Gatekeeping theory from social fields to social networks.

DeIuliis, David

Gatekeeping theory refers to the control of information as it passes through a gate (Shoemaker & Vos, 2009). The gate is guarded by gatekeepers, who make decisions about what information to let through and what to keep out (Lewin, 1947b). In making these decisions, gatekeepers exercise power over those on the other side of the gate. The intellectual origins of gatekeeping can be traced to Kurt Lewin, a Berlin-born social scientist who applied the methods of individual psychology to the whole social world. Lewin approached gatekeeping as just one of many interrelated phenomena that together make up a social field. To understand gatekeeping, one had to understand the whole field. Lewin's student, David Manning White, was the first to apply the concept of gatekeeping to mass communication. White's (1950) analysis of the gatekeeping decisions of one newspaper editor, called Mr. Gates, focused on the subjective factors that influence gatekeeping decisions. Following White (1950), the field of communication has most often conceptualized gatekeeping as the selection of news, where a small number of news items pass a gate manned by journalists. In making their selections, gatekeepers construct social reality for the gated (Shoemaker, 1991).

The World Wide Web has presented new challenges to these traditional models of gatekeeping, where raw content passes uni-directionally through a gate manned by journalists before reaching the reading public. The ability of users to create and disseminate their own content has uprooted and inverted the roles of gatekeeper and gated. However, if the mass of information on the Web necessitates some form of gatekeeping, what does it look like? Brown (1979) and, more recently, Shoemaker and Vos (2009) and Shoemaker and Reese (2014), call for a return to Lewin. They argue that Lewin's field theory remains relevant for gatekeeping. Much early gatekeeping research followed White (1950) and left Lewin's field theory behind (Shoemaker & Vos, 2009). For Shoemaker and Vos (2009), gatekeeping must reconnect with its origins in field theory and add an audience channel to old models of gatekeeping. In several articles and chapters, Karine Barzilai-Nahon disagrees. For Barziali-Nahon (2009), adding new channels to old models does not adequately account for the dynamism of gatekeeping in new media, and the changed roles of gatekeeper and gated. Barzilai-Nahon proposes a new concept, the "gated," and a new theory, network gatekeeping, to model the dynamism of gatekeeping on new media.

In this review essay, I first describe the intellectual origins of gatekeeping theory in Lewin's field theory. I then trace the development of gatekeeping theory from early debates about its focus to modern questions about its applicability to new media.

I then outline the conceptual apparatus of network gatekeeping theory, and situate network gatekeeping theory within communication studies. Finally, I apply network gatekeeping in the context of Digg, Twitter, and Facebook, three social networks that demonstrate the capacity for not only selecting, but also repeating, channeling, and manipulating information.

1. Intellectual Origins of Gatekeeping Theory

A. Lewin and field theory

The father of gatekeeping theory is Kurt Lewin (Shoemaker & Vos, 2009). Born in Poland in 1890 and raised in Berlin, Lewin was a pioneer of applied psychology in the United States. From his early days in Berlin, Lewin developed a conceptual apparatus for studying human behavior and motives with methodological rigor. Lewin developed best practices for social scientists to understand the social world with the same precision that natural scientists understand the physical world (Cartwright, 1951). In Lewin's time, social science was torn between speculative theory and objective empiricism. Lewin called for a middle ground that makes generalizations about the social world from observations of human behavior. For Lewin, the task of social science should be to conceptualize the world. This process of conceptualization, the translating of phenomena in the world to concepts in the mind (Cartwright, 1951, p. ix), is the heart of scientific inquiry and the building block of theory. Good conceptualization oscillates between qualitative and quantitative, general and particular, part and whole, and group and individual.

The theory of gatekeeping emerged organically out of this process of conceptualization. Lewin conceptualized the social world as a relationship between individuals and groups. Each individual constitutes a "lifespace," which consists of the individual and the individual's environment (Lewin, 1947a). Groups too, comprise lifespaces, made up of the group and its environment. Together, the life spaces of the social world make up a "social field" (Lewin, 1947a). In a social field, the lifespaces of every individual and group coexist within one "ecological setting" (Lewin, 1951, p. 14). The relationship of social field to individual life-space determines human behavior. The social scientist defines a life space by identifying its individual parts, then determining how they relate to the whole social field in space and time (Cartwright, 1951). In defining a lifespace, the social scientist acts as a gatekeeper who "determines specifically what things are to be included in the representation of any given life space at any particular time" (Cartwright, 1951, p. xi). With a series of "in" and "out" decisions about what to include and exclude from an individual life space (Lewin, 1947b, p. 145; White, 1950, p. 383), the social scientist defines the relationship between life space and social field, and individual psychology and social interaction (Shoemaker, 1991).

Gatekeeping decisions about the life space were informed by the intellectual climate in the early 20th century. For Lewin, the individual life space included the thought processes of the individual (Gestalt psychology), as well as the environment that the individual perceives (Husserlian phenomenology), and any unconscious states that affect the individual's psychology (Freudian psychoanalysis). Lewin argued that needs, goals, and cognitive structures, as well as political, economic, and legal processes, must be included in the life space because they directly affect behavior. Remote events and movements have minimal effect and need not be included. The life space moves through history and is affected by past experiences, but only the system as a whole can show effects at any one time (Cartwright, 1951, p. xiii). Lewin's work had a ripple effect on the social sciences, extending centrifugally the limits of the life space.

Similarly, the idea of gatekeeping is inherent within Lewin's conception of the social sciences, informed by Gestalt psychology and logical positivism, as well as the philosophy of Ernst Cassirer. Logical positivism arose in Germany in the 1920s as a response to the idealism of German philosopher Georg Wilhelm Friedrich Hegel (1770-1831), whose idealist philosophy emphasized abstract Spirit over concrete reality (Ayer, 1966). The movement of logical positivism brought philosophy back down to earth by grounding human knowledge in only experience and reason. Logical positivists rejected metaphysics, the study of the underlying nature of things, as meaningless (Ayer, 1966). All theoretical statements are also meaningless until verified. Logical positivists assigned meaning to words according to their practical use and considered language only a representation of the true nature of things. For logical positivists, the language of science should be the common language of human knowledge (Ayer, 1966). Like logical positivism, Gestalt psychology emerged in Germany in the 20th century as a response to certain schools of psychology that divided psychological experience into isolated and distinct parts (Kohler, 1969). Grounded in phenomenology, the study of human conscious experience, Gestalt psychology argues that the sum of psychological experience is other than any of its parts, and understanding any of the parts requires an understanding of the whole (Kohler, 1969).

Lewin's conception of the social sciences was also heavily influenced by Cassirer (1874-1945), a prominent German philosopher and cultural theorist who taught Lewin philosophy at the University of Berlin. In his three-volume Philosophy of Symbolic Forms (1923, 1925, 1929), Cassirer grounded knowledge in human culture. Humans make sense of the world through signs and sign systems (1923). For Cassirer, these signs form the fundamental unit of scientific analysis (Lewin, 1949). Cassirer's philosophy of symbolic forms looked to early, more primitive, forms of human knowledge expressed in natural language. For Cassirer, religion and art emerge from myth, and science from natural language (Lewin, 1949). Symbolic forms progress from emotional expressions of experience with no distinction between appearance and reality, to more abstract representations of types and forms of appearances in natural language, to pure signification of reality in theoretical science (Cassirer, 1923). Cassirer encouraged Lewin to expand the reach of psychology beyond the limits of individual psychology (Lewin, 1949).

In response, Lewin wanted to develop a more democratic method of social management (Cartwright, 1951). Lewin saw an opportunity to combine the objectivity of science with the objectives of people to understand this turmoil. Concepts in individual psychology, such as force fields, fluctuations, and phase spaces, could be used for social processes, and the mathematical tools used to study quasi-stationary equilibria in economics could be used in cultural settings (Lewin, 1951). For Lewin, economic equilibria such as supply and demand were conceptually similar to social processes such as the productivity of a work team. By employing experimental and mathematical procedures of the natural sciences, the social sciences can achieve the same level of specificity as the natural sciences (Lewin, 1951).

However, the Jewish-born Lewin faced a number of challenges in the years following World War II. As the crimes committed by Hitler's Third Reich came to light, some natural scientists acknowledged the potential for social events to shape the natural world. However, not all social scientists agreed about the objective reality of the social world. Some concepts, like leadership, could not shake a "halo of mysticism" that kept them outside the purview of social sciences (Lewin, 1947a, p. 7). In the natural sciences, debates about reality concern the most elementary of physical phenomena, like atoms and electrons. In the social sciences, debates concern the reality of the whole social world (Lewin, 1947a). Lewin argued that, if the body as a whole has different properties from individual molecules, and molecules from individual atoms, the dynamic social whole will have different properties from individuals. Both are as equally real as atoms and the human body. Social science just had to identify the right constructs, then measure them in relation to a social whole, then infer laws about human behavior (Lewin, 1947a). Lewin's methods have informed modern multivariate analysis, which measures causal relationships between two variables by controlling for alternative explanations (Shoemaker & Vos, 2009).

Lewin also saw a disconnect between theory and practice (Cartwright, 1951). Practitioners of the social world, such as priests, parents, and politicians, tend to favor intuition and experience over conceptual analysis and social rules (Lewin, 1947a). For these practitioners, events are complex reflections of the whole environment, not the whims of specific individuals. Lewin argues that social scientists should look for these reflections as underlying conditions of the social world by situating life spaces within social fields, then monitoring changes in the social field caused by individual or group behavior. For instance, to understand the steps leading to war between two nations, the social scientist starts with the lifespace of each nation, then observes what each nation does, then reevaluates the goals, standards, and values of each nation. Both the scientist and practitioner approach elements of the social world as part of a dynamic whole.

