Next Article in Journal
Community-Aware Two-Stage Diversification for Social Media User Recommendation with Graph Neural Networks
Previous Article in Journal
Oversampling Algorithm Based on Improved K-Means and Gaussian Distribution
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Information
Volume 17
Issue 1
10.3390/info17010026
information-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Mapping Fake News Research in Digital Media: A Bibliometric and Topic Modeling Analysis of Global Trends

by
Yuh-Shan Ho
1,
Fatma Yardibi
2,*,
Murat Ertan Dogan
2 and
Huseyin Kusetogullari
3
1
CT HO Trend, 3F.-7, No. 1, Fuxing N. Rd., Songshan Dist., Taipei City 105611, Taiwan
2
Department of New Media and Communication, Faculty of Communication, Akdeniz University, Campus Konyaaltı, Pınarbaşı Mah., Dumlupınar Boulevard, 07070 Antalya, Turkey
3
Department of Computer Science, Blekinge Institute of Technology, 37141 Karlskrona, Sweden
*
Author to whom correspondence should be addressed.
Information 2026, 17(1), 26; https://doi.org/10.3390/info17010026
Submission received: 18 November 2025 / Revised: 20 December 2025 / Accepted: 23 December 2025 / Published: 31 December 2025
(This article belongs to the Section Information Processes)
Browse Figures
Versions Notes

Abstract

This study aims to identify research trends in communication regarding the phenomenon of “fake news” in digital media. Fake news has become a rapidly growing and significant area of research in communication studies in recent years. Published studies were collected from the Science Citation Index Expanded database. The analysis included the annual distribution of publications, citation metrics, leading journals, countries, institutions, and authors. To explore the conceptual structure, topic modeling was conducted using text mining techniques along with DBSCAN and k-means clustering methods. The United States is a leader in the field, both as a producer country and in terms of technology implementation. Vraga, Bode, Tully, Hameleers, and Tandoc are among the most influential authors. The most cited studies specifically focus on misinformation in the health sector and political disinformation and manipulation during elections. Topic modeling analyses show that the literature mainly clusters around health disinformation, political communication, and verification technologies. The findings have important implications for communication policies, media literacy, and fact-checking technologies. Research that systematically examines fake news from a communication perspective, both in performance and conceptual structures, is scarce in the literature. The resulting thematic clusters provide valuable insights for future research.

1. Introduction

New media refers to digital communication technologies of the 21st century that allow for the creation and sharing of content as digital data. Platforms such as social media, websites, and online forums have overtaken traditional media as the main sources of information [1]. Digital media is transforming how information is created and shared. This change involves many aspects that enhance modern life. Social media platforms, in particular, are seen as tools that enable communication and interaction by offering easy access to information [2]. The rapid growth of digital technologies has sped up information sharing but has also created new problems. Social media, blogs, and news sites make it easier to access information while also being important sources of disinformation [3]. Before social media became popular, false information spread slowly and was limited to certain media outlets. While it could still have a strong effect, the lack of instant sharing sets that time apart from today, showing how technology and digital media have changed the way fake news spreads [4]. The definition of fake news remains contested and has evolved over time. Initially associated with comedy and satire, it is now linked to political manipulation and disinformation [5]. The term refers to the intentional dissemination of misleading content, often serving political, economic, or social agendas [6]. This phenomenon has the potential to influence public opinion and compromise the integrity of democratic processes [7]. The impact of disinformation on democratic institutions goes beyond individual information control.
Researchers like [8] have stated that such news stories help to strengthen misinformation environments and cause greater polarization in public opinion [8,9]. The complexity and impact of fake news have made it a prominent topic in academic research, and several bibliometric studies have been conducted on the subject. Missau’s (2024) bibliometric study on misinformation, disinformation, and fake news analyzed data from Scopus and WoS separately, rather than merging them as is common in the literature [10]. This approach is useful for comparing databases but may limit generalizability. The study focused solely on fake news in terms of scope. Akram et al. (2022) [11] conducted a systematic review of 264 publications from the WoS database using bibliometric methods. Focusing on cognitive abilities and the relationship between social media and fake news, the study found that political interests play a key role in spreading systematic disinformation to influence cognition, build political support networks, and reinforce narratives [11]. Arias et al. (2023) reviewed 46 studies in Ibero-American journals focusing on scholarly communication studies on post-truth and fake news [12]. Pandey and Gosh conducted keyword co-occurrence analyses for 1364 studies in their literature review by associating fake news only with social media rumors [4]. While Vijayan et al. [13] mentioned they conducted a search using the keywords “fake news,” “misinformation,” “media,” and “journalism” in the Scopus database, the exact Boolean structure of the query and any potential language or domain restrictions are not clearly provided in the article. The study can be broadly considered to address and evaluate the phenomenon of fake news within the context of journalism [13]. While these studies generally share a common purpose, they differ in terms of data sources, scope, and the subfields of focus. While studies on fake news have typically addressed the issue within a social and political context, the role of digital media in the formation and propagation of fake news has been examined in a more limited manner [6,14,15].
Bibliometric analysis is a powerful method for identifying research trends and knowledge gaps by examining the scope, trends, and impact of publications on a specific topic [16]. The primary original contribution of our study is that it surpasses the thematic mapping offered by standard bibliometric techniques and integrates topic modeling with bibliometric analysis through advanced computational algorithms to improve structural analysis. Thus, by providing an in-depth analysis of the relationship between digital media and fake news, we aim to establish a foundation for future research.

