ONLINE STORYTELLING Studying Homo narrans in several online habitats by
analyzing the framing of the Wikileaks Iraq video in
newspapers, blogs and tweets
Paper for the research seminar: Citizen/Journalism: User Generated Content and the Consequences for Journalism Module: LJX031M10 Lecturer: Dr. Todd S. Graham Paper by Jelmer Mommers (s1624180) [email protected] June 6, 2010 Master of Journalism, University of Groningen
The most striking phenomenon in Western journalism, in both praxis and theory, is
the obstinate, the diehard metaphysical faith that language is transparent. Or, put
somewhat differently: The fault lies in the refusal of journalists, but also of students
of journalism, to put the profession where it belongs, that is, within the context of
human expression, of expressive activity. It is the refusal to deal with and judge
newswriting for what it is in essence – storytelling.
Roeh, 1989: 162
Choices of words and their organization into news stories are not trivial matters.
They hold great power in setting the context for debate, defining issues under
consideration, summoning a variety of mental representations, and providing the
basic tools to discuss the issues at hand.
Pan & Kosicki, 1993: 70
Regardless of how many pundits attempt to frame this tragedy within the vagaries
of a "war is hell" narrative, Collateral Murder will prove to be a landmark event in
the reportage of the Iraq war, as it forces the viewer, in the most visceral way
possible, to simultaneously confront both the deplorable unreality of American
aggression and the grim fate of those caught within its scope.
Douglas Haddow, 2010 (np)
1
Contents page
Contents 1 1. Introduction 2 2. Literature review 2 3. This research 6 On Wikileaks 7 The case 7 4. Methodology 9 Framing analysis 9 Units of analysis 11 Categorization of data 11 Validity 12
Stability 12 5. Limitations 12
Researcher bias 13 Shortcomings in sampling 13 Shortcomings of the analysis 14
6. Results 14 Syntactical structure 15 Initial reception 15 Follow-up 16 Rhetorical/Lexical proporties 17 Description of the video 17 Description of incident 18 Denomination of victims 18 Culprit 19 Source 19 7. Conclusions 19 References 22 Appendices Please note that appendices 5, 6, 7 and 8 do not appear in the print version of this paper. To consult them, please surf to jelmermommers.nl/online-storytelling and dowload them there
1. Summary: selection of data, units of analysis 25
2. Selection and sampling processes 26 3. Exploration of the Wikileaks Collateral Murder (WCM) frame 29 4. Coding schemes 31
a. For newspaper articles, blog postings 31 b. For tweets 34
5. Frame analysis examples (separate) separate
6. Data (separate) separate
a. Copies of selected newspaper articles separate
b. Copies of selected blog postings separate
c. Tweets April 5 separate
d. Tweets April 7 separate
7. Analysis (separate) separate
a. Codings of newspaper articles and blog postings separate
b. Codings of tweets separate
8. Retest data (separate) separate
a. Newspapers and blogs separate
b. Tweets separate
9. Figures showing some results of the analysis 37
2
1. INTRODUCTION Humans are essentially storytellers. That was the basic proposition of a 1985 article by Walter Fisher,
in which he argued that all human communication should be understood within the narrative
paradigm, a paradigm that nullifies the distinction between mythos and logos and returns to the
concept of logos as it was proposed by the (pre-Socratic) ancients, i.e. one concept that incorporates
all aspects of human expression and communication, namely: story, reason, rationale, conception,
discourse and/or thought. The paradigm holds that all meaningful communication is a form of
storytelling.
The following paper is the result of taking this paradigm online and studying homo narrans
(man the storyteller) as he used different platforms to make sense of one concrete news event, in this
case the leaking of the Wikileaks Iraq shooting video (discussed later).
2. LITERATURE REVIEW “Journalists are professional storytellers. They do not write articles, they write stories” (Bell, 236).
Working from this perspective, journalism becomes more like other kinds of public expression
(Zelizer, 2004:128). It can be regarded as a form of cultural production, in the same league as film
making and fine arts. And like all acts of cultural production, journalism relies “on conventions to
accomplish work while approximating a shared understanding with audiences and while working
within social, cultural and economic constraints” (Lowrey & Latta, 2008: 185).
The news accounts journalists construct “shape in decisive ways our perceptions of the „world out
there‟ beyond our immediate experience. For many of us, our sense of what is happening in the
society around us, what we should know and care about from one day to the next, is largely derived
from the news stories they tell” (Allan, 1999: 77).
“[This] role of storyteller is a significant one”, notes Bell (2001: 236), “both in language behavior
and in society at large. Much of humanity‟s most important experience has been embodied in stories.”
Those readers still doubting the omnipresence of story in journalism should study its lingo. As
Bell notes (idem): “A good journalists „gets good stories‟ or „knows a good story‟. A critical news editor
asks: „Is this really a story?‟ Where‟s the story in this‟?”
The storytelling perspective proposed here argues against the idea of „journalism as information‟
(cf. Zelizer, 2004:129). It also argues strongly against any notion of a reality „out there‟ independent
of human observation, one which can be represented objectively in text, working by normative
standards and professional conventions. From the storyteller perspective of journalism, there is no
one „true story‟ that can be told „correctly‟ (cf. Roeh, 1989: 163). Codifying an event to make it fit into
standardized news discourse (cf. Allan, 1999: 81) does not make that discourse the truest mirror of
reality. There are only interpretations.
However, it would be a misunderstanding to deny that some stories make more sense or better
reflect reality than others. Incidentally, this is not what (this interpretation of) Fisher holds. What
Fisher does argue, however, is that no one form of discourse has “final jurisdiction” over others
(1985:87). While sidelining the philosophical debate over fact and fiction, this paper confirms that
there is a difference between true and fictional stories. The point is that whether or not a story is
believed to be true depends on the producer, the medium and the audience. As Roeh explains, within
the storytelling perspective, (research) questions are “located in the world of culture and meaning-
production, rather than in the world of true facts and distorted facts” (Roeh, 1989: 168).
As Schudson notes, “all news stories are stories, but some are more storylike than others” (2003:
186). Hard news, for instance, is not narrative driven. “The classic hard news story operates more to
convey useful information efficiently than to build a shared world with readers emotionally. At this
end of journalistic writing, the reporter mimics a piece of machinery that conveys relevant
information with accuracy” (idem). However, this hard news story should still be regarded as story,
because different from a „simple chronological account‟, it seeks coherence and meaning: “a story has
3
a point, and it exists within a cultural lexicon of understandable themes” (Bird & Dardenne, 2009:
207). I would challenge anybody seeking to counter the idea of journalism as story, to produce one
article or television news item that has no point, that seeks no coherence, that creates no meaning.
What then, is it that journalists do? According to Schudson, the reporter‟s job is to “make meaning
(…). The writer must construct a tale, one whose understanding requires a reader or viewer to
recognize not the sum of facts but the relationships among them” (Schudson, 2003: 177). As meaning
makers, journalists have authority (Barnhurst, 2003: 2). They “„make events mean‟ by providing
frameworks for understanding unexpected or unusual events, thus bringing them within the bounds
of the normal and comprehensible” (Hall et.al. (1978) as paraphrased in Lawrence, 1996: 141).
Importantly, these frameworks for understanding should usually already be familiar with audiences
to be effective. In other words: journalists have to retell familiar stories to be popular storytellers. Or
as Bird & Dardenne put it, “the impulse to tell stories may lead journalists to frame the world in
conventional ways that often reinforce existing ideologies” (2009: 208).
With all this storytelling by journalists and others in the business of cultural production, one
would almost forget the first proposition of this paper. It does not say: “Journalists are essentially
storytellers.” It said: “Humans are essentially storytellers.” Although in their professional role,
journalists may have a privilege over others to tell what convention deems „true stories‟, they are –
like everybody else – in the business of making meaning, producing narratives, and thereby
constructing the world (cf. Bird & Dardenne, 2009: 209).
The narrative paradigm was posited by Fisher as a proposal against the privileging of “experts”
and their discourse. Fisher hoped that our (re)new(ed) understanding of man as storyteller would
result in a concept of logos that approximates that of the ancients (i.e. one that incorporates story,
reason, rationale, conception, discourse and thought in one concept), one that “regards all humans
and their communication as not irrational and as deserving of respect” (1985:79 (italics added)). It
is with this notion in mind that this story begins.
As always, the media landscape is changing. With the advent of the internet, it has been changing
rapidly. Print media are in decline, online media are on the rise (cf. e.g. Pew, 2009). One pressing
issue for online journalism is its business model: media companies have yet to find a way to pay for
quality (investigative) journalism with online revenues, not to mention making a profit.
The internet has given non-journalists the tools to produce their own content (user-generated
content (UGC)) on platforms like Blogger, Wordpress and Twitter (basically a platform for micro-
blogging). Some of these „amateurs‟, as they have been positively (Lessig, 2007) and negatively
described (Keen, 2007), are taking on an active role in (online) public debate. Some are stepping on
to turf that used to be controlled by professional journalists only. These non-professional storytellers
are getting actively involved in all stages of journalism. In the old days, readers had the option to
react to an article or news item once it had been produced and printed or broadcasted (by mail,
usually). Now, online users can get involved in all stages of journalistic production, and make their
voice heard before professional gatekeepers do their work. Journalists are no longer the filters they
used to be (Goode, 2009: 1291). In this sense, all this online activity is a “challenge of top down
storytelling” (Deuze, 2006: 72). Online news journalism by non-professional journalists is
dismantling a carefully cultivated hierarchical relationship between (mass) media consumers and
producers (ibidem: 65). Even without going into the idea of Bourdieu‟s field theory (1992), it is
perfectly understandable that traditional media outlets have reacted negatively to these changes. If
online amateurs can do (part of) a journalist‟s job, then what is the point?
One prominent answer to the changes in the media landscape over the past 20 years has been the
idea of “news as conversation” (cf. Gillmor, 2004; Goode, 2009: 1295). “The lines will blur between
producers and consumers, changing the role of both in ways we‟re only beginning to grasp now.”
(Gillmor, 2004: xiii). “News reception is about process, not text, as “the story” emerges in
4
conversation with the news narrative as framing structure. In making sense of the news, we involve
others in the negotiation of meaning and its cultural significance emerges through everyday
interaction” (Bird (2003) as cited in Bird & Dardenne, 2009: 212).
In this conversation, power matters. The texts that meaning makers manufacture, be they
journalists or moviemakers, are shaped by the power relations that surround them (cf. Bourdieu,
2005). The creation of meaning is central to the idea of communication. But besides the socio-
cultural or organizational aspect of creating meaning, there is always the question of raw power. Who
has the power to create meaning? As stated before, journalists have authority to tell true stories, to
manufacture true interpretation of events. But who is able to contest when the „right‟ meaning is not
being constructed? These are questions of power and when studying different (new) media and
content producers, we should keep a sharp eye on the power relations that shape their stories. Power
is imprinted in every act of cultural production.
There are several ways to interpret the influence of context on the production of story. One is a
sociocognitive understanding of news discourse, as posited by Van Dijk. This perspective “explains
how underlying ideologies control more specific group attitudes and how personal mental models of
journalists about news events control activities of news making (…)” (2009: 195).
Schudson distinguishes two views to the manufacture of story: the cultural view or the social-
organizational view (2003: 182). “Organizational views take news to be a manufactured product
created anew each day in interactions among firms, markets, and resources. A cultural view is more
impressed with the things news workers inherit than with what they create; it emphasizes not social
production so much as the symbolic determinants of news in the relation between facts and symbols.”
(ibidem, 183)
In my view, these perspectives on the influence of context on story overlap. For the purpose of
this paper, the main point to be made is that the context of cultural production matters. As Roeh
notes, “stories entail given points of view” and “storytelling assumes given contexts of communication
and of meaning-production” (1989: 168). Later on in this paper, frames will be analyzed as the key
manifestation of both story and context of the cultural production. In this view, the cultural producer
is primarily a product of his surroundings.
Form is by no means neutral. The meaning of any text is both a result of its content (words,
sentences, propositions, ideas) and its form (medium, (story-)structure, lay-out). As Goode notes:
“[the] formal properties of any medium embody „codes‟ which carry political consequences” (2009:
1303). Schudson gives a clear example of this phenomenon. He describes a changing role of
journalists in the U.S. and concludes that “with the establishment of the summary lead as newspaper
convention, journalists began to move from being stenographers, or recorders, to interpreters”
(Schudson, 1995: 59). In other words: a change in function of journalists brought about a change in
form of the journalistic product (i.e. story). Concludes Schudson: “It is a very different matter to say
that the news reflects the social world by describing it, and to say that it reflects the social world by
incorporating it into unquestioned and unnoticed conventions of narration. When a changed political
reality becomes part of the very structure of news writing, then the story does not “reflect” the new
politics but becomes part of the new politics itself. There is not only a narration of politics in the
news; the news is part of the politics of narrative form” (ibidem, 66).