Lewin conceptualized social processes objectively as the result of conflicting forces. Lewin's goal was to isolate, identify, and measure the forces, then manipulate the forces to achieve a desired objective. For instance, to change the level of productivity of a work team, a manager must unfreeze the current level of performance, then introduce a new level, then refreeze the new level in the same way as the old (Lewin, 1947a). The productivity of a work team is a "quasi-stationary equilibrium" that fluctuates around an average level (Lewin, 1947a, p. 6). The level of productivity of a work team proceeds in a predictable way despite changing demands and worker turnover. The fluctuation is due to equal and opposite forces, such as the orders of managers and laziness of employees, that push productivity above or below the average level. The productivity of the work team ceases to be quasi-stationary when forces push the level of fluctuation outside the area of the average. When such a force is strong enough to change the average level of fluctuation, opposing forces may bring the average back to the previous level. However, the force may also be so strong that it moves the average level of fluctuation permanently, as in the case of social revolutions (Lewin, 1947a).

Changing the level of fluctuation around a quasi-stationary equilibrium requires attention to the social whole within which the equilibrium fluctuates. The level of fluctuation is often associated with a social custom or habit that resists change, such as procrastination in the workplace. In this case, additional force is needed to break the "inner resistance" of habit and ensure that the level of fluctuation does not revert back to its previous level (Lewin, 1947a). The standards of conduct within a group may also acquire a positive force of their own that ousts individual members of the group who deviate too far from the group standards. In this case, the social equilibrium itself becomes a standard of individual conduct, pushing a work team member to keep up with its pace of productivity. Lewin accounts for not only the relationship between the individual and work team, but also the relationship between the work team and whole cultural context (Cartwright, 1951).

B. Lewin and gatekeeping theory

Lewin articulated his field theory in two articles published in the journal Human Relations in 1947. These papers, along with several other theoretical papers, were collected in 1951 and posthumously published as Field Theory in Social Science. In the first paper, Lewin developed the constructs of social fields and quasi-stationary equilibria, described above. In the second, Lewin proposed "gatekeeping" as a way to examine how objective problems, such as the movement of goods and people, are affected by subjective states and cultural values. In this famous article, Lewin shifts his focus to the social channels that connect individuals to social fields, and the ways to make change at the level of not only a work team, but also society as a whole.

When making widespread social change, such as changing the eating habits of a population, it is impractical and expensive to educate every member of the community. Instead, one needs to target the most influential members of the community who are in a "key position" to spread the message and model the desired behavior (Lewin, 1947b). The key position will depend on the desired social change. For instance, in the late 1940s American housewives were in a key position to change the eating habits of their families. Rather than educating every single American about nutritional value of orange juice, one should look to the person who buys the food for the family.

In this context, Lewin introduced his theory of gatekeeping using the terms channel, section, force, and gate. A channel determines what obstacles an item will face from discovery to use (Lewin, 1947b). Lewin mapped two channels through which food passes on its way to the American kitchen table, the grocery store and the garden. From the grocery store, food is purchased, put in the icebox or pantry, prepared for consumption, and placed on the kitchen table. From the garden, food is planted and picked, then put in the pantry, prepared for consumption, and placed on the kitchen table. Decisions about what to buy or grow will determine what foods enter the channels. The points at which decisions are made within a channel are sections. Once a potato is chosen from either the grocer or garden, the cook must decide how to serve it, what to serve it with, and how much to save for tomorrow. Of all the potato plants grown in a given year, only a few make it through all of the sections of a channel to a dinner table.

Decisions in each section of the channel are guided by forces. Purchasing decisions at the grocer will be affected by positive forces in favor of buying, such as low prices and personal preference, and forces against buying, such as high price or low nutrition. If the total forces in favor of buying the food outweigh the total forces against buying the food, the food will be bought, or the reverse. Consider pork chops. As you stand in the grocery store, deciding whether to buy the pork chops, there is a strong force in favor of buying the pork chops, if you like pork chops, but also a strong force against buying the pork chops, if they are expensive. Lewin's housewife is similarly conflicted.

Once the housewife decides to buy or grow the food, it enters either the grocer or garden channel, and is pushed through the channel by forces that change direction once a decision is made. If the forces in favor of buying expensive pork chops prevail, the forces formerly keeping the pork chops from the table now push them through the channel, because a housewife would not want to waste expensive pork chops. The point at which this force changes direction, from keeping food out of a channel to keeping it in the channel, is a gate (Lewin, 1947b). At this critical point, the character of the force changes. For instance, an elite university admissions board may produce strong forces against admission by admitting only the highest scoring students but, once they are admitted, the university helps the students to graduate in order to keep its matriculation rate high.

The decisions about what items to let in and keep out of a channel are made by "gatekeepers" (Lewin, 1947b, p. 145), such as the housewife and university admission board. Lewin (1947b) argued that social change was the product of forces acting on these gatekeepers. To understand the forces, one must first identify the gatekeepers, then change, or change the mind of the gatekeepers. For Lewin, gatekeeping is always situational, subject to changing circumstances in the present, and ideological, informed by long- term value systems in the past and future (Cartwright, 1951). With this understanding of gatekeeping, Lewin answered his own call to combine the concepts and methods of natural science and economics with social science. The subjective forces acting on the housewife directly impact objective output, such as units of food sold (Lewin, 1947b). Lewin's approach to gatekeeping in particular and social sciences in general was fundamentally practical. Lewin saw himself as a "gatekeeper of civilization" (White, 1950, p. 390) who laid tracks through the vast expanses of human knowledge. Lewin was a careful and thoughtful theoretician who subscribed to the "method of careful approximation" (Cartwright, 1951, p. xiv). For Lewin, methods must always match questions. With this in mind, Lewin proposed the concepts of the gate and gatekeeper in order to inspire the next generation of social scientists.

2. Gatekeeping Theory after Lewin

A. Mr. Gates

After Lewin, the trajectory of gatekeeping theory progressed in the spirit of its creator, rippling outward from individual factors to organizational and external factors, to the entire social milieu in which gatekeeping occurs (Shoemaker & Vos, 2009). Wilbur Schramm, one of the founders of the field of communication studies (White, 1950, p. 383), had written that "no aspect of communication is so impressive as the enormous number of choices and discards which have to be made between the formation of the symbol in the mind of the communicator, and the appearance of a related symbol in the mind of the receiver" (p. 289). A year later, David Manning White became the first to explicitly apply Lewin's gatekeeping theory to the mass media with "one of the first studies of its kind" in Journalism Quarterly (White, 1950, p. 383). With his article, "The 'Gate Keeper': A Case Study in the Selection of News," White tried to understand how one media gatekeeper, called "Mr. Gates," "operates his gate" (White, 1950, p. 383).

Staffed by anchors, reporters, editors, and correspondents, the first gate of the media is the point at which initial decisions about the newsworthiness of events are made (White, 1950). The impact of these decisions is easy to see, said White, by reading two accounts of a controversial policy issue from opposing political perspectives. The media gatekeeper determines not only what events the public knows about, but also how the public thinks about events based on the gatekeeper's own experiences, attitudes, and expectations (p. 384). In a chain of communication from discovery to dissemination, a news item passes through many people who all make "in" and "out" decisions based on their own backgrounds (p. 383). At the end of this chain is Mr. Gates, a middle-aged man who edits a morning paper with a circulation of 30,000 in a Midwestern city of 100,000 citizens (p. 384).

Mr. Gates was responsible for choosing a select few stories from the "mass" of wire copy he received every day (White, 1950, p. 384). The stories that Mr. Gates saw had already made it through several gates manned by reporters, rewriters, and lower-level editors. It was up to Mr. Gates to make the final decision. Over a period of one week, Mr. Gates saved all of the copy that crossed his desk, and wrote justifications on every piece of rejected copy. Mr. Gates rejected 90% of the 12,400 column inches of wire copy he received. Among the reasons given by Mr. Gates for rejecting events were, "BS," "Propaganda," "He's too red," and "Don't care for suicide stories," (p. 386). Also, reasons such as, "No space" and "Would use-if space," were paired with subjective judgments such as "Better story," or "Lead more interesting" (p. 387).

White interpreted these findings as evidence of how "highly subjective, how reliant upon value judgments based on the gatekeeper's own set of experiences, attitudes and expectations the communication of 'news' really is" (p. 387).

White has become synonymous in gatekeeping literature with individual influences on gatekeeping decisions (Shoemaker & Vos, 2009). However, he was also prescient in ascribing to gatekeepers the power to construct social reality: "in his position as 'gatekeeper,' the newspaper editor sees to it (even though he may never be consciously aware of it) that the community shall hear as a fact only those events which the newsman, as the representative of his culture, believes to be true" (White, 1950, p. 390). These two themes run through all subsequent gatekeeping research.

B. Levels of analysis

In the decades after Lewin, gatekeeping research has attended to both individual and cultural factors. There are four levels of analysis in Lewin's field theory: microsystem, or immediate context; mesosystem, or collection of immediate contexts; exosystem, or external institutional standards; and macrosystem, or culture. Gatekeeping occurs in a microsystem (e.g., White, 1950), in a mesosystem as the product of competing interests among news outlets, in an exosystem of journalistic standards and organizational policies, and in a macrosystem of cultural influences. Following Lewin, Shoemaker and Reese (1996) and Shoemaker and Vos (2009) identified five levels of analysis for the study of gatekeeping: individual, communication routines, organizational, social institutions, and social system.

