Introduction

The media ecosystem has experienced a 180-degree turn. Social media allow greater access to multimedia information in an interactive environment. The appearance of new media platforms has taken place parallel to the economic and financial difficulties experienced by traditional media, whose monolithic ability to demarcate widespread public perception is threatened, due in part to the emergence of a growing number of alternative sources, which allow citizens to collect and share information without having to resort to journalists or other professionals who usually establish the routes of information (^{Shaw, Hamm & Knott, 2000}).

From the perspective of Agenda Setting, the analysis of the relationship between traditional media and new virtual spaces has witnessed growing momentum ever since 2005 with the publication of Lee’s doctoral dissertation about the uses and effects of online media on public opinion (^{Lee, 2005}). Even when blogs and networks have gained ground in the circulation of information, there is not an unequivocal answer to whether new media dispute their agenda with old media or whether they copy their agendas, and especially whether they have the ability to create public opinion.

A detailed reading of studies on the interrelation between old and new media shows that the findings of researchers that tested the link between traditional elite media, blogs and Twitter, veered sharply from the original 1968 Chapel Hill thesis (^{McCombs & Shaw, 1972}). However, to explore the forces of power stemming from said relationship, it is worth highlighting that the simple presence of the content of one agenda within another agenda does not yield information about its capacity of influence. ^{Charron (1998)} defines the term “influence” as the capacity to configure any media coverage so it serves and supports a political leader’s agenda. ^{Ferree, Gamson, Gerhards and Rucht (2002)} agree that the fact that the voice of an individual actor or group appears in the news does not necessarily imply that it holds the opportunity to provide interpretation and meaning to the events in which it is involved. These authors refer to “standing” as the fact of having a voice in the media: “It refers to gaining the status of a media source whose interpretations are directly or indirectly quoted” (p. 86).

We agree with these authors that there is a need to incorporate tools for measuring an agenda’s capacity to influence another agenda. This theoretical and methodological position leads us to pose the following question: If Agenda Setting theory studies the relationship between media, political, and public agendas, under which of these areas do new media fall?

Theoretical developments are usually a part of a larger empirical study, but in some cases they are an “entire project in itself “especially when conceptions are hotly contested ... More than anything else, explication is about theorizing” (^{Kiousis, 2002, p. 356}). Since agenda setting analyses have reached heterogeneous results, the purpose of this study is to present a review of the literature that discusses the core theoretical and empirical premises achieved thus far for Agenda Setting, and propose a new demarcation of the boundaries in this terrain. The organization and reading of the bibliography has been structured according to three axes, which answer this study’s basal question: Do blogs and Twitter express public opinion, and therefore, promote and allow an interactive relationship between the media and the public, or are they a hierarchical organization dominated by a handful of actors (^{Calvo, 2015}), a channel that serves to perpetuate the agenda setting power of specific actors with influential capacity?

Social media as an expression of public opinion?

Who drives the agenda-setting process? If social media have come to occupy that leading role, will the tenets of Agenda Setting have to be adapted to a new model? Preliminary findings confirmed the hypothesis that newspapers maintain their influential capacity if one looks at how online conversations developed in South Korea (^{Lee, Lancendorfer & Lee, 2005}) and how digital versions of mass media succeeded —with the exception of the abortion issue— in setting the agenda of electronic bulletin boards (EBB) during the 1996 presidential elections in the United States. ^{Roberts, Wanta and Dzwo’s (2002)} supposition is that traditional media provide information on issues that people subsequently discuss in said news fora.

This evidence has also been observed in adjacent fields. In a study performed in 2006, ^{Yu and Aikat (2006)} found a revelatory correlation between the home pages of major online newspaper publications, online television channels and online news search engines. ^{Coleman and McCombs (2007)} analyzed the differences in agenda setting effects by relating Internet use to audience age in Louisiana and North Carolina. While media influence was weaker for intense Internet users and younger generations, it remained significant. Therefore, Coleman and McCombs concluded that the use of Internet did not eliminate the impact that the media have on audiences.

