Shifting Circles: Reconceptualizing Shoemaker and Reese’s Theory of a Hierarchy of Influences on Media Content for a Newer Media Era


By Susan Keith

WJMCR 29 (February 2011)

Introduction | Gatekeeping Roots | Hierarchy of Influences | Reconceptualizing | Conclusion


Drawing on Merrill Morris and Christine Ogan’s 1996 call for researchers to pursue theoretically grounded approaches to the Internet’s “many different configurations of communication,” this article considers two ways in which  Shoemaker and  Reese’s model of a hierarchy of influences on media content might be reconsidered in examinations of newer-media content. The paper argues that the Shoemaker and Reese model does not account well for a longstanding lack of consistent routines in multimedia/newer media professional journalism or for individual Web content produced by individuals not laboring for organizations. Then the paper proposes a reconceptualization of the hierarchy of influences model that would account for the interplay between levels of influence on media content. This version of the model would allow for levels to be breached or eliminated when warranted by the realities of a media world in upheaval.


In a 1996 article1 that has been cited more than 560 times,2 Indiana University professor Christine Ogan and then-Ph.D. candidate Merrill Morris called for mass communication researchers to do a better job of considering the Internet as a mass medium. If scholars persisted in studying principally  “the traditional forms of broadcast and print media that fit much more conveniently into … theories of mass communication,” the authors wrote, “their theories about communication will become less useful. Not only will the discipline be left behind, but it will also miss an opportunity to explore and rethink answers to some of the central questions of mass communications research, questions that go to the heart of the model of source-message-receiver with which the field has struggled.”3

Much has changed since that article was simultaneously published in a joint issue of the International Communication Association’s  Journal of Communication and the fourth issue of the then-new Journal of Computer Mediated Communication. Mass communication researchers—as well as researchers from related fields, including journalism studies, media studies, computer-mediated communication, organizational communication and health communication—have embraced the Internet as a research subject in a variety of ways, often treating it at least implicitly as a mass medium. There is still room, however, for scholars to, as Morris and Ogan wrote, “rethink assumptions and categories.”4

One area in which this reconceptualizing would appear to be especially fruitful—even nearly 15 years later—is theory. In their essay, Morris and Ogan argued that mass communication researchers’ study of the Internet was constrained in the early years of the Internet’s diffusion because they had not found appropriate theoretical models.  The authors encouraged scholars to approach the Internet through the lenses of critical mass, diffusion of innovations, interactivity, uses and gratifications, social presence and network analysis.  They wrote that doing so would “allow researchers both to continue to use the structures of traditional media studies and to develop new ways of thinking about those structures.”5 Indeed, since 1996, researchers have studied the Internet using all the perspectives Morris and Ogan highlighted,6 especially interactivity.7

Study of the Internet grounded in those frameworks has most often focused on content and audiences, as has research on traditional media.8 One likely reason: The hurdles researchers have faced in obtaining access to mass media producers and production sites in offline environments are just as high or higher in online news media. Over the past 15 years, mainstream print and broadcast media have struggled with the organization, staffing, and even physical production locations of their Internet units. Among online news media outlets, for example, there is no standard staffing or process for preparing news for the Web.9 Even within individual newsrooms, routines have changed so often that dozens of routines for producing, editing, posting, and overseeing Web content may have been used and abandoned since the mid-1990s. That very tumult around online content production, however, suggests that a topic that has received some scrutiny10 deserves more.

One theoretical framework that would seem to be useful for studying online media production is  Shoemaker and  Reese’s theory of a hierarchy of influences on media content,11 which has been used in many studies of print and broadcast content and production.12 The theory—which, despite its content-centric name, focuses in large part on factors related to production—uses a visual model of concentric circles to describe the relative influences on media content of individuals, routines, organizations, factors outside the medium and ideology.13 This article proposes, however, that the model could benefit from being somewhat reconceptualized for studies of the Internet, particularly online news production, so that it 1) better acknowledges the persistent absence of established routines in some new media sectors and 2) better accounts for the newer-media content-production powers of individuals. It is hoped that in this way the article will contribute to what Morris and Ogan saw as an intellectual advantage of Internet growth: “allow[ing] scholars to rethink, rather than abandon, definitions and categories.”14