2. Methodology

The data for this study were obtained from the Clarivate Analytics Web of Science Core Collection (WoS Core), specifically the online version of the Social Sciences Citation Index (SSCI), with the most recent update as of 21 December 2024. In 2023, the Journal Citation Reports (JCR) indexed 3541 journals that included citation references across 58 Web of Science categories within SSCI. Of them, 96 journals were classified in the Web of Science category of “communication”. The search employed quotation marks (“ ”) and the Boolean operator “OR” to ensure that at least one search keyword appeared in the Topic (TS) fields, which included the title, abstract, author keywords, and Keywords Plus, in the Web of Science category of communication, covering the period from 1992 to 2023. The keywords that constituted the selection criteria were identified through a literature review. The search keywords related to fake news (“fake news,” “disinformation,” “false news,” “false information,” “unverified information,” and “misinformation”) and digital media (“new media,” “digital media,” “social media,” and “digital medium”) were used. In total, 812 documents containing the specified search keywords in the Topic (TS) were identified, of which 743 documents (92% of 812 documents) were published between 2011 and 2023. Complete records from SSCI, along with the number of citations for each document per year, were verified and downloaded into Excel, where additional coding was conducted. The impact factors (IF2023) of the journals were sourced from the Journal Citation Reports (JCR) published in 2023.
The concept of “front page” (title, abstract, and author keywords) was used as a filtering mechanism to improve Topic (TS) search strategies in the WoS Core for bibliometric analysis [17]. This filter effectively minimizes the inclusion of unrelated publications in bibliometric analyses. Significant discrepancies were observed when applying the “front page” filter in bibliometric research on topics published in SSCI. For instance, the filter identified a 12% deviation in studies on evolutionary economic geography [18] and 28% in research on medication-adherence. Ultimately, 625 documents (84% of 743 documents) that contained search keywords in their “front page” were classified as research publications on fake news in digital media. This study performed additional categorization based on relevant authors, institutions, and countries to obtain more precise and accurate results in the analysis of scientific research. While the SSCI database designates the reprint author as the corresponding author, we opted to use the term “corresponding author” instead. For articles with multiple corresponding authors, each author, institution, and country was counted separately. Records with missing affiliation names in SSCI were checked and updated. Affiliations from England, Scotland, and Wales were grouped under the United Kingdom, and those from Turkiye were standardized as Turkey. Institutional names were also unified (e.g., University of Wisconsin with University of Wisconsin–Madison, Nanyang Technological University with its Singapore variant, and University of Minnesota with Twin Cities). The evaluation of publications in this study was conducted using three citation indicators:
Cyear: citations received in a specific year (e.g., C2023) [19].
TCyear: total citations from publication year to the end of 2023 (e.g., TC2023) [20].
CPPyear: average citations per publication, calculated as CPP2023 = TC2023/TP, where TP denotes the total number of publications.
CPPyear can be utilized across various contexts, including but not limited to different document types, languages, Web of Science categories, journals, countries, institutions, and authors. The advantage of using TC2023 and CPP2023, compared with the conventional “total citations” metric from the Web of Science Core Collection, is that these indicators maintain theoretical invariance and ensure repeatability.
Six publication indicators were used to evaluate the publication performance of countries and institutions: total publications (TPs), single-country or single-institution publications (IPs), internationally or inter-institutionally collaborative publications (CPs), first-author publications (FPs), corresponding-author publications (RPs), and single-author publications (SPs) [21].
To further explore the study’s conceptual structure, topic modeling was applied to the authors’ keywords of the selected publications, in conjunction with traditional bibliometric analysis, to uncover latent thematic patterns. This study employed DBSCAN followed by k-means to enhance the robustness of clusters and improve the quality of topic separation. Unlike probabilistic topic modeling techniques such as LDA or HDP, which depend on Dirichlet priors [22] and often require a predetermined number of topics, DBSCAN is a density-based method that can detect noise [23], outliers, and clusters of varying shapes. This feature enabled us to detect naturally dense areas of keywords without the need to cluster all data points. To achieve more understandable thematic groups, the primary results from DBSCAN were enhanced with k-means, which optimizes separation based on centroids and enhances the uniformity within clusters. By combining both methods, this strategy successfully merges DBSCAN’s sensitivity to density with k-means’ ability to partition globally, leading to clusters that are more stable and semantically coherent in high-dimensional text data.
k-means clustering was performed with k set to 5 clusters, chosen iteratively for semantic coherence and consistent visualization. The model was executed as k-means (n_clusters = 5, random_state = 42). DBSCAN was utilized with parameters eps = 0.3 and min_samples = 5. These settings were determined empirically by examining cluster separation in the t-SNE embedding space and adjusting to prevent excessive fragmentation while maintaining significant dense areas. DBSCAN was applied as DBSCAN (eps = 0.3, min_samples = 5).

3. Results and Discussion

3.1. Characteristics of Document Types

The average number of citations per publication in a document type (CPPyear) and the average number of authors per publication (APP) have been proposed as basic metrics to determine the characteristics of document types in a research field [24]. A total of 625 fake news in digital media-related documents were published in SSCI between 2011 and 2023, spanning seven document types, as detailed in Table 1. Of these, 600 were articles, comprising 92% of the total documents, with an average of 2.7 authors per article. Among the document types, “editorial materials” (20 documents) had the highest CPP2023, averaging 22 citations per publication. The CPP2023 for reviews was found to be 19 citations per publication, which was lower than the 21 citations for articles with a CPP2023 [25]. The contributions of different document types can vary significantly. Articles, typically structured with sections such as introduction, methods, results, discussion, and conclusion, are often recommended as the primary focus for bibliometric research in a specific field. In the case of fake news in digital media, a total of 600 articles were published in three different languages. English was the dominant language, with 580 articles (97% of the total), followed by Spanish (19 articles) and Dutch (1 article).

3.2. Characteristics of Publication Outputs

To understand the development trends and impacts of publications in a specific research topic, the annual number of articles (TPs) and their average number of citations per publication (CPPyear) can be analyzed year by year [26]. From 1991 to 2023, a total of 600 articles related to fake news in digital media were indexed in the Web of Science category of “communication” in the SSCI database. The CPP2023 for these articles was 21 citations per article, with the highest individual article receiving 531 citations by Bennett and Livingston (2018) [14]. The first article related to fake news in digital media in the Web of Science category of communication was published in 2011, entitled “Online authority communication during an epidemic: A Finnish example” by Tirkkonen and Luoma-aho (2011) [27]. Figure 1 illustrates that the number of fake news in digital media-related articles rose sharply from seven articles in 2017 to 166 in 2023. The years 2015 exhibited the highest CPP2023, with 198 citations per article, largely due to the only article in 2015 entitled “In related news, that was wrong: The correction of misinformation through related stories functionality in social media” [28] with a TC2023 of 298 citations, ranking 4th in this field. Leticia Bode and Emily K. Vraga’s study, “ In related news, that was wrong: The correction of misinformation through related stories functionality in social media” drew attention for its innovative approach to combating misinformation on social media platforms [28]. This study was a pioneering empirical study of how social media use contributes to the reinforcement or correction of misinformation, and it quickly became a focus of attention.

3.3. Journals

In 2021, the characteristics of journals based on their CPPyear (average number of citations per publication in a journal) and APP (average number of authors per publication in a journal) were reported as basic metrics for understanding journal performance within a specific research topic [29]. This approach has been applied in journals in the SSCI database [30]. Table 2 highlights the top 10 most productive journals, with their impact factors (IF2023), CPP2023, and APP. The Social Media + Society journal (IF2023 = 5.5) published the highest number of articles, with 70 articles, accounting for 12% of the 600 articles-related to fake news in digital media. A comparison of the top 10 most productive journals reveals notable differences in citation impact and authorship patterns. Twenty-one articles published in Health Communication (IF2023 = 3.0) had the highest CPP2023, with an average of 43 citations per article. In contrast, 52 articles in the International Journal of Communication (IF2023 = 1.9) averaged only 6.8 citations. The average number of authors per publication (APP) varied as well, ranging from 3.7 authors per article in Health Communication to 2.4 authors in Information Communication & Society. In terms of journal impact factor (IF2023), the Journal of Communication, with an IF2023 of 6.1, ranked 2nd in the Web of Science category of communication, published eight articles.