The individual level of analysis concerns the characteristics of individual gatekeepers, or the communicative products of individuals such as blog posts, emails, webpages, statuses, updates, podcasts, etc. The communicative routines level of analysis concerns the practices of a profession embodied in instincts and news values and judgments (e.g., Berkowitz, 1990). For instance, values of timeliness, proximity, and newsworthiness, and the practice of inverted pyramid style of journalistic writing, represent the field of journalism as a whole, not the preferences of an individual journalist. The organizational level of analysis makes some news organizations different from others (Shoemaker & Vos, 2009). The forces affecting gatekeeping decisions at a small rural newspaper will differ from those at large national operations. The social institution level of analysis focuses on forces that act on an organization, such as advertisers, governments, and activist groups. Media outlets may tailor their content to appease one or all of these social institutions. Finally, the social system level of analysis concerns how more abstract forces, such as ideology, culture, economics, and politics, affect the gatekeeping process (Shoemaker & Vos, 2009).

C. Gatekeeping after Mr. Gates

Some researchers have found support for White's (1950) emphasis on the individual and subjective (e.g., Chang & Lee, 1992). Bleske (1991) replicated White's study with a woman, Ms. Gates, to determine the affects of technological development and changing gender norms on individual gatekeeping decisions. Bleske found that Mr. and Ms. Gates assigned the same relative importance to human interest and national and international politics stories. While Bleske's results were similar to White's, they also show the importance of industry standards if a woman with the same job makes the same decisions 50 years later.

More recently, Enli (2007) replicated White's study in the context of the Norweigian current affairs program SevenThirty. In SevenThirty, viewers can respond to content by texting a moderator. If chosen, the text will be displayed on the screen or presented to the program's hosts as discussion prompts. The moderator act as a "boundary gatekeeper" by selecting which text messages to air (Shoemaker & Vos, 2009, p. 113). Over a two month period, only 15% of the nearly 1100 text messages sent to SevenThirty appeared on the show. All messages were edited for length and thematic content. Enli (2007) identified the opposing forces at work in the individual gatekeeping decisions as journalistic norms and participatory ideals. Messages that were too sympathetic or antipathetic were rejected, as were personalized and elitist messages.

For Dimmick (1974), on the other hand, gatekeepers value their sources, colleagues, opinion leaders, reference institutions, and organizational policies over their own intuition. Similarly, Gieber (1956, 1960, 1964), Westley and MacLean (1957), and Chibnall (1975) emphasized the routines and practices of journalism. For Gieber (1956), structural considerations, space limitations and a "straitjacket of mechanical details" (p. 432) limit the decisions a gatekeeper can make. Under an impending deadline, Mr. Gates cannot make subjective judgments about every piece of copy that crosses his desk. Shoemaker and his colleagues also argue that gatekeeping is more influenced by the "routinized practices of news work" than by any personal beliefs of the gatekeeper. Routine forces serve as heuristics, or mental shortcuts, for making gatekeeping decisions under deadlines (Shoemaker, Eichholz, Kim, & Wrigley, 2001, p. 235).

Cassidy (2006) found that routine forces had a greater impact on the gatekeeping decisions of both online and print journalists than individual factors. Cassidy shows that impending deadlines, as well as the online ethos of instantaneity, make White's individual factors less relevant. For Cassidy, immediacy is now a journalistic norm. Decades earlier, McNeilly (1959) and Bass (1969) had emphasized the roles of the journalist in gatekeeping decisions. McNeilly (1959) modeled international news gatekeeping as an "obstacle course" (p. 230) of gatekeepers, from correspondents to copy editors, who all make decisions about whether and how a story should pass through their gate. In McNealy's model, gatekeepers could provide feedback to other gatekeepers, and stories could emerge at different stages of the gatekeeping process independently of the initial story. Bass (1969) divided the gatekeeping process into two levels: news gatherers--the writers, reporters, and local editors who turn raw news into news copy--and the news processers--the editors, copy readers, and translators--who turn news copy into a completed product.

While some (e.g., Halloran, Elliott, & Murdock, 1970) followed Bass (1969) by focusing on raw news gatherers, others (e.g., Chibnall, 1975) argued that news stories are shaped from raw observations. For instance, Westley and MacLean (1957) had modeled gatekeeping as a dynamic process where information moves from sender to receiver through a media channel to an audience. At any time in the process, there are multiple senders sending information through various media channels, each with its own chain of communication and series of gatekeepers. Information can bypass the sender and flow directly to the channel or, if an audience member experiences an event directly, may bypass both sender and channel. The dynamism of Westley and MacLean's model attracted many subsequent gatekeeping theorists.

D. Contemporary gatekeeping models

In Gatekeeping, Pamela Shoemaker (1991) defined gatekeeping as the "process by which the billions of messages that are available in the world get cut down and transformed into the hundreds of messages that reach a given person in a given day (p. 1). With this definition, Shoemaker focused on the actual decisions of gatekeepers. Shoemaker and her colleagues later defined gatekeeping in the spirit of Lewin, as the "overall process through which social reality transmitted by the news media is constructed, and is not just a series of in and out decisions" (Shoemaker, et al., 2001, p. 233). Shoemaker extended gatekeeping beyond micro-level decisions to the whole construction of social reality (Shoemaker & Vos, 2009), to what "exists" for those subjected to a gatekeeping process (Lewin, 1947a, p. 6). Shoemaker et al. (2001) account for the many forces that affect gatekeeping decisions and for the power of the mass media to construct social reality. They answer Lewin's call to extend the limits of the life space.

Shoemaker and Vos (2009) defined gatekeeping as the "process of culling and crafting countless bits of information into the limited number of messages that reach people every day" (p. 1). Shoemaker and Vos echo White (1950) that gatekeeping decisions make many versions of the same material reality (2009, p. 2), and that the news reported by various media outlets is very similar (Shoemaker & Cohen, 2006). White's (1950) chain of communication consists of reporters, rewriters, and wire editors. Shoemaker and Vos' (2009) gatekeeping process begins when a news worker is exposed to an event and ends with selection of the most newsworthy and dissemination as news. However, another gatekeeping process begins when an audience member chooses what to consume. Gatekeeping is not arbitrary or random, but the result of deliberate decisions from exposure to dissemination (Shoemaker & Vos, 2009). News items that make it through all gates draw the "cognitive maps" of news consumers (Ranney, 1983), and set the agenda for what it is important to think about (McCombs & Shaw, 1976).

With this in mind, Shoemaker and Vos (2009) synthesized the extant models of gatekeeping into a model of the gatekeeping field. They argue that the constructs of gates, gatekeepers, forces, and channels are as relevant now as they were for Lewin. In their model, raw information flows through three gatekeeping channels: source, media, and audience. Information enters the source channel through experts, observers, participants, commentators, and interested parties. Information enters the media channel through reporters, editors, production staff, interactive staff, and editorial and marketing assistants. The source and media channels converge as news content. Audience members then take what they want from the field of news content. Information enters the audience channel through Twitter feeds, Facebook posts, smart phone cameras, or any communication technology that records events. Information that is odd or unusual, of personal relevance, or a threat to public well-being is most likely to make it through the audience channel and reach the public (Shoemaker & Vos, 2009). In this model, the audience is a gatekeeper that allows only attention-grabbing information through the channel, and attention-grabbing has replaced newsworthiness as a marker of journalistic credibility (Shoemaker & Vos, 2009).

In this model, messages move through communication organizations such as blogs, newspapers, news outlets, television networks, and public relations agencies. Within each organization, boundary gatekeepers make initial decisions about what information to let into their channel. Once information passes these boundary gatekeepers, it moves to internal gatekeepers, who make decisions based on journalistic routines and standards. The internal gatekeepers then pass information to boundary output gatekeepers, who make final decisions about how to the present information based on feedback from the audience. Gates bracket each gatekeeper, and forces surround each gate. All gatekeepers weigh the influences of organizational socialization, conception of their role, cognitive heuristics, and values, attitudes, and ethics (Shoemaker & Vos, 2009, p. 115). The source, media, and audience channels all exist within a journalistic field.

Shoemaker and Vos (2009) argue that future studies of gatekeeping theory must be attentive to history and transcend their own social system while accounting for globalization. Furthermore, gatekeeping theorists should mix descriptive and interpretive accounts of gatekeeping with quantitative studies. They should follow Shoemaker's (1991) sociological gatekeeping research by employing new statistical techniques to gatekeeping as a single variable in a sociological field. Shoemaker and Vos (2009) call for theorists to account for journalism practice, and for journalists to pay more attention to gatekeeping theory.

E. The gated

Barzilai-Nahon (2009) echoes Shoemaker and Vos (2009) in noting that more attention to the audience is needed. However, Barzilai-Nahon calls for a new construct--the gated--and a new model that accounts for its dynamics. She bases her call on a review of all articles published from 1995 to March 2007 in Communication Research; Information, Communication and Society; Journalism and Mass Communication Quarterly; and New Media and Society. Of the 2800 articles published in these journals during the 12-year period, 98 articles either mentioned gatekeeping or used it as an implicit conceptual foundation. Journalism and Mass Communication Quarterly had the highest percentage of articles containing a gatekeeping concept at 4.4%, followed by New Media and Society at 4.2%, Information, Communication and Society at 1.2%, and Communication Research at 0.6%. The field of communication had the highest total percentage of articles focusing on or referencing gatekeeping of all the disciplines studied. However, most articles only alluded to or cited gatekeeping without analysis (Barzilai-Nahon, 2009).

Barzilai-Nahon (2009) writes that communication scholars, like Shoemaker, see gatekeepers as trained and trusted elites, while information scientists, like Barzilai-Nahon, see gatekeepers as embedded within a larger community. Also, communication scholars tend to focus on individual characteristics of gatekeeper elites, and information scientists on the relationships between individual gatekeepers and larger networks. Within both fields, however, Barzilai-Nahon saw increasing attention to new media technologies in gatekeeping research, but without a compatible theoretical foundation. With its attention to new media technologies, gatekeeping research has shifted from the process of gatekeeping to the capacity of gatekeepers to construct social reality. For instance, Hardin (2005) found that sports editors base gatekeeping decisions on their perceptions of the reading audience, not demographics of the audience itself. Their perceptions are often sexist and serve to perpetuate patriarchy in sports.

