossible to disaggregate the impact of social media from other events, the call for the V-Day was only spread through social media—mainly blogs—and word-of-mouth. It was completely ignored by mass media and, indeed, major newspapers did even ignore the protest itself. In addition to spreading the call, social media platforms such as Meetup were heavily used to conduct the local organization in each town. Thus, as with the 1999 Seattle protests it is pretty clear that social media can be used to successfully organized massive events; still, organization does not emerge and has to be carefully crafted at the beginning.
Pepe and di Gennaro answered a second important question: whether the V-Day call was top-down organized by Grillo or if it was a networked movement. When analyzing the network of blogs discussing V-Day they obviously found that Beppe Grillo’s blog was the most central node, but also that the network was rather dense and robust and, thus, that Grillo was not the only actor spreading the call. Moreover, taking into account that the goal of the V-Day was not only to gather people but to collect signatures for a popular initiative, and also the heavy use of Meetup, it is clear that most of the organization was performed by local actors. In this regard, it seems that the V-Day was closer to the netwar concept than the EZLN strategy. Spain, New York, Istanbul and Hong Kong: Occupation

Peaceful occupation of public spaces is not a new form of protest; although it had somewhat disappeared from the public’s conscience since Tiananmen. However, the images of Tahrir Square during the Egyptian Revolution of 2011 swept the world and occupation was again a fashionable tactic.
Though, I will not cover the Arab Spring here; instead, I will briefly review a number of relevant occupation movements inspired by Tahrir and the role social media played on them. The underlying causes for each of those occupations were different—namely, austerity in Spain, economic inequality in the US, authoritarianism in Turkey, and electoral reforms in Hong Kong. Thus, we can safely assume that common features will mostly apply to any similar contentious action no matter its causes.
Spain: Toma la calle (Take over the street)
Let’s start with southern Europe. In 2010, countries such as Greece, Portugal or Spain were facing the results of austerity measures adopted because of the Great Recession: unemployment was rising, and public services characteristic of the welfare state were suffering huge budget cuts. Demonstrations and strikes were common in Greece and, in early 2011, a number of movements were demanding changes from their respective governments, the so-called outraged (geração à rasca in Portugal, indignados in Spain and aganaktismenoi in Greece). Chronologically speaking, protests started earlier in Greece and were followed by Portugal and Spain. However, it was Spain the first democratic country where citizens occupied the squares with tent camps and, thus, it was the starter of the outraged movement; therefore, a brief review of the history of the indignados and their social media use is needed.
Rather surprisingly, the trigger was an anti-piracy law. By late 2010 the socialist government was preparing a law to heavily restrict file-sharing services. Although contested—mainly by online activists—it was eventually passed in early 2011 with the support of a number of parties besides the government’s socialist party—among them the conservative party. In response, an online campaign was started: (don’t vote for them). The website provided a manifesto and links to a wiki, chats, mailing lists and different resources; the goal was to persuade voters to choose an option other than the two major parties in the upcoming local elections. The hashtag #nolesvotes was heavily promoted and, slowly, the it was co-opted by different organizations and, thus, discussion drifted from anti-piracy laws to the many other reasons for discontent with the government and political elites.
Other organizations emerged about the same time such as Juventud sin Futuro (Youth without Future), Democracia Real Ya (DRY, Real Democracy Now), and later Toma la Plaza (Take over the Square). DRY was the one that eventually called for demonstrations all over Spain for May 15th (15M), a week before the elections. Their manifesto asked for a major change to reduce inequality and improve governance—it also introduced the key term indignado. In addition to the website they used web forums extensively, besides Twitter, Facebook and YouTube. They also linked to blogs of local groups organizing the different marches. Most of the other mentioned organizations endorsed and spread the call.
The 15M demonstrations were a huge success and about 130,000 persons marched on 60 cities across the country. The call was able to mobilize individuals not usually taking part in contemporary demonstrations in Spain (Anduiza et al., 2014): they were younger, more educated, and less politically involved than other demonstrators; women and jobless were a much larger fraction than in other demonstrations; and they confirmed that the main channel of diffusion for the call had been word-of-mouth and SNS.
So far there was nothing particularly unusual about 15M. However, in Madrid, once the march finished, a group of about 90 persons went back to Puerta del Sol [L]  [L] La Puerta del Sol is not just a central square in Madrid, it is Spain’s kilometre zero, a communication hub with metro and commuter train stations, and last but not least the focus of a New Year’s Eve celebration which is nationally broadcast yearly. to debate, and decided to establish a permanent camp at the square at least until the elections. That first night 50 persons stayed in the square—less than one week later there would be 28,000 (Sánchez, 2015). On the first day a blog, a streaming channel to broadcast the assemblies and a Twitter account were created (Acampada Indefinida en Sol, 2011); shortly after, Facebook and YouTube were incorporated.
The 15M movement had started and they were using social media from the very beginning but mainly to inform, get outreach and attract support; the real activity was conducted on the ground in the assemblies [M]  [M] The general assemblies held during 15M were its signature and they were later “exported” to the Occupy movement. They were key to sustain the movement and, obviously, they are a fully offline old-school activist action. where proposals were made, debated and adopted or rejected. Indeed, Theocharis et al. (2013) found that Twitter was seldom used by 15M activists for political participation, logistics or coordination; information diffusion was the most common application.
Be that as it may, the news about the camp spread online and triggered additional camps in other cities making the 15M movement global. Though, an analysis of the communication patterns of activists on Twitter (Borge-Holthoefer et al., 2011) revealed that the movement was highly fragmented: activists were communicating mostly within their local group and seldomly with activists in Sol.
Although these findings may be somewhat discouraging we must acknowledge that coordinating a global demonstration would have been impossible without social media, that camps emerged due to the news spreading online, and the movement was sustained thanks to a continuously updated stream of multimedia information; asking for meaningful deliberations being held online would have been too much.
Still, there was another occasion when social media helped 15M: After just two days of occupation police tried to forcefully evict the camp; the action was quickly publicized on Twitter and broadcast on YouTube and, in response, thousands of activists gathered at Sol (Hughes, 2011).
So, in short, 15M was organized through online calls and local groups on the ground, and the information spread online got attention and support, and triggered additional camps. Still, deliberation and decisions were taken in physical meetings. 15M illustrates that new social movements employ social media when it fits their needs: sometimes it is crucial and others irrelevant.
New York: Occupy Wall Street
The occupation of squares in Spain earned attention from mass media worldwide and during the rest of 2011 and following years peaceful occupation was heavily used by a large array of movements, starting with Occupy Wall Street (OWS) and following with other cities in the US and the rest of the world.
As with many other movements, several groups took part in the initial stages of OWS taking advantage of prior experience or plans for similar actions (Kroll, 2011). Still, the call to action, the name for the movement, and the iconic poster of the ballerina on the bull were the product of Adbusters, a Canadian anti-consumerist environmentalist organization that had started other campaigns such as Buy Nothing Day.
On February 2011, Kono Matsu published a post [N]  [N] in Adbusters blog digressing on what was needed to have “a million man march on Wall Street”. He argued that people in the US had reasons to uprise and that they needed to start organizing and planning, in the same way that the apparently spontaneous uprisings in the Middle East had been organized in advance.
On May 2011, a new post [O]  [O] praised the so-called Spanish Revolution (i.e., 15M), particularly their improvement over the Tahrir model: Instead of just occupying a space transforming it into an exercise of direct democracy by means of the assemblies. Indeed, OWS would later apply the assembly model that was new in the American protest repertoire (Kroll, 2011).
On July 2011, the actual call to action was issued [P]  [P] A date was set (September 17), a name and hashtag chosen (#OcuppyWallStreet), an explicit call to camp on the site was made and the initial demand stated: “that Barack Obama ordain a Presidential Commission tasked with ending the influence money has over our representatives in Washington”. The day after the post was published, a comment announced a subreddit and a website to organize the action ( Another comment pointed out to the global website and its corresponding Facebook page which was informing about the different occupied squares worldwide (in that case inspired by the 15M movement). Shortly after, a new blog post announced the independent website; slowly profiles in other social media platforms were included (e.g., Facebook, Twitter or Tumblr).
By August 2011, General Assemblies a la 15M were being held in New York [Q]  [Q] the attendees discussed matters such as logistics for the camp, the actual demands to issue, the organization of future general assemblies, or the different committees required to organize the whole campaign. Again, as the other movements discussed above, participants went offline to held face-to-face meetings. Also on August 2011, it was announced that other movements were to occupy their respective stock markets: Madrid and San Francisco were confirmed and other cities were rumored to join [R]  [R]
During early September, Adbusters called [S]  [S] for a “meme war” to get a critical mass of inspiring visuals that could be used in the different camps. Finally, on September 17 300 persons (Gerbaudo, 2012: p. 102) camped in Zuccotti Park, close to Wall Street.
Hence, as a social media ignited action OWS did not started well, to say the least (Gerbaudo, 2012: pp. 113-117). Later, its purpose and demands were redefined into the eventual “We are the 99%”; however, such repurposing of the movement was conducted on the ground and not online (Gerbaudo, 2012: p. 102). Besides, social media was being used mostly to spread information (Theocharis et al., 2013) and to make the events on the ground resonate within a broader audience (Gerbaudo, 2012: p. 117). Its purported advantages to organize movements, attract activists and spark debate were missing from OWS. Actually, one of the Adbusters’ responsibles of the OWS call recently said in an interview:
“During Occupy, we experience it [the negative side of social media]: things started to look better on social networks than in real life. Then people started to focus on social media and to feel more comfortable posting on Twitter and Facebook than going to an Occupy event. To me this is the biggest risk: to become spectators of our own protests.” (Vallone, 2015).
Istanbul: Diren Gezi Parki (Resist, Gezi Park)
While the spark in Spain was an anti-piracy law in Turkey it was a park (Kuymulu, 2013). On May 2013, protesters opposing a plan to reform Gezi Park in Istanbul in order to build a shopping mall prevented the demolition, they used social media to call for other activists to defend the park and established a camp inside. Their only goal was to defend a public good from being destructed to benefit a minority. Three days later, the police raid the camp with water cannons and tear gas; the protesters had to leave the park but they gathered in the nearby Taksim Square. The news of police brutality spread through social media because major news outlets were not informing about the events (Kuymulu, 2013; Haciyakupoglu and Zhang, 2015). Tens of thousands of persons responded to this and gathered in the square. Again the police brutally tried to crash the demonstrators which, in turn, triggered additional demonstrations in other places in Istanbul and other Turkish cities.
The protesters taking part in the subsequent demonstrations and occupations were moved not by the park, but due to deeper discontent with Turkish political affairs (Tastan, 2013) such as “restrictions on liberties, government interference in their daily lives, and the Prime Minister’s authoritarian rule”—all of which was painfully illustrated in the brutal repression of the original camp.
Haciyakupoglu and Zhang (2015) interviewed 21 participants in the protests to better understand their use of social media. This work is particularly interesting because it reveals an important flaw in many of the research surveyed up to this moment. Most researchers have relied on data crawled from public profiles in social media such as Twitter and Facebook. However, Haciyakupoglu and Zhang found that protesters in Turkey heavily relied on Whatsapp to obtain trustworthy information or double check information received from other social media platforms; closed groups in Facebook were used for the same purposes, but to a lesser extent because of bandwidth issues on the ground.
Twitter certainly played a role in Gezi Park protests; however, it was used as a source of real time information but not to exchange sensitive communication. Indeed, most of those interviewed by Haciyakupoglu and Zhang abstained from tweeting for fear of their tweets being reframed or spread within social circles unknown to them. Facebook was considered better suited for dissemination purposes because of the unlimited length of the posts and, especially, because protesters felt that they were able to better manage their privacy settings. Still, Facebook use was mainly limited to out of the field reporting: protesters found its interface too cumbersome to use on the ground. Finally, SMS were also heavily used for a number of reasons: to save battery, to communicate even without Internet connection, and to interact with protesters not using smartphones.
These findings have profound implications. First, we must reconsider popular depictions of so-called social media revolutions. Second, we must be very cautious when interpreting research that is focused in one single platform, particularly Twitter.
I said above that protesters are bricoleurs; Turkish activists are a vivid example: they exploited the tools that best suited their actions depending on factors such as trust, privacy, usability, or as mundane as battery life. They moved across them seamlessly, and while they were extremely active on some, they were passive consumers in others; indeed, the most public the platform the less active they appeared.
This means that many of the communication among activists was not visible for the casual user—including scholars but, more importantly, the authorities [T]  [T] Activists tend to fear the police is monitoring what is published in social media and, thus, they may avoid a given platform for certain kinds of communications (Penney and Dadas, 2013).. Assuming such a behavior is not uncommon in other protests, it may be that many of the claims about lack of political discourse or coordinating actions in social media are incorrect— they may be occurring “underground” by means of Whatsapp, SMS, email or direct messages in Twitter. Again, the monoculture of a few social media platforms results a problematic research practice. Though, researchers have no means of accessing private data and, thus, they can only aim at surveying protesters to get a glimpse of their practices.
Hong Kong: Umbrella Revolution
On July 1, 1997 Hong Kong was transferred from the United Kingdom’s rule to that of China. Purportedly, citizens of Hong Kong would not notice the change and, indeed, they were supposed to freely elect their Chief Executive (mostly equivalent to the former Governor) by 2017 (Kaiman, 2014). However, in 2014 an electoral reform was announced that was felt by many as too restrictive with regard to the possible candidates running for election (Ibid.) and breaching the handover agreement with the UK (Moore, 2014).
Among the different organizations lobbying to get free elections in Hong Kong and opposing that reform it was Occupy Central with Love and Peace (OCLP), started in 2013. Although clearly linked with other Occupy movements they saw occupation as the last resort, and other civic activities were conducted well before the announcement of the electoral reform.
As other movements described before, OCLP has relied on a mixture of tactics, using social media (blogs, Twitter, Facebook, Instagram or YouTube) to spread their message [U]  [U] For instance, the Twitter account (@OCLPHK) is basically used for broadcasting not for interaction with other users. Its follower base grew rather slowly before the occupation and jumped from less than 5,000 followers before the camp to 15,000 immediately after. During the succeeding months it grew quickly reaching a plateau of about 27,000 followers at the eviction of the camps. and get support from individuals and other organizations, both local and abroad. Still, the main actions were on the ground acts of civil disobedience, being the most visible one the occupation of certain areas of the economic district of Hong Kong by September 2014.
To the best of my knowledge the only scholar analysis of the unfolding of the movement in social media has been conducted by Fu and Chan (2015). They employed snowball sampling to collect Hong Kong based Facebook pages with posts covering the timeline of the protests. By means of shared posts they produced a network and tried to ascertain the different communities within. Unsurprisingly they detected OCLP as a central actor and, as with other movements, they found a number of additional civic organizations and independent media endorsing or coalescing with the movement. In contrast to other movements, however, they found a very active set of pages which opposed the movement and were clearly pro-Beijing. It is still an open question whether this was the result of the movement opposing a measure by an authoritarian regime, or if it can be in the future a common situation also in democratic countries.

5.3.2 London: rioting

With the exception of the WTO Seattle protests and isolated acts of violence by minor groups, most of the cases discussed so far were notably peaceful. When violence emerged it tended to be exercised by the authorities and not the protesters [V]  [V] As a matter of fact, police violence tend to trigger additional support by citizenry and attract more activists to the actions (it happened in 15M, OWS, Gezi and Tahrir). However, to the best of my knowledge nobody has suggested that riot gear is a democracy catalizer.. However, contentious actions can be extremely violent—think for instance of a civil war or a revolution—and some recent events have been linked to social media, such as the 2011 London riots (Fuchs, 2012). Actually, prime minister David Cameron (2011) officially suggested the possibility of taking down social media sites:
“Mr Speaker, everyone watching these horrific actions will be stuck by how they were organised via social media.
Free flow of information can be used for good. But it can also be used for ill.
And when people are using social media for violence we need to stop them.
So we are working with the Police, the intelligence services and industry to look at whether it would be right to stop people communicating via these websites and services when we know they are plotting violence, disorder and criminality.”
This time the spark was police violence: On August 5, 2011 Mark Duggan was shot dead by police in Tottenham, London. The next day about 200 people, including local civic leaders and Duggan’s family members gathered in front of Tottenham police station demanding answers on his death. Police did not engage with them and in the following hours tension escalated and eventually a riot erupted (Lewis, 2011). Amid the ensuing chaos, vandalism, arson and looting spread not only in Tottenham but in other neighborhoods (BBC, 2011) and, in a rather shocking reaction, police claimed that social media such as Twitter were being used by the rioters to coordinate their actions (Ibid.)
At this point we may wonder whether the events after the demonstration are actually contentious political actions or merely gratuitous acts of violence. It is true that there were looting; however, it amounted for just one third of the incidents, and in some areas for less than a tenth (Platts-Fowler, 2013). The rioters that were not taking part in the looting were risking their physical security and legal punishment apparently for nothing, but actually to engage with the police. Waldman (2015) very vividly argued when discussing the 2015 Baltimore riots that such a behavior responds to altruistic punishment and, thus, it must be read in political terms (even though rioters are not making any political claim).
Altruistic punishment consists of inflicting costs on third parties although such action is also costly for the punisher and produces no gain (Fehr and Gächter, 2002). Surprisingly, altruistic punishment is crucial for future cooperation among individuals and when it is not possible cooperation diminishes. Besides, negative emotions are a crucial proximate cause to trigger altruistic punishment.
Taking that into account, the thesis by Waldman is clearly plausible and totally applicable (although backwards) to the London riots. Moreover, the riots must be put within a broader context of inequality and collective discontent in the UK (cf. Lewis et al., 2011); such a scenario can eventually produce collective violent actions if there is a motivation, such as police violence (Fuchs, 2012).
With regard to the mass media arguments about social media fueling the riots, they were as missed as those suggesting it triggered the Arab Spring or it was crucial for the anti-austerity protests. As Fuchs (Ibid.) puts it:
“There is today a generation that has entirely grown up with the use of these technologies [internet and mobile phones]. In such a society, communication technologies do not cause riots, revolutions, or rebellions; but rather discontented people will make use of all means necessary and available, including communication technologies, in order to achieve their goals.”
So, the London riots were definitely contentious actions, and yes, they used social media—particularly BlackBerry Messenger (Lewis et al., 2012: pp. 30-33). Yet, as with the rest of events described up to this point, such use is not definitory of the actions but merely of the social context where they took place (Fuchs, 2012). Still, there are two problematic consequences of the purported interaction of social media and rioting.
First, the fact that a democratic government seriously discussed the possibility of taking down, on the basis of social unrest, media that allows citizens free speech.
Second, the blossoming of technological “solutions” that simply ignore the social causes of unrest and contention, and instead go straightforwardly to propose permanent monitoring to forecast the need of police response (e.g., Davies et al., 2013; Cadena et al., 2015; Muthiah et al., 2015).
Both facts reveal a worrisome trend towards authoritarianism in democratic countries that will be discussed at large in the next chapter.

5.3.3 What about the outcomes?

Contentious politics, particularly protests, deal with claims that activists aim to get realized. What can we say about the demands of aforementioned movements? And what about the impact of social media to achieve them?
To start with, social media was not crucial for all of those movements. It is debatable whether cell phone messages were the catalyzer of the People Power II movement in Philippines. I have also shown that for most of the occupation movements social media was part of a complex mix which also required plenty of work on the ground—actually, the social media orchestration of OWS by Adbusters would probably failed if not for the local NYC activists, and still only 300 persons attended the first day.
Yet, there were some actions that had to unavoidably rely on social media and succeeded to attract participants: V-Day, the 13M concentration, the 15M marches or the Gezi Park protests. Finally, EZLN successfully employed Internet to leverage international support and attention to force the Mexican government to negotiate.
With regard to the realization of the claims and the importance of social media to achieve them results are less encouraging.
The Umbrella Revolution and OWS simply faded without getting any of their demands satisfied. The 15M marches did not seem to affect the local elections, and it is debatable whether the 13M concentration had any impact in the 2008 general election. The V-Day gathered the required signatures to submit a popular initiative law project, but it ended up in legislative limbo. EZLN failed to trigger a new Mexican revolution and oust PRI from power, although it got some reforms on indigenous issues and a certain level of de facto self-government in areas of Chiapas.
It may be that the 15M movement is the most successful case not because of reaching their demands—they failed to produce a concrete set—but because of increasing civic participation; this has led to the creation of new political parties and popular coalitions which are slowly entering in local and regional governments.
Of course it would be unfair to blame social media for the failures of some of those movements, but also to praise it for their partial successes given that all of them have involved a huge number of activities conducted on the ground and in face-to-face meetings. So, in short, social media is an activist tool but it is not the only one nor the most important.

5.4 Social media is not a democratization force

At the moment of this writing, there are few people claiming that social media can boost democracy but the truth is that a good amount of hope was put on that idea:
“Judging by the popular press, in January 2011 Twitter and Facebook [...] upended decades of Arab authoritarianism. [...] [S]ocial media prompted hundreds of thousands, and then millions, of Tunisians and Egyptians to pour into the streets and peacefully demand change. Brittle authoritarian regimes had little choice but to comply, and in this way, social media irrevocably changed the future of the Middle East.” (Alterman, 2011)
I must state again the obvious: technology does not free people, people free themselves (Diamond, 2010; Howard and Hussain, 2011). I must state it because, sooner or later, there will be another wave of optimism about the umpteenth CMC tool which will free people from tyranny’s yoke. I am certain of such a revival because we have been already there: before Twitter and Facebook it was the Web, and before the Web it was the Internet, all of them would bring democracy.
In the following subsections I will cover the rather unfounded hopes that time and again have been put on so-called liberation technologies. First, I will briefly show how the mere existence of Internet was supposed to bring democracy; then I will broadly survey both the Iranian protests of 2009 and the Arab Spring of 2010-2011, and how social media was depicted as a democratic booster in both cases.

5.4.1 Internet was supposed to bring democracy too

Diamond (2010) defines liberation technology as “any form of information and communication technology (ICT) that can expand political, social, and economic freedom” and offers as examples the computer, the Internet, the mobile phone and, of course, social media. It must be noted that he explicitly argues against assuming that technologies on themselves are enough—he considers such approach “technological utopianism”. Still, it has been pretty common to attribute some sort of silver bullet nature to most of those technologies, which is dangerous in the long term.
A good example of that wishful thinking is a RAND report (Kedzie, 1997) which argues that the introduction of electronic communications in authoritarian regimes coincided with their collapse, using the Soviet Union as a case study. The report is also worrisome because it recommends the US and their allies to actively encourage telecommunications as a sure way to bring democracy to authoritarian countries. The following paragraph shows how misguided such premises were:
“Whether the dictator admits it or not, E is in a bad situation. In some countries, like China with its ‘intranet,’ leaders might prefer to believe they can utilize information and communication technologies to serve their own economic needs while sufficiently sterilizing against the untoward social effects. The evidence and analysis here suggests that, in the long run, these hopes are unfounded.”
China had introduced Internet just three years before that report and, at the moment of this writing—about 20 year later—expecting that liberal democracy will just “occur” in China because of Internet (or social media, for that matter) seems exceedingly optimistic.
That line of thought is problematic for a number of reasons. First, it focuses on technology (the mean) instead of politics (the end); second, it may divert resources that could be better employed in other areas to develop ineffective projects—e.g., the ZunZuneo SNS in Cuba [W]  [W] ZunZuneo was a SNS targeted at the Cuban market and which operated over SMS—yet, messages were sent to a phone number located in Spain. Although it was dubbed as the “Cuban Twitter”, its operation was much more similar to WhatsApp groups. Users were able to create and subscribe to different groups and thus send and receive messages through them. They could also subscribe to different topics such as News, Sports, Music, etc. According to Butler et al. (2014) the service was funded through USAID and was eventually intended to stir a “Cuban Spring”; however, USAID (Herrick, 2014) denied such claims. Anyway, ZunZuneo had stopped operating around March 2012 and before that moment it had worked unevenly.; third, it can easily backfire by suggesting authoritarian regimes that citizens using such kind of tools are dissidents [X]  [X] Although the number of Iranian Twitter users during the 2009 protests was well below 20,000, the U.S. State Department asked Twitter to delay a scheduled outage that, purportedly, could disrupt the protests (Pleming, 2009). or spies and, thus, triggering retaliation on them (e.g., Morozov, 2012: p. 157; or Kelly et al., 2014: p. 415).
Although that RAND report is epitome of the myth of the Internet as a liberating force it was not the only one. For instance, Hill and Hughes (1999) argued that Usenet newsgroups could be used “as a relatively safe form of political expression against less-democratic, even repressive, regimes” and that Usenet was “absolutely impossible to globally censor”.
Bennahum (1997) claimed that students taking part in the 1996-1997 protests in Serbia had labeled them as the Internet Revolution because the role it was playing, and that “[t]here [in Serbia] the democrats just may win.”
Even Margolis and Resnick (2000: p. 210) who were highly skeptical about the Internet having any real impact on liberal democracies claimed that the Internet would foster democracy in authoritarian regimes. They argued that nations that were “both poor and governed despotically” could compete with richer nations in the information economy thanks to the Internet, or that “the Internet also will help dissident groups in authoritarian societies” because “it is difficult for an authoritarian regime to monopolize the sources of public information”.
Unfortunately, loose political talk in Usenet contesting a repressive regime is not safe; it may be that Usenet cannot be globally censored but it can be blocked in a given country; Serbia did not reach democracy in 1997 but 3 years and one (another) war later; poor nations are not really competing in the information economy; and, as I will show in next chapter, authoritarian regimes are exceedingly imaginative to disrupt Internet and to exploit it to their advantage. In this regard, Augusto Monterroso’s story is painfully accurate: the dinosaur is still here.
Of course, there were discordant voices (e.g., Shapiro, 1999; Kalathil and Boas, 2001) that compellingly argued that there is nothing inherently democratic (or autocratic) in communication technologies, that technologies can be regulated and controlled, or that citizens could eventually be identified online. In short, that “an authoritarian regime can counter the challenge posed by Internet use and even utilize the Internet to extend its reach and authority” (Kalathil and Boas, 2001).
It goes without saying that they have been mostly ignored, or it may be that social media pundits thought that this time it was going to be different. Either way, social media has been claimed to be able to foster democracy on the basis of similar premises to those used earlier for the Internet.