Form influences our perception of story. News papers, for instance, are masters of creating the
form in which „true stories‟ can be told. In this sense, the form of the news paper, the lay-out (e.g.
using big or small pictures), the structure of particular stories, are all ideological. They reflect ways of
looking at the world, they reflect ideas about truth and objectivity. A good example is the separation
of op-eds and news in a newspaper. This form speaks the ideology of objectivity. Thus, news media
provide the forms in which “true declarations” appear. (Schudson, 1995: 54) This is what Broersma
calls “performative discourse” (in press).
5
I have relaid here this connection between form, content and ideology, because it follows logically
that a new medium, a new form of communication, always challenges ideological givens, i.e. that any
new medium has political consequences that may go unnoticed for some time, but are very real in the
everyday lives of people as more of them start to use the medium. “The manifold scrambled,
manipulated, and converged ways in which we produce and consume information worldwide are
gradually changing the way people interact and give meaning to their lives” (Deuze, 2006: 66).
Obviously, the internet is playing a major role in this process. It is having profound influence on
communication, journalism and politics. Roles are changing. Borders are being redrawn. This is
fundamental of “online culture”, it involves a changing relationship between the consumers and
producers of news. It facilitates open publishing initiatives and individualized storytelling (Deuze,
2006: 64).
It should be noted that like any other medium, the internet itself is not neutral:
[All] technology is routed in ideology. (…) The Internet has (...) a very concrete cultural ideological
context. It represents the fusion of the old military industrial complex of southern California, and the
hippy ethic of northern California. (…) [Both] the military industrial complex, which invented the internet
or at least financed the internet in the late 50's, and the counter culture of the late 60's, in san Francisco
were libertarian or hostile to authority, hostile to the state, hostile to traditions. These people were tied
together by the libertarianism, by their hostility to traditional forms of authority.
(Keen, 2009, np)
And because “technology is simply a reflection of human will [and technology] isn't accidental,
(…) doesn't come about in a vacuum”, Keen concludes that “the Internet itself reflects a hostility
towards authority.” The notion that the internet has “no centre”, that it cannot be controlled, is a
reflection of its anti-hierarchical nature. On the other hand, Keen recognizes the paradoxical
relationship between this ideal and the immense power of a few internet companies like Google and
Facebook.
The point to be made here is that like other media, the internet is not a neutral medium but an
ideological one, and so the producers that function within its context are influenced by its nature.
This helps explain, as we will see later, why Wikileaks, a whistle-blower Web site with the mission to
“break the world open” is really such a natural phenomenon of the internet. It is very hostile to any
authority whatsoever. Whether this is good or bad is another question.
This also explains, for instance, the perceived “outsider status” of bloggers. They are seen by their
users “as conduits to raw information, somehow less corrupted by power than their predecessors
[from traditional journalistic outlets, JM]” (Tremayne, 2007: xvi). One important characteristic of
blogs is their opposition to mainstream media, notes Deuze (2006: 65). This can be understood
within Bourdieu‟s field theory, i.e. that new competitors in a „field‟ (in this case the journalistic field,
or that of storytellers) naturally oppose the status quo (2005). It can also be understood with Keens
proposition of the anti-authoritative nature of the internet (2009). Both perspectives lend a greater
understanding of what is actually happening when “the people formerly known as the audience” (a
term by Rosen, 2006) go online and start producing their own stories, their own narratives. “Instead
of relying on journalists, public relations officers, marketing communications professionals, and
other professional storytellers to make sense of our world, we seem to become quite comfortable in
telling and distributing our own versions of those stories (…)” (Deuze, 2006: 66). Like the
professional journalist‟s transformation from a stenographer to an interpreter (see Schudson above),
this is in itself an ideological shift; it is political.
Accordingly, professional journalists have not stood idly by as meaning makers on the internet
challenged their authority as storytellers. The new content producers have been criticized and seen as
a danger by professional journalists. Many journalists are of the opinion that the internet threatens
quality of journalism (O‟Sullivan & Heinonen, 2008: 366). As Hermida (2010: 4) relays, some
6
journalists‟ opposition to a service like Twitter “reflects the unease in adopting a platform which
appears to be at odds with journalism as a „„professional discipline for verifying information‟‟ (Project
for Excellence in Journalism, nd). “Services like Twitter are a challenge to a news culture based on
individual expert systems and group think over team work and knowledge-sharing” (Singer,2004, as
cited in Hermida, 2010: 4).
Others have suggested a complementary relationship between weblogs and traditional journalism
(Reese et.al, 2007), between old and new media. This complementary relationship is the one this
paper endorses. Different media producers serve different purposes, and all should be free from
elitist, normative judgments. Different media can and do co-exist, they influence one another but they
do not necessarily compete for the same functions (Tremayne, 2007). No blogger has ever set out to
be The New York Times.
Meaning makers do not function in a vacuum. On all sorts of online platforms, they influence and
react to one another. Their normative ideas about story may be different, leading some to believe that
they are different. But the borders between different forms of cultural production – take the genre-
lines within journalism – are purely conventional and constantly changing. Historically, the idea of
„objective journalism‟ as it is favored by institutional media was normalized in reaction to other forms
of storytelling, like the movement of „New Journalism‟ (Roggenkamp, 2005). It is important to
recognize the nature of these conventions and the corresponding normative judgments, because it
opens up the possibility of ignoring them in favor of openly comparing the stories that different
media producers tell. When we consider all people as storytellers, we consider all media as platforms
for this activity. When we comparatively study the texts in these media, we learn about their politics
of form, about ourselves and by extension we learn about the (social, organizational, cultural, and
political) context that surrounds media producers.
Without normatively judging the „journalistic quality‟ of one medium over another (a normative
choice in its own right), this paper simply examines how people use different online platforms to
make sense of the world.
3. THIS RESEARCH
In this research the meaning making function of different online storytellers will be analyzed by
taking a one case approach and studying its reception on different online platforms. More specifically,
the research takes the case of the Wikileaks Iraq video and analyzes how storytellers made sense of it.
This will tell us how – in relation to this event – different media platforms facilitate meaning making.
The study is not interested in determining whether in a normative sense, one form of content
is more journalistic than another. Instead, the study identifies the functions of these different media
platforms by examining the reception of the Wikileaks Iraq video. The study provides, if you will, a
snapshot of online news breaking and storytelling. Before discussing the case studied here, the source
should be introduced.
On Wikileaks
The video that prompted this research was obtained, decrypted and leaked by Wikileaks, a
whistleblower Web site “run by an international collective dedicated to untraceable document-
leaking” (Lynch, 2010:1).
Wikileaks was launched in early 2007. The Web site “publishes anonymous submissions and
leaks of sensitive documents from governments and other organizations, while preserving the
anonymity of its sources.” (Wikipedia, 2010). For its material, Wikileaks is completely dependent on
what it calls its „courageous sources‟.
Lynch notes: “As a site whose mandate is to publish any leak they deem significant, without
regard for political impact, violation of privacy or breach of copyright law, Wikileaks has been a
controversial project since its inception” (2010: 1). It should be noted that within the libertarian, anti-
7
authoritative nature of many new online media platforms, Wikileaks makes perfect sense (see last
chapter).
Wikileaks cannot be considered a regular journalistic enterprise from a normative point of
view, because, as co-founder Julian Asssange told the New York Times: if he were forced to choose
between being a journalist or an advocate, he would choose advocate (2010, April 6). The
organization, therefore, cannot be said to share the ethic of institutional media working within the
Anglo-Saxon model of „objective journalism‟. “Wikileaks is aggressively proactive in its attempts to
bring their leaks to the attention of journalists.” (Lynch, 2010:3)
This has not been the only reason for a tense relationship with mainstream media. Also,
“Wikileaks is unsettling to journalists because it represents a radical shift in the way information is
collected and distributed in the media landscape” (Lynch, 2010:2).
“Wikileaks poses a challenge to traditional journalistic practice, in particular to the ways in
which investigative journalists have cultivated source relationships and the ways in which media
outlets have established themselves as the arbiters of fact. And the friction rubs both ways: the
collective that runs Wikileaks has expressed frustration in their dealings with traditional media, in
particular with the press‟s seeming indifference to some of their leaked material” (Lynch, 2010: 3).
According to a survey performed by Lynch (2010), journalists have mixed feelings about
Wikileaks but most show at least some approval of the project and acknowledge that is was useful to
them at some point. “Other journalists reported being frustrated with how the unconventional nature
of the site thwarted their usual reporting practices” (ibidem: 9). In terms of gathering information,
reporters are Wikileaks‟ competitors and were described as such by the Web site in 2007 (ibidem:4).
It is important to note the essential differences and tensions between Wikileaks and
traditional, institutional media because they may explain some of the findings of this research.
The case
The case studied here is the leaking of a video by Wikileaks. The video, shot from a helicopter gun-
camera, shows the shooting and killing of alleged Iraqi insurgents and a Reuters reporter and his
driver by American forces in Baghdad in 2007. The men on the ground were believed to be armed
insurgents. Only after the incident did it show that the RPG (rocket launcher) soldiers had spotted on
one of the men, was in fact the zoom lens of the Reuters photographer‟s camera. Some of the other
victims were armed. When a minivan arrived on the scene to carry away the wounded, the U.S.
soldiers targeted the van too, injuring two children inside and killing others.
This incident had been reported by several media (including the Washington Post and the
New York Times (both 2007, July 13)) and by David Finkel in his book The Good Soldiers. The scene
in the video was thus not news in itself, although to many who were unaware of previous coverage, it
may have been.
The footage of the incident is shocking in several respects. First, the images are graphic and
gripping because they show not only the shooting and killing of armed men, but also of wounded
people and of people coming to their rescue who seem to pose no imminent threat to the security of
American forces.
Another shocking dimension of the video was the audio of military radio transmissions
between soldiers in the helicopters and controllers on the ground. Before opening fire, for instance,
one of the soldiers commands to “light „em all up”. After the shooting, a soldier calls on his buddy to
“look at those dead bastards”. When friendly ground troops arrive, one of the soldiers in the
helicopter chuckles after seeing a truck drive over a corpse. “I think they just drove over a body”.
Wikileaks published two versions of the video: one edited to nearly 18 minutes, and one full 39
minutes version showing the before and after of the incident described above. This has important
implications, because Wikileaks would later be criticized for editing the images in such a way that
they lack the context of ongoing fighting in the Baghdad neighborhood.
8
The question of what the actual news was, remains open and cannot be answered definitively,
as the different interpretation by media will show. What was noteworthy, though, was that Reuters
had tried in vain to obtain the footage ever since the incident. The news organization had even filed a
freedom of information act (FOIA) request, but without success (Reuters, 2008). This led many,
including Wikileaks, to believe the army was covering up a (series of) war crime(s).
Consequently, this was the way Wikileaks framed the video, giving it the title “Collateral
Murder” and claiming that not only did the video show murder of at least some unarmed civilians, the
two Reuters reporters and those coming to their rescue, but also that the US military had tried to
cover this up. It is important to note that in this sense, the way Wikileaks published the news was far
from „neutral‟. The video was also given a clear political slant. Take for instance the opening of the
video, with a George Orwell quote: “Political language is designed to make lies sound truthful and
murder respectable, and to give the appearance of solidity to pure wind.”
Wikileaks held a press conference at the National Press Club in Washington D.C. on the
morning of April 5, 2010, announcing the publication of the video. At 4:57 pm, Wikileaks tweeted
that it had published the video: “LEAK: Cover-up of murder of civilians and journalists through US
forces in Iraq. Video online at: http://collateralmurder.com”. This is where this research takes off.
Readers are advised to watch the video to gain a full understanding of the following analysis. The
research questions are:
R1: How was the leaking of the Wikileaks Iraq video framed on American newspaper Web
sites, professional news blogs, and on Twitter?
R2: If the framing on the different online platforms differs, what does this tell us about these
platforms and about the people using them?
4. METHODOLOGY The reception of this leaked video („the news‟ in the following) was analyzed on different meaning
making platforms, in an effort to sample both old and new organizations and applications on the
internet. Newspapers were chosen as a representative of old, institutional media that have gone
online. Professional news blogs were chosen as a representative of new media online, which share
many characteristics with newspapers, in that they too have become professional organizations that
employ professional journalists. As Lowrey & Latta note: “[The] more relevant bloggers become in
terms of audience and influence, the more their production routines resemble those of professional
journalists” (2008: 185).