However, with the development of this kind of investigations, the agenda setting function of mass media toward an undeniably homogeneous audience starts to lose ground for several reasons. Firstly, an increased heterogeneity of the use of media by audiences challenges the idea that they should be seen as homogeneous and passive users. ^{Berger and Freeman (2011)} reinforce this hypothesis from the perspective of Active Audience Theory, which claims that it is not natural for individuals to accept dominant messages: they are singular audiences capable of interpreting, rejecting and challenging the media. This gives us motives to doubt whether this aspect is at all relevant to the discussion of the foundations of Agenda Setting. While the term “public opinion” and its measurement methods denote the idea of homogeneity, because polls are understood as being representative of the public’s perception of relevant issues, the fact remains that already in the late 70’s ^{Shaw and McCombs (1977)} classified variables that intervened between audiences and media —the so-called contingent conditions, such as education, socioeconomic status, the need for orientation, interpersonal communication, among others— which led them to identify differences in media effects according to each particular audience. Therefore, audiences in the past where already thought of as being heterogeneous.

Secondly, also worth mentioning is the time required for the media agenda to be absorbed and accepted by the public. McCombs and his followers refer to “time lag” as “the optimal time that a topic must be covered in the media before the public considers it as important” (^{Coleman, McCombs, Shaw & Weaver, 2009, p. 155}). During the height of traditional media, consensus was not reached in studies on time lag. Not only because of the varied and random selection of time periods (^{Winter & Eyal, 1981}), but also because of observed differences relating to the type of issue analyzed, its experiential relationship with the audience, the social composition of interviewees and the political context in which said influence is being measured. Although there has been a heated debate between those scholars who maintain that agenda setting effects are more significant when an issue being covered lasts over a greater time interval, while others maintain that the greatest levels of influence occur when information has recently been assigned priority by the media (^{Aruguete, 2015}), time lag is assumed to range from one to eight weeks, with an average of three weeks. As of the eighth week all the way up to the twenty-seventh week, the issue loses its salience until it completely disappears (^{Coleman et al., 2009}).

Conversely, the range of time needed for an issue to sink in with the public is not a component applicable to Agenda Setting studies within the new media environment. The emergence of Web 2.0 enables the production and circulation of user generated content in almost real time. Moreover, the new interactive culture in which these actors perform alongside information producers, not only as news consumers but also as producers, fosters opportunities for participation in the construction of online spaces (^{Gane & Beer, 2008}). In this sense, we note that the concepts of participation and interactivity need to be reworked. The purpose of this article is not to open a debate on these notions, but to expose the fact that these concepts were not assumed uncritically given that they have been vehemently discussed (^{Gane & Beer, 2008}; ^{Kiousis, 2002}).

Interactivity may be understood from a technically focused perspective of media systems (^{Manovich, 2001}), from a social point of view and as human experience framed by the use of these systems (^{Kiousis, 2002}), from the power dynamics that structure communication access (^{Schultz, 2000}) or even from a political conception tied to concrete opportunities of governmentalism and citizenship in a given society (^{Gane & Beer, 2008}).

“While some scholars see interactivity as a function of the medium itself, others argue that interactivity resides in the perceptions of those who participate in communication” (^{McMillan, 2000, p. 71}). We include this distinction because we agree with ^{Kiousis (2002)} that these terms are conceptually structured and placed under analysis in accordance with the way in which different scholars chose to conceptualize them. On this particular issue, it is worth noting that the ongoing influx of new communication technologies as well as the ability to consume information at the same time as it is produced, do not guarantee the democratic and participative flow of information. We insist that it is not the intention of this paper to open a theoretical debate about the terms participation and interactivity, but rather to stress that the time lag of agenda setting has varied due to the new communication dynamics, without necessarily regarding technological convergence in information production and consumption as being analogous to democratic participation. The analysis of this phenomenon should not evade the political dimension and socioeconomical context that shape specific opportunities of access to information and communication.

^{Thorndyke (2012)} posed the following question: “Do conversations existing on social media sites represent the agenda set forth by popular media sources?”, whilst intuitively sensing that trending topics on Twitter would copy issues from the headlines of the BBC, The New York Times and Al Jazeera. If such imitation took place and the most prominent conversations in social networks were to revolve around the headlines of major media, Twitter could become an exact measurement of public opinion (^{Thorndyke, 2012}).

