Abstract
Media literacy is increasingly gaining importance as a multidimensional concept reflecting the social, cultural, and pedagogical transformations of the digital age. The aim of this study is to examine academic production in media literacy using bibliometric indicators and text-mining methods to reveal structural trends and conceptual orientations in the literature. The study examined 3,209 articles published between 1980 and 2025 in the Web of Science database, following the PRISMA protocol and employing a two-stage method. In the first stage, publication trends, author, journal, institution, and country-based outputs were evaluated using bibliometric analysis; in the second stage, four main themes emerging in the literature were identified by modeling topics using the LDA algorithm, a text mining technique. The findings show that publication volume has increased steadily, the field’s interdisciplinary nature has strengthened, and themes centered on fake news resistance, cultural representation, body image in adolescents, and digital learning processes have become dominant in the literature. The study emphasizes that media literacy is not only an area of academic inquiry but also carries strategic value in the design of educational policies, digital citizenship practices, and multicultural representation, and offers practical contributions.
Plain Language Summary
This study looks at how research on media literacy has developed over time and what topics are most commonly explored. Media literacy means the ability to understand, evaluate, and use media in a smart and responsible way. The researchers analyzed more than 3,000 academic articles published between 1980 and 2025 in the Web of Science database. They first looked at trends in publications, such as how many articles were written each year, which countries and authors were most active, and which journals published the most on this topic. Then, they used a computer-based method called topic modeling to identify the main themes discussed in media literacy research. The results show that media literacy has become a more popular and important subject in recent years. It is not just about understanding media anymore—it now includes different areas such as education, digital citizenship, social issues, and cultural diversity. The study found four major themes in the research: (1) media education and learning, (2) digital skills and safety, (3) social and cultural representation, and (4) critical thinking in media use. Based on these findings, the authors suggest that future media literacy efforts should be more connected to other areas like adult education, cultural inclusion, and digital citizenship. They also recommend more collaboration between different academic fields and researchers to better understand and support media literacy in everyday life. This study helps show where the field has been and where it might go next.
Introduction
The new media environment, shaped by the cultural codes of the digital age, has transformed individuals’ relationships with information, text, and reality; media is no longer merely a channel of communication, but has become a cultural and epistemological ground that shapes individuals’ ways of thinking, value systems, and social positioning (Hobbs, 2001; McLuhan, 1994). To such an extent that individuals are no longer passive recipients of media (Kress, 2010); they have become active subjects who produce, share, and comment on content. This transformation has provided the basis for a structure extending from individual identity to social belonging (Adams & Hamm, 2001; Anderson, 1983; Potter, 2004). In this new position, media functions not only as a provider of information but also as a force that reconstructs individuals’ daily lives, social memory, and even political orientations (Buckingham, 1993; Sholle & Denski, 1995). In this context, the need to critically evaluate the relationship individuals establish with media has made media literacy one of the primary pedagogical and social issues in contemporary societies.
Media literacy theory has several approaches. Potter (2004, 2022) focuses on cognitive processes, mental filters, and decision-making, with a focus on individual awareness. Barton and Hamilton (1998) view literacy as a set of social relations between the individual and the text. Sholle and Denski (1995) connect media literacy with critical pedagogy. They encourage people to become subjects who transform media. Masterman (1985) says critical consciousness is at the core of media literacy. Lewis and Jhally (1998) argue that individuals need to question dominant media discourses. These approaches show that media literacy is more than a technical skill. It is also a pedagogical, cultural, and political field.
The concept of media literacy has been shaped within a deep historical, philosophical, and cultural-political context, rather than being a superficial technical skill. The critical foundations of the concept were first laid by the Frankfurt School’s analyses of the culture industry; Adorno and Horkheimer (2002) argued that modern communication tools function as structures that render individuals passive and suppress their capacity for inquiry, thereby revealing the media’s function as an ideological apparatus. This intellectual groundwork was further systematized by Althusser’s (1971) approach to the state’s ideological apparatuses; the media was evaluated as an institutional space playing a central role in the reproduction of social order and in constructing individuals’ intellectual consent. Within this framework, media literacy is positioned not merely as a practice of content analysis but as an intellectual stance that enables individuals to develop critical awareness of the cultural and ideological codes into which they are systematically immersed.
This critical orientation has gained a pedagogical dimension in the Latin American context through Freire’s (1970) understanding of “critical consciousness”; it has become a process of awareness that increases individuals’ intellectual and social agency by overcoming passivity in the face of media texts. Freire’s dialogic education model structured media literacy as a praxis that prioritizes liberation. This approach was theoretically integrated with media education, particularly in the UK, by Masterman (1985). Based on an analysis of the ideological nature of media representations, this model aimed to enable citizens to take critical and democratic positions regarding media. With Buckingham’s (1993) contribution, this line of thought gained a more contemporary interpretation, redefining media literacy not only as textual analysis but also as a multi-layered learning process that permeates individuals’ cultural identity construction processes, shapes social affiliations, and influences social positioning.
In recent years, media literacy literature has expanded further, particularly with the cultural-political tensions of the digital age. The role of social media platforms in content distribution has led users to become trapped in echo chambers and experience deepening discursive polarization, increasing the need for critical analysis of these environments (Santos Albardía et al., 2025; Wright, 2023). Current academic debates emphasize that media literacy cannot be addressed solely at the level of content verification or visual literacy; rather, it is necessary to develop the capacity to understand how information production is structured within a socio-cultural context, which actors manage this process, and the individual’s position within these networks (Mihailidis & Viotty, 2017; Peña-Fernández et al., 2024). At this point, recent studies have shown that digital empowerment processes significantly shape how such networks function, underscoring the importance of technological infrastructures in mediating information flows (Sun et al., 2024, 2025). Furthermore, media literacy is now being rethought not only in terms of individual learning but also in the context of citizenship, ethical responsibility, and social participation; it is transforming into a critical reading practice shaped through discourse analysis, the exposure of power relations, and the reinterpretation of cultural representations (Kellner & Share, 2019). In this context, media literacy is evolving into a comprehensive theoretical framework in contemporary literature that questions ideological structures, seeks to transform cultural subjectivities, and centers social justice.