F. Gatekeeping on the Web

Both Barzilai-Nahon (2009) and Singer (2006) identified the comparison of old and new media as another theme in gatekeeping research. The Web has expanded the reach of gatekeeping to anyone with an Internet connection. Sources outside the scope of traditional journalism, and without its professional standards, have taken their place alongside the giants of journalism as destinations for news (Singer, 2006). However, journalists maintain that their privileged position as gatekeepers is safe, hardened by a "cultural understanding" that they are the most qualified to gatekeep (Singer, 2006). Nevertheless, journalists have acknowledged the increasingly prominent role of non-journalists or untrained citizen journalists to make decisions about newsworthiness. Working from Shoemaker and Vos' (2009) model of journalistic field, Singer (2014) identifies the audience as secondary gatekeepers who judge the contributions of journalists and other users.

Singer (2014) attempted to characterize the decisions of these secondary gatekeepers and to determine the criteria they use to make decisions about the value and quality of content. A majority of the 138 newspaper websites studied by Singer (2014) allowed users to flag inappropriate comments. A smaller majority of papers allowed users to rate or recommend content. The vast majority of papers allowed readers to rate the merit of content through social networking or social bookmarking tools. With this affordance, users "identify what they see as worthwhile material for their own personal use, communicate that assessment to others, and republish or otherwise disseminate their selected items to a mass audience" (p. 66). They are gatekeepers as White (1950) understood the term.

While communications scholars have theorized extensively about the old and new media dichotomy, information scientists tend to focus on the identity of gatekeepers within new media (Barzilai-Nahon, 2009). Other themes identified by Barzilai-Nahon include the influence of gatekeepers on production of cultural artifacts and portrayal of various social groups, as well as how the gatekeeping process works, and normative questions about who should be gatekeepers. Overall, Barzilai-Nahon calls the period from 1995-2007 a period of stagnation for gatekeeping theory, where traditional gatekeeping theory cannot keep up with changing communication environments. Barzilai-Nahon's work is a response to an interdisciplinary failure to ask important questions about gatekeeping on the assumption that its 50-yearold foundations are firmly in place.

Barzilai-Nahon (2009) sees an unwillingness to rethink the foundations of gatekeeping theory as a failure of theory building. However, Shoemaker's early work, especially Shoemaker, Tankard, and Lasorsa's 2004 book, How to Build Social Science Theories, addresses this very issue. Nevertheless, Barzilai-Nahon (2009) argues that Shoemaker's gatekeeping theory is unfaithful to her own rules for theory building. She argues that current definitions of gatekeeping are too disparate and contradictory to ground a mature and adaptable theory, and that a new theory of gatekeeping is needed for several reasons. First, the Web redefines the roles of "gate," "gatekeeper," and "gated." The traditional model of gatekeeping needs more than tweaking because its conceptual apparatus is no longer applicable. Second, gatekeeping theory has been held back by disciplinary boundaries. An interdisciplinary concept needs a theory that can learn from each discipline rather than using terminology understood only within each discipline. Third, even within the same discipline, there is no shared vocabulary for speaking about the gated. While early gatekeeping models through Shoemaker and Vos (2009) speak of the audience, Barzilai-Nahon (2009) was the first to apply a label to the "gated," those subjected to a gatekeeping process. Barzilai-Nahon also found that the majority of scholars were concerned with editorial decisions made by editorial staffs and journalists about what items were newsworthy enough to be disseminated. Later, the view of gatekeeping as the preservation of culture through construction of social reality also became commonplace.

This is consistent with the work of Shoemaker and Reese (1996), who emphasized the historical and social content of the media. Just as Lewin applied the methods of individual psychology to social phenomena, Shoemaker and Reese (1996) applied the methods of media effects to cultures and media organizations. Their 2014 book, Mediating the Message in the 21st Century, is an updated edition of their influential 1991 and 1996 editions of Mediating the Message. For Shoemaker and Reese (2014), the symbolic environment is made up of messages. These media messages are not an objective mirror of reality, but a co-creative construction of reality. The early editions of Mediating the Message focused on theory building through not only effects of the media on people, but also the influences of individuals, routine practices, media organizations, social institutions, and social systems on the production of the messages that make up the symbolic environment.

From the perspective of media effects, media content acts on individual psychology as an independent variable. From Shoemaker and Reese's (1991, 1996, 2014) perspective, media content is treated as a dependent variable, acted on by a variety of independent variables within a social field. In the 1991 and 1996 versions of Mediating the Message, Shoemaker and Reese model influences of media production as a two-dimensional bullseye, with individual influences in the center, followed by routines, media organizations, social institutions, and social systems in concentric circles (Shoemaker & Reese, 2014). In Mediating the Message in the 21st Century, Shoemaker and Reese (2014) model influences on the media as a three-dimensional wedding cake, with individual newlyweds on top, and social systems as the bottom and biggest layer of cake. While the early versions focused on individual influences, the new model can be approached in two ways. From the perspective of the individual, the lower layers empower individuals to succeed. From a media sociological perspective, the individual is perched precariously atop a supporting structure that, if destroyed, will bring down the whole cake. From either perspective, each layer of the cake part of a larger social field. By three-dimensionalizing their model, Shoemaker and Reese (2014) answer their own call for a return to Lewin, and Barzilai-Nahon's (2009) call for more thorough explication of the theory's foundation.

However, neither Shoemaker and Reese (2014) nor Shoemaker and Vos (2009) cite Barzilai-Nahon, and Barzilai-Nahon (2009) cites Shoemaker only in passing. Both agree, however, that gatekeeping decisions have shifted from "in" and "out" (Lewin, 1947b, p. 145) decisions to "more or less" rules of presentation (Shoemaker & Reese, 2014). On the Web, for instance, news portals like Google News employ a set of rules (Lewin, 1947b) to collect news articles from various outlets and display them on a single interface. Algorithms organize the articles into topics (e.g., World News, Local, Sports, Science) and rank them according to recency, source credibility, and newsworthiness (Kalyanaraman & Sundar, 2008). In this context, the people in a "key position" to make gatekeeping decisions are the mathematicians who manipulate the algorithms. The experiences of the user replace the experiences of the news worker as the determining factor in the gatekeeping "decisions" of a nonhuman news portal: "The Web is literally a web woven collectively by all citizens on the Internet, resulting in a massive amount of information being disseminated by both professional gatekeepers and laypersons. For casual users interested in efficiently obtaining news and information on the net, this proves burdensome because they now have the arduous additional task of sifting through information of unknown pedigree and determing its veridicality instead of simply attending to news of established credibility" (Kalyanaraman & Sundar, 2008, p. 239). The audience member now makes decisions about the credibility of information and sources, but the algorithm is the gatekeeper as White (1950) understood the term.

Barzilai-Nahon (2009) argues that these new forms of gatekeeping call for a new model. The space for information is finite, making it "necessary to have established mechanisms which police these gates and select events to be reported according to specific criteria of newsworthiness" (Bruns, 2003, p. 1). However, in social and information networks that serve as both hosts and conduits of information, there is no lack of space and few established mechanisms to police ambiguous gates. The mass of information in this "contextual vacuum" (Sundar & Nass, 2001, p. 57) necessitates some form of gatekeeping, either individual or institutional. For Lewin, the housewife is in a position to be a gatekeeper, but it is the social scientist with the power to make social change by studying her gatekeeping decisions. This is no longer the case. Laypersons have the power to change the social world (Shoemaker & Vos, 2009). Whereas Lewin looked backward to the steps leading to war, today's social scientist tracks the organization of a revolution in real time. Barzilai-Nahon's (2009) empowers the researcher to reclaim the position of gatekeeper.

3. Network Gatekeeping Theory

To outline the theoretical foundation of and terms associated with network gatekeeping, this review relies on Barzilai-Nahon (2008). In this article, among the most thorough articulations of her theory, Barzilai-Nahon proposes a new conceptual framework for gatekeeping. Rather than simply selecting which news items should pass through a gate, network gatekeepers aim to: (1) "lock-in" the gated within gatekeeper's network, (2) isolate the gated within a network to protect their norms and information, and (3) allow for an uninterrupted flow of information within "network boundaries" (p. 1496). With these reoriented outcomes in mind, Barzilai-Nahon then defines the conceptual infrastructure of network gatekeeping consisting of the gate, the gated, gatekeeping mechanism, network gatekeeper, and gatekeeping.

A. Definitions

The gate is the "entrance to or exit from a network" (Barzilai-Nahon, 2008, p. 1496). Although the gate is fluid and dynamic within social and information networks, it most often rests at the point at which the gated enter or exit the network. The gated, then, is simply "the entity subject to gatekeeping" (p. 1496). In social and information networks, gatekeeping need not be forced or imposed on the gated. Because they likely have other options across the Web, the gated may willingly agree to the terms of gatekeeping in order to enter a network depending on their relationship with the gatekeepers. Several factors affect these relationships among gatekeepers and the gated. First, political power dictates the extent to which gatekeepers can control the gated. The dynamics of information control often break down according to the political interests of involved parties. Second, the capability of the gated to produce information dictates the dynamics of network gatekeeping. While communicative technologies allow anyone to not only create but also disseminate their own content, access to and mastery of these devices will vary. The platforms on which users can disseminate their content will also vary, affecting relationships among gatekeepers and the gated. Third, relationships among gatekeepers and gated will determine the level of gatekeeping present, with more direct and reciprocal ties resulting in less gatekeeping and more indirect and uni-directional ties leading to more gatekeeping. Fourth, the existence of alternative sources of information changes the gated according to the makeup of the gatekeeping mechanism.