Gatekeeping and the Roots of a Hierarchy of Influences

Shoemaker and Reese’s theory of a hierarchy of influences on media content grows out of and is closely related to gatekeeping theory. Gatekeeping theory, in turn, developed out of a metaphor used by sociologist Kurt Lewin in describing how food moved through a series of “gates,” controlled by one or more gatekeepers, on its way from garden or grocery to the table.15 The metaphor was popularized in a media context by David Manning White, who became friends with Lewin while pursuing a Ph.D. at the University of Iowa.16 In 1950, White published research on the gatekeeping functions performed by a Midwestern newspaper wire editor, whom he dubbed “Mr. Gates” for his role in determining which wire-service copy was published and which was left out.17

White found that Mr. Gates’ news selection was “highly subjective … reliant on value-judgments based on the ‘gate keeper’s’ own set of experiences, attitudes and expectations,”18 a finding contextualized by White’s later admission that in his own copy-editing work at the newspaper, his “‘liberal’ political orientation” had affected his ability to edit a conservative columnist objectively.19 A 1966 replication of the study with the same wire editor still employed at the same newspaper, found that he reported similar subjectivity in news choices,20 as did wire editors in a 1997 study of wire-photo selection at a Northeastern newspaper21 and metro editors in a 2003-2005 study of story selection at a Korean-language daily newspaper in the United States.22 In most of those cases, however—as well as “Gates”-type studies of the news selections of daily newspaper and TV station staffers,23 television news producers,24 and a female wire editor25 —other influences were at work as well, including journalists’ perceptions of audience preferences,26 journalists’ knowledge of audience demographics,27 publisher preferences,28 wire service news rankings,29 and the news values of a corporate owner.30

White’s work has rightly been criticized31 for focusing on a newsroom position that probably had somewhat less power to influence the news report than it might have appeared on the surface, given that the wire editor (a position that no longer exists in many newsrooms) would have, no doubt, been subject to content prescriptions by more senior editors. However, White’s research undeniably influenced decades of gatekeeping scholarship. As Roberts has noted in tracing the evolution of gatekeeping theory,32 that scholarship has been dominated for nearly the past 20 years by Shoemaker, who wrote an influential 1991 book on gatekeeping,33 updated in 2009 with  Vos.34 In the newer book, Gatekeeping Theory, Shoemaker and Vos conceive of gatekeeping occurring at or being evident at five levels:35

  • Individual, in which selection of content is influenced by individual communicators, such as Mr. Gates.
  • Communication routines, where content selection is influenced by “norms of selection that have evolved over the history of mass communication.”36
  • Organizational, where group decision-making in specific settings affects content.
  • Social institution, where content is influenced by such factors as markets, audiences, advertisers, and interest groups.
  • Social system, in which content is influenced by the social systems or ideologies of societies.

Outlining a Hierarchy of Influences on Media Content

Those levels of analysis are similar in conception, if not always in name, to the levels of outlined in Shoemaker and Reese’s work on influences on mass media content, explained in the 1991 and 1996 editions of Mediating the Message: Theories of Influences on Mass Media Content.37 (A new version of the book, which the authors had been unable to persuade the original publisher to produce, is planned for 2011.)38 In outlining their conceptual framework, the authors attempted to refocus mass communication research in several important ways. First, they sought to nudge it away from a tendency to view the individual media practitioner as the sole arbiter of media content, noting that such a focus reflected U.S. cultural, methodological, and theoretical biases.<39 Second, the authors sought to alter the longstanding media effects tradition in mass communication research, which views media messages as independent variables that can influence some dependent variable, such as audience knowledge, attitude, or behavior. Instead, the hierarchy of influences model portrayed media content itself as a dependent variable influenced by multiple other factors.40 Third, Shoemaker and Reese attempted to make clear through their framework the problems with “studies that make observations at one level of analysis and interpret those findings at a higher level,”41 such as research that examines the behavior or attitudes of selected individuals and then draws conclusions from that data about media organizations as a whole. Finally, the authors sought to encourage other researchers to consider the connections among the multiple levels of influence on media content.42

As outlined in Mediating the Message, those levels of influence were, from the inner part of the model outward: individual, media routines, organizational, extramedia, and ideological. Only the last two differ markedly from the names used in Shoemaker and Vos’ 2009 book43 ; Shoemaker and Reese used “extramedia” and “ideological” for the fourth and fifth levels,44 while Shoemaker and Vos used “social institution” and “social system.”45 Descriptions of those levels indicated, however, that both books refer to essentially the same concepts.