On the other hand, in the digital age, media literacy has become a holistic area of competence that encompasses not only individuals’ ability to access information and analyze content, but also their awareness of how this content is produced, circulated, and presented algorithmically (Scharrer et al., 2022; Striphas, 2015). In this sense, algorithmic content ranking (Gillespie, 2018) and data-driven surveillance systems (Zuboff, 2019) have made individuals’ positions in the media environment more complex, giving rise to an expanded definition of media literacy that integrates data ethics, digital transparency, and resistance to disinformation. Indeed, during the COVID-19 pandemic, when the potential of misinformation to harm democratic processes became even more evident, the European Union prioritized media literacy as a strategic intervention area and focused on strengthening critical thinking and verification skills (Grizzle et al., 2021; Tiernan et al., 2023). In this context, advances in artificial intelligence techniques, one of the most notable developments of our time, play a dual role in the disinformation ecosystem. While generative AI tools such as ChatGPT accelerate content production and blur the boundaries between fact and fabrication, automatic classification and machine learning algorithms support the detection and verification of misinformation (Bhutani et al., 2019; Hetler, 2024; Singh et al., 2021). However, human intervention remains crucial due to ethical issues, algorithmic biases, and a lack of transparency (Bostrom & Yudkowsky, 2018; Santos, 2023). Thus, media literacy must now strongly focus on questioning and verifying information obtained from AI-mediated digital environments.
Within the context of the key concepts highlighted in the LDA modeling that underpins this study, it is noteworthy that the media literacy literature has developed a deep sensitivity to the social, cultural, and pedagogical issues of the digital age. Approaches to the problem of accuracy and information pollution encountered in online environments primarily view media literacy not as limited to content consumption, but as a field where critical thinking and verification mechanisms must be developed (Lewandowsky et al., 2017; Wardle & Derakhshan, 2017). Additionally, analyses of the effects of gender roles, cultural representations, and popular media products on children and young people emphasize the cultural reading and critical interpretation functions of media literacy (Gill, 2007; Lemish, 2015). Studies focusing on the relationships between young people’s body image, mental health, and social media use demonstrate that the literature has expanded to include psychosocial dimensions (de Vries et al., 2019; Tiggemann & Slater, 2014). Finally, research addressing the impact of media tools on learning processes in digitalized educational environments positions media literacy as a pedagogical tool and associates it with learners’ cognitive development (Hobbs, 2011; Livingstone, 2018). In parallel, current hybrid analysis models have demonstrated how nonlinear dynamics can be captured in digitally mediated systems, reinforcing the potential of advanced methods for analyzing complex structures (Di et al., 2023, 2024). This thematic framework demonstrates that the field of media literacy is approached holistically, encompassing cultural analysis, critical awareness, and ethical responsibility (see Figure 1).
LDA-based conceptual map of media literacy.
Considering all of these traditional, contemporary, and LDA-based theoretical frameworks, it is evident that academic work on media literacy has become increasingly diverse. Indeed, with contributions from different disciplines, the field has gained thematic richness. At this point, this richness necessitates systematic literature mapping and the identification of thematic clusters. In this regard, Asadzandi et al. (2013) analyzed the literature published in Scopus between 1996 and 2011 on media literacy, revealing developments in publication language, country distribution, and author profiles. Similarly, Wang et al. (2019) examined the thematic axes of media literacy research using network analysis and clustering methods; Zhang et al. (2020) highlighted the power of interdisciplinary approaches by comparing European and non-European practices in the context of educational programs. Meanwhile, Bapte (2021) mapped prolific authors, keyword patterns, and citation relationships using Web of Science data, while Kumar et al. (2021) evaluated growth rates and international collaborations in their analysis of SpringerLink data. While these studies reveal general trends, they also underscore the need for a more in-depth examination of thematic patterns and conceptual density through advanced analyses, such as word modeling.
In this context, the present study aims to examine scientific output on media literacy using bibliometric indicators and to map it multidimensionally through text-mining-based analyses (Aggarwal & Zhai, 2012). The Latent Dirichlet Allocation (LDA) topic-modeling technique, which constitutes the methodological focus of the present study, enables the simultaneous analysis of implicit thematic clusters, conceptual transitions, and interdisciplinary orientations in the literature (Blei et al., 2003; Zhao et al., 2011). This method reveals both the quantitative aspects of the literature and its depth, making thematic tendencies and trends in media literacy more visible. In this regard, the original contribution of this study lies in its examination of the scientific literature on media literacy, with a focus on analysis and modeling, providing both a conceptual framework and a methodological roadmap for future research.
Within this framework, the fundamental research questions to be answered are defined as follows:
As shown in Figure 1, the concepts highlighted in the LDA-based analysis were selected because they reflect the multidimensional nature of media literacy literature. Indeed, the “Disinformation and Verification” dimension shows that information pollution and the search for accuracy are central to media literacy in the digital age. Meanwhile, the “Pedagogical Context” dimension emphasizes the close connection between the concept and educational processes, as well as the role of media literacy in developing individuals’ critical thinking skills. The “Psychosocial Dimension” highlights the effects of media environments on individuals’ social relationships, identity construction, and psychological processes, revealing that media literacy encompasses not only cognitive but also emotional and social dimensions. Finally, the concept of “Cultural Interpretation” underscores the importance of understanding media content within cultural contexts and analyzing diverse identities, values, and forms of representation. In this context, it is argued that the concepts derived from the LDA algorithm reflect the fundamental thematic axes of media literacy discussed in contemporary literature and enrich its theoretical richness.