3.4. Publication Performances: Countries and Institutions

It is widely recognized that two authors, the first and the corresponding authors, are considered the most contributed authors in a research article [31]. A total of 600 articles were published by authors affiliated from 60 countries including 454 single-country articles (76% of 600 articles) published by authors from 47 countries with a CPP2023 of 21 citations per publication and 146 internationally collaborative articles (24%) published by authors from 52 countries with a CPP2023 of 19 citations. The results demonstrated that international collaboration slightly decreased citations in the fake news in digital media research.
Six publication indicators and the six related citation indicators (CPP2023) [32] were applied to compare the top 11 productive countries with 13 articles or more, including five in Europe, two in America, two in Asia, and one in Oceania and Africa, respectively (Table 3).
The USA dominated in the six publication indicators with a TP of 299 articles (38% of 600 articles), an IPC of 159 articles (35% of 454 single-country articles), a CPC of 70 articles (48% of 146 internationally collaborative articles), an FP of 194 articles (32% of 600 first-author articles), an RP of 194 articles (32% of 600 corresponding-author articles), and an SP of 48 articles (34% of 140 single-author articles). It was not surprising that the USA dominated the six publication indicators in research topics in the SSCI; for example, the role of social media platforms (particularly Facebook and Twitter) in spreading disinformation during the 2016 US presidential election is an important turning point in the debate on digital media and fake news [9,33]. In addition, the USA has been a center where digital media platforms have not only been produced but also culturally disseminated and have naturally been the focus of studies in terms of both case studies and research.
Compared to the top 11 productive countries in Table 3, the USA also had the highest CPP2023 of 29 and 32 citations per publication for total articles and single-country articles, respectively. Singapore, with a CPC of 15 articles, an FP of 23 articles, and an RP of 24 articles, had the highest CPP2023 of 26, 30, and 30 citations per publication for internationally collaborative articles, first-author articles, and corresponding-author articles, respectively. Spain, with an SP of 13 articles, had the highest CPP2023 of 31 citations per publication for single-author articles. Belgium had a lower CPP2023 for all six types of articles. Development trends in the publication of the top five productive countries are presented in Figure 2. The USA ranked at the top from 2017 to 2023. China published its first article in 2020 and reached 33 articles, ranking fourth in 2023. Spain increased in the last two years to reach 24 articles, ranking second in 2023.
The USA began conducting its first academic studies on digital media platforms in 2014 and has maintained its leading position in this field since then. 2014 stands out as the year when digital media platforms such as Facebook and Twitter became widespread on a global scale [34,35]. At the institutional level, the institution where the corresponding author is identified may be the home base of the study or the origin of the paper [19]. Concerning institutions, 307 fake news in digital media articles (51% of 600 articles) originated from single institutions with a CPP2023 of 18 citations per publication, while 293 articles (49%) were institutional collaborations with a CPP2023 of 24 citations. The results showed that institutional collaborations increased citations in fake news research in digital media. Table 4 highlights the top 10 most productive institutions, along with their publication characteristics. Among these, five were in the USA, two in the UK, while Australia, the Netherlands, and Singapore were each represented by one institution. The University of Oxford in the UK (U Oxford) ranked the top in two publication indicators with a TP of 24 articles (4.0% of 600 articles) and a CPI of 20 articles (6.8% of 293 inter-institutionally collaborative articles). The University of Amsterdam in Netherlands (U Amsterdam) ranked the top in three publication indicators with an IPI of 13 articles (4.2% of 307 single-institution articles), an FP of 18 articles (3.0% of 600 first-author articles), an RP of 19 articles (3.2% of 600 corresponding-author articles), and an SP of four articles (4.9% of 140 single-author articles). Compared to the top 10 productive institutes in Table 4, the Georgetown University in the USA (Georgetown U) with a TP of 12 articles, a CPI of 12 articles, an FP of three articles, and an RP of three articles, had the greatest CPP2023 of 81, 81, 126, and 126 citations per publications for total articles, inter-institutionally collaborative articles, first-author articles, and corresponding-author articles,, respectively. The University of Minnesota in the USA with an IPI of one article and an SP of one article, had the greatest CPP2023 of 86 citations per publication for single-institution articles and single-author articles respectively.

3.5. Publication Performances: Authors

For articles related to fake news in digital media, the average number of authors per publication (APP) was 2.7 authors, with the maximum reaching 20 authors in a single article by Chadwick et al. [36] Among the 600 articles related to fake news in digital media in the Web of Science category of communication in the SSCI, 80% were authored by groups of one to three individuals. Specifically, 181 articles (30%) were written by two authors, 158 articles (26%) by three authors, and 140 articles (23%) by one author. Table 5 lists the top nine productive authors with more than five articles, presenting their publication and citation indicators, along with their Y-index constants [37].
E.K. Vraga was the most prolific author, with 14 fake news articles in digital media-related articles, including seven first-author articles (ranked second), and seven corresponding-author articles (ranked second), without single-author articles, followed by L. Bode (12 articles), M. Tully (11 articles), and M. Hameleers (10 articles). M. Hameleers with an FP of 10 articles, an RP of 10 articles, and an SP of three articles, ranked at the top in first-author articles, corresponding-author articles, single-author articles respectively (see Table 5).
Among the top nine authors, L. Bode with a TP of 12 articles, an FP of 4 articles, and an RP of 4 articles, had the highest CPP2023 of 106, 169, and 169 citations per publication for total articles, first-author articles, and corresponding-author articles, respectively. J. Lukito with an SP of one article had the highest CPP2023 of 50 citations per publication for single-author articles.
Six of the top nine authors, including M. Hameleers, E.K. Vraga, E.C. Tandoc, J. Lee, L. Bode, and M. Tully, were identified as having also the top nine high publication potential based on the Y-index.
A total of 600 fake news in digital media-related articles were analyzed using the Y-index. These 600 articles were contributed by 1298 authors. Of these, 732 authors (56%) had neither first-author nor corresponding-author articles, resulting in a Y-index of (0, 0). In contrast, 56 authors (4.3%) published only corresponding-author articles with h = π/2 (RP > 0), 12 authors (0.92%) published more corresponding-author than first-author articles π/2 > h > π/4 (FP > 0), 436 authors (34%) published an equal number of first- and corresponding-author articles h = π/4 (FP > 0 and RP > 0), eight authors (0.62%) published more first-author articles than corresponding-author articles π/4 > h > 0 (RP > 0), and 54 authors (4.2%) published only first-author articles with h = 0 (FP > 0).
In the polar coordinates (Figure 3), the Y-index (j, h) distribution of the top 61 authors in fake news in digital media research, with j ≥ 0, was demonstrated. Each point represents a single author or a group of authors. For example, C. Vaccari, A. Bruns, M. Bastos, and S. Valenzuela shared the same Y-index of (5, 0.9828). M. Hameleers with a Y-index of (20, π/4), exhibited the highest publication potential in fake news in digital media-related research. Hameleers, who authored the most first-author articles, corresponding-author articles, and single-author articles, indicating his focus not only on supervising but also active to perform fake news in digital media research. E.K. Vraga (14, π/4) and E.C. Tandoc (13, 0.8622) also had high publication potential in the field. Emily Vraga from the University of Minnesota, Michael Hameleers from the University of Amsterdam, and Edson Tandoc from the Nanyang Technological University are academics who have made significant contributions to the field with their work examining the spread of fake news and disinformation on social media, analyzing the social impacts of these processes, and proposing strategies for correcting misinformation.
Vraga has conducted studies combining theory and practice on how to effectively correct misinformation on social media [28]. She has made significant contributions with her experimental studies on the importance of misinformation and correcting misperceptions in health issues [38,39]. The close collaboration between Emily Vraga and Leticia Bode has propelled both authors to the pinnacle of the most prominent authors in the field of digital media and fake news.
Hameleers has made significant contributions to the development of the field with his pioneering work on visual and textual disinformation as well as the interaction between social media and democracy [40,41]. His work bridges the applied and theoretical by examining not only the effects of disinformation but also strategies to reduce these effects [42]. He sheds light on both the academic and practical aspects of these issues by analyzing the relationship between media literacy, trust and disinformation in detail [41].
Edson Tandoc’s significant contributions to the field of fake news research include his examination of the dissemination of disinformation on social media, addressing this phenomenon at the individual and platform levels [6,43]. He has enriched the existing literature, especially on topics such as the definition of fake news, social media users’ reactions to fake news, and the social functions of fake news [3,43].
A. Bechmann, C.M. Lunga, K.P. Chen, L.M. Romero-Rodríguez, M.J. Riedl, P. Borah, and P. Malhotra (3, 1.107), C. Pérez-Curiel, I. Freiling, J. Farkas, J. Hodson, K. Koc-Michalska, O. Vinhas, and S. Bradshaw (3, 0.4636), and S. Altay (3, 0) shared the same Y-index of j of 3, signifying equal publication potential but with different publication. All of the 15 authors located on the same curve (j = 3) in Figure 3. Bechmann and other six authors published more corresponding-author articles than first-author articles with an h of 1.107. Pérez-Curiel and the other six with an h of 0.4636 published more first-author articles than corresponding-author articles with an h of 0.4636. Altay published only first-author articles with an h of 0. This indicates that Altay is still active in the field of fake news in digital media research then Pérez-Curiel and the other six.
A total of 40 authors, including M. Hameleers (20, π/4), E.K. Vraga (14, π/4), L. Bode and other three authors (8, π/4), A. Duffy and other six authors (6, π/4), and A. Chadwick and other twenty-six authors (4, π/4) shared the same publication characteristics, represented by a diagonal line in the Y-index graph. Hameleers led with the highest publication potential with a j of 20, followed by Vraga with a j of 14, Bode et al. with a j of 8, Duffy et al. with a j of 6, and Chadwick et al. with a j of 4.
The position on the graph, whether along one of the curves or a line from the origin, represents distinct groups of authors with varying publication potential or characteristics. However, potential biases, such as authors sharing the same name or using different names over time, might affect the analysis [44].