A gatekeeping mechanism is a "tool, technology, or methodology used to carry out the process of gatekeeping" (Barzilai-Nahon, 2008, p. 1496) in any of 10 ways. First, channeling mechanisms such as search engines, directories, categorizations, and hyperlinks direct and attract the gated to various other networks. Second, censorship mechanisms delete or exclude undesired information or users from an existing network. Third, internationalization mechanisms tailor and translate information to local customs and conventions. Fourth, security mechanisms manage the authenticity of and access to confidential or sensitive information. Fifth, cost-effect mechanisms assign values to entering and exiting a network and using the information it provides. Sixth, value-adding mechanisms allow users to customize and contextualize their information on the network. Seventh, infrastructural mechanisms control access to the network at algorithmic and infrastructural levels. Eighth, user interaction mechanisms govern a network's level of interactivity, modality, and navigability (e.g., Sundar, 2008). Ninth, editorial mechanisms govern content decisions in much the same way as traditional gatekeeping theory and, finally, regulation meta-mechanisms at state, national, or governmental levels may overrule any of the other mechanisms, depending on the makeup of its network gatekeepers (Barzilai-Nahon, 2008).

A network gatekeeper is an "an entity (people, organization, or government) that has the discretion to exercise gatekeeping through a gatekeeping mechanism" (Barzilai-Nahon, 2008, p. 1497). Barzilai-Nahon identifies two dimensions of network gatekeepers: an authority dimension, which classifies network gatekeepers according to their level of authority over the gated, and a functional dimension, which organizes network gatekeepers according to the level of control they exercise over the gatekeeping mechanism. Within the authority dimension are governmental, industry regulator, internal authority, and individual levels. At the governmental level, network gatekeepers are governments. Different types of governments will use gatekeeping mechanisms in different ways, with non-democratic states more likely to use gatekeeping mechanisms to limit access to information (e.g., age-limits on pornographic content). At the industry regulator level, public or private entities can establish and enforce gatekeeping mechanisms within a given industry, either in collaboration with or independent of governmental regulation (e.g., cable television controls). At the internal authority level, an organization itself exercises control of the gatekeeping mechanism (e.g., Facebook privacy controls). At the individual level, individuals monitor their own or their families access to information (e.g., parents limit their children's access to television or Internet content).

Within the functional dimension, network gatekeepers can be infrastructure providers, authority site properties, or network administrators. First, infrastructure providers, including network, Internet, and carrier service companies, determine the speed and flow of information passing through a network. Second, authority site properties and their search, portal, or content providers act as gatekeepers by controlling which information Internet users see first or most often. Lastly, network gatekeepers may be designated network administrators or content moderators (e.g., newspaper employee tasked with regulating an online message board), or also individuals who play a network gatekeeping role (e.g., YouTube users who flag inappropriate material)

Finally, network gatekeeping is the "process of controlling information as it moves through a gate" (p. 1496) through not only selection of news, but also addition, withholding, display, channeling, shaping, manipulation, timing, localization, integration, disregard, and deletion of information. Three of these capacities--channeling manipulation, and repetition--are of particular importance for conceptualizing network gatekeeping on social networks such as Facebook, Digg, and Twitter, platforms to which network gatekeeping applies (Barzilai-Nahon, 2008, p. 1497) but has not yet been considered.

B. Gatekeeping on social networks

A 2010 report by the Pew Research Center revealed that 75% of people who read news online get it through social networks. On social networks, users can participate in the gatekeeping process by offering feedback and comments on a particular selection, even if they do not post content themselves, and by forwarding, sharing, and posting links to news stories. Also, traditional news outlets may be only the first or, as is often the case among Digg.com users, the last link in a chain of sources and hyperlinks where news is not only selected by editors, but also funneled through the Web, where it is "amplified, sustained, and potentially morphed as it is re-circulated, reworked, and reframed by online networks" (Goode, 2009, p. 1293) in several ways, identified by Flew and Wilson (2008) as content work, networking, community work, and technical work.

First, although its goals differ from traditional journalism, the content work of network gatekeepers resembles traditional gatekeeping in that network gatekeepers edit, create, and disseminate content that conforms to journalistic standards and norms. Second, networking establishes relationships with other users and outlets to build a close-knit group of connections. Drawing attention to obscure news sites by book-marking their articles on sites like Digg and forwarding or retweeting links to news articles on Facebook or Twitter can also be considered networking (Flew & Wilson, 2008). Third, community work includes skills such as registering on a site, creating a profile, and posting content. Similarly, technical work consists of tasks related to the technological affordances of a particular medium which, as Barzilai-Nahon (2008) contends, exist within a horizon of gatekeeping power.

To "explore what new modes of gatekeeping power may be emerging" (Goode, 2009, p. 1295), this review article now explicates network gatekeeping on Digg.com, Facebook, and Twitter, three social networks that represent network gatekeeping's capacity for manipulating, channeling, and repeating information.

C. Network gatekeeping on Digg

In the context of network gatekeeping, Barzilai-Nahon (2008) defines manipulation as "changing information by artful or unfair means to serve the gatekeeper's purpose" (p. 1497). When users of the news aggregator Digg.com submit news articles to the site, they can be either rewarded, if the submission is highly "dugg" and promoted to the front page, or punished, if the submission receives little attention. Although the site advertises itself as an editor-free "place where people can collectively determine the value of content" (Digg, 2010), users perform their own network gatekeeping by manipulating the aesthetics of news articles, as well as digging, burying, sharing, or commenting on others' submissions. Digg may be free of editors as conceptualized by traditional gatekeeping theory, but its users' success is dependent on their network gatekeeping. With this in mind, this section first outlines the uses of Digg and, second, further explicates the dynamics network gatekeeping on the site.

All content on Digg is submitted to a community of registered users. Users then "digg" articles that interest them. If a given submission receives enough diggs, it is promoted to the site's front page along with a marker of the user who initially submitted it. Because a front page story on Digg.com can result in an increase of at least 12-15,000 visitors to the site of the news outlet that produced the story (Cohn, 2007), Digg.com buttons have become ubiquitous in online news and social networking environments. Since developing a custom widget that ranks the top five most dug stories on its website, Time Magazine's presence on Digg.com has risen more than twofold, and its Digg-driven clicks increased from 500,000 to 1.3 million (Shields, 2009). The site has caused similar traffic increases for Newsweek.com and Wired.com (Shields, 2009), making it an invaluable resource for advertisers targeting Digg.com's tech-savvy audience in search of customizable options.

The site features a number of customizable options, including a choice between seeing the most recent content or the top content from the last one, seven, 30, or 365 days. In the most recent option, articles are ranked according to the recency with which they were made popular or received enough diggs to appear on the homepage. In the top content option, articles are ranked according to their number of diggs. In either option, articles are accompanied by a number of comments, choices to share or bury, the news outlet that produced the article, and the username of the Digg member who submitted it. While clicking on the share icon allows users to share a link to the article by email, Facebook, or Twitter, users can also initiate and respond to comments on the article, either digg or bury each comment, and sort comments by oldest, newest, most controversial, and most dugg. Clicking on the username of the user who submitted the article leads to that user's profile, which features statistics such as number of diggs, submissions, and comments--their network gatekeeping scorecard.

Halavais (2009) argues that this scorecard encourages further participation on the site (p. 445). By sampling 30,000 of Digg's 2.8 million users and downloading all of the comments they made on the site and the total number of diggs and buries each comment received, Halavais found that comments by experienced users were generally positively correlated with both diggs and buries. This indicates that on Digg those with experience are more likely to receive a reaction, either positive or negative. However, while most users strove for positive feedback and reinforcement through a large number of diggs and small amount of buries, some, through racial and religious slurs, insults, and profanity, sought to become as little liked by other users as possible. He suggests that the rewards that encourage participation on the site also enforce a "process that trains users to behave in ways that conform to community standards and expectations" (Halavais, 2009, p. 457). In other words, in the absence of traditional gatekeeping standards, network gatekeepers develop their own.

While comments containing the word "liar" were likely to be buried, especially when used in reference to another Digg user, comments by users who supported their arguments with credible sources were likely to be dugg, despite the site's pride in operating without editorial authority (Halavais, 2009). Because the word "Digg" itself was associated with editorial authority, articles with "Digg" in them were likely to be buried, along with criticisms of spelling or grammar, two common editorial tasks (Halavais, 2009). Although the level at which these processes occur is unclear, in the context of network gatekeeping they represent an abstraction of the relationship between gatekeepers and the gated in traditional gatekeeping theory, with Digg users relying on the wisdom of others to become good editors, then using that same wisdom to perpetuate their own editorial influence. Despite its claim to egalitarian editorship, users of Digg have various levels of confidence in their fellow gatekeepers, much like traditional gatekeeping theory would predict. However, whereas in traditional gatekeeping theory editors can be condescending toward their readers and lack confidence in readers' gatekeeping ability (Gladney, 1996), online networks such as Digg reverse this relationship by allowing the audience to determine the efficacy of each gatekeeper, an arrangement addressed in two controversial redesigns.

In August 2010 and July 2012, the site underwent significant redesigns, which made it aesthetically similar to Facebook and functionally comparable to Twitter. Although the front page of the site has arguably been made more credible by the infusion of news items from traditional sources such as the New York Times, Washington Post, BBC, and CNN, Digg's most ardent users were unhappy with is the perceived infiltration of editorial authority from a select number of publishers (Bohn, 2012). Also, the bury button, which allowed users to give submissions a low rating, was removed. Nevertheless, gatekeeping on the site remains fundamentally the same, and understanding the dynamics of network gatekeeping by Digg users will shed additional light on the complex and multi-layered motivations of network gatekeepers.

