Abstract
The seemingly global nature of English-language hashtags often obscures activism from outside the Global North (GN). This systematic review explores geographic representation in this field (N = 315 articles) through an investigation of case study location, author affiliation, methods of data collection and analysis, and researched social media platforms. The results show a preponderance of GN/Majority cases and non-region-specific social media groupings such as hashtag publics, particularly in research employing digital methods. As such, extant research in the field has disproportionately produced what we term Northern Visibilities—groups and movements based in GN countries (above all the United States) and platforms popular within them. We use the findings of the review to critically interrogate notions of the Global South in digital social research and provide recommendations for rectifying geopolitical underrepresentation and promoting more inclusive research practice.
Introduction
Recent studies have problematized the dominance of the Global North (GN) within media and communications studies and called for a de-westernization of the discipline to incorporate more cosmopolitan perspectives (Badr & Ganter, 2021; Bosch, 2022; Waisbord, 2022). While some scholars argue that a “decolonial turn” in digital media research is occurring (Couldry & Mejias, 2021), it remains unclear whether this has shaped the study of digital activism (hereafter abbreviated as DA). Extant research suggests that “Western” social media platforms, as well as activism from GN countries, are over-represented within the field (e.g., Mahl et al., 2023; Özkula et al., 2022; Ruess et al., 2021). Similar arguments have been made about the research methods applied in communication research (Bosch, 2022; Schoon et al., 2020).
While previous meta-analyses and systematic reviews have explored either the relationship between Information and Communication Technologies and political engagement (e.g., Boulianne et al., 2023; Matthes et al., 2019) or the platforms being studied in DA research (Özkula et al., 2022), there has been no research to date exploring the geographic foci of empirical DA research. That is even though knowing where knowledge of DA is being produced, and by whom, is crucial to addressing issues of representation (and therefore power imbalances) within this field. It may therefore further democratize future research by allowing for a “critique of extractive theorising” (Dutta & Pal, 2020, p. 350).
This article sets out to address this gap by presenting the results of a systematic review of DA research published between 2011 and 2018. Specifically, it explores the representation of Global South (GS) and semi-periphery case studies in DA journal articles, the methods and platforms that feature in this body of research, and where the author(s) were based at the time of publication.
The article is organized as follows. In the first section, we explore definitions of the term “Global South” and assess the geographical representativeness of DA research to date. The next section presents an overview of the systematic review of DA research undertaken for this article, including a discussion of how geographic location was determined within this study. Finally, the results of the systematic review are presented and the implications are discussed. We argue that extant research on DA predominantly produces Northern Visibilities, irrespective of the methods used in these studies, although some differences in geographic representation exist between methodological approaches applied in the field.
Communication Research and the GS
It is something of a truism that the GS is underrepresented within the social sciences. Yet, the term is often used imprecisely and in ways that stereotype certain countries or regions (see Pinheiro, 2024). In geographic terms, the GS broadly refers to Latin America, Asia, Africa, and Oceania (Dados & Connell, 2012), or low- or middle-income countries within these regions (Clarke, 2018). Subject to a complex definitional history in line with evolving epistemic politics, the term has developed as a shorthand or “meta-category of world politics” (see Haug et al., 2021) for countries often described as developing, under-developed, BIPOC (Black Indigenous and People of Color), or “Third World” (see Haug et al., 2021; Khan et al., 2022; Pinheiro, 2024). In this way, the GS is a manifestation of long-standing geopolitical power dynamics and inequalities that privilege the colonizing high-income countries of the GN (Dados & Connell, 2012; Dutta & Pal, 2020; Haug et al., 2021; Khan et al., 2022). Beyond geopolitical regions, the term therefore acknowledges people’s lived experiences in extractive spaces of the GS (Dutta & Pal, 2020) as well as wider “moral geographies of inequality” (Pinheiro, 2024).