The fake-news resistance theme identified by the LDA analysis is directly related to disinformation and verification practices, one of the most current issues in media literacy. This finding shows that this dimension has become dominant due to increasing social concerns about media reliability and the spread of misinformation. Similarly, the digital learning theme is closely linked to the pedagogical context; this trend reveals that media literacy is not only an individual skill but also a pedagogical tool integrated into education systems. The findings under the social impact heading focus on individuals’ identity construction, social belonging, and interactions with media; in this respect, they reflect the psychosocial dimension in the conceptual diagram. Finally, the theme of cultural interpretation highlights the analysis of media content through its representational forms and cultural codes, underscoring the field’s conceptual developments as a cultural endeavor. Therefore, these four themes are not merely technical classifications but also high-level conceptual dimensions that reflect the current problem areas and theoretical orientations of media literacy research.
Main Problem of the Study
The main problem of this study is that the subject of media literacy, which is frequently addressed across different social science disciplines, has not yet been mapped holistically at the structural level within the scientific literature. Although there are various bibliometric analyses of media literacy, the structural and temporal analysis of thematic content has not been adequately addressed. In this context, this study aims to analyze the media literacy literature using the Latent Dirichlet Allocation (LDA)-based topic modeling method grounded in bibliometric data. Through LDA, dominant trends, thematic clusters, and topic transformations over time will be revealed, thereby enabling the conceptual development and contextual evolution of media literacy in the social sciences literature to be examined in depth.
Definitions
Transformation: This stage is the most important phase of text mining, as unstructured data are structured and transformed into numerical representations (Aggarwal & Zhai, 2012). For example, unifying text into a single language set, removing capitalization distinctions, and eliminating HTML extensions occur at this stage.
Tokenization: This is the stage where data are processed to obtain meaningful statistical and linguistic outputs. For example, resolving spacing between words and correcting misleading punctuation take place here.
Normalization: This stage reduces data sparsity and standardizes text classification tasks for modeling.
Filtering: In this final stage, clustering and classification results are refined. The main tasks involve removing stop-words and excluding high-frequency but semantically low-value terms to ensure quality and objectivity in textual representation (Manning et al., 2008).
Method
In line with the main research problem, two different techniques were used to answer the study’s research questions. First, a bibliometric analysis was applied to the obtained data to reveal general trends in academic accumulation on media literacy and to structurally map the documents. Bibliometric analysis not only provides a broad perspective on research dynamics but also forms the foundation for the text-mining dataset used in the second stage. Text mining yields highly efficient and meaningful results by structuring large-scale unstructured text data and transforming it into usable information. In addition, text mining enables researchers to uncover hidden patterns in large, unstructured datasets (Cheerkoot-Jalim & Khedo, 2020).
This research is classified as a descriptive content analysis methodologically. Descriptive content analysis, commonly used in content analysis classifications, involves explaining trends and results by including all studies on a predetermined topic in the dataset and by strengthening findings with quantitative indicators.
The method adopted in this study diverges from classical content analysis or descriptive approaches by not only classifying data but also providing a critical framework that reveals epistemological orientations, conceptual priorities, and power asymmetries in knowledge production. Indeed, this approach goes beyond a technical application and enables the reinterpretation of the field’s intellectual landscape, adding a critical dimension that distinguishes it from traditional models.
Data Collection and Limitations
The Web of Science (WOS) database was used to systematically examine media literacy literature. The term “media literacy” was used as a keyword, and only journal articles were included. The search was conducted within “title, abstract, and keywords,” and English was selected as the language. No time limitation was applied; all publications from the earliest media-literacy record in the database up to 20.04.2025 were included.
The methodological rationale behind data and software choices was evaluated holistically. The high level of duplication observed when comparing Scopus and WOS posed a risk to data consistency, leading to the selection of WOS as the sole data source due to its standardized metadata structure and suitability for bibliometric analysis. In addition, limitations in RStudio (bibliometrix/biblioshiny) and Orange Data Mining in extracting Google Scholar data were considered, as scattered and incomplete bibliographic information could threaten methodological integrity. Therefore, the study focused on mapping media-literacy literature through high-impact journals indexed in SSCI/AHCI, and WOS was deemed the most compatible source for this purpose due to its interdisciplinary structure and index coverage.
No disciplinary restrictions were imposed; instead, fields such as communication, education, psychology, and cultural studies intersecting with media literacy were considered within a holistic framework. This ensured that the field was analyzed not only from a single disciplinary perspective but also from an interdisciplinary viewpoint. Absolute publication counts were preferred in trend analysis to indicate the field’s general trajectory, which was deemed sufficient to reflect growth direction despite not reflecting relative disciplinary impact.
Sampling
Since specific criteria were applied to select publications from databases, purposive sampling, a nonprobability sampling method, was used. In purposive sampling, the researcher may include studies that best represent the topic and serve the research purpose based on judgment (Pamuk, 2017).
PRISMA Flow Chart
The screening process was conducted according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) protocol and summarized as follows (Figure 2).
PRISMA protocol.
In the study, the literature review was conducted in accordance with the PRISMA 2020 protocol. In the first stage, records retrieved from the WOS database were combined, yielding a total of 4,924 records for evaluation. Since there were no duplicate records, publications flagged as inappropriate by automation systems, or studies that needed to be excluded for technical reasons, all records were included in the screening process.
During the screening phase, 3,209 reports were accessible, while 1,715 sources could not be accessed due to technical or access restrictions. Accessible studies were carefully reviewed within the framework of pre-determined inclusion and exclusion criteria. These criteria were defined as follows:
Inclusion Criteria:
Articles included in the WOS database,
Studies directly addressing the topic of media literacy,
Theoretical, empirical, or comparative studies relevant to the research objectives,
Articles published in English.
Exclusion Criteria:
Publication types other than articles (book chapters, conference proceedings, reports, etc.) (n = 1,008),
Publications published in languages outside the scope of the research (n = 691),
Reports lacking sufficient bibliographic information or methodological details (n = 16).
Following this comprehensive screening process, 3,209 articles were included in the study’s analysis phase. In the final stage, the suitability of the articles included in the study was confirmed not only by automated filters but also by content and methodology assessments conducted by two experts in the field. This ensured both the scientific reliability and the methodological robustness of the resulting dataset.