What distinguishes the framework offered in Mediating the Message from that outlined in Shoemaker and Vos’s gatekeeping book is the implication in the former book that some influences are generally more powerful than others. Although not explicitly stated, that idea of different levels having stronger or weaker influences on media content is inherent in the word “hierarchy” in the title of the Shoemaker and Reese book, in the name of their conceptual framework, and in the examples they cite. For instance, Shoemaker and Reese write:
[C]onsider the gatekeeper label commonly applied to mass media decision makers. … Whether in news or entertainment industries, the media gatekeeper must winnow down a larger number of potential messages to a few. The book publisher chooses from many possible titles; the network programmer selects from among several ideas for sitcoms, serials, and dramas to compose a prime-time schedule; and the newspaper editor must decide on a handful of stories to run on the front page. … But are those decisions made at the whim of the individual? Perhaps. The popular notion is that they are. These gatekeepers, however, represent their respective professions and organizations. As such the occupational setting limits their decisions. To understand these limits, we have to consider the media system within which people work, including the routines and craft norms that are so much a part of systematic information-gathering. The standardized, recurring patterns of news and entertainment content result in large part from these routine practices. These routines ensure that the media systems will respond in predictable ways and cannot be easily violated. They form a cohesive set of rules and become integral parts of what it means to be a media professional.46 

In those words, Shoemaker and Reese suggest that, in most cases, the power of media routines to influence media content is inherently greater than the power of individuals to influence media content. Similarly, when discussing the ideological level of analysis, the authors note that it is more powerful than any of the other levels of influence, writing that it “subsumes all the others we have been talking about and, therefore, is the most macro of the levels in our hierarchy of influences model.”47

The notion that some factors influencing media content are more powerful than others is further reinforced by the visual model that Shoemaker and Reese use to illustrate their theoretical framework. It displays the five levels of influence on media content as concentric circles, with the individual content producer occupying the center core, media routines occupying the layer just beyond the core, organizational influences making up the layer beyond that, extramedia influences filling the next layer, and ideological influences occupying the outermost layer.48 The visual implication is that individual communicators generally will not be able to consistently exert idiosyncratic influence on a particular type of media content unless there is no established media routine governing production of that type of content, organizational and extramedia constraints do not apply, and production of the content somehow escapes ideological constraints. Otherwise, as Reese has put it, the individual is “progressively hemmed in by more and more layers of constraint.”49

Re-conceptualizing the Hierarchy of Influences for a Digital Age

The hierarchy of influences model has been criticized for being somewhat U.S.-centric;50 for being somewhat unclear about the reasons behind its organization, especially placement of the organizational level above the media routines level;51 and for failing to fully account for “how one can empirically separate the effect of one level from that of another; or how one can accurately compare those effects with one another.”52 At least one other scholar has proposed a variation on the model,53 and Reese himself has noted that frameworks like the one he and Shoemaker developed “cannot capture all of the complex interrelationships involved in the media”54 or provide a comprehensive picture of media production and content. One area in which a gap seems particularly evident is with multimedia and emerging media. This section describes two such situations and suggests ways that the model and, especially, its visual representation might be reconceived to better fit these emerging realities.

Emerging media and a lack of established routines

One of the most powerful ideas in Shoemaker and Reese’s theoretical framework is the notion that routines embedded in the craft or professional culture of media production—stemming from the needs of information suppliers, content processors, and media consumers55 >—can be more powerful in determining media content than individual journalists’ backgrounds and experiences. Taking the example of television journalism, the authors note:

Television newspeople change jobs more often than print journalists. This creates a continual turnover in personnel, which make easily learned routines essential for smooth organization continuity. Television news requires careful coordination of complex technologies (e.g., videotape editing, microwave and satellite transmissions) requiring specialized roles, scheduling, and other routinized procedures to bring it off smoothly. In addition, competition has led stations to rely on news consultants who prescribe formulaic guidelines for the number of stories and their length.56

Or, as a former TV journalist explained to the author of this paper and a colleague:57

You slam together your story, and you better give your editor 15 minutes in a live truck . . . with a machine that’s probably going to break at least twice in that 15 minutes. And you better establish a live shot . . . . And you better hope they can receive it at [the station]. And you better hope that [a competitor] isn’t parked next to you with a more powerful signal, so when they power up, they blast you off the air . . . And then if you want graphics, well, you’ve got to convey the graphics you want to at least two people over there. And then two people work on it and then they put them down on tape and hope it rolls on time.58

It is easy to see that in such a setting, where producing content effectively depends on following successive steps in well-honed routines, individuals are constrained in many ways from substantially and idiosyncratically altering media content.