Twitter was chosen as a place to look for spontaneous meaning making by ordinary people,
who without having taken a degree in journalism know to tell stories.
The focus is on American storytellers. The current research design can be replicated for other
countries (e.g. Iraq) later.
Concretely, the top 10 most trafficked newspaper Web sites and the 10 top most popular
political/news blogs in the U.S. were selected, along with 200 Twitter updates („tweets‟) about the
news. These numbers are not representative of all American online newspapers, blogs and tweeters,
but they do give a general impression. (It is unclear whether or not the tweets were – like the other
productions – written by Americans. More on this later.)
Especially with tweets, it was important to look at a relatively big number because, as
Hermida (2010: 5) has argued, “the value [of Twitter] does not lie in each individual fragment of news
and information, but rather in the mental portrait created by a number of messages over a period of
time.”
Both (1) the initial reception of the news and (2) the first follow-up after the initial reception
were analyzed. This would have meant looking at 20 newspaper articles, 20 blog postings, and 200
tweets – 100 forming the initial reception, another 100 forming the follow-up. However, some
9
newspapers and blogs did not produce either an initial reception or a follow-up (or both), so the total
number of selected productions was slightly smaller. Also, two blog postings that were selected for
the follow-up category – the second posting about the news – were in fact initial receptions in terms
of content. They were thus analyzed as such.
The table in appendix 1 summarizes the selection of data and units of analysis. Please see
appendix 2 for more details on the selection and sampling processes.
Framing analysis
To analyze the content of the selected media productions, this research has resorted to framing
analysis.
A very general definition of framing was proposed by Gamson & Modigliani as a “central
organizing idea or story line that provides meaning to an unfolding strip of events” (cited in Entman
et.al., 2009: 175). This definition according to Entman et.al. (idem) does not provide sufficient basis
for consistent measurement or theory, but it does clarify the main idea of this branch of analysis.
More specifically, it is useful to distinguish between issue-specific frames and generic frames (as
proposed by de Vreese, 2005, cited in ibidem: 176). The issue specific frames “are pertinent only to
specific topics or events. (…) Generic frames transcend thematic limitations as they can be identified
across different issues and contexts” (idem).
Basically, “[the] idea is that framing and frames are primary means through which people
make sense of a complicated world” (Entman et.al, 2009: 179). Frame analysis, then, “examines the
selection and salience of certain aspects of an issue by exploring images, stereotypes, metaphors,
actors and messages” (ibidem: 180).
This type of analysis fits especially well with the concept of news discourse as story. The
method is essentially language-based (cf. Zelizer, 2004: 128). Framing provides “an important
pathway for thinking about language use in conjunction with the intersection between journalists and
their public.” (ibidem: 142). The idea of studying the use of language in news discourse closely – as
framing analysis does – can be situated within the tradition of critical linguistics, which – among
other things - holds that “journalists and other media workers can never evade the power structures
which shape the vocabulary and other aspects of the way the language makes sense.” (Matheson,
2005: 5). Language can never be free from power. Framing analysis as interpreted here thus connects
notions put forward in the literature review about storytelling, ideology and language. Frames reflect
the people and the (organizational) context which produced them: the powers around a media
producer (journalist, blogger, tweeter) and his/her personal bias and interpretation are imprinted in
the texts they produce. As far as a media producer controls the selection of a frame, it represents the
exercise of his/her power.
“Framing processes occur at four levels: in the culture, in the minds of elites and professional
political communicators; in the texts of communications; and in the minds of individual citizens”
(Entman (1993, 2004) as cited in Entman et.al, 2009:176). The issue-specific frames studied in this
research resonate with broader frames that are present not only in specific media discourses, but also
in the minds of people (both producers and consumers). A characteristic of these frames is that they
are diachronic in nature: “A framing message has particular cultural resonance; it calls to mind
currently congruent elements or schemas that were stored in the past.” (Entman et.al., 2009: 177).
This research focuses on the issue-specific frames surrounding the Wikileaks Iraq video and takes the
qualitative approach to framing, as clarified by the work of Pan & Kosicki (1993).
Pan & Kosicki identify four categories of framing devices in news discourse, representing four
structural dimensions of that discourse. Analysis on every structural level yields insight into the
nature of the frame that connects different signifying elements into a coherent whole. This analysis
does not uncover the actual meaning of the story as it perceived by audiences. “Rather, it generates a
data matrix of signifying elements that might result in different interpretations.” (Hall, 1980, as
paraphrased in Pan & Kosicki, 1993: 69). The methodological benefit of this analysis is that it can be
10
replicated for different discourses on one issue and that the results can be aggregated to observations
on a higher level (idem). This is exactly what the current research intends to do.
The structural dimensions include (cf. Pan & Kosicki, 1993: 59-63):
1. syntactical structure (or macro syntax, i.e. form of content which can be analyzed without
semantic analysis),
2. script structure (or story grammar, i.e. elements that organize events or actions),
3. thematic structure (or hypothesis testing, i.e. putting forward propositions)
4. and rhetorical structure (or stylistic choices, i.e. choice of words, categorizations, metaphors).
The current study has analyzed the syntactical structure of the different media generically, i.e. once
for every media category (cf. Entman et.al., 176). Specific script structures were not analyzed in this
study, because the main proposition that all media productions should be regarded as stories has
already been put forward. This includes an understanding of the „story grammar‟ of all news
discourse, namely that it puts people, events, themes, places and scenes into a coherent (and often
dramatic) story. (This is not to deny that different types of stories exist.) That thematic and rhetorical
structure will be analyzed for each media production.
Units of analysis
Identifying (issue-)specific frames in the selected media productions was done by studying the
thematic and rhetorical elements in the units of analysis, which are the headline and lead paragraph
for articles and blog postings, or the whole tweet alternatively.
The headline and lead paragraph were chosen as the units of analysis because they represent
the “heart” of the story (Schudson, 2003: 182). “The importance of the lead or first paragraph in
establishing the main point of a news story is clear. (…) It summarizes the central action and
establishes the point of the story”. It is where people “get the main point of a story from reading a
single opening sentence”. The headline, then, “is an abstract of the abstract. The lead pares the story
back to its essential point, and the headline abstracts the lead itself” (Bell, 2001: 239). Or as Allan
puts it (1999:83):
Headline: represents the principal topic or ‘key fact’ at stake in the account. To the extent that it is
recognized as performing this function by the readers, it is likely to influence their interpretation of
the account to follow. In this way, then, it helps to set down the ideological criteria by which the
reader is to ‘make sense’ of what follows.
News lead: typically the opening paragraph or two providing a summary or abstract of the account’s
essential ‘peg’ or ‘hook’ which projects, in turn, ‘the story’ in a particular direction or ‘angle’. The five
Ws and H (the who, what, where, when, why and how most pertinent to the event) will likely be in the
lead or first paragraph.
The „ideological criteria‟ that Allan mentions are important. The selection of the story angle is not a
neutral choice, rather a product of interpreting and filtering on the basis of both professional and
political norms and beliefs. The „inverted pyramid‟ story-format puts the information deemed most
important and „newsworthy‟ on top. “[The] account proceeds to structure the remaing details in a
descending order of discursive (and usually ideological) significance” (Allan, 1999: 83).
Categorization of data
In an iterative (inductive) process of collecting, coding and categorizing data, a list of working frames
for the codation of general thematic choices was generated. These were used to code the better part
of all data, after which they were reviewed and partly adapted, resulting in a definitive list of thematic
frames. Categories for rhetorical/lexical choices were used to code more specific, local properties of
11
the news discourse. Both the working frames and rhetorical/lexical categories were put down in a
coding scheme and used as variables (cf. Entman et.al., 2009: 180).
An example of a general news frame identified in the initial reception of the news is the
“Wikileaks publication frame”, meaning the discourse in the unit of analysis emphasized the
publication of the video over its revelations. Another example is the “Wikileaks Collateral Murder
(WCM) frame”, which means the article or posting copied the Wikileaks frame to convey the news.
Examples of coding categories for specific rhetorical/lexical choices include but are not
limited to: “victim denomination” (e.g. „Iraqis‟, „journalists‟, „civilians and reporters‟), “description of
incident” (e.g. „attack‟, „shooting‟, „killing‟, „murder‟) and “description of video” (e.g. „combat‟,
„shocking, „brutal‟).
When combined, analysis of these basic news- or issue-frames and rhetorical/lexical choices
yields insight into the thematic and rhetorical structure of the media productions at hand.
Again, these frames represent the primary angle to the news as identified in headline and lead only.
The analysis does not cover the rest of the discourse, because it is primarily interested in the „issue-
framing‟ by the content producers (Entman et.al, 2009: 182). What do the different meaning makers
emphasize? This question of emphasis is central to the idea of framing, as framing according to
Entman „„essentially involves selection and salience. To frame is to select some aspects of a perceived
reality and make them more salient in a communicating text‟‟ (1993: 52). Or put differently: “[The]
intended meaning of a news story [which this type of frame-analysis shows, JM] has the capability of
directing attention as well as restricting the perspectives available to audiences” (Hall, 1980;
Tuchman, 1978, as paraphrased in Pan & Kosicki, 1993: 59).
A full exploration of the WCM frame can be found in appendix 3. A detailed description of all frames
and coding categories can be found in the coding scheme, in Appendix 4. Appendix 5 shows two
complete examples of the framing analysis. Appendix 6 shows all data used for the analysis, i.e. all
articles (6a), postings (6b) and tweets (6c, 6d). Appendix 7 contains the actual analysis; it shows the
coding of the data in the coding schemes.
Validity
Validity of the results will be ensured by providing thick, rich description of the material as much as
possible within the given constraints (of length) of this paper, and by openly reflecting on the
shortcomings of the study, as well as disclosing the assumptions, beliefs and biases that might have
influenced the researcher‟s interpretation. (These validity procedures borrow from both the
constructivist (interpretative) and critical paradigm (Creswell & Miller, 2000: 125).)
Stability
Stability of measurement was assessed by retesting 25% of all data, two weeks after the original
analysis. This method is not ideal to provide evidence for strong reliability claims, but within the
context of this research, it is the best test available. According to Riffe et.al., accordance should be
found on 80% of all tested cases (1998: 128). Babbie & Baxter (2004:248) claim 70% is enough. A
sequence of random numbers was used to select cases for the retest. It should be noted that in case of
the newspapers and blogs, the researcher still remembered part of the last coding session.
The retest suggests high stability, i.e. the researcher seems to be consistent, though not
completely. As for the newspaper and blog analysis, agreement was found in 87% of the retested data
(with a total of 62 codings in 10 articles/blog postings). In the retesting of tweets, 95% agreement was
found on a total of 80 codings (in 50 tweets). These percentages were not corrected for chance
agreement because, frankly, the researcher does not know how to use a statistical correlation test to
do so. Test-retest data can be found in appendix 8.
12
5. LIMITATIONS
Before moving to the results of the analysis, some notes on the research, disclosure of assumptions,
beliefs and biases that might have influenced the researcher‟s interpretation, and a discussion of
concrete limitations of the method employed will follow. The researcher is of the opinion that these
should be conveyed before turning to the results.
For the record, it should be stated that the researcher has thought about the ethical
implications of this study. However, no obstacles worth an extensive review were encountered. The
only point that could be raised is that of the privacy of tweeters, who although tweeting in public,
often feel they are communicating with their friends and followers only. One could argue that tweets
used for research should thus be made anonymous, protecting the privacy of their authors. However,
this researcher is of the opinion that was is public, is public. Secondly, the current study is not
suggesting judgment of the content of tweets. In this respect, there really is not much case for offense.
Researcher bias
Determining issue-frames and categories was an interpretative process, as was the interpretation of
the framing analysis that follows. Because of this interpretative aspect, the reader has the right to
know what bias in the mind of the researcher might have skewed the interpretation one way or the
other.
To start, I found the Wikileaks video horrible to watch. It made me hate war more than I had
consciously hated it before.
Being a blogger and tweeter myself, I was personally involved in the reception of the news. I
wrote about it on my blog and I tweeted about it (mostly in Dutch). I was disappointed with
mainstream media coverage of the Wikileaks video, especially with the framing of it in some (Dutch)
newspapers. I felt the coverage failed to adequately describe how ugly the images and the comments
of the soldiers were. A lot of the media I read framed the issue within the “war-is-hell” narrative, as
Haddow (2010) calls it. I personally feel that this frame is a missed opportunity to discuss the horror
of war more broadly. The video could have been the start of a conversation. Within big media, I did
not witness that conversation as much as I would have personally „liked‟. Still, I realize that there is
no „wrong‟ or „right‟ interpretation of the news.