5.4.2 The Iranian protests of 2009 and the Arab Spring

On June 12, 2009 Iran held presidential elections with the reformist Mir-Hossein Mousavi running against the incumbent hardliner Mahmoud Ahmadinejad. The expectations were high since many Iranians hoped for a change in the regime. Following the count, Ahmadinejad was announced as the winner of the elections; Mousavi claimed that the result is a “dangerous charade”; and thousands of his supporters took to the street claiming electoral fraud and clashing with the police.
Two days later, Supreme Leader Khamenei asked for an investigation of those allegations. Still, marches and demonstrations persisted during the following days with massive rallies in Tehran and other cities. The clashes with the regime were bloody and several protesters died; eventually, foreign journalists were banned from covering the demonstrations. Marches and demonstrations went on unabated even though the Supreme Leader had declared the elections legitimate; tension increased, more protesters were killed, hundreds of them arrested and some persons even accused of espionage. From that moment the tension between the Iranian opposition and the government has persisted (Financial Times, 2010; Reuters, 2010).
Later that year, on December 17, 2010, a street vendor from a small city in Tunisia burned himself to death in protest for his cart being confiscated by the police. That event triggered a number of demonstrations and clashes with the regime that spread all over the country; despite the reprisals and deaths, protesters persisted on their actions and by January 14, 2011 President Ben Ali (in office since 1987) fled the country. In the meantime protests had erupted in Algeria, and later they took place in Egypt, Libya and other countries (cf. Dabashi, 2012).
The Arab Spring had been born. Surprisingly, in spite of Tunisia being the only country where the uprisings have produced some advances towards democracy, it was the Egyptian Revolution the one that has grabbed more attention. It started shortly after Ben Ali departed Tunisia: On January 25 a Day of Rage was called and thousands of persons demonstrated against Hosni Mubarak. During the following weeks Tahrir Square in Cairo was occupied by close to a million protesters asking for the resignation of Mubarak; as in Tunisia there were bloody clashes with regime forces and thugs but, in the end, Mubarak resigned.
Those events captured not only the attention of the international public but also their imagination and, as it was aforementioned, they inspired most of the occupation movements that have occurred since 2011.
Still, there was an additional factor that was even more attractive for the public and the press: purportedly, the Iran protests and the Arab Spring had been fueled by the use of social media and, thus, Iran was a “Twitter Revolution” (The Washington Times, 2009) and Egypt a “Facebook Revolution” (Hauslohner, 2011). Of course, that was an oversimplification but the underlying assumption about social media being a trigger factor for democracy still deserves attention—despite being similar to prior claims about the Internet and the Web.
The least that can be said about such a claim regarding Iran is that it was ill-founded.
At the peak of the protests, there were less than 20,000 Twitter users in Iran (Evans, 2009), a country with a population over 70 million. Still, the level of hype in the press had reached the point of leading the U.S. State Department to ask Twitter to delay a scheduled outage so the protests could ensue undisturbed (Pleming, 2009); and a former top adviser for George W. Bush, Mark Pfeifle (2009) asking the Nobel Peace Prize for Twitter and its founders.
What could have made the press missed the shot so badly? It seems that it was a mixture of wishful thinking with a lack of on the ground information. When the protests erupted heavy censorship was exerted on the mainstream media, and foreign journalists were banned from covering the demonstrations. Hence, those foreigners looking for information relied on Twitter which, at the beginning of the protests was not blocked. We may safely assume that Farsi is not widely speak by pundits and, thus, that they relied on the information tweeted by English speaking Iranians. In that case, the only source of information about the protests was a minuscule fraction of the protesters which hardly represented all of the opposition, much less the whole Iranian population (Morozov, 2009a).
One of the first persons attracted by those “news” and spreading them was the blogger Andrew Sullivan (2009). He quickly claimed that the revolution would be twittered, and provided links to Iranian Twitter users so everyone could witness the unfolding of the purported revolution. After Sullivan other news outlets followed and, eventually, American politicians. At that moment, an availability cascade was unfolding in the West but not a revolution in Iran.
The truth is that Iranian Twitter users were too few to trigger the protests, much less to coordinate them; moreover, they were not a representative sample of the population and, thus, their point of view depicted a huge uproar while, actually, a substantial amount of Iranians were perfectly satisfied with the electoral outcome.
Still, the cheerleading approach in the West towards the role of Twitter in the protests badly backfired, although not on the Westerners but on the Iranian tweeps. For instance, simply using social media such as Facebook or Twitter was a matter of suspicion and, besides, all of the information—including pictures and videos—plus the links to other activists were there for the secret services to collect (Morozov, 2009a). Actually, claiming that a tool can overthrow a regime only makes it more paranoid and authoritarian (Morozov, 2010).
With regard to the Arab Spring, some scholars (e.g., Howard et al., 2011; Zhuo et al., 2011) have pointed out that social media were crucial for those uprisings, arguing that they were used as deliberation realms, coordination tools and broadcasting instruments to reach the international community, and that they also helped to spread tactics and knowledge from one country to another. Of course, none of these authors claim that social media were the cause of the Arab Spring but cautiously suggest that they greatly facilitated it and accelerated it.
This cyber-optimistic opinion is not shared by everybody, however. To start with, some authors criticize the mere labeling of those uprisings with the technology used by some of the activists (Mejías, 2010). Others reduced to the role of social media to provide a tactical advantage when coordinating actions (Lynch, 2011; Barrons, 2012). In addition to that, a few suggest that mainstream media (concretely satellite television) were responsible for legitimizing the actions of the protesters and, in turn, amplified and reinforced the events (Alterman, 2011; Lynch, 2011; Nanabhay and Farmanfarmaian, 2011).
Moreover, most authors who limit the impact of social media in the Arab Spring argue that there were longstanding reasons for the uprisings such as unemployment, poverty or growing inequalities (Lynch, 2011; Fuchs, 2012; Lim, 2012), and that political debate did not only precede social media (Wolfsfeld et al., 2013) but that it was being held in other realms in addition to social media.
Of course, nobody says that social media did not play a role in the Arab Spring; however it was nuanced and intermingled with other factors. As a matter of fact, Hussain and Howard (2012) have shown that Internet and mobile phone were key factors in some of the revolts, but always in combination with other variables (mainly economical, cf. Brancati, 2013), and with very different weights depending on the country of interest. That research is consistent with other studies (Faris, 2010; Karagiannopoulos, 2012): actually, if we compare “social media revolts” in countries such as Iran, Egypt, Kenya or Ukraine, it seems clear that the success or failure of the protests is much more related with the context of each country than with social media use.
On top of this, as with the 2009 Iran Protests, the depiction of the Arab Spring as a social media driven event has been more a Western artifact than a reality. For instance, Twitter was much more useful as a source of information for foreigners interested in the events that for national activists (Aday et al., 2013). Moreover, social media activists who were eventually made “spokespersons” of the uprisings for the West media have been criticized for being detached not only of the common people (Gerbaudo, 2012: pp. 70-74) but of the real action of the uprisings:
“Almost all Tahrir was safe, except for the front lines, which witnessed at least two major battles [...] in which hundreds of protesters died. [...] Most of those who fought on the front lines came from the lower classes of Cairo and the provincial areas of Egypt. [...] Westernized, middle-class youth in jeans and T-shirts were tweeting or updating their Facebook status about the battles and taking photos or making videos of the fighters. The whole event was mistakenly then called a Facebook and Twitter revolution to celebrate the Westernized side of it, but none of those who were posting to Facebook or tweeting were severely injured or died in the protests.” (Abul-Magd, 2012)
Finally, I would like to close this section with two important caveats. First, when activists raise against a regime which has got a tight control of the Internet, and which is ruthless enough with its own people, social media usage is more dangerous than useful. Second, if there is a prime example of social media failing to spark democratic reforms it is China: an authoritarian regime comfortably coexisting with social media platforms used by hundreds of millions of citizens, and that does not need to exert any pressure comparable to that of Iran or Syria. Both negative aspects of social media under authoritarian regimes will be discussed at length in the next chapter.

5.5 Social media and terrorism

Finally, this chapter would be incomplete if it did not include terrorism because, simply put, terror acts are “a form of of political communication, intended to send a message to a particular constituency” (McNair, 2011: p. 9). On top of that, it is undeniable that terrorist organizations are using social media to spread their message and, moreover, social media offer terrorists other opportunities beyond mere communication.
Terrorist’s use of the internet is not new and predates social media (Tsfati and Weimann, 2002). Indeed, the main goals of terrorists in the internet still apply to social media; namely: (1) to spread their message to get support, recruit militants, and raise funds; (2) to provide training materials; and (3) to plan and coordinate their actions. Notwithstanding, social media have introduced some novelties in terrorists tactics and strategies.
To start with, video-sharing sites have enormously simplified the release of recordings by their authors, their distribution by supporters, and their access by sympathizers and wanna-be terrorists. Indeed, despite efforts by websites such as YouTube to remove this kind of videos the truth is that they are widely available although highly ephemeral. It is also of interest that plenty of the available videos are either dubbed in English or close-captioned in that language (Weimann, 2009). That is, terrorist groups are targeting global audiences not only to communicate their message, but also to get supporters and recruit militants outside their countries of origin. In fact, there is growing evidence that consumption of terrorist videos is one important source of self-radicalization. This, in turn, drives homegrown terrorism (Conway and McInerney, 2008), particularly so-called “lone wolf” attacks (Weimann, 2012). It is also of interest that most users uploading this kind of videos are not the actual authors (Conway and McInerney, 2008), and that most of the material is produced by a relatively small set of publishers (Klausen et al., 2012).
In addition to video-sharing sites, terrorist groups also rely on SNS. Facebook is quite popular to create groups or pages supporting terrorist organizations; those pages are used to provide information and graphic material, links to material in other platforms (Weimann, 2014), and, to a lesser extent, to create links between sympathizers (befriending). The latter, however, is usually discouraged because it could put at risk the militant network with just one single member getting arrested (Weimann, 2009). Still, it seems that some groups are being used to recruit militants abroad (Farwell, 2014). Finally, Facebook is also very relevant for terrorists as an information gathering tool: they visit the pages of users who are military personnel or police forces to look for potentially sensitive information they may inadvertently release (Weimann, 2009).
It seems to be Twitter, however, one of the most useful SNS for terrorists (e.g., Katz, 2014). According to Weimann (2009) it can be an effective coordination tool [Y]  [Y] Farwell (2014) has casted doubts on such kind of use, at least among ISIS members; he claims that electronic communication is minimum among top leaders and that they rely on “couriers to deliver command-and-control messages by hand”., it allows militants to interact with each other, and it is used to release “official” messages from the terrorist organizations. The latter is particularly worrisome given the importance that mass media are giving to real time coverage; in this sense, Weimann (2014) points out that media “may take tweets as a legitimate news source”. In other words, the terrorists’ message and information are straightforwardly addressed to the audience of mass media. Although unusual there have been also cases where a terrorist attack was “broadcasted” through Twitter (Sullivan, 2014), thus, making of the SNS a terrorizing tool. According to Klausen (2014), Twitter is also a used “to build a transnational community of violent extremism” and as a tool for indoctrination and recruitment of foreign fighters. As with Facebook, links to material hosted in different platforms is also spread using Twitter.
I would like to close this section with one of the most alarming consequences of social media use by terrorist organizations: the possibilities they offer for self-radicalization of sympathizers. Given its broad appeal, social media reach enormous audiences and the kind of materials released (particularly video) are very effective when those exposed are relatively young. What starts at first as mere sympathy can slowly evolve towards online militancy (especially if some terrorist acts as a mentor, which is perfectly possible online) and, in a few but not negligible number of cases, motivate those persons to act in the physical world up to the expectatives they have grown online [Z]  [Z] The interested reader should consult the work by Markus and Nurius (1986) about the Possible Selves theory to better understand the underlying psychological bases on which self-radicalization works. (Brachman and Levine, 2011). In other words, materials (text, images, tutorials, videos) are working 24x7 to engage sympathizers and eventually trigger attacks with little effort from the terrorists.

5.6 Conclusions

Social media is a new tool in the tactic repertoire of contentious politics. However, it is not being used because of the impact it can have on such actions; instead, it is being used because it is an everyday commodity in the lives of activists. Thus, social media is employed if and when it is useful for a given action and ignored when it is not the case. Still, social media affordances can have an impact on contentious actions and on their analysis by scholars and governments.
Activists should be aware that social media makes easier to participate in movements and actions but, at the same time, it fosters an individualistic participation. This can negatively affect the bonding with the rest of participants, dilute (or even inhibit) collective identity and, in the long run, undermine the action or the movement by making it chaotic and unproductive.
With regard to organizational purposes it seems obvious that CMC are essential for transnational movements or actions. However, the larger the action the more important the need to involve multiple pre-existing organizations with on-the-ground knowledge. Indeed, it seems pretty clear that any action solely based on social media is likely to fail even before starting. All the actions and movements described in this chapter that achieved relative success either involved experienced organizations and activists, or call for volunteers to steer local on the ground activities. Still, many of those movements failed to exploit social media to create a fully global movement. Social media helped to spread the ideas and tactics but was not used to coordinate the ensuing debates and produce coherent demands. 15M or Occupy, for instance, were widespread but they operated on a very local and fragmented way.
Researchers aiming to study contentious politics and social movements from digital traces must be extremely cautious. Activists do not have a platform of choice and use whatever best fits their purposes; thus, by just analyzing data from a single platform the picture is not only incomplete but highly skewed. Besides, it is likely that actual activists are relatively quiet in open social media and that most of the material is produced by non-activists and observers detached from the action.
With regard to the use of social media during contentious actions we must separate contention under democratic and under authoritarian rule. When used by social movements operating in relatively democratic regimes, social media can be an extremely powerful tool to get outreach and support, not only regional but even international. Needless to say, by simply deploying a website or using social media a movement is not going to be the next EZLN. The message needs to resonate with as many persons as possible so it is accepted, endorsed and shared.
Social movements operating under highly repressive authoritarian regimes cannot expect to safely rely on social media, particularly on platforms where posts are open to everybody or that can be easily infiltrated. Under such circumstances they are ripe for censorship, harassment, arrests and violence. Of course, an activist in such kind of country should know the risks; however, it may be that some users feel less riskier by blogging or tweeting when it may be extremely dangerous for them and their peers.
With regard to the concrete case of the Arab Spring, social media played a role but a minor one within an ensemble of factors. It did not trigger the protests neither fueled them. Social media was a useful tool for some protesters but not for everyone in those countries. It provided powerful imagery that was used by mainstream media and thus reinforced the protests but, it is Tunisia who is slowly advancing towards democracy and not the telegenic Facebook Revolution of Egypt.
Therefore, cheerleading the purported benefits of a tool to bring democracy when they are not realistic does not benefit anyone, in particular the activists. Doing that when holding office is a disservice. However, it is much more worrisome that democratic officials suggest controlling social media platforms or even shut them down in cases of social unrest; I will debate this trend towards authoritarianism in democratic countries in the next chapter.
Finally, terrorist groups seem to be the movements that are the closer to be leaderless, and their use of the Web and social media as self-radicalization instruments is extremely worrisome because it is extremely efficient. Taking into account what we have learnt from social movements, it seems that the most sensible way of defusing terrorism online is to make its message unattractive; that, however, is not a technological problem but a social and political one. On another hand, social movements could learn how to exploit the underlying psychological basis of self-radicalization to be more effective when recruiting volunteers.

6 Social media and authoritarianism

“Now, there’s no question China has been trying to crack down on the Internet—good luck.
That’s sort of like trying to nail Jello to the wall.”
Bill Clinton
“You can’t stop it. Vladimir Putin can’t stop it, and the Middle East leaders can’t stop it.
Social media is tied to youth and cannot be slowed down. It can help bring democracy.”
Colin Powell
“Them’s my principles; but if you don’t like them—I kin change them!”
The archetypal American legislator
in New Zealand Tablet, 18 October 1873. Volume I, Issue 25, p. 8

6.1 Introduction

In the previous chapter I discussed the purported impact of social media in democratic uprisings—particularly in Iran and later the Arab Spring. On that regard I showed that social media played some role but it was not crucial and, moreover, those ‘revolutions’ fell short of achieving democratic reforms—with the notable exception of Tunisia.
As a matter of fact, authoritarian regimes worldwide have shown that they can be extremely resourceful when tackling with new technologies threatening their power: they can disrupt them, they can exploit them against dissidents, and eventually they can even promote them while at the same time preserving the status quo. Thus, in this chapter I will focus on a number of aspects of the interaction between social media and authoritarian regimes.
First, I will review some basic authoritarian measures regarding social media, such as disruption—e.g., by blocking access or censoring contents—or counter mobilization of dissidents. I will also cover the impact of sympathizers and regime co-opters in social media.
Then, I will study how authoritarian regimes can use social media to monitor their citizens, and to harass and repress dissenting individuals.
Hereafter, I will analyze how different regimes have been able to allow their citizens an apparently free use of social media which is not free at all; while, at the same time, they apply rather limited violence on people.
Finally, a section is devoted to the trends towards online authoritarianism in countries which are not undemocratic; there I will cover the cases of Russia and Turkey, and I will also briefly survey initiatives in purportedly strong democracies which can derive in close to authoritarian behaviors.
For illustrative purposes I will refer to events and legislation from different countries, such as China, Iran, Syria, Russia, Turkey, France, Spain, the United Kingdom and the United States. I have chosen those countries for a number of reasons: China, Iran and Syria are the worst offenders of online freedom; while Russia and Turkey are swiftly degrading it (Kelly et al., 2014); France, Spain, UK and US are liberal democracies that aim to introduce important legislative changes to fight extremism which are extremely similar to methods of online authoritarians. Furthermore, both the US and the UK have engaged in massive surveillance which is also undermining of democratic ideals.
Yet, before discussing those topics we firstly need to understand why certain regimes are violently repressive while others are not because, depending on that, their approaches to exploit social media greatly differ.

6.1.1 Violence is the last refuge of the incompetent

Many of those living in liberal democracies seem to believe that authoritarian regimes are capricious, and that they brutally repress all of the citizens. Those, in turn, are in total opposition to the regime and are willing to overthrow it, provide it they have the necessary resources—and maybe some external help. Yet, reality is much more complex: not all authoritarian regimes are equally arbitrary or openly violent, not all citizens and societal groups oppose their governments to the same degree (or at all), not all of those uncontent are willing to uprise, and not always external resources are a panacea.
Indeed, the stability of authoritarian regimes can be explained by three major processes which can also shed light on the different approaches of such regimes to social media. Those processes are repression, legitimation and co-optation (Gerschewski, 2013).
Repression hardly needs any definition since, as aforementioned, it is commonly associated with authoritarian regimes; however, repression alone does not explain their longevity and, besides, it is very costly (Gerschewski, 2013). Furthermore, repression is to some extent inversely related with the legitimacy of a regime: “States that lack legitimacy devote more resources to maintaining their rule and less to effective governance, which reduces support and makes them vulnerable to overthrow or collapse” (Gilley, 2006b).
Thus, for authoritarian regimes it is on their self-interest to legitimate themselves: their survival chances improve, and their people are easier to govern. Needless to say, winning legitimacy is not easy—and in some cases exceedingly difficult—but not impossible. According to Gilley (2006a) there are three major causal bases for legitimacy [A]  [A] These bases apply for any kind of state, not only authoritarian ones.: good governance, democratic rights, and welfare gains; furthermore, democratic rights are not as crucial as the other two:
“[I]t is notable that democratic rights, while certainly qualifying as one of the most important causes of legitimacy, turn out to be roughly on par with welfare gains, and both of these are far less important than good governance.”
So, it is easy to see how legitimation has been earned by certain authoritarian regimes: they have been able of improving their citizens’ welfare, and they exert “good” governance in the sense of offering public services, keeping order, providing a sometimes reasonable judiciary, and so on. Some of them may even hold elections (Levitsky and Way, 2001) which do not only serve legitimation purposes but also contribute to the third ‘pillar of stability’ (Gerschewski, 2013): co-optation, whose goal is to tie different actors to the regime elites. Other co-optation mechanisms include “patronage, clientelism, and corruption” which are common among the elites ruling the domestic social media platforms.
Finally, many authoritarian countries claim to grant certain rights such as free speech and freedom of assembly to their citizens (e.g. China, Malaysia, or Singapore). Needless to say, such freedoms are greatly limited but they still allow citizenry for certain online deliberation.
Taking all of this into account, we can broadly distinguish two different kinds of authoritarian regimes. On one hand, we have those lacking public legitimacy which need to devote most of their efforts to control their population both offline and online. In these regimes social media—if available at all—is mainly used as a tool for monitoring and repression, and censorship and disinformation are pervasive. On the other hand, we have regimes where repression does not affect the majority of citizens, and legitimation and co-optation processes are much more important. The approach of such regimes to social media is much more complex. Needless to say, some services are blocked and censorship is still used; however, self-regulation by service providers and citizens is preferred; moreover, online deliberation is tolerated to a certain extent, although that and other online behaviors are extensively regulated by law; finally, surveillance, harassment and repression exist but are applied with lower intensity—i.e., a large part of the users is largely unaffected.

6.2 The bases of Internet authoritarianism

No matter the kind of authoritarian regime, there are a number of practices that all of them apply at one time or another to control internet in general and social media in particular; namely: service disruption, content blocking and censorship, counter mobilization of dissenters (usually by means of cronies spreading disinformation), and leveraging of co-opters (e.g., mobilizing regime sympathizers or requiring the collaboration of private companies).

6.2.1 Disruption and blocking, filtering and censorship

The most obvious and straightforward way of avoiding citizenry to use social media of any kind is shutting down the Internet; i.e., denying access to everybody within the whole country or a particular region. It goes without saying that such an action is not trivial and, furthermore, it has also an impact on the regime and on the country’s economy. Because of that, such an approach has been seldom used; still, there have been two prominent cases: the Chinese government shutdown the Internet in Xinjiang Uyghur Autonomous Region for 10 months after ethnic riots (Hogg, 2010); later, during the Egypt revolution of 2011, Mubarak’s government cut off Internet for a few days on late January (Karagiannopoulos, 2012).
Less drastic—and, thus, much more common—is degrading the service to avoid certain uses. For instance, Iran applied bandwidth throttling during the 2009 presidential election protests to avoid dissenters using any Internet service (Karagiannopoulos, 2012; Aday et al., 2015); Syria switched off 3G services sporadically, particularly in the most contentious areas to limit “the ability of dissidents to upload videos of protests to YouTube” (Preston, 2011).
An even more selective approach consists of just blocking the unwanted platforms. For instance, Iran has blocked, among hundreds of websites, Twitter, Facebook (Kelly et al., 2014: p. 414), and blog hosting platforms (Zarwan et al., 2005); Twitter and Facebook are also blocked in China (Bamman et al., 2012; King et al., 2013) in addition to YouTube (Helft, 2009); Turkey—not an authoritarian country yet—has blocked, among other services, Twitter and YouTube (e.g., Kelly et al., 2014: p. 796; or Akkoc, 2015).
Depending on the underlying blocking methods (cf. Deibert et al., 2008) circumvention by users may be more or less feasible. For instance, Cardullo (2015) describes how Turkish users spread such kind of information to avoid Twitter blockages. Needless to say, more sophisticated filtering systems are increasingly difficult to skip and, hence, only the most savvy—and politically motivated users—try to bypass them. According to Roberts et al. (2010) “no more than 3% of Internet users in countries that engage in substantial filtering use circumvention tools.”
Blocking is a brush stroke approach to filtering and, thus, content deemed inappropriate by the regime can still be accessed by users. For instance, Iran dissenters used Goodreads—a website mainly devoted to discuss books—for political discussion until it was blocked (Morozov, 2009a). Such game of cat and mouse between dissenters and the regime is certainly unavoidable; however, authoritarian governments have implemented additional measures to filter particular content even in unblocked sites. Such methods tend to rely on black lists of keywords which are used to avoid access to either URLs or contents containing them (e.g., Zittrain and Edelman, 2003). Their main strength is that they can be adapted on the fly and target topics as they arise; on the other hand, they may require the collaboration of the companies running the different services, and they also require constant supervision to detect new “sensitive” themes.
Finally, authoritarian regimes are not only interested in limiting access to particular pieces of information; they are even more concerned about their citizens producing and, worse, spreading certain types of ideas. Needless to say, by severely limiting access to blogs, microblogs, or video sharing sites a good amount of users are devoid of any chance or producing such contents and releasing them to the international community. Moreover, the same black lists can be used to monitor the available platforms for banned contents and remove them (e.g., Bamman et al., 2012). However, those methods can be easily avoided by changing the vocabulary and, thus, to find and remove certain contents humans (i.e., censors) must enter the loop.
It goes without saying that research about censorship in authoritarian regimes is limited, and it has required a number of educated guesses. Nevertheless, the available literature depicts an interesting picture about which kind of contents and under which circumstances are tolerated, even when they seem sensitive.
Bamman et al. (2012) found that keyword-based content removal greatly depends on the current state of affairs and geographical location. For instance, microblogs by users in the provinces of Tibet and Qinghai undergo a much higher ratio of censorship than those in Shanghai or Beijing. King et al. (2013) provided further details by analyzing the contents of microposts beyond their keywords. Their results are quite surprising: they argue that messages extremely critic of the state and its leaders are no heavily censored; according to them censorship is actually targeted at “curtailing collective action by silencing comments that represent, reinforce, or spur social mobilization, regardless of content.” Given the importance of such findings, and its idiosyncratic relation with Chinese online practices, we will back to them in the section devoted to networked authoritarianism.
Still, one could think that the job of censors is daunting—particularly when facing contents that are avoiding the blacklisted keywords—and thus unproductive. Quite to the contrary, the truth is that the sharing culture of social media is actually helping censorship (Morozov, 2012: pp. 99-10). For instance, if a number of known dissidents link to a particular content or profile, censors may assume that the target is suspicious and it deserves further inspection; on another hand, “people who are unpopular probably don’t even deserve to be censored; left to their own devices and nearly zero readers, they will run out of blogging energy in a month or so” (p. 167).

6.2.2 Sympathizers, counter mobilization and collaborators

Even the most cruel regimes have supporters, and those enjoying certain legitimacy have much larger numbers of sympathizers. This means that social media is not only used by those contesting the regime—who are greatly limited because of filtering and censorship—but also by those who agree with the government’s policies and may feel the need to defend them.
For instance, Aday et al. (2010) show how the Iranian government encourage its sympathizers to be active in social media promoting its policies and, in addition to that, “fram[ing] its opponents’ use of the Internet and other media as reflecting their domination by foreign interests”. Actually, the act of vilifying the mere use of foreign social media is extremely powerful and not unique of Iran (Aday et al., 2010; Morillon and Julliard, 2010: p. 18); in fact, it has been observed in Venezuela (Rich, 2011), Azerbaijan (Pearce and Kendzior, 2012), or Turkey (Moore, 2015). It shows, again, that authoritarian regimes prefer their citizens to exercise self-control with regard to social media than controlling its use.
Moreover, this kind of counter mobilization actions serves a number of additional purposes such as reframing the discourse of the opposition—and not only their use of foreign services; spreading disinformation; influencing the local and international public opinion; and demobilizing the dissenters. Given the importance of counter argumentation in social media, it is hardly surprising to find that authoritarian regimes rely on paid staff and automated tools to conduct such campaigns and, thus, overcome the opposition.
Actually, the governments of China (e.g., Fung, 2002), Iran (Karagiannopoulos, 2012), Syria (Khamis et al., 2012), or Russia (Alexander, 2015; or Toler, 2015) have infiltrated chat rooms, social networking sites, or created blogs in an attempt to tip the balance of public opinion to their advantage. In some cases, such as Syria, they have even impersonated actual dissidents to make disinformation more credible, but also to capture other activists (Khamis et al., 2012).
Another kind of action that is increasingly common among supporters of authoritarian regimes are Distributed Denial-of-Service (DDoS) attacks. Such attacks disrupt services beyond the borders of the instigator and, hence, their main goal is not blocking access for nationals, but showing force and, sometimes, try to force the victims to change their behavior (e.g., attacking an online journal or a blogger to stop discussing a given topic). A well-known example is the series of attacks on LiveJournal, WordPress, Facebook, and Twitter that started in late-2008 and peaked on August 2009. Eventually, it was found that the target were not the services themselves but just one single user: a Georgian blogger writing under the screen name of Cyxymu who is particularly critical with the Russian involvement in Georgia (Morozov, 2009c; Doernberg, 2010). However, it goes without saying that in most DDoS attacks the perpetrators and their underlying motives can only be speculated.
Finally, we must note that many of the aforementioned measures—such as blocking, keyword filtering, or content removal—require the active collaboration of private companies—some of them multinationals headquartered in liberal democracies. Such collaboration takes place on the grounds of legality (Kalathil and Boas, 2010) but, the truth is, that foreign companies could chose not entering a given market; in other words, many providers of online services have chosen profit over principles (MacKinnon et al., 2006).
Nevertheless, and to their credit, the most popular social media platforms are blocked in China and Iran [B]  [B] At the moment of this writing Blogger, Facebook, Twitter, and YouTube are blocked in both China and Iran. Tumblr and Flickr are blocked in China but not in Iran, while MySpace and WordPress are blocked in Iran but not in China. Quite surprisingly LiveJournal is not blocked in none of those countries., so we may assume they have not yielded to those governments’ demands. Unfortunately, domestic services such as VK (a Russian SNS similar to Facebook) or Sina Weibo (a Chinese microblogging site) are much more dependant of their respective markets, they have much closer ties with their respective local elites and, thus, they are legally pressured to comply with censorship and surveillance policies (Kalathil and Boas, 2010). Such kind of collaboration implies that state censorship has been effectively privatized (MacKinnon et al., 2006); this has increased arbitrariness (MacKinnon, 2009) because the degree of censorship depends on each company and its fear of fines and reprisals.
This state of affairs is deeply worrisome not only because it is currently affecting the lives of millions of people but because of its future impact. As MacKinnon et al. (2006) put it: if private companies “actively collaborate with political censorship in China, it will be difficult for them to turn down similar requests made by other governments seeking to control their citizens.” Given the measures that governments worldwide are adopting because of their war on terror, the behavior of private companies can be unsettling with regard to current democracies. I will cover such issues in the closing section of this chapter.