Whether they exhibit a similar condescension toward other users as editors show toward readers in print media, and whether these perceptions have changed with the design of the site, will be a crucial first step in determining if sites like Digg.com are polarizing or uniting network news consumers. Digg's popularity, at least in part, has been attributed to the democratic ideals of equality and egalitarianism (e.g., Hargittai, 2000) but, because "there seems to be prima facie evidence of a powerful core of 'elite' at work" on the site, the relationships among its users has been termed an artistocracy, a popularity contest, and a Digg mafia (Goode, 2009, p. 11). Contrastingly, Digg users have argued that the most popular users earn their status through skill and hard work (Goode, 2009) or, from the perspective of network gatekeeping, through their effectiveness as gatekeepers. While this effectiveness is a function of artful manipulation of information, users may also perform a network gatekeeping function by channeling news through a social network, one of the least understood motivational mechanisms of Digg (Halavais, 2009) and the hallmark of Twitter.

D. Network gatekeeping on Twitter

For Barzilai-Nahon (2008), channeling refers to "conveying or directing information into or through a channel" (p. 1497). Introduced in 2006, Twitter is a microblogging service that has become a source of immediate, instantaneous news. It allows users to act as network gatekeepers by channeling news through the site. This environment provides an ideal venue for examining the gatekeeping decisions of both followers and the followed, or gatekeepers and the gated. This section will, first, review the uses and influence of Twitter and, second, discuss several studies with implications for network gatekeeping.

Any user of Twitter can follow or be followed by any other without any necessary interaction or mutual approval. Followers receive all tweets from those they follow, which appear on the user's profile chronologically. There is a well-defined language on the site which promotes brevity and conciseness within a 140 character limit. "Retweeting," or forwarding the tweets of other users without their knowledge and beyond their scope, has become a popular means of disseminating news items (Cha, Haddadi, Benevenuto, & Gummadi, 2010) and reinforcing a message (Watts & Dodd, 2007). Much like the sharing function on Facebook, links to stories or tweets themselves can be retweeted independently of their originator in real time. Although those looking to Twitter for news may miss newsworthy items among the "other chatter going on," the biggest advantage of the site is this element of instantaneity (Weinberg, 2008, para. 3).

In one of the first studies to explore the implications of instantaneity in the entire "Twitter-sphere," Kwak, Lee, Park, and Moon (2010) examined 41.7 million profiles, 1.47 billion social interactions, and 106 million tweets to study, among many other variables, the distributions of reciprocity between followers and the followed, or gatekeepers and the gated. They found that most tweets are not reciprocated, but there is some evidence of homophily among users. Similarly, Cha and colleagues (2010) identified three types of a related concept--influence--that Twitter users may attain. First, in-degree influence refers simply to the number of followers a user has. It is a straightforward and overt marker of that user's known audience (Cha et al., 2010). Second, retweet influence is the number of retweets that bear a particular user's name. It is a more subtle way of tracking users' influence outside of their network of followers, and measures their ability to produce content likely to be enjoyed by a large number of users. Third, mention influence is measured by the number of times a user's name is included in a tweet or retweet, and indicates the "ability of that user to engage others in a conversation" (p. 3). All of these types of influence are played out in network gatekeeping decisions, and largely determine the makeup of gatekeepers and the gated.

More recently, Xu and Feng (2014) examine conversations between traditional journalists and Twitter users to determine the identity of gatekeepers and the political power of the gated in terms of online connectivity and political and issue involvement. They found that politically active Twitter users reached out most often to journalists with similar political leanings. Also, most of the interactions between journalists and citizens on Twitter occurred more than once, but were most often initiated and retweeted by citizens. Xu and Feng see network gatekeeping as inclusive and empowering of average citizens who may not have had the opportunity to interact with journalists, even though they reach out most often to those they agree with. The inclusiveness and openness of social media may expose citizens to new viewpoints but may also harden their existing opinions.

Leavitt, Burchard, Fisher, and Gilbert (2009) engaged this makeup by measuring the influence of 12 of the most popular Twitter users over a 10-day period. They found that while celebrities were mentioned more often, news outlets were more influential in getting their information retweeted. Although Weng, Yao, Leonardi, and Lee (2010) found high levels of reciprocity in a nonrandom sample of nearly 7,000 Twitter users, Cha et al. (2010) found only 10% reciprocity in a random sample of users. These mixed results concerning levels of reciprocity have implications for network gatekeeping because, as Barzilai-Nahon's (2008) theory posits, one of the advantages of social networks like Twitter is the ability of the gated, in this case average Twitter users, to interact with gatekeepers, the 12 most popular Twitter users. Low reciprocity on Twitter would indicate a hierarchical model of gatekeeping, not the horizontal model proposed by Barzilai-Nahon (2005).

Cha et al. (2010) provided an answer to this inconsistency by examining which particular network gatekeeping activities result in the most influence in what topic and at what time. They found that while news sites, politicians, athletes, and celebrities were highest in in-degree influence, news sites, content aggregators, and business sites were highest in re-tweet influence and, for the most part, celebrities were highest in mention influence. Also, because fewer than 30% of "mentions" contained links to original sources, mentions are identity-driven, and retweets more content-driven because they almost always contained a link to the original source (Cha et al., 2010). Cha and colleagues also found that the most influential users of Twitter were public figures, websites, and content aggregators, and that there was little overlap between the three types of influence.

However, ordinary users can also become influential gatekeepers. In the same study, 20 of the most followed users who discussed a single news topic were examined. Although unknown prior to the news topic they discussed, users who tweeted consistently about one topic increased their influence scores the most over the course of a particular event. This implies that users can become more influential gatekeepers by focusing on a single topic and tweeting detailed, insightful things about that topic rather than merely conversing with other users. This finding was confirmed by Huberman, Romero and Wu (2008), who examined over 300,000 Twitter users and discovered that, although the number of followers did increase with the number of posts, those users with many posts do not necessarily have many followers, making number of friends a more indicative marker of influence than number of followers.

For the purposes of network gatekeeping, these results show that two users who are linked on Twitter need not be interacting. Research that has looked only at traffic on the site without measuring influence has found that the top 10% of Twitter users post over 90% of total tweets on the site (Cheng & Evans, 2009), prompting Goode (2009) to posit that "social networks are not flat; they are hierarchical; and they are not as conversational as we often assume" (p. 1293). However, how gatekeepers interact with the gated--and who exactly plays these roles--has not been considered in the context of news sharing and network gatekeeping. This review essay now turns to Facebook, the world's largest social network (Kaplan & Haenlein, 2010), to lay out directions for future research in network gatekeeping theory.

E. Network gatekeeping on Facebook

Facebook is a social network that allows users to post pictures, comments, and status updates visible to self-chosen Facebook friends. Facebook users make choices about what information to add, withhold, and disregard, and how to shape, localize, and manipulate the information they channel through their profile. While much research has addressed motivations for using social networking sites and personal web pages (e.g., Banczyk, Kramer & Senokozlieva, 2008; Papacharissi, 2002), little has specifically analyzed Facebook in the context of network gatekeeping. Some of the most applicable work to network gatekeeping theory has been done in the context of online news sharing, and the heuristics that motivate news sharing and consumption on Facebook.

Within a given network of Facebook friends, there exist a small number of users who consistently share links to news stories. When online news consumers go to Facebook for their news, these users are performing a gatekeeping function. The other members of a given network can easily go elsewhere for their news but, if they consistently follow one friend's links to news stories, that friend is a network gatekeeper, either voluntarily or unknowingly. Facebook users voluntarily include personal information on their profiles as a function of their trust in the site. Many teenagers are willing to sacrifice privacy for constant connectivity and are more likely to give personal information to a perceived "friend" online even if the friend is fake. More recently, Facebook users have adapted to the public nature of the Internet with a reluctance to share private information on their public profiles.

Facebook users make judgments about to friendships and connections based on bandwagon heuristics and authority heuristics (Sundar, 2008). In the context of online news sharing, the bandwagon heuristic posits that, "if others think that this is a good story, then I should think so too" (Sundar, 2008, p. 83). Much e-commerce research has shown the power of the "bandwagon effect," whereby products that are recommended by a large number of users are more likely to be purchased than those with no or a small number of recommendations. For instance, the more and more positive reviews a book has, the higher the sales (Chevalier & Mayzlin, 2006). Essentially, ecommerce websites recognize the need not only to sell books to consumers, but to "enable users to sell them to each other" interpersonally (Sundar, Xu, & Oeldorf-Hirsch, 2009, p. 3457). Many online consumers recognize this profusion of bandwagon cues and consider a site without them incredible or unrepresentative (Sundar, Xu, & Oeldorf-Hirsch, 2009). In the context of Facebook, the number of "likes" on a profile will influence perceptions of the profile owner based on the bandwagon effect.

When the New York Times website displays the day's most e-mailed, searched and blogged articles, and the Washington Post features a Facebook application displaying friends' activity, readers assign agency to a mass of other users and trigger the bandwagon heuristic. In the context of online news, readers may make judgments of the quality and credibility of articles, and the people and issues they are written about, under the blind direction of other anonymous users conveyed through interface cues. Sundar, Knobloch-Westerwich, & Hastall (2007) explored the effect of three such cues: the source of the article, number of related articles, and how recently it was posted, and found that although the source of an article was not considered credible, it was nevertheless rated as credible and newsworthy when associated with a large number of related articles, indicating the influence of a bandwagon heuristic. Similarly, in a study of an online news portal, Sundar and Nass (2001) found that users were more likely to choose and spend more time reading articles that had been strongly recommended by many other users. Likewise on Facebook, where users with lots of friends are seen as authorities.