Data Pre-Processing Phase
The data collected during the data collection phase were analyzed using RStudio, and a structural mapping of all articles on media literacy was performed. For the text mining in the second stage, the data obtained in RStudio were analyzed with the Orange Data Mining software, and topic modeling was performed using the LDA (Latent Dirichlet Allocation) algorithm. However, a preprocessing stage is needed to clean the data obtained from the RStudio program of misleading, repetitive, and unnecessary words. In this context, the raw data set was filtered through a four-stage process to ensure the accuracy and reliability of the pre-analysis results.
Data Analysis
Bibliometric and Topic Modeling
It is a method that has attracted a lot of attention in recent years, especially thanks to advances in processing scientific big data and in research capacity. Especially in scientific studies, bibliometric analysis is preferred for revealing trends in article or journal performance, collaboration patterns, and research components. (Donthu et al., 2021). In this study, the bibliometric analysis method was preferred to examine the status of media literacy in the scientific field and to identify trends in articles within the specified period. With bibliometric analysis, the chronological order of the articles that make up the data pool was made and then an in-depth understanding and disclosure of the journal that publishes the most on the subject, the local effects of the sources, the annual citation numbers and rates, the author who produces the most publications on media literacy, and the countries that make the most scientific contributions to the subject was carried out. In this study, the “biblioshiny” software in the “bibliometrix” package of the R Studio program was used for bibliometric analysis, and the data were supported with visuals.
Another analysis method used in the study was topic modeling, an applied form of text mining. In the topic modeling process performed in the Orange Data Mining program, the results from the bibliometric dataset are processed a second time, and the Latent Dirichlet Allocation (LDA) algorithm is used to reveal latent themes across thousands of texts and hidden topics in the dataset.
Trend Analysis
Latent Dirichlet Allocation (LDA) is a generative probabilistic model in text mining that estimates the probability of each descriptive word and word set for each topic (Blei et al., 2003). LDA, which allows documents to contain multiple topics to varying degrees, allows the researcher to systematically uncover changes and shifts in the prevalence and composition of various topics over time. In this context, the trend analysis method was applied to the topics identified by topic modeling during the data analysis. Trend analysis was used to calculate the interest levels and relative acceleration levels of the topics.
Findings
To address the first research question, our study traces the evolution of media literacy scholarship from its early conceptual origins to its contemporary theoretical articulations. Within this temporal examination, particular emphasis is placed on identifying the shifting epistemological foundations, pedagogical orientations, and socio-technological influences that have shaped the field across successive periods. By systematically analyzing the keywords scholars have employed over time, the investigation further illuminates how core concepts have been introduced, contested, expanded, and refined in response to emerging media environments, educational needs, and critical debates.
In this context, general information about the articles on media literacy published in WOS databases between 01.01.1980 and 20.04.2025 is given in Table 1.
Basic Statistical Data on Media Literacy.
Table 1 shows the statistical results of academic studies on media literacy in the WOS database. When the statistical view of this study, limited to articles in academic documents, is examined, it is seen that 3,209 articles were published by 1,197 sources. Again, considering these data, it was determined that the first article on media literacy was published in the WOS database in 1980 and the last article in 2025. The dataset includes publications published between 1980 and 2025, enabling the examination of the historical development and conceptual transformation of media literacy research over time.
A total of 1,197 sources published articles related to media literacy during the examined period. The annual growth rate of the publications was calculated as 11.83%, indicating an increase in publication output over time. These findings reflect the expansion of media literacy research across different publication sources and disciplines. The average age of the articles published on media literacy was 6.49 years. The average citation score of the articles included in the analysis was calculated as 13.74.
When the keywords (DE) of the articles and the automatically derived Keywords Plus terms (ID) were analyzed, 2,565 Keywords Plus terms and 6,654 author keywords were identified in the dataset. The Keywords Plus terms derived from WOS indicate the conceptual diversity of the literature, while the author keywords reflect the thematic focus areas of the studies. These keywords represent an important indicator of the descriptive direction and focal points of the articles. These keywords reflect the thematic diversity and interdisciplinary structure of the studies included in the dataset.
Within the scope of the study, 835 authors contributed to 3,209 articles as single authors, while 6,521 authors preferred co-authorship. This ratio indicates that co-authorship is more prevalent in scientific documents related to the field. Approximately 15.5% of the publications were produced through international collaborations. This rate demonstrates that there are multinational academic interactions and knowledge exchange in the field.
Table 2 shows the distribution of articles on media literacy by year. According to this table, the first article on media literacy in WOS was published in 1980. Until 1998, publication output remained relatively stable; publication counts increased in 1998, 2004, 2007, 2012, 2016, and 2020. The number of publications on media literacy increased steadily across the examined periods.
Scientific Production Related to Media Literacy.
In Figure 3, three area graphs are presented in relation to the subject of the study. In the figure, the authors with the highest number of publications on media literacy are shown on the left, the focus words in the titles of these authors’ articles are shown in the middle, and the keywords used by these authors are shown on the right. The size of the rectangles in each field indicates the contribution to the field, while the lines between the fields represent the intensity and level of these relationships.
Three domains graph for media literacy.
Addressing the second research question, this section aims to delineate the structural contours of scholarly production in the media literacy domain by examining its publication venues, leading contributors, and conceptual anchors. Through a systematic bibliometric inquiry, key journals that serve as principal dissemination platforms for media literacy research will be identified, thereby revealing the disciplinary homes, intellectual communities, and editorial directions shaping the field’s epistemic discourse. In parallel, the analysis will spotlight influential authors whose scholarly trajectories and cumulative contributions have played a formative role in defining and expanding the scope of media literacy inquiry.
An analysis of Table 3 shows that Erica Weintraub Austin has the highest scientific productivity in media literacy. This author, who produced a total of 26 articles on the subject, was followed by Fedorov, Hobbs, and Levitskaya in terms of proportional contributions. With 22 articles, Alexander Fedorov ranked second in the quantitative ranking of scientific contributions but ranked first in the proportionally decimalized number of articles. This ratio indicates that the researcher contributed more intensively to certain studies in the field than other authors did.