3.6. The Top Ten Most Frequently Cited Articles in Fake News in Digital Media Research

Total citations in the WoS Core are updated periodically. To minimize bias, the total number of citations from the publication year through the most recent year (TCyear) was taken directly from the database. However, a highly cited article does not always have consistently high annual citations; therefore, examining citation histories is important [26]. It is necessary to understand citation history of a highly cited article. The citation histories of the top ten most frequently cited fake news in digital media articles are shown in Figure 4 (see Supplementary Table S1 for details).
Bode and Vraga’s seminal 2015 [28] study signifies a paradigm shift in the realm of misinformation correction strategies, underscoring the efficacy of such methodologies. This pioneering study illuminates the potential of social media platforms to provide users with the tools necessary to rectify misinformation. This experimental study, which aims to combine these two areas by considering the role that social media can play in correcting misinformation, is the longest-term, as can be seen in Figure 4, where the impact continues to this day. As demonstrated in Figure 4, Bennett and Livingston’s “The Disinformation Order: Disruptive Communication and the Decline of Democratic Institutions” has continued to rise rapidly since its publication in 2018 [14].
The seven of the top ten most frequently cited articles were also ranked the top ten in the most impactful in the most recent year of 2023, which were summarized as follows:
  • Bennett and Livingston’s “The disinformation order: Disruptive communication and the decline of democratic institutions” (TC2023 = 531; C2023 = 125) stands as the most influential publication in this field [14]. The study offers a comprehensive framework for understanding how disinformation undermines democratic institutions, linking theoretical insights with real-world cases such as the Trump presidency and the Brexit campaign.
  • Bode and Vraga’s “See something, say something: Correction of global health misinformation on social media” (TC2023 = 320; C2023 = 73) ranked second in total and third in annual citations [38]. Rather than critiquing social media’s role in spreading health misinformation, the study demonstrated its corrective potential through a simulated Facebook News Feed experiment that incorporated both algorithmic and user-generated interventions. The findings highlighted the importance of social media campaigns in debunking false or misleading health information and providing sources that support accurate facts. It has maintained its importance since its publication in 2018, and the number of citations, which peaked especially during the global pandemic, has continued to rise to date. (see Figure 4).
  • Casero-Ripollés’ “Impact of COVID-19 on the media system. Communicative and democratic consequences of news consumption during the outbreak” (TC2023 = 310; C2023 = 74) ranked third in total and second in annual citations [45]. The study explored how the pandemic transformed news consumption, media credibility, and citizens’ ability to detect fake news, highlighting its broader social and communicative impacts. Although citation activity has slightly decreased after the pandemic, the article remains one of the most influential works in the field (see Figure 4).
  • Bode and Vraga’s “In related news, that was wrong: The correction of misinformation through related stories functionality in social media” (TC2023 = 298; C2023 = 53) ranked fourth in total citations and sixth in annual citations [28]. This seminal 2015 study marked a paradigm shift in misinformation correction, emphasizing the potential of social media platforms to empower users in addressing false information. Its innovative approach has ensured continued relevance, keeping it among the top ten most influential works in the field.
  • Vargo, Guo, and Amazeen’s “The agenda-setting power of fake news: A big data analysis of the online media landscape from 2014 to 2016” (TC2023 = 271; C2023 = 46) ranked fifth in total citations [15]. Analyzing online media from 2014 to 2016, the study examined the rising impact of fake news during the U.S. presidential elections, emphasizing its societal and political effects. The effects of fake news on society and the consequences of these effects have garnered significant attention not only in academic circles but also in the media, politics, and the public.
  • Vraga and Bode’s “Using expert sources to correct health misinformation in social media” (TC2023 = 222; C2023 = 56) ranked sixth in total and fifth in annual citations [39]. This pre-pandemic experimental study highlighted the importance of expert sources in correcting health misinformation on digital platforms. Its relevance increased during the COVID-19 outbreak, emphasizing its significance in health communication literature.
  • Vaccari and Chadwick’s “Deepfakes and disinformation: Exploring the impact of synthetic political video on deception, uncertainty, and trust in news” (TC2023 = 197; C2023 = 71) ranked seventh in total and fourth in annual citations [46]. This pioneering study investigated how synthetic political videos influence deception, uncertainty, and trust in news. Unlike pandemic-related misinformation studies, it demonstrated sustained and growing academic interest in deepfake technologies, establishing a solid foundation for future research on the evolving dynamics of political communication (see Figure 4). Deepfakes have such high potential to pose serious threats such as public opinion manipulation, geopolitical tensions, chaos in financial markets, fraud, slander, and identity theft that studies in this area will continue to attract attention among academics [47].

3.7. Research Foci and Topic Modeling

In recent years, innovative approaches have been developed to analyze word distributions in article titles, abstracts, author keywords, and Keywords Plus to reveal research focuses and trends [17]. These keywords were used to build a word bank that identified major research areas and their development. After receiving the dataset, pre-processing was performed to clean and organize the text for effective topic modeling. This pipeline ensures the data is sanitized and formatted at each step to make it suitable for analysis. Detailed descriptions of each step are provided in Supplementary S2.
DBSCAN Clustering: DBSCAN is a density-focused clustering technique that excels in identifying clusters of various shapes and managing noise within the dataset. It clusters data points that are densely packed together while marking outliers in regions of low density.
Parameter Selection: The DBSCAN algorithm requires two key parameters: Epsilon (ε). This parameter defines the radius of the neighborhood around a data point. We determined the optimal ε.
MinPts: This parameter specifies the minimum number of data points required within the ε neighborhood for a point to be considered a core point. We set MinPts to [Value of MinPts] based on [Justification for MinPts, e.g., domain knowledge and the size of the dataset].
We applied DBSCAN to the [Data Representation, e.g., TF-IDF matrix] of the preprocessed text data to identify similar clusters [Entity being clustered, e.g., documents or reviews]. The algorithm identified core points, border points, and noise points, allowing us to distinguish between dense clusters and outliers. The resulting clusters were evaluated using [Evaluation Metrics, e.g., silhouette score or visual inspection] to assess the quality of the clustering.
k-means Clustering: k-means clustering is a centroid-based method that aims to partition data into k distinct clusters, where each data point belongs to the cluster with the nearest mean (centroid).
The main parameter in k-means clustering is k, representing the number of clusters. Using the same [data representation, e.g., TF-IDF matrix] as in the DBSCAN analysis, k-means iteratively assigns data points to the nearest centroid and updates the centroids until convergence. As illustrated in Figure 5, five distinct clusters were identified. To achieve a more comprehensive understanding of the data structure, both DBSCAN and k-means algorithms were applied. While DBSCAN detects clusters of arbitrary shapes and outliers, k-means efficiently partitions the data into k groups. The final clustering outcome of k-means is presented in Figure 6 (see Supplementary S3 for detailed topic clusters and keyword interpretations). The recurring keywords across the clusters illuminate important thematic connections in the study of media, information, and public discourse. The frequent occurrence of “digital,” “media,” and “information” in several clusters indicates a comprehensive focus on the influence of digital technologies on the distribution and perception of information. The terms “news” and “credibility” bring attention to concerns about the trustworthiness of information, particularly in the realms of journalism and social media. The convergence of “political,” “public,” and “election” highlights the role of digital media in shaping political conversations, public opinion, and polarization. Additionally, the repeated appearance of “cognitive” and “theory” in discussions about media and information credibility suggests an interest in understanding how individuals process and evaluate information. Although cluster 4 seems less connected to the others, its emphasis on “scientific” and “conflict” points to a distinct theme related to public trust in science and societal divisions. Collectively, these connections underscore the intertwined nature of digital media, information credibility, political engagement, and public perception in contemporary discourse.
The results of k-means clustering analyses revealed five discrete clusters. To facilitate the conceptualization of the keywords compiled in the research, a thematic topic was assigned to each cluster, based on its constituent keywords.