But what about media milieu in which routines have yet to develop or fully solidify? The hierarchy of influences model does not provide an account of such situations, although change theory by Lewin, who devised the gatekeeping metaphor, implies that they might exist in media in transition. Lewin also developed a theory of organizational change that posited that there are three steps in successful group change: “unfreezing” the status quo, transitioning to a new state, and “refreezing” that new state.59 In Lewin’s view, however, it takes a great force to propel change in the first place and then a force greater than the forces resisting change to move an organization out of transition into a new state that has the potential to be solidified. In addition, Lewin argued, change is not likely to “take” unless permanence is planned. Lewin’s theory of change has sometimes been viewed as somewhat simplistic, and some organizational communication scholars have suggested modifications to it.60 The framework nevertheless powerfully suggests that it may take a long time for new routines to solidify when media are going through an era of transition, as they have been in the first decade of the 21st century. That would further seem to leave some opportunity open for individual media producers to affect content to a greater degree than in a media setting where firmly established routines ruled.

Research on cross-platform newsroom convergence and online media supports that suggestion. In 2000, for example, Zavoina and Reichert reported that data from a survey of daily newspaper photo editors and Web directors showed that presently there is no industry standard to handle dual [print and online] publishing. Whereas some online news sites are independent from the chain of command and protocol that exists within their traditional news department counterparts, other online news sites work hand-in-hand with the existing photo (visuals) staff for the hard copy edition.”61

Other studies of newsrooms in transition have also reported periods of rapid change that resulted in a lack of long-term fixed routines for newer media practices. In 2006, Silcock and Keith reported that news production routines in place in 2003, when they began a study of cross-platform newsroom convergence operations among the Arizona Republic, KPNX-TV, and the Web site, had been significantly altered by 2005.62 Similarly, Keith reported in 2009 that interviews with copy editors in the U.S. and other predominantly English-speaking countries indicated that routines for editing copy and producing pages were changing rapidly in some newsrooms, as newspapers facing sharp print circulation declines sought to cut expenses by repeatedly eliminating staff members and consolidating or outsourcing print news production duties.63

In online journalism, routines remain so unsolidified—perhaps because early enthusiasm for the Internet’s potential for hypertextuality, multimediality and interactivity has not yet been matched by what journalism organizations actually do on the Web64 —that it did not seem unusual for online journalist Gina Chen to propose in a 2009 blog post “what a typical day should/could be like in an online-first newsroom.”65 (In contrast, routines for specific types of print and broadcast media are so well established that describing an ideal day likely would not be considered blog-worthy.) Approaches to creating and overseeing the creation of online news differ across national boundaries,66 within nations,67 and even within regions.68

Although some of these differences may be the result of the persistent power of traditional digital routines,69 it seems likely that there are increasingly situations where a lack of the sort of established routines that Shoemaker and Reese describe allows individuals to have greater influence on newer media than was originally foreseen in the hierarchy of influences model. That does not mean that the model is no longer useful, but merely that it needs to be reconceptualized to account for periods like the present one when media routines are undergoing great upheaval. One way to account for such situations is to rework the visual model of the hierarchy of influences.

Imagine that rather than being presented as a series of static concentric circles arranged around a core (the individual level of influence), the model is presented as a dynamic system of layers, any one of which can be “breached” by forces bubbling up from below. Where the influence of one factor, such as media routines, is weak—as it appears to be with some types of online journalism in the early 21st century—that layer of influence can thin to the point that it allows innovative content producers to burst out of the subsumed core position to which the model relegated individual influence on content. In that case, individual content producers might find themselves constrained by fewer influences, perhaps only organizational factors, extramedia influences, and ideology. (See Figure 1.) Keith reported one such example in 2009: A copy editor at a small Midwestern daily started using Twitter to promote her newspaper’s content70 —at about the time when the social media usage of some journalists, including those at The Washington Post, was being restricted71 —not because someone told her to but because no one told her not to. In the absence of media routines- or organizational-level constraints, a single individual journalist was able to have great influence over the manner in which her newspaper’s content was presented.

