Building on the introduction’s overview of digital challenges, this section contextualises the study’s rationale by examining how cybersecurity behaviour, digital resilience and adolescent vulnerability converge, warranting a focused bibliometric investigation. Cybersecurity behaviour plays a key role in fostering digital resilience and online safety practices. To comprehend the factors that affect online behaviour, researchers utilise behavioural theories such as the theory of planned behaviour (TPB). From a TPB perspective, findings indicate that comprehending the repercussions of online risks is crucial for educating individuals about online safety (Alanazi, Freeman & Tootell 2022). To meet the challenges and online dangers to which individuals, including teenagers, are susceptible, it is crucial to have digital resilience (Sharma et al. 2022).

This suggests that the capacity for resilience utilises technological concepts to develop adaptive skills for managing unforeseen disruptions and is vital in cyberspace for bolstering cybersecurity (Bahmanova & Lace 2024). Digital resilience can serve as an effective protective mechanism for individuals who may face online risks, especially adolescents (Pan et al. 2024; Qi & Yang 2024a). Their degree of digital resilience will probably have a major impact on adolescents’ success in the digital age, as well as on their social interactions, academic performance and overall life satisfaction (Qi & Yang 2024a). Digital resilience is usually addressed in the context of both personal and organisational competence (Fischer et al. 2023; Mahmood, Chadhar & Firmin 2024). This research utilises the individual capability perspective to examine how research has advanced in the area of digital resilience. It is on this premise that the study investigates the mapping of scientific production and conceptual structure among adolescents in digital resilience and cybersecurity based on thematic developments and trajectories, and prolific scholars.

This study is guided by the following research questions:

RQ1: How has digital resilience research grown in terms of scientific output and annual publication trends over time?

RQ2: What thematic developments and trajectories characterise digital resilience research, particularly concerning adolescents?

RQ3: What are the patterns of scholarly collaboration, affiliation and citation impact within the digital resilience research field?

RQ4: What opportunities exist for interdisciplinary collaboration to advance digital resilience research focused on adolescents?

The data for analysis were retrieved from the Scopus-curated database, which offers better coverage of citation databases in the literature (Waltman 2016). The combination of operators, filters and search string is shown in the left-side block of Figure 1. The OR Boolean operator was used to perform a compound search and concatenation. In addition to the keywords searched, the search string used asterisk (*) wildcards to ensure that all potential keywords were included in the results. The search term “digital resilien*,” for example, produced results that included “digital resilience,” “digital resiliency,” or “digital resilient,” among other terms.

The dataset, consisting of 483 documents in comma-separated values (CSV) file format, was extracted from the Scopus database. No duplicates were found. After reviewing the keywords of the documents and examining the abstracts of those whose keywords showed a lack of alignment with the research key terms and focus, 9 documents were removed, leaving 474 for analysis in this study. Nine of the documents were removed because their content was not related to the study’s focus and instead concerned supply chain cybersecurity. Table 1 presents the rationale for terms used in search strings in this study.

Bibliometric analysis was conducted using Bibliometrix (version 4.2.1) through the Biblioshiny web interface version 4.4.3. Statistical analysis was performed using the R version 4.4.0 with the Bibliometrix package (Naples, Italy) (Aria & Cuccurullo 2017). Using algorithms for statistical and science mapping analysis, this software offers a suite of tools for quantitative bibliometric studies. Additionally, it has Biblioshiny, a web-based interface that enables the import of data from Scopus or Web of Science databases in CSV, BibTeX or plain text formats. The software does not facilitate the integration of data produced from separate databases for this investigation (Agbo, Oyelere & Suhonen 2021). Authors use acronyms, synonymous phrases and spelling or term variants, which may limit the ability to generalise the research findings across research trends (Mohsen, Fu & Ho 2017).

Therefore, different variations of authors’ keywords were reviewed and combined into corresponding single words (Mak et al. 2024) and uploaded into the Biblioshiny web interface as synonyms in the analysis of the dataset, including ‘cyber security’, ‘cyber-security’ used as ‘cybersecurity’, ‘cyber safety’, ‘cyber-safety’ as ‘cyber safety’, and ‘COVID-19 pandemic’ as ‘COVID-19’. Also, full terms were used in place of abbreviations, including AI and natural language processing (NLP). This enabled the Bibliometrix application to provide a consolidated report on terms with multiple forms or abbreviations.

Table 2 provides an overview of dataset information, including document contents, authors and author collaborations. There were 474 documents in the dataset, which covered the years December 2006 to January 2025. The focus of the published materials was scoped using 1302 author keywords (DE). Individual authors generated 82 documents out of 1275. Among the document kinds are 53 books, 54 book chapters, 88 conference papers and 279 articles.