6.2.3 Social media as an instrument of surveillance and repression

In previous section we have seen how authoritarian governments try to avoid citizens interacting with information that is considered a threat to the status quo; that calls to action tend to be heavily censored; and how sympathizers and cronies conduct counter mobilization.
All of such measures try to disrupt the online actions of dissenters but it goes without saying that authoritarians do not stop there. If some individual or group persist on their attempt to access, produce or share banned contents, or they attempt to organize collective actions aim at changing the regime, they are certain to suffer some kind of retaliation.
The Iranian blogosphere is an epitome of such a situation; it quickly grew during the 2000s and it reached a degree of freedom unparalleled by traditional media and, of course, the offline public sphere in Iran. Indeed, it still holds “opinions that the country’s newspapers and other media would never run” (Zarwan et al., 2015). Because of that, a huge amount of blogs and websites have been closed (Aday et al., 2015), and online journalists and bloggers are continuously intimidated and some of them imprisoned (Rahimi, 2003; Zarwan et al., 2015). However, such kind of intimidation is not unique of Iran and it is also commonly suffered by bloggers in Syria (Zarwan et al., 2015) and other countries.
Obviously, security forces do not find dissenting bloggers by accident, they actively look for them. The blogosphere and other social media services are thoroughly monitored in Iran—particularly after the 2009 protests (Christensen, 2009; Karagiannopoulos, 2012)—and China, but other countries—such as Bahrain, Egypt, Kazakhstan, Turkey or Uganda—are implementing similar systems or plan to do it (Kelly et al., 2014).
However, authoritarian regimes are not only using social media to monitor online dissenting behavior, they are exploiting social media to police offline actions, too. Morozov is particularly pessimistic in this regard arguing that “both Twitter and Facebook give Iran’s secret services superb platforms for gathering open source intelligence about the future revolutionaries” (2009b). Similar approaches were used in Egypt (Gallagher, 2011), or Belarus (Morozov, 2009a). In this later case the tracking of protesters in LiveJournal helped the regime to the point of police appearing at the demonstration sites before them. Once there, protesters were photographed, and using those pictures and others obtained from social media they were identified and, later, interrogated and threatened. A similar approach was followed by supporters of the Iranian regime who crowdsourced the identification of street protesters in social media photos for their detention (Aday et al., 2010; Morozov, 2010).
In addition to those approaches to identification, the same methods that Morozov described to find censurable contents can be used to find central individuals within the opposition; he says: “If tweets of a particular user are retweeted more often than average, it’s a good idea for the government to start watching that individual closely and learn more about his or her social network.” (2012: p. 167). This approach also allows security forces to find dissenters abroad and, thus, extend the regime threats to those living in exile (Morozov, 2010 and 2012; Karagiannopoulos, 2012).
The truth is that at this moment social media makes surveillance tasks so simple that it is increasingly risky for dissenters to use them (Karagiannopoulos, 2012). Besides, the pressure of feeling constantly tracked, joined to the threats and detentions tends to greatly demobilize the opposition (Morozov, 2009a; Karagiannopoulos, 2012; Pearce and Kendzior, 2012).

6.2.4 Social media under networked authoritarianism

So, virtually all authoritarian regimes disrupt access to online services, block certain platforms, filter foreign content, censor content published by their citizens, subject them to surveillance, and gather intelligence from social media to threaten some dissenters, and detain and imprison others. In addition to these actions, a few regimes feature additional traits that provide them a semblance of legitimacy, and their citizens a caricature of freedom. Such regimes—whose prime example is China—implement the so-called networked authoritarianism which MacKinnon (2010) described this way:
“In the networked authoritarian state [...] a wide range of conversations about the country’s problems nonetheless rage on websites and social networking services. The government follows online chatter, and sometimes people are even able to use the Internet to call attention to social problems or injustices, and even manage to have an impact on government policies. As a result, the average person with Internet or mobile access has a much greater sense of freedom—and may even feel like they have the ability to speak and be heard—in ways that weren’t possible under classic authoritarianism. At the same time, in the networked authoritarian state there is no guarantee of individual rights and freedoms. People go to jail when the powers-that-be decide they are too much of a threat—and there’s nothing anybody can do about it. Truly competitive, free and fair elections do not happen. The courts and the legal system are tools of the ruling party.”
The truth is that literature discussing social media use in China describes a number of cases where netizens were able to raise their voice, report misconduct by government officials, and get their demands satisfied.
For instance, Diamond (2010) describes how the death of Sun Zhigang while detained sparked a wave of indignation in Chinese Internet; that eventually lead to a formal investigation and to the eventual sentencing of those responsible of his death.
Esarey and Qiang (2008) recount the Xiamen protests of 2007 against the construction of a chemical factory sparked by a blogger and which lead to the termination of the project. These protests were not unique, and Qiang (2011) describes other events which triggered collective online protests in China.
Esarey and Qiang (2008) also reported how Chinese users are able to criticize governmental policies and even corruption of officials in online forums without being censored. Indeed, Chinese social media brims with political deliberation although with certain idiosyncratic features (Jiang, 2010) that have prompted the name of “authoritarian deliberation” (He, 2006).
This kind of deliberation is rational-critical debate held mainly among citizens, but also between citizens and officials, which takes place under an authoritarian regime. It is not aimed to influence the major policies but, still, it can exert real impact on local governance. Hence, by allowing the discussion of certain topics an authoritarian regime can obtain some of the benefits of deliberation while not introducing actual democratic practices. For instance, deliberation may help to “solve difficult issues and delegate power” and it “reduce[s] social conflicts and the level of opposition” (Jiang, 2010).
So, Chinese netizens are able to discuss certain political issues, voice their concerns and criticism without facing severe censorship, and even organize campaigns and demonstrations; could it be that democracy is emerging in China because of social media? Unfortunately, the answer is ‘no’.
First, all of these actions fall short of liberal democracy standards. For instance, deliberation cannot be held on every conceivable topic. It is true that the Chinese State Council claims that “[t]he Chinese government [...] guarantees the citizens’ freedom of speech on the Internet”; however, it also enumerates a long list of contents which cannot be produced or spread such as those “subverting state power and jeopardizing national unification; damaging state honor and interests; [...] spreading rumors, disrupting social order and stability; [...] humiliating or slandering others, trespassing on the lawful rights and interests of others; and other contents forbidden by laws and administrative regulations.” (SCIO, 2010).
When criticism is voiced it is disguised in humor [C]  [C] We must not dismiss the use of humor: there are situations in which humor is the only available expressive tool, and it also implies the existence of an underlying collective identity (‘t Hart, 2007). and vague language, without directly attacking the state and its policies. Actually, bloggers tend to avoid risks, and there is still repression for “those criticizing the state or state policy directly, [...] advocating mass political action, or [...] airing views that openly conflict with party ideology” (Esarey and Qiang, 2008).
The collective actions described in the literature are mostly anecdotal and, moreover, they are far from being social movements. Cai (2010: pp. 184-186) has labeled them as “collective resistance” which is characterized by its short-life, and raising demands that are nonpolitical and non threatening for the state. Moreover, we must remember that King et al. (2013) found that calls to action tend to be censored in China, something that Esarey and Qiang (2008) agree on. Arguably, the documented demonstrations and protests triggered in Chinese social media are the exception which proves the rule.
The second line of argumentation against these behaviors being the prelude to democracy is that all of them are far more useful to the regime than to their citizens.
For instance, reporting corruption in social media serves two purposes: on one hand, to detect and remove problematic officials that otherwise would be unaccountable to the regime (Hassid, 2012); on the other hand, to earn legitimacy. After all, excessive corruption is delegitimizing and, at the same time, “a regime where scandals occur is not totally illegitimate” because it implies some freedom of speech (Dogan, 1992).
Authoritarian deliberation and self-expression also provide some legitimacy to the regime given a façade of democratic practices while, at the same time, being a “safety valve” (Esarey and Qiang, 2008; Hassid, 2012).
Of course, when users discuss topics that the authorities deemed inappropriate, their criticism is too harsh or it targets top level officials, or their call to actions are considered a threat, repression is assured. Actually, according to Kelly et al. (2014) “China has imprisoned more internet users than any other nation”. With regard to social media, the grip is very tight, particularly after passing a law targeted at SNS which establish “5,000 views or 500 reposts as a new threshold for prosecuting false, defamatory, or ‘harmful’ comments online”. Because of this hundreds of Weibo users have been detained and a few top bloggers indicted.
Under these circumstances users of social media dare to express themselves, even in critical terms, but they want to avoid repression which, although not entirely random, still exhibits great arbitrariness. As a consequence, the users are indirectly encouraged to exercise self-censorship and effectively deterred from pushing the limits of which is acceptable. As Kalathil and Boas (2010) explain:
“Such measures reduce the need for authoritarian governments to explicitly control the Internet with firewalls and other measures. Many casual web surfers in authoritarian countries understand the boundaries of politically acceptable Internet use, and authorities exploit this understanding to create an environment where comprehensive censorship is not necessary.”
To sum up, networked authoritarians obtain the best of public discussion while, at the same time, they numb meaningful political deliberation and defuse real threats to the regime. This approach is not only applied in China, but also in Singapore (Rodan, 2003; Kalathil and Boas, 2010) or Azerbaijan (Pearce and Kendzior, 2012). In fact, most Asian and post-Soviet authoritarian regimes seem interested in approaches similar to those of China and Singapore

6.3 Authoritarianism is trending on social media

At the moment of this writing democracy is not specially healthy; if we pay attention to the Democracy Index developed by The Economist Intelligence Unit (2014) during the years 2006, 2008, and 2010-2014 we find plenty of reasons for concern.
For instance, countries such as the United States, the United Kingdom, or my own, Spain, are close to the bottom of the full democracy [D]  [D] In full democracies civil liberties are granted and they have got additional traits that boost democracy: good governance, independent and diverse media, and independent and strong judiciary (The Economist Intelligence Unit, 2014: p. 38). category and they are in the brink of being flawed democracies [E]  [E] Flawed democracies hold free and fair elections and grant basic civil liberties. However, they suffer important problems in areas such as governance, political culture, and political participation (ibid).. As a matter of fact, France was labeled as such in the period 2010-2013, and Italy in all of the reports except for 2008 when it scored as full democracy.
Moreover, the scores for many countries and the averages of most regions have declined during the last years and, in some cases, the fall is dramatic. For instance, Russia has degraded from being an hybrid regime [F]  [F] In hybrid regimes elections are not free nor fair because of substantial irregularities. Pressure and even harassment on the opposition and on the media are common. Corruption is general. Weaknesses of flawed democracies are exacerbated (ibid). to be an authoritarian [G]  [G] In authoritarian regimes there is little or none political pluralism, media and the judiciary are not independent, civil liberties are not granted, elections if held are neither free nor fair, censorship and repression are common, democratic institutions (if any) are purely decorative (ibid). one, Ukraine has degraded from being a flawed democracy to be an hybrid regime. Other countries while still not changing category exhibit worrisome trends; for example, Turkey, Hungary, or Greece.
The causes of such trends vary from country to country but the consequences are obvious: democracy is being eroded and it is affecting—in some places faster, in other slower—civil liberties, including online behaviors such as the use of social media. This, conjoined with the global economic crisis, the corresponding contentious response of the people and rise of populist parties, and the ongoing war on terror, means that many governments have increased their interference over Internet in general and social media in particular. In this section I will cover the cases of Russia and Turkey, and I will give a glimpse on a number of worrisome laws and initiatives in Western democracies.

6.3.1 The case of Russia

The first Democracy Index by The Economist Intelligence Unit (Kekik, 2007), which studied the health of democracies in 2006, devoted one paragraph to Russia:
“[A]t present a hybrid regime, with a trend towards curtailment of media and other civil liberties. A potentially highly flawed parliamentary election at the end of 2007 would reflect a further intensification of the country’s apparent slide in an authoritarian direction.” (p. 6)
The succeeding reports increasingly devoted more attention to the changes in Russia. First, they noted that those trends were not only increasing, but that “[m]ost Russians appear unperturbed by the trend towards authoritarianism” (The Economist Intelligence Unit, 2008: p. 10). Later, they confirmed the definite authoritarian nature of Russian regime on the grounds “of the cynical decision by Vladimir Putin to return to the presidency and because of deeply flawed parliamentary elections” (The Economist Intelligence Unit, 2011: p. 10). Finally, in their last report—at the moment of this writing—they labeled Russia as a “Putocracy” and summarized the main traits of that variety of personalistic authoritarianism. With regard to its approach to online media the report says:
“Recent legislation requires information technology (IT) companies to hold Russian user data in servers based in the country, giving law enforcement greater powers of surveillance over citizens’ communications. The authorities have further tightened their control of the media, restructuring state-owned outlets, replacing the management of popular online-news services, and limiting foreign ownership. Alexei Navalny, Russia’s most prominent opposition politician, was given a five-year suspended prison sentence for alleged embezzlement in 2013 and placed under house arrest in February 2014.” (The Economist Intelligence Unit, 2014: p. 24).
Indeed, legislation about setting data centers in Russia seems particularly targeted at SNS such as Twitter and Facebook. The online news service whose management was replaced was VK—a SNS similar to Facebook. VK’s founder refused to provide data about users to the Russian government, and he was shortly after dismissed by shareholders close to Putin. Finally, Alexei Navalny reached political prominence because of his blog in LiveJournal where he was extremely critic of Putin. Hence, Putin’s Russia is not merely attacking information technologies but concretely the social media ecosystem.
This trend towards outright authoritarianism has been also documented in a series of reports by the Berkman Research Center, and supported by additional sources. Thus, at early 2010 the Russian blogosphere was described as “a free and open space for Russians of all political stripes to discuss politics, criticize or support government, fight corrupt practices and officials, and to mobilize others around political and social causes” (Etling et al., 2010). However, the same report also warned that should the blogosphere remain as an active watchdog of the state, there would be “greater government involvement and contestation in the networked public sphere.”
Two years later the situation seemed very similar (Alexanyan et al., 2012): the blogosphere was extremely active, still apparently unfettered by the government, and the only alternative to controlled media to (1) get objective information, and (2) give a voice to those opposing the government. However, they also reported (1) that netizens believed that online surveillance was pervasive, (2) that “offline attacks and threats against journalists and others critical” of the regime and its co-opters were common, and (3) that DDoS attacks against the online platforms of the opposition were frequent. They also theorized that should Putin win the 2012 presidential election the situation could worsen.
Indeed, Putin won the election and the situation did worsen, mainly as a result of increasingly restrictive regulations for Internet services (Tselikov, 2014; Kelly et al., 2014). Thus, websites can be blocked or shut down without judicial oversight for a variety of infractions—e.g., hosting allegedly “extremist” content or calling for “illegal” protests; bloggers with more than 3,000 daily visits must be registered at the federal government; and security forces have got more legal instruments to require access to user data, including the obligation for companies storing Russian users’ data to have data centers at Russian soil.
Furthermore, government attempts to spread disinformation in social media are evident at this moment. For instance, using Twitter bots to spread pro-government rhetoric (Alexander, 2015) and setting up fake websites that pretend to be Ukrainian and, thus, reframe that conflict (Toler, 2015). Finally, a system to monitor social media and detect the organization of unsanctioned demonstrations and other acts of “disobedience” has been recently announced by a think tank close to the government (Lokot, 2015).
So, in short, social media in Russia was a realm of deliberation mainly used by the opposition to denounce the regime’s abuses and organize a response; in less than a decade it has been mostly crushed down with a variety of measures of technical, legal, and also illegal nature. At this moment, social media users opposing the Russian government are subject to surveillance, threaten and attacked both online and offline, their voice is muzzled, and their attempts to collectively organize hindered.

6.3.2 The case of Turkey

The first mention about Turkey in the Democracy Index (apart of appearing in tables) was in 2010: it was reported as one of three Western European countries whose media freedom score had worsen in the 2008-2010 period, the others being France and Italy (The Economist Intelligence Unit, 2010). No further mentions occurred until the 2014 report (The Economist Intelligence Unit, 2014) where it was devoted a whole sidebar. As with Russia, one major problem of Turkey was the concentration of power in one single person—the president and former prime minister, Recep Tayyip Erdogan. In contrast to Russia the report does not cover the many attacks that Erdogan’s government has addressed to social media.
The truth is that until very recently the situation in Turkey was not as bad as in Russia: website blocking had required court orders that were somewhat substantiated, although the reasons were ad hoc. Moreover, Turkish higher courts had ruled a number of times against concrete bans arguing they were illegal, and even the president Abdullah Gul criticized the ban of YouTube—surprisingly using Twitter—while Erdogan was prime minister.
Yet, this does not mean that attacks have been sporadic or ineffective; actually, blocking of social media sites have been taking place in Turkey at least from 2007, and it has addressed YouTube, Google Sites, Blogger, WordPress, and quite recently Twitter—particularly after the Gezi protests (Kelly et al., 2014: p. 796). Bans have lasted from days to months, and in some cases (as Google Sites) years. Needless to say, users have tried to circumvent the bans (e.g., Cardullo, 2015), and social media companies have appealed the measures (e.g., Gadde, 2014). However, circumvention is increasingly difficult and, after all, the goal is to demonize social media so most citizens simply avoid it; furthermore, appeals are mostly pointless and companies eventually comply with the requests to be unblocked [H]  [H] For instance, Twitter ban of 2014 was lifted after the the suspension of those accounts that had released audio recordings revealing the corruption of close associates of Erdogan (Sezer, 2015)..
All of this reveals that Erdogan is extremely concerned about social media being used as a watchdog against the abuses and corruption of his regime, and also as a realm for collective organization (Orucoglu, 2015). Indeed, he has even “vowed to «wipe out Twitter» and called social media the «worst menace to society»” (Kelly et al., 2014: p. 6). Such ideology is clearly present in a leaked report from the Turkish National Security Council which considers social media a threat to Turkish domestic security comparable to that of terrorism (Moore, 2015).
Furthermore, the recent amendments of Law No. 5397 (the so-called 2007 Internet Act) have made much easier to block websites, puts greater burdens on the service providers regarding their liability for hosting content, and extends the compulsory length of data retention. Other regulation allows security forces to access users data without court orders and provides broader and vaguer reasons for surveillance (Kelly et al., 2014: p. 796).
All of that, joined to the detention and imprisonment of social media users accused of defamation or religious offense (Kelly et al., 2014: p. 796), the intimidation of journalists and bloggers (Orucoglu, 2015), and the creation of a so-called “social media army” with hundreds of experts and tens of thousands of automated accounts (Orucoglu, 2015) means that, at least with regard to social media, Turkey has effectively transformed into an authoritarian regime.

6.3.3 The case of “full democracies”

This trend is not unique to Russia or Turkey; other countries, even some which are considered full democracies, have implemented or are studying similarly worrisome approaches on the grounds of fighting terrorism and other malaises affecting social media.
In South Korea a real name verification law was passed on 2007 and put into effect in 2012; it required of users posting in popular websites and newspapers to previously register their actual names. The law effectively encourage self-censorship of many users (Cho et al., 2012) and it was in force during 5 years(!) before it was declared unconstitutional.
The United Kingdom has proposed to ban individuals from broadcasting contents, including posts in social media, by means of so-called Extremist Disruption Orders (Travis, 2014; Wintour, 2015).
On 2014, the French legislature passed the Law 2014-1353 strengthening the provisions relating to the fight against terrorism. Like similar regulations in other countries, this one is also a source of concern for social media users. For instance, using social media is an aggravating circumstance if facing charges of encouraging or glorifying terrorism. The mere consultation of online sites hosting information to conduct terrorist acts or glorifying them, can be considered a terrorist act. Finally, the law allows administrative authorities to ask service providers to block access to sites hosting such kind of materials.
In Spain, the two major parties have recently agreed to modify the Criminal Code, purportedly to fight extremism and terrorism (Morris et al., 2015), but the amendments are vague enough so that virtually every form of pressure on public authorities, even using online media, could be considered a form of terrorism (Sánchez Almeida, 2015). Also in Spain, the ruling party has amended the Law on Public Security effectively restricting both freedom of assembly and of speech, having a strong impact on social media:
“Making a statement on social media that could be ‘perceived’ as inciting others to commit violent attacks will be outlawed, even if the statement cannot be directly linked to an act of violence.” (Amnesty International, 2015).
Such legislation implies that Spaniards (or at least certain groups and individuals) are to be subject to online monitoring. Actually, the Spanish government has recently solicited from private contractors the development of a system to detect and profile potential extremists and terrorists by combining different kinds of information, including a great variety of social media platforms including Twitter, web forums or blogs (Águeda, 2015).
It goes without saying that Spain is not precisely spearheading social media monitoring, not even among full democracies. Indeed, obtaining open source intelligence from social media, particularly SNS, is not precisely new, although it has been only recently when it has received a name of its own: SOCMINT (Omand et al., 2012). The work by Omand et al. compellingly describes the attractiveness of SOCMINT and its technical feasibility; it also tries to address the ethical and legal constraints that the gathering of such kind of information should face. They deserve the merit of clearly exposing that SOCMINT in democratic countries is, to say the least, problematic. However, their arguments on that regard are rather weak.
For instance, they do not ask governments to determine whether SOCMINT is legal or moral in a liberal democracy, instead they offer the following advice (emphasis mine):
“To meet the challenge of legitimacy, the public must broadly understand and accept why, when and with what restrictions SOCMINT is undertaken.”
It is also problematic that the ethical dilemma raised by open SOCMINT is not faced by what they call intrusive SOCMINT [I]  [I] According to Omand et al. (2012) open non-intrusive SOCMINT “should not be able to identify individuals, be used as a means of criminal investigation, or puncture the privacy wishes of the user”. Any SOCMINT not working under such assumptions would be intrusive SOCMINT.. They argue that intrusive SOCMINT would be conducted under the legal protection of the different laws regarding “national security and the prevention and detection of crime” which, they insist, are well accepted by the public.
That line of argumentation is questionable. Certainly, these issues are not a concern for most of the society in democracies but, still, there are groups that strongly disagree with the supposed “sound ethical principles” underlying electronic surveillance and social media monitoring. Indeed, the scarce information that has reached the public about the nature of mass surveillance systems such as PRISM, UPSTREAM or TEMPORA (King, 2014) makes hard to believe that SOCMINT of any kind will be founded on legal, moral and ethical principles.

6.4 Conclusions

So, in short, social media is a two-edged sword for activists and dissenters. On one hand it provides them with tools to reach an audience and organize themselves. On the other hand it can be an extremely useful instrument for surveillance and repression.
Governments dispose of a great variety of tools of legal and technical nature to disrupt access to the services, filter the accessed content, and even remove contents produced by their citizens. In addition to that, sophisticated regimes can monitor, disrupt and repress the opposing minority while, at the same time, follow a “softer” approach for the rest of the population which could eventually believe that it is enjoying some level of freedom and democratic discussion.
These authoritarian manners towards social media are pervasive in non democratic countries. Unfortunately, there is a worrisome trend in countries which are not strictly undemocratic, and even in full democracies, to pass legislation and implement technical measures whose purported aim is to fight terrorism, but which are a clear threat to anyone engaging in contentious politics.
Unquestionably, free speech, internet and social media should not be gamed to the advantage of extremists and terrorists. However, it is sometimes difficult to ascertain if a given individual or a group are such, or they are simply labelled that way by supporters of the status quo.
It should not be difficult for the reader to imagine a new version of Martin Luther King labeled in a dystopian near future as an extremist and, thus, banned from social media in most countries, and even facing charges of terrorism in many of them. I acknowledge that this fiction corresponds to a maximalist interpretation of such regulations; I also know that many hope that if such a thing happened, the public would raise their voice against the abuse, and social media operators would refuse to cooperate. I must disagree with such optimistic interpretation.
To start with, most citizens feel largely unaffected by this kind of situations. Actually, users in Russia or China (Fallows, 2008), for instance, seem to be relatively happy with their internet; and the most common response to the ever increasingly stringent limits to privacy in Western countries is the fallacious “I’ve got nothing to hide” (Solove, 2007).
On the other hand, private owners of social media platforms have already shown that their challenges to unjust legal systems are ineffective, and that they eventually comply with such requests. If they conduct themselves in such a way in countries where they have a relatively minor physical presence, it is implausible to think that they would bluntly reject the advances of the countries where they are headquartered.
Finally, this situation is disturbing because advanced democracies are losing legitimacy as champions of liberty. They simply cannot pass laws to exploit social media to monitor and repress their own protesters in ways strikingly similar to those of authoritarian regimes [J]  [J] I left as an exercise for the reader to find the differences between the Russian regulation about hosting extremist contents and calls for unsanctioned rallies, and the UK Extremist Disruption Orders, the French Law 2014-1353, and the reform of the Spanish Law on Public Security which has worried Amnesty International., and then ask those countries to respect the rights of their citizens (Morozov, 2012: p. 220). Actually, this loss of legitimacy, combined with the exceptional measures to address online extremism and terrorism, badly bode for the prospects of liberal democracy: after all, we have already seen the relatively fast pace in which hybrid regimes can deteriorate and eventually transform into authoritarian ones.

