Trends of Agenda Setting Research: A Bibliometric and A Thematic Meta-Analysis

Agenda-setting studies continue to experience an evolutionary process. It goes beyond its initial assumption, which is the transfer of meaning from the media agenda to the public agenda, and expands to experience replication. Recent literature studies regarding the mapping of agenda-setting studies have not been carried out much. Therefore, this study aims to find trends in agenda setting research in the global scope based on agenda setting data from 2014 to 2022. Trends in the global scope are interesting to study to see agenda-setting studies today. This study uses the Bibliometric Analysis and a thematic theme analysis approaches. The research show that politics-related topics have dominated over the last eight years. The use of network agenda-setting (NAS) and agenda-setting intermedia (IAS) theory, content analysis and survey, and Twitter are essentials part of this study. The entire development of digital media is slowly leaving conventional media. Therefore, future studies, in the presence of a variety of media platforms, need to design alternative models and methodologies that can explain the power of influence of each media in shaping the agenda-setting effect.


INTRODUCTION
conducted by Zhou (2016) looks at a comparison of agenda-setting in China and Korea. The differences in the Korean study reflect similarities with the US study in terms of research focus. Zhou found studying in Korea more topical and less interested in social issues.
On the other hand, research in China concentrates more on social issues and is usually atheoretical, i.e., adopting phenomena or field facts and lacking in methodological diversity. In common, many studies from the two countries have turned to the internet and social media (Zhou, et al., 2016). The study conducted by Luo (2018) examines the regulation of media effects on the public agenda through a strict meta-analytic approach to analyzing empirical agendas. Studies published from 1972 to 2015 and 67 studies that met the inclusion criteria for the analysis resulted in a moderate grand mean effect size of 0.487 (Luo, et al., 2019). The results of the Intermedia Agenda Setting (IAS) Study published between 1997 and 2019 reveal the reciprocal flow from traditional media to new media, such as online media and social media (Su, 2021). Another study of agenda-setting with the topic of immigration in Spain between 2015-2020. The study's results revealed the relevant aspects of immigration when forming an agenda, the relationship between the media and public agendas, and the effect on the perception of Spanish public opinion (Tirado-Espín, 2020). However, the literature study is still local and limited to one pattern: media effects or one topic.
Another study, using a thematic meta-analysis approach , examined theoretical, topical, and methodological trends in agenda-setting research from 1972 to 2015. Findings indicated that the number of agenda-setting research studies had increased over time, along with expanding research topics, media, methods, and other uses of theory. However, the research does not include research impact factors, a collaboration between researchers, the spread of theory to various countries, and text corpus models. In addition, the study was only up to the 1972 to 2015 deadline.
To summarize the literature review of previous agenda-setting research. The research only focuses on one pattern, such as the study of the effect of agenda setting, intermedia agenda setting (IAS), the pattern of themes regarding immigration, and local scope. Meanwhile, the methodology only uses one approach, namely a quantitative or qualitative approach.
On the other hand, the development of communication technology marked by the massive digital platform-based media, such as online and social media, has shifted the position of conventional media, such as newspapers and television (Hanna, et al., 2011). Digital transformation has opened up multiple platforms and new media, such as online and social media. This interactive platform transforms the public into consumers and producers of information (Stone & Wang, 2019). The consequences will change the relationship between the media and public agendas. Therefore, to fill the gaps in previous research and see trends in agenda-setting research, it is important to clearly describe the latest developments in agenda-setting studies to analyze and reevaluate this research area. Through a bibliometric approach and thematic meta-analysis on a research basis, agenda-setting between 2014 and 2022 can become a foundation that helps better understand the evolution of knowledge so that it can build a solid foundation for advancing agenda-setting theory. This research enables and empowers researchers to understand knowledge integratively, identify gaps, get updates, and position research contributions.