^{Rubio García also takes Twitter as a “reflection of the public agenda” (2014, p. 249)}. Following the Agenda Setting hypothesis, she examined whether traditional media set the agenda of issues among Spanish citizens. She found a strong correlation between predominant issues in comparing the digital versions of El País and El Mundo newspapers and the commentaries of Twitter users in Spain.

In contrast, the correlations tested by Thorndyke did not arrive at the same conclusion. While strong homogeneity was found in the coverage given to the riots in the Middle East and to the global financial crisis by Al Jazeera, the BBC and The New York Times, these affairs were not given a prominent place in Twitter for the period analyzed. The salient topics in public opinion would seem to correspond more to traditional media rather than to social media. “Twitter is still being used mainly as a place to follow celebrity gossip and popular culture, more so than world events and news. Twitter’s ability to set the agenda is not as strong as traditional news media” (^{Thorndyke, 2012, p. 13}).

In the era of digital media, with myriad sources of information and expanded social media, the question on the ability of dominant media to continue establishing the agenda comes up again and again. Along these lines, ^{Artwick (2012)} brought into question the conventional postulates of the agenda setting effects of big media in her study on the sources of influence in Twitter during the period when the body of Virginia Tech student Morgan Harrington was found, after she disappeared in Virginia on October 17^th 2009 outside a Metallica concert.

Artwick examined the links in Tweets about the Harrington case as an indicator of newsworthy influence, and studied retweeted messages as elements that measure interpersonal communication. The retweets accounted for an active audience in social media, whose interpersonal conversations played a significant role in the new media. With this study, Artwick confirmed her prediction that information posted on this microblog principally derived from alternative sources, which influenced the content of over 60% of links and over 50% of retweets referring to this particular case between January 26^th and 28^th in 2010. Specifically, the weblogs and the alternative sources not containing mainstream news set the agenda on Twitter, thus confirming findings in earlier studies in which the incipient emergence of social media as sources of influence was made evident (^{Artwick, 2012}).

Differences in the findings of ^{Thorndyke (2012)} and ^{Artwick (2012)} are clear in the topics addressed by each of these authors. The non-experiential issues introduced by the BBC, The New York Times, and Al Jazeera (^{Thorndyke, 2012}) display no significant correlation with the agenda in Twitter. By contrast, the Harrington case —an experiential and obtrusive topic— displayed a positive correlation with that of Twitter. ^{Zucker (1978)} suggests that obtrusive issues —those that touch on issues that are experienced by readers— obstruct the effect that the media agenda has on people. In turn, Twitter users do not require a media stimulus to discuss issues which are meddled in everyday’s life. To summarize, the type of issues to be considered are a fundamental yet contingent determinant of the comparisons between studies of traditional media and social networks.

Users’ behavior appeared to model a hybrid version of the two-step flow of communication. But instead of the information flowing from mainstream media to opinion leaders, and from opinion leaders to followers (^{Lazarsfeld, Berelson & Gaudet, 1948}), Twitter received information from different sources (^{Artwick, 2012}).

In his study on the interaction between politicians, media and citizens about the #Nisman case on the Twitter social network, ^{Calvo (2015)} arrived at findings that were very different from those of Artwick. Twitter is a hierarchical network that consolidates the social standing of certain political actors —political leaders and far-reaching mass communications— through the spread of information which the public approves. Specifically, he observes high levels of “identification” between the mass media and social networks in that the information transmitted is consistent with that offered by traditional media (quoted in ^{Aruguete, 2016})². “Polarization (political polarization, such as the one currently observed in Argentina) is not simply a reflection of manipulations and maquinations ... it also results from the information offering that voters have at their disposal” (^{Calvo, 2015, p. 60}).

A common denominator in these studies is a return to the idea of active users whose perceptions are no longer permeated by unique sources of influence, fundamentally by a press that acts authoritatively (^{Shaw, Hamm & Knott, 2000}). This is even less viable when new media devices not only contribute to the diversification of sources according to individual interests, but also forge empowerment opportunities whilst challenging official dominant agendas, more than has ever happened in any other period. The Twitter agenda is constructed from an array of influences, of which The New York Times is an important influence, but an influence whose agenda setting capacity has been diluted (^{Kushin, 2010}).