Concluding remarks

“Technology should not try to produce an appropriate democracy;
democracy should try to produce an appropriate technology.
Experts in technology are not experts in the appropriate public uses of technology.”
Benjamin R. Barber

The current state of politics and social media

Along the prior chapters I explored the different aspects of social media and politics, explicitly stating the different issues and pointing out the cases where excessive optimism backfired. A summary may be of interest but, instead of organizing it in the fashion of the chapters, I think it is much more useful to structure it according to the agents that are affected by (or creating) the different problems, namely: citizens, corporate and political elites, researchers, and terrorists.
Citizens are using social media for conventional and contentious political participation, both individually and collectively. At the moment of this writing the main issues affecting citizens political participation in social media are slacktivism [K]  [K] “Beginning with ‘S’, political participation—Slacktivism?” (in chapter 1 Political participation)., lack of meaningful political debate [L]  [L] “Is democratic deliberation taking place in social media?” (in chapter 3 Public opinion)., biases in participation [M]  [M] “The normalization scenario” (in chapter 1 Political participation)., and diluted collective identities [N]  [N] “Collective and connective action” (in chapter 5 Contentious politics).. All of these issues can be traced, in part at least, to three reasons: the affordances of current social media, its private ownership, and its perception by the public.
Corporate elites have colonized social media and they have commoditized most of the platforms [O]  [O] “Media socialia. Cui bono?” (in chapter 3 Public opinion).. That greatly affects the affordances on which users rely and, thus, political actions are using social media in spite of that, and not thanks to it. The commoditization of social media means that the interfaces that citizens are using are not the best to conduct meaningful debate, they may be fostering slacktivism, and they encourage individualistic participation [P]  [P] “Public opinion in social media is complicated” (in chapter 3 Public opinion). that in the long run weakens social movements [Q]  [Q] “Collective and connective action” (in chapter 5 Contentious politics)..
Moreover, corporate ownership of social media means that the monitoring and censorship that citizens in authoritarian regimes suffer has been privatized [R]  [R] “Sympathizers, counter mobilization and collaborators” (in chapter 6 Social media and authoritarianism). and it is being exerted by companies that are, for the most part, based on democratic countries. This situation is on one hand hypocritical and on another one worrisome for citizens in democratic countries. After all, we are witnessing in democratic countries an increasing trend towards authoritarian measures [S]  [S] “Authoritarianism is trending on social media” (in chapter 6 Social media and authoritarianism). regarding social media—of course on the basis of fighting terrorism, and perhaps social unrest. If social media companies tend to comply with authoritarian countries where they just have branch offices, what response can we expect in the countries where they are headquartered?
Political elites in democratic countries are using social media mainly for broadcasting and propaganda [T]  [T] “Mainstream political parties” and “Citizens. This is government” (both in chapter 2 Political actors). purposes, particularly during electoral campaigns [U]  [U] “Social media in electoral campaigns” (in chapter 4 Elections).. For their most part they try to keep a façade of interaction with citizens but it tends to be ineffective to get meaningful feedback from them. Although Q&A sessions, virtual town halls and petitioning systems are laudable, some of those actions are more spectacular than useful, while others—e.g., e-petitioning systems or any other crowdsourced approach—are prone to manipulation by committed small groups [V]  [V] “Government. This is the people” (in chapter 2 Political actors)..
With regard to political elites in authoritarian regimes they have two main approaches. The most repressive countries tend to block access to main social media platforms [W]  [W] “Disruption and blocking, filtering and censorship” (in chapter 6 Social media and authoritarianism)., censor contents [X]  [X] Ibid. and actively monitor their citizens looking for dissenters [Y]  [Y] “Social media as an instrument of surveillance and repression” (in chapter 6 Social media and authoritarianism).. Other countries allow certain social media platforms—usually owned by local companies—and some amount of political comment. That approach is useful for a number of reasons [Z]  [Z] “Social media under networked authoritarianism” (in chapter 6 Social media and authoritarianism).: it acts as a safety valve to release public pressure; it can be useful to detect cases of corruption in officials that, otherwise, would be virtually unaccountable; and, in turn, it earns the regime a certain amount of legitimacy among their citizens. Except for a minority, most citizens in networked authoritarian countries are satisfied with the freedom they enjoy in social media [A]  [A] “Conclusions” (in chapter 6 Social media and authoritarianism)..
The postures hold by political elites in democratic countries about the approach of authoritarian regimes towards social media ranges from the hypocritical to the counterproductive.
On one hand, it is hypocritical to ask for free access to the Internet and social media platforms in authoritarian regimes while monitoring their own citizens or discussing ways to penalize the political use of social media in their own countries [B]  [B] “Authoritarianism is trending on social media: The case of «full democracies»” (in chapter 6 Social media and authoritarianism).. It is even more sanctimonious when actions to restrict the freedom of users in authoritarian countries tend to be conducted by Western companies.
On another hand, claiming that social media on its own is able to overthrow a regime is unrealistic and, at the same time, dangerous for dissenters using it [C]  [C] “The Iranian protests of 2009 and the Arab Spring” (in chapter 5 Contentious politics).. On top of that, previously mentioned networked authoritarianism [D]  [D] “Social media under networked authoritarianism” (in chapter 6 Social media and authoritarianism). is the living proof that technology on its own does not bring democracy.
Researchers working in the intersection of politics and social media have to acknowledge both the issues of the data they are working with, and the responsibilities they held towards fellow citizens.
Regarding the former, social media users are not representative of the population at large [E]  [E] “Gauging public opinion from social media: Limitations” (in chapter 3 Public opinion).; thus, social media is a conflictive data source to conduct public opinion mining. In addition to that, participants in collective contentious actions move seamlessly between platforms and offline behaviors [F]  [F] “Istanbul: Diren Gezi Parki (Resist, Gezi Park)” (in chapter 5 Contentious politics).. That means that social media can provide at best a partial picture; and one highly skewed if we limit ourselves to data crawled from the most convenient platform.
Social media researchers with STEM backgrounds have mostly engaged in simplistic technological solutions to social problems without further enquiry. In this regard it is worrisome to see the abundance of methods trying to forecast social unrest and protests by means of social media [G]  [G] “London: rioting” (in chapter 5 Contentious politics).; not only due to the topic, but because of the obvious “solution” that is normally suggested: more police. I have no solution to offer for social unrest or rioting—I am not a social scientist, after all—but I am positive that algorithmically-driven extra policing or preemptive repression is not one.
Finally, the most worrisome actions of terrorists in social media are those aiming at self-radicalization of sympathizers [H]  [H] “Social media and terrorism” (in chapter 5 Contentious politics).. Arguably, monitoring, surveillance and censorship could be used to avoid the access of users to such materials. However, it would probably only push them to less visible fora and would not help to avoid radicalization in the long run. As I said in chapter 5 the solution may pass through reducing the attractive that terror ideologies held for some individuals; however, that is a social problem, not a technological one.

Pressing issues and long term approaches

As researchers we need to discern which of the problems affecting social media in politics can be realistically faced from those that cannot; only then we can try to look for solutions to the former in the hope that they can help to improve the latter.
With regard to citizenry, we cannot push them to increase their participation in politics, to engage more frequently with other persons that think differently than them, to improve the coherence of their discourse, or to be more active on the ground than online. However, we can devise social media platforms that increase the chances of serendipitously finding bridges between one’s ideas and dissimilar ones; we can introduce mechanisms to help meaningful debate to emerge or, at least, to allow users to build their own realms for such kind of debate; and we can develop tools to help committed activists to recruit other persons and mobilize them.
We cannot “decommoditize” social media, nor “decolonize” it from corporate elites; much less oblige them to behave in a different way regarding the pressure of different governments, including those in authoritarian countries. However, we can work to offer free alternatives to privately owned social media. Such alternatives should incorporate the mechanisms enumerated above to decrease polarization, increase meaningful debate and mobilize users. Moreover, they should be distributed to avoid control by any government (no matter its nature), and federated to facilitate their interconnection but also to allow users to integrate them, if they want it, with their profiles in commercial social media services. Finally, those tools should be designed to allow citizens in authoritarian countries use them without fear of surveillance or censorship.
With regard to political elites, we cannot expect of them to change their tactics. In authoritarian countries we can just hope to offer their citizens the best possible tools to allow them free speech albeit in a virtual realm. In democratic countries it is possibly unrealistic for every citizen to interact with top officials. However, we can aim to offer elected representatives tools to allow them a meaningful and not cumbersome interaction with their constituency. Moreover, tools to gather the opinion of the people in order to inform the policy making processes should be further developed; particularly to avoid cheating by committed or joyous groups.
Social media researchers should fully acknowledge the limitations of the data they are working with, and accept that sometimes it is unfeasible to understand an event by just analyzing social media data. In those cases, alternative research methods are required and we should not stubbornly focus on social media data. Moreover, social media researchers with STEM backgrounds should try to avoid simplistic technological solutions to social problems and, instead, join forces with social scientists to first understand the underlying causes of the phenomenon they are studying.
Finally, I have no suggestions regarding terrorism; besides, I fully recognize that the kind of tools I have enumerated above would not only help politically active users and activists in democracies, or dissenters in authoritarian regimes, but also terrorists. Still, I think that terrorism is not a technological problem; any measure and countermeasure can eventually be overthrown by terrorists. Besides, technologies developed to fight terrorism are much more frequently addressed towards dissenters.

Lines of research to explore in the near future

The ideas sketched in prior section have a number of important limitations: they are too tenuous in their description, too grandiose in their goals, too ambitious to be handle by a few researchers, and too likely to fail. Though, they still serve two purposes: On one hand, they set the direction towards our daily research should aim; on another hand, they can be used as a rule against which decide the lines of work that are more useful to attain the goals stated above.
In that regard, I consider that the following are amongst the most interesting ones: (1) Alternative social media, (2) online deliberation environments, (3) liquid democracy, and (4) crowdsourcing and swarms.
Alternative social media (cf. Gehl, 2015) includes federated platforms (cf. Halpin and Bria, 2014) such as GNU Social [I]  [I] or Diaspora [J]  [J], p2p software such as Twister [K]  [K] or Ricochet [L]  [L], and darknet services (Gehl, 2014) such as BlackBook or TorBook [M]  [M] http://blkbook3fxhcsn3u.onion/ and http://torbookntnjusnqd.onion/, respectively; you will need a Tor enabled browser to reach them..
The purported advantages of alternative social media are numerous: services are decentralized and, hence, they are more robust; communication is normally fully encrypted and users are anonymous so, in theory, it should be safer for users—particularly in authoritarian countries; and, finally, services are not run for profit or, in the case or p2p software, they are run entirely in the user’s computer.
Still, there are a number of limitations: for many users it is costly to quit a commercial SNS because their whole social network is there; instead of trusting a well known private company users must trust a group of unknown people running the service; as with many open source projects they may be trickier to use than commercial platforms; darknet services are not simple to use for most users under normal conditions and they are commonly demonized by the press, in authoritarian countries its use is even more difficult (Winter and Crandall, 2012); finally, being open source, it is entirely possible for malicious actors to deploy such services to gather personal information or detect dissenters.
Moreover, the user interfaces of those services attempt to clone the user experience of commercial SNS (see Figure XX); therefore, the problems due to affordances in commercial social media are translated to alternative social media. Because of that, research on online deliberation environments is crucial. In this regard, there have been some recent and interesting developments, such as the applications DemocracyOS and Loomio (Mancini, 2015; Firth, 2015), or the work conducted within the European Project D-CENT [N]  [N] Decentralised Citizens ENgagement Technologies. (e.g., Brekke et al., 2014; Aragón et al., 2014; Rogers et al., 2015). Still, there is prior research whose findings should be carefully considered and incorporated in future work; for instance, the report by Price and Cappella (2002) about pre-scheduled and moderated online deliberation, or the work by Wright and Street (2007) on incorporating deliberating features “by design”.
Delegative or liquid democracy is another area that deserves attention. Simply put, it is an hybrid between direct democracy and representative democracy. Every individual is granted a vote; however, depending on the issues they can decide to trust their vote to another individual. For instance, one person can decide to vote on health issues but trust their decisions on energy to another person. That way, people just devote time to decide their vote on issues they are interested and confident enough to take an informed decision; for the rest of issues they delegate their vote. Delegation is transitive and, thus, in a liquid democratic system some individuals may play a role close to that of current representatives. Still, there are some important differences: that “power” is usually trusted on concrete issues and it can be lost very quickly.
The current rendering of liquid democracy is commonly attributed to Bryan Ford (2002) although he acknowledges (Ford, 2014) that the idea is not really new, and traces it back to Lewis Carroll (1884). In any case, recent research has suggested different implementations for liquid democracy (e.g., Boldi et al., 2011; Hardt et al., 2015) and there exist a number of software platforms to apply it in real settings. Maybe the best known is LiquidFeedback [O]  [O] (Behrens et al., 2014) which has been heavily used by the Pirate Party; that software does not only provide proxy voting but also mechanisms to submit and debate propositions—i.e., online deliberation.
This and similar approaches could be extremely interesting for social movements to concrete their demands, political parties develop their manifestos and, why not, governments to incorporate in meaningful ways the opinion of citizens into policy making processes. Needless to say, naïve implementations would be ripe for fraud and, thus, strong cryptography will be needed. In this regard, the reader is referred to the recent book by Melanie Swan about the impact of block chain in liquid democracy and other forms of digital government (Swan, 2015: pp. 44-52).
Up to this moment I have only discussed fully online mechanisms; however, we have learnt that on the ground actions is as important—if not more—than online ones. Meetup was discussed in a number of cases (e.g., the US 2004 Presidential elections or Beppe Grillo’s V-Day and M5S) but we know very little about the actual inner working of those local groups on the ground. That is unfortunate because activists of any kind would greatly benefit from some systematization of offline actions—particularly if it could be somehow integrated with the tools mentioned above.
It goes without saying that on the ground political actions have little to do with social media, and everything with activism; still, I recommend the reader the book by Rick Falkvinge (2013), founder of the Swedish Pirate Party. I am positive that many of the tactics discussed in that manual could be integrated into social media applications.


In the introductory chapter I stated my goals for this book. First, to persuade you that the complacency of cyber-optimism (i.e., social media will make us all freer) is naïve and dangerous. Second, to convince you that it is still possible to transform social media into a set of tools to increase participation, foster dialogue, and improve democracy. I am reasonably sure that I reached my first goal but, unfortunately, I have some doubts regarding the second one.
Still, among the three scenarios depicted by Barber (1998) the only one that is acceptable—no matter how improbable—is precisely that one where technology is neither a tool of oppression or embraced as a panacea, but a tool to help democracy. However, Barber sounds a note of caution:
“Technology can then help democracy, but only if programmed to do so and only in terms of the paradigms and political theories that inform the program. Left to the market, it is likely only to reproduce the vices of politics as usual.
By the same token, if those who deploy the technologies consciously seek a more participatory, deliberative form of strong democracy and a newly robust civil society, they can also find in telecommunications innovation an extraordinarily effective ally. The trouble with the zealots of technology as an instrument of democratic liberation is not that they misconceive technology but that they fail to understand democracy.”
In case the reader is a practitioner of the field I have a few extra pieces of advice:
P.S. Yes, the book was about social media but above all it was about politics.


[1] ARAGÓN, Pablo, et al. D4.3 Technical Design of Open Social Web for Crowdsourced Democracy. D-CENT Decentralised Citizens ENgagement Technologies, 2014. Available at: Last access: July 21st, 2015.

[2] ABBASI, Mohammad-Ali; LIU, Huan. Measuring user credibility in social media. In Social Computing, Behavioral-Cultural Modeling and Prediction. Springer Berlin Heidelberg, 2013. p. 441-448.

[3] ABUL-MAGD, Zeinab. Occupying Tahrir Square: the myths and the realities of the Egyptian revolution. South Atlantic Quarterly, 2012, vol. 111, no 3, pp. 565-572.

[4] ACAMPADA INDEFINIDA EN SOL. Acampada indefinida en sol (Indefinite camp in Sol), 2011. Available at:*/ Last access: July 18th, 2015.

[5] ADAMIC, Lada A.; GLANCE, Natalie. The political blogosphere and the 2004 US election: divided they blog. In Proceedings of the 3rd international workshop on Link discovery. ACM, 2005. p. 36-43.

[6] ADAY, Sean, et al. Blogs and bullets: New media in contentious politics. 2010. Available at: Last access: June 23rd, 2015.

[7] ADAY, Sean, et al. Watching from afar: Media consumption patterns around the Arab Spring. American Behavioral Scientist, 2013, p. 0002764213479373.

[8] ÁGUEDA, Pedro. Interior encarga un sistema para buscar potenciales terroristas en las redes sociales y con la huella digital., 2015. Available at: Last access: June 27th, 2015

[9] AKKOC, Raziye. Turkey blocks access to social media and YouTube over hostage photos. The Telegraph, 2015. Available at: Last access: June 25th, 2015.

[10] AL ZAMAL, Faiyaz; LIU, Wendy; RUTHS, Derek. Homophily and Latent Attribute Inference: Inferring Latent Attributes of Twitter Users from Neighbors. In ICWSM. 2012.

[11] ALBRECHT, Steffen; LÜBCKE, Maren; HARTIG-PERSCHKE, Rasco. Weblog campaigning in the German Bundestag election 2005. Social Science Computer Review, 2007, vol. 25, no 4, p. 504-520.

[12] ALDRICH, John H. Political Parties in and out of Legislatures. In RHODES, R. A. W.; BINDER, Sarah A.; ROCKMAN, Bert A. (Eds.) The Oxford Handbook of Political Institutions. Oxford University Press, 2006, pp. 555-576.

[13] ALDRICH, John H. Why parties? The origin and transformation of party politics in America. Chicago: University of Chicago Press, 1995.

[14] ALEXANDER, Lawrence. Social Network Analysis Reveals Full Scale of Kremlin’s Twitter Bot Campaign. Global voices online, 2015. Available at: Last access: June 24th, 2015.

[15] ALEXANYAN, Karina, et al. Exploring Russian Cyberspace: Digitally-mediated collective action and the networked public sphere. Berkman Center Research Publication, 2012, no 2012-2.

[16] ALLISON, Lincoln. Politics. In MCLEAN, Iain; MCMILLAN, Alistair. The Concise Oxford Dictionary of Politics (3 ed.). Oxford University Press, 2009.

[17] ALLPORT, Floyd H. Toward a science of public opinion. Public opinion quarterly, 1937, vol. 1, no 1, p. 7-23.

[18] ALSUMAIT, Loulwah; BARBARÁ, Daniel; DOMENICONI, Carlotta. On-line lda: Adaptive topic models for mining text streams with applications to topic detection and tracking. In Data Mining, 2008. ICDM’08. Eighth IEEE International Conference on. IEEE, 2008. p. 3-12.

[19] ALTERMAN, Jon B. The revolution will not be tweeted. The Washington Quarterly, 2011, vol. 34, no 4, p. 103-116.

[20] AN, Jisun; QUERCIA, Daniele; CROWCROFT, Jon. Fragmented social media: a look into selective exposure to political news. In Proceedings of the 22nd international conference on World Wide Web companion. International World Wide Web Conferences Steering Committee, 2013. p. 51-52.

[21] ANCU, Monica; COZMA, Raluca. MySpace politics: Uses and gratifications of befriending candidates. Journal of Broadcasting & Electronic Media, 2009, vol. 53, no 4, p. 567-583.

[22] ANDUIZA, Eva; CRISTANCHO, Camilo; SABUCEDO, José M. Mobilization through online social networks: the political protest of the indignados in Spain. Information, Communication & Society, 2014, vol. 17, no 6, p. 750-764.

[23] ANSTEAD, Nick; O’LOUGHLIN, Ben. Semantic polling: the ethics of online public opinion. 2012.

[24] ANSTEAD, Nick; O’LOUGHLIN, Ben. The emerging viewertariat and BBC question time: Television debate and real-time commenting online. The international journal of press/politics, 2011, p. 1940161211415519.

[25] ARAGÓN, Pablo, et al. Communication dynamics in Twitter during political campaigns: The case of the 2011 Spanish national election. Policy & Internet, 2013, vol. 5, no 2, p. 183-206.

[26] ARAGÓN, Pablo, et al. D4.3 Technical Design of Open Social Web for Crowdsourced Democracy. D-CENT Decentralised Citizens ENgagement Technologies, 2014. Available at: Last access: July 21st, 2015.

[27] AUSSERHOFER, Julian; MAIREDER, Axel. National politics on Twitter: Structures and topics of a networked public sphere. Information, Communication & Society, 2013, vol. 16, no 3, p. 291-314.

[28] AUTY, Caroline. UK elected representatives and their weblogs: first impressions. In AUTY, Caroline (ed.). Aslib Proceedings. Emerald Group Publishing Limited, 2005. p. 338-355.

[29] BAKARDJIEVA, Maria. Do clouds have politics? Collective actors in social media land. Information, Communication & Society, 2015.

[30] BAKKER, Tom P.; DE VREESE, Claes H. Good news for the future? Young people, Internet use, and political participation. Communication Research, 2011, p. 0093650210381738.

[31] BAKLIWAL, Akshat, et al. Sentiment Analysis of Political Tweets: Towards an Accurate Classifier. NAACL 2013, 2013, p. 49.

[32] BAKSHY, Eytan; MESSING, Solomon; ADAMIC, Lada. Exposure to ideologically diverse news and opinion on Facebook. Science, 2015, p. aaa1160.

[33] BALOG, Krisztian; MISHNE, Gilad; DE RIJKE, Maarten. Why are they excited?: identifying and explaining spikes in blog mood levels. In Proceedings of the Eleventh Conference of the European Chapter of the Association for Computational Linguistics: Posters & Demonstrations. Association for Computational Linguistics, 2006. p. 207-210.

[34] BAMMAN, David; O’CONNOR, Brendan; SMITH, Noah. Censorship and deletion practices in Chinese social media. First Monday, 2012, vol. 17, no 3.

[35] BARAB, Jordan. Labor Blogs: Bright Future for a New Medium. Perspectives on Work, 2007, p. 30-32.

[36] BARBER, Benjamin R. Three scenarios for the future of technology and strong democracy. Political Science Quarterly, 1998, vol. 113, no 4, p. 573-589.

[37] BARBERÁ, Pablo. A new measure of party identification in Twitter. evidence from Spain. In Proc. 2nd Annual General Conference of EPSA. 2012.

[38] BARBERÁ, Pablo. Birds of the same feather tweet together: Bayesian ideal point estimation using Twitter data. Political Analysis, 2015, vol. 23, no 1, p. 76-91.

[39] BAROMETER, Edelman Trust. 2012 Edelman trust barometer executive summary. 2012.

[40] BAROMETER, Edelman Trust. 2014 Edelman trust barometer executive summary. 2014.

[41] BARRONS, Genevieve. ‘Suleiman: Mubarak decided to step down #egypt #jan25 OH MY GOD’: Examining the use of social media in the 2011 Egyptian revolution. Contemporary Arab Affairs, 2012, vol. 5, no 1, p. 54-67.

[42] BARRY, Ellen. Protests in Moldova explode, with help of Twitter. The New York Times, 2009. Available at: Last access: July 16th, 2015.

[43] BARTLETT, Jamie, et al. New political actors in Europe: Beppe Grillo and the M5S. 2013.

[44] BARTLETT, Jamie. Populism, Social Media and Democratic Strain. European Populism, 2014, p. 99.

[45] BAUERLEIN, Mark. The dumbest generation: How the digital age stupefies young Americans and jeopardizes our future (or, don’t trust anyone under 30). Penguin, 2008.

[46] BAUMGARTNER, Jody C. Humor on the Next Frontier Youth, Online Political Humor, and the JibJab Effect. Social Science Computer Review, 2007, vol. 25, no 3, p. 319-338.

[47] BAUMGARTNER, Jody C.; MORRIS, Jonathan S. MyFaceTube politics: Social networking web sites and political engagement of young adults. Social Science Computer Review, 2009.

[48] BAXTER, Graeme; MARCELLA, Rita; VARFIS, Evaggelos. The use of the internet by political parties and candidates in Scotland during the 2010 UK general election campaign. In Aslib Proceedings. Emerald Group Publishing Limited, 2011. p. 464-483.

[49] BBC. London riots: Looting and violence continues. BBC News, 2011. Available at: Last access: July 18th, 2015.

[50] BEAUCHAMP, Nick. Predicting and interpolating state-level polling using Twitter textual data. In New Directions in Analyzing Text as Data Workshop. 2013.

[51] BEHRENS, Jan, et al. The principles of LiquidFeedback. Interaktive Demokratie e. V., 2014.

[52] BEKAFIGO, Marija Anna; MCBRIDE, Allan. Who tweets about politics? Political participation of Twitter users during the 2011 gubernatorial elections. Social Science Computer Review, 2013, vol. 31, no. 5, pp. 625-643.

[53] BELL, Emily. Separating the bloggers from the tossers. The Guardian, 2006. Available at: Last access: July 7th, 2015.

[54] BENEVENUTO, Fabricio, et al. Detecting spammers on twitter. In Collaboration, electronic messaging, anti-abuse and spam conference (CEAS). 2010. p. 12.

[55] BENNAHUM, David S. The Internet Revolution. WIRED, 1997, vol. 5, p. 122-129. Available at: Last access: June 23rd, 2015.

[56] BENNETT, W. Lance. The personalization of politics. Political identity, social media, and changing patterns of participation. The ANNALS of the American Academy of Political and Social Science, 2012, vol. 644, no 1, p. 20-39.

[57] BENNETT, W. Lance; SEGERBERG, Alexandra. The logic of connective action: Digital media and the personalization of contentious politics. Information, Communication & Society, 2012, vol. 15, no 5, p. 739-768.

[58] BENNETT, Winston. Communicating global activism. Information, Communication & Society, 2003, vol. 6, no 2, p. 143-168.

[59] BENVIE, Drew. Social media. Wikipedia, 2006. Available at: Last access: May 25th, 2015.

[60] BERCOVICI, Jeff. Who coined “social media”? Web pioneers compete for credit. Forbes, 2010. Available at: Last access: May 25th, 2015.

[61] BERGSMA, Shane, et al. Broadly Improving User Classification via Communication-Based Name and Location Clustering on Twitter. In HLT-NAACL. 2013. p. 1010-1019.

[62] BERMINGHAM, Adam; SMEATON, Alan F. On using Twitter to monitor political sentiment and predict election results. 2011.

[63] BERNAYS, Edward L. The engineering of consent. The Annals of the American Academy of Political and Social Science, 1947, p. 113-120.

[64] BERTOT, John Carlo, et al. Social media technology and government transparency. Computer, 2010, no 11, p. 53-59.

[65] BIMBER, Bruce, et al. Reconceptualizing collective action in the contemporary media environment. Communication Theory, 2005, vol. 15, no 4, p. 365.

[66] BIMBER, Bruce; DAVIS, Richard. Campaigning online: The Internet in US elections. Oxford University Press, 2003.

[67] BISHOP, Bradford H.; HILLYGUS, D. Sunshine. Campaigning, debating, advertising. The Oxford Handbook of American Public Opinion and the Media, 2013, pp. 204-219.

[68] BLODGETT, Bridget; TAPIA, Andrea. Do avatars dream of electronic picket lines? The blurring of work and play in virtual environments. Information Technology & People, 2011, vol. 24, no 1, p. 26-45.

[69] BLUMENTHAL, Sidney. The permanent campaign. New York: Simon and Schuster, 1982. Cited by HECLO, Hugh. Campaigning and governing: A conspectus. In ORNSTEIN, Norman J.; MANN, Thomas E. (Eds.) The permanent campaign and its future. Washington, DC: American Enterprise Institute, 2000, pp. 1-37.

[70] BODE, Leticia, et al. Candidate Networks, Citizen Clusters, and Political Expression Strategic Hashtag Use in the 2010 Midterms. The ANNALS of the American Academy of Political and Social Science, 2015, vol. 659, no 1, p. 149-165.

[71] BODE, Leticia. Facebooking it to the polls: a study in online social networking and political behavior. Journal of Information Technology & Politics, 2012, vol. 9, no 4, p. 352-369.

[72] BOLDI, Paolo, et al. Viscous democracy for social networks. Communications of the ACM, 2011, vol. 54, no 6, pp. 129-137.

[73] BONCHEK, Mark S. Message to candidates: It’s about community, stupid, 1996. Available at: Last access: July 5th, 2015.

[74] BOND, Robert M., et al. A 61-million-person experiment in social influence and political mobilization. Nature, 2012, vol. 489, no 7415, p. 295-298.

[75] BOND, Robert; MESSING, Solomon. Quantifying Social Media’s Political Space: Estimating Ideology from Publicly Revealed Preferences on Facebook. American Political Science Review, 2015, vol. 109, no 01, p. 62-78.

[76] BORDIGNON, Fabio; CECCARINI, Luigi. Five stars and a cricket. Beppe Grillo shakes Italian politics. South European Society and Politics, 2013, vol. 18, no 4, p. 427-449.

[77] BORGE-HOLTHOEFER, Javier, et al. Structural and dynamical patterns on online social networks: the spanish may 15th movement as a case study. PloS one, 2011, vol. 6, no 8, p. e23883.

[78] BORONDO, J., et al. Characterizing and modeling an electoral campaign in the context of Twitter: 2011 Spanish Presidential election as a case study. Chaos: an interdisciplinary journal of nonlinear science, 2012, vol. 22, no 2, p. 023138.