METHODS
This study uses a bibliometric approach and thematic meta-analysis. Bibliometrics aims to map well-established fields' cumulative and evolutionary scientific knowledge by understanding large amounts of unstructured data (Zhong & Liu, 2022). There are several strategies applied in this research mapping. First, look at the highest number of citations from a journal article. Citation analysis is a science mapping technique that assumes frequently cited publications with similar themes (Hjørland, 2013). Second, a collaboration between writers and countries. This would show agenda-setting research interacting with each other, including associated author attributes such as affiliated institutions and countries (Donthu, et al., 2021). This proves that agenda-setting studies have spread throughout the world. Third, Corpus text assumes words that often appear together have a thematic relationship with each other (Inamdar et al., 2020). For example, corpus text will show the frequency of words often appearing in the research title and abstract.
A thematic meta-analysis is an approach that helps identify, analyze, and interpret patterns of meaning through a systematic meta-synthesis of relevant research studies (Cheng et al., 2021). Metaanalysis is a method for identifying, analyzing, and reporting patterns (themes) in data to reflect the reality that can open or reveal the surface of a 'reality' (Guzzo et al., 1987). This study will reveal patterns regarding topics, methodologies, other theories, and the media used in agenda-setting research.
In data collection, I developed a meta-analysis protocol with the PRISMA model (Preferred Reporting Items for Systematic Reviews and Meta-analyses), meanwhile, regarding the agendasetting research publication data focusing on databases originating from Scopus. Therefore, Scopus is a database containing abstracts and citations from peer-reviewed journal articles and literature from various publisher sources by applying strict publication standards. Furthermore, every literature/journal and website registered with Scopus has passed a review test by "content selection" and "advisory board" from researchers and librarians from various disciplines.
Searching for the Scopus database using the Publish or Perish application, I typed in the keyword "agenda setting" from 2014 to 2022. Meanwhile, articles as units of analysis are the results of research. Therefore, journal articles from editorial notes, reviews, and books are not part of this study. This limitation is made because publications based on research results can describe patterns according to the research objectives. The results of research agenda-setting research from 2014 to 2022, 1325 journal articles were collected. A strict screening was carried out, 148 journal articles were obtained as research samples, and 86 publishers came from as the unit of analysis. Development with the PRISMA model can be seen in Figure 1.
Bibliometric data analysis uses R Studio with data visualization on researcher collaboration and interaction between countries through Viosviewer. While data analysis for A thematic meta-analysis applies a coding scheme with criteria, such as research trends (topics), trends in agenda-setting theory, use of other theories, media used, and methodology.
For the level of data reliability, two coders were involved in the coding scheme and procedure, and the coders were carried out by two authors of this article for the inter-coder reliability test; the two coders coded 25 articles randomly selected. The coders engaged in coding practice independently and applied the coding scheme to the pretest.  The inter-coder reliability score using Cohen's Kappa was 0.885 for research trends, 0.735 for research topics, 0.841 for research methods, 0.787 for media, and 0.768 for the use of other theories. The overall inter-coder reliability score for all variables was 0.803.

Bibliometric Analysis
It starts from the assumption that the higher the number of citations an author has, the more likely they are to be cited, and it gives direction to new research. The results of the bibliometric analysis of agenda-setting research from the ten most cited references from 2014 to 2022 are agenda-setting research regarding the dynamics of communication on political issues in traditional media and social media (Neuman et al., 2014;Camaj, 2014). The following most cited quote is to test the power of fake news during elections, and themes regarding elections, especially political campaigns, are of concern and are often cited by agenda-setting researchers. Outside of political campaigns, other themes are also of concern to researchers. This is in line with the predictions of McCombs (2005). Agendasetting studies have expanded to other fields or topics outside of politics (McCombs, 2005). Such as public policy (Luo, 2014), sports (Denham, 2014), corporate communications, health (McGetrick, et al., 2019), environment (Ritter, 2020), and disaster communications (Valenzuela et al., 2017). In the agenda-setting study, there has been interaction from researchers from various countries. Figure 2 shows the pattern of researcher collaboration. Eighty-one researchers collaborated, divided into 5 clusters (each marked with the same color relationship). The Kiosis, Spiro Cluster (2014) became the center of research collaboration because many researchers collaborated eight times. They were followed by the Kim cluster, Ji Young, Zhang cluster, Tianduo, Barbara Myslikm cluster and Ragas cluster, and Mathew. To see the spread of agenda-setting research globally. Figure 3 explains 23 countries are divided into 6 clusters, namely the USA, Belgium, China, South Korea, Austria, and Sweden, which always collaborate. The United States of America (USA) is a center that produces a lot of research on agenda setting and doing a lot of collaboration with other countries.