Although it continues to have currency among policymakers, GS remains a contentious term because it perpetuates dichotomies of developed-developing and colonizer-colonisee. It often fails to acknowledge the existence of “Norths” in the GS and “Souths” in the GN, as well as other minority or semi-periphery regions (Clarke, 2018; Milan & Treré, 2019). Hence the term “Global Majority” has recently been used as a deterritorialized descriptor for countries previously identified as being in the GN. Even so, some scholars have embraced the ambiguity of the term GS. Milan and Treré (2019, p. 325), for instance, characterize it as a “plural entity” with multiple and often conflicting meanings reflecting “geopolitical and sociocultural shifts.” While we acknowledge the limitations of the term GS in accurately reflecting global and local power differentials (e.g., compared to Global Majority) or middling countries (e.g., compared to semi-periphery), it is used here to reflect current discourses in the field. In this vein, this article will use GS as a shorthand to examine the geographical representativeness of DA research alongside other terms including Global Majority and the semi-periphery.
The underrepresentation of the GS has frequently been problematized in media and communications research, for example, in the African (Moyo & Mutsvairo, 2018; Wasserman, 2018) and Latin-American contexts (Gómez-Cruz et al., 2023; Medrado & Rega, 2023). Other studies have drawn attention to the western-centric nature of information and communication technologies, as demonstrated by the countries from which their designers originate (Arora, 2019). Much of this critical research has framed these digitally mediated exploitations as a manifestation of “data colonialism,” a term used to describe how user data are processed at scale to produce economic value (Couldry & Mejias, 2021, 2019). In the same vein, cognate theories of databased geographies’ (Arora, 2019), “technocolonialism” (Madianou, 2022), and “neocolonial media culture” (Bosch, 2022,p. 299) illustrate how the growth of “big data” has often been linked to the exploitation of those residing in the GS.
Problematically, knowledge about the GS is typically produced through the lens and paradigms of the GN (Waisbord, 2022). Wasserman (2018, p. 213) suggests that “media research in Africa needs to be conducted from the perspective of the Global South, rather than just using Africa as ‘raw data’ to support theories developed in the Global North.” Scholars have criticized the overall absence of South-to-South dialogues and theory from the South in this research (Dutta & Pal, 2020; Gómez-Cruz et al., 2023; Medrado & Rega, 2023). This issue is compounded by academic publishing processes that provide greater visibility to communities in the GN than the GS, for example, through journal preferences, common publication languages, editorial functions, as well as regional differences in publication methods, formats, and types (see, for example, Bosch, 2022; Nishikawa-Pacher et al., 2022). Non-white scholars also continue to be underrepresented within communication studies in terms of publication, citation rates, and membership of editorial boards (Chakravartty et al., 2018). It can therefore be argued that more media and communications research is needed not only on but also from the GS.
DA Scholarship and GS Visibility
This article focuses specifically on the visibility of the GS within DA research. Although its conceptual boundaries have long been contentious (see Kaun & Uldam, 2018; Özkula, 2021), the term serves as a useful descriptor for research at the nexus of protest, activism, and technology. DA can be broadly defined as the use of digital technologies for “political conduct aiming for reform or revolution by non-state actors and new socio-political formations such as social movements, protest organisations and individuals and groups from civil society, that is by social actors outside government and corporate influence” (Karatzogianni, 2015, p. 1). Digital technologies have facilitated the emergence of new arenas and repertoires for these actors while also creating new opportunities for activist engagement across the globe, especially in North America and Western Europe. In addition to communication studies, DA research is also situated in a wide range of disciplines including anthropology, political science, and sociology (Kaun & Uldam, 2018, p. 2099). In theory, digital social research should therefore have opened up the field for more case study diversity due to new spaces, actors, and practices in digital environments—actions that are often more readily accessible to researchers and can be researched remotely.
In particular, the evolution of “digital methods” should theoretically create more opportunities for DA research on the GS. These are methods that draw on natively digital objects, methods, and data (Rogers, 2019; Venturini et al., 2018). They are not “digitized methods,” such as interviews and surveys that have migrated online. Instead, digital methods typically draw on automated approaches or software to collect, interpret, or visualize data (e.g., social network analysis and data mining). However, it remains to be seen whether their use correlates with more representative DA research. The often prohibitive cost of these software packages, as well as their configuration for the study of “Western” social media platforms such as Facebook, Twitter/X,2 and Instagram, suggests that they are primarily intended for privileged GN researchers. Thus, while digital methods research may focus on the GS, it is rarely conducted by researchers from these regions. This may have implications for research findings too; early studies suggest that researcher location affects the granularity and type of data collected using software-based approaches; that is, remote research does not produce the same results as in-situ research (e.g., Borra et al., see Digital Methods Initiative, 2022). A related issue is that such research often omits relevant contextual information that may only be revealed through mixed methods or the addition of qualitative research conducted on the ground (Bosch, 2022; Digital Methods Initiative, 2022). This suggests that the impact of DA within the GS may not be fully captured by those studying it remotely using digital methods.