Ethical clearance to conduct this study was obtained from the Department of Applied Information Systems Research Ethics Committee of the University of Johannesburg (Ref. No. 2025AIS003).

On average, digital resilience research accounts for 8.84% of scientific output from December 2006 to January 2025, as measured by the Bibliometrix R package shown in Table 2. Table 3 shows that one book, published in 2006, discusses burgeoning issues, types of cybercrime, disputes and remedies related to the Internet’s effects on society (Schell 2006). No contribution was made in 2007; however, three research works were published in 2008, which focused on the rise of cyberbullying using instant messaging, e-mail, chat rooms and other digital messaging platforms due to advancements in computer and mobile technology (Kowalski, Limber & Agatston 2008; Mason 2008; Shariff 2008). By recognising the risks of the Internet and cyberbullying, these studies laid a foundation, underscoring the need for resilience in society and possibly marking the inception of digital resilience research. In 2009, five research works were published, paving the way for digital resilience research with a focus on information systems security effectiveness (Coronado et al. 2009), online piracy (Fisk 2009), website filtering (Stol et al. 2009), cyberbullying (Shariff 2009) and online information privacy (Chai et al. 2009). Of interest is Chai et al.’s (2009) study on digital information privacy-influencing factors, which surveyed 285 pre-teens and early teens and found that children’s perceptions of the importance of information privacy and their self-efficacy influenced their information privacy behaviours on the Internet.

The years 2007 and 2010 were unproductive, while 2006 had only one article, 2008 had three, and 2011 saw eight research articles – signalling the growing interest in digital resilience research.

As build-up themes in digital resilience research, these studies expanded the scope of research to include Internet use safety and co-regulation, cyber risks, self-disclosure (SD) technologies, and factors like online privacy, cyber ethics, cyber safety, cybercriminal psychology and cybercrime (Kirwan & Power 2013; Lowry, Cao & Everard 2011; Marsden 2011; Martin & Rice 2011; Nigam & Collier 2011; Pusey & Sadera 2011; Slovak & Singer 2011; Valcke et al. 2011).

The results show a year-by-year alternating growth pattern: there was no growth in 2012, only 5 publications were produced, but it increased to 11 in 2013, demonstrating growing interest in digital resilience research. Publications increased steadily in 2014 with 16 works, but dropped to 15 in 2015. This alternating growth pattern persisted, but in 2019 and 2020, there was remarkable growth, with 32 and 51 articles, respectively, following a resurgence of interest in 2018 with 25 articles published. The coronavirus disease 2019 (COVID-19) pandemic lockdown may have spurred this newfound interest and growth, as people turned to cyberspace for social and professional needs (Dockery & Bawa 2020; Kaufmann, Straganz & Bork-Hüffer 2020). However, the research tempo slowed down in 2021, with 48 articles, as people’s lives readjusted to hybrid social habits post-COVID-19 (Alam, Rampes & Ma 2021; Smrdelj & Lenarčič 2024). Articles increased impressively to 54 and 75 in 2022 and 2023, respectively, but 2024 saw a low article publication (71) – indicating readjustment to year-by-year alternating growth. Only five articles had been recorded in January 2025. Given the steady growth trend, digital resilience research will continue to expand, as shown by the geometric progression of annual scientific output in Figure 2.

Table 4 shows the top 20 most cited references based on local and global citations. Among the analysed articles, Pusey and Sadera’s (2011) work on cyber ethics, cyber safety and cybersecurity, published in the Journal of Digital Learning in Teacher Education, received 20 local citations and 60 global citations and was the most impactful in the field of digital resilience. Also, Chai et al.’s (2009) work on online information privacy, published in the Institute of Electrical and Electronics Engineers (IEEE) Transactions on Professional Communication, had 16 local citations versus 93 global citations and had a significant impact.

Von Solms and Von Solms’ (2014) publication, with 10 local citations and 29 global citations, offers an informative guide to enable primary school teachers in Africa to teach cyber safety basics through a set of easy-to-use, self-explanatory and publicly available online video cartoons that can be used to discuss and engage learners in cyber safety principles. Rahman et al. (2020) article on cybersecurity education had 9 local citations and 71 global citations and was also influential.

