The present analysis examined news values and the form of news on Twitter, during the period of January 25-February 25 and focusing on the Egyptian uprisings that prompted the resignation of ex-President Mubarak, as depicted via the #egypt hashtag. As the most prominent tag used during this period of turmoil, this tag also featured a majority of tweets cross-posted to other frequently used tags, such as #Jan25 or #Tahrir, thus expanding our sample further. Archives were obtained from the online archive service of Twapperkeeper, an online tool for capturing public timelines, or archives, of tweets more extensive than the ones provided by the Twitter API. The archives constructed included tweets generated during the aforementioned time period, and contained the text of the tweet, hashtags, keywords, date and time stamps in various formats, and miscellaneous bits of backend information that are recorded based on user set preferences. Usernames were also included, but were removed from the file for further analysis.
The Twapperkeeper datasets are delivered in standard comma separated value format (CSV), which sustains several inconsistencies and noise. We used a variety of programming scripts and filters to organize the data set into a workable format. We collected a total of approximately 1.5 million tweets from the #egypt tag, and measured frequency of tweets shared during the period analyzed via R, an open-source software1 program. For the purposes of subsequent content and discourse analysis, we filtered out tweets containing Arabic characters, which resulted in eliminated approximately 400,000 tweets. We eliminated the tweets for practical reasons, as the foreign characters were unfortunately not recognizable by the content analysis tools. Still, given that our focus is on global news, listening practices, and news values, the sample we worked with fit our study objectives of studying who tweets, but also who is able to listen, and what they are listening to. We analyzed a total of approximately 1.1 million tweets utilizing Latin characters, some of which were multilingual.
Computerized Content Analysis
A random sample of 9,000 tweets was drawn from the #egypt corpus of tweets. This sample was analyzed using centering resonance analysis (CRA), a mode of computer-assisted network-based text analysis that represents the content of large sets of texts by identifying the most important words that link other words in the network (Corman & Dooley, 2006; Corman, Kuhn, McPhee, & Dooley, 2002). The analysis of a just a portion of the tweets collected was necessary due to software limitations.
CRA calculates the words’ influence within texts and sets of texts, using their position in the text’s structure of words (Dooley & Corman, 2002). This influence is based on words’ coefficient of betweenness centrality, defined by Corman et al. (2002) as “the extent to which a particular centering word mediates chains of association in the CRA network” (p. 177). As Dooley and Corman (2002) stated, “words with high betweenness, and thus influence, add coherence to the text by connecting strings of words that otherwise would not be connected” (p. 123). The results of aggregating the possible centers or nodes (the most influential words) in a message denote the author’s intentional acts regarding word choice and message meaning. The concept of resonance also allows us to compare sets of text to detect similarities and differences. As Corman et al. (2002) stated the more two texts frequently use the same words in influential positions, the more word resonance they have. The more word resonance they have, the more the communicators used the same words, and the more those words were prominent in structuring the text’s coherence. (p. 178). Based on these concepts, the tweets were analyzed to detect the most influential words.
Qualitative textual analysis techniques (e.g., Fairclough 1995, 2000; van Dijk 1997) pursue a deep explanation of meaning by observing and recording patterns present in a mediated text. The qualitative analysis sought to verify, expand, and illuminate the quantitative findings of the content analysis. This study examined discourse (as defined by Wood & Kroger, 2000 and Fairclough,1995) as a text, using the Wood and Kroger definition of discourse as “all spoken and written forms of language use (talk and text) as social practice” (p. 19). Therefore, the aim of this textual analysis was to understand the “systematic links between texts, discourse practices, and sociocultural practices” (Fairclough 1995: 17). In identifying news values, we used Hartley’s definition of news values as ever-evolving and reflective of news stories and not news events themselves. Our goal was to understand how the medium of Twitter is employed in turning events into news stories, told by the hybrid and networked publics of journalists and citizens working concurrently. In analyzing the text, we referred back to this definition and prior categorizations of news values, identified in previous research and detailed in the previous section.