[79] BOSHMAF, Yazan, et al. The socialbot network: when bots socialize for fame and money. In Proceedings of the 27th Annual Computer Security Applications Conference. ACM, 2011. p. 93-102.

[80] BOULIANNE, Shelley. Does Internet use affect engagement? A meta-analysis of research. political communication, 2009, vol. 26, no 2, p. 193-211.

[81] BOURDIEU, Pierre. Public opinion does not exist. Communication and class struggle, 1979, vol. 1, p. 124-310.

[82] BOUTET, Antoine, et al. What’s in Your Tweets? I Know Who You Supported in the UK 2010 General Election. In ICWSM. 2012.

[83] BOYD, danah; CRAWFORD, Kate. Critical questions for big data: Provocations for a cultural, technological, and scholarly phenomenon. Information, communication & society, 2012, vol. 15, no 5, p. 662-679.

[84] BRACHMAN, Jarret M.; LEVINE, Alix N. You too can be Awlaki. Fletcher F. World Aff., 2011, vol. 35, p. 25.

[85] BRADLEY, Margaret M.; LANG, Peter J. Affective norms for English words (ANEW): Instruction manual and affective ratings. Technical Report C-1, The Center for Research in Psychophysiology, University of Florida, 1999.

[86] BRANCATI, Dawn. Pocketbook Protests Explaining the Emergence of Pro-Democracy Protests Worldwide. Comparative Political Studies, 2013, p. 0010414013512603.

[87] BREKKE, Klara Jaya, et al. D1.2 Communities’ Requirements and Social Design. D-CENT Decentralised Citizens ENgagement Technologies, 2014. Available at: Last access: July 21st, 2015.

[88] BRIONES, Rowena L., et al. Keeping up with the digital age: How the American Red Cross uses social media to build relationships. Public Relations Review, 2011, vol. 37, no 1, p. 37-43.

[89] BROERSMA, Marcel; GRAHAM, Todd. Social media as beat: tweets as a news source during the 2010 British and Dutch elections. Journalism Practice, 2012, vol. 6, no 3, p. 403-419.

[90] BRUNDIDGE, Jennifer; RICE, Ronald E. Political engagement online: Do the information rich get richer and the like-minded more similar. Routledge handbook of Internet politics, 2009, p. 144-156.

[91] BRUNS, Axel, et al. # ausvotes: How Twitter covered the 2010 Australian federal election. Politics and Culture, 2011, vol. 44, no. 2, pp. 37-56.

[92] BRUNS, Axel; HIGHFIELD, Tim. Political networks on Twitter: tweeting the Queensland state election. Information, Communication & Society, 2013, vol. 16, no 5, p. 667-691.

[93] BUCHSTEIN, Hubertus. Bytes that bite: The Internet and deliberative democracy. Constellations, 1997, vol. 4, no 2, p. 248-263.

[94] BURGER, John D., et al. Discriminating gender on Twitter. In Proceedings of the Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics, 2011. p. 1301-1309.

[95] BURKART, Patrick. Pirate Politics: The New Information Policy Contests. MIT Press, 2014.

[96] BURNAP, Pete, et al. 140 Characters to Victory?: Using Twitter to Predict the UK 2015 General Election. Social Science Research Network. Available at: Last access: July 10th, 2015.

[97] BURNHAM, Peter. Trade unions. In MCLEAN, Iain (Ed.) Oxford concise dictionary of politics. Oxford University Press UK, 2003.

[98] BUTLER, Desmond; GILLUM, Jack; ARCE, Alberto. US secretly created ‘Cuban Twitter’ to stir unrest. Associated Press, Apr, 2014, vol. 4, p. 12. Available at: Last access: June 23rd, 2015.

[99] CADENA, Jose, et al. Forecasting Social Unrest Using Activity Cascades. PloS one, 2015, vol. 10, no 6, p. e0128879.

[100] CAI, Yongshun. Collective resistance in China: Why popular protests succeed or fail. Stanford University Press, 2010.

[101] CALDARELLI, Guido, et al. A multi-level geographical study of Italian political elections from Twitter data. 2014.

[102] CALHOUN, Craig J. Habermas and the public sphere. MIT press, 1992.

[103] CAMERON, David. PM statement on disorder in England. Prime Minister’s Office, 2011. Available at: Last access: July 18th, 2015.

[104] CAMERON, Michael P.; BARRETT, Patrick; STEWARDSON, Bob. Can Social Media Predict Election Results? Evidence from New Zealand. Journal of Political Marketing, 2014.

[105] CAMMAERTS, Bart. Critiques on the participatory potentials of Web 2.0. Communication, Culture and Critique, 2008, vol. 1, no 4, p. 358-377.

[106] CARDULLO, Paolo. Hacking multitude’and Big Data: Some insights from the Turkish ‘digital coup. Big Data & Society, 2015, vol. 2, no 1, p. 2053951715580599.

[107] CARR, David. Hashtag Activism, and Its Limits. The New York Times, 2012. Available at: Last access: July 15th, 2015.

[108] CARROLL, Lewis. The principles of parliamentary representation. Harrison and Sons, 1884.

[109] CARVALHO, Paula, et al. Clues for detecting irony in user-generated contents: oh...!! it’s so easy;-). In Proceedings of the 1st international CIKM workshop on Topic-sentiment analysis for mass opinion. ACM, 2009. p. 53-56.

[110] CASTELLS, Manuel, et al. Electronic communication and socio-political mobilization: A new form of civil society. Global civil society, 2005, vol. 6, p. 266-285.

[111] CASTELLS, Manuel. Networks of outrage and hope: Social movements in the Internet age. John Wiley & Sons, 2013.

[112] CASTELLS, Manuel. The power of identity: The information age: Economy, society, and culture. John Wiley & Sons, 2010.

[113] CASTILLO, Carlos; MENDOZA, Marcelo; POBLETE, Barbara. Information credibility on twitter. In Proceedings of the 20th international conference on World wide web. ACM, 2011. p. 675-684.

[114] CASTILLO, Carlos; MENDOZA, Marcelo; POBLETE, Barbara. Predicting information credibility in time-sensitive social media. Internet Research, 2013, vol. 23, no 5, p. 560-588.

[115] CERON, Andrea, et al. Every tweet counts? How sentiment analysis of social media can improve our knowledge of citizens’ political preferences with an application to Italy and France. New Media & Society, 2014, vol. 16, no 2, p. 340-358.

[116] CERON, Andrea; CURINI, Luigi; IACUS, Stefano M. Using Sentiment Analysis to Monitor Electoral Campaigns Method Matters—Evidence From the United States and Italy. Social Science Computer Review, 2015, vol. 33, no 1, p. 3-20.

[117] CHADWICK, Andrew. Web 2.0: New challenges for the study of e-democracy in an era of informational exuberance. ISJLP, 2008, vol. 5, pp. 9-41.

[118] CHEN, Lu; WANG, Wenbo; SHETH, Amit P. Are twitter users equal in predicting elections? A study of user groups in predicting 2012 US Republican presidential primaries. In Social informatics. Springer Berlin Heidelberg, 2012. p. 379-392.

[119] CHRISTENSEN, Christian. Iran: Networked dissent. Le Monde Diplomatique, 2009.

[120] CHU, Zi, et al. Who is tweeting on Twitter: human, bot, or cyborg?. In Proceedings of the 26th annual computer security applications conference. ACM, 2010. p. 21-30.

[121] CHUNG, Jessica Elan; MUSTAFARAJ, Eni. Can collective sentiment expressed on twitter predict political elections?. In AAAI. 2011.

[122] COGBURN, Derrick L.; ESPINOZA-VASQUEZ, Fatima K. From networked nominee to networked nation: Examining the impact of Web 2.0 and social media on political participation and civic engagement in the 2008 Obama campaign. Journal of Political Marketing, 2011, vol. 10, no 1-2, p. 189-213.

[123] COHEN, Raviv; RUTHS, Derek. Classifying Political Orientation on Twitter: It’s Not Easy!. In ICWSM. 2013.

[124] COLEMAN, Stephen. Blogs and the new politics of listening. The Political Quarterly, 2005, vol. 76, no 2, p. 272-280.

[125] CONOVER, Michael D., et al. Predicting the political alignment of twitter users. In Privacy, Security, Risk and Trust (PASSAT) and 2011 IEEE Third Inernational Conference on Social Computing (SocialCom), 2011 IEEE Third International Conference on. IEEE, 2011b. p. 192-199.

[126] CONOVER, Michael, et al. Political Polarization on Twitter. In ICWSM. 2011b.

[127] CONROY, Meredith; FEEZELL, Jessica T.; GUERRERO, Mario. Facebook and political engagement: A study of online political group membership and offline political engagement. Computers in Human Behavior, 2012, vol. 28, no 5, p. 1535-1546.

[128] CONWAY, Maura; MCINERNEY, Lisa. Jihadi video and auto-radicalisation: Evidence from an exploratory YouTube study. En Intelligence and Security Informatics. Springer Berlin Heidelberg, 2008. p. 108-118.

[129] CORNFIELD, Michael, et al. Buzz, blogs, and beyond: The Internet and the national discourse in the fall of 2004. Pew Internet & American Life Project. 2005.

[130] COWLING, David. Election results: How did pollsters get it so wrong? BBC News, 2015. Available at: Last access: July 10, 2015.

[131] CROUCH, Colin. Coping with Post-democracy. London: Fabian Society, 2000.

[132] CRUMP, Jeremy. What are the police doing on Twitter? Social media, the police and the public. Policy & Internet, 2011, vol. 3, no 4, p. 1-27.

[133] DABASHI, Hamid. The Arab Spring: Delayed Defiance and the End of Postcolonialism. Zed Books Limited, 2012.

[134] DAHLBERG, Lincoln. Computer-mediated communication and the public sphere: A critical analysis. Journal of Computer-Mediated Communication, 2001a, vol. 7, no 1.

[135] DAHLBERG, Lincoln. Democracy via Cyberspace Mapping the Rhetorics and Practices of Three Prominent Camps. New media & society, 2001b, vol. 3, no 2, p. 157-177.

[136] DAHLBERG, Lincoln. Extending the public sphere through cyberspace: The case of Minnesota E-Democracy. First Monday, 2001c, vol. 6, no 3.

[137] DAHLBERG, Lincoln. The corporate colonization of online attention and the marginalization of critical communication?. Journal of Communication Inquiry, 2005, vol. 29, no 2, p. 160-180.

[138] DAHLBERG, Lincoln. Which Social Media? A Call for Contextualization. Social Media+ Society, 2015, vol. 1, no 1, p. 2056305115578142. Available at: Last access: May 25th, 2015.

[139] DAHLGREN, Peter. The Internet, public spheres, and political communication: Dispersion and deliberation. Political communication, 2005, vol. 22, no 2, p. 147-162.

[140] D’ALESSIO, Dave. Adoption of the World Wide Web by American political candidates, 1996-1998. Journal of Broadcasting & Electronic Media, 2000, vol. 44, no 4, p. 556-568.

[141] DAVIES, Toby P., et al. A mathematical model of the London riots and their policing. Scientific reports, 2013, vol. 3.

[142] DAVIS, Richard, et al. The Internet in US election campaigns. Routledge handbook of Internet politics, 2009, p. 25-39.

[143] DAVIS, Richard. The web of politics: The Internet’s impact on the American political system. Oxford University Press, Inc., 1999.

[144] DAVIS, Steve. Presidential campaigns fine-tune online strategies. Journalism Studies, 2005, vol. 6, no 2, p. 241-244.

[145] DEAN, Jodi. Blog theory: Feedback and capture in the circuits of drive. Polity, 2010.

[146] DEAN, Jodi. Communicative capitalism: circulation and the foreclosure of politics. Cultural Politics, 2005, vol. 1, no 1, p. 51-74.

[147] DEARING, James W.; ROGERS, Everett. Agenda-setting. Sage publications, 1996.

[148] DEIBERT, Ronald, et al. Access denied: The practice and policy of global internet filtering. MIT Press, 2008.

[149] DELLA PORTA, Donatella; DIANI, Mario. Social movements: An introduction. Blackwell Publishing, 2006.

[150] DELLI CARPINI, Michael X. Gen. com: Youth, civic engagement, and the new information environment. Political communication, 2000, vol. 17, no 4, p. 341-349.

[151] DELWICHE, Aaron. Agenda–setting, opinion leadership, and the world of Weblogs. First Monday, 2005, vol. 10, no 12.

[152] DEMARTINI, Gianluca, et al. Analyzing Political Trends in the Blogosphere. In ICWSM. 2011.

[153] DERRIDA, Jacques. Specters of Marx, trans. Peggy Kamuf. Routledge, 1994.

[154] DI GENNARO, Corinna; DUTTON, William. The Internet and the public: Online and offline political participation in the United Kingdom. Parliamentary Affairs, 2006, vol. 59, no 2, p. 299-313.

[155] DIAKOPOULOS, Nicholas A.; SHAMMA, David A. Characterizing debate performance via aggregated twitter sentiment. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. ACM, 2010. p. 1195-1198.

[156] DIAMOND, Larry. Liberation technology. Journal of Democracy, 2010, vol. 21, no 3, p. 69-83.

[157] DIAZ, Fernando, et al. Online and social media data as a flawed continuous panel survey. 2014. Available at: Last access: June 17th, 2015.

[158] DIGRAZIA, Joseph, et al. More tweets, more votes: Social media as a quantitative indicator of political behavior. 2013. PLoS ONE, vol. 8, no. 11, p. e79449.

[159] DIMAGGIO, Paul, et al. Digital inequality: From unequal access to differentiated use. Social inequality, 2004, p. 355-400.

[160] DIMAGGIO, Paul, et al. Social implications of the Internet. Annual review of sociology, 2001, p. 307-336.

[161] DOERNBERG, Donald L. Sovereignty in the Age of Twitter. Vill. L. Rev., 2010, vol. 55, p. 833.

[162] DOGAN, Mattei. Conceptions of legitimacy. Encyclopedia of government and politics, 1992, vol. 1, p. 116-26.

[163] DREIER, Peter. Organizing in the Obama Era: A Progressive Movement or a New Progressive Era. J. Marshall L. Rev., 2008, vol. 42, pp. 685-722.

[164] DREZNER, Daniel W.; FARRELL, Henry. Web of influence. Foreign Policy, 2004, p. 32-41.

[165] DRUCKMAN, James N.; KIFER, Martin J.; PARKIN, Michael. The Technological Development of Congressional Candidate Web Sites How and Why Candidates Use Web Innovations. Social Science Computer Review, 2007, vol. 25, no 4, p. 425-442.

[166] DUFFY, Bobby. Spinning the Election: Who is setting the Agenda in the UK General Election 2015? Ipsos MORI, 2015. Available at: Last access: July 8th, 2015.

[167] DUGGAN, M., et al. Social Media Update 2014, Pew Research Center, 2015.

[168] EAGLETON-PIERCE, Matthew. The internet and the Seattle WTO protests. Peace Review, 2001, vol. 13, no 3, p. 331-337.

[169] EASTON, David. An approach to the analysis of political systems. World politics, 1957, vol. 9, no 03, p. 383-400.

[170] ELMER, Greg, et al. “Blogs I Read”: Partisanship and Party Loyalty in the Canadian Political Blogosphere. Journal of Information Technology & Politics, 2009, vol. 6, no 2, p. 156-165.

[171] ELMER, Greg. Live research: Twittering an election debate. New media & society, 2012, p. 1461444812457328.

[172] ENLI, Gunn Sara; SKOGERBØ, Eli. Personalized campaigns in party-centred politics: twitter and Facebook as arenas for political communication. Information, Communication & Society, 2013, vol. 16, no 5, p. 757-774.

[173] ESAREY, Ashley; QIANG, Xiao. Political expression in the Chinese blogosphere: Below the radar. 2008.

[174] ESULI, Andrea; SEBASTIANI, Fabrizio. Sentiwordnet: A publicly available lexical resource for opinion mining. In Proceedings of LREC. 2006. p. 417-422.

[175] ETLING, Bruce, et al. Public discourse in the Russian blogosphere: Mapping RuNet politics and mobilization. Berkman Center Research Publication, 2010, no 2010-11.

[176] EVANS, Mark. A Look at Twitter in Iran. Sysomos blogs, 2009. Available at: Last access: July 20th, 2015.

[177] EVELAND JR, WILLIAM P.; DYLKO, IVAN. Reading Political Blogs During the 2004 Election Campaign: Correlates and Political Consequences. Blogging, citizenry, and the Future of Media, 2012, p. 105.

[178] EZLN. Declaración de la Selva Lacandona (Declaration from the Lacandon Jungle). 1993. Available at: Last access: July 17th, 2015.

[179] FALKVINGE, Rick. Swarmwise: The tactical manual to changing the world. CreateSpace Publishing Platform, 2013.

[180] FALLOWS, Deborah. Few in China complain about internet controls. Pew Internet & American Life Project. Pew Research Center Publications: March, 2008, vol. 27.

[181] FANELLI, Daniele. “Positive” results increase down the Hierarchy of the Sciences. PLoS ONE, 2010, vol. 5, no 3.

[182] FARINA, Cynthia R.; NEWHART, Mary; HEIDT, Josiah. Rulemaking vs. democracy: Judging and nudging public participation that counts. Mich. J. Envtl. & Admin. L., 2012, vol. 2, p. 123.

[183] FARIS, David. Revolutions without revolutionaries? Social media networks and regime response in Egypt. Publicly accessible Penn Dissertations, 2010.

[184] FARWELL, James P. The Media Strategy of ISIS. Survival, 2014, vol. 56, no 6, p. 49-55.

[185] FEENBERG, Andrew; FREEDMAN, Jim. When poetry ruled the streets: The French May events of 1968. Suny Press, 2010.

[186] FEHR, Ernst; GÄCHTER, Simon. Altruistic punishment in humans. Nature, 2002, vol. 415, no 6868, p. 137-140.

[187] FENTON, Natalie; BARASSI, Veronica. Alternative media and social networking sites: The politics of individuation and political participation. The Communication Review, 2011, vol. 14, no 3, p. 179-196.

[188] FERGUSON, Ross. Diving in the shallow end. In Spinning the Web: online campaigning in the 2005 general election. Hansard Society, 2005, pp. 31-35.

[189] FINANCIAL TIMES. Timeline: Iran’s post election protests. Financial Times, 2010. Available at: Last access: July 20th, 2015.

[190] FINK, Colin, et al. Twitter, Public Opinion, and the 2011 Nigerian Presidential Election. In Social Computing (SocialCom), 2013 International Conference on. IEEE, 2013. p. 311-320.

[191] FIRTH, Niall. We the people. New Scientist, 2015, vol. 226, no 3018, pp. 38-41.

[192] FLESHER FOMINAYA, Cristina. The Madrid bombings and popular protest: misinformation, counter-information, mobilisation and elections after ‘11-M’. Contemporary Social Science, 2011, vol. 6, no 3, p. 289-307.

[193] FONSECA, António. Modeling political opinion dynamics through social media and multi-agent simulation. In First Doctoral Workshop for Complexity Sciences. 2011.

[194] FOOT, Kirsten A., et al. Comparing web production practices across electoral web spheres. In The internet and national elections: A comparative study of web campaigning, 2007, pp. 243-260.

[195] FOOT, Kirsten A.; SCHNEIDER, Steven M. Web campaigning. Cambridge, MA: MIT press, 2006.

[196] FORD, Bryan. Delegative democracy revisited. Blog post, 2014. Available at: Last access: July 21st, 2015.

[197] FORD, Bryan. Delegative democracy. Draft paper, 2002. Available at: Last access: July 21st, 2015.

[198] FRANCH, Fabio. (Wisdom of the Crowds) 2: 2010 UK election prediction with social media. Journal of Information Technology & Politics, 2013, vol. 10, no 1, p. 57-71.

[199] FREELON, Deen G. Analyzing online political discussion using three models of democratic communication. New Media & Society, 2010.

[200] FREELON, Deen. Discourse architecture, ideology, and democratic norms in online political discussion. New Media & Society, 2015, vol. 17, no 5, p. 772-791.

[201] FREELON, Deen; KARPF, David. Of big birds and bayonets: hybrid Twitter interactivity in the 2012 Presidential debates. Information, Communication & Society, 2015, vol. 18, no 4, p. 390-406.

[202] FU, King-Wa; CHAN, C.H. Networked collective action in the 2014 Hong Kong Occupy Movement: analysing a Facebook sharing network. In Proceedings of the 2nd International Conference on Public Policy (ICPP 2015), 2015.

[203] FUCHS, Christian. Social media, riots, and revolutions. Capital & Class, 2012, vol. 36, no 3, p. 383-391.

[204] FUKUYAMA, Francis. At the ‘End of History’ Still Stands Democracy. Wall Street Journal, 2014. Available at: Last access: May 24th, 2015.

[205] FUKUYAMA, Francis. The End of History and the Last Man. 1992.

[206] FUNG, Anthony. One city, two systems: Democracy in an electronic chat room in Hong Kong. Javnost-The Public, 2002, vol. 9, no 2, p. 77-93.

[207] GADDE, V. Challenging the access ban in turkey. The Official Twitter Blog, 2014. Available at: Last access: June 27th, 2015.

[208] GALLAGHER, Ian. Egyptian Police Use Facebook and Twitter to Track down Protesters’ Names before ‘Rounding Them Up.’. Daily Mail, 2011. Available at: Last access: June 23rd, 2015.

[209] GANZ, Marshall. Organizing Obama: Campaign, Organizing, Movement. In American Sociological Association Annual Meeting in San Francisco, August (unpublished). 2009. Available at: Last access: July 6th, 2015.

[210] GARRIDO, Maria; HALAVAIS, Alexander. Mapping networks of support for the Zapatista movement. Cyberactivism: Online activism in theory and practice, 2003, pp. 165-184.

[211] GATES, Christopher. Eulogy for Politwoops. Available at: Last access: June 10th, 2015.

[212] GAURAV, Manish, et al. Leveraging candidate popularity on Twitter to predict election outcome. In Proceedings of the 7th Workshop on Social Network Mining and Analysis. ACM, 2013. p. 7.

[213] GAYO AVELLO, Daniel. All liaisons are dangerous when all your friends are known to us. In Proceedings of the 22nd ACM conference on Hypertext and hypermedia. ACM, 2011a. p. 171-180.

[214] GAYO-AVELLO, Daniel. A meta-analysis of state-of-the-art electoral prediction from Twitter data. Social Science Computer Review, 2013, p. 0894439313493979.

[215] GAYO-AVELLO, Daniel. A meta-analysis of state-of-the-art electoral prediction from Twitter data. Social Science Computer Review, 2013, vol. 31, no. 6, pp. 649-679.

[216] GAYO-AVELLO, Daniel. Don’t turn social media into another ’Literary Digest’ poll. Communications of the ACM, 2011b, vol. 54, no 10, p. 121-128.

[217] GAYO-AVELLO, Daniel. Don’t turn social media into another’Literary Digest’poll. Communications of the ACM, 2011, vol. 54, no 10, p. 121-128.

[218] GEARY, Joanna. Political discovery on Twitter. Twitter blog. 2015. Available at: Last access: July 8th, 2015.

[219] GEHL, Robert W. Power/freedom on the dark web: A digital ethnography of the Dark Web Social Network. new media & society, 2014, p. 1461444814554900.

[220] GEHL, Robert W. The Case for Alternative Social Media, 2015. Forthcoming in Social Media + Society. Available at SSRN:

[221] GEIGER, R. Stuart. Does Facebook Have Civil Servants? On Governmentality and Computational Social Science. In proceedings of CSCW 2015 Ethics Workshop, 2015.

[222] GERBAUDO, Paolo. Tweets and the streets: Social media and contemporary activism. Pluto Press, 2012.

[223] GERBAUDO, Paolo; TRERÉ, Emiliano. In search of the ‘we’ of social media activism: introduction to the special issue on social media and protest identities. Information, Communication & Society, 2015.

[224] GERSCHEWSKI, Johannes. The three pillars of stability: legitimation, repression, and co-optation in autocratic regimes. Democratization, 2013, vol. 20, no 1, p. 13-38.

[225] GIBSON, Rachel K.; LUSOLI, Wainer; WARD, Stephen. Online Participation in the UK: Testing a ‘Contextualised’ Model of Internet Effects. The British Journal of Politics & International Relations, 2005, vol. 7, no 4, p. 561-583.

[226] GIBSON, Rachel K.; MCALLISTER, Ian. Do online election campaigns win votes? The 2007 Australian “YouTube” election. Political Communication, 2011, vol. 28, no 2, p. 227-244.

[227] GIBSON, Rachel K.; NEWELL, James L.; WARD, Stephen J. New parties, new media: Italian party politics and the Internet. South European Society and Politics, 2000, vol. 5, no 1, p. 123-136.

[228] GIBSON, Rachel K.; WARD, Stephen J. UK political parties and the Internet “politics as usual” in the new media?. The Harvard International Journal of Press/Politics, 1998, vol. 3, no 3, p. 14-38.

[229] GIBSON, Rachel; CANTIJOCH, Marta. Conceptualizing and measuring participation in the age of the internet: Is online political engagement really different to offline?. The Journal of Politics, 2013, vol. 75, no 03, p. 701-716.

[230] GIBSON, Rachel; WARD, Stephen. Party democracy on-line: UK parties and new ICTs. Information, Communication & Society, 1999, vol. 2, no 3, p. 340-367.

[231] GIBSON, William. Neuromancer. Ace, 1984.

[232] GIGLIETTO, Fabio. If Likes Were Votes: An Empirical Study on the 2011 Italian Administrative Elections. In ICWSM. 2012.

[233] GILENS, Martin; PAGE, Benjamin I. Testing theories of American politics: Elites, interest groups, and average citizens. Perspectives on politics, 2014, vol. 12, no 03, p. 564-581.

[234] GILES, Jim. Fake tweets by ’socialbot’ fool hundreds of followers. New Scientist. 2011. Available at:

[235] GILLEY, Bruce. The determinants of state legitimacy: Results for 72 countries. International political science review, 2006a, vol. 27, no 1, p. 47-71.

[236] GILLEY, Bruce. The meaning and measure of state legitimacy: Results for 72 countries. European Journal of Political Research, 2006b, vol. 45, no 3, p. 499-525.

[237] GILLEY, Bruce. The right to rule: how states win and lose legitimacy. Columbia University Press, 2013.

[238] GILPIN, Dawn R.; PALAZZOLO, Edward T.; BRODY, Nicholas. Socially mediated authenticity. Journal of communication management, 2010, vol. 14, no 3, p. 258-278.

[239] GLENN, Norval D.; HILL, Lester. Rural-urban differences in attitudes and behavior in the United States. The Annals of the American Academy of Political and Social Science, 1977, vol. 429, no 1, p. 36-50.

[240] GODARD, John. What has happened to strikes?. British Journal of Industrial Relations, 2011, vol. 49, no 2, p. 282-305.

[241] GOLBECK, Jennifer; GRIMES, Justin M.; ROGERS, Anthony. Twitter use by the US Congress. Journal of the American Society for Information Science and Technology, 2010, vol. 61, no 8, p. 1612-1621.

[242] GOLBECK, Jennifer; HANSEN, Derek. Computing political preference among twitter followers. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. ACM, 2011. p. 1105-1108.

[243] GONZÁLEZ-BAILÓN, Sandra; BANCHS, Rafael E.; KALTENBRUNNER, Andreas. Emotions, Public Opinion, and US Presidential Approval Rates: A 5-Year Analysis of Online Political Discussions. Human Communication Research, 2012, vol. 38, no 2, p. 121-143.

[244] GONZÁLEZ-BAILÓN, Sandra; PALTOGLOU, Georgios. Signals of Public Opinion in Online Communication A Comparison of Methods and Data Sources. The ANNALS of the American Academy of Political and Social Science, 2015, vol. 659, no 1, p. 95-107.