With 7 local citations and corresponding global citations of 21, 348 and 14, respectively, other noteworthy works include Kritzinger’s (2016) work on cyber safety within schools; Lowry et al.’s (2011) study on privacy concerns; and Martin and Rice’s (2012) work on children’s cyber safety and protection. In this cohort, Kritzinger’s (2016) study focuses on the state of cyber safety awareness among South African teachers and pupils and suggests short-term initiatives, such as using posters, to raise awareness. Lowry et al. (2011) examined connections between the use of SD technology and culture and found that cross-cultural dimensions are important predictors of online awareness and privacy concerns, attitudes towards, intentions to use, and actual use of SD in instant messaging. Interestingly, Lowry et al.’s (2011) work has a high global impact (348 citations) versus local impact (7 citations), indicating more influence on other research fields but less so in the digital resilience field, versus the locally impactful work of Pusey and Sadera’s (2011) – 20 local citations. Therefore, citation results show heightened interest in and the influence of publications, as well as interdisciplinary opportunities, which will enhance the digital resilience research field.

The article titles can be summarised into five niche research areas of interest regarding adolescents’ digital resilience. The first theme focused on cyber safety with variations such as awareness, culture and gamification, education, interventions, policy, stakeholder perspective and digital skills inequality (Dodel & Mesch 2018; Kritzinger 2015, 2016, 2017a, 2017b; Martin & Rice 2012; Valcke et al. 2011; Von Solms & Von Solms 2014). Cybersecurity is the second research area, alongside education and teacher preparedness (Pusey & Sadera 2011; Rahman et al. 2020). The third niche focused on digital citizenship with technology perception, education and professional development (Martin et al. 2019, 2020, 2022). The fourth area is privacy in conjunction with youth behaviour, co-design, digital transparency and SD behaviour. The fifth area focused on social media literacy development and privacy behaviour (Heravi et al. 2018; Pangrazio & Cardozo-Gaibisso 2020).

This study also examined how the topics that scholars in the field of digital resilience discuss have evolved over time by analysing the authors’ keywords, temporal occurrence frequencies, co-occurrence networks, trends and thematic maps. Trending topics and scholarly focus within a field are identified by analysing publication keywords (Chen & Xiao 2016; Song et al. 2019). This exercise provides an overview of a scientific discipline, while the thematic progression map of scientific production organises the study topics into basic and motor themes (Chen & Xiao 2016; Melega 2022).

Figure 3 illustrates the development and connections among topics categorised into three analytic periods, spanning from December 2006 to January 2025. The line thickness indicates the importance of the relationships between the keywords, and the block size indicates the importance of the topic. Interestingly, only two topics were discussed most frequently during the 2006–2011 period: the Internet and privacy. In the second period (2012–2017), the Internet topic was maintained from the first period and, with a weaker connection, merged with the cybercrime topic. While the privacy topic has a strong connection to the security theme. Six additional topics emerged in the second period, including cyber safety, information privacy, social engineering, e-safety, protection motivation theory and awareness, indicating an exploration of digital resilience research.

There is a significant connection between the research topics of the second and third periods (2018–2025) because closely related themes fall under multiple research categories. Three of the second period themes – cyber safety, awareness and cybercrime – merged into more than one theme in the third period. The cyber safety theme was strongly connected to cybersecurity, but showed a weak connection to sexting. The awareness topic also merged with cybersecurity and privacy. While the cybercrime theme progressed to the third period, it also blended with the sexting topic. The security theme merged with a strong connection to the privacy theme only; the information privacy and the protection motivation theory themes merged into the privacy concerns topic; and the social engineering, e-safety and Internet themes merged with the digital competence theme with strong connections. The number of themes was reduced to six, three themes fewer than in the second period, indicating a stabilisation of basic terms in the digital resilience research field.

Cybercrime and sexting received less attention than digital competence, privacy concerns and cybersecurity, which were the most discussed themes and exceeded those of the preceding period, yet still indicated the development of basic themes in the digital resilience research field. Overall, these observations indicate the evolving trend of themes in the digital resilience field and its interdisciplinary nature, including digital competence, cybersecurity and sexting.

Figure 4 displays the dynamic view of keyword occurrences over time. Every coloured line represents a keyword, and the frequency of that keyword when the dataset was retrieved is included in parentheses. The field of digital resilience has gradually expanded in topics during the past 18 years, even though no new terms emerged in 2010, 2018, 2019, and 2021–2025. The terms discussed in 2009 include privacy, cybercrime and the Internet. The fact that privacy, cybercrime and the Internet first surfaced in 2009 indicates that these topics were essential in the development of the digital resilience field. Cybercrime, cyber safety, cyberbullying and security themes emerged in discussions of digital resilience in 2011.