These ideas on the coverage made me think of this research in the first place. I wondered
whether other media step in when mainstream media do not facilitate the emotional processing of the
images and this conversation on the horrors of war. This question may have led me to look for
otherness on blogs and on Twitter, as opposed to the generally more traditional newspapers.
However, I have done an honest attempt to answer the research question independent of these ideas.
I have not consciously ignored signs that did not fit my implicit hypothesis.
Shortcomings in sampling
The sampling techniques that were used to gather the data were not flawless. For the sampling of
tweets, the research was completely dependent on Google search results. The Google update search is
a new service and it is unclear what „rules‟ Google applies for the scanning of Twitter. In general,
Google is not a neutral search tool that scans the whole internet. According to some estimations, its
robots scan only 25% of the web. Others estimate this percentage to be 50% (Van Ess, 2009: 34).
Either way, the search results Google produces are never complete. Secondly, Google „ranks‟ pages on
a number of aspects, including the number of links to it. If a similar mechanism is in place for the
scanning of Twitter, this would mean that Google update search only scans popular tweets or
tweeters. This dependency on Google‟s scanning technique of Twitter is far from ideal, but within the
boundaries of this research, acceptable. It should be noted that the Twitter conversation studied here
may not be representative of the actual activity on the whole of Twitter.
13
Also, it is unclear whether or not the Tweeters researched here were American citizens, like
the storytellers at newspapers and blogs (supposedly). This showed when one of my own tweets –
written in English – appeared in the Google search results. To take a more extreme and unlikely
example: some of the anti-war statements in the sample may have been written by Iraqis, in English,
and thus show up in an English Google update search. Manually, it is possible to check each tweeter
for his/her „location‟. But this would have taken too much time within the boundaries of this
research.
Similarly, there is a somewhat unfair mechanism in place for the selection of blogs. Because
blogs were only selected if they had in fact reported on the news (see appendix 2), popular news blogs
that ignored the news were not part of the sample. In the case of top trafficked newspaper Web sites,
however, those that did not report on the news online were included in the sample. Unfortunately,
without this sampling technique for blogs, it was impossible to come up with a definitive list that was
both popular and in this case, relevant.
In general, the best has been done to provide representative and relevant samples with the
given limitations of online sampling.
Shortcomings of the analysis
Blogs at newspaper Web sites were excluded from the analysis. This is a debatable choice. In the end,
the argument that made for excluding blogs on newspaper Web sites was that the research looks at
the „official‟ reception of the news, one that is prominent on the website. It is assumed that blog
postings on newspaper Web sites do not make it to the (digital) front page and are mostly read by
specialists and „fans‟. Accordingly, they cannot be seen as representative of the reception of the news
by the journalistic organization of the newspaper. However, it should be noted that newspapers are
incorporating blogs on their websites, and that this has consequences for their online storytelling.
One could argue that comparing news articles and blog posts to tweets is an exercise in apples
and oranges, because in the case of tweets, the whole thing is taken into account, whereas the articles
and blogs are cut short after the lead. However, as was pointed out several times, the research looks
at the story angle. Accordingly, a tweet should be compared to the headline of an article or blog
posting. The analysis identifies the main issue-frame.
This research does not examine the motives of storytellers. The consequence of a radical
storyteller perspective is that a pr-spokesperson‟s story is analyzed in the same terms as an
independent journalist‟s story. After all, everybody is a storyteller and no form of discourse has „final
jurisdiction‟ over others (see literature review). Because obviously these motives do matter, future
study should tackle this problem by distinguishing somehow between storyteller motives (such as
commercial, political, entertaining, etc.).
This research although recognizing the interactive nature of online news production, has not
taken into account the actual interaction between different media. The content producers were
analyzed in isolation. Especially blogs and tweets are a place for hyper-linking, but the analysis did
not include a category to check where they linked to (e.g. „personal blog‟, „newspaper Web site‟).
Future research should incorporate this aspect so that an impression of the entangled nature of the
online storytelling arises. This includes an understanding of the globalised character of the
conversation. Although it was studied as being so, in reality the online storytelling is not local. Again,
this hints to future research.
Comments about the size of the samples are beside the point, as this study makes no claims of
representativeness on a quantitative basis.
14
6. RESULTS The following results provide the answer to the first research question:
R1: How was the leaking of the Wikileaks Iraq video framed on American newspaper Web sites,
professional news blogs, and on Twitter?
Syntactical structure
The syntactical structure of the different platforms in this research is markedly different. Newspapers
typically produce articles like they would for the newspaper, but instead publish them online (first).
The textual structure is much the same as a print article would be, starting with a lead after the
headline. However, newspaper Web sites are, like blogs, much more interactive than a print article
and they provide the journalists with new reporting tools, like hyper-linking and embedding video.
Blog postings are similar to online newspaper articles in terms of their syntactical structure.
Twitter is a completely different story. It provides users („tweeters‟) with the possibility to
produce short 140 character messages, which is precisely the length of this sentence. Tweeters who
follow one another can than see each other‟s messages and reply or republish the message to their
own followers („retweet‟). Most tweets are public, i.e. also accessible for unsubscribed visitors, and
include links to content on other online platforms. All in all, Twitter facilitates instant but short
communication. Many journalistic organizations, including Wikileaks, have created an account to
promote their content via Twitter.
One very important distinction between newspapers and the other platforms, is that
conventionally, newspapers separate opinions (in op-eds) from facts (in articles). Blogs make this
distinction less markedly. In form, tweets do not differentiate at all. One tweet can be about a pet,
while the next carries political content. In these online spaces, the conventional separation in form
between „fact‟ and „fiction‟ seems to fade. As for the journalistic ideology of blogs, it shows that in
comparison to newspapers, they have a different conception of news and objectivity. The ideal of
objective reporting is not represented in the same form as it is in online newspapers. While these
notions are somewhat rough and generalised, they are relevant for the interpretation of the results to
come. Of other political consequence are the lay-out, the use of images, the size of headlines, and so
on. These can vary wildly for both online newspaper Web sites and blogs. It is beyond the scope of
this research to discuss these differences in detail.
It is important to note that these platforms co-exist in an endless online space and that there
are plentiful linkages between them. This is the macro-syntax of any of these platforms, if you will:
they are interactive and interconnected.
Initial reception
Of the 10 newspapers, 2 did not produce an initial reception of the news. One of them, The Wall
Street Journal, did link to other websites who reported the news. Of the 8 remaining initial
receptions of the news by newspapers, a majority (6) emphasized the content of the video in their
initial reception, while framing the news on their own terms, i.e. not copying the WCM frame. Take
the example of The New York Times:
15
One newspaper initially emphasized the publication of the video over its revelations. Another
communicated the news within the 2007 incident frame when first reporting on the news,
emphasizing what happened three years ago before speaking of the release of the video.
The blogs were decidedly less uniform in their first reporting on the news. To start, three blogs
emphasized the publication of the video by Wikileaks in their first reporting, taking as the main news
angle the fact that the video got published, only then followed by what it showed. Take for instance
the headline of The Daily Beast: “Wikileaks releases Iraq-Killing Video”. After this headline, a short
posting clarifies what the video shows, but the main news according to the headline is the publishing
of the video.
Remarkably, two blogs communicated the news by directly debating the content of the video –
asking whether it showed collateral murder – and thereby assuming the reader was already aware of
the news. Take the example of the Hot Air Blog, who headlined its posting: “Video: Collateral murder,
or the risk of war zones?” One blog communicated the news by framing it within the precedent of
Reuters trying to obtain the video in vain.
Of the five remaining blogs, only one primarily framed the news exclusively in their own
terms, i.e. not copying any part of the WCM frame in their initial reception. Three blogs
communicated the news within the WCM frame, while attributing that frame to Wikileaks – making
clear that this interpretation is not necessarily theirs (using for instance quotation marks). An
example is the HuffingtonPost report, which read the headline: “Wikileaks VIDEO Exposes 2007
„Collateral Murder‟ In Iraq”. Two other blogs copied the WCM frame without distancing themselves
from it, implying the WCM frame is the „right‟ interpretation of the events shown in the video. An
example is the posting by Truthdig, which was titled: “Collateral Murder in Iraq”.
On Twitter, there were many different first reactions to the news. 75 out of the 100 tweets
analyzed for the initial reception were concerned somehow with breaking the news, or „helping‟ to do
so. No less than 45% of these tweeters copied the WCM frame directly by retweeting the Wikileaks
news tweet (cited above), or by phrasing the news in their own terms but clearly within the WCM
frame. Another 7% of these news breaking tweeters communicated the news within the WCM frame,
while attributing that frame to Wikileaks and thereby taking a subtle distance. One tweeter directly
refuted the WCM frame by saying “you say zoom lens, I say RPG” – defending the soldiers‟ actions.
The remaining 47% of news breaking tweeters communicated the news in their own terms, i.e.
without clearly referring to or copying from the WCM frame.
The 46 tweets that were not breaking the news communicated about it in many ways. Eight
tweeters reacted emotionally, saying for instance the video was “sickening” or that “no video has ever
depressed me more”. Three others made comments about the coverage of the news, noting for
instance that “this story is exploding!” Two tweeters criticized mainstream media, e.g. “American
media show ignorance over Wikileaks leaked video of soldiers shooting civilians in Iraq…” Twelve
others made statements that were interpreted as being political or anti-war, like for instance the
phrase “END WARS NOW!” or the label “America‟s disgrace” or the statement that “someone should
be accountable”. Other tweets included citations from the soldiers in the video (3), recommendations
for watching it (4), words of thanks to Wikileaks (4), and statements about the video authenticity (7).
Comparing these platforms, this analysis shows that in their initial reception, newspapers
were more likely to frame the news in their own terms, not copying Wikileaks‟ suggestion of
“Collateral Murder”. Half of the blogs communicated the news within the WCM frame, either with
attribution or without, while the other half chose a different interpretation altogether. Almost half of
the tweets that were analyzed here copied the WCM frame literally, mostly by retweeting. This is a
considerable difference in the choice of the main news angle, especially with the newspaper Web
sites. In other tweets, a variety of reactions to the news could be found, including for instance
emotional reactions and anti-war or political statements.
16
Follow up
Two newspapers did not follow up on the news online. Three others followed up on the news by
publishing an (identical) AP story which quoted a senior military official saying the video was real.
The five other newspaper follow-ups were stories about the White House response to the video, the
fact that the military supposedly lost its copy of the video, a story about the military criticizing the
lack of context in the edited version of the video, one about the role of social media like Wikileaks in
current day journalism, and one on Wikileaks itself.
One blog also followed-up on the news by conveying that an official had confirmed its
authenticity. Another told the story of the U.S. Army‟s CENTCOM publishing a redacted version of its
own investigation into the 2007 incident. Two other blog postings were stories debating the incident
and what the video shows. Finally, another two blogs published posts about comedian Stephen
Colbert who interviewed Wikileaks‟ editor in chief, Julian Assange.
The follow up on Twitter was as diverse as the initial reception and even more so, because in
the initial reception, almost half of all tweets were copies (often retweets) of the WCM frame. The
topics of the April 7 tweets are shown in figure a (Appendix 9).
First and foremost, comparing the platforms on their follow-up shows a rich diversity of
stories. Among these, remarkable findings include the fact that (1) three newspapers published the
same AP story, and that (2) two blogs discussed what the video showed, while in this sample, no
newspaper dedicated a follow-up story to do this. Twitter facilitated a very diverse follow-up in terms
of content, with tweets often pointing to coverage on other platforms.
Please note that neither the image of the initial reception nor that of the follow up provides a
complete overview of all internet activity surrounding the news. For instance, Web sites of major
television news networks were not part of the sample of this study, although they are far more
popular than newspaper Web sites for the consumption of news (Pew, 2010). Besides, the analysis
above has only looked at the principal thematic choices in headlines and leads.
Rhetorical/Lexical properties
The above analysis of the initial reception and the follow-up of the news on the different platforms
was focused on the thematical choices made by the different meaning makers, it analyzed the basic
themes and topics of their stories. Now, in which terms did they tell these stories? The analysis of the
frames is not complete without looking at the rhetorical and lexical properties of the texts, i.e. the
choice of words and stylistic devices to convey the news.