In corpus text, the highest frequency of words that often appear from titles and abstracts, of the top 5 words in public at 1.20%, political at 1.02%, the issue at 0.86%, social at 0.85%, and Twitter at 070%. This means that agenda-setting research during the 2014-2022 period reviews more on the effects that occur on the public as a result of media agendas regarding political issues from social media Twitter, Figure 4. This explains the orientation of researchers in looking at the impact of the media is no longer focused on conventional media such as newspapers, television, or radio. Rather, this research looks into in digital-based media, such as social media and online media.

A Thematic Meta-Analysis
Results of thematic meta-analysis for agenda-setting theory trends often used by researchers. The coding process is based on, first, the evolutionary stages of agenda-setting, namely, the immediate effects, the attribute agenda-setting, the psychology of agenda-setting effects, the sources of the media agenda, and the consequences of agenda-setting effects (McCombs, 2005). Second, the theoretical replication of agenda setting, namely agenda building, Network Agenda Setting (NAS), Need For Orientation (NFO), and agenda melding . Therefore, the theme criteria for the trend of agenda-setting theory in this study are (1) First Level Agenda setting. At this level, the news media significantly influences the focus of public attention. Special issues are considered by members of the public as the top priority at all times . The transfer of salience from the news media to the public is often documented as an initial step in forming public opinion. (2) The agenda-setting attribute is the Second Level Agenda Setting. The agenda attribute study examines the salience of these two attribute dimensions. First, the substantive dimension describes the specific characteristics of issues on the media and public agenda. Second, the affective dimension is positive, negative, or neutral. The agenda attribute emphasized by the media affects the salience of this attribute in public opinion (Cheng & Chan, 2015). (3) Psychology of agenda-setting effects is understanding the power of agenda-setting effects with the concept of the Need For Orientation (NFO) (Camaj, 2014). (4) Agenda building, this theory focuses on investigating the interactions between the media and news sources (e.g., decision-makers, political actors, policymakers, the media, and the public) influencing the formation of the media agenda (Kiousis et al., 2016). (5) Intermedia Agenda Seeting (IAS), This theory explains how content transfers between news media (Harder et al., 2017). (6) Network Agenda Setting, this theory explains the importance of the reciprocal relationship between constructs or associative networks regarding specific topics which can be transferred from the media agenda to the public agenda Woo et al., 2020). This concept shows the importance of network objects and media attributes on public network effects (Sulistyanto, et al., 2019). Therefore, the agenda-setting network involves the effects of the media agenda network on the public agenda network (Guo, 2012). (7) Agenda Melding, this theory explains that individuals are motivated to affiliate to form agendas and leave their original environment, because of media connections (Shaw et al., 1999).
The coding results regarding agenda-setting theory are presented in Figure 5. Network agenda setting became the dominant theory used by researchers by 41 (27.7%). The second rank is Intermedia Agenda Setting (IAS) with 33 (22.3%), and the third, is First Level Agenda Setting with 27 (18.6%). Meanwhile, theories that are less interesting to researchers, such as Need For Orientation, and Agenda Melding, are equal to or less than 5 (3.38%). This is because both theories focus on public activity, while the media landscape develops rapidly and is convergent, so it is the choice of researchers to prioritize testing media effects. Figure 6 describes the use of other theories for understanding media and public phenomena. The use of other theories is constantly increasing and varied. In previous studies, framing and priming were prominent in supporting agenda-setting research, namely 60 (54.%) and 22 (19.8%). The results of the agenda-setting research for the 2014-2022 period, the most dominant media exposure theory is 21 (25.61%), while the Preeminent Theory is 10 (12. 2%). These two theories are new in agendasetting research, replacing priming theory by 11 (13.41%) and framing by 4 (4.88%).