3.7.1. Topic 1: Digital Media and Audience Engagement

Supporting Words: digital, analysis, audience, practices, fact, sharing, mobile, comparative, news, information
This topic focuses on the ways in which audiences interact with and engage with news and information in digital settings. It involves examining digital platforms, audience behaviors, and the methods of consuming and sharing information. The focus on “fact” and “sharing” underscores the conflict between trustworthy information and the spread of misinformation. This aligns with the document’s overarching theme, which addresses the increase in fake news in digital media. The supporting terms indicate the transition from traditional news consumption to digital platforms, where audience engagement is pivotal in the distribution of both accurate and inaccurate information. The reference to “analysis” highlights the research methods employed to investigate these issues, such as the bibliometric analysis mentioned in the document.

3.7.2. Topic 2: Social and Political Implications of Media

Supporting Words: media, social, news, political, online, communication, journalism, public, digital, information
This topic explores the wider societal and political effects of media, especially in the context of the digital era. It focuses on how media shapes public dialogue, political communication, and journalism. The cluster underscores the interconnection between media, society, and politics, acknowledging the power of information and its potential for both beneficial and harmful outcomes. This topic is closely linked to the document’s examination of political misinformation and its influence on democratic processes. The supporting words directly address concerns about the manipulation of information for political advantage and the diminishing trust in traditional journalism. The document references studies that investigate the role of social media in political disinformation campaigns, which are encompassed within this topic.

3.7.3. Topic 3: Information Credibility and Cognitive Aspects

Supporting Words: information, data, model, theory, media, credibility, cognitive, right, network, anti.
This topic centers on assessing the credibility of information and the cognitive mechanisms involved in this evaluation. It covers theoretical frameworks, models, and data analysis related to how individuals perceive, interpret, and judge the trustworthiness of information. The terms “cognitive,” “theory,” and “model” imply an academic and psychological approach to understanding information processing. The mention of “anti” could indicate studies on resistance to fact-checking or the spread of counter-narratives. This is linked to the document’s discussion on correcting misinformation. Understanding how people evaluate credibility and the cognitive biases that influence their judgments is vital for developing effective strategies to combat the spread of false information. The research on misinformation correction strategies, as discussed in the document, often integrates cognitive and psychological theories.

3.7.4. Topic 4: Crisis Communication and Societal Challenges

Supporting Words: crisis, election, polarization, publics, research, COVID, electoral, people, 19, vaccine.
This topic centers on communication during times of crisis, particularly focusing on societal challenges such as elections, polarization, and the COVID-19 pandemic. It highlights the role of communication in shaping public opinion, addressing crises, and navigating complex social issues. The specific references to “COVID,” “vaccine,” and “election” pinpoint contemporary challenges that have been heavily influenced by the spread of misinformation. The document explicitly mentions the impact of misinformation in health crises, such as the COVID-19 pandemic, and political misinformation’s influence on elections. This topic directly reflects these concerns and aligns with research exploring the role of communication in managing crises and mitigating the negative effects of misinformation. The authors Vraga and Bode, mentioned in the document, have conducted research on correcting health misinformation, a key aspect of this topic.

3.7.5. Topic 5: Social Issues and Dynamics

Supporting Words: America, scientific, celebrity, class, disgust, poor, conflict, consensus, cure, cures.
This topic covers a range of social issues and dynamics, including social class, conflict, and scientific discourse. It suggests an exploration of how information and misinformation influence social perceptions, inequalities, and debates. The presence of words like “celebrity,” “disgust,” and “poor” indicates a focus on social phenomena and their representation in media. While this topic is broader than the others, it still connects to the document’s theme. Misinformation can exacerbate social divisions, influence perceptions of different social groups, and distort scientific consensus. The document implicitly addresses these issues by highlighting the impact of misinformation on public discourse and societal challenges.
Topic modeling has uncovered a complex structure in fake news research. Within the scope of these theme issues, the literature on fake news can be evaluated through studies conducted at the individual, societal, and institutional levels. At the individual level, studies categorized under information credibility and cognitive aspects as well as digital media and audience engagement, primarily examine how users perceive, process, and interact with information within digital environments. These studies investigate the cognitive processes that influence exposure to and acceptance of misinformation, as well as the development of trust and engagement behaviors. On a broader societal level, research areas like the social and political impacts of media and social issues and dynamics focus on how fake news affects public discourse, political division, and social unity. These themes examine how misinformation goes beyond individual beliefs to influence collective narratives, institutional discourse, and power structures. Finally, the crisis communication and societal challenges cluster includes governance-focused studies that address situations where misinformation significantly harms public health and safety, such as health crises, emergencies, or major social unrest.

4. Conclusions

The phenomenon of ‘fake news’ in digital media has emerged as a significant research topic within communication studies. It continues to develop by providing specific theoretical frameworks. The USA, one of the most influential countries in terms of fake news in digital media, holds a leading position both as a “game maker” that produces these technologies and as a “player” that implements them. Vraga, Bode, Tully, Hameleers, and Tandoc were the most influential authors who have made significant contributions to forming the general framework of the literature. The findings showed that the most cited research focused on health misinformation, political misinformation, and election manipulation. Correcting false health information with reliable sources and using social media as both a dissemination and corrective tool are crucial for effective public health communication, particularly during pandemics. The surge of fake news during the pandemic highlighted its critical role in crisis communication. Individuals, societies, and organizations-including governments and the WHO must develop strategies to ensure the dissemination of accurate information in such contexts. In crisis communication, it is vital to understand how quickly and widely information spreads through social media. Preventing the misuse of these platforms requires specific actions. Strengthening collaboration between digital media platforms and governments, along with strategies to increase public awareness, is crucial. Research that encourages the sharing of accurate information and supports the use of trustworthy tools helps build more resilient societies against misinformation. Efforts to improve the effectiveness of these strategies are expected to grow. Moreover, the study’s findings indicated that with the emergence of technologies such as deepfakes and artificial intelligence, research focusing on algorithms and strategies for detecting content accuracy will become increasingly significant. In particular, the production of political and policy content that can create a mass effect, closely approximating reality, with artificial intelligence, or the spread of disinformation content that creates a crisis environment, has the potential to drag societies into chaos.
The five themes identified through topic modeling show that fake news in digital media is a complex and interdisciplinary challenge for communication science. Fake news is not just about news accuracy; it affects a wide range of issues, from participation practices and political processes to cognitive psychology, crisis management, and social equality. Future studies are expected to explore both theoretical contributions, such as trust, the public sphere, and information processing, and applied contributions, including verification technologies and media literacy, from the perspective of communication science.
Future research might expand on these thematic patterns by explicitly grounding them in established communication theories, such as framing or agenda-setting, to promote deeper theoretical integration.
The study has some limitations. This study’s analysis only used the SSCI from Clarivate Analytics’ Web of Science Core Collection database, which limits the scope of the database. The WoS Core Collection tends to focus on English-language journals. This provides a global perspective but not a local perspective. For a local assessment, information from country-language databases should be used. Despite these disadvantages, the study provides insightful observations about the status and future directions of academic literature on fake news in digital media in the field of communication.