Ten Authors Who Produced the Most Documents on Media Literacy.
When Table 4 is analyzed, it is observed that communication sciences journals keep the subject more actively on the agenda than other social sciences journals and include a greater number of articles on the topic. It is particularly noteworthy that the number of articles published in Media Literacy and Academic Research is more than twice that of other journals.
Ten Journals With the Highest Number of Articles on Media Literacy.
In Table 5, the five most-cited articles in the citation-based analysis of the media literacy literature are listed. The five highly cited articles identified in the analysis address media literacy from different perspectives and emphasize both theoretical and practical aspects of the field. In particular, the most-cited article, The Virtual Sphere: The Internet as a Public Sphere, demonstrates that classical texts that approach media literacy from a theoretical perspective still serve as reference points in the literature. However, the relatively lower normalized citation rate of this study compared to more recent works suggests that, while foundational theoretical texts continue to exert long-term influence, contemporary problem-based approaches have greater short-term academic resonance.
Ranking of the Five Most Cited Articles on Media Literacy.
According to Table 6, Washington State University had the highest publication count among the listed institutions. The institutional distribution reflects the participation of universities from different countries in media literacy research.
Ten Academic Institutions and Number of Articles on Media Literacy.
In response to the third research question, this map examines the collaborative architecture that underpins media literacy scholarship at the international level. Through network-based bibliometric indicators, the analysis identifies the principal institutional hubs, transnational research linkages, and prevailing modes of scholarly co-production that have facilitated knowledge exchange and collective advancement in the field.
Figure 4 presents the cooperation network among countries. The United States occupied a central position in the collaboration network and showed a high level of international co-authorship connections within the dataset.
Network map of countries’ cooperation on media literacy.
Figure 5 shows the stages of the topic modeling process for media literacy. Accordingly, the data set in the corpus was first cleaned with the Preprocess Text process and made suitable for topic modeling. The data prepared for analysis in this process is then used to identify the main topics in the texts using the Latent Dirichlet Allocation (LDA) algorithm.
Steps of the topic modeling analysis.
In topic modeling, the documents in the data stack are a mixture of multiple topics. These topics can be independent and different from each other. The main regulator of this difference comes from the word-text relationship in the machine learning process. On the other hand, the topic modeling process is also related to the researcher’s familiarity and expertise with the topic. In fact, the ideal topics that emerge from the word-text relationship are shaped by the researcher’s expertise. Although the researcher’s expertise on the subject is sufficient to determine the ideal number of topics, it remains a subjective judgment. To eliminate subjectivity, LDA models use evaluation criteria to determine the ideal topic.
Coherence and Log Perplexity are evaluation metrics used in the LDA modeling process. In the topic modeling process, Coherence measures the semantic consistency of words within topics, while Log Perplexity is a probabilistic metric that assesses the ability to predict new data. The lower the Perplexity value, the higher the model’s generalization performance (Röder et al., 2015). In this context, the ideal number of topics was determined by examining Coherence and Perplexity scores. Accordingly, Coherence and Perplexity values from 1 to 20 were compared to determine the ideal number of topics.
Table 7 shows the Coherence and Plurality scores for all topics (1-20) to determine the ideal number of topics related to media literacy. High Coherence scores, which indicate semantic consistency, are desirable for determining the ideal topic. Perplexity, a measure of probability estimation, should be low for better modeling. In this context, when the above data are used to explain the study’s modeling, the number of topics should be 4.
Log Perplexity and Coherence Values in the Process of Determining the Number of Topics.
The topic modeling process was performed using the Latent Dirichlet Allocation (LDA) algorithm. Prior to analysis, the dataset underwent text preprocessing steps (lowercase conversion, stop-word removal, punctuation and special-character cleaning, lemmatization, etc.) to make it suitable for LDA. Bold values indicate the four-topic model was selected as the optimal solution because it achieved the lowest perplexity value (135.22131) while also producing the highest coherence score (0.41065), indicating the best balance between model fit and semantic interpretability. Therefore, the four-topic solution was adopted for the subsequent analyses.
Determining the Parameters
The Latent Dirichlet Allocation (LDA) model used in the study was run with default hyperparameters (α and β). No manual adjustments were made to the parameters because the aim of the research was to objectively evaluate the topics generated by the algorithm under standard conditions using the coherence and log perplexity metrics.
Verifying the Accuracy of the Model
The accuracy of the model was verified in three stages:
Numerical evaluation: Coherence and Perplexity values were tested for 1 to 20 topics. The results showed that 4 topics were ideal in terms of both semantic integrity and model fit.
Expert review: The topic clusters obtained were examined by two independent researchers and then validated by two academics who are experts in the field.
Content alignment: The key terms in the topic clusters were compared with the content of the articles, and conceptual consistency was observed.
As shown in Table 8, the four-topic model produced a Coherence value of 0.41065 and a Perplexity value of 135.22131. The resulting four-topic model has been validated not only through numerical indicators but also through comparisons among researchers and independent reviews by two experts in the field.
Number of Topics and Headings Revealed as a Result of Text Mining.
As a result of this triple validation (statistical + researcher + expert), it was concluded that the ideal number of topics is 4.
Thanks to this methodological process, the identified topics have been validated not only algorithmically but also within the framework of academic expertise, ensuring both the internal consistency and external validity of the model.
To address the fourth research question, text mining and LDA-based topic modeling techniques were applied to identify the main thematic structures in the media literacy literature. The analysis revealed the dominant topics and thematic trends within the dataset. By employing text mining procedures and Latent Dirichlet Allocation (LDA) modeling, the analysis moves beyond surface-level descriptors and citation-based indicators to uncover deeper conceptual layers, discursive patterns, and thematic constellations that have shaped the field’s scholarly agenda. This analytical approach enables the identification of core research clusters, emerging thematic directions, and domain-specific epistemological tendencies, thereby illuminating how media literacy has been theorized, operationalized, and contextualized across diverse academic and sociocultural environments.