[245] GOODIN, Robert E. Democratic deliberation within. Philosophy & Public Affairs, 2000, vol. 29, no 1, p. 81-109.

[246] GOOLSBY, Rebecca. Social media as crisis platform: The future of community maps/crisis maps. ACM Transactions on Intelligent Systems and Technology (TIST), 2010, vol. 1, no 1, p. 7.

[247] GORDON, Jill. John Stuart Mill and the “marketplace of ideas”. Social theory and practice, 1997, p. 235-249.

[248] GOULD, Roger V. Historical Sociology and Collective Action. In Julia Adams, Elisabeth S. Clemens, and Ann Shola Orloff, eds, Remaking Modernity: Politics, History, and Sociology, 2005, p. 286-299.

[249] GRAELLS-GARRIDO, Eduardo; LALMAS, Mounia; BAEZA-YATES, Ricardo. Finding Intermediary Topics Between People of Opposing Views: A Case Study. arXiv preprint arXiv:1506.00963, 2015.

[250] GRAHAM, Todd, et al. Between broadcasting political messages and interacting with voters: The use of Twitter during the 2010 UK general election campaign. Information, Communication & Society, 2013, vol. 16, no 5, p. 692-716.

[251] GRAHAM, Todd; JACKSON, Dan; BROERSMA, Marcel. New platform, old habits? Candidates’ use of Twitter during the 2010 British and Dutch general election campaigns. new media & society, 2014, p. 1461444814546728.

[252] GRANKA, Laura. Using online search traffic to predict US presidential elections. PS: Political Science & Politics, 2013, vol. 46, no 02, p. 271-279.

[253] GREENGARD, Samuel. The first Internet president. Communications of the ACM, 2009, vol. 52, no 2, p. 16-18.

[254] GRIER, Chris, et al. @ spam: the underground on 140 characters or less. In Proceedings of the 17th ACM conference on Computer and communications security. ACM, 2010. p. 27-37.

[255] GROW, Gabrielle; WARD, Janelle. The role of authenticity in electoral social media campaigns. First Monday, 2013, vol. 18, no 4.

[256] GRUZD, Anatoliy; ROY, Jeffrey. Investigating political polarization on Twitter: A Canadian perspective. Policy & Internet, 2014, vol. 6, no 1, p. 28-45.

[257] GUEORGUIEVA, Vassia; SIMON, Rita J. Voting and Elections the World Over. Lexington Books, 2009.

[258] GULATI, Girish J.; WILLIAMS, Christine B. Closing the gap, raising the bar: Candidate web site communication in the 2006 campaigns for congress. Social Science Computer Review, 2007, vol. 25, no 4, p. 443-465.

[259] GULATI, Girish J.; WILLIAMS, Christine B. Congressional candidates’ use of YouTube in 2008: Its frequency and rationale. Journal of Information Technology & Politics, 2010, vol. 7, no 2-3, p. 93-109.

[260] GUO, Chao; SAXTON, Gregory D. Tweeting social change: How social media are changing nonprofit advocacy. Nonprofit and Voluntary Sector Quarterly, 2013, p. 0899764012471585.

[261] GUPTA, Manish; ZHAO, Peixiang; HAN, Jiawei. Evaluating Event Credibility on Twitter. In SDM. 2012. p. 153-164.

[262] GUSTAFSSON, Nils. The subtle nature of Facebook politics: Swedish social network site users and political participation. New Media & Society, 2012, vol. 14, no 7, p. 1111-1127.

[263] HAAS, Tanni. From “Public Journalism” to the “Public’s Journalism”? Rhetoric and reality in the discourse on weblogs. Journalism Studies, 2005, vol. 6, no 3, p. 387-396.

[264] HABERMAS, Jürgen. Contributions to a discourse theory of law and democracy. Cambridge: Polity Press, 1996.

[265] HABERMAS, Jürgen. The structural transformation of the public sphere: An inquiry into a category of bourgeois society. MIT press, 1991.

[266] HACIYAKUPOGLU, Gulizar; ZHANG, Weiyu. Social Media and Trust during the Gezi Protests in Turkey. Journal of Computer-Mediated Communication, 2015, vol. 20, no. 4, pp. 450-466.

[267] HALPIN, Harry; BRIA, Francesca. D4.1 - State of the Art of social networking systems, identity ecosystem and social data stores. D-CENT Decentralised Citizens ENgagement Technologies, 2014. Available at: Last access: July 21st, 2015.

[268] HAMPTON, Keith, et al. Social media and the ‘spiral of silence’. Pew Research Center, Washington, DC. 2014.

[269] HARDAKER, Claire. Trolling in asynchronous computer-mediated communication: From user discussions to academic definitions. 2010.

[270] HARDT, Steve; LOPES, Lia CR. Google Votes: A Liquid Democracy Experiment on a Corporate Social Network. Technical Disclosure Commons, 2015. Available at: Last access: July 21st, 2015.

[271] HARDY, Bruce W.; SCHEUFELE, Dietram A. Examining differential gains from Internet use: Comparing the moderating role of talk and online interactions. Journal of Communication, 2005, vol. 55, no 1, p. 71-84.

[272] HARFOUSH, Rahaf. Yes We Did! An inside look at how social media built the Obama brand. New Riders, 2009.

[273] HARGITTAI, Eszter. Is Bigger Always Better? Potential Biases of Big Data Derived from Social Network Sites. The ANNALS of the American Academy of Political and Social Science, 2015, vol. 659, no 1, p. 63-76.

[274] HARGITTAI, Eszter; GALLO, Jason; KANE, Matthew. Cross-ideological discussions among conservative and liberal bloggers. Public Choice, 2008, vol. 134, no 1-2, p. 67-86.

[275] HARLOW, Summer. Adapting, Adopting and Diffusing: Leveraging Web 2.0 Tools for Activism in Mexico. Journal of Latin American Communication Research, 2013, vol. 3, no 1, p. 3-35.

[276] HASKELL, John. Direct democracy or representative government?: dispelling the populist myth. 1959. Reprint. Westview Pr, 2001.

[277] HASSID, Jonathan. Safety valve or pressure cooker? Blogs in Chinese political life. Journal of Communication, 2012, vol. 62, no 2, p. 212-230.

[278] HAUSLOHNER, Abigail. Is Egypt about to have a Facebook revolution?. Time Magazine, 2011, vol. 24. Available at:,8599,2044142,00.html Last access: July 20th, 2015.

[279] HAWTHORNE, Joshua; HOUSTON, J. Brian; MCKINNEY, Mitchell S. Live-tweeting a presidential primary debate: exploring new political conversations. Social Science Computer Review, 2013, p. 0894439313490643.

[280] HAY, Colin. Why we hate politics. John Wiley & Sons, 2013.

[281] HAYES, Rebecca A.; SMOCK, Andrew; CARR, Caleb T. Face [book] Management: Self-Presentation of Political Views on Social Media. Communication Studies, 2015, no ahead-of-print, p. 1-20.

[282] HE, Baogang. Western theories of deliberative democracy and the Chinese practice of complex deliberative governance. Search for deliberative democracy in China, 2012, p. 133-148.

[283] HE, Yulan, et al. Quantising opinions for political tweets analysis. In: LREC 2012, Eighth International Conference on Language Resources and Evaluation, 2012.

[284] HELFT, Miguel. YouTube blocked in China, Google says. The New York Times, 2009. Available at: Last access: June 25th, 2015.

[285] HEMPHILL, Libby; CULOTTA, Aron; HESTON, Matthew. Framing in Social Media: How the US Congress Uses Twitter Hashtags to Frame Political Issues. Available at SSRN 2317335, 2013.

[286] HERRICK, Matt. Eight Facts About ZunZuneo. Available at: Last access: June 23rd, 2015.

[287] HESS, Aaron. Resistance up in smoke: Analyzing the limitations of deliberation on YouTube. Critical Studies in Media Communication, 2009, vol. 26, no 5, p. 411-434.

[288] HEVERIN, Thomas; ZACH, Lisl. Twitter for city police department information sharing. Proceedings of the American Society for Information Science and Technology, 2010, vol. 47, no 1, p. 1-7.

[289] HILL, Kevin A.; HUGHES, John E. Is the internet an instrument of global democratization?. Democratization, 1999, vol. 6, no 2, p. 99-127.

[290] HILTZ, Starr Roxanne. The network nation: Human communication via computer. Mit Press, 1993.

[291] HIMELBOIM, Itai; MCCREERY, Stephen; SMITH, Marc. Birds of a feather tweet together: Integrating network and content analyses to examine cross-ideology exposure on Twitter. Journal of Computer-Mediated Communication, 2013, vol. 18, no 2, p. 40-60.

[292] HINDMAN, Matthew. The real lessons of Howard Dean: Reflections on the first digital campaign. Perspectives on Politics, 2005, vol. 3, no 01, p. 121-128.

[293] HOFFMAN, Lindsay H. Participation or communication? An explication of political activity in the Internet age. Journal of Information Technology & Politics, 2012, vol. 9, no 3, p. 217-233.

[294] HOGAN, Bernie; MELVILLE, Joshua. Social Media. International Encyclopedia of the Social & Behavioral Sciences, 2nd edition, Vol. 22. Oxford: Elsevier. pp. 421–425.

[295] HOGG, Chris. China Restores Xinjiang Internet. BBC News, May, 2010, vol. 14, p. 2010.

[296] HOLMES, O. W. Abrams v. United States, 250 U.S. 616 (1919). 1919.

[297] HONG, Sounman; NADLER, Daniel. Which candidates do the public discuss online in an election campaign?: The use of social media by 2012 presidential candidates and its impact on candidate salience. Government Information Quarterly, 2012, vol. 29, no 4, p. 455-461.

[298] HOSCH-DAYICAN, Bengü, et al. How Do Online Citizens Persuade Fellow Voters? Using Twitter During the 2012 Dutch Parliamentary Election Campaign. Social science computer review, 2014, p. 0894439314558200.

[299] HOWARD, Philip N., et al. Opening closed regimes: what was the role of social media during the Arab Spring?. Available at SSRN 2595096, 2011.

[300] HOWARD, Philip N.; HUSSAIN, Muzammil M. The Upheavals in Egypt and Tunisia: The role of digital media. Journal of democracy, 2011, vol. 22, no 3, p. 35-48.

[301] HUBERTY, Mark Edward. Multi-cycle forecasting of congressional elections with social media. In Proceedings of the 2nd workshop on Politics, Elections and Data. ACM, 2013. p. 23-30.

[302] HUBERTY, Mark. Awaiting the second big data revolution: from digital noise to value creation. Journal of Industry, Competition and Trade, 2015, vol. 15, no 1, p. 35-47.

[303] HUGHES, Neil. ‘Young People Took to the Streets and all of a Sudden all of the Political Parties Got Old’: The 15M Movement in Spain. Social Movement Studies, 2011, vol. 10, no 4, p. 407-413.

[304] HUSSAIN, Muzammil M.; HOWARD, Philip N. Democracy’s Fourth Wave? Information technologies and the fuzzy causes of the Arab Spring. Information Technologies and the Fuzzy Causes of the Arab Spring (March 27, 2012), 2012.

[305] INGBER, Stanley. The marketplace of ideas: a legitimizing myth. Duke Law Journal, 1984, p. 1-91.

[306] IOANNIDIS, J. P. A. Why most published research findings are false. PLoS Med, 2005, vol. 2, no 8, p. e124.

[307] IOSIFIDIS, Petros. The public sphere, social networks and public service media. Information, Communication & Society, 2011, vol. 14, no 5, p. 619-637.

[308] JACKSON, Daniel; THORSEN, Einar. UK Election Analysis 2015: Media, Voters and the Campaign. The Centre for the Study of Journalism, Culture and Community. 2015. Available at: Last access: June 17th, 2015.

[309] JACKSON, Nigel A.; LILLEKER, Darren G. MPs and E-representation: Me, MySpace and I. British Politics, 2009, vol. 4, no 2, p. 236-264.

[310] JACKSON, Nigel. Dipping their big toe into the blogosphere: The use of weblogs by the political parties in the 2005 general election. In Aslib Proceedings. Emerald Group Publishing Limited, 2006. p. 292-303.

[311] JACKSON, Nigel. Representation in the blogosphere: MPs and their new constituents. Parliamentary Affairs, 2008, vol. 61, no 4, p. 642-660.

[312] JACKSON, Nigel; LILLEKER, Darren. Microblogging, constituency service and impression management: UK MPs and the use of Twitter. The Journal of Legislative Studies, 2011, vol. 17, no 1, p. 86-105.

[313] JAEGER, Paul T.; BERTOT, John Carlo; SHILTON, Katie. Information policy and social media: Framing government—citizen web 2.0 interactions. En Web 2.0 Technologies and Democratic Governance. Springer New York, 2012. p. 11-25.

[314] JAEGER, Paul T.; PAQUETTE, Scott; SIMMONS, Shannon N. Information policy in national political campaigns: A comparison of the 2008 campaigns for President of the United States and Prime Minister of Canada. Journal of Information Technology & Politics, 2010, vol. 7, no 1, p. 67-82.

[315] JANOSKI-HAEHLEN, Emily M. The Courts Are All A ’Twitter’: The Implications of Social Media Use in the Courts. Valparaiso University Law Review, 2011, vol. 46, no 1, p. 43-68.

[316] JENNINGS, M. Kent; ZEITNER, Vicki. Internet use and civic engagement: A longitudinal analysis. Public Opinion Quarterly, 2003, vol. 67, no 3, p. 311-334.

[317] JENSEN, Jakob Linaa. Public Spheres on the Internet: Anarchic or Government-Sponsored–A Comparison. Scandinavian political studies, 2003, vol. 26, no 4, p. 349-374.

[318] JENSEN, Michael J.; ANSTEAD, Nick. Psephological investigations: Tweets, votes, and unknown unknowns in the republican nomination process. Policy & Internet, 2013, vol. 5, no 2, p. 161-182.

[319] JIANG, Min. Authoritarian deliberation on Chinese Internet. Electronic Journal of Communication, 2010, vol. 20.

[320] JIN, Xin, et al. The wisdom of social multimedia: using flickr for prediction and forecast. In Proceedings of the international conference on Multimedia. ACM, 2010. p. 1235-1244.

[321] JOHNSON, Martin. Non-profit organisations and the Internet. First Monday, 1999, vol. 4, no 2.

[322] JOHNSON, Thomas J.; ZHANG, Weiwu; BICHARD, Shannon L. Voices of convergence or conflict? A path analysis investigation of selective exposure to political websites. Social Science Computer Review, 2010.

[323] JUNGHERR, Andreas. Predictor of Electoral Success and Public Opinion at Large. In Analyzing Political Communication with Digital Trace Data. Springer International Publishing, 2015. p. 189-210.

[324] JUNGHERR, Andreas. The Media Connection. In Analyzing Political Communication with Digital Trace Data. Springer International Publishing, 2015. p. 155-188.

[325] JUNGHERR, Andreas; JÜRGENS, Pascal; SCHOEN, Harald. Why the Pirate Party won the German election of 2009 or the trouble with predictions: A response to tumasjan, a., sprenger, to, sander, pg, & welpe, im “predicting elections with twitter: What 140 characters reveal about political sentiment”. Social Science Computer Review, 2012, vol. 30, no 2, p. 229-234.

[326] JURGENS, David, et al. Geolocation prediction in twitter using social networks: A critical analysis and review of current practice. In Proceedings of the 9th International AAAI Conference on Weblogs and Social Media (ICWSM). 2015.

[327] JÜRGENS, Pascal; JUNGHERR, Andreas; SCHOEN, Harald. Small worlds with a difference: New gatekeepers and the filtering of political information on twitter. In Proceedings of the 3rd International Web Science Conference. ACM, 2011. p. 21.

[328] KAIMAN, Jonathan. Hong Kong’s umbrella revolution - the Guardian briefing. The Guardian, 2014. Available at: Last access: July 18th, 2015.

[329] KALATHIL, Shanthi; BOAS, Taylor C. Open networks, closed regimes: The impact of the Internet on authoritarian rule. Carnegie Endowment, 2010.

[330] KALATHIL, Shanthi; BOAS, Taylor C. The Internet and state control in authoritarian regimes: China, Cuba and the counterrevolution. First Monday, 2001, vol. 6, no 8.

[331] KALSNES, Bente; KRUMSVIK, Arne H.; STORSUL, Tanja. Social media as a political backchannel: Twitter use during televised election debates in Norway. Aslib Journal of Information Management, 2014, vol. 66, no 3, p. 313-328.

[332] KANG, Byungkyu; O’DONOVAN, John; HÖLLERER, Tobias. Modeling topic specific credibility on twitter. In Proceedings of the 2012 ACM international conference on Intelligent User Interfaces. ACM, 2012. p. 179-188.

[333] KANG, Seok; NORTON, Hanna E. Nonprofit organizations’ use of the World Wide Web: are they sufficiently fulfilling organizational goals?. Public Relations Review, 2004, vol. 30, no 3, p. 279-284.

[334] KARAGIANNOPOULOS, Vasileios. The role of the internet in political struggles: some conclusions from Iran and Egypt. New Political Science, 2012, vol. 34, no 2, p. 151-171.

[335] KARLSEN, Rune. A platform for individualized campaigning? Social media and Parliamentary candidates in the 2009 Norwegian election campaign. Policy & Internet, 2011, vol. 3, no 4, p. 1-25.

[336] KARLSEN, Rune. Followers are opinion leaders: The role of people in the flow of political communication on and beyond social networking sites. European Journal of Communication, 2015 vol. 30, no. 3, pp. 301-318.

[337] KATZ, James E.; BARRIS, Michael; JAIN, Anshul. The social media president: Barack Obama and the politics of digital engagement. Palgrave Macmillan, 2013.

[338] KATZ, Rita. Follow ISIS on Twitter: A Special Report on the Use of Social Media by Jihadists. Insite Blog on Terrorism & Extremism, 2014. Available at: Last access: July 19th, 2015.

[339] KEDZIE, Christopher. Communication and democracy: Coincident revolutions and the emergent dictators. 1997.

[340] KEKIC, Laza. The Economist Intelligence Unit’s index of democracy. The Economist, 2007, vol. 21, p. 1-11.

[341] KELLNER, Douglas. Habermas, the public sphere, and democracy: A critical intervention. Perspectives on Habermas, 2000, p. 259-288.

[342] KELLY, S. et al. Freedom on the Net 2014—Tightening the Net: Governments Expand Online Controls. 2014

[343] KENIX, Linda Jean. Nonprofit organizations’ perceptions and uses of the Internet. Television & New Media, 2008.

[344] KENSKI, Kate; STROUD, Natalie Jomini. Connections between Internet use and political efficacy, knowledge, and participation. Journal of broadcasting & electronic media, 2006, vol. 50, no 2, p. 173-192.

[345] KERBEL, Matthew R.; BLOOM, Joel David. Blog for America and civic involvement. The Harvard International Journal of Press/Politics, 2005, vol. 10, no 4, p. 3-27.

[346] KHAMIS, Sahar; GOLD, Paul B.; VAUGHN, Katherine. Beyond Egypt’s “Facebook Revolution” and Syria’s “YouTube Uprising:” Comparing Political Contexts, Actors and Communication Strategies. Arab Media & Society, 2012, vol. 15.

[347] KHATUA, Aparup, et al. Can# Twitter_Trends Predict Election Results? Evidence from 2014 Indian General Election. In System Sciences (HICSS), 2015 48th Hawaii International Conference on. IEEE, 2015. p. 1676-1685.

[348] KIM, Yonghwan. The contribution of social network sites to exposure to political difference: The relationships among SNSs, online political messaging, and exposure to cross-cutting perspectives. Computers in Human Behavior, 2011, vol. 27, no 2, p. 971-977.

[349] KIM, Young Ae; AHMAD, Muhammad A. Trust, distrust and lack of confidence of users in online social media-sharing communities. Knowledge-Based Systems, 2013, vol. 37, p. 438-450.

[350] KING, Eric. BILETA 2014 privacy and digital communications. International Review of Law, Computers & Technology, 2015, vol. 29, no 1, p. 4-15.

[351] KING, Gary; PAN, Jennifer; ROBERTS, Margaret E. How censorship in China allows government criticism but silences collective expression. American Political Science Review, 2013, vol. 107, no 02, p. 326-343.

[352] KLANG, Mathias; NOLIN, Jan. Disciplining social media: An analysis of social media policies in 26 Swedish municipalities. First Monday, 2011, vol. 16, no 8.

[353] KLASTRUP, Lisbeth; PEDERSEN, Pia Svejgaard. Blogging for election: The use and function of blogs as communication tool in a Danish Parliament election campaign. Internet research annual, 2006, vol. 4, p. 27-40.

[354] KLAUSEN, Jytte, et al. The YouTube Jihadists: A social network analysis of Al-Muhajiroun’s propaganda campaign. Perspectives on Terrorism, 2012, vol. 6, no 1.

[355] KLAUSEN, Jytte. Tweeting the Jihad: Social media networks of Western foreign fighters in Syria and Iraq. Studies in Conflict & Terrorism, 2015, vol. 38, no 1, p. 1-22.

[356] KLINGER, Ulrike. Mastering the art of social media: Swiss parties, the 2011 national election and digital challenges. Information, Communication & Society, 2013, vol. 16, no 5, p. 717-736.

[357] KLOTZ, Robert J. The sidetracked 2008 YouTube senate campaign. Journal of Information Technology & Politics, 2010, vol. 7, no 2-3, p. 110-123.

[358] KLUVER, Randolph, et al. (ed.). The Internet and national elections: A comparative study of Web campaigning. Routledge, 2007.

[359] KOBAYASHI, Tetsuro; IKEDA, Ken’ichi. Selective exposure in political web browsing: Empirical verification of ‘cyber-balkanization’ in Japan and the USA. Information, Communication & Society, 2009, vol. 12, no 6, p. 929-953.

[360] KOHUT, A.; RAINIE, L. Internet election news audience seeks convenience, familiar names. Pew Research Internet Project. 2000.

[361] KOLTSOVA, Olessia; KOLTCOV, Sergei. Mapping the public agenda with topic modeling: The case of the Russian livejournal. Policy & Internet, 2013, vol. 5, no 2, p. 207-227.

[362] KRAUT, Robert, et al. Internet paradox revisited. Journal of social issues, 2002, vol. 58, no 1, p. 49-74.

[363] KRAUT, Robert, et al. Internet paradox: A social technology that reduces social involvement and psychological well-being?. American psychologist, 1998, vol. 53, no 9, p. 1017.

[364] KROH, Martin; NEISS, Hannes. Internet Access and Political Engagement: Self-Selection or Causal Effect?. APSA 2009 Toronto Meeting Paper. 2009.

[365] KROLL, Andy. How occupy wall street really got started. Mother Jones, 2011, vol. 17.

[366] KRUEGER, Brian S. A comparison of conventional and Internet political mobilization. American Politics Research, 2006, vol. 34, no 6, p. 759-776.

[367] KRUEGER, Brian S. Assessing the Potential of Internet Political Participation in the United States A Resource Approach. American Politics Research, 2002, vol. 30, no 5, p. 476-498.

[368] KUYMULU, Mehmet Baris. Reclaiming the right to the city: Reflections on the urban uprisings in Turkey. City, 2013, vol. 17, no 3, p. 274-278.

[369] L 2014-1353, Loi renforçant les dispositions relatives à la lutte contre le terrorisme, Quatorzième Législature, France, 2014.

[370] LAMPOS, Vasileios; PREOTIUC-PIETRO, Daniel; COHN, Trevor. A user-centric model of voting intention from Social Media. In ACL (1). 2013. p. 993-1003.

[371] LANSDALL-WELFARE, Thomas; LAMPOS, Vasileios; CRISTIANINI, Nello. Nowcasting the mood of the nation. Significance, 2012, vol. 9, no 4, p. 26-28.

[372] LAPPAS, Georgios; KLEFTODIMOS, Alexandros; YANNAS, Prodromos. Greek parties and Web 2.0. Elections, Campaigning and Citizens Online, 2010.

[373] LARSSON, Anders Olof; MOE, Hallvard. Studying political microblogging: Twitter users in the 2010 Swedish election campaign. New Media & Society, 2012, vol. 14, no 5, p. 729-747.

[374] LASSEN, David S.; BROWN, Adam R. Twitter: The electoral connection?. Social Science Computer Review, 2011 vol. 29 no. 4 419-436

[375] LAWSON-BORDERS, Gracie; KIRK, Rita. Blogs in campaign communication. American Behavioral Scientist, 2005, vol. 49, no 4, p. 548-559.

[376] LAZER, David, et al. The parable of Google Flu: traps in big data analysis. Science, 2014, vol. 343, no 14 March.

[377] LEE, Byoungkwan; LANCENDORFER, Karen M.; LEE, Ki Jung. Agenda-setting and the Internet: The intermedia influence of Internet bulletin boards on newspaper coverage of the 2000 general election in South Korea. Asian Journal of Communication, 2005, vol. 15, no 1, p. 57-71.

[378] LEE, Eric. How the Internet Is Changing Unions. WorkingUSA, 2000, vol. 4, no 2, p. 56-72.

[379] LEE, Kyumin, et al. Content-driven detection of campaigns in social media. In Proceedings of the 20th ACM international conference on Information and knowledge management. ACM, 2011. p. 551-556.

[380] LESSIG, Lawrence. The new road to the White House: How grassroots blogs are transforming presidential politics. Wired Magazine, 2003, vol. 11, no 11.

[381] LESTON-BANDEIRA, Cristina. The impact of the internet on Parliaments: a legislative studies framework. Parliamentary Affairs, 2007, vol. 60, no 4, p. 655-674.

[382] LÉVI-STRAUSS, Claude. The savage mind. The Garden City Press Limited, 1966.

[383] LEVITSKY, Steven; WAY, Lucan. The rise of competitive authoritarianism. Journal of democracy, 2002, vol. 13, no 2, p. 51-65.

[384] LEWICKI, Don; ZIAUKAS, Tim. The digital tea leaves of election 2000: The Internet and the future of presidential politics. First Monday, 2000, vol. 5, no 12.

[385] LEWIS, Paul, et al. Reading the riots: investigating England’s summer of disorder. 2011.

[386] LEWIS, Paul. Tottenham riots: a peaceful protest, then suddenly all hell broke loose. The Guardian, 2011. Available at: Last access: July 18th, 2015.

[387] LEYVA GARCÍA, Heriberto. Reforma al artículo 27 de la Constitución Política de los Estados Unidos Mexicanos (Amendment to article 27 of the Political Constitution of the United Mexican States). Revista de la Facultad de Derecho de México, 1992, no. 183-184. Available at: Last access: July 17th, 2015.

[388] LILLEKER, Darren G. Key concepts in political communication. Sage, 2006.

[389] LILLEKER, Darren G.; JACKSON, Nigel A. Towards a more participatory style of election campaigning: The impact of Web 2.0 on the UK 2010 general election. Policy & Internet, 2010, vol. 2, no 3, p. 69-98.

[390] LILLEKER, Darren G.; KOC-MICHALSKA, Karolina. Online political communication strategies: MEPs, e-representation, and self-representation. Journal of Information Technology & Politics, 2013, vol. 10, no 2, p. 190-207.

[391] LILLEKER, Darren G.; MALAGÓN, Casilda. Levels of interactivity in the 2007 French Presidential candidates’ Websites. European Journal of Communication, 2010, vol. 25, no 1, p. 25-42.

[392] LIM, Merlyna. Clicks, cabs, and coffee houses: Social media and oppositional movements in Egypt, 2004–2011. Journal of Communication, 2012, vol. 62, no 2, pp. 231-248.

[393] LIN, Jimmy. On Building Better Mousetraps and Understanding the Human Condition Reflections on Big Data in the Social Sciences. The ANNALS of the American Academy of Political and Social Science, 2015, vol. 659, no 1, p. 33-47.