According to this research, many new theories are used in agenda-setting research, such as Mediated Public Diplomacy Theory (Zhang et al., 2017), Situational Crisis Communication Theory (Han et al, 2017), The Extended Parallel Process Model (Chen, 2019), and others.  Figure 7 shows the results regarding the topic or theme that is the object of the agenda-setting study. Topics regarding political campaigns stood out compared to other topics by 29 (19.6%), followed by topics regarding health by 14 (9.46%), and environment by 13 (8.78%). The dominance of political campaigns shows that agenda-setting researchers are more interested in the topic when the agenda-setting theory is built regarding general elections. In particular, political campaigning is an interesting habit to study. This relates to the strengthening of democratic values. Studies on general elections, presidential debates, or political scandals are the themes that are often researched. The flexibility of agenda-setting theory can be used in viewing various communication phenomena, opening up opportunities for new topics that have never been researched, for example, regarding companies or corporations (Cheung, 2020;Park et al., 2019;Badham, 2019), history (Cannon & Cannon, 2019), or also studies on public policy. Furthermore, Twitter was the most widely used by agenda-setting researchers by 32 (21.62%). Newspapers were also still widely used in agenda-setting research, by 28 (18.92%) and online media by 22 (14.86%), such as shown in Figure 8. This study takes 2014-2022 period because that period is based on the author's assumption that there is a transformation in media use from conventional media to digital media platforms, including Twitter. In 8 years, we collected 32 agenda-setting studies using Twitter as the medium. The Publish or Perish application can collect data on all articles that discuss agenda setting within the Global scope-the process of collecting data through the keywords "Agenda Settings and Twitter" on Publish or Perish. The use of agenda-setting and Twitter is the focus of searching for agenda-setting media usage patterns so that the results are automatically filtered. In the future, it is estimated that social media and online media will increase, leaving traditional media, such as newspapers and television. A new finding on the use of media is the use of new media, such as making policy (Vo et al., 2019) amendments (Finke, 2016), books (Fuhlhage et al., 2017), database content (Lee, 2015), stakeholder thinking (Vos, 2014), news videotapes (Melek, 2019). Although, the average new media is still relatively small at 1 (0.68%). The use of new media explained in the agenda-setting study shows the use of any method; most importantly, the media affects the public. Another finding is the combined use of media. This is the research objective to measure the comparison of media effects. The combination of newspapers, television, radio, social media and online media is a combination of media use that enriches the variety of agenda-setting research.
In looking at the methodologies often used in research, as shown in Figure 9, content analysis and surveys still dominate at 78(52.7%), while content analysis and network Social Analysis are at 19 (12.8%). The combined use of methods, namely content analysis, time series, and surveys, and the three use network analysis are at 18 (12.2%). There are several other prominent methods, such as interviews and time series. The combination of the use of these methods is common in agenda-setting research. However, some researchers combine three methods, using time series, content analysis, and surveys in one study. For example, the study on dynamics examined the effects of intermedia agenda setting among the Twitter feeds of the leading candidates for the 2012 US presidential election; there was a symbiotic relationship between agendas in Twitter posts and traditional news, with varying degrees of intensity and differences in time lag by issue (Conway et al., 2015).
Likewise, the application of new methods that did not exist before, such as time lag design, historical analysis, semantic network analysis, and vector autoregression (VAR). Using new methods, explaining the research agenda setting, is open to replicating the method. Examples of the use of this method, such as research on online searches, can be an essential element in forming public opinion (Lee et al., 2016).

Discussion
Overall the results of the literature review are different from previous studies. There are several significant findings in this study. In bibliometric analysis, the term "politics", especially political campaigns, has been widely studied by agenda-setting researchers. This confirms that agenda-setting research cannot be separated from the initial domain around general elections. However, agendasetting research also extends to other fields such as sports and health communication, the environment, and others. This is under the predictions of McCombs (2014); agenda-setting research will move in two directions, namely centrifugal and centripetal. Centrifugal is an extension of the domain of agenda-setting outward focus on public issues. Meanwhile, the centripetal research trend explains sticking to the core agenda-setting theory .