Intermedia Agenda Setting. The encounter between old and new media in the 2.0 world

At the beginning of this century, the mainstream idea was that elite media, which tended to legitimize one another, had a greater effect over independent political blogs that were “forced” to include events in their postings that already had media coverage (^{Adamic & Glance, 2005}; ^{Cornfield, Carson, Kalis & Simon, 2005}; ^{Haas, 2005}). Close to half of the news links appearing in blogs gathered information from dominant media (^{Reese, Rutigliano, Hyun & Jeong, 2007}), while 20% collected information from other blogs that may have, at the same time, turned to other newspapers as a source (^{Murley & Smith, 2005}). “Political bloggers rarely do any original reporting, and as a result, they tend to rely primarily on established media outlets for their information” (^{Wallsten, 2010, p. 166}).

In early 2000, weblogs became a potential alternative source of information to traditional media. In that context, ^{Messner and Distaso (2008)} raised the question of which of these two media had greater influence over the other. In a study about the use of sources by traditional media and weblogs between 2000 and 2005, they found an increased number of references to weblogs in news coverage on politics, entertainment, business and sports. This tendency was strengthened in 2002. They also noted that whereas weblogs did not resort to news sources regularly, when they did, newspapers, television, radio and magazines took the lead as media outlets in almost half of the cases.

Accordingly, while at the beginning of the century weblogs rarely appeared in traditional media, with time they gained legitimacy as a source of information. “While overall the public has not taken widespread notice of weblogs, the traditional media have clearly increased their attention to this new journalistic format” (^{Messner & Distaso, 2008, p. 454}). However, the growing recurrence to new media was not a guarantee of the greater democratization of information flows. Indeed, in this new landscape a “blogger aristocracy” persisted with a greater power of influence (^{Meraz, 2009}). The same is true for the Twitter network, where certain voices enjoy higher levels of hierarchy and impose themselves onto others, thus producing aristocracies in social networks that tend to consolidate their social relevance (^{Aruguete, 2016}).

^{Meraz (2009}, ²⁰¹¹⁾ seeks to unravel the potential of citizen media —more precisely, independent political weblogs— in order to invert the existing power structure between citizens and traditional media. In her first study she analyzed the use of hyperlinks in: 1) two dominant media, The New York Times and The Washington Post; 2) eleven newsroom political weblogs of these two newspapers; and 3) eighteen top independent political weblogs across the political spectrum. She reached conclusions that, although very consistent, may not be held as definitive or irrefutable. While The New York Times and The Washington Post maintain relative dominance as top sources for the top 20 weblogs, their propensity to exert their agenda setting power is evident in the short term while not in the long term. Their influence on citizen media is far from being “universal”, and they are one force among many competing influences. “Citizen media’s efficacy is in its aggregate effect, an effect which is able to blunt traditional media’s singular agenda setting effect” (^{Meraz, 2009, p. 701}).

In a subsequent study, ^{Meraz (2011)} took issue emphasis and selection as a parameter to gauge influence in both types of media, since she understood that investigating the influence of media sources through the analysis of hyperlinks meant arriving at limited conclusions. In this case, her suspicion that traditional channels acted as agenda setters in citizen media was partially confirmed since she did not corroborate this influence in the contents of left-leaning weblogs (^{Meraz, 2011}). In parallel, she researched the impact of diverse political ideologies amongst political weblogs and found interesting revelations: left wing weblogs were not only able to generate an alternative agenda, but they also set the issues of the agenda of right wing weblogs. The resistance that these weblogs had in reproducing the dominant agenda was the counterpart of the progressive dilution of traditional elite media’s “singular” agenda setting influence (^{Meraz, 2011}).

Within the study of intermedia agenda setting, a bidirectional process between traditional and new media is observed (Wanta & Foote, cited in ^{Parmelee, 2014}). Kushin bases his study on the idea of a bidirectional relationship whilst analyzing media effects in the new interactive scenario, specifically, between The New York Times and Twitter online publications. According to his hypothesis, said bidirectional influence takes place between information published over the course of a single day and across several days. His findings, however, do not corroborate this presumption. To begin with, social media effects on traditional media takes place in less than half of exchanges, while the remaining cases did not display a clear influence of one media over the other. This does not preclude that there exists a subtle influence of Twitter over The New York Times web; this platform “may not only play a role in how journalists gather news, but may have a minor influence on what issues journalists choose to cover” (^{Kushin, 2010, p. 135}).