The sample for discourse analysis was assembled through a composite approach, using stratified sampling to construct a corpus of 150,000 tweets, or roughly a little over 10% of the total sample, which were read and analyzed in greater detail for the purposes of the discourse analysis. The files were also extensively perused to get a feel for the pace and progression of the twitter stream. The selected tweets were then read over, several times, to identify news values using the afore mentioned framework. Notes were taken regarding language use, tone, presence or absence of traditional news values and news values previously identified in research, focus, and differences and similarities in how people shared information over Twitter. We looked for thematic patterns, repetition, and redundancy in the trends that we observed. Finally, notes and findings were categorized and further analyzed, in light of previous research on news values and the form of news. The combined quantitative and qualitative approach sought to expand validity and reliability.
Two graphs were created to visually represent the sheer amount of information shared via Twitter during the 2011 uprisings in Egypt. As previously mentioned, a total of approximately 1.5 million tweets from the #egypt tag from January 25 to February 25.
Insert Graph 1 and 2 here
Both graphs present the amount of tweets shared in intervals of 5 minutes. Graph 1 shows over 10,000 tweets using the #egypt tag were exchanged on February 11, day of Mubarak’s resignation, on average, every 5 minutes. Graph 2 offers a zoomed view of graph 1 in order to make possible a closer examination of the volume of tweets shared during the entire period examined. The frequency analysis of the entire sample is employed to trace the rhythms of news storytelling on Twitter. The computerized content and discourse analyses are used to identify prominent news values and to examine how they shape the form of news storytelling on twitter. Findings from the frequency, content and discourse analysis are summarized and combined, in response to our two main research questions below.
Hybridity of Old and New(er) News Values
The content and discourse analysis both indicated that the stream of news reflected a mix of traditional news values and values specific to the platform of Twitter. The discourse analysis suggested that the types of events covered and the tone of the coverage mimicked the tendency of traditional media to emphasize all of the following news values, as defined by McQuail (2002): large scale of events, closeness to home, clarity of meaning, short time scale, relevance, consonance, personification, significance, and drama and action. The only value identified in traditional media but not present in the twitter feed was that of negativity. Otherwise, and at varying degrees, information and opinions featured regularly on the twitter streams tended to revolve around larger scale events, in proximate locations, were intent on providing clarity and accuracy, prioritized more recent events, were reflective of drama and action, and associated specific persons with aspects of a story.
The computer-mediated text analysis of the #egypt tag showed similar patterns. The quantitative approach adopted in this analysis—centering resonance analysis—is designed to back out patterns of meanings found on precise mathematical rules, avoiding in this way coder bias and sometimes manifesting unexpected findings (Oliveira & Murphy, 2009). However, in this analysis the patterns found corroborate with the qualitative analysis.
Insert Figure 1 here
Words were arranged on the map based on how influential or central they were. The most influential words are those in black boxes; words with slightly lesser influence have gray boxes; and less influential words are unboxed. The lines in the map depicted levels of associations among words, with darker lines depicting stronger associations2 (Corman & Dooley, 2006).
The whole configuration of the map shows one tightly connected cluster of meanings, where the information tweeted about the uprisings in Egypt reflects many traditional news values. The most influential words and the tweets behind them illustrate the following aspects: closeness to home (represented by the words Egypt and Egyptian), personification (Mubarak), significance and relevance (revolution), and drama and action (people and protest).
The discourse analysis further illuminated these trends, suggesting that differences lie not in the news values that are prevalent, but in who makes the decisions based on the same news values. So, whereas in a traditional news room, it is the professional hierarchy and ethos that drives how these news values are applied to judge and cover events, in the case of Twitter, these judgments were made collectively and organically.
Because of the collectivity in the decision making, we noticed some variation in what these values mean for the context of Twitter. So, while the stream focused on larger scale protests, there was a pronounced tendency to ensure that smaller scale protests, occurring in cities peripheral or remote to Cairo were not neglected or undercovered. Proximal locations were covered, but attention was also called to connecting with remote areas of Egypt under turmoil. References were also made to uprisings in neighboring states like Libya, and reaction to the uprisings from potential strategic partners, like the US. In effect, this trend was identified through the computer-mediated text analysis of the #egypt tag too. Having Libya among the most influential words in the corpus of tweets studied highlights the fact that the stories shared via Twitter called attention to the political context of neighboring regions.