This study shows that agenda-setting research is currently dominated in the political field. This is caused by two things. First, there is media convergence from traditional media, such as newspapers and television, to digital media platforms (Gilardi et al., 2022). Consequently, digital media has reduced the gatekeeping power of traditional media. Second, issues related to political communication do not depend on traditional media coverage policies. The presence of digital media has placed individuals or groups able to act as political actors as well as the publics who interact in communicating political issues on digital media platforms. Individual or group views become a public sphere discussion channel for conveying aspirations, views on political policies, or even digital social movements. Thus, both the media agenda and the public have the power to influence each other. This is different from the previous conclusion of Kim and McComb (2007). Based on a study of the 2002 Governor and Senator elections in Texas, Kim and McComb concluded that the importance of the candidate-political attribute relational network in the media agenda would be positively associated with the public importance of this attribute network. The attribute centrality of political candidates in the media attribute network will be positively associated with the attribute centrality in the public network agenda (Kim & McCombs, 2007).
In future studies, it is important to study the influence of the public or netizens in shaping the media agenda, as described in agenda building. For example, political events regarding political actors that are uploaded on social media by the public and have the potential to go viral will make references to online or mainstream media as news sources. This study, at least, can explain the power of the public agenda in influencing the media agenda. This is clearly different from the initial assumptions of the agenda-setting theory, which explains the strength of the effects of the media agenda on the public agenda.
In a thematic meta-analysis of theoretical trends, third-level or network agenda setting is getting stronger. The researcher's interest in setting network agendas is due to internet which has risen the new media alternatives. Likewise, the agenda-setting of intermedia has increased from previous studies. This is because digital platforms, especially social media, provide an environment where people can communicate about important issues in the public's minds (Castells, 2007). Online discussions can add meaning and news value to specific issues, affecting the prominence of issues in conventional news media, such as newspapers. Moreover, the production process of traditional news media has changed since the advent of social media. Currently, many journalists use social media as an essential source of information (Park et al., 2019). As a result of the deteriorating financial situation among traditional news media, social media is becoming increasingly important as the primary source of news (Su & Borah, 2019).
Nevertheless, it does not mean that researchers rule out other theories. In recent theoretical trends, research on agenda setting is varied enough to combine internet media with other theories, such as agenda setting at the first level, second level, or agenda building. The findings of this study confirm that there has been a change in the dominance of the use of traditional or first-level agendasetting theory to be replaced with third-level theory or network agenda-setting. This shift is an evolutionary change from agenda-setting research. This aligns with McCombs (2014) statement regarding the evolution of agenda-setting theory .
For future studies, it is essential to design alternative models and methodologies that do not consider intermedia agenda-setting a strictly linear process. Therefore, a method is needed to determine which news analysis platform appears first so that it can track and follow events in detail. Such in-depth methods will open up far more accurate insights into how and when media outlets influence each other.
This study shows that political themes are still dominant, but other themes, such as health, public policy, sports, and history, are also essential to the study. This finding is also consistent with McCombs's (2005) prediction that the themes or sub-fields of study in agenda-setting research will extend to various fields outside public opinion (McCombs, 2005). The political theme is still dominant, consistent with the new direction of agenda-setting research.  describes one of the theoretical expansions: testing and validating theories in the digital and political landscape . This theoretical expansion demonstrates the vital link between digital media platforms and politics. For example, when a political actor tweets an issue, it will generate millions of views, retweets, favorites, and comments on Twitter.
Additionally, political actor tweets generate news coverage and can result in hundreds of news articles being written. This phenomenon shows that digital media is increasingly important in political communication. This research explains the relationship, such as the influence of salience decision issues on voting choices (Kiousis et al., 2015) and elections (Baumann et al., 2018;Pedro-Carañana, 2020;Carazo-Barrantes, 2021).
On the other hand, fake news is thought to impact the agenda-setting effect. Fake political news on websites can signal the importance of the news to the public. In certain cases, fake news can inform that specific issues are worthy and signal more coverage of these issues by credible media. For example, in the 2016 US presidential election, fake news stories about Donald Trump and accusations of raping a teenage girl .