As defined by Sundar, Oeldorf-Hirsch, and Xu (2008), the authority heuristic posits that "experts' statements can be trusted" (p. 3455). In both face-to-face and technologically-mediated communication, deference to an authority figure "is likely to directly confer importance, believability, and pedigree to the content provided by that source and thereby positively impact its credibility" (Sundar, 2008, p. 84). Even in the context of online news aggregators and portals, each article is accompanied by the news outlet that produced it, allowing Facebook users to make credibility judgments about other users. On Facebook, authority heuristics often compete with bandwagon heuristics. If heuristics are influencing the perceptions of Facebook users, then cues that trigger both the authority and bandwagon heuristics should "directly impact user perceptions of message credibility" (Sundar, Knobloch-Westerwick, & Hastall, 2007) By juxtaposing the two heuristics, Sundar, Oeldorf-Hirsch, and Xu (2009) found that both are psychologically relevant, but bandwagon cues are generally more persuasive than authority cues, but only when consistent.

The distinction between authority and bandwagon heuristics has implications for content-sharing on Facebook. By sharing content on Facebook produced outside Facebook, Facebook users blur the line between editor and user, bandwagon and authority. For instance, The Washington Post, one of the most credible and recognizable American newspapers, has its masthead in the upper left-hand side of its online interface. On the right side is a Facebook Network News application, which allows users to view either the most popular stories of the day accompanied by the number of people who have shared them or a summary of their Facebook friends' news-viewing activity. By logging into Washingtonpost.com using Facebook Connect, users can share, like, and comment on content, as well as see all the content their friends have shared, liked, or commented on. Also, users can read content recommended by their network, see what Washington Post content is most popular across Facebook, and keep a profile page showing the content with which their Facebook friends have interacted. Depending on whether Facebook profile owners psychologically consider themselves editors or part of a community of users, juxtaposing these two ontologically distinct editorial roles may not only allow the Washington Post to become much more social, as the site posits, but also much less credible.

4. Conclusion

Gatekeeping theory began with individual gatekeepers and rippled outward to organizational and institutional routines, to the entire social field in which gatekeeping occurs. Network gatekeeping theory, too, must attend to the distinction between individual gatekeepers and network gatekeeping. A social networking site like Facebook, if users are considered sources of content themselves, could be considered individual gatekeeping but, if users are conceptualized as part of a community of users interacting with other profile owners, it would be considered collective gatekeeping. These distinctions will determine the nature of relationships among the gated and gatekeeper, dictate the capacities for gatekeeping in various gatekeeping mechanisms, and direct the motivations and practices of network gatekeepers. For instance, within a given network of Facebook friends, there exist small number of users who consistently share links to news stories. When online news consumers go to Facebook for their news, this small number of users is performing a gatekeeping function. Although the other members of that network can easily go elsewhere for their news--they have alternatives, in terms of network gatekeeping--if they consistently follow a prominent news sharer's links and psychologically consider that user a news source (see Sundar & Nass, 2001), the user is a network gatekeeper, and a very powerful one.

Equally ambiguously, when "individual users control information on their social networking site (SNS) profiles," Facebook users may engage individual gatekeeping but, because SNS "offers wall posts and other interactions between profile owners and their social networks," (Hu & Sundar, 2010, p. 105), they may also be gatekeeping collectively. Little research to this point has addressed how, if Facebook users are considered editors of their own content, their relationship with friends in their network relates to the condescending and hierarchical relationship between editors and audiences in traditional journalism (Gladney, 1996) or, as more people use social networks as news sources (Purcell, Rainie, Mitchell, Rosenstiel, & Olmstead, 2010), whether Facebook users' selection of content on their Facebook profiles begins to resemble their selective consumption of news. A similarity would not only complicate Barzilai-Nahon's (2005, 2008, 2009) network gatekeeping theory, but also have implications for how heavy Facebook users interact in real life, given Facebook's purported ability to alleviate the tedium of face-to-face communication (Ellison, Steinfield, & Lampe, 2011).

Because the popularity of user-generated news sites is, at least in large part, attributable to the "democratic ideals of equality, accountability, transparency, and empiricism" (Keegan & Gergle, 2010, p. 134), it is important to know if users of these collective sites abide by the same democratic principles when making individual decisions about the quality, credibility, and representativeness of online news and, more fundamentally, the gatekeeping ability of fellow online news consumers. Answers to these questions about collective vs. individual gatekeeping on Digg, Facebook, and Twitter will ultimately reveal whether these technologies merely indicate a shift in gatekeeping practices on the Web or signify a more fundamental and consequential transformation of the way news in produced and consumed in a digital environment.

The Web poses paradigmatic challenges not only to news production and consumption, but also to traditional understanding of gatekeeping theory. Barzilai-Nahon's network gatekeeping theory responds to the challenge by rebuilding the infrastructure of gatekeeping theory through the gate, gated, gatekeeping, network gatekeeper, and gatekeeping mechanism. Network gatekeeping theory extends traditional gatekeeping theory beyond selection of news to addition, withholding, display, channeling, shaping, manipulation, timing, localization, integration, disregard, and deletion of information. Social networks like Digg, Twitter, and Facebook allow for more open and diverse exchange of information. At the same time, with no trained editors, the sites may more closely resemble a supermarket tabloid than a social network. Whatever the outcome, this review of gatekeeping in general and of Barzilai-Nahon's network gatekeeping theory in particular hopes to provide a first step towards a holistic understanding of network gatekeeping, one that allows researchers to keep up with the ever-changing online news landscape and better equips communication practitioners to map the trajectory of information on the Web.

References

Ayer, A.J. (1966). Logical positivism. New York, NY: Free Press.

Banczyk, B., Kramer, N. C., & Senokozlieva, M. (2008). "The wurst" meets "fatless" in MySpace. The relationship between personality, nationality, and self-presentation in an online community. Paper presented at the 58th annual meeting of the International Communication Association, Montreal, Canada.

Barzilai-Nahon K. (2005). Network gatekeeping theory. In K. Fisher, S. Erdelez, & E. McKechnie (Eds.), Theories of information behavior: A researcher's guide (pp. 247-254). Medford, NJ: Information Today.

Barzilai-Nahon, K. (2008). Toward a theory of network gatekeeping: A framework for exploring information control. Journal of the American Society for Information Science and Technology, 59, 1493-1512.

Barzilai-Nahon K. (2009). Gatekeeping: A critical review. Annual Review of Information Science and Technology, 43, 433-478.

Bass, A. (1969). Redefining the gatekeeper concept: A U.N. radio case study. Journalism Quarterly, 46, 59-72.

Berkowitz, D. (1990). Refining the gatekeeping metaphor for local television news. Journal of Broadcasting & Electronic Media, 36(1), 45-60.

Bleske, G. L. (1991). Ms. Gates takes over: An updated version of a 1949 case study. Newspaper Research Journal, 12(4), 88-97.

Bohn, D. (2012, July 31). Digg launches redesign and new app with editorial curation, no comments. Verge. Retrieved October 31, 2013, from http://www.the-verge.com/2012/7/31/3207670/digg-redesign-live

Brown, R. M. (1979). The gatekeeper reassessed: A return to Lewin. Journalism Quarterly, 59 (3), 595-601.

Bruns, A. (2003). Gatewatching, not gatekeeping: Collaborative online news. Media International Australia Incorporating Culture and Policy: Quarterly Journal of Media Research and Resources, 107, 31-44.

Cartwright, D. (1951). Foreword. In D. Cartwright (Ed.), Field theory in social science: Selected theoretical papers (pp. vii-xx). New York: Harper.

Cassidy, W. (2006). Gatekeeping similar for online, print journalists. Newspaper Research Journal, 27(2), 6-23.

Cassirer, E. (1923). Philosophie der symbolischen Formen. Erster Teil: Die Sprache. Berlin: Bruno Cassirer. (Translated as The Philosophy of Symbolic Forms. Volume One: Language. New Haven: Yale University Press, 1955.)

Cassirer, E. (1925). Philosophie der symbolischen Formen. Zweiter Teil: Das mythische Denken. Berlin: Bruno Cassirer. (Translated as The Philosophy of Symbolic Forms. Volume Two: Mythical Thought. New Haven: Yale University Press, 1955.)

Cassirer, E. (1929). Philosophie der symbolischen Formen. Dritter Teil: Phanomenologie der Erkenntnis. Berlin: Bruno Cassirer. (Translated as The Philosophy of Symbolic Forms. Volume Three: The Phenomenology of Knowledge. New Haven: Yale University Press, 1957.)

Cha, M., Haddadi, H., Benevenuto, F., & Gummadi, K. (2010). Measuring user influence in Twitter: The million follower fallacy. Proceedings of the International Conference on Weblogs and Social Media, Washington, D.C. Retrieved January 14, 2015 from http://twitter.mpi-sws.org/icwsm2010_fallacy.pdf.

Chang, T., & Lee, J. (1992). Factors affecting gatekeepers' selection of foreign news: A national survey of newspaper editors. Journalism Quarterly, 69(3), 554-561.

Cheng, C., & Evans, T. (2009, June). An in-depth look inside the Twitter world. Sysomos. Retrieved October 31, 2013, from http://www.sysomos.com/docs/Inside-Twitter-By-Sysomos.pdf.

Chevalier, J., & Mayzlin, D. (2006). The effect of word of mouth on sales: Online book reviews. Journal of Marketing Research, 43, 345-354.

Chibnall, S. (1975). The crime reporter: A study in the production of commercial knowledge. Sociology, 9(1), 49-66.

Cohn, D. (2007). Digg this. Columbia Journalism Review, 45, 8-9.

Digg (2010). What we do. Retrieved January 14, 2015 from http://www.jrwhipple.com/findit/%28find4.html.

Dimmick, J. (1974). The gate-keeper: An uncertainty theory. Journalism Monographs, 37.