Table 8 presents the four main topics related to media literacy that emerged during the topic modeling process. These topics show the focus and general tendency of all articles on the subject. In addition, the topic weight ratios indicate the relative importance of articles on media literacy. According to the topics revealed, the title “The effect of media literacy on students’ learning process in the digital age” (37.29%) has the highest weight ratio. This topic is followed by “Media literacy against fake and false news on social media” (26.08%). “Social media use, media literacy and body image perception in adolescents” (22.30) and “Analyzing gender and culture in social media and films through media literacy” (14.30) were identified as other topics based on the mass they formed. The word cloud used to inform these topics is shown in Figure 6 below.
Figure 6 shows the word cloud of the 4 main topics that emerged during the topic modeling process. In this cloud, the concepts of “media” used 15,212 times and “literacy” used 9630 times are the basic concepts of media literacy, followed by “education” (4574), “use” (4219), “social” (4054), “study” (4047), “inform” (3284), “digit” (3274), and “student” (3078).
LDavis word cloud.
Analysis of Table 9 shows that significant acceleration has occurred after 2000 across all topics examined. Especially the topic numbered 4, “The effect of media literacy on students’ learning process in the digital age,” has shown a steady increase compared to the other topics. The publication trend shows a gradual increase in studies focusing on media literacy and educational processes.
Distribution of the Subjects of the Articles According to Years.
Figure 7 shows the volumetric changes and general trends of the topics uncovered in the study over time. Accordingly, while there is a continuous increase in the topic “media literacy against fake and false news on social media,” this increase also grows in volume. The observed increase became more pronounced after 2000 and continued across the subsequent periods examined in the dataset. When the topic “Analyzing gender and culture in social media and films through media literacy” is examined, an inverse trend emerges in terms of volume compared to other topics. This indicates that academic interest in this topic has begun to decline relative to other subjects.
Volumetric changes and trends of the topics revealed in articles on media literacy over time.
The topic “Social media use, media literacy and body image perception in adolescents” appears to have a very high volume compared to the other topics. Additionally, it was determined that academic interest in this topic is consistently increasing. The findings indicate a growing number of studies focusing on social media use and body image perception in adolescents. It was also determined that the relationship between learning processes and media literacy, under the topic “The effect of media literacy on students’ learning process in the digital age,” is increasing in volume. The increase continued during the post-pandemic period characterized by expanded digital education practices.
When the R2 (coefficient of determination) values for the topics were examined in Figure 5, the topic “Media literacy against fake and false news on social media” produced an R2 value of 0.6767. This value indicates a moderate level of explanatory power within the model, with approximately 67% of the observed variance explained. The topic “Analyzing gender and culture in social media and films through media literacy” produces awareness in gender and cultural analysis of media texts, with an R2 score of 0.6598. However, the R2 value is not sufficiently high, indicating a weaker explanatory relationship.
When the topic “Social media use, media literacy and body image perception in adolescents” was analyzed, the model produced an R2 value of 0.87, indicating a high level of explanatory power for this topic.
According to the coefficient of determination for the topic “The effect of media literacy on students’ learning process in the digital age,” studies in media literacy explain 89% of students’ learning processes in the digital age.
Figure 8 shows the trends in the topics over time. In this context, it is observed that academic trends shift in focus over different periods. Accordingly, while the topic “Media literacy against fake and false news on social media” displays a fluctuating pattern, the topic “Analyzing gender and culture in social media and films through media literacy” did not demonstrate a steady increase; instead, it peaked in the second and fourth periods but entered a downward trend in the fifth period. On the other hand, it was determined that the topic “Adolescents’ use of social media, media literacy and body image perception” gained importance steadily and significantly throughout all five periods. The publication trend indicates a continuous increase in studies focusing on adolescents, social media use, and body image perception. Finally, it is also observed that while interest in the topic “The effect of media literacy on the learning process of students in the digital age” was very high in the early periods, it decreased significantly over time
Trend changes in the topics covered by media literacy articles over time.
Figure 9 shows the acceleration of change in scientific academic orientations related to media literacy. Accordingly, the topics “Social media use, media literacy and body image perception in adolescents,”“Analyzing gender and culture in social media and films through media literacy,” and “Media literacy against fake and false news in social media” exhibit a similar positive acceleration trend, while the topic “The effect of media literacy on students’ learning process in the digital age” demonstrates a negative trend. This graph also provides insight into the topics that researchers focusing on media literacy are likely to prioritize in the future.
Accelerational changes in the trends of the topics of articles on media literacy.
In relation to the fifth research question, this study aims to elucidate the intellectual architecture of media literacy scholarship and to derive forward-looking implications for the field’s evolution. In this context, the patterns revealed through thematic mapping, co-citation structures, and conceptual clusters will be interpreted to delineate the foundational knowledge domains, emerging lines of inquiry, and interconnected scholarly trajectories that collectively constitute the field’s epistemic landscape. Attention will be devoted to identifying conceptual convergences and divergences, latent research niches, and transformative themes informed by technological, sociocultural, and pedagogical developments. Given the strategic relevance of this analysis, detailed reflections on the implications of these findings for future research agendas and theoretical development will be presented in the concluding sections.
Discussion, Conclusion, and Recommendations
The analyses conducted within the scope of the study, to put it in the most general terms, revealed that academic production in the field of media literacy should be rethought in a way to address both the target audiences in the field of education and adult individuals, public actors, and all social groups exposed to the directive influence of media content. The findings indicate that media literacy is no longer a matter of education limited to students, but rather a lifelong learning process. Media literacy has become an integral part of lifelong education policies, especially as adults who consume content intensively on digital platforms have become vulnerable to misinformation. Indeed, in the post-pandemic period, the increase in online training and the central position of digital channels in daily life have made it necessary to expand media literacy through voluntary, state-sponsored, or sponsored programs. At this point, it is essential to increase academic production on media literacy, especially in the field of adult education, and to evaluate trainers’ qualifications against high-quality standards.