The stream leaned toward relevant news and opinions, even though the architecture permits irreverence. Comments that were irrelevant or unrelated were simply not retweeted or ignored, and thus organically eliminated from the process of forming the dominant news story. The topical organization of the hashtag, including the fact that the tags were created specifically for the purpose of covering these events facilitated this focus. The nature of the events tweeted also facilitated the prevalence the news values of ethnocentrism, altruistic democracy and moderatism, which have also been previously identified by Gans (1984) as characteristic of western media. Several tweets reflected pride in the Egyptian ethnic identity, selfless declarations and actions in favor of democracy, and many urgent calls to cover events carefully, accurately, and not rush to judgment. It is not uncommon to encounter altruism and a measure of national pride during political uprisings, as well as an emphasis on using the media at hand to communicate the accurate and authentic version of what is happening to potential audiences and publics. At these times, individuals are recast as journalists. They function based on what they have been socialized to recognize as accepted news values, but they adapt them to the context, what the situation calls for, and their own perspective.
There, were however, specific news values that emerged, that were specific to the platform of Twitter and the context of the uprising, and we spend the next few paragraphs describing those. In line with Hartley’s definition of news values as evolving, but shaping news events into news stories, we identified the four following prominent news values:
Instantaneity. We use the term instantaneity to describe the drama of events unfolding, being recorded and reported online instantly. The ability to live-tweet events as they happen presents the primary appeal of Twitter. At times when mainstream media are restricted in their ability to report, or disseminate information, it is because of this ability that platforms like Twitter rise to prominence. It is also this instantaneity that exposes the temporal incompatibility of Twitter with our conventional definitions of what is news, what separates fact from opinion, and subjectivity from objectivity. Instantaneity, or, the coverage of things that happen as they happen, reigns over the twitter news stream, and individuals are free to tweet their observations. The tone and the language used emphasizes this tendency, with individuals retweeting and requesting instant updates.
The network map of the #egypt tag also supports this interpretation, with a number of other hashtags, such as #libya, #tahrir, and #mubarak emerging as the most influential words. The rhythms of updates posted reflect the consumption with instantaneity, with updates streaming every few seconds, and during certain events, on every second. The tendency to instantly communicate to as many publics as possible is also reflected in the urgency of the language employed and the repetition of instant reports from the ground, in ways that seek to affirm and spread word of mouth retellings of what is going on. Graph 2 illustrates these tendencies, with updates and retweets every second. The repetition of events on the one hand mimics the tendency of media to repeat breaking news and on the other hand, is afforded by the platform, which permits endorsement of information through repetition and crossposting. Tweets frequently use words that convey urgency, like now, live, happening now, and link to sites that offer live streaming of the events. Moreover, the constancy of the updates combined with the tone of the language drum up the heartbeat of a news feed and the movement the feed reflects and mediates.
Crowdsourced elites. It is common for news coverage to award priority to elite nations, organizations, or individuals. While there is no priority granting authority in the organically generated stream of news on Twitter, it is common for elite news organizations and specific individuals to be featured prominently in the stream of news. This is typically facilitated via the logic of tweeting and retweeting stories or news that come from prominent news organizations or individual citizens who provide constant news updates. #Egypt was characterized by patterns through which elites and individuals achieved prominence. In the first few hours that the stream was active, the news feed was populated by tweets of a general nature, commentary, some fact and some opinion. Gradually, as events and protests escalated, media elites started to participate in the news feed regularly, typically through cross-posting headlines and links to stories they were running on their web sites. At the same time, individual citizens, reporting events live or reporting reports of events on a regular basis emerged as primary or adjunct sources of information. Elite status was awarded to those citizens through the practice of retweeting, but also through directly encouraging others to follow the timelines of specific bloggers, activists and ordinary citizens who tweeted constant updates. The tweets contributed from mainstream news sources typically assumed the objective and laconic tone of a headline, with the occasional exception of live tweets produced by journalists, through their individual accounts and not the generic outlet stream, as they were observing events taking place on site. Well known examples include the tweets filed by journalists like Ben Wedeman (@bencnn), Ivan Watson (@ivancnn), Nick Robertson (@nicrobertsoncnn), which were frequently integrated into the taped or live news broadcasts produced for the station affiliate. On occasion, these tweets would integrate fact with opinion, typically integrating reports of events with moderate and careful expressions of solidarity. For example, reporters frequently retweeted expressions of solidarity texted by Egyptians, as a way of reporting public sentiment. These conformed both to the news values of the parent news organization and the evolving values of the news stream. While media elites frequently dominated blocks of the feed through constant tagged updates, they were only awarded leader status through retweets or mentions.