Social media platforms are also particularly challenging research sites for scholars seeking to redress the GN/GS imbalance. These platforms led to a “Data Golden Age” by making a vast trove of user data available to researchers (Tromble, 2021, p. 2). Yet, these data traces often do not contain content posted by users who choose to limit their visibility by setting their privacy settings to the maximum level (Ronzhyn et al., 2022). Furthermore, sites like Facebook and Twitter have the power to edit out marginalized communities through their moderation policies while also restricting academic access to these data through changes to their Application Programming Interfaces (APIs). For those studying DA in the GS, this means that the data available to them via online platforms is at best incomplete, and many voices may be absent from content collected via these sites.
It therefore remains to be seen whether DA research is subject to “digital bias,” skewed representations attributed to non-contextualized data selection, as well as restrictions in the data made available to researchers (Marres, 2017). Scholars have criticized the prevalence of Twitter data in digital research and the focus on hashtag publics (e.g., Burgess & Bruns, 2015; Hargittai, 2020; Matamoros-Fernández & Farkas, 2021; Özkula et al., 2022). While Twitter remains an important platform due to its use by journalists and politicians to break stories (Šimunjak, 2022), its affordances and patterns of use do not necessarily map onto other online platforms. Not only is its primary demographic highly educated and wealthy users in the GN, but the platform is rarely the primary or only communicative space for social movements or protest campaigns (Blank, 2017; Hino & Fahey, 2019; Lorentzen & Nolin, 2017; Özkula et al., 2022). More generally, a focus on the most popular sites such as Facebook, Twitter, and Instagram is often at the expense of alternative spaces of resistance (Bosch, 2022, p. 299). As such, researchers’ decisions about what platforms to research may mean they miss out on other data which is harder to locate and collect.
This article addresses these questions through a systematic review of DA literature. Previous meta-analyses in the field have tended to focus on the relationship between internet use and political participation (Boulianne et al., 2020, 2023; Chae et al., 2019) or the effects of media exposure on political engagement (Matthes et al., 2019). A range of reviews have also identified the methods (e.g., Borah, 2017; Neumayer & Rossi, 2016; Özkula et al., 2022; Proferes et al., 2021) and platforms (Pearce et al., 2019; Puschmann & Pentzold, 2020) commonly used in this field. This article adds to this emergent literature by exploring the geographical representativeness of empirical DA research published between 2011 and 2018 (= the years following the Arab Spring), with a specific focus on the visibility of the GS in this work.
Materials and Methods
Research Design and Hypotheses
This article presents the results of a mixed-methods systematic review of empirical DA journal articles (N = 315). While the review was largely quantitatively oriented, it was not a meta-analysis similar to those conducted by Shelley Boulianne and colleagues (Boulianne et al., 2020, 2023; Boulianne & Theocharis, 2020). Principles from high-quality systematic reviews were implemented to create focused research questions, hypotheses, clearly defined inclusion and exclusion criteria, and a review protocol (Cochrane Library, 2019; for full research protocols and PRISMA flowchart, see Özkula et al., 2022).
Based on extant literature, the following research question was set:
In support of this inquiry, three hypotheses were tested:
Sampling Strategy
The corpus was created by running queries spanning 21 relevant keywords describing digitally enabled activism (e.g., DA, online activism) on the Scopus database (see appendix in Özkula et al., 2022), and results were cross-checked against accumulated databases and top listings in Google Scholar. Scopus was chosen as the primary database due to its functionality which allowed several search terms to be searched concurrently and for filtering the corpus according to abstract, discipline, publication type, and year of publication. Queries were limited to searches of article titles, abstracts, and keywords, producing 2,668 articles within the “Social Sciences” category. The corpus was then filtered to leave only DA articles drawing on empirical data, which were published in peer-reviewed academic journals.