The terms ‘cyber threats’ and ‘children and education’ emerged in 2012 and 2013, respectively, indicating a research focus on educational aspects and minor groups. In 2014, awareness, online safety and digital literacy were included in discussions of digital resilience research; in 2015, digital citizenship; and in 2016, information security and adolescents, showing research interests yet focused on minor groups. In 2017, social media was only introduced into the discussion of digital resilience research. In 2020, the terms COVID-19, privacy concerns and AI were added, marking the end of new-term additions in the digital resilience field. Notably, the terms adolescents, AI, cyber threats and digital literacy had the lowest occurrence frequency (8), while cybersecurity (55), cyber safety (37), cyberbullying (33) and privacy (32) were the most discussed terms in 2024–2025.

Common themes and potential avenues for future research can be suggested by co-occurrence networks (Napasti et al. 2025). Keyword relationships on the network of published terms that co-occur in digital resilience are visualised in Figure 5. Research on cyber safety, education, curricula and awareness channels, as well as topics such as cybercrime, cyber ethics, cyber threats, gender, children, adolescents, youth, parents, online safety and digital resilience, is reflected in the cyber security cluster. The goals of the research developed within the cybersecurity network cluster were to instil awareness and digital resilience in children, youth and parents, with a focus on gender, and to promote education on cybersecurity, online safety and cyber safety. The cybersecurity node also links with the COVID-19 node, which is interconnected with online and higher education nodes.

The cluster of cyberbullying displays research that focused on social media, sexting, child pornography, the Internet and NLP using machine learning as interventions. Through the link between cyber safety and cyberbullying nodes, cybersecurity relates to the cyberbullying research cluster. The privacy cluster focuses on terms such as security, safety and the Internet, which relate to the cyberbullying research cluster through the social media node. The privacy node also links with the cybersecurity node and the privacy calculus node. Furthermore, the privacy calculus node links to privacy concerns and information disclosure, demonstrating engagement with privacy calculus theory to address these issues. The privacy calculus node links to the information disclosure node, which, in turn, links to the privacy paradox node, indicating research interests in privacy–paradoxical behaviour in information disclosure.

The information disclosure node further links to the awareness node in the cybersecurity cluster. The teacher education cluster includes digital competence and e-safety, linking to the cybersecurity cluster through the teacher education and digital competence nodes. The information security node links to the self-efficacy node, which also links with the protection motivation theory, signifying research on self-efficacy through the lens of the protection motivation theory. The information security node also links to the cybersecurity node, indicating research efforts in addressing information security from the perspectives of self-efficacy and protection motivation theory. The teachers’ node links to the cybersecurity node and digital literacy, which links to digital citizenship, while digital citizenship and digital literacy are also connected to the cyber safety node. These links indicate research on the need for teacher training that includes digital literacy and citizenship. These findings outline research directions and growth, the diversity of research areas and the interdisciplinarity of the digital resilience field.

The top 20 trending themes are shown in Figure 6, signifying their influence on digital resilience. From 2018 to 2023, the cybersecurity topic trended, with 55 occurrences, making it a major topic of discussion in the field, along with cyberbullying (33 occurrences) and awareness (18 occurrences) in 2020. In 2019, cyber safety (37 occurrences) and digital literacy and cyber threats each had 8 occurrences.

The privacy topic has been discussed from 2019 to 2023, peaking at 32 occurrences in 2022 as a key topic of discussion in the research area. Other topics that peaked in 2022 include social media (23 occurrences) and children (19 occurrences). The cybercrime theme trended between 2014 and 2023, peaking at 20 occurrences in 2021; other themes included privacy concerns (11 occurrences) and AI (8 occurrences).

The digital resilience theme began trending in 2023, peaking at six occurrences in 2024, and NLP in 2023, while youth and child pornography had five occurrences, respectively, in 2014.

The thematic map of subjects contributing to the development of the digital resilience research field is visualised in Figure 7. Based on themes of centrality and density, the thematic map is divided into four quadrants (Q1–Q4). Driving themes or motor themes, are represented in the upper-right quadrant (Q1), indicating that they are well developed and important to the subject. Although they are less developed, the underlying themes (basic themes) in the lower-right quadrant (Q4) are just as important. The upper-left quadrant (Q2) contains specialised themes, also known as niche themes, which are associated with other, less central issues and are of lesser importance to the main research field. Lastly, themes that are transitioning or of marginal prominence are found in the lower-left quadrant (Q3). Each node clusters related themes, and only the topmost themes in each cluster are shown as labels, abbreviated when they overlap.