The following analysis is concerned with the description of the video (was it a „combat‟ video
or was it „harrowing‟?), the labels or descriptions of the incident (was it an „attack‟ or a „massacre‟?),
the denomination of the victims (were they „men‟ or „unarmed civilians‟?), the denomination of the
culprit (was a US helicopter or were American soldiers shooting in the video?) and the description of
the source („Wikileaks‟ or „a Web site‟?).
The headlines and leads of all newspapers and all blogs were analyzed on these properties, but
not all tweets. Only those tweets that were not copying (i.e. retweeting) the WCM frame and that were
involved with breaking the news, were analyzed on these properties because (1) analyzing the (exact)
frame copies (i.e. retweets) would yield identical data time and time again and (2) these tweets do not
represent the story telling characteristics of tweeters, because they are copies of somebody else‟s
story. Tweets that were not concerned with breaking the news were excluded for this part of the
analysis too, because they would not be on the topic of the video itself and thus not use the terms
analyzed here. In short, 35 April 5 non-WCM news tweets were coded for their rhetorical/lexical
properties.
Description of the video
See figure b in appendix 9: Newspapers and blogs in about 35% of all (34) cases described the video
as a “killing video”, a “shocking video” or a “brutal video” (or a similar variation to these terms).
17
These powerful words were far less common on Twitter, where most tweeters described it either as
the “Wikileaks video” (56% of 18 tweets) or a “combat video” (28%). (17 other tweets did not specify a
description of the video at all.) Some newspapers were unique to describe the quality of the video,
calling it “gritty” (10%). Newspapers were also more likely to call the video “leaked” (25%), rather
than the “Wikileaks video” (only 5%). These results may be explained by the nature of these
platforms. On Twitter, users have only very limited space to articulate the news and adjectives or
other descriptives may be left out, unless when functional (like „Wikileaks‟ or to a lesser extent,
„combat‟). Conversely, blogs and newspapers have several chances – in the headline and the lead – to
describe the video with an adjective. It is remarkable how blogs were far more likely to use the
descriptive “Wikileaks” than newspapers, who in this sample seemed relatively more comfortable
with the term “combat” or “leaked”.
Description of incident
See figure c in appendix 9: Newspapers, and to a slightly lesser extent, blogs, were more likely than
tweeters to describe the incident shown in the video as an „attack‟, a „firefight‟ or an „engagement‟.
This type of technical description for the images in the video can be interpreted as distant: it allows
for some distance between the culprit and the victims because it were not „soldiers shooting‟ but „an
attack killing‟. Newspapers were more likely than blogs, and blogs more likely than tweeters, to
describe the incident as a „shooting‟. However, the most used category of description for all platforms
was a term focused on the act of killing. This term was used most by tweeters (59% of 29 tweets),
followed by newspapers (40% of 20 newspaper articles). About 25% of blogs and tweeters described
the video in Wikileaks‟ terms or in line with Wikileaks terms (i.e. „murder‟, „slaying‟, „indiscriminate
targeting and killing‟, „atrocity‟), with blogs attributing these descriptions to Wikileaks in 5 out of 7
cases. Only one newspaper article described the incident in terms that were in line with the powerful
WCM terms mentioned before, referring to the troops as „unleashing devastation and killing (…)‟.
Apart from this one instance, the most graphic descriptions in the newspaper headlines and leads
were focused on the act of killing: „shooting and killing‟, „gunning down‟. The difference with the blogs
and tweeters in this respect is considerable, their language in many cases was more direct, with blogs
and two tweeters using WCM frame terms by describing the incident as an „atrocity‟ or a „massacre‟. It
is important to note that in this sample, the most confronting descriptions of the incident were
almost exclusively used on new online platforms such as blogs and Twitter.
Denomination of victims
See figure d in appendix 9: Finally, the analysis has shown that the victims were denominated in a
wide variety of ways. In the headline and the lead paragraph, newspapers were most likely (in 46% of
13 cases) to name those killed in the video “journalists” or “Reuters employees” only (i.e. ignoring in
the first instance the civilian and/or insurgent casualties). Others (23%) called the victims a “group of
men, some of whom were unarmed”. Finally, some newspapers (15%) named those killed in the video
„Iraqis and reporters‟, while others chose to denominate them as „journalists, and possibly armed
men‟ (8%), or simply as „Iraqis‟ or „civilians‟ (8%).
The blogs used different denominations in their headlines and lead paragraphs. One part (of
16 cases) specified all victims as „reporters‟ only (13%) or „Iraqis/civilians‟ (6%). More than half of all
bloggers denominated the victims as „Iraqis and reporters‟ (56%). Finally, 25% of bloggers described
the victims of the shooting as „Iraqis, journalists, and a van collecting the wounded‟.
Tweeters were (like newspapers) most likely (in 37% of 27 cases) to specify the victims of the
killings as „reporters‟ only. Almost 30% of tweeters described the victims as „Iraqis‟ or „civilians‟.
18,5% denominated those killed in the video as „Iraqis and reporters‟. Remarkably, 11% of tweeters
(equaling 3 tweets in the sample) described the victims as „unarmed civilians and journalists‟.
It is remarkable that the denomination that can be considered most complete and specific,
„Iraqis, journalists, and a van collecting the wounded‟, was used only by bloggers and once by a
18
tweeter (who mentioned the shooting at children). To the other meaning makers, some of the victims
were apparently not part of „the who‟ that is typically contained in the lead. Nobody put the shooting
at the people and children in the van on first place.
Many newspapers were quick to emphasize the killing of Reuters reporters, emphasizing the
loss of journalists over the loss of Iraqi civilians or the wounding of children. This is not to say these
newspapers neglected to mention them altogether (e.g. later in the article), but it does say that they
chose not to emphasize these victims alongside the reporters, something 4 out of 16 blog postings did
choose to do. It is remarkable too that some tweeters described the victims as „unarmed civilians and
journalists‟, when clearly at least some of the men in the video were armed. Most precise descriptions
(„some unarmed‟, „possibly armed men‟) could be found on the newspaper Web sites. Contrary to
these precise descriptions, however, 6 out of 13 newspapers described the victims primarily and only
as journalists, when clearly the Reuters employees were not the only victims.
Culprit
Finally, the analysis has determined how the culprit – the one shooting in the video – was
denominated. It showed that on all media platforms, content producers were most likely (67-81%) to
denominate the culprit in an impersonal way, e.g. as „U.S. Army‟, „American helicopter‟ or „Apache‟.
The other (beginnings of) stories analyzed here denominated the culprit more directly as (American)
people, calling the shooters „U.S. forces‟, „soldiers‟ or „troops‟ among other things.
Source
For newspapers and blogs, the analysis has coded the denomination of the source. Most productions
on these platforms denominated the source fully as „Wikileaks‟, or for instance „the whistle-blower
Web site Wikileaks‟. One newspaper hid the Wikileaks name (calling the source „a website‟), while
seven newspapers and three blogs produced one article or posting that did not name the source of the
video at all. It can be concluded that in the coverage analyzed in this sample, the source (i.e.
Wikileaks) of the video was emphasized more often in the headline or lead paragraph of blog postings
than that of newspaper articles.
7. CONCLUSIONS Images have the power to disrupt dominant narratives. They may not (always) have a direct influence
on politics and policy-making, but they might have a “considerable impact on popular imagination
and historical consciousness” (Papadopoulos, 2008: 6). “Iconic news photographs might have long-
term repercussions for the shaping of public consciousness and national memory (…)” (ibidem:23). I
would argue this goes for video as well as for photographs. We should recognize images‟ power to
directly engage with publics. As Papadopoulos argues in the case of the Abu Ghraib torture
photographs, images can function as their own frame, i.e. not have as a main function to support
another frame in a news text (ibidem: 6).
The Wikileaks Iraq video, however, was directly incorporated within a frame. The first ever
time audiences saw the images, they were already part of Wikileaks‟ proposed story of „collateral
murder‟. From the title to the tweet that notified people of the leak, Wikileaks interpreted the images
every step of the way, and it gave them political meaning before audiences had the chance to do so. By
all means, this was forceful meaning making – it was an act of storytelling outside of conventional
journalistic platforms.
And it was quite successful, at that. On Twitter, the WCM frame was widely reproduced. Half
of all blogs copied it too, while some emphasized the source of the frame by attribution. Wall (2009:
21) has found that much of the collecting, constructing and distributing of information on war via
social media such as blogs still echoes the military‟s preferred narratives, irrespective of “who the
storytellers happens to be”. The reception of the Wikileaks video shows that new media organizations
can now successfully bypass these „preferred narratives‟ in favor of their own story.
19
The analysis of the newspaper‟s interpretation of the news showed the undoing of the Wikileaks
Collateral Murder frame. The vocabulary used in newspaper articles was markedly different from the
original „collateral murder‟ story. The terms used to describe the incident in newspaper articles were
less graphic than those used in blogs and in tweets, which stayed closer to the WCM frame. The
newspapers, like the tweeters, barely emphasized the victims of the van-shooting as the first issue,
many conversely naming only the Reuters reporters as victims in the headline and lead. Wikileaks
tweeted that for them, “the focus on the Iraq massacre response should be the cover-up and the
van/missile attack”. Pronouncing this preference on the framing of the video so explicitly, makes it all
the more interesting to see how effectively (especially) newspapers were at removing the WCM frame
and telling the story in their own terms. With some exceptions, the newspaper articles removed the
sharp edges in Wikileaks‟ discourse, translating the news event into a discourse that was less
confronting. Blogs were in many ways in between much of the dominant meaning making in
newspapers and the diverse conversation on Twitter. But on comparison, they stayed much closer to
the WCM frame than newspapers did. Like Twitter, they were less uniform in their reception than
newspapers.
Professional journalists‟ unwillingness to retell the story that Wikileaks proposed might be
understood by the uneasy relationship between the mainstream press and the online collective.
Wikileaks‟ aggressively proactive behavior when promoting its leaks (and in this case, its story) might
irritate journalists from traditional news outlets. Mainstream media in this research (i.e. the
newspapers) may not have been indifferent to Wikileaks‟ material, they were indifferent to Wikileaks‟
proposed story.
There is one big if to these claims. The cases researched here were all news articles, never op-
ed pieces. The op-ed pages may have literally copied the WCM frame, and this research would have
never known. But the finding of this research and the following conclusions do apply to newspaper‟s
primary discourse.
To conclude, the second research question will be answered by combining and comparing the
results from the analysis above with the points raised in the literature review, thereby aggregating
findings to a higher level of understanding.
R2: If the framing on the different online platforms differs, what does this tell us about these
platforms and about the people using them?
On functions
When regarding it as the professional journalists‟ job to bring events within the bounds of the normal
and comprehensible, to fit new stories within existing narratives, it is no surprise that the WCM
frame was discarded. News discourse often works to depoliticize the issues at hand (cf. Allan, 1999:
77). When information is placed in a hard news story, this involves undoing it of subjective emotions
and primary reactions. One could even argue that newspapers had no other choice then to discard the
WCM frame, because it was full of blatantly “subjective language”. Copying it would have comprised
the impartial stance the organizations working within this paradigm seek. The analysis shows that
newspapers efficiently exercise this objectifying function, that within their paradigm, they act as
cleaners for the status quo (which exists by virtue of existing narratives). A similar point could be
made for the blogs that communicated the WCM frame, but took a subtle distance by attributing it to
Wikileaks.
On Twitter, this is not the case. Twitter functioned as a much more effective echo chamber for
the WCM frame. Seeing the results of this research, Twitter shows the potential to be a diverse
platform for the distribution of content, whereas blogs and newspapers specialize in one function to
serve their audiences (or advertisers). However, it should be noted that many tweets link to story on
other platforms and in this sense, Twitter is as much a place for conversation as it is for aggregation
(or advertising).
20
The literature review suggested a complementary relationship between media. This
proposition is supported here because no one platform fulfilled all identified functions. Different
functions – for instance the re-framing by newspapers, the emotional and political communication on
Twitter – can co-exist online, which suggests that all are necessary.
Talking of media functions raises questions on the future of (online) investigative journalism.
It has been noted that amateurs step in and take over (part of) the functions of journalists, while
rarely making them completely abundant. But with the leaking of the Wikileaks Iraq video, it shows
that exclusively online organizations can perform tasks that were once the exclusive domain of
traditional media. Generally, the internet is not seen as a good place for investigative journalism
because there is no business model in place to pay for it. But Wikileaks is showing traditional outlets
that online journalistic enterprises can perform their tasks and tell their own stories. Lynch (2010)
suggests that some professional journalists are using Wikileaks as a safehouse for sensitive
information. This development paves the way for seeing the internet as a place for collaborative
journalism by all sorts of meaning makers. For the future of journalism, this is a hopeful sign.