Coding
These articles were manually coded to identify the methods of data collection and analysis, including whether digital data was used as a source. For intercoder reliability, ambiguous cases were highlighted and discussed by the three coders before agreement was reached on how they should be coded (cf. Özkula et al., 2022). The authors and countries represented were recorded, with a specific focus on whether they were situated within the GS or semi-periphery regions. Coding categories included the choice of case study and location, the countries of authors’ institutional affiliation, and the platforms being studied. This helped capture the diversity and geographical representation of DA research while enabling a comparison between studies using digital methods and those deploying traditional approaches such as focus groups and surveys. While acknowledging its conceptual fuzziness, GS was used in this study to refer to countries that lack privilege according to development indicators provided by organizations like the World Bank. In this way, the term GS is used pragmatically for the purposes of this research rather than as an endorsement for the terminological GS–GN dichotomy.
Abstracts were coded manually in Excel according to their methodological attributes such as methods of data collection and analysis, digital data sources (i.e., studied platforms), case studies, and author affiliation. Articles were excluded if they included no information on methods of data collection or analysis, focused on formal political actors such as governments, parties, or processes (e.g., elections), or did not have DA as their primary focus. Based on these criteria, the final corpus consisted of 315 articles published between 2011 and 2018 (coding protocol: Özkula et al., 2022). For comparison, these articles were split into (a) the traditional dataset (TD): 117 articles that used exclusively traditional research methods (e.g., interviews and surveys), accounting for over a third of the entire dataset (37.1%), and (b) the digital dataset (DD): 198 articles which included some form of digital data as their source (but not necessarily solely), accounting for 62.9% of the dataset.
Methodological Challenges
A number of challenges need to be acknowledged. First, Scopus was used as the primary database for collecting articles and this may have affected the results to an extent. Tennant (2020) provided evidence that academic search engines and indexing sites such as these are not strictly global databases of knowledge as they are biased against non-Western and non-English language scholarship. These latent database dynamics suggest that article visibility is subject to many of the GN–GS biases discussed earlier in this article, an effect that cannot be mitigated by cross-referencing the results with those generated by other academic search engines. Second, differences in methodological writing also meant that certain information (e.g., API access, use of software packages) was sometimes omitted and had to be judged on the basis of contextual information contained in the article. Third, articles excluded due to missing information might have been formatted as such due to the requirements of specific journals, disciplines, and local publishing conventions. Finally, there were ambiguities in terms of defining DA, what constituted empirical research, and which countries were located in the GS, that were addressed through discussion between the coders. These interpretations played a key role in determining whether articles should be included in the review, and were therefore based on judgments about these topics. The use of three coders strengthened the decision-making vis-a-vis the coding process and enabled more robust judgments to be made on whether articles met the inclusion criteria and their respective characteristics.
Results
H1: Empirical DA Research Predominantly Focuses on Case Studies From GN Contexts
Case Study Country Distribution Across the Entire Dataset (N = 315)
Across the dataset, most articles featured case studies located within GN countries. A clear majority (265 articles, or 84.13 %) focused on a case study based within one country. In total, 56 countries were referred to in these articles, with the four large economies of the Anglosphere—Australia, Canada, the United Kingdom, and the United States (US) featuring prominently. Analysis of the top 10 countries showed that the United States was the most frequent location in these articles (30.6%), followed by the United Kingdom (4.9%), Egypt (4.5%), and Hong Kong (4.2%) (see Figure 1).
Top 12 countries that featured in DA articles using one case study country (N = 265).
While GS and semi-periphery countries like Nigeria, Turkey, and Taiwan also featured in the corpus, they were largely in the minority. Indeed, over three-quarters of the countries identified (43, 76.8%) featured fewer than five times, and just over a third (21, 37.5%) appeared only once in the dataset. The latter included advanced economic countries in Western Europe such as Finland and the Netherlands, as well as those in the GS such as Sudan, Tanzania, and Uganda.
A similar trend was evident in the small number of articles (N = 21) focusing on DA within more than one country. Once more, the US was the most common case study country (7 articles, 33.3%), followed by the UK and Spain (in 5 articles each, or 23.81%). Overall, the “Big Four” accounted for 47.6% of the multi-country studies, including several articles which directly compared DA from at least two of these Anglosphere countries.