The four main themes that emerged from the LDA analyses indicate that media literacy is shaped by different dimensions in today’s era of crisis. The first theme, “Media literacy against fake and false news on social media,” calls for strengthening individuals’ digital resistance capacity against the growing threat of disinformation globally. Information pollution and manipulative content, which have increased especially during the pandemic, have underscored the importance of young individuals’ ability to access, sort, and interpret information in the digital environment (Anstead et al., 2025; Fabbro & Gabbi, 2024). At this point, media literacy needs to be supported not only by a protectionist educational approach, but also by constructive and participatory strategies that enable individuals to interact with the media consciously. Especially given new threats such as identity representation on digital platforms, fake accounts, and algorithmic manipulation, it is understood that media literacy should be structured to empower individuals and strengthen their critical reflexes rather than merely protect them. This theme clearly demonstrates how media literacy is positioned not only as a defensive barrier against misinformation and disinformation but also as a strategic tool that empowers individuals and strengthens their critical thinking skills (Orhan, 2023). Indeed, the rapid circulation of fake news, particularly through social media platforms, has brought media literacy to the forefront as an indispensable critical skill in democratic societies. On the other hand, developments of global significance, such as the 2016 USA elections, the Brexit referendum, and the COVID-19 pandemic, have shaped the evolution of this field, deepened the literature on combating misinformation, and ensured that media literacy has become a critical resistance mechanism in contemporary societies.
The second theme, “Analyzing gender and culture in social media and films through media literacy,” points to how media content reproduces the politics of representation and cultural norms. Especially through films and social media posts, gender roles, ethnic identities, and cultural codes are presented within the framework of certain patterns; social norms are reconstructed through these presentations (Grajek, 2025). In this context, it becomes essential to consider media literacy as an educational tool that develops cultural analysis skills and promotes a perspective on social equality and justice (D’Haenens & Joris, 2025). In this sense, the individual’s capacity to question not only the media’s content but also the normative orientations underlying it is a very important issue. Here, it is understood that media literacy guides not only information verification processes but also cultural representation and gender studies. Indeed, how women, men, and different identities are presented in media content can be analyzed more critically with media literacy skills (Salleh et al., 2023). Thus, individuals could question the normative orientations underlying representation practices. On the other hand, feminist media studies and cultural studies have played a decisive role in shaping this approach, leading to the more intensive use of media literacy as a methodological framework in visual culture analysis since the 2000s.
The third theme, “Social media use, media literacy and body image perception in adolescents,” revealed the impact of media content on individual identity construction. In this regard, especially the body norms and idealized lifestyles displayed on social media platforms cause serious distortions in young individuals’ body perception (Czubaj et al., 2025; Mazzeo et al., 2024). In fact, this situation increases psychological vulnerability in adolescents. At this point, the aim of media literacy in establishing and sustaining social order becomes much more important, bringing the individual to a reasonable level of consciousness that can both analyze media content and maintain a healthy distance from it. In this respect, it is clear that media literacy should be considered a field of cognitive, emotional, and social skills. The relationship between social media use and body image during adolescence highlights the psychosocial aspect of media literacy. Idealized body images and lifestyles circulating on visually intensive platforms, in particular, undermine young people’s self-perception, leading to a loss of self-confidence and intense social comparison tendencies. Indeed, the increase in these vulnerabilities has made it imperative to position media literacy not only as a cognitive skill but also as an emotional and social skill. In this context, media literacy stands out as a tool that allows young individuals to maintain a critical distance from content and strengthens their psychological resilience. On the other hand, the evolution of this theme has taken shape at the intersection of psychology, education, and communication disciplines in recent years, becoming even more pronounced with the proliferation of platforms such as Instagram and TikTok.
The fourth theme, “The effect of media literacy in the learning process of students in the digital age,” reveals that the relationship between education and media has become more visible as digital education tools have become widespread. As a matter of fact, distance education practices, which have become compulsory during the pandemic, have made media literacy not only a learning object but also a learning tool (Turner, 2024). In this respect, media literacy is now considered one of the main variables that determine the capacity of education to digitalize (Salleh et al., 2023). However, if these media interactions are not integrated with students’ values of digital citizenship, critical reading, and ethical sensitivity, there is also a risk that the individual will become a mere technical user. The dominant theme is the role of media literacy in education. Especially in the digital age, media literacy is a critical competency for students’ access to, verification of, and learning from information. This theme has evolved alongside the integration of media literacy into formal education curricula, digital pedagogies, and lifelong learning. With the digital transformation after 2010, this field has gained significant prominence in the literature.
On the other hand, when considered in the context of international collaborations, the findings reveal that the media literacy literature still has a production dynamic centered in the United States. However, the low rate of international co-authorship makes it difficult to construct a polyphonic discourse on media literacy on a global scale. This situation shows that media content is mostly analyzed through Western frameworks, with insufficient inclusion of diverse sociocultural contexts in knowledge production. Especially in this period of heightened alignment between individuals’ virtual and real identities, there is a serious rupture between traditional social values and perceptions of digital citizenship, leading to a weakening of democratic participation in digital spaces.
The prominence of the United States as a central country is not merely a quantitative data point but also reflects the relationship between academic production and geopolitical balances and hegemony in the field of knowledge. In this context, the US’s academic centrality can be explained by its publication policies setting global trends, the distribution of research funds, and the epistemological influence of Anglo-Saxon literature. Therefore, this finding has been evaluated as a critical indicator that reflects not only an academic orientation of media literacy studies, but also power relations on a global scale and the center-periphery distinction in knowledge production.
Particularly in the literature, studies that are limited to bibliometric analysis tend to stand out. For example, studies in the field (İli, 2025; Kutlu-Abu & Arslan, 2023) mapped the media literacy literature using bibliometric indicators but offered limited insights into conceptual patterns, as they did not delve into the depth of the texts’ content. Our study goes beyond this level by employing text mining methods, generating data not only through publication counts and citation relationships but also through thematic orientations and conceptual clusters hidden within the texts. This methodological choice has allowed us to shed light not only on numerical distributions but also on the conceptual dynamics shaping the depths of the literature while explaining why four themes have become dominant.