Ellison, N., Steinfield, C., & Lampe, C. (2011). Connection strategies: Social capital implications of Facebook-enabled communication practices. New Media & Society, 13, 873-892.

Enli, S. G. (2007). Gatekeeping in the new media age: A case study of the selection of text-messages in a current affairs programme. Javnost--The Public, 14(2), 47-62.

Flew, T., & Wilson, J. (2008). Citizen journalism and political participation: The Youdecide 2007 project and the 2007 Australian Federal Election. Australian Journal of Communication, 35(2), 17-37.

Gieber, W. (1956) Across the desk: A study of 16 telegraph editors. Journalism Quarterly, 33(4), 423-432.

Gieber, W. (1960). How the 'gatekeepers' view local civil liberties news. Journalism Quarterly, 37(1), 199-205.

Gieber, W. (1964). News is what newspapermen make it. In L. A. Dexter & D. M. White (Eds.), People, society, and mass communication (pp. 172-182). New York, NY: Free Press.

Gladney, G. (1996). How editors and readers rank and rate the importance of 18 traditional standards of newspaper excellence. Journalism & Mass Communication Quarterly, 73, 319-331.

Goode, L. (2009). Social news, citizen journalism, and democracy. New Media & Society, 11, 1287-1305.

Halavais, A. (2009). Do dugg diggers digg diligently? Information, Communication, & Society, 12, 444-459.

Halloran, J., Elliott, P., & Murdock, G. (1970). Demonstrations and communication: A case study. Baltimore: Penguin.

Hardin, M. (2005). Stopped at the gate: Women's sports, 'reader interest,' and decision-making by editors. Journalism & Mass Communication Quarterly, 82(1), 62-77.

Hargittai, E. (2000). Open portals or closed gates: Channeling content on the world wide web. Poetics: Journal of Empirical Research on Culture, the Media, and the Arts, 27, 233-254.

Hogg, T., & Lerman, K. (2012). Social dynamics of Digg. EPJ Data Science, 1(5). Retrieved January 9, 2015 from http://www.epjdatascience.com/content/1/1/5.

Hu, Y., & Sundar, S. S. (2010). Effects of online health sources on credibility and behavioral intentions. Communication Research, 37 (1), 105-132.

Huberman, B., Romero, D., and Wu, F. (2008). Social networks that matter: Twitter under the microscope. First Monday, 14(1).

Kalyanaraman, S., & Sundar, S. S. (2008). Portrait of the portal as a metaphor: Explicating Web portals for communication research. Journalism & Mass Communication Quarterly, 85(2), 239-256.

Kaplan, A., & Haenlein, M. (2010). Users of the world, unite! The challenges and opportunities of social media. Business Horizons, 53, 59-68.

Keegan, B., & Gergle, D. (2010). Egalitarians at the gate: One-sided gatekeeping practices in social media. Proceedings of the 2010 ACM conference on Computer Supported Cooperative Work (pp. 131-134). Savannah, GA. Retrieved January 14, 2015 from http://research.microsoft.com/enus/um/redmond/groups/connect/cscw_10/docs/p131.pdf

Kohler, W. (1969). Gestalt psychology: The definitive statement of the Gestalt theory. New York: Liveright.

Kwak, H., Lee, C., Park, H., & Moon, S. (2010). What is Twitter, a social network or a news media? Paper presented at the International World Wide Web Conference, Raleigh, North Carolina.

Leavitt, A., Burchard, E., Fisher, D., & Gilbert, S. (2009). The influentials: New approaches for analyzing influence on Twitter. Web Ecology Project. Retrieved January 9, 2015 from http://tinyurl.com/lzjlzq.

Lewin, K. (1947a). Frontiers in group dynamics: Concept, method, and reality in social science; social equilibria. Human Relations, 7(1), 5-40.

Lewin, K. (1947b). Frontiers in group dynamics: II. Channels of group life, social planning, and action research. Human Relations, 7(2), 143-153.

Lewin, K. (1949). Cassirer's philosophy of science and the social sciences. In A. Schelipp (Ed.). The philosophy of Ernest Cassirer (pp. 269-88). Evanston, IL: Library of Living Philosophers.

Lewin, K. (1951). Field theory in social science: Selected theoretical papers. New York: Harper.

McCombs, M., & Shaw, D. (1976). Structuring the "unseen environment." Journal of Communication, 26(2), 18-22.

McNeilly, J. (1959). Intermediary communicators in the international flow of news. Journalism Quarterly, 36(1), 23-26.

Papacharissi, Z. (2002). The virtual sphere: The Internet as a public sphere. New Media & Society, 4, 9-27.

Purcell, K., Rainie, L., Mitchell, A., Rosenstiel, T., & Olmstead, K. (2010). Understanding the participatory news consumer. Pew Research Center. Retrieved October 1, 2013, from http://www.pewinternet.org/~/ media/Files/Reports/2010/PIP_Understanding_the_ Participatory_News_Consumer.pdf.

Porter, L. V., & Sallot, L. M. (2003). The Internet and public relations: Investigating practitioners' roles and World Wide Web use. Journalism & Mass Communication Quarterly, 80, 603-622.

Ranney, A. (1983). Channels of power: The impact of television on American politics. New York: Basic Books.

Schroeder, R., & Kralemann, M. (2005). Journalism ex machina--Google News Germany and its news selection processes. Journalism studies, 6(2), 245-247.

Schultz, U., & Boland, R. (2000). Knowledge management technology and reproduction of knowledge work practices. Journal of Strategic Information Systems, 9, 193-212.

Schramm, W. (1949). Mass communications. Urbana, IL: University of Illinois Press.

Shields, M. (2009). News sites digg deeper. Media-Week, 79, 5-66.

Shoemaker, P. (1991). Gatekeeping. Newbury Park, CA: Sage Publications.

Shoemaker, P., & Cohen, A. (2006). News around the world: Content, practitioners and the public. New York, NY: Routledge.

Shoemaker, P., Eichholz, M., Kim, E., & Wrigley, B. (2001). Individual and routine forces in gatekeeping.

Journalism & Mass Communication Quarterly, 78, 233-246.

Shoemaker, P., & Reese, S. (1991). Mediating the message: Theories of influence on mass media content. New York, NY: Longman.

Shoemaker, P., & Reese, S. (1996). Mediating the message: Theories of influence on mass media content (2nd ed.). London and New York, NY: Routledge.

Shoemaker, P., & Reese, S. (2014). Mediating the message in the 27st century: A media sociological perspective. London and New York: Routledge.

Shoemaker, P., Tankard, J., & Lasorsa, D. (2004). How to build social science theories. Thousand Oaks CA: Sage.

Shoemaker, P., & Vos, T. (2009). Gatekeeping theory. London and New York, NY: Routledge.

Singer, J. (1997). Still guarding the gate? The newspaper journalist's role in an online world. Convergence: The Journal of Research Into New Media Technologies, 7, 72-89.

Singer, J. (2006). Stepping back from the gate: Online newspaper editors and the co-production of content in campaign 2004. Journalism & Mass Communication Quarterly, 83(2), 265-280.

Singer, J. (2014). User-generated visbility: Secondary gatekeeping in a shared media space. New Media & Society, 76(1), 55-73.

Sturges, P. (2001). Gatekeepers and other intermediaries. Aslib Proceedings, 53, 62-67.

Sundar, S. (2008). "The MAIN model: A heuristic approach to understanding technology effects on credibility." In

M. Metzger and A. Flanigan (Eds.). Digital Media, Youth, and Credibility (pp. 73-100). The John D. and Catherine T. MacArthur Foundation Series on Digital Media and Learning. Cambridge, MA: The MIT Press.

Sundar, S. S., Knobloch-Westerwick, S., & Hastall, M. R. (2007). News cues: Information scent and cognitive heuristics. Journal of the American Society for Information Science and Technology, 58(3), 366-378.

Sundar, S., & Nass, C. (2001). Conceptualizing sources in online news. Journal of Communication, 57, 52-72.

Sundar, S. S., Oeldorf-Hirsch, A., & Xu, Q. (2008). The bandwagon effect of collaborative filtering technology. CHI '08 extended abstracts on Human factors in computing systems (CHI EA '08). New York: ACM, 3453-3458.

Sundar, S. S., Xu, Q., & Oeldorf-Hirsch, A. (2009). Authority vs. peer: How interface cues influence users. Proceedings from the 27 th international conference on human factors in computing systems (CHI EA '09). New York: ACM, 4231-4236.

Watts, D., & Dodd, P. (2007). Influentials, networks, and public opinion formation. Journal of Consumer Research, 34, 441-458.

Weinberg, T. (2008, January 25). Every social network is different: Here's what you need to know. Techipedia. Retrieved October 1, 2013, from http://www .techipedia.com/2008/every-social-network-differs/.

Weng, J., & Lee, F. (2011). Event detection in Twitter. Proceedings from the Fifth International AAAI Conference on Weblogs and Social Media, Barcelona, Spain, 401-408. Retrieved January 14, 2015 from http://www.cse.fau.edu/~xqzhu/courses/cap6777 /event.detection.twitter.pdf.

Weng, J., Yao, Y., Leonardi, E., & Lee, B. (2011). Event detection in Twitter. Technical report, HP Labs. Retrieved January 26, 2015 from http://www.hpl.hp .com/techreports/2011/HPL-2011-98.pdf.

Westley, B., & MacLean, M. (1957). A conceptual model for communications research. Journalism Quarterly, 34, 31-38.

White, D. (1950). The "gate keeper": A case study in the selection of news. Journalism Quarterly, 27(4), 383-390.

Xu, W., & Feng, M. (2014). Talking to the broadcasters on Twitter: Networked gatekeeping in Twitter conversations with journalists. Journal of Broadcasting & Electronic Media, 58(3), 420-437.

David DeIuliis

deiuliisd@duq.edu