Kushin’s study is interesting in that he incorporates other factors that have an effect on this new relationship. We will only discuss a few of them. As noted previously, ^{Kushin (2010)} argues that the nature of an issue that enters the agenda influences the type of effect it will achieve on the perception of audiences, in terms of its proximity to personal experience: “The nature of the event may have precluded the influence of Twitter resulting in an overall lack of intermedia agenda setting from Twitter to The New York Times for these issues” (p. 131).

^{McCombs (1993)} postulated that the agenda setting ability of the mass media is what grants them a powerful role: they promote social consensus around the configuration of an agenda and create a “sense of community” (p. 64). ^{Baum and Groeling (2008)} observe that online media offer audiences greater capability of becoming isolated in alternative perspectives and therefore contribute to the creation of heterogeneous media agendas. The challenge that this hypothesis poses lies not only in the fact that the fragmentation of agendas neutralizes the influence of mainstream media but there is a return to the thesis of selective perception postulated in the 1940’s and 1950’s by Lazarsfeld and Hovland, among others. Selective perception postulates that attention will be directed towards those supportive elements of a message, as far as voters are actively predisposed to avoid exposure to political information that is contrary to their beliefs and are receptive to information that supports their beliefs (^{McCombs, 1993}). In the present media landscape, “the growing ability of the individual to enact selective perception” (^{Kushin, 2010}, p. 128) is a key factor that deserves being rethought in the context of agenda setting. While in their seminal study ^{McCombs and Shaw (1972)} called into question the selective perception thesis, they also warned in subsequent studies about the need to conjugate sociological and psychological variables so as to include the affective dimensions of public opinion judgments. Shortly after, ^{McCombs and Weaver (1973)} also recognized that media influence rests on a psychological level, that is, inside every individual, since they identified significant differences across individual perceptions about the issues covered.

For this reason, we agree with ^{Kushin (2010)} that in the current media landscape “the debate between these two camps should be reexamined” (p. 128), but we disagree with the idea that the viability of agenda setting “may pose a key threat” (p. 128), since it was recognized early on at the beginnings of this perspective that media were not sufficient cause to explain the behavior of people, thus it was “necessary to give special consideration to those personal and social characteristics that mediated to this effect” (^{Aruguete, 2015, p. 83}).

Additionally, we must not lose track of the consistency —that is, the similarities found in the coverage of different media (^{Reese & Danielian, 1989})— between both types of media that include issues with the greatest coverage. This behavior allows us to infer that the influence of mass media in Twitter may not be direct and can take place at different stages. For example, through interpersonal conversations or by searching weblogs whose posts used content derived from elite media.

Agenda building in the new media environment

Social media have been studied as an expression of dominant discourse in public opinion, but also as an instrument for the political system to shape news coverage and even more so, to challenge journalism standards and practices. In the domain of agenda building, two lines of analysis are viable. Firstly, the degree of influence that the political agenda exercises —or not— on the media agenda, and secondly, the uses that journalists give to new media (^{Lasorsa, Lewis & Holton, 2012}; ^{Singer, 2005}). These two lines are addressed in this section. On the one hand, the interest involves examining the type of relationship that exists between the political agenda and the media agenda in the current environment, placing special emphasis on the notion of “influence” as discussed by ^{Charron (1998)}. On the other hand, the goal is to measure, on the basis of several studies, whether or not certain traditional media journalism standards have been reviewed.

Concluding remarks

The purpose of this article has been to systematically review a series of studies on the relationship between new and old media, and to explore the theoretical and methodological perspectives they set forth, while it presumes that citizen media have the ability of impacting the news agenda and of disseminating information traffic while challenging the mass media’s centralized role as news deciders.

It is a fact that new media have gained ground in the dispute for setting the agenda. However, in view of the heterogeneous findings observed, it is impossible to provide definite answers to the questions that were formulated at the beginning of this study: Do social networks set conversation topics or do they repeat the agenda of topics proposed by elite media? Does the agenda setting power claimed by official information sources persist in the new media environment? Do the new media constitute a real challenge for traditional journalism standards or do they serve as a normalizing tool? Finally, are trending topics an adequate measurement of the expression of public opinion?