A parallel and more vocal stream of opinion leaders emerged, consisting of bloggers, activists and intellectuals with some prior involvement with online activism that was associated with the uprisings. These included senior Google executive @ghonim, or Wael Ghonim, who had been secretly incarcerated by Egyptian police for 11 days and interrogated regarding his work as the administrator of the Facebook page, "We are all Khaled Saeed", which had helped spark the revolution. They also featured citizens with little to none prior involvement with activism, as was the case with @gsquare86 or Gigi Ibrahim and Mona Seif or @monasosh, both activists/bloggers who rose to prominence through documenting events. And they also included individuals who were not in Egypt during the entirety of the uprisings but who received and retweeted reports, together with their own opinions and comments, as was the case of Mona Eltahawy @monaeltahawy. The discourse analysis revealed that organically emerging leaders interacted with media elites, through processes of retweeting, mentions, and commenting, but differed in the form of their updates, with organic leaders frequently being more openly emotive and media elites trying to balance the values of the parent news organization with the drama of the reports forwarded on twitter.
Solidarity. Tweets documenting events and expressing opinion reflected overwhelming expressions of solidarity. The emergent news streams were characterized by a hybridity of new reports and solidarity, so much so that it became difficult to separate fact from expression of camaraderie, and doing so perhaps misses the point. The network map of prominent words emerging in #egypt reflects this solidarity, in the dense connections that place ‘revolution’ and ‘people’ in the core, connect them to sites of struggle (‘tahrir’) and unity against the cause of the struggle (‘mubarak’, appearing both as a word and a tag), as well as unity for the country of Egypt (also prominent as a word and a tag). The centrality of ‘revolution,’ compared to the presence but peripheral position of ‘protest’ suggests a tendency to affirm this movement as revolutionary, and thus distinguish it from protests that might connect publics but not result in decisive breaks with past hierarchies of governance.
These tendencies are further affirmed by the discourse analysis, revealing a confluence of solidarity and news sharing that is not uncommon during the course and escalation of movements. For example, tweets frequently featured calls like “Its time to come back NOW and join your fellow brothers and sisters,” or “If the dove is a symbol of peace the #Twitter Bird is a symbol of freedom,” or “Muslims and Christians Work Together in a New Egypt,” and “#Libya and #Egypt one hand together ..#Revolution until victory against all dictators” that typically ended with a link to additional content; a photograph, blogpost, live stream, or just a list of several relevant tags and users to follow. The solidarity findings are consistent with previous research, which points to greater social cohesion and a measure of homophily among individuals sharing both topical interests and geolocation.
Ambience. Finally, the constant pace, frequency, and tone of tweets contributed to and constructed an ambient information sharing environment. We term this a news value, because not only does the architecture of the medium invite the constancy and continuity that constitute ambience, but also because the continuous updates, even if redundant, contributed to the creation of a live and lively environment that sustained online and offline expressions of the movement. For example, as graphs 1 and 2 illustrated, on February 11, the day of Mubarak’s resignation, thousands of tweets repeated the same news, before, leading up to, but also well after the event of the resignation had been widely disseminated, even by mainstream news outlets. These tweets did not constitute news updates, but sustained an always-on news environment. They were also focused on communicating personal news, personal emotion, and a genre of news that we term affective.
Affect refers to emotion that is subjectively experienced, and has been connected to processes of premediation, enabled by newer media, that frequently anticipate news or events prior to their occurrence (Grusin,2010). #Egypt is characterized by mounting, emotive anticipation, expressed through posts that are shared to inform, but also frequently simply for the sake of opinion expression and release. These constant and repetitive stream of updates sustains a lively stream of news that is always on, and thus mediates a networked movement that never sleeps. Drawing emotive tweets from those in Egypt and supporters from abroad, tweets conveyed news, solidarity and emotion (“Proud of you Egyptians! Over 20k Ideas and More than 630k votes. Everyone is thinking what should be Egypt 2.0 http://bit.ly/hF5F65 “), sustain cohesion even when there is no news to report (“Good morning sunshine... Good moring my sweet lovely Egypt :) #Egypt #Jan25“), communication emotion, opinion and affection in 140 characters or less (“Seeing amazing footage on AJA ppl are helping the army clean #Tahrir. Oh #Egypt I love u #Jan25 http://dlvr.it/GQ53L“ ), and also invite others to maintain an ambient stream of news that is accurate (“Triple-check news before you retweet. At least today. This is not a video game #Jan25 #Egypt #Tahrir #jan24”). In response to our second research question, concerning the form of news, we explore affective news further.