On form
The form that storytelling takes on Twitter facilitates process and negotiation of meaning. It is not
one tweet, but the mental portrait that all tweets construct over time that matters to users (cf.
Hermida, 2010). The form of the newspaper Web site is much more definitive. It suggests that much
of the meaning making has already been done. Professional news blogs are somewhat in between
these two opposites, with blogs usually being more interactive than newspaper Web sites.
Earlier, this paper proposed that because the internet is not a neutral medium, it is likely to
influence the producers that function within its libertarian context. This proposition can be both
endorsed and refuted with the current results in hand. First, it can be endorsed because of Wikileaks‟
political inclination. Wikileaks poses as a radical outsider, ready to break the world open. In fact, the
story researched here took on the form of its creator. The proposed story fits with the storyteller‟s
profile and with the anti-authoritative nature of the internet as identified by Keen (2009).
But second, this proposition can be refuted when looking at the newspapers. They did not
fundamentally change their function when they went online, although they might have shifted to a
more interactive and accessible form of storytelling. The medium may be the message, but within its
borders, many messages can be communicated.
On power
It was stated in the literature review that frame analysis by extension lays bare the imprint of power
and the context in which the storytellers function. Looking at the results from this perspective, what
can be concluded? Most interestingly, it shows that newspapers have the power to re-frame the
proposed story. Their function was to undo the „bare facts‟ from the story Wikileaks proposed, and to
use them to retell the story using within another, less confronting narrative.
The power of blogs in this context can be interpreted in two ways. Either, they had the power
to resist re-framing the Wikileaks story into a cleaner version, or they were powerless to resist
Wikileaks‟ framing.
Twitter had the potential to be all of the above, as it was both an aggregator for the stories
other online media produced, and a platform for the negotiation of meaning.
The meaning makers in this research used their power to make meaning differently. At
newspapers and professional blogs, few write for many to read. The meaning makers at these
organizations have relatively much power to (re-)create stories. On Twitter, many write for relatively
few to read. But the tweeters have the power of the collective to make meaning. Their conversation is
open and the negotiation of meaning among users creates story.
21
On political consequences
What are the political consequences of the framing in the online media that were studied here?
Basically, the research has shown that on the internet, several frames can co-exist. No dominant
frame was identified across all platforms. This finding is important because it has consequences for
the interpretation of the news by readers. These readers are offered different vocabularies to think
about the news. They are confronted with different denominations of key actors and events in the
video. These different frames construct different worlds „out there‟ – they direct attention to other
parts of the information and they restrict the number of perspectives – which obviously has major
political consequences.
It should be noted that there is no way of knowing how exactly the proposed frames on the
different platforms were interpreted by their consumers. We can say, however, that framing heavily
influences readers‟ responses to news discourse because “on most matters of social or political
interest, people are not generally [very] well-informed and cognitively active (…). [Framing] therefore
heavily influences their responses to communications (…)” (Entman, 1993: 56).
This arguably had major consequences for people using Twitter and directly encountering the
WCM frame, which was widely spread there. Communication on Twitter is quick, unfiltered and
direct. These characteristics offer the suggestion of consuming „raw information‟. In this case,
however, the WCM story was anything but „raw‟, it had been finely edited before making it online.
Tweeters may have been unaware of the extent to which Wikileaks framed the issue in the video,
heavily influencing their interpretation. Arguably, this suggestion was perpetuated by the very nature
of video, which like photojournalism is “cloaked in [a] mantle of objectivity” (Schwarts, 1992:108,
cited in Allan, 1999:87).
Moving further, the fact that several frames co-existed can be interpreted as having
detrimental effects on the shared culture of a society. If people do not exchange and negotiate story,
they become isolated from one another. If it were true that people use only one medium to get the
news, these findings would endorse Keen‟s fear of “the risk of a relativistic world devoid of common
stories [as a result of the opposing, co-existing frames on the Web (JM)]” (cited in Bird & Dardenne,
2009: 213).
However, research shows that only 7% of Americans get the news from only one medium on a
typical day. Twitter is not commonly used as a news source: in a research by Pew it was the least cited
news source. Conversely, “[some] 46% of Americans say they get the news from four to six media
platforms on a typical day” (Pew, 2010). It can thus be concluded that the different framing on the
online platforms might have influenced the same readers and that the different frames do not exist in
isolation but rather are re-mixed by both consumers and producers of news.
On Homo narrans
In the perspective of this paper, content producers are seen as a product of their surroundings, not as
free agents who – except for some superficial choices – have the power to select the stories they wish
to tell. This perspective makes for a different interpretation of the frames used. For newspapers and
blogs, we can imagine the professional, institutional context of their story production. For Twitter,
there is no way of knowing the institutional context that controlled the meaning makers.
When not viewing the human meaning makers who‟s content was studied in this research as
different in a personal way – when seeing them all as storytellers within different institutional
contexts – this research shows most of all that storytellers will tell one story or the other dependent
on the context of their production, and dependent on the medium they use. People on different
platforms produced different meaning making. We are left to conclude that if anything, Homo
narrans is flexible.
22
References
Allan, S. (1999). The cultural politics of news discourse. In Allan, S. News culture (2nd ed. pp.77-97).
Maidenhead: Open University Press.
Andén-Papadopoulos, K. (2008). The Abu Ghraib torture photographs. News frames, visual culture, and the
power of images. Journalism. Theory, practice and criticism, 9:1, 5-30.
Barnhurst, K.G. (2003). The makers of meaning: National public radio and the new long journalism, 1980-
2000. Political communication, 20:1, 1-22.
Babbie, E. & Baxter, L. A. (2004). The basics of communication research. Belmont: Wadsworth/Thomson
Learning.
Bell, A. (2001). News stories as narratives. In Jaworski, A. & Couplands, N. (Eds.), The discourse reader (pp.
236-251). London: Routledge.
Bird, E. S. en Dardenne, R. W. (2009). Rethinking news and myth as storytelling. In K. Wahl-Jorgensen & T. Hanitzsch (Eds.), The handbook of journalism studies. (pp. 205-218). New York: Routledge.
Bourdieu, P. (1992). Enkele eigenschappen van velden [Some characteristics of fields]. In idem, Opstellen over
smaak, habitus en het veldbegrip [Essays on taste, habitus and the concept of field], (pp.171-178). Amsterdam:
Van Gennep.
Bourdieu, P. (2005). The political field, the social science field, and the journalistic field. In Benson, R. &
Neveu, E. (Eds.), Bourdieu and the journalistic field (pp.29-47). Cambridge: Polity.
Broersma, M. (in press). Journalism as performative discourse. The importance of form and style in journalism.
In Rupar, V (Ed.), Reading the newspaper. Journalism and Meaning-making.
Creswell, J.W. & Miller, D.L. (2000). Determining validity in qualitative inquiry. Theory into practice, 39:3,
124-130.
Deuze, M. (2006). Participation, remediation, bricolage: Considering principal components of a digital culture.
The information society, 22, 63-75.
Ebizma (2010, June). Top 25 most popular political news websites. Retrieved May 3, 2010 from
http://www.ebizmba.com/articles/political-news-websites.
Entman, R. (1993). Framing: Toward Clarification of a Fractured Paradigm. Journal of Communication, 43, 51-
58.
Entman, R., Matthes, J. & Pellicano, L. (2009). Nature, sources, and effects of news framing. In K. Wahl-
Jorgensen & T. Hanitzsch (Eds.), The handbook of journalism studies. (pp. 175-190). New York: Routledge.
Fisher, W.R. (1985). The narrative paradigm: In the beginning. Journal of Communication, 35:4, 74-89.
Gilmor, D. (2004). We the media. Grassroots journalism by the people for the people. Sebastopol, CA: O‟Reilly.
Goode, L. (2009). Social news, citizen journalism and democracy. New media society, 11:8, 1287-1305.
Haddow, D. (April 7, 2010). Grim Truths of Wikileaks Iraq video. Retrieved April 7, 2010, from
http://www.guardian.co.uk/commentisfree/libertycentral/2010/apr/07/wikileaks-collateral-murder-iraq-
video
Hermida, A. (2010). Twittering the news. Journalism practice, first published on march 11, 2010 (iFirst)
Keen, A. (2007). The cult of the amateur. How blogs, Myspace, Youtube, and the rest of today’s user-
generated media are destroying our economy, our culture, and our values. New York: Doubleday.
23
Keen, A. (October 16, 2009) Rushes sequences – Andrew Keen Interview – USA (Video). Retrieved May 30,
2010, from http://www.bbc.co.uk/blogs/digitalrevolution/2009/10/rushes-sequences-andrew-keen-i.shtml
Krippendorff, Klaus (2004). Content analysis. An introduction to its methodology. Thousand Oaks: SAGE
publications.
Lawrence, R.G. (1996). Managing meaning: Media, officials and police brutality. Political Communicatoin, 13:1,
140-142.
Lessig, L. (March 31, 2007). Keen‟s “The cult of the amateur”: BRILLIANT! Retrieved May 30, 2010, from
http://www.lessig.org/blog/2007/05/keens_the_cult_of_the_amateur.html
Lowrey W. & Latta, J. (2008). The routines of blogging. In Patterson, C. & Domingo, D. (Eds.), Making online
news: The ethnography of new media production (pp.185-197). New York: Peter Lang.
Lynch, Lisa (2010). “We're Going To Crack The World Open”. Forthcoming in Journalism Practice. First
published digitally on: March 24, 2010 (iFirst).
Matheson, D (2005). Introduction (Excerpt). In idem, Media discourses. Analysing media texts (pp.1-12).
Maidenhead: Open University Press.
Nielsen (2008, January26). News release on Web traffic to top 10 online newspapers. Retrieved May 3, 2010
from http://www.nielsen-online.com/pr/pr_090127.pdf
O‟Sullivan, J & Heinonen, A. (2008). Old values, new media. Journalism practice, 2:3, 257-371.
Pan, Z. & Kosicki, G (1993). Framing analysis. An approach to news discourse. Political Communication 10, 55-
57.
Pew internet and American life project (March, 2010). Understanding the participatory news consumer.
Summary of findings. Retrieved April 15, 2010, from http://www.pewinternet.org/Reports/2010/Online-
News.aspx?r=1
Pew research center for the people and the press (2009). Newspapers face a challenging calculus. Retrieved
April 15, 2010, from http://pewresearch.org/pubs/1133/decline-print-newspapers-increased-online-news
Reese et.al (2007). Mapping the blogosphere: Professional and citizen-based media in the global news arena.
Journalism, 8, 235-261.
Reuters (2008, July 11). Reuters seeks US army video of staff killed in Iraq. Retrieved May 10, 2010 from
http://www.reuters.com/article/idUSL05399965.
Riffe, D. & Lacy, S. & Fico, F. (1998). Analyzing media messages using quantitative content analysis in
research. Mahwah, NJ: Lawrence Erlbaum Associates.
Roeh, I. (1989). Journalism as storytelling, coverage as narrative. American behavioral scientist, 33, 162-168.
Roggenkamp, K. (2005). Narrating the news: New journalism and literary genre in late nineteenth-century
American newspapers and fiction. Kent: Kent State University Press.
Rosen, J. (June 26, 2006). The people formerly known as the audience. Retrieved June 5, 2010, from
http://journalism.nyu.edu/pubzone/weblogs/pressthink/2006/06/27/ppl_frmr.html
Schudson, M. (1995). The politics of the narrative form. In idem, The power of news (pp. 59-66). Harvard:
Harvard University Press.
Schudson, M. (2003). News as literature and narrative. In: The sociology of news (pp.177-193). New York & London: W.W. Norton.
Tecnorati (May 3, 2010). Top 100 blogs. Retrieved May 3, 2010 from http://technorati.com/blogs/top100/
24
Tecnorati (May 3, 2010). Top 100 political blogs. Retrieved May 3, 2010 from
http://technorati.com/blogs/directory/politics
The New York Times (2007, July 13). 2 Iraqi journalists killed as U.S. forces clash with militias. Retrieved June
6, 2010 from http://www.nytimes.com/2007/07/13/world/middleeast/13iraq.html?_r=2
The New York Times (2010, April 6). Iraq Video Brings Notice to a Web Site. Retrieved April 6, 2010 from
http://www.nytimes.com/2010/04/07/world/07wikileaks.html
The Washington Post (2007, July 13). U.S., Shiite fighters clash in Baghdad. Retrieved June 6, 2010 from
http://www.washingtonpost.com/wp-dyn/content/article/2007/07/12/AR2007071202357.html
Tremayne, M. (2007). Introduction: Examining the blog-media relationship. In idem (Ed.), Blogging,
citizenship, and the future of media (pp.ix-xix). New York & London: Routledge.