Supplementary Materials

The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/info17010026/s1. Supplementary Table S1: Top ten most frequently cited fake news in digital media articles; Supplementary S2: Data Preprocessing Procedures for Topic Modeling; Supplementary S3: Topic Modeling Clusters and Keyword Interpretations.

Author Contributions

Conceptualization, Y.-S.H. and F.Y.; methodology, Y.-S.H. and F.Y.; software, Y.-S.H., F.Y. and M.E.D.; validation, Y.-S.H., F.Y. and M.E.D.; formal analysis, F.Y.; investigation, Y.-S.H., F.Y., M.E.D. and H.K.; resources, F.Y. and M.E.D.; data curation, F.Y. and M.E.D.; writing—original draft preparation, Y.-S.H., F.Y., M.E.D. and H.K.; writing—review and editing, Y.-S.H., F.Y., M.E.D. and H.K.; visualization, F.Y., Y.-S.H. and M.E.D.; supervision, Y.-S.H.; project administration, Y.-S.H. and F.Y.; funding acquisition, H.K. All authors have read and agreed to the published version of the manuscript.

Funding

This research is funded by the Blekinge Institute of Technology.

Institutional Review Board Statement

Not applicable.

Informed Consent Statements

Not applicable.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request. All bibliometric data were obtained from Web of Science Core Collection, following its terms of use.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Johnson, R.M. Social Media and Free Speech: A Collision Course That Threatens Democracy. Ohio North. Univ. Law Rev. 2023, 49, 5. [Google Scholar]
  2. Jones, N. From Social Media to Social Ministry: A Guide to Digital Discipleship; Zondervan: Grand Rapids, MI, USA, 2020. [Google Scholar]
  3. Duffy, A.; Tandoc, E.; Ling, R. Too Good to Be True, Too Good Not to Share: The Social Utility of Fake News. Inf. Commun. Soc. 2020, 23, 1965–1979. [Google Scholar] [CrossRef]
  4. Pandey, S.; Ghosh, M. Bibliometric Review of Research on Misinformation: Reflective Analysis on the Future of Communication. J. Creat. Commun. 2023, 18, 149–165. [Google Scholar] [CrossRef]
  5. Rodríguez-Ferrándiz, R. An Overview of the Fake News Phenomenon: From Untruth-Driven to Post-Truth-Driven Approaches. Media Commun. 2023, 11, 15–29. [Google Scholar] [CrossRef]
  6. Tandoc, E.C.; Lim, D.; Ling, R. Diffusion of Disinformation: How Social Media Users Respond to Fake News and Why. Journalism 2020, 21, 381–398. [Google Scholar] [CrossRef]
  7. Lazer, D.M.J.; Baum, M.A.; Benkler, Y.; Berinsky, A.J.; Greenhill, K.M.; Menczer, F.; Metzger, M.J.; Nyhan, B.; Pennycook, G.; Rothschild, D.; et al. The Science of Fake News. Science 2018, 359, 1094–1096. [Google Scholar] [CrossRef]
  8. Wardle, C. Fake News. It’s Complicated. First Draft 2017, 16, 1–11. [Google Scholar]
  9. Allcott, H.; Gentzkow, M. Social Media and Fake News in the 2016 Election. J. Econ. Perspect. 2017, 31, 211–236. [Google Scholar] [CrossRef]
  10. Missau, L.D. The Role of Social Sciences in the Study of Misinformation: A Bibliometric Analysis of Web of Science and Scopus Publications (2017–2022). Tripodos 2024, 141–166. [Google Scholar]
  11. Akram, M.; Nasar, A.; Arshad-Ayaz, A. A Bibliometric Analysis of Disinformation through Social Media. Online J. Commun. Media Technol. 2022, 12, e202242. [Google Scholar] [CrossRef]
  12. Arias, D.; González Pardo, R.E.; Cortés Peña, O. Post-Truth and Fake News in Scientific Communication in Ibero-American Journals. Comun. Y Soc. 2023, 20. [Google Scholar] [CrossRef]
  13. Vijayan, V.; Thomas, T.; Nellanat, P.D. Mapping Fake News and Misinformation in Media: A Two-Decade Bibliometric Review of Emerging Trends. Insight-News Media 2025, 8, 734. [Google Scholar] [CrossRef]
  14. Bennett, W.L.; Livingston, S. The Disinformation Order: Disruptive Communication and the Decline of Democratic Institutions. Eur. J. Commun. 2018, 33, 122–139. [Google Scholar] [CrossRef]
  15. Vargo, C.J.; Guo, L.; Amazeen, M.A. The Agenda-Setting Power of Fake News: A Big Data Analysis of the Online Media Landscape from 2014 to 2016. New Media Soc. 2018, 20, 2028–2049. [Google Scholar] [CrossRef]
  16. Adalı, G.; Yardibi, F.; Aydın, Ş.; Güdekli, A.; Aksoy, E.; Hoştut, S. Gender and Advertising: A 50-Year Bibliometric Analysis. J. Advert. 2025, 54, 411–431. [Google Scholar] [CrossRef]
  17. Wang, C.-C.; Ho, Y.-S. Research Trend of Metal–Organic Frameworks: A Bibliometric Analysis. Scientometrics 2016, 109, 481–513. [Google Scholar] [CrossRef]
  18. Ho, Y.-S. Comments on the Reply to the Rebuttal to: Zhu, Jin, & He ‘On Evolutionary Economic Geography: A Literature Review Using Bibliometric Analysis’, European Planning Studies Vol. 27, Pp 639–660. Eur. Plan. Stud. 2019, 27, 1241–1243. [Google Scholar] [CrossRef]
  19. Ho, Y.-S. Top-Cited Articles in Chemical Engineering in Science Citation Index Expanded: A Bibliometric Analysis. Chin. J. Chem. Eng. 2012, 20, 478–488. [Google Scholar] [CrossRef]
  20. Wang, H.M.; Hui, Z.F.; Yuh, S.H. Comparison of universities’scientific performance using bibliometric indicators. Malays. J. Libr. Inf. Sci. 2011, 16, 1–19. [Google Scholar]
  21. Ho, Y.; Kahn, M. A Bibliometric Study of Highly Cited Reviews in the Science Citation Index Expanded TM. Asso. Info. Sci. Technol. 2014, 65, 372–385. [Google Scholar] [CrossRef]
  22. Scientific, L.L. Advancements in Dynamic Topic Modeling: A Comparative Analysis of LDA, DTM, GibbsLDA++, HDP and Proposed Hybrid Model HDP with CT-DTM for Real-Time and Evolving Textual Data. J. Theor. Appl. Inf. Technol. 2024, 102, 4344–4360. [Google Scholar]
  23. Deng, D. DBSCAN Clustering Algorithm Based on Density. In Proceedings of the 2020 7th International Forum on Electrical Engineering and Automation (IFEEA), Hefei, China, 25–27 September 2020; IEEE: New York, NY, USA, 2020; pp. 949–953. [Google Scholar]
  24. Monge-Nájera, J.; Ho, Y.-S. El Salvador Publications in the Science Citation Index Expanded: Subjects, Authorship, Collaboration and Citation Patterns. Rev. Biol. Trop. 2017, 65, 1428–1436. [Google Scholar] [CrossRef]
  25. Usman, M.; Ho, Y.-S. A Bibliometric Study of the Fenton Oxidation for Soil and Water Remediation. J. Environ. Manag. 2020, 270, 110886. [Google Scholar] [CrossRef] [PubMed]
  26. Ho, Y.-S. The Top-Cited Research Works in the Science Citation Index Expanded. Scientometrics 2013, 94, 1297–1312. [Google Scholar] [CrossRef]
  27. Tirkkonen, P.; Luoma-aho, V. Online Authority Communication during an Epidemic: A Finnish Example. Public Relat. Rev. 2011, 37, 172–174. [Google Scholar] [CrossRef]