In a context where reaching established findings is impossible due to the lack of unequivocal questions about this relationship, we tried to explore the extent to which the circulation dynamics of information created by the new media —particularly in weblogs and Twitter— is distorting the boundaries of the traditional postulates of Agenda Setting theory. The questions that underlie the studies that were analyzed vary in terms of the location of the weblogs or social networks in the different agenda: political, media or public.

The research studies that analyze these platforms as an expression of public discourse have reached varied conclusions. While there is a degree of consensus on the idea that the forces of power favor elite media, they do not hold the power to establish issues in a generalized way. In other words, a direct influence of the mass media over Twitter is not observed, but rather a two-step information process occurs such as Lazarsfeld had established. “Twitter users who tweeted information received it from a combination of sources” (^{Artwick, 2012, p. 19}).

Furthermore, there is no room for the optimistic view on the new ecosystem that accepts a non-critical replacement of the role of gatekeeper by that of the “gatewatcher”. Two elements back up this observation. On the one hand, the ^{Singer (2005)} and ^{Lasorsa et al. (2012)} studies show that journalists working for national newspapers, national television news divisions and cable news networks, use their weblogs and social networks as normalizing tools, so they are less likely to relinquish their gatekeeping role or to provide accountability and transparency. Consequently, the grounds for a cyber-urban producer to emerge, who can exert countervailing pressure on conventional journalism routines, has not been fully prepared.

On the other hand, McCombs observes that the multiple agendas available for a broad segment of the public have generated a change in the relationship that the public has with regards to information. As a consequence, the end of the ability of the media to fix the set of priorities is foreseen, to the extent that audiences possess a single external media agenda with a highly personalized configuration, which they actively build thanks to the vast wealth of online news and information (^{McCombs, 2005}). Meanwhile, in 2005 the father of the theory asked himself: How many people tune into such a diverse set of agendas offered by such a vast array of web sites? Two hypothetical responses are viable.

Firstly, the idea that Internet really has such a broad and fragmented audience remains to be proven. Seen through the lens of the digital divide, increased access to web 2.0 has not managed to decrease the gap: best-educated and most wealthy young adults dominate cybernetic audiences (^{Salwen, Garrison & Driscoll, 2005}). Another feature of this presumption is the occasional use of the web for obtaining information.This is a premise that must be questioned, just a few years after it was posed and especially with the explosion of social networks. A third aspect relates to the concentration of supply. While the five major newspapers in the United States with the largest circulation accumulated 21.5% of circulation among the top hundred newspapers, the websites of the top five newspapers represent 41.4% of the total of links found on Internet for these 100 newspapers (Hamilton, cited in ^{McCombs, 2005}). A finding subsequently confirmed by ^{Boczkowski (2010)} years later, when he observed that the Internet, a platform that was thought out to facilitate access to information, in reality is subject to a greater degree of concentration than that observed in traditional media.

In the area of intermedia agenda setting the growing presence of social media in the coverage of elite media is starting to be observed. ^{Kushin (2010)} shows that even while Twitter doesn’t play a decisive role in the way journalists gather information, they have a lesser influence on the topics that the press decides to cover, with the exception of obstrusive issues.

The real threat to the viability of agenda setting as we know it (^{Coleman et al., 2009}) stems from the existence of an active audience with the ability of promoting selective perception in the process of information consumption. Herein resides the most significant resistance towards the media’s agenda power, in the fact that they do not hold the monolithic power of homogeneously fixing the issues that matter. However, said multiplicity of sources does not imply destroying the propensity of indexing topics and points of view —according to this theory, the media tend to prioritize the affairs and perspectives of political elites and of those institutions with most power (^{Bennett, 1990}, ²⁰¹²)— that still maintain media coverage. Even in those cases where a significant recurrence of new media is observed as an information input, the persistence of a blogger aristocracy and the proliferation of social media accounts of political leaders and think tanks inhibit the desired democratization of the flow of information. However, the influence of the media should not be analyzed on the abstract. It is necessary to observe the political and sociocultural context in which the media drafts and sets the agenda —both the traditional and the social media— and how they frame a particular viewpoint on public issues. That is, we take as an assumption that “news messages cannot be read, apprehended, or studied in isolation of the context in which they are produced” (^{Aruguete, 2013, pp. 208-209}).