[394] LINDNER, Ralf; RIEHM, Ulrich. Broadening Participation Through E-Petitions? An Empirical Study of Petitions to the German Parliament. Policy & Internet, 2011, vol. 3, no 1, p. 1-23.

[395] LIPINSKI, Daniel; NEDDENRIEP, Gregory. Using “New” Media to Get “Old” Media Coverage How Members of Congress Utilize Their Web Sites to Court Journalists. The International Journal of Press/Politics, 2004, vol. 9, no 1, p. 7-21.

[396] LIPPMANN, Walter. Public opinion. 1922. Reprint. Dover Publications, 2004.

[397] LIU, Bing. Sentiment Analysis—Mining Opinions, Sentiments, and Emotions. Cambridge University Press, 2015.

[398] LIVNE, Avishay, et al. The Party Is Over Here: Structure and Content in the 2010 Election. ICWSM, 2011, vol. 11, p. 17-21.

[399] LOKOT, Tetyana. Russia Launches ‘Predictive System’ for Monitoring Protest Activity Online. Global Voices, 2015. Available at: Last access: June 23rd, 2015.

[400] LOVEJOY, Kristen; SAXTON, Gregory D. Information, community, and action: how nonprofit organizations use social media. Journal of Computer-Mediated Communication, 2012, vol. 17, no 3, p. 337-353.

[401] LUCORE, Robert E. Challenges and opportunities: unions confront the new information technologies. Journal of Labor research, 2002, vol. 23, no 2, p. 201-214.

[402] LUI, Catherine; METAXAS, P. Takis; MUSTAFARAJ, Eni. On the predictability of the US elections through search volume activity. In Proceedings of the IADIS International Conference on e-Society, 2011.

[403] LYNCH, Marc. After Egypt: The limits and promise of online challenges to the authoritarian Arab state. Perspectives on politics, 2011, vol. 9, no 02, pp. 301-310.

[404] MACKINNON, Rebecca et al. “Race to the Bottom”—Corporate Complicity in Chinese Internet Censorship. Human Rights Watch, vol. 18, no. 8 (C), 2006. Available at: Last access: June 23rd, 2015.

[405] MACKINNON, Rebecca. China’s Censorship 2.0: How companies censor bloggers. First Monday, 2009, vol. 14, no 2.

[406] MACKINNON, Rebecca. Networked authoritarianism in China and beyond: Implications for global internet freedom. Liberation Technology in Authoritarian Regimes, Stanford University, 2010.

[407] MAGRO, Michael J. A review of social media use in e-government. Administrative Sciences, 2012, vol. 2, no 2, p. 148-161.

[408] MAHMOOD, Tariq, et al. Mining Twitter big data to predict 2013 Pakistan election winner. In Multi Topic Conference (INMIC), 2013 16th International. IEEE, 2013. p. 49-54.

[409] MANCINI, Pia. Why it is time to redesign our political system. European View, 2015, vol. 14, no 1, pp. 69-75.

[410] MANETTO, Francisco. Iglesias afirma que “el famoso SMS del ‘Pásalo’ se gestó” en su entorno (Iglesias claims that "the famous ’Pass It On’ SMS was conceived" by people close to him). El País, 2014. Available at: Last access: July 17th, 2015.

[411] MARCHETTI, Rita; CECCOBELLI, Diego. Twitter and Television in a Hybrid Media System: The 2013 Italian election campaign. Journalism Practice, 2015, no ahead-of-print, p. 1-19.

[412] MARGOLIS, Michael; RESNICK, David. Politics as Usual: The Cyberspace Revolution. Sage Publications, 2000.

[413] MARGOLIS, Michael; RESNICK, David. Politics as Usual: The Cyberspace Revolution’. Sage Publications, 2000.

[414] MARGOLIS, Michael; RESNICK, David. Politics as Usual: The Cyberspace ’Revolution’. Sage Publications, 2000.

[415] MARKUS, Hazel; NURIUS, Paula. Possible selves. American psychologist, 1986, vol. 41, no 9, pp. 954-969.

[416] MARTINEZ-TORRES, Maria Elena. Civil society, the Internet, and the Zapatistas. Peace Review, 2001, vol. 13, no 3, p. 347-355.

[417] MCCHESNEY, Robert W. The Internet and US communication policy-making in historical and critical perspective. Journal of Communication, 1996, vol. 46, no 1, p. 98-124.

[418] MCCOMBS, Maxwell E.; SHAW, Donald L. The agenda-setting function of mass media. Public opinion quarterly, 1972, pp. 176-187.

[419] MCLEAN, Iain (Ed.). Oxford concise dictionary of politics. Oxford University Press UK, 2003.

[420] MCLEAN, Iain; MCMILLAN, Alistair. The concise Oxford dictionary of politics. OUP Oxford, 2009.

[421] MCNAIR, Brian. An introduction to political communication. Taylor & Francis, 2011.

[422] MCPHAIL, Clark; MCCARTHY, John D. Protest mobilization, protest repression, and their interaction. Repression and mobilization, 2005, p. 3-32.

[423] MCSWEENEY, Dean. The Front-Runner Falls The Failure of Howard Dean. Party Politics, 2007, vol. 13, no 1, p. 109-126.

[424] MEEKS, Lindsey. Blogs and Politicians. Available at: Last access: July 5th, 2015.

[425] MEHNDIRATTA, Pulkit, et al. Elections Again, Twitter May Help!!! A Large Scale Study for Predicting Election Results Using Twitter. In Big Data Analytics. Springer International Publishing, 2014. p. 133-144.

[426] MEIJER, Albert; THAENS, Marcel. Social media strategies: Understanding the differences between North American police departments. Government Information Quarterly, 2013, vol. 30, no 4, p. 343-350.

[427] MEJÍAS, Ulises. The Twitter Revolution Must Die. International Journal of Learning and Media, 2010, vol. 2, no. 4, pp. 3-5.

[428] MEJOVA, Yelena; SRINIVASAN, Padmini; BOYNTON, Bob. GOP primary season on twitter: popular political sentiment in social media. In Proceedings of the sixth ACM international conference on Web search and data mining. ACM, 2013. p. 517-526.

[429] MERGEL, Ines. Social media adoption and resulting tactics in the US federal government. Government Information Quarterly, 2013, vol. 30, no 2, p. 123-130.

[430] MERGEL, Ines. The use of social media to dissolve knowledge silos in government. The future of public administration, public management, and public service around the world, 2011.

[431] MESSNER, Marcus; GARRISON, Bruce. Study shows some blogs affect traditional news media agendas. Newspaper Research Journal, 2011, vol. 32, no 3, p. 112.

[432] METAXAS, Panagiotis T.; MUSTAFARAJ, Eni. Social Media and the Elections. Science, 2012, vol. 338, no 6106, p. 472-473.

[433] METAXAS, Panagiotis T.; MUSTAFARAJ, Eni; GAYO-AVELLO, Daniel. How (not) to predict elections. In Privacy, Security, Risk and Trust (PASSAT) and 2011 IEEE Third Inernational Conference on Social Computing (SocialCom), 2011 IEEE Third International Conference on. IEEE, 2011. p. 165-171.

[434] METAXAS, Panagiotis Takas; FINN, Samantha; MUSTAFARAJ, Eni. Using TwitterTrails. com to Investigate Rumor Propagation. In Proceedings of the 18th ACM Conference Companion on Computer Supported Cooperative Work & Social Computing. ACM, 2015. p. 69-72.

[435] METAXAS, Panagiotis; MUSTAFARAJ, Eni. The battle for the 2008 US Congressional Elections on the Web. 2009.

[436] MICONI, Andrea. Italy’s “Five Stars” movement and the role of a leader: Or, how charismatic power can resurface through the web. New Media & Society, 2014, p. 1461444814520872.

[437] MIGA, Andrew. Dean fans flog blog, rip Kerry to threads. Boston Herald, August 13 2003, p. 6.

[438] MILBRATH, Lester W. Political participation. In The handbook of political behavior. Springer US, 1981. p. 197-240.

[439] MISLOVE, Alan, et al. Understanding the Demographics of Twitter Users. ICWSM, 2011, vol. 11, p. 5th.

[440] MITCHELL, Amy; GOTTFRIED, Jeffrey; MATSA, Katerina Eva. Millennials & Political News: Social Media - the Local TV for the Next Generation? Pew Research Center, 2015.

[441] MITCHELL, Amy; HITLIN, Paul. Twitter reaction to events often at odds with overall public opinion. Pew Research Center, 2013, vol. 4.

[442] MOCANU, Delia, et al. Collective attention in the age of (mis) information. Computers in Human Behavior, 2015.

[443] MONTALVO, Jose G. Re-examining the evidence on the electoral impact of terrorist attacks: The Spanish election of 2004. Electoral Studies, 2012, vol. 31, no 1, p. 96-106.

[444] MOORE, Jack. Turkey Lists Social Media as Top Threat in Leaked Document. Newsweek, 2015. Available at: Last access: June 24th, 2015.

[445] MOORE, Malcolm. China ’has breached terms of Hong Kong handover’. The Telegraph, 2014. Available at: Last access: July 18th, 2015.

[446] MORILLON, L.; JULLIARD, J. Enemies of the Internet: Web 2.0 versus Control 2.0. Reporters Without Borders. 2010.

[447] MOROZOV, Evgeny. How dictators watch us on the web. Prospect magazine, 2009a, vol. 18, no 11, p. 2009. Available at: Last access: June 23rd, 2015.

[448] MOROZOV, Evgeny. Iran: Downside to the" Twitter Revolution". Dissent, 2009a, vol. 56, no 4, p. 10-14.

[449] MOROZOV, Evgeny. Iran: Downside to the" Twitter Revolution". Dissent, 2009b, vol. 56, no 4, p. 10-14.

[450] MOROZOV, Evgeny. The brave new world of slacktivism. Foreign Policy, 2009, vol. 19, no 05. Available at: Last access: July 15th, 2015.

[451] MOROZOV, Evgeny. The brave new world of slacktivism. Foreign Policy, 2009b, vol. 19, no 05. Available at: Last access: July 15th, 2015.

[452] MOROZOV, Evgeny. The Internet: A room of our own?. Dissent, 2009c, vol. 56, no 3, p. 80-85.

[453] MOROZOV, Evgeny. The net delusion: The dark side of Internet freedom. PublicAffairs, 2012.

[454] MOROZOV, Evgeny. Why the internet is failing Iran’s activists. Prospect Magazine, 2010, vol. 166. Available at: Last access: July 20th,2015.

[455] MOROZOV, Evgeny. Why the internet is failing Iran’s activists. Prospect Magazine, 2010, vol. 166. Available at: Last access: June 23rd, 2015.

[456] MORRIS, Sarah et al. Spain to pass measures to tackle homegrown Islamist militants. Reuters, 2015. Available at: Last access: June 27th, 2015.

[457] MORSTATTER, Fred, et al. Is the Sample Good Enough? Comparing Data from Twitter’s Streaming API with Twitter’s Firehose. 2013. In Proceedings of ICWSM.

[458] MOSSBERGER, Karen; WU, Yonghong; CRAWFORD, Jared. Connecting citizens and local governments? Social media and interactivity in major US cities. Government Information Quarterly, 2013, vol. 30, no 4, p. 351-358.

[459] MOUFFE, Chantal. Deliberative democracy or agonistic pluralism?. Social research, 1999, p. 745-758.

[460] MURLEY, Bryan; ROBERTS, Chris. Biting the hand that feeds: Blogs and second-level agenda setting. in Media Convergence Conference, Provo, UT. 2005.

[461] MURPHY, Joe, et al. Social Media in Public Opinion Research: Report of the AAPOR Task Force on Emerging Technologies in Public Opinion Research. American Association for Public Opinion Research, 2014.

[462] MURTHY, Dhiraj; PETTO, Laura R. Comparing Print Coverage and Tweets in Elections A Case Study of the 2011–2012 US Republican Primaries. Social Science Computer Review, 2015, vol. 33, no 3, p. 298-314.

[463] MUSTAFARAJ, Eni, et al. Vocal minority versus silent majority: Discovering the opinions of the long tail. In Privacy, Security, Risk and Trust (PASSAT) and 2011 IEEE Third Inernational Conference on Social Computing (SocialCom), 2011 IEEE Third International Conference on. IEEE, 2011. p. 103-110.

[464] MUSTAFARAJ, Eni; METAXAS, Panagiotis. From Obscurity to Prominence in Minutes: Political Speech and Real-Time Search. 2010.

[465] MUTHIAH, Sathappan, et al. Planned Protest Modeling in News and Social Media. In Proceedings of Innovative Applications of Artificial Intelligence, Twenty-Seventh IAAI Conference, 2015.

[466] NAGLER, Jonathan; TUCKER, Joshua A. Drawing Inferences and Testing Theories with Big Data. PS: Political Science & Politics, 2015, vol. 48, no 01, p. 84-88.

[467] NAHON, Karine, et al. Fifteen minutes of fame: The power of blogs in the lifecycle of viral political information. Policy & Internet, 2011, vol. 3, no 1, p. 1-28.

[468] NANABHAY, Mohamed; FARMANFARMAIAN, Roxane. From spectacle to spectacular: How physical space, social media and mainstream broadcast amplified the public sphere in Egypt’s ‘Revolution’. The Journal of North African Studies, 2011, vol. 16, no 4, p. 573-603.

[469] NEUMAN, W. Russell, et al. The Dynamics of Public Attention: Agenda-Setting Theory Meets Big Data. Journal of Communication, 2014, vol. 64, no 2, p. 193-214.

[470] NEWMAN, Nathan. Is labor missing the Internet third wave?. WorkingUSA, 2005, vol. 8, no 4, p. 383-394.

[471] NISBET, Matthew C.; SCHEUFELE, Dietram A. Political talk as a catalyst for online citizenry. Journalism & Mass Communication Quarterly, 2004, vol. 81, no 4, p. 877-896.

[472] NIXON, Paul; JOHANSSON, Hans. Transparency through technology: the Internet and political parties. Digital democracy: Discourse and decision making in the information age, 1999, p. 135-153.

[473] NOELLE-NEUMANN, Elisabeth. The spiral of silence a theory of public opinion. Journal of communication, 1974, vol. 24, no 2, p. 43-51.

[474] NORRIS, Pippa. Digital divide: Civic engagement, information poverty, and the Internet worldwide. Cambridge University Press, 2001.

[475] NORRIS, Pippa. New Social Movements, Protest Politics, and the Internet. In Democratic phoenix: Reinventing political activism. Cambridge University Press, 2002, pp. 188-212.

[476] NORRIS, Pippa. Who Surfs Cafe Europa? Paper for the Annual Meeting of the American Political Science Association, Atlanta. 1999.

[477] NORTON, Philip. Four models of political representation: British MPs and the use of ICT. The Journal of Legislative Studies, 2007, vol. 13, no 3, p. 354-369.

[478] O’SULLIVAN, Patrick B.; FLANAGIN, Andrew J. Reconceptualizing ‘flaming’and other problematic messages. New Media & Society, 2003, vol. 5, no 1, p. 69-94.

[479] OBAMA, Barack (POTUS). “Hello, Twitter! It’s Barack. Really! Six years in, they’re finally giving me my own account.” 18 May 2015, 5:38 p.m. Tweet. Available at: Last access: May 25th, 2015.

[480] OBAR, Jonathan A.; ZUBE, Paul; LAMPE, Clifford. Advocacy 2.0: An analysis of how advocacy groups in the United States perceive and use social media as tools for facilitating civic engagement and collective action. Journal of Information Policy, 2012, vol. 2, p. 1-25.

[481] O’CONNOR, Brendan, et al. From Tweets to Polls: Linking Text Sentiment to Public Opinion Time Series. ICWSM, 2010, vol. 11, no 122-129, p. 1.2.

[482] of-followers.html. Last access: June 11th, 2015.

[483] OLIVEIRA, Gustavo Henrique Maultasch; WELCH, Eric W. Social media use in local government: Linkage of technology, task, and organizational context. Government Information Quarterly, 2013, vol. 30, no 4, p. 397-405.

[484] OMAND, David; BARTLETT, Jamie; MILLER, Carl. Introducing social media intelligence (SOCMINT). Intelligence and National Security, 2012, vol. 27, no 6, p. 801-823.

[485] ORCUTT, Mike. Twitter mischief plagues mexico’s election. MIT Technology Review. 2012. Available at: Last access: June 11th, 2015.

[486] ORTIZ, Isabel, et al. World Protests 2006-2013. Initiative for Policy Dialogue and Friedrich-Ebert-Stiftung, 2013. Available at: Last access: July 16th, 2015.

[487] ORUCOGLU, Berivan. After Gezi—How Erdogan uses information to control society. In The New Authoritarians: Ruling Through Disinformation (pp. 14-19). Legatum Institute, 2015. Available at: Last access: June 24th, 2015.

[488] OSER, Jennifer; HOOGHE, Marc; MARIEN, Sofie. Is online participation distinct from offline participation? A latent class analysis of participation types and their stratification. Political Research Quarterly, 2013, vol. 66, no 1, p. 91-101.

[489] OSIMO, David. Web 2.0 in government: Why and how. Institute for Prospective Technological Studies (IPTS), JRC, European Commission, EUR, 2008, vol. 23358.

[490] ÖSTMAN, Johan. Information, expression, participation: How involvement in user-generated content relates to democratic engagement among young people. New Media & Society, 2012, vol. 14, no 6, p. 1004-1021.

[491] OUNIS, Iadh, et al. Overview of the trec-2011 microblog track. In Proceeddings of the 20th Text REtrieval Conference (TREC 2011). 2011.

[492] OWEN, Diana; DAVIS, Richard. Presidential communication in the Internet era. Presidential Studies Quarterly, 2008, vol. 38, no 4, p. 658-673.

[493] PAGE, Benjamin I. The mass media as political actors. PS: political science & politics, 1996, vol. 29, no 01, p. 20-24.

[494] PANG, Bo; LEE, Lillian. Opinion mining and sentiment analysis. Foundations and trends in information retrieval, 2008, vol. 2, no 1-2, p. 1-135.

[495] PAPACHARISSI, Zizi. A private sphere: Democracy in a digital age. Polity, 2010.

[496] PAPACHARISSI, Zizi. Democracy online: Civility, politeness, and the democratic potential of online political discussion groups. New Media & Society, 2004, vol. 6, no 2, p. 259-283.

[497] PARK, Chang Sup. Does Twitter motivate involvement in politics? Tweeting, opinion leadership, and political engagement. Computers in Human Behavior, 2013, vol. 29, no 4, p. 1641-1648.

[498] PARMELEE, John H. The agenda-building function of political tweets. New Media & Society, 2013, p. 1461444813487955.

[499] PATTERSON, Thomas E. The news media: An effective political actor?. Political Communication, 1997, vol. 14, no 4, p. 445-455.

[500] PEARCE, Katy E.; KENDZIOR, Sarah. Networked authoritarianism and social media in Azerbaijan. Journal of Communication, 2012, vol. 62, no 2, p. 283-298.

[501] PENNACCHIOTTI, Marco; POPESCU, Ana-Maria. A Machine Learning Approach to Twitter User Classification. ICWSM, 2011a, vol. 11, p. 281-288.

[502] PENNACCHIOTTI, Marco; POPESCU, Ana-Maria. Democrats, republicans and starbucks aficionados: user classification in twitter. In Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining. ACM, 2011b. p. 430-438.

[503] PENNEY, Joel; DADAS, Caroline. (Re) Tweeting in the service of protest: Digital composition and circulation in the Occupy Wall Street movement. New Media & Society, 2013, p. 1461444813479593.

[504] PEPE, Alberto; DI GENNARO, Corinna. Political protest Italian-style: The blogosphere and mainstream media in the promotion and coverage of Beppe Grillo’s V-day. First Monday, 2009, vol. 14, no 12.

[505] PEW RESEARCH CENTER. McCain vs. Obama on the Web: A Study of the Presidential Candidate Web Sites. 2008. Available at: Last access: July 6th, 2015.

[506] PFEIFLE, Mark. A Nobel Peace Prize for Twitter?. The Christian Science Monitor, 2009. Available at: Last access: July 20th, 2015.

[507] PHILLIPS, Kevin. Virtual Washington. Time, 145,(12), 1995, p. 65-68.

[508] PIMENTA, Fernanda; OBRADOVIC, Darko; DENGEL, Andreas. A Comparative Study of Social Media Prediction Potential in the 2012 US Republican Presidential Preelections. In Cloud and Green Computing (CGC), 2013 Third International Conference on. IEEE, 2013. p. 226-232.

[509] PL 2110, Projet de loi renforçant les dispositions relatives à la lutte contre le terrorisme, Quatorzième Législature, France, 2014. Available at: Last access: June 24th, 2015.

[510] PLASSER, Fritz; PLASSER, Gunda. Global political campaigning: A worldwide analysis of campaign professionals and their practices. Greenwood Publishing Group, 2002.

[511] PLATTS-FOWLER, Deborah. Beyond the loot. Papers from the British Criminology Conference, 2013, vol. 13, pp. 17-32.

[512] PLEMING, Sue. US State Department speaks to Twitter over Iran. Reuters, June, 2009, vol. 16. Available at: Last access: June 25th, 2015.

[513] POOR, Nathaniel. Mechanisms of an online public sphere: The website Slashdot. Journal of Computer-Mediated Communication, 2005, vol. 10, no 2.

[514] POSTMES, Tom; BRUNSTING, Suzanne. Collective action in the age of the internet mass communication and online mobilization. Social Science Computer Review, 2002, vol. 20, no 3, pp. 290-301.

[515] PRESTON, Jennifer. Seeking to disrupt protesters, Syria cracks down on social media. New York Times, 2011, vol. 22.

[516] PRICE, Vincent; CAPPELLA, Joseph N. Online deliberation and its influence: The electronic dialogue project in campaign 2000. IT & Society, 2002, vol. 1, no 1, p. 303-329.

[517] PUTNAM, Robert D. Bowling alone: The collapse and revival of American community. Simon and Schuster, 2000.

[518] PUTNAM, Robert D. Tuning in, tuning out: The strange disappearance of social capital in America. PS: Political science & politics, 1995, vol. 28, no 04, p. 664-683.

[519] QAZVINIAN, Vahed, et al. Rumor has it: Identifying misinformation in microblogs. In Proceedings of the Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics, 2011. p. 1589-1599.

[520] QIANG, Xiao. The rise of online public opinion and its political impact. Changing media, changing China, 2011, p. 202-24.

[521] RAFAEL, Vicente L. The cell phone and the crowd: Messianic politics in the contemporary Philippines. Philippine Political Science Journal, 2003, vol. 24, no 47, p. 3-36.

[522] RAHIMI, Babak. Cyberdissent: the internet in revolutionary Iran. Middle East, 2003, vol. 7, no 3, p. 102.

[523] RAINIE, Harrison; WELLMAN, Barry. Networked: The new social operating system. Cambridge, MA: Mit Press, 2012.

[524] RAINIE, Lee, et al. Social media and political engagement. Pew Internet & American Life Project, 2012.

[525] RAINIE, Lee. Social media and voting. Pew Research Center Internet and American Life Project, 2012.

[526] RAINIE, Lee; HORRIGAN, John B., CORNFIELD, Michael; The Internet and campaign 2004. Pew Internet & American Life Project, 2005.

[527] RAINIE, Lee; SMITH, Aaron. Social networking sites and politics. Washington, DC: Pew Internet & American Life Project. 2012.

[528] RAINIE, Lee; WELLMAN, Barry. Networked: The New Social Operating System. MIT Press, 2012.

[529] RAMESH, Randeep; HERN, Alex. Conservative party deletes archive of speeches from internet. The Guardian, 2013. Available at: Last access: July 7th, 2015.

[530] RAO, Delip, et al. Classifying latent user attributes in twitter. In Proceedings of the 2nd international workshop on Search and mining user-generated contents. ACM, 2010. p. 37-44.

[531] RATKIEWICZ, Jacob, et al. Detecting and Tracking Political Abuse in Social Media. In ICWSM. 2011a.

[532] RATKIEWICZ, Jacob, et al. Truthy: mapping the spread of astroturf in microblog streams. In Proceedings of the 20th international conference companion on World wide web. ACM, 2011b. p. 249-252.

[533] RAZZAQ, Muhammad Asif; QAMAR, Ali Mustafa; BILAL, Hafiz Syed Muhammad. Prediction and analysis of Pakistan election 2013 based on sentiment analysis. In Advances in Social Networks Analysis and Mining (ASONAM), 2014 IEEE/ACM International Conference on. IEEE, 2014. p. 700-703.

[534] REBER, Bryan H.; KIM, Jun Kyo. How activist groups use websites in media relations: Evaluating online press rooms. Journal of Public Relations Research, 2006, vol. 18, no 4, p. 313-333.

[535] RESNICK, David. Politics on the Internet: The normalization of cyberspace. New Political Science, 1997, p. 47-68. Available at: Last access: May 27th, 2015.

[536] REUTERS. TIMELINE - Iran’s Ahmadinejad since 2009 re-election. Reuters, 2010. Available at: Last access: July 20th, 2015.

[537] RHEINGOLD, Howard. The virtual community: Homesteading on the electronic frontier. MIT press, 1993.

[538] RICH, Frank. Wallflowers at the Revolution. The New York Times, 2011, vol. 5.

[539] ROBERTS, Hal, et al. 2010 circumvention tool usage report. The Berkman Center for Internet & Society. 2010.

[540] ROBERTS, Marilyn; WANTA, Wayne; DZWO, Tzong-Horng Dustin. Agenda setting and issue salience online. Communication Research, 2002, vol. 29, no 4, p. 452-465.

[541] ROBERTSON, David. The Routledge dictionary of politics. Routledge, 2004.

[542] RODAN, Garry. Embracing electronic media but suppressing civil society: authoritarian consolidation in Singapore. The Pacific Review, 2003, vol. 16, no 4, p. 503-524.

[543] ROGERS, Gareth, et al. D5.3 Implementation of collaborative policy-making tool. D-CENT Decentralised Citizens ENgagement Technologies, 2015. Last access: July 21st, 2015.

[544] RONFELDT, David, et al. The Zapatista" social Netwar" in Mexico. Rand Corporation, 1999.

[545] ROPER, Juliet. New Zealand Political Parties Online: the World Wide Web as a tool for democratization or for political marketing?. New Political Science, 1997, p. 69-84.

[546] ROSANVALLON, Pierre. Counter-democracy. Politics in an Age of Distrust, 2008.

[547] ROSENSTIEL, Tom; JURKOWITZ, Mark. Twitter and the Campaign. Pew Research Center, 2011. Available at: Last access: July 9th, 2015.

[548] ROSENSTIEL, Tom; MITCHELL, Amy; JURKOWITZ, Mark. The Final Days of the Media Campaign 2012. Pew Research Center, 2012b. Available at: Last access: July 9th, 2015.

[549] ROSENSTIEL, Tom; MITCHELL, Amy; JURKOWITZ, Mark. Winning the Media Campaign 2012. Pew Research Center, 2012a. Available at: Last access: July 9th, 2015.

[550] ROSS, Karen; BÜRGER, Tobias. Face to face (book) Social media, political campaigning and the unbearable lightness of being there. Political Science, 2014, vol. 66, no 1, p. 46-62.

[551] SÆBØ, Øystein. Understanding Twitter use among parliament representatives: A genre analysis. In Electronic participation. Springer Berlin Heidelberg, 2011. p. 1-12.

[552] SÁNCHEZ ALMEIDA, Carlos. El triunfo del terrorismo. El Mundo, 2015. Available at: Last access: June 27th, 2015.

[553] SÁNCHEZ, Juan Luis. Los primeros 40 de Sol (The first 40 persons in Sol)., 2015. Available at: Last access: July 18th, 2015.