  28. Bode, L.; Vraga, E.K. In Related News, That Was Wrong: The Correction of Misinformation through Related Stories Functionality in Social Media. J. Commun. 2015, 65, 619–638. [Google Scholar] [CrossRef]
  29. Ho, Y.-S. A Bibliometric Analysis of Highly Cited Publications in Web of Science Category of Emergency Medicine. Signa Vitae 2021, 17, 11–19. [Google Scholar]
  30. Omega, P.D.; Parung, J.; Yuwanto, L.; Ho, Y.-S. A Bibliometric Analysis of Positive Mental Health Research and Development in the Social Science Citation Index. Int. J. Ment. Health Promot. (IJMHP) 2024, 26, 817–836. [Google Scholar] [CrossRef]
  31. Riesenberg, D.; Lundberg, G.D. The Order of Authorship: Who’s on First? JAMA 1990, 264, 1857. [Google Scholar] [CrossRef] [PubMed]
  32. Ho, Y.-S.; Mukul, S.A. Publication Performance and Trends in Mangrove Forests: A Bibliometric Analysis. Sustainability 2021, 13, 12532. [Google Scholar] [CrossRef]
  33. Grinberg, N.; Joseph, K.; Friedland, L.; Swire-Thompson, B.; Lazer, D. Fake News on Twitter during the 2016 U.S. Presidential Election. Science 2019, 363, 374–378. [Google Scholar] [CrossRef]
  34. Duggan, M.; Ellison, N.B.; Lampe, C.; Lenhart, A.; Madden, M. Social Media Update 2014. Pew Res. Cent. 2015, 19, 1–2. [Google Scholar]
  35. Paul, T.; Puscher, D.; Strufe, T. The User Behavior in Facebook and Its Development from 2009 until 2014. arXiv 2015, arXiv:1505.04943. [Google Scholar] [CrossRef]
  36. Chadwick, A.; Kaiser, J.; Vaccari, C.; Freeman, D.; Lambe, S.; Loe, B.S.; Vanderslott, S.; Lewandowsky, S.; Conroy, M.; Ross, A.R.N.; et al. Online social endorsement and Covid-19 vaccine hesitancy in the United Kingdom. Social Soc. Media + Soc. 2021, 7, 20563051211008817. [Google Scholar] [CrossRef]
  37. Ho, Y.-S.; Fülöp, T.; Krisanapan, P.; Soliman, K.M.; Cheungpasitporn, W. Artificial Intelligence and Machine Learning Trends in Kidney Care. Am. J. Med. Sci. 2024, 367, 281–295. [Google Scholar] [CrossRef]
  38. Bode, L.; Vraga, E.K. See Something, Say Something: Correction of Global Health Misinformation on Social Media. Health Commun. 2018, 33, 1131–1140. [Google Scholar] [CrossRef] [PubMed]
  39. Vraga, E.K.; Bode, L. Using Expert Sources to Correct Health Misinformation in Social Media. Sci. Commun. 2017, 39, 621–645. [Google Scholar] [CrossRef]
  40. Hameleers, M.; Powell, T.E.; Van Der Meer, T.G.L.A.; Bos, L. A Picture Paints a Thousand Lies? The Effects and Mechanisms of Multimodal Disinformation and Rebuttals Disseminated via Social Media. Political Commun. 2020, 37, 281–301. [Google Scholar] [CrossRef]
  41. Hameleers, M.; Brosius, A.; De Vreese, C.H. Whom to Trust? Media Exposure Patterns of Citizens with Perceptions of Misinformation and Disinformation Related to the News Media. Eur. J. Commun. 2022, 37, 237–268. [Google Scholar] [CrossRef]
  42. Dan, V.; Paris, B.; Donovan, J.; Hameleers, M.; Roozenbeek, J.; Van Der Linden, S.; Von Sikorski, C. Visual Mis- and Disinformation, Social Media, and Democracy. J. Mass Commun. Q. 2021, 98, 641–664. [Google Scholar] [CrossRef]
  43. Tandoc, E.C.; Lim, Z.W.; Ling, R. Defining “Fake News”: A Typology of Scholarly Definitions. Digit. Journal. 2018, 6, 137–153. [Google Scholar] [CrossRef]
  44. Chiu, W.-T.; Ho, Y.-S. Bibliometric Analysis of Tsunami Research. Scientometrics 2007, 73, 3–17. [Google Scholar] [CrossRef]
  45. Casero-Ripollés, A. Impact of COVID-19 on the Media System. Communicative and Democratic Consequences of News Consumption during the Outbreak. Prof. Inf. 2020, 29, e290223. [Google Scholar] [CrossRef]
  46. Vaccari, C.; Chadwick, A. Deepfakes and Disinformation: Exploring the Impact of Synthetic Political Video on Deception, Uncertainty, and Trust in News. Soc. Media + Soc. 2020, 6, 2056305120903408. [Google Scholar] [CrossRef]
  47. Gambín, Á.F.; Yazidi, A.; Vasilakos, A.; Haugerud, H.; Djenouri, Y. Deepfakes: Current and Future Trends. Artif. Intell. Rev. 2024, 57, 64. [Google Scholar] [CrossRef]
Information 17 00026 g001
Figure 1. Number of fake news in digital media articles and average number of citations per publication by year.
Figure 1. Number of fake news in digital media articles and average number of citations per publication by year.
Information 17 00026 g001
Information 17 00026 g002
Figure 2. Development trends of the top five productive countries.
Figure 2. Development trends of the top five productive countries.
Information 17 00026 g002
Information 17 00026 g003
Figure 3. Top 61 authors with Y-index (j ≥ 3).
Figure 3. Top 61 authors with Y-index (j ≥ 3).
Information 17 00026 g003
Information 17 00026 g004
Figure 4. The citation histories of the top ten most frequently cited fake news in digital media articles.
Figure 4. The citation histories of the top ten most frequently cited fake news in digital media articles.
Information 17 00026 g004
Information 17 00026 g005
Figure 5. Plots of the k-means and DBSCAN clustering.
Figure 5. Plots of the k-means and DBSCAN clustering.
Information 17 00026 g005
Information 17 00026 g006
Figure 6. Plots of the clusters for topic modeling.
Figure 6. Plots of the clusters for topic modeling.
Information 17 00026 g006
Table 1. Citations and authors according to the document type.
Table 1. Citations and authors according to the document type.
Document TypeTP%AUAPPTC2023CPP2023
Article6009216172.712,47621
Early access284.3873.1140.50
Editorial material203.1482.443322
Review192.9593.121811
Book review121.8131.110.083
Correction10.1522.000
Proceedings paper10.1511.044.0
TP stands for the total count of publications, AU signifies the number of authors involved, and APP indicates the average number of authors per publication. TC2023 refers to the total citations recorded in the WoS Core Collection by the end of 2023, while CPP2023 represents citations normalized by publication count.
Table 2. The top 10 most productive journals.
Table 2. The top 10 most productive journals.
JournalTP (%)R (IF2023)AUAPPTC2023CPP2023
Social Media + Society70 (12)4 (5.5)2042.9135419
International Journal of Communication52 (8.7)57 (1.9)1282.53546.8
Profesional De La Informacion41 (6.8)39 (2.6)1253.089822
New Media & Society39 (6.5)14 (4.5)1183.0155740
Digital Journalism35 (5.8)8 (5.2)882.569920
Media and Communication34 (5.7)35 (2.7)832.42467.2
Information Communication & Society25 (4.2)17 (4.2)602.486335