Van Dijk, T.A. (2009). News, Discourse, and Ideology . In K. Wahl-Jorgensen & T. Hanitzsch (Eds.), The
handbook of journalism studies. (pp.191-204). New York: Routledge.
Van Ess, H. (2009). De Google code [The Google code]. Amsterdam: Pearson Education Benelux.
Wall, M. (2009). War reporting 2.0. Social media and soldier content. A paper presented at the International
Communication Association annual conference. Chicago, May 2009.
Wikileaks (2010). Collateral murder. Wikileaks Web site for the release of the video.
http://www.collateralmurder.com/.
Wikipedia (2010). Entry for Wikileaks. Retrieved April 7, 2010 from http://en.wikipedia.org/wiki/Wikileaks
Zelizer, B. (2004). Taking journalism seriously.Thousand Oaks, CA: Sage Publications, Inc.
25
Appendix 1: Summary; Selection of data, units of analysis
Platform Selected media productions (all dates 2010) Unit of analysis
10 most traficked U.S.
newspaper Web sites
New York Times, USA
Today, Washington Post,
Los Angeles Times, Wall
Street Journal, Daily
News, Chicago Tribune,
New York Post, Boston
Globe, San Fransisco
Chronicle.
Initital reception: first article published about
the news on the website of the selected
newspaper (presumably published April 5)
N=8
Headline and lead
paragraph of articles
N=16
Drawn from 10
newspapers; initital
reception (N=8) and follow-
up (N=8) produced by same
newspaper
First follow-up to the initital reception: second
article published about the news on the websites
of the selected newspapers (presumably
published April 5, 6, or 7)
N=8
10 popular U.S. blogs
Huffington Post, Hot Air
Blog, The Daily Beast,
Think Progress, Politico,
Talking Points Memo,
Salon, Mediaite, The
Dangerroom at Wired,
Truthdig
Initital reception: first posting published about
the news on the selected blogs (presumably
published April 5)
N=12
Headline and lead
paragraph of postings
N=18
Drawn from 10 blogs;
initital reception (N=12)
and follow-up (N=6)
produced by same blog
First follow-up to the initital reception: second
posting published about the news on the selected
blogs (presumably published April 5, 6, or 7)
N=6
Twitter Initital reception: stratisfied sample (N=100)
from Google search results for
Twitter updates
Query: wikileaks iraq video OR "collateral
murder";
Date: April 5
Whole tweet
N=200
Drawn from 200 tweeters;
initial reception and follow-
up produced by different
tweeters First follow-up on the initial reception:
stratisfied sample (N=100) from Google search
results for:
Twitter updates
Query: wikileaks iraq video OR "collateral
murder";
Date: April 7
Excluded: tweets that break the news for the first
time
26
Appendix 2 : Selection and sampling processes
Although this research takes a qualitative approach and will not claim representativeness (i.e.
generalizability) on a quantitative basis, it is important to select a sample as objective as
possible in terms of content. With the given limitations of searching for and archiving from
the different media, this research has intended to provide a representative image of the
activity on the different platforms. This means the researcher has not been cherrypicking
articles or postings, but that all possible checks were put in place so that the sample under
study would actually be representative of the activity on the given platforms.
There are three different types of content productions:
Newspaper articles
Postings on blogs
Updates on Twitter
In the following, the selection of specific cases will be clarified.
Newspapers: the top ten most traficked US newspaper websites according to the most
recent Nielsen/NetRatings available (2008) were selected:
1. New York Times (http://www.nytimes.com )
2. USA Today (http://www.usatoday.com )
3. Washington Post (http://www.washingtonpost.com )
4. Los Angeles Times (http://www.latimes.com )
5. Wall Street Journal (http://www.online.wsj.com )
6. Daily News (http://www.nydailynews.com )
7. Chicago Tribune (http://www.chicagotribune.com )
8. New York Post (http://www.nypost.com )
9. Boston Globe (http://www.boston.com )
10. SFGate.com / San Fransisco Chronicle (http://www.sfgate.com )
Blogs: the most popular and influential news blogs that covered the Wikileaks Iraq video
were selected. A list of blogs was compiled using the combined data of Technorati (a site
that measures a blog‟s standing & influence in the blogosphere by looking at linking
behavior, categorization and other associated data over a short, finite period of time)
overall and political top 100 rankings (as of May 3, 2010); an Ebizma ranking for most
popular blogs (2010); and a New York Times blog list (2010).
o The focus was on „holistic‟ news blogs that cover all sorts of news themselves. They
are seen in this research as journalistic organizations.
o Aggregators were excluded.
o One-man blogs at traditional media institutions were excluded. Andrew Sullivan
of The Daily Dish, for instance, writes for The Atlantic‟s website, but his blog
although very popular cannot be seen as an online journalism organization.
o One-issue blogs were left out (e.g. Media Matters, Right Wing Watch).
o Blogs were only included if they had in fact reported on the Wikileaks Iraq video.
27
The selected blogs are:
1. Huffington Post (http://www.huffingtonpost.com )
2. Hot Air Blog (http://www.hotair.com )
3. The Daily Beast (http://www.thedailybeast.com )
4. Think Progress (http://www.thinkprogress.com )
5. Politico (http://www.politico.com )
6. Talking Points Memo (http://www.talkingpointsmemo.com )
7. Salon (http://www.salon.com )
8. Mediaite (http://www.mediaite.com )
9. The Dangerroom at Wired (http://www.wired.com/dangerroom )
10. Truthdig (http://www.truthdig.com )
Twitter: selected were updates that show up in a Google update search for the query
wikileaks iraq video OR "collateral murder". Google provides the possiblity to show
Twitter updates between very precise time frames:
From the newspaper websites and blogs, the (1) initial reception of the news and (2) the first
follow-up on the initial reception were located via search and archived for analysis. (Postings
appearing on newspaper blogs were excluded, because this research is interested in the first
official drafts of the news, i.e. newspaper articles.)
It was impossible to take the exact same approach for tweets because one can never be sure
that a tweeter who shows up in Google search results reported on the news more than once
(more than one tweet). Via search, it is simply impossible to find only tweeters who followed
28
up on the initial reception, i.e. to filter out those people that just produced one tweet on the
topic. (Manually, this is possible but it would simply take too much time.)
Therefore, for tweeters, a different approach had to be taken to resemble the initial reception
and the first follow-up. This was done by selecting from different time slots, i.e. tweets
produced on the day the news broke (April 5, 2010) and tweets produced two days after
(April 7, 2010). It may be expected that a tweet two days after the breaking of the news,
contains some sort of follow-up – it is assumed that most tweeters were already up to date on
the initial breaking of the news. This is supported by the fact that tweeters are very active
news followers: “The average Twitter user is two to three times more likely to visit a news
website than the average person” (Farhi, 2009, cited in Hermida, 2010: 2).
For the tweets, it was impossible to establish whether or not the producers were American.
SAMPLING FROM TWEETS
The sampling from the newspaper Web sites and blogs was straighforward in the sense that
the first and second article or posting on topic were selected. The sampling of tweets had to
be done by using a specific targeted approach further described in the appendices containing
the tweets (appendices 6c and 6d). In general terms, it should be stated that a stratisfied,
systematic sample with a random start was used. Because all tweets were collected and sorted
sorted by time of publishing on the selected days, this stratisfied sample was the best way to
ensure a representative sample from tweets published on different times of that day. A
random sample from the list of tweets would have contained the risk of over-representing
some section of the list, i.e. some time of the day (cf. Babby & Baxter, 2004: 154).
29
Appendix 3: Exploration of the Wikileaks Collateral Murder (WCM) frame
1. The following is the text that was published with the video on
http://www.collateralmurder.com/. Terms coined here by Wikileaks that were found in other
media too are highlighted. Especially the first three paragraphs are illustrative of the WCM
frame.
Overview
5th April 2010 10:44 EST WikiLeaks has released a classified US military video depicting the
indiscriminate slaying of over a dozen people in the Iraqi suburb of New Baghdad -- including two
Reuters news staff.
Reuters has been trying to obtain the video through the Freedom of Information Act, without
success since the time of the attack. The video, shot from an Apache helicopter gun-sight, clearly
shows the unprovoked slaying of a wounded Reuters employee and his rescuers. Two young children
involved in the rescue were also seriously wounded.
The military did not reveal how the Reuters staff were killed, and stated that they did not know how the children were injured.
After demands by Reuters, the incident was investigated and the U.S. military concluded that the actions of the soldiers were in accordance with the law of armed conflict and its own "Rules of Engagement".
Consequently, WikiLeaks has released the classified Rules of Engagement for 2006, 2007 and 2008, revealing these rules before, during, and after the killings.
WikiLeaks has released both the original 38 minutes video and a shorter version with an initial analysis. Subtitles have been added to both versions from the radio transmissions.
WikiLeaks obtained this video as well as supporting documents from a number of military whistleblowers. WikiLeaks goes to great lengths to verify the authenticity of the information it receives. We have analyzed the information about this incident from a variety of source material. We have spoken to witnesses and journalists directly involved in the incident.
WikiLeaks wants to ensure that all the leaked information it receives gets the attention it deserves. In this particular case, some of the people killed were journalists that were simply doing their jobs: putting their lives at risk in order to report on war. Iraq is a very dangerous place for journalists: from 2003- 2009, 139 journalists were killed while doing their work.
2. The following are a few tweets by Wikileaks that enforce and elaborate on the issue-frame
put forward above.
http://twitter.com/wikileaks/status/11641921146:
30
http://twitter.com/wikileaks/status/11644350489:
http://twitter.com/wikileaks/status/11677894090:
http://twitter.com/wikileaks/status/11678711175:
http://twitter.com/wikileaks/status/11700518875:
3. Summary of the WCM Frame
Using this text and these tweets, the WCM frame can be summarized as follows:
The video shows murder by US troops. The incident is a massacre. The victims are
Iraqis and civilians, or over a dozen people -- including two Reuters staff.
The focus on the news should be on the „murderous attack on the van/wounded‟ and
on „the cover-up‟ of the incident.
Media should not be „strawmanning the camera/rpg confusion‟ as the issue in the
video.
31
Appendix 4a: Coding scheme for newspaper articles, blog
postings
Thematical Categories T1-T14
Possibilities, examples Coding rules
INITITAL RECEPTION The news is:
If not identical, headline frame overrules lead paragraph frame
T1. Wikileaks publication First main news is not the video itself, but Wikileaks publishing it. “Wikileaks publishes, releases, posts video of [Bagdhad attack that killed two journalists].”
T2. Video itself a. Own frame News angle independent from WCM frame (…) Video shows killing(s), shooting, etc.
b. Very similar to WCM frame
WCM frame with attribution, subtle distance Video shows „what Wikileaks calls‟ „Collateral Murder’ OR Video shows alledged cover-up OR Video shows „murder‟ in Iraq
c. WCM frame Exact WCM frame Video shows Collateral Murder in Iraq, and/or cover-up
T3. Incident News angle refers to the incident two years ago. “Two years ago…”
T4. Precedent News angle is the precedent to the publishing: Reuters had in vain tried to get the video
T5. Discussion of video News angle is a direct discussion of the content of the video: Collateral murder or the risk of war zones? Cover-up by army or something else?
FOLLOW-UP The follow-up:
T6. Video authenticity The video is real, authentic, say sources “(Senior) official(s) say(s) the video is real”
T7. White House responds The White House responds to the video. T8. Lost tape Army, military cannot find its version of leaked tape T9. CENTCOM Army CENTCOM releases documents of own investigation
into 2007 incident T10. Journalism Social media influence on journalism / war reportage / future
of journalism T11. Video criticism Criticism from US army, military, commentators: images lack
context T12. Wikileaks Focus on Wikileaks, the organization, precedents, history T13. Discussion over incident, what video shows
Does the video show murder? How to interpret the video?
T14. Colbert interviews Wikileaks‟ Assange
Rhetorical & Lexical Categories R1-R5
Possibilities, examples Coding rules
For both initital reception and follow-up
Multiple adjective categories possible. Please note the term “2007 video” is not coded here.