The first quadrant (Q1) displays well-developed subjects in the discourse of the digital resilience field, and the co-design node clusters cyber scams, interventions and qualitative research, to name but a few. Online education, confirmatory factor analysis and digital transformation are all clustered in the higher education theme. The machine learning node includes NLP, cyberbullying detection and bidirectional encoder representations from transformers. This node slightly transcends Q4, indicating that themes are still developing while encompassing fundamental concepts of digital resilience. Much like the machine learning node, the digital resilience slightly transcends into Q4 and includes information literacy, elementary education and intervention themes. The digital competence node includes teacher education, e-safety and cybersecurity education themes, and it also extends to Q4. Developing themes are presented in Q4, including curriculum, child pornography, information security awareness, cybersecurity, victimisation and privacy. These themes are still developing, yet are important in digital resilience research. The curriculum node includes secondary education, learning and computer science themes, and it transcends to Q4. This indicates that the themes in the curriculum node are still developing but have begun to structure the field. The cybersecurity node clusters cyber safety, cyberbullying and social media as developing but fundamental concepts. The child pornography node includes sexting, cyberspace and freedom of expression themes, and slightly transcends Q1 as still developing, but also contributes to structuring the field. The information security awareness node clusters gender studies and phishing themes, while the privacy topic clusters security, privacy concerns and information disclosure themes. The cybervictimisation cluster is emerging and evolving within digital resilience as the field transitions to Q3. The second quadrant (Q2) presents case-study nodes, including cybersecurity education, action research and active audience theory, all of which are also specialised topics in other fields. The digital practice node also includes young people as a specialised theme in other fields. Motivation, fake news and technology themes are presented in Q3 as emerging topics.

This study also identified the countries of origin, predominant themes, and leading authors reflected in the reviewed studies. Figure 8 presents a three-field plot, arranged in author countries (leftmost column), author-published keywords (middle column) and authors (rightmost column). With greater volumes on privacy, cybersecurity, social media, children, digital citizenship and cyber safety, the USA leads with up to 219 author keywords. With 114 themes, including a primary focus on privacy, children, education, cybersecurity, security and cybercrime, the UK follows the USA. Malaysia, with 104 themes, focused on cybersecurity, social media, cybercrime, cyberbullying, cyber safety and security. South Africa, with 80 themes, focused on cyber safety, cybersecurity and awareness, and other impactful countries include the Netherlands, India, China, Germany and Spain. Themes that still need attention include the Internet, adolescents, information security, awareness, online safety, cyber safety, privacy concerns and digital citizenship, as indicated by the sizes of their blocks.

This study used the Bibliometrix R package with the Biblioshiny web interface to evaluate data from the Scopus-indexed database. As a result, the author acknowledges the limits of relying on a single source and the language biases involved.

This study conducted a bibliometric analysis of the digital resilience research field over time, assessing trends, thematic focus and scientific productivity. The dataset was downloaded solely from the Scopus-indexed database because the Bibliometrix R package used in this work does not support integrating data from multiple databases (Agbo et al. 2021). As a result, the author notes the limits of relying on a sole source and potential language biases. As Ng et al. (2022) also note, effectively managing the ongoing rise in cybercrime to prevent, investigate and prosecute these crimes requires coordinating state tactics with academia to initiate research-driven policies and training. The findings highlighted research trends and a field-wide growth trajectory, with opportunities for interdisciplinary studies, spanning thematic domains such as cyber safety, cybersecurity, machine learning, fake news (disinformation), child pornography and digital competence to enhance digital resilience research. This analysis is valuable for informing researchers and stakeholders about opportunities for future research in areas such as adolescents, recovery capabilities, knowledge acquisition and adaptation within the digital resilience field. The research identified thematic discussions that also addressed privacy, cybersecurity, social media, children, digital citizenship, education, cybercrime and cyberbullying. Other themes that warrant attention in the realm of digital resilience encompass adolescents, cyber safety, information security, awareness, the Internet and emerging topics such as motivation, fake news, cybervictimisation, sexting and cybercrime. Wongmahesak, Singh and Kaunert (2025) note that the distinctive features of cybercrime victims have evolved, with younger individuals, particularly adolescents, becoming increasingly vulnerable. This vulnerability is evidenced by the rise in cyberbullying and online fraud among youths due to their frequent use of social media and gaming platforms. Therefore, policies, educational institutions, individuals and society must monitor and mitigate potential harm while maximising the benefits of cyberspace and the new norms in education, work and social life that technology brings. This necessitates changes in our educational approach, including a focus on cognitive abilities (such as critical thinking and problem solving) and safety; teaching cyber safety and digital resilience in curricula; and embracing modern technologies while keeping abreast of necessary risk-mitigation skills.