Journalism Practice21 (3.5)49 (2.2)522.544521
Health Communication21 (3.5)29 (3.0)773.789943
Political Communication20 (3.3)12 (4.6)623.170835
TP denotes the total count of articles, while % represents the proportion of each journal within the entire collection of fake news research in digital media. AU and APP stand for the number of authors and the average number of authors per article, respectively. R indicates the journal’s ranking among 96 Communication journals in the 2023 Web of Science list, and IF2023 signifies the journal’s impact factor for 2023. Citation performance is assessed using TC2023 (total citations up to the end of 2023) and CPP2023 (citations per article).
Table 3. Top 11 productive countries with TP of 13 articles or more.
Table 3. Top 11 productive countries with TP of 13 articles or more.
CountryTPTP (n = 600)IPC (n = 454)CPC (n = 146)FP (n = 600)RP (n = 600)SP (n = 140)
R (%)CPP2023R (%)CPP2023R (%)CPP2023R (%)CPP2023R (%)CPP2023R (%)CPP2023
USA2291 (38)291 (35)321 (48)221 (32)291 (32)291 (34)16
UK782 (13)233 (8.4)232 (27)243 (9.2)223 (9.3)222 (10)14
Spain693 (12)162 (13)178 (8.2)142 (10)162 (11)163 (9.3)31
Australia394 (6.5)234 (5.9)248 (8.2)224 (5.3)274 (5.3)244 (6.4)26
China335 (5.5)145 (3.5)6.93 (12)205 (4.5)145 (4.2)1014 (1.4)3.0
Singapore296 (4.8)287 (3.1)305 (10)266 (3.8)306 (4.0)305 (2.9)21
The Netherlands296 (4.8)185 (3.5)146 (8.9)237 (3.5)197 (3.8)185 (2.9)20
Germany268 (4.3)199 (2.0)123 (12)239 (2.7)188 (3.5)235 (2.9)14
Canada259 (4.2)108 (2.6)126 (8.9)8.98 (2.8)119 (2.7)1014 (1.4)18
South Africa1310 (2.2)1111 (1.1)1511 (5.5)8.817 (1.0)1514 (1.2)169 (2.1)20
Belgium1310 (2.2)7.311 (1.1)4.211 (5.5)9.317 (1.0)3.519 (1.0)3.55 (2.9)6.8
TP signifies the total number of articles published, with TP R (%) indicating both the ranking and the percentage contribution. IPC R (%) and CPC R (%) denote the rankings and percentages for publications originating from a single country and those involving international collaboration, respectively. FP R (%), RP R (%), and SP R (%) relate to articles authored by first authors, corresponding authors, and single authors. The citation impact is represented by CPP2023, which represents citations normalized by publication count.
Table 4. Top 10 most productive institutions.
Table 4. Top 10 most productive institutions.
6TPTP (n = 600)IPI (n = 307)CPI (n = 293)FP (n = 600)RP (n = 600)SP (n = 140)
R (%)CPP2023R (%)CPP2023R (%)CPP2023R (%)CPP2023R (%)CPP2023R (%)CPP2023
U Oxford241 (4.0)267 (1.3)261 (6.8)263 (2.0)224 (1.8)254 (1.4)36
U Amsterdam222 (3.7)211 (4.2)166 (3.1)271 (3.0)221 (3.2)211 (2.9)20
NTU203 (3.3)362 (3.3)365 (3.4)352 (2.3)432 (2.5)404 (1.4)19
UTA164 (2.7)125 (1.6)6.24 (3.8)156 (1.3)7.66 (1.3)7.64 (1.4)10
UWM164 (2.7)463 (2.3)686 (3.1)284 (1.8)563 (2.0)524 (1.4)26
U Minnesota146 (2.3)3853 (0.33)862 (4.4)347 (1.2)418 (1.2)5820 (0.71)86
QUT127 (2.0)323 (2.3)4519 (1.7)135 (1.5)385 (1.5)382 (2.1)46
Georgetown U127 (2.0)81N/AN/A3 (4.1)8131 (0.50)12633 (0.50)126N/AN/A
U Iowa109 (1.7)3753 (0.33)7.06 (3.1)4118 (0.67)2518 (0.67)2520 (0.71)7.0
CUL910 (1.5)1353 (0.33)1.09 (2.7)1518 (0.67)1414 (0.83)1120 (0.71)1.0
TP refers to the total number of articles produced, with TP R (%) representing both the ranking and the relative share. IPI R (%) and CPI R (%) indicate the rankings and proportions of collaborations within a single institution and between multiple institutions, respectively. FP R (%), RP R (%), and SP R (%) highlight the contributions of first authors, corresponding authors, and single authors. The research impact is assessed using CPP2023, representing citations normalized by publication count. N/A is used when data is unavailable. Institutional abbreviations are provided for clarity. U Oxford: University of Oxford, UK. U Amsterdam: University of Amsterdam, The Netherlands. NTU: Nanyang Technological University, Singapore. UTA: University of Texas Austin, USA. UWM: University of Wisconsin Madison, USA. U Minnesota: University of Minnesota, USA. QUT: Queensland University of Technology, Australia. Georgetown U: Georgetown University, USA. U Iowa: University of Iowa, USA. CUL: City University London, UK.
Table 5. Top nine productive authors with more than five articles.
Table 5. Top nine productive authors with more than five articles.
AuthorTP (n = 600)FP (n = 600)RP (n = 600)SP (n = 689)hRank (j)
Rank (TP)CPP2023Rank (FP)CPP2023Rank (RP)CPP2023Rank (SP)CPP2023
E.K. Vraga1 (14)1022 (7)922 (7)92N/AN/Aπ/42 (14)
L. Bode2 (12)1065 (4)1694 (4)169N/AN/Aπ/45 (8)
M. Tully3 (11)345 (4)254 (4)256 (1)7.0π/45 (8)
M. Hameleers4 (10)141 (10)141 (10)141 (3)6.3π/41 (20)
E.C. Tandoc5 (8)563 (6)702 (7)61N/AN/A0.86223 (13)
J. Lee5 (8)144 (5)114 (4)9.06 (1)3.00.67474 (9)
C. Vaccari7 (6)6917 (2)1029 (3)69N/AN/A0.982816 (5)
J. Lukito7 (6)2917 (2)5020 (2)506 (1)50π/420 (4)
A. Chadwick7 (6)6917 (2)10420 (2)104N/AN/Aπ/420 (4)
TP indicates the total number of articles. FP, RP, and SP represent first-author, corresponding-author, and single-author publications, respectively. CPP2023 reflects the average citation impact per article. The Y-index parameters j and h capture publication potential and authorship characteristics. N/A denotes missing information.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Ho, Y.-S.; Yardibi, F.; Dogan, M.E.; Kusetogullari, H. Mapping Fake News Research in Digital Media: A Bibliometric and Topic Modeling Analysis of Global Trends. Information 2026, 17, 26. https://doi.org/10.3390/info17010026

AMA Style

Ho Y-S, Yardibi F, Dogan ME, Kusetogullari H. Mapping Fake News Research in Digital Media: A Bibliometric and Topic Modeling Analysis of Global Trends. Information. 2026; 17(1):26. https://doi.org/10.3390/info17010026

Chicago/Turabian Style

Ho, Yuh-Shan, Fatma Yardibi, Murat Ertan Dogan, and Huseyin Kusetogullari. 2026. "Mapping Fake News Research in Digital Media: A Bibliometric and Topic Modeling Analysis of Global Trends" Information 17, no. 1: 26. https://doi.org/10.3390/info17010026

APA Style

Ho, Y.-S., Yardibi, F., Dogan, M. E., & Kusetogullari, H. (2026). Mapping Fake News Research in Digital Media: A Bibliometric and Topic Modeling Analysis of Global Trends. Information, 17(1), 26. https://doi.org/10.3390/info17010026

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Article metric data becomes available approximately 24 hours after publication online.
Back to TopTop