32
R1. Adjective before „video‟ OR „footage‟ OR „images‟, description immediately after (e.g. the video, which is graphic,…)
a. SOURCING Wikileaks(‟), Wikileaks “Collateral Murder” video [only with brackets], Wikileaks Iraq video, United States military
b. CATEGORIZING AS WAR IMAGES Combat, war, Apache gunship, guncamera
c. EMPHASIZING LEAKED/CLASSIFIED CHARACTER Classified, until now classified, long-sought, leaked, decrypted
d. EMPHASIZING KILLING, SHOOTING (Iraq)-killing, Iraq (shooting)
e. LABELLING CONTROVERSIAL Contraversial
f. EMPHASIZING SHOCKING CHARACTER Disturbing, shocking, graphic
g. EMPHASIZING BRUTAL CHARACTER Brutal, grisly, harrowing, disgusting, deadly, sickening, depressing, deeply disturbing, gruesome
h. PERTAINING TO QUALITY Gritty
Others: no adjective Multiple labels possible R2. Labels for / description of incident
a. TECHNICAL TERM Attack (that killed), firefight, air attack, engagement, strike
b. FOCUSSED ON THE ACT OF SHOOTING Shooting (at), firing (repeatedly) (on), opening fire on
c. FOCUSSED ON THE ACT OF KILLING Shooting and killing, killing(s), gunning down, gunned down, …gunnfire kills
d. WCM FRAME WITH ATTRIBUTION What Wikileaks calls „collateral murder‟, „slaying, „coverup‟
e. WCM FRAME COPY Murdering / murder, indiscriminate targeting and killing, slaying [without brackets], massacre, cover-up
f. IN LINE WITH WCM FRAME Slaughtering / slaugther, uleashing devastation and killing, atrocity
g. HORRIBLE MISTAKE Grave and horrible mistake
Only one source per unit of analysis. Please note, when no source was named at all, no coding.
L1. Source a. FULL SOURCE Wikileaks, the web site Wikileaks, the whistleblower web site Wikileaks
b. SOURCE NAME HIDDEN A (whistleblowers) website
No source Only one vicitim category per unit of analysis, choose the one
with most determined victims L2. Victims As named in the title and lead paragraph
a. IRAQIS / CIVILIANS People in Iraq, Iraqis, (Iraqi) civilians, (group of) men, more than a dozen Iraqis
b. A GROUP OF MEN, SOME UNARMED (…a group of men, some of whom were) unarmed (men)
c. ONLY JOURNALIST(S) ((Two) Reuters) journalists, employees, news photographer (and his driver)
d. JOURNALISTS AND (POSSIBILY) ARMED MEN Men who are believed to be holding guns and ((two) Reuters) journalists
e. IRAQIS/PEOPLE AND REPORTERS
33
Iraqis, group of men and journalists, group of men including journalists, “shooting into Iraqi crowd, including two Reuters employees”, “killing reporters and (dozens of) others”
f. IRAQIS, JOURNALISTS and VAN COLLETING WOUNDED / CHILDREN Iraqis, group of men and journalists, group of men including journalists, “shooting into Iraqi crowd, killing two Reuters employees” and a van collecting the wounded( innocent civilians, children)
Only one culprit category per unit of analysis. Naming forces overrules naming the above
L3. Culprit a. COUNTRY, ARMY or HELICOPTER US, US Army, US military, US Army,military/American helicopter, Apache
b. FORCES US (or American) forces (or soldiers, or troops, pilot, personell) “Video shows soldiers firing in on…”
34
Appendix 4b: Coding scheme for tweets
Thematical Categories
TW1-16
Possibilities, examples Coding rules
April 5 reception on Twitter
Tweet can fit into multiple categories.
TW1. Emotional Emotional reactions, comments after viewing. Any one emotionally loaded word makes tweet fit this category “Shocked after seeing”, “this is depressing”, “makes me sad”, “angry”, “Goddamn”, “nuts”, “OMG!”
TW2. News notification (typically includes link)
a. Own frame News is framed in own words, no clear traces of WCM frame “Video shows killing”
For these frames, also code rhetorical/lexical choices blow
b. WCM frame with attribution
WCM frame with attribution, subtle distance „Murder‟ between brackets, “Wikileaks calls it “Collateral Murder””, “footage believed to show cover-up”…
c. Exact WCM frame (copy)
Exact WCM frame (copy) Wikileaks LEAK retweet / Own words but clearly in Wikileaks frame: Video shows Collateral Murder in Iraq, and/or cover-up Note the term “indiscriminate slaying” was introduced by Wikileaks.
d. Rejecting WCM frame
“You say zoom lens, I say RPG” – implying the soldiers were in the right (no Collateral Murder)
TW3. Media reaction a. Neutral notifications
Neutral notifications or questions on story progress “Anybody seen official reactions yet?” “Video has now been watched 1,5 million times”
b. Mainstream media criticism
Mainstream media coverage criticism (includes critical questions) “Media ignoring this video” “NYTimes falsely implies…” “Where‟s the mainstream media?”
TW4. Anti-war, political comments
Comments implying dislike or disgust from war and conflict, and the government waging it, in general “This is how you win hearts and minds”, “war is terrible”, “we wonder why they hate us”, “love your government”, “someone should be accountable”, “human rights”
TW5. Citation from video
Quoting helicopter pilots from the video “Ha, ha, I hit them”
TW6. Recommendation after viewing
People telling one another to watch the video “MUST SEE!”, Shouldn‟t more people watch this?
TW7. Thanking Wikileaks
People thanking Wikileaks for leaking the video. “God bless Wikileaks”
TW8. Video authenticity
Video is real, says authority, AP
Other Link to torrent, Wikileaks also publishes rules of Engagement, Calling the
video “funny”
Rhetorical & Lexical Categories TWR1-2, TWL1-2
Possibilities, examples Coding rules
For own frames in news notification tweets
Multiple adjective categories possible. Please note the term “2007 video” is not coded here.
35
TWR1. Adjective before „video‟ OR „footage‟ OR „images‟, description immediately after (e.g. the video, which is graphic,…)
a. SOURCING Wikileaks(‟), Wikileaks “Collateral Murder” video [only with brackets], Wikileaks Iraq video, United States military b. CATEGORIZING AS WAR IMAGES Combat, war, Apache gunship, guncamera c. EMPHASIZING LEAKED/CLASSIFIED CHARACTER Classified, until now classified, long-sought, leaked, decrypted d. EMPHASIZING KILLING, SHOOTING (Iraq)-killing, Iraq (shooting), deadly
e. LABELLING CONTROVERSIAL Contraversial
f. EMPHASIZING SHOCKING CHARACTER Disturbing, shocking, graphic
g. EMPHASIZING BRUTAL CHARACTER Brutal, grisly, harrowing, disgusting, deadly, sickening, depressing, deeply disturbing, gruesome
h. PERTAINING TO QUALITY Gritty
NO ADJECTIVE Multiple labels possible TWR2. Labels for / description of incident
a. TECHNICAL TERM Attack (that killed), firefight, air attack, engagement, strike
b. FOCUSSED ON THE ACT OF SHOOTING Shooting (at), firing (repeatedly) (on), opening fire on, raid
c. FOCUSSED ON THE ACT OF KILLING Shooting and killing, killing(s), gunning down, gunned down, …gunnfire kills
d. WCM FRAME WITH ATTRIBUTION What Wikileaks calls „collateral murder‟, „slaying, „coverup‟
e. WCM FRAME COPY Murdering / murder, indiscriminate targeting and killing, slaying [without brackets], massacre, cover-up
f. IN LINE WITH WCM FRAME „Atrocity‟
g. ALTERNATIVE Grave and horrible mistake
Others: no description of incident, f.i. Tweet: “Wikileaks posts combat video” Only one vicitim category per unit of analysis, choose the one with most
determined victims TWL1 .Victims As named in the tweet
a. IRAQIS / CIVILIANS People in Iraq, Iraqis, (Iraqi) civilians, (group of) men, more than a dozen Iraqis
b. UNARMED CIVILIANS / IRAQIS Unarmed Iraqis, (Iraqi) civilians, (group of) men
c. A GROUP OF MEN, SOME UNARMED (…a group of men, some of whom were) unarmed (men)
d. ONLY JOURNALIST(S) ((Two) Reuters) journalists, employees, news photographer (and his driver)
e. JOURNALISTS AND (POSSIBILY) ARMED MEN Men who are believed to be holding guns and ((two) Reuters) journalists
f. UNARMED CIVILIANS / MEN AND JOURNALISTS Murder of unarmed civilians and journalists
e. IRAQIS/PEOPLE AND JOURNALISTS Iraqis, group of men and journalists, group of men including journalists, “shooting into Iraqi crowd, including two Reuters employees”, “killing reporters and (dozens of) others” (no „armed‟ or „unarmed‟)
f. IRAQIS, JOURNALISTS and VAN COLLETING WOUNDED / CHILDREN
36
Iraqis, group of men and journalists, group of men including journalists, “shooting into Iraqi crowd, killing two Reuters employees” and a van collecting the wounded( innocent civilians, children)
OTHERS: NO VICTIMS MENTIONED
Only one culprit category per unit of analysis. Naming forces overrules
naming the above TWL2. Culprit a. COUNTRY, ARMY or HELICOPTER
US, US Army, US military, US Army,military/American helicopter, Apache b. FORCES
US (or American) forces (or soldiers, or troops, pilot, personell) “Video shows soldiers firing in on…”
Others: no mention of „culprit‟
April 7 follow-up on Twitter
TW8. White House responds
The White House responds to the video.
TW9. Lost tape Army, military cannot find its version of leaked tape
TW10. Journalism, media
Social media influence on journalism / war reportage / future of journalism
TW11. Video criticism
Criticism from US army, military, commentators: images lack context
TW12. Wikileaks Focus on Wikileaks, the organization, precedents, history includes for instance tweets referring to NYTimes article “Notice to a Web Site” and notifications that Wikileaks has 10.000+ new followers
TW13. Broad discussion, questions over incident, what video shows
Does the video show murder? How to interpret the video? What are your views? Includes links to coverage on other platforms, f.i. Assange on Russia today, Al Jazeera with the families, DemocracyNow interview, Douglas Haddow‟s popular article on “Grim truths”, Glenn Greenwalds “Iraq slaughter not an aberration”
TW14. Coverage notifications
Neutral notifications on story progress or questions “Anybody seen official reactions yet?” “Video has now been watched 1,5 million times”
TW15. Mainstream media criticism
Mainstream media coverage criticism “Media ignoring this video” “NYTimes falsely implies”
TW16. Iraq outrage Outrage in Iraq after the video got published TW17. Connection to other wars
“Incident happened over 90 times in Gaza”
TW18. Video authenticity confirmed / questioned
Video is real officials say, video may not be real military says
Other comments “Where is the outrage?” “War crime takes two days to reach you”
37
Newspapers (N=20) Blogs (N=14) Non-WCM news tweets (N=18)
BRUTAL
SHOCKING
KILLING
GRITTY
CONTRAVERSIAL
LEAKED
COMBAT
WIKILEAKS
Figure B. Description of video (%)
Appendix 9: figures showing some results of the analysis
White house response
Military's lost video Journalism & social media
Criticism of the Video
Wikileaks
Broad discussion of video contents and
consequences, with links to
coverage on other platforms including
f.i. Salon, The Guardian, Al-
Jazeera, NYTimes
Coverage progress
notifications
Mainstream media criticism
News about outrage in Iraq
over video
Connection to other war (Gaza)
Video authenticity confirmed /
debatedOther
Figure A. Story topics of Twitter follow up, April 7 2010 (N=100)
38
Newspapers (N=13) Blogs (N=16) Non-WCM news tweets (N=27)
JOURNALISTS AND (POSSIBILY) ARMED MEN
GROUP OP MEN, SOME UNARMED
UNARMED CIVILIANS AND JOURNALISTS
IRAQIS, JOURNALISTS AND VAN COLLECTING WOUNDED / CHILDREN
IRAQIS AND REPORTERS
ONLY JOURNALISTS
IRAQIS / CIVILIANS
Figure D. Denomination of victims in newspapers, blogs and non-WCM news tweets (%)
Newspapers (N=20) Blogs (N=29) Non-WCM news tweets (N=29)
HORRIBLE MISTAKE
IN LINE WITH WCM FRAME
WCM FRAME COPY (MURDER)
WCM FRAME WITH ATTRIBUTION
FOCUS ON ACT OF KILLING
FOCUS ON ACT OF SHOOTING
TECHNICAL TERM
Figure C. Description of incident (%)
Top Related