The Form of Affective News
The shape and rhythm of #egypt is reflective of a form that we characterize as affective, for a number of reasons. Early tweets resembled conversation openers, in that they were too general or too specific; inquisitive and anticipatory in a phatic mode. The streams commenced in an amorphous manner, with tweets like “#Egypt 's street awakening tomorrow #Jan25 #Revolution” or “Egypt is about to have a Facebook revolution,” both retweeting and endorsing the sentiment conveyed in a Time magazine article by the same title. Slowly, the pace of posting attained regularity, with several tweets posted first in a matter of hours, then minutes, then seconds, as reflected through the rhythms of posts depicted in graphs 1 and 2. Eventually opinion leaders emerged, in the form of frequent posters whose tweets were retweeted, and through this means attained visibility, and potentially, credibility. Mainstream media began to chime in, especially as the protests attained greater visibility, and despite the fact that internet access was shut down. Once internet access was reinstated or workarounds became available, the stream regained regularity, and more voices joined the conversation, from Egypt, neighboring countries, some countries in Europe, and primarily, from the US. The texts expressed opinions and reported facts, but rarely new ones.
Tweets blended emotion with opinion, and drama with fact, reflecting deeply subjective accounts and interpretations of events, as they unfolded. Perhaps this is an illustration of what Robinson (2010) had termed ‘finding one’s own place in the story’, in discussing blogged accounts of Katrina catastrophes. Revolving around this drama of instantaneity, tweets were personal, emotive, and involved the sharing of opinion and fact without distinguishing between the two. The drama of instantaneity almost created the perception that events were occurring at a pace faster than that of reality, or as one individual put it on January 25, 2011: “amazing how #social media make #history happens faster…#egypt #Tunisia.”
The progression of tweets was reflective of patterns of repetition and mimicry, similar to trends observed between and within mainstream news organizations (Boczkowski, 2010). Prominent and popular tweets were reproduced and endorsed, contributing to a stream that did not engaged the reader cognitively, but primarily emotionally. Frequently, the same news was repeated over and over again, with little or no new cognitive input, but increasing affective input. This tone of many tweets was deeply emotive but on occasion reflective of the expressive habits of western media, as tweets from western media were frequently quoted with commentary or simply retweeted. The result reflects a confluence of a conversational norms, one that is frequently effected through the oral practices of conversation, reflexivity and reciprocity and opinion expression and listening.
The habitus that developed as individuals fell into familiar practices of information and opinion sharing but also adapted to the context of the platform and the situation reflected a blend of both oral and print cultures. Herein lies the unique character of this platform, the mass communication potential of which is frequently overemphasized at the expense of its interpersonal dimensions. Availing themselves of the affordances of the platform for both interpersonal and mass communication, users engaged in practices of authoring, listening, and editing digital-word-of-mouth-news. The news prodused blended information from mainstream outlets with rumor, fact, opinion, and corrections or edits to prior reports. Calls for and pledges of solidarity and mobilization were extended and received from all directions, and as the internet shut down, posts became more other-directed, concerned with getting the word out.
In response to the internet shutting down and the state monitoring online practices, the stream became more populated with foreigners, or those with secure access to the stream. The resulting stream became even more dense and emotion-filled, characterized by repetition, restating, resaying, and similar expressive patterns that we are more accustomed to encountering in the oral traditions of interpersonal communication. Links to multimedia stories, mainstream and independent media resembled the interpersonal gestures of pointing, nudging, and affirming. They also featured insider Twitter jokes, like “A government that is scared from #Facebook and #Twitter should govern a city in Farmville but not a country like #Egypt #Jan25,“ or “Deleting Dictator...Deleting Installation files...Some files could not be removed. Country still being used...Aborted.#Egypt #Mubarak,” that adapted cosmopolitan references to the local context. The blend of humor, news sharing, opinion expression, and emotion is reflective of the affective patterns of interpersonal conversations. In this manner, affective and ambient news streams might not be perceived as journalistic substitutes, but rather, as alternatives to existing journalistic traditions.