Based on this information, the present study provides an in-depth analysis of academic production in media literacy and reveals how this concept has evolved in the new media environment, shaped by the transformative effects of the digital age. Indeed, bibliometric data showed that academic interest in the field has steadily increased over the last four decades, while LDA-based analyses revealed that media literacy is divided into thematic clusters including educational, social, cultural, and technological dimensions. In fact, the study found that media literacy is no longer just a skill area for students, but has become a vital competence for adults, educators, and all social groups that interact intensively with the media.
Especially in the current period of accelerated digitalization after the pandemic, media literacy is positioned as a fundamental democratic tool that enables individuals not only to understand content but also to critically distance themselves from it, question cultural representations, and participate as digital citizens. In fact, the incompatibility between the individual’s virtual identity and public identity weakens traditional democratic practices and limits digital participation in terms of pluralism and representation. In this context, media literacy should be considered not only as an individual awareness but also as a strategic field of education for the sustainability of democratic societies.
Based on this framework, the following recommendations are presented in light of the data obtained within the scope of the study:
Recommendations for Practice
To make media literacy more inclusive, it is essential to view it as a lifelong learning area not only for students but also for adults, the elderly, and all individuals who actively consume content in digital environments. Especially as the dependency on digital media has increased due to the pandemic, it is evident that adults have become vulnerable to threats of information pollution, disinformation, and content manipulation in these environments. Therefore, it is recommended that media literacy be integrated not only into the formal education system but also delivered to all segments of society through voluntary, state-supported, or sponsorship-based non-formal education programs. Media literacy should not be limited to formal educational institutions. Adult media literacy modules should be developed through Public Education Centers affiliated with the relevant Ministry, local lifelong learning units of municipalities, and civil society organizations. These programs, aligned with the European Union’s Digital Citizenship Education Strategy and UNESCO’s Media and Information Literacy Framework, will help close the digital skills gap and build a resilient society against disinformation.
In addition, the knowledge levels of media literacy instructors (i.e., trainers) should be evaluated against international reference standards, and continuous in-service training programs for these individuals should be expanded. This is because media literacy requires not only knowledge of the content but also the ability to question this content at educational, cultural, and political levels. At this stage, blending internationally established media literacy programs, such as the UNESCO Framework for Media and Information Literacy, with local practices will be decisive in enhancing educators’ competencies. The quality of media literacy education is directly related to educators’ qualifications. Therefore, the competencies of teachers and educators in media literacy should be measured and accredited against international standards (e.g., UNESCO MIL indicators and the Council of Europe Digital Competence Framework). Higher education institutions should support these processes through pedagogical certification programs and continuing professional development modules.
Recommendations for Policy
On the other hand, it is noteworthy that academic production on media literacy is still largely centered in the US; however, international cooperation networks are insufficient to diversify this production. For this reason, it is suggested that multilingual, multicultural, and international academic collaborations in the field of media literacy be expanded, and that equitable publication and project opportunities be created to enable participation from diverse social contexts in knowledge production. In particular, the construction of common academic platforms where the themes of migration, gender, cultural plurality, and representation can be discussed by geographically balanced circles is an important necessity in this context. To break this unipolar structure, multinational funding mechanisms such as Horizon Europe, Erasmus+, and similar projects should be used more effectively, and international joint publications and research centers should be established. In this way, media literacy will cease to be a Western-centric field of discussion and become a global discipline enriched by knowledge production from different geographical areas.
Media literacy education policies should be structured not only with a protectionist reflex, but also with a constructive framework that focuses on how individuals can use digital media more effectively, consciously, and ethically. In this context, media literacy should be taken to a level of consciousness where the individual sees digital media not as a threat but as a means of freedom of expression, democratic participation, and cultural plurality. In this respect, media literacy should be structured as a strategic field of education that shapes the digital future of democratic societies. In this context, concrete steps such as integrating media literacy into the curriculum, offering digital ethics courses, and developing critical media reading skills should be incorporated into national education policies.
Media literacy policies should be expanded to include social themes such as migration, gender, cultural pluralism, and representation. Multilingual educational materials should be prepared, especially for immigrants, minorities, and diverse cultural communities; public institutions, universities, and international organizations should collaborate in their development. This approach will ensure that media literacy is positioned as an egalitarian field of education that increases democratic participation.
Within the scope of this study, determining integrative media literacy policies requires a holistic approach to the field. Indeed, moving media literacy policies away from a reliance on the formal education system and designing them as a multi-layered strategy that extends from preschool to the elderly population will enable the management of media risks faced by different age and socioeconomic groups under one roof. In this context, establishing permanent mechanisms for coordinated work among state institutions, academic research centers, and civil society organizations will align national strategies with local initiatives, fostering the development of sustainable policies rather than fragmented applications.
On the other hand, redefining media literacy beyond communication studies to include disciplines such as law, political science, education, psychology, and sociology will create a more inclusive policy framework by ensuring that themes such as fake news resistance, cultural representation, democratic participation, and digital ethics feed into each other. Furthermore, supporting multicultural collaboration networks and aligning them with global standards will shift media literacy policies from a solely national agenda to a shared product of global information circulation.
Recommendations for Future Research
Finally, media literacy is expected to go beyond merely protecting individuals from misinformation and to foster active digital citizenship in democratic societies; thus, these policies are positioned as a strategic area that strengthens social resilience and contributes to the construction of democratic capacity.
Footnotes
Ethical Considerations
This study does not involve any human or animal subjects, biological materials, or experimental interventions. Only secondary bibliometric data were used, and no direct data collection from individuals was conducted. The research data were obtained from the Web of Science (WoS) and Scopus databases. Therefore, ethical approval is not required. The study was conducted in accordance with the principles of academic research and publication ethics.
Author Contributions
All authors have contributed equally at all stages of the manuscript. All authors have read and approved the final version of the work.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.