These trends were also reflected in continental and regional distributions. North American case studies, that is, Canada and the US were dominant in the corpus, followed by Asia, Europe, Africa, Latin America, and Oceania (see Table 1). These findings demonstrated the strong representation of GN case studies in the dataset, though considerably more so in the case of North America than Europe.
Regional Distribution of Articles in Corpus Using a Single Case Study Country (N = 265).
It was not possible to attribute case studies to specific countries or regions in a minority of the articles analyzed (22, or 8.3%). In cases such as #BLM (Black Lives Matter), many DA articles were coded as “n/a” in terms of region because they focused on hashtag publics that were not explicitly linked to a specific geographic region. In contrast, #ferguson (part of the BLM movement) indicated a US focus and was reflected in the characteristics of the data analyzed. A very small number of articles (5, or 1.9%) also focused on regions such as MENA (Middle East and North Africa) or the European Union but did not identify specific countries within these entities.
Overall, H1 was proven accurate in light of the number of case studies in the corpus drawn from the US and other global majority regions.
H2: Empirical DA Research Drawing Exclusively on Traditional Data Sources is Less Diverse in Terms of Geographical Focus Than Those Using Digital Data Sources
Case Study Country Distribution: DD Versus TD
The dominance of the GN was particularly evident in the DD compared to the TD. The former (N = 198) had a higher proportion of articles focused on the US than the latter (34.2% compared to 25.2%) and the Big Four overall (46.2% compared to 32.7%; see Table 2). The TD also displayed a higher degree of regional diversity in case study countries through a larger proportion of single-country articles (91.3% vs. 79.8% in DD) and a higher number of individual countries studied (31.8% vs. 27.9% in DD), above all in Asia and South America. As such, those DA articles applying traditional methodologies were more likely to focus on case studies outside the GN than those involving the study of digital data.
Sample Comparison: Traditional Versus Digital Dataset.
Digital Methods Approaches in Comparison (N = 73)
A further stage of coding was conducted on a subset of the articles which relied exclusively on digital methods for data collection (N = 73; i.e., software-based methods drawing on natively digital methods, tools, and data). The vast majority of these were single-platform studies (84.9%), with 72.6% of these primarily focused on Twitter or using its data alongside other digital sources. Hashtags accounted for nearly half of these as the primary focus (48%) and a further 20.6% considered them as part of their wider methodological focus. As such, articles using digital methods were more likely to focus on data collected from Twitter (and particularly hashtag networks) compared to the other articles relying on exclusively digital data (which also included methodological approaches such as digital ethnography).
In terms of regional distribution, articles using digital methods correlated with a higher preponderance of US case studies (30.1%) compared to the traditional dataset (25.2%). However, there were few differences between the digital methods dataset and TDs, including in the number of articles that had non-region specific case studies (10.1% in TD compared to 10.2% in the digital methods dataset). In this way, DA articles based on digital methods did not appear to be any more or less diverse in terms of their geographical representation than those using other internet research methods (e.g., qualitative research online). Nevertheless, as discussed earlier, Twitter’s demographic profile is largely privileged, English-speaking, and US-based. Twitter network studies were particularly prevalent among articles using digital methods. This suggests that methodological approaches relying on digital methods in data collection show a higher preponderance of the GN, particularly where no regional focus is indicated. These findings are congruent with H2 and suggest that, beyond the TD-DD distinction, articles using digital methods in data collection largely (and even more so than articles using traditional data sources) represent the GN.
H3: Empirical DA Research is Primarily Published by Lead Authors Based in GN Institutions
Author Affiliation
The analysis of the institutional affiliations of lead authors provided further evidence of how DA research published during this period was viewed through the lens of the GN. Scholars based in US research institutions accounted for just over a third of the articles in the entire dataset (36.5%), followed by the other Big Four countries: United Kingdom (10.2%), Australia (7.0%), Canada (4.1%), and then a mixture of global majority, semi-periphery, and GS regions (Germany, Hong Kong, the Netherlands, and Singapore/Sweden; see Table 3).
Lead Author Affiliations in Articles (Digital and Traditional Datasets), Descending by Percentage.
Some differences were observed between the TD and DD. The lead authors of articles using digital data sources were more likely to be based in universities and research institutes in the GN (e.g., US, United Kingdom, Australia, Germany, the Netherlands, and Sweden) than those