[554] SANDERS, Eric; VAN DEN BOSCH, Antal. Relating Political Party Mentions on Twitter with Polls and Election Results. In DIR. 2013. p. 68-71.

[555] SANG, Erik Tjong Kim; BOS, Johan. Predicting the 2011 dutch senate election results with twitter. In Proceedings of the Workshop on Semantic Analysis in Social Media. Association for Computational Linguistics, 2012. p. 53-60.

[556] SARTORI, Giovanni. Parties and party systems: A framework for analysis. 1976. Reprint. ECPR press, 2005.

[557] SAXTON, Gregory D.; WATERS, Richard D. What do stakeholders like on Facebook? Examining public reactions to nonprofit organizations’ informational, promotional, and community-building messages. Journal of Public Relations Research, 2014, vol. 26, no 3, p. 280-299.

[558] SAYYADI, Hassan; HURST, Matthew; MAYKOV, Alexey. Event Detection and Tracking in Social Streams. In ICWSM. 2009.

[559] SCHLACHTER, Eric. Cyberspace, the Free Market and the Free Marketplace of Ideas: Recognizing Legal Differences in Computer Bulletin Board Functions. Hastings Comm. & Ent. LJ, 1993, vol. 16, p. 87.

[560] SCHLOZMAN, Kay Lehman; VERBA, Sidney; BRADY, Henry E. Weapon of the strong? Participatory inequality and the Internet. Perspectives on Politics, 2010, vol. 8, no 02, p. 487-509.

[561] SCHNEIDER, Steven M. Creating a Democratic Public Sphere Through Political Discussion A Case Study of Abortion Conversation on the Internet. Social Science Computer Review, 1996, vol. 14, no 4, p. 373-393.

[562] SCHRADIE, Jen. Qualitative Political Communication| Labor Unions, Social Media, and Political Ideology: Using the Internet to Reach the Powerful or Mobilize the Powerless?. International Journal of Communication, 2015, vol. 9, p. 21.

[563] SCHUDSON, Michael. Why conversation is not the soul of democracy. Critical Studies in Media Communication, 1997, vol. 14, no 4, p. 297-309.

[564] SCHWEITZER, Eva Johanna. Election Campaigning Online German Party Websites in the 2002 National Elections. European Journal of Communication, 2005, vol. 20, no 3, p. 327-351.

[565] SCIO (State Council Information Office). The Internet in China. 2010. Available at: Last access: June 24th, 2015.

[566] SHAH, Dhavan V., et al. Information and expression in a digital age modeling Internet effects on civic participation. Communication research, 2005, vol. 32, no 5, p. 531-565.

[567] SHAH, Dhavan V.; KWAK, Nojin; HOLBERT, R. Lance. "Connecting" and "Disconnecting" With Civic Life: Patterns of Internet Use and the Production of Social Capital. Political Communication, 2001, vol. 18, no 2, p. 141-162.

[568] SHAMMA, David A.; KENNEDY, Lyndon; CHURCHILL, Elizabeth F. Tweet the debates: understanding community annotation of uncollected sources. In Proceedings of the first SIGMM workshop on Social media. ACM, 2009. p. 3-10.

[569] SHAPIRO, Andrew L. The internet. Foreign Policy, 1999, p. 14-27.

[570] SHAPIRO, Robert Y.; MAHAJAN, Harpreet. Gender Differences in Policy Preferences: A Summary of Trends from the 1960s to the 1980s. Public Opinion Quarterly, 1986, vol. 50, no 1, p. 42-61.

[571] SHARP, Gene. 198 methods of nonviolent action. The politics of nonviolent action, 1973, vol. 2.

[572] SHENON, Philip. Mobile Phones Primed, Affluent Thais Join Fray. New York Times, 1992, vol. 20. Available at: Last access: July 16th, 2015.

[573] SHI, Lei, et al. Predicting US primary elections with Twitter. Social Network and Social Media Analysis: Methods, Models and Applications Workshop, 2012.

[574] SHOSTAK, Arthur B. CyberUnion: Empowering labor through computer technology. ME Sharpe, 1999.

[575] SHOSTAK, Arthur B. The CyberUnion handbook: Transforming labor through computer technology. ME Sharpe, 2002.

[576] SHULMAN, Stuart W. The case against mass e-mails: Perverse incentives and low quality public participation in US federal rulemaking. Policy & Internet, 2009, vol. 1, no 1, p. 23-53.

[577] SKORIC, Marko, et al. Tweets and votes: A study of the 2011 singapore general election. In System Science (HICSS), 2012 45th Hawaii International Conference on. IEEE, 2012. p. 2583-2591.

[578] SKOVSGAARD, Morten; VAN DALEN, Arjen. Dodging the gatekeepers? Social media in the campaign mix during the 2011 Danish elections. Information, communication & society, 2013, vol. 16, no 5, p. 737-756.

[579] SLOAN, Luke, et al. Who tweets? Deriving the demographic characteristics of age, occupation and social class from twitter user meta-data. PloS one, 2015, vol. 10, no 3, p. e0115545.

[580] SMALL, Tamara. Canadian politics in 140 characters: Party politics in the Twitterverse. Canadian parliamentary review, 2010, vol. 33, no 3, p. 39-45.

[581] SMITH, Aaron. 22% of online Americans used social networking or Twitter for politics in 2010 campaign. Report of the Pew Internet Research Center, 2011.

[582] SMITH, Aaron. Cell phones, social media and campaign 2014. Pew Internet & American Life Project, 2014.

[583] SMITH, Aaron. The Internet’s role in campaign 2008. Pew Internet & American Life Project, 2009, vol. 15.

[584] SNEAD, John T. Social media use in the US Executive branch. Government Information Quarterly, 2013, vol. 30, no 1, p. 56-63.

[585] SOBACI, Mehmet Zahid; KARKIN, Naci. The use of twitter by mayors in Turkey: Tweets for better public services?. Government Information Quarterly, 2013, vol. 30, no 4, p. 417-425.

[586] SOLER, Juan M.; CUARTERO, Fernando; ROBLIZO, Manuel. Twitter as a tool for predicting elections results. In Proceedings of the 2012 International Conference on Advances in Social Networks Analysis and Mining (ASONAM 2012). IEEE Computer Society, 2012. p. 1194-1200.

[587] SPENCER, Tessa. The potential of the Internet for non-profit organizations. First Monday, 2002, vol. 7, no 8.

[588] SPIERINGS, Niels; JACOBS, Kristof. Getting personal? The impact of social media on preferential voting. Political Behavior, 2014, vol. 36, no 1, p. 215-234.

[589] STANLEY, J. Woody; WEARE, Christopher. The Effects of Internet Use on Political Participation Evidence From an Agency Online Discussion Forum. Administration & Society, 2004, vol. 36, no 5, p. 503-527.

[590] STANYER, James. Online campaign communication and the phenomenon of blogging: an analysis of web logs during the 2005 British general election campaign. In GUNTER, Barrie (ed.). Aslib Proceedings. Emerald Group Publishing Limited, 2006. p. 404-415.

[591] STANYER, James. The British public and political attitude expression: the emergence of a self-expressive political culture?. Contemporary Politics, 2005, vol. 11, no 1, p. 19-32.

[592] STRANDBERG, Kim. A social media revolution or just a case of history repeating itself? The use of social media in the 2011 Finnish parliamentary elections. New Media & Society, 2013 vol. 15, no. 8, pp. 1329-1347.

[593] STRANGELOVE, Michael. The Internet, electric gaia and the rise of the uncensored self. Computer-Mediated Communication Magazine, 1994, vol. 1, no 5, p. 11.

[594] STRECK, John. Pulling the plug on electronic town meetings: Participatory democracy and the reality of the Usenet. The Politics of Cyberspace (New Political Science Reader), 1998, Routledge, pp. 18-47.

[595] STREECK, Wolfgang; HASSEL, Anke. Trade unions as political actors. International handbook of trade unions, 2003, pp. 335-365.

[596] STROMER-GALLEY, Jennifer. Diversity of political conversation on the Internet: Users’ perspectives. Journal of Computer-Mediated Communication, 2003, vol. 8, no 3.

[597] SULLIVAN, Andrew. The revolution will be twittered. The Daily Dish, 2009. Available at: Last access: July 20th, 2015.

[598] SULLIVAN, Rachel. Live-tweeting terror: a rhetorical analysis of @HSMPress_ Twitter updates during the 2013 Nairobi hostage crisis. Critical Studies on Terrorism, 2014, vol. 7, no 3, p. 422-433.

[599] SUNSTEIN, Cass R. Republic. com 2.0. Princeton University Press, 2007.

[600] SWAN, Melanie. Blockchain: Blueprint for a New Economy. O’Reilly Media, Inc., 2015.

[601] SWEETSER, Kaye D.; GOLAN, Guy J.; WANTA, Wayne. Intermedia agenda setting in television, advertising, and blogs during the 2004 election. Mass Communication & Society, 2008, vol. 11, no 2, p. 197-216.

[602] SWEETSER, Kaye D.; LARISCY, Ruthann Weaver. Candidates make good friends: An analysis of candidates’ uses of Facebook. International Journal of Strategic Communication, 2008, vol. 2, no 3, p. 175-198.

[603] T HART, Marjolein. Humour and social protest: An introduction. International Review of Social History, 2007, vol. 52, no S15, p. 1-20.

[604] TABOADA, Maite, et al. Lexicon-based methods for sentiment analysis. Computational linguistics, 2011, vol. 37, no 2, p. 267-307.

[605] TALBOT, David. How Obama really did it. Technology Review, 2008, vol. 111, no 5, p. 78-83.

[606] TAPSCOTT, Don. Grown up digital. New York: McGraw-Hill, 2009.

[607] TARROW, Sidney. Contentious Politics. The Wiley-Blackwell Encyclopedia of Social and Political Movements, 2013.

[608] TARROW, Sidney. Power in Movement: Social Movements and Contentious Politics. Cambridge University Press. 2011.

[609] TASTAN, Coskun. The Gezi Park protests in Turkey: A qualitative field research. Insight Turkey, 2013, vol. 15, no 3, pp. 27-38.

[610] TAYLOR, Matthew; MUIR, Hugh. General election 2010: the defeat of the BNP. The Guardian, 2010. Available at: Last access: July 7th, 2015.

[611] TAYLOR, Maureen; KENT, Michael L. Congressional web sites and their potential for public dialogue. Atlantic Journal of Communication, 2004, vol. 12, no 2, p. 59-76.

[612] THAPEN, Nicholas A.; GHANEM, Moustafa M. Towards Passive Political Opinion Polling using Twitter. In SMA@ BCS-SGAI. 2013. p. 19-34.

[613] THE ECONOMIST INTELLIGENCE UNIT. Democracy Index 2010: Democracy in retreat. The Economist, 2010.

[614] THE ECONOMIST INTELLIGENCE UNIT. Democracy index 2011: Democracy under stress. The Economist, 2011.

[615] THE ECONOMIST INTELLIGENCE UNIT. Democracy Index 2014—Democracy and its discontents. The Economist, 2014.

[616] THE ECONOMIST INTELLIGENCE UNIT. The Economist Intelligence Unit’s index of democracy 2008. The Economist, 2008, vol. 29.

[617] THEOCHARIS, Yannis, et al. Using Twitter to mobilise protest action: transnational online mobilisation patterns and action repertoires in the occupy Wall Street, Indignados and Aganaktismenoi movements. 2013. Available at SSRN:

[618] THOMASSEN, Jacques. Chapter 5: Opinions about Political Issues. European Social Survey Core Questionnaire Development. European Social Survey, 2001.

[619] THORNTON, Alinta L. Does the internet create democracy?. Ecquid Novi: African Journalism Studies, 2001, vol. 22, no 2, p. 126-147.

[620] TILLY, Charles. Social Movements, 1768-2004. Paradigm Publishers, 2004a.

[621] TILLY, Charles. Terror, terrorism, terrorists. Sociological Theory, 2004b, vol. 22, no 1, p. 5-13.

[622] TILLY, Louise A. Paths of proletarianization: Organization of production, sexual division of labor, and women’s collective action. Signs, 1981, p. 400-417.

[623] TOLBERT, Caroline J.; MCNEAL, Ramona S. Unraveling the effects of the Internet on political participation?. Political research quarterly, 2003, vol. 56, no 2, p. 175-185.

[624] TOLER, Aric. Fake ‘Ukrainian’ News Websites Run by Russian ‘Troll Army’ Offshoots. Global voices online, 2015. Available at: Last access: June 24th, 2015.

[625] TRAVIS, Alan. What are Theresa May’s new ‘extremism disruption orders’? The Guardian, 2014. Available at: Last access: June 27th, 2015.

[626] TRERÉ, Emiliano. Reclaiming, proclaiming, and maintaining collective identity in the# YoSoy132 movement in Mexico: an examination of digital frontstage and backstage activism through social media and instant messaging platforms. Information, Communication & Society, 2015, pp. 1-15.

[627] TRILLING, Damian. Two Different Debates? Investigating the Relationship Between a Political Debate on TV and Simultaneous Comments on Twitter. Social Science Computer Review, 2015, vol. 33, no 3, p. 259-276.

[628] TRUMPER, Diego Saez; MEIRA, Wagner; ALMEIDA, Virgilio. From total hits to unique visitors model for election’s forecasting. 2011.

[629] TSAKALIDIS, Adam, et al. Predicting elections for multiple countries using Twitter and polls. Intelligent Systems, IEEE, 2015, vol. 30, no 2, p. 10-17.

[630] TSAKALIDIS, Adam, et al. Predicting elections for multiple countries using Twitter and polls. Intelligent Systems, IEEE, 2015a, vol. 30, no 2, p. 10-17.

[631] TSAKALIDIS, Adam, et al. UK Elections 2015 - Our predictions. 2015b. Available at: Last access: July 10th, 2015.

[632] TSELIKOV, Andrey. The Tightening Web of Russian Internet Regulation. Berkman Center Research Publication, 2014, no 2014-15.

[633] TSFATI, Yariv; WEIMANN, Gabriel. www. terrorism. com: Terror on the Internet. Studies in Conflict and Terrorism, 2002, vol. 25, no 5, p. 317-332.

[634] TUFEKCI, Zeynep. Big Questions for Social Media Big Data: Representativeness, Validity and Other Methodological Pitfalls. 2014. ICWSM.

[635] TUMASJAN, Andranik, et al. Predicting Elections with Twitter: What 140 Characters Reveal about Political Sentiment. ICWSM, 2010, vol. 10, p. 178-185.

[636] TURNEY, Peter D. Thumbs up or thumbs down?: semantic orientation applied to unsupervised classification of reviews. In Proceedings of the 40th annual meeting on association for computational linguistics. Association for Computational Linguistics, 2002. p. 417-424.

[637] TURNEY, Peter D.; LITTMAN, Michael L. Measuring praise and criticism: Inference of semantic orientation from association. ACM Transactions on Information Systems (TOIS), 2003, vol. 21, no 4, p. 315-346.

[638] TUROFF, Murray. The anatomy of a computer application innovation: Computer mediated communications (CMC). Technological Forecasting and Social Change, 1989, vol. 36, no 1, p. 107-122.

[639] TWITTER. The Twitter Government and Elections Handbook. 2014. Available at: Last access: July 8th, 2015.

[640] UNANKARD, Sayan, et al. Predicting Elections from Social Networks Based on Sub-event Detection and Sentiment Analysis. In Web Information Systems Engineering–WISE 2014. Springer International Publishing, 2014. p. 1-16.

[641] UTZ, Sonja. The (potential) benefits of campaigning via social network sites. Journal of Computer-Mediated Communication, 2009, vol. 14, no 2, p. 221-243.

[642] VACCARI, Cristian. “Technology is a commodity”: The Internet in the 2008 United States presidential election. Journal of Information Technology & Politics, 2010, vol. 7, no 4, p. 318-339.

[643] VACCARI, Cristian. From the air to the ground: The Internet in the 2004 US presidential campaign. New media & society, 2008, vol. 10, no 4, p. 647-665.

[644] VACCARI, Cristian. Surfing to the Elysee: The Internet in the 2007 French elections. French Politics, 2008, vol. 6, no 1, p. 1-22.

[645] VALENZUELA, Sebastián; KIM, Yonghwan; DE ZÚÑIGA, Homero Gil. Social networks that matter: Exploring the role of political discussion for online political participation. International journal of public opinion research, 2012, vol. 24, no 2, p. 163-184.

[646] VALENZUELA, Sebastián; PARK, Namsu; KEE, Kerk F. Is there social capital in a social network site?: Facebook use and college students’ life satisfaction, trust, and participation. Journal of Computer-Mediated Communication, 2009, vol. 14, no 4, p. 875-901.

[647] VALLONE, Giuliana. Protestos precisam mais do que multidões para ter sucesso, diz Occupy (Protests need more than crowds to succeed, says Occupy). Folha de S. Paulo, 2015. Available at: Last access: July 19th, 2015.

[648] VAN AELST, Peter; WALGRAVE, Stefaan. New media, new movements? The role of the internet in shaping the ‘anti-globalization’ movement. Information, Communication & Society, 2002, vol. 5, no 4, p. 465-493.

[649] VAN DETH, Jan W. A conceptual map of political participation. Acta Politica, 2014, vol. 49, no 3, p. 349-367.

[650] VAN LAER, Jeroen; VAN AELST, Peter. Internet and social movement action repertoires. Information, Communication & Society, 2010, vol. 13, p. 1147-1165.

[651] VAN OSCH, Wietske; COURSARIS, Constantinos K. Social media research: An assessment of the domain’s productivity and intellectual evolution. Communication books, 2014, vol. 81, no 3, p. 285-309.

[652] VARLEY, Pamela. Electronic democracy. Technology Review, 1991, vol. 94, no 8, p. 42-51.

[653] VENKATARAMAN, Mukundan; SUBBALAKSHMI, K. P.; CHANDRAMOULI, R. Measuring and quantifying the silent majority on the internet. In Sarnoff Symposium (SARNOFF), 2012 35th IEEE. IEEE, 2012. p. 1-5.

[654] VERBA, Sidney, et al. Voice and equality: Civic voluntarism in American politics. Cambridge, MA: Harvard University Press, 1995.

[655] VERBA, Sidney; NIE, Norman. Participation in America: Political Democracy and Social Equality. 1972.

[656] VICENTE-MERINO, Ma Rosa. Websites of parliamentarians across Europe. The Journal of Legislative Studies, 2007, vol. 13, no 3, p. 441-457.

[657] VITAK, Jessica, et al. It’s complicated: Facebook users’ political participation in the 2008 election. CyberPsychology, behavior, and social networking, 2011, vol. 14, no 3, p. 107-114.

[658] VOLKOVA, Svitlana; COPPERSMITH, Glen; VAN DURME, Benjamin. Inferring user political preferences from streaming communications. In Proceedings of ACL. 2014. p. 186-196.

[659] VOSEN, Simeon; SCHMIDT, Torsten. A monthly consumption indicator for Germany based on Internet search query data. Applied Economics Letters, 2012, vol. 19, no 7, p. 683-687.

[660] WALDMAN, Steve Randy. There is a name for this. Interfluidity, 2015. Available at: Last access: July 18th, 2015.

[661] WALLSTEN, Kevin. Agenda setting and the blogosphere: An analysis of the relationship between mainstream media and political blogs. Review of Policy Research, 2007a, vol. 24, no 6, p. 567-587.

[662] WALLSTEN, Kevin. The Blogosphere’s Influence on Political Discourse: Is Anyone Listening?. In annual meeting of the Midwest Political Science Association, Illinois. 2007b.

[663] WANG, Gang, et al. Serf and turf: crowdturfing for fun and profit. In Proceedings of the 21st international conference on World Wide Web. ACM, 2012. p. 679-688.

[664] WANG, Yazhe; CALLAN, Jamie; ZHENG, Baihua. Should We Use the Sample? Analyzing Datasets Sampled from Twitter’s Stream API. ACM Transactions on the Web (TWEB), 2015, vol. 9, no 3, p. 13.

[665] WARD, Stephen; GIBSON, Rachel; LUSOLI, Wainer. Online participation and mobilisation in Britain: Hype, hope and reality. Parliamentary affairs, 2003, vol. 56, no 4, p. 652-668.

[666] WARD, Stephen; LUSOLI, Wainer. ‘From weird to wired’: MPs, the Internet and representative politics in the UK. Journal of Legislative Studies, 2005, vol. 11, no 1, p. 57-81.

[667] WARD, Stephen; LUSOLI, Wainer. Dinosaurs in cyberspace? British trade unions and the Internet. European Journal of Communication, 2003, vol. 18, no 2, p. 147-179.

[668] WARD, Stephen; LUSOLI, Wainer. Logging on or switching off? The public and the internet at the 2005 general election. Spinning the Web, London: Hansard Society, 2005, p. 13-21.

[669] WASHINGTON TIMES, The. EDITORIAL: Iran’s Twitter revolution. The Washington Times, 2009. Available at: Last access: July 16th, 2009.

[670] WATERS, Richard D., et al. Engaging stakeholders through social networking: How nonprofit organizations are using Facebook. Public Relations Review, 2009, vol. 35, no 2, p. 102-106.

[671] WATERS, Richard D.; JAMAL, Jia Y. Tweet, tweet, tweet: A content analysis of nonprofit organizations’ Twitter updates. Public Relations Review, 2011, vol. 37, no 3, p. 321-324.

[672] WATTS, Duncan J. Computational social science: Exciting progress and future directions. The Bridge on Frontiers of Engineering, 2013, vol. 43, no 4, p. 5-10.

[673] WEBER, Ingmar; POPESCU, Ana-Maria; PENNACCHIOTTI, Marco (Eds.) Proceedings of the 2nd Workshop on Politics, Elections and Data. ACM, 2013.

[674] WEBER, Lori M.; LOUMAKIS, Alysha; BERGMAN, James. Who participates and why? An analysis of citizens on the Internet and the mass public. Social Science Computer Review, 2003, vol. 21, no 1, p. 26-42.

[675] WEIMANN, Gabriel. Lone wolves in cyberspace. Journal of Terrorism Research, 2012, vol. 3, no 2.

[676] WEIMANN, Gabriel. New Terrorism and New Media. Wilson Center Common Labs, 2014.

[677] WEIMANN, Gabriel. Terror on Facebook, Twitter, and YouTube. Brown J. World Aff., 2009, vol. 16, p. 45.

[678] WEISS, J. Blogs Shake the Political Discourse: Website Bloggers Changing the Face of Political Campaigns. Boston Globe, 2003. Available at: Last access: July 8th, 2015.

[679] WELLMAN, Barry, et al. Does the Internet increase, decrease, or supplement social capital? Social networks, participation, and community commitment. American behavioral scientist, 2001, vol. 45, no 3, p. 436-455.

[680] WILHELM, Anthony G. Virtual sounding boards: How deliberative is on-line political discussion?. Information Communication & Society, 1998, vol. 1, no 3, p. 313-338.

[681] WILLIAMS, Christine; GULATI, Girish. What is a social network worth? Facebook and vote share in the 2008 presidential primaries. American Political Science Association, 2008.

[682] WILLIAMS, Rachel. Tories remove web forum on MPs’ looks. The Guardian, 2007. Available at: Last access: July 7th, 2015.

[683] WILLIAMSON, Andy. The effect of digital media on MPs’ communication with constituents. Parliamentary Affairs, 2009, vol. 62, no 3, p. 514-527.

[684] WILSON, Theresa; WIEBE, Janyce; HOFFMANN, Paul. Recognizing contextual polarity in phrase-level sentiment analysis. In Proceedings of the conference on human language technology and empirical methods in natural language processing. Association for Computational Linguistics, 2005. p. 347-354.

[685] WINNER, Langdon. Mythinformation. The whale and the reactor, 1986, The University of Chicago Press.

[686] WINTER, Philipp; CRANDALL, Jedidiah R. The Great Firewall of China: How It Blocks Tor and why It Is Hard to Pinpoint. ;login: 2012, vol. 37, no. 6, pp. 42-50.

[687] WINTERS, Jeffrey A.; PAGE, Benjamin I. Oligarchy in the United States?. Perspectives on Politics, 2009, vol. 7, no 04, p. 731-751.

[688] WINTOUR, Patrick. David Cameron to unveil new limits on extremists’ activities in Queen’s speech. The Guardian, 2015. Available at: Last access: June 27th, 2015.

[689] WOJCIESZAK, Magdalena E.; MUTZ, Diana C. Online groups and political discourse: Do online discussion spaces facilitate exposure to political disagreement?. Journal of Communication, 2009, vol. 59, no 1, p. 40-56.

[690] WOLFSFELD, Gadi; SEGEV, Elad; SHEAFER, Tamir. Social media and the Arab spring politics comes first. The International Journal of Press/Politics, 2013, vol. 18, no 2, p. 115-137.

[691] WOODWARD, Will. Tories unveil their secret weapon: ’webcameron’. The Guardian, 2006. Available at: Last access: July 7th, 2015.

[692] WRIGHT, Scott. Government-run Online Discussion Fora: Moderation, Censorship and the Shadow of Control. The british journal of politics & international relations, 2006, vol. 8, no 4, p. 550-568.

[693] WRIGHT, Scott. Political blogs, representation and the public sphere. In GUNER, Barrie (ed.). Aslib Proceedings. Emerald Group Publishing Limited, 2009. p. 155-169.

[694] WRIGHT, Scott; STREET, John. Democracy, deliberation and design: the case of online discussion forums. New Media & Society, 2007, vol. 9, no 5, p. 849-869.

[695] XENOS, Michael; FOOT, Kirsten. Not your father’s Internet: The generation gap in online politics. Civic life online: Learning how digital media can engage youth, 2008, p. 51-70.

[696] XENOS, Michael; VROMEN, Ariadne; LOADER, Brian D. The great equalizer? Patterns of social media use and youth political engagement in three advanced democracies. Information, Communication & Society, 2014, vol. 17, no 2, p. 151-167.

[697] YASSERI, Taha; BRIGHT, Jonathan. Can electoral popularity be predicted using socially generated big data?. it-Information Technology, 2014, vol. 56, no 5, p. 246-253.

[698] YU, Bei; KAUFMANN, Stefan; DIERMEIER, Daniel. Exploring the characteristics of opinion expressions for political opinion classification. In Proceedings of the 2008 international conference on Digital government research. Digital Government Society of North America, 2008. p. 82-91.

[699] ZAGHENI, Emilio; WEBER, Ingmar. Demographic research with non-representative internet data. International Journal of Manpower, 2015, vol. 36, no 1, p. 13-25.

[700] ZARWAN, Elijah, et al. False Freedom—Online Censorship in the Middle East and North Africa. Human Rights Watch, vol. 17, no. 10 (E), 2005. Available at: Last access: June 23rd, 2015.

[701] ZHUO, Xiaolin; WELLMAN, Barry; YU, Justine. Egypt: the first internet revolt?. Peace Magazine, 2011, vol. 27, no 3, p. 6-10.

[702] ZICKUHR, Kathryn. Generations online in 2010. Pew Internet & American Life Project, 2010, vol. 16.

[703] ZITTEL, Thomas. Political representation in the networked society: the Americanisation of European systems of responsible party government?. The Journal of Legislative Studies, 2003, vol. 9, no 3, p. 32-53.

[704] ZITTRAIN, Jonathan. The future of the internet--and how to stop it. Yale University Press, 2008.

[705] ZITTRAIN, Jonathan; EDELMAN, Benjamin G. Internet filtering in China. IEEE Internet Computing, March/April, 2003.

[706] ŽIŽEK, Slavoj. Violence: Six sideways reflections. Profile books, 2009.

[707] ZUCKERMAN, Ethan. New media, new civics?. Policy & Internet, 2014, vol. 6, no 2, p. 151-168.