At the first level of agenda setting, fake news can shape the public agenda; however, not for attribute agenda and network agenda setting. This is proven by studies; fake news almost never succeeds in sending false information or attributes to credible online media (Vargo & Guo, 2018). Therefore it is important to carry out further studies on the effects of fake news related to politics by combining intermedia agenda setting (IAS) with network agenda setting. For example, during the presidential election campaign, fake news can become a reference source for other media and form network associations for media agendas and public agendas.
This study shows the increasing expansion of research focusing on online and social media use, or intermedia combining mainstream media with online and social media. The method's implication is using the network settings program (NAS). This model asserts that news media not only tell what to think and how to think about it but also set network agendas to determine how the public relates to different messages of the shift (Guo & Vargo, 2017).
Although not yet dominant, the use of NAS in this review is slowly being used in agenda-setting studies. Due to the internet's progress, which can present new media alternatives, the NAS model is important as an essential future. Therefore, for future agenda-setting studies, it is important to consider applying the NAS model to other media and communication concepts such as schemes, framing, and agenda-building. Researchers can also use the issue ownership network framework to predict public opinion in other communication contexts.
The new theory used is a breakthrough in enriching agenda-setting research. For example, the extended parallel process model, stakeholder theory, issue ownership theory, and community structure theory are some new theories that have emerged. Other theories McCombs (2005) noted as part of evolutionary theory, such as framing, the need for orientation is still essential in agenda-setting research (McCombs, 2005). Interestingly, network theory is related to the emergence of internet media by looking at media effects from an associative perspective. That is, considering whether the associations built can be transferred between agendas (Guo & Vargo, 2017).
The dominant trend of the methodology is to expand the agenda-setting study. This aligns with the increasing use of third-level agenda-setting theory and intermedia agenda-setting. This finding aligns with Su's study (2021); the application of the network agenda-setting model has increased in recent years with data analysis methods, such as time lag correlation and time series. Likewise, in the intermedia agenda-setting study, previous research confirms the flow from one traditional media to another. In contrast, more recent research reveals the flow from conventional media to new media and the reciprocal relationship (Su, 2021).
In agenda-setting research, methodology is more dominant in quantitative research (McCombs, 2018). For future studies, a qualitative approach is recommended. This approach can better understand different public agendas and reveal what sources journalists use in selecting issues and attributes to cover .
Content analysis is an option to determine important priority issues in agenda setting and surveys to measure the public agenda. The dominance of content analysis is related to the use of newspaper media. This is because most agenda-setting studies are textual in nature (Luqman, 2019). The central unit of analysis is text content (like text from news stories, tweets, and so on) or content translated into text (such as interviews and surveys). Meanwhile, visual content widely spread on social media and online media has received less attention. Photos uniquely highlight events or attributes related to issues and stories (Newton, 2013). This is because "visual agenda setting" seeks to overcome the influence of the news media in setting the visual agenda (Lough, 2018). Therefore, future research using visual data, for example, with a visual framing approach, will enrich the variety of agenda-setting research.
The results of the research findings show the evolution of theoretical trends, study focus, methodology, media, and the application of other theories in agenda-setting studies. Although the method is still dominant in using content analysis and surveys. For future studies, it is necessary to pay attention to the time lag to know the effectiveness of media effects. The idea of "time" has acquired a different connotation. With news websites, live blogs, and social media, news publication highly depends on a fixed schedule (Karlsson & Strömbäck, 2010). News also need not be communicated as a "finished" product (Vonbun et al., 2016). In practice, journalists can present news snippets as they occur via social media channels, such as Twitter, and then follow up for more indepth reports with full articles on their media websites. The use of content analysis on the effects of the media agenda will be more challenging if we pay attention to the comparison of social media with online media or mainstream media, such as newspapers or television.
In reviewing the literature with meta-analysis, there are several limitations. The sample used in this research is limited to Scopus-indexed articles. Even though several agenda-setting studies are not affiliated with Scopus indexers, they are worth considering and can provide descriptions and contributions to agenda-setting studies in the future.