One might also imagine that a layer higher in the hierarchy than that of media routines could be thinned to the point of breaking. During a revolution, change of government, or other prolonged period of national unrest, ideological influences on media content might be so changed or extramedia influences, such as legal restrictions, so altered that at least for a brief time, a media outlet’s willingness to publish specific content in an uncertain environment—an organizational-level influence—would be the highest level of constraint on media content.72

The hierarchy of influences and the Web’s lone wolves

Another common 21st century media reality that the Shoemaker and Reese model does not account for very well is that much online content is produced by individual bloggers, social media users, and Web site operators working outside traditional organizational constraints. Between 2002 and early 2009, the Web site Technorati73 indexed 133 million blogs,74 and by mid-November 2010, the social networking site Facebook had more than 500 million active users.75 As of the writing of this sentence in mid-November 2010, there had been more than 29.7 billion Twitter posts produced.76 Although some of that online content involved surely was produced for organizations (news media outlets, public relations or marketing firms, universities, retailers and other businesses, etc.), a significant percentage would have been created by individuals not operating on behalf of any organization.

Those individuals’ content obviously would be subject to ideological-level influences. People blogging, Tweeting, or using Facebook in the United States, for example, can do so relatively freely because the U.S. social and legal systems place a high value on free expression. Individual online content producers would be subject to extramedia constraints in the form of privacy and defamation laws or the opinions of their online readers, “friends,” and “followers.” To some extent, they also may be subject to media routines, because fairly widely followed conventions have developed for some types of individual online content production, such as Twitter posts.77 But many individual producers of online content are not subject to organizational constraints because they are not creating on behalf of an organization—at least not in the same way that a journalist creates content for a news organization.

That suggests that scholars using the hierarchy of influences model in researching newer media work by individuals should feel free to eliminate the organizational level of the model from consideration if warranted. Shoemaker and Reese never suggest that any researcher studying media content should consider all the levels of influence described in their model; in fact, they acknowledge that “any given study cannot address all these levels at once.”78 But Shoemaker and Vos do suggest that “more studies need to span multiple levels of analysis,”79 an observation that tends to emphasize the static nature of at least the visual representation of the theoretical framework. A better approach for scholars studying newer media produced by individuals would be to think of the layers of the hierarchy of influences visual model as not only breach-able but dynamic enough to, in some cases, disappear. So a visualization of a model of the hierarchy of influences on the content of online media produced by a Web “lone wolf,” an individual not affiliated with an organization, might leave out the organizational level altogether while maintaining the dynamic, level-breaching capabilities of the model described in the previous section. (See Figure 2.)


This article was written in the spirit of Morris and Ogan’s essay “The Internet as Mass Medium,” which called for researchers to recognize the Internet’s “many different configurations of communication.”80 The article throws a spotlight on two common types of communication configuration: multimedia/newer media professional journalism and individual digital content produced by individuals not formally working for organizations. The article argues that those types of communication are not well accounted for by the hierarchy of influences model, a frequently used and compelling theoretical framework for examining media production and content. By proposing small reconceptualizations that would allow the model to be seen as a more dynamic and fluid framework for considering the interplay between levels of influence on media content, this article aims to help preserve Shoemaker and Reese’s important work.

As this article was being finished, Shoemaker and Reese were preparing a third edition of Mediating the Message, in a process that Shoemaker had described as “writing anew as much as … revising the old text.”81 That update to the model, the first since widespread diffusion of online media, will no doubt take into account the Internet and other newer media. Until the third edition of Mediating the Message is released, however, researchers using the hierarchy of influences model in studies of digital media should consider viewing the model as more dynamic than it may seem to be on a printed page.

This effort to reconceptualize a popular media theory should also serve as a call to other scholars to think about whether and how existing media-related theories, theoretical frameworks, and conceptual frameworks may need to be reimagined for a digital era. Many theories traditionally used to understand media were conceptualized when “mass media” was a clearly definable concept and the so-called “mainstream media,” now in decline, held greater power. That suggests that we should add to Shoemaker and Vos’s call for researchers to “think creatively about applying … theory to a changing world and to adapt research methodology to keep pace”82 a call to critically engage theoretical models’ fitness for the 21st century.

Susan Keith is an assistant professor in the Department of Journalism and Media Studies in the School of Communication and Information at Rutgers University.

Leave a comment

Your email address will not be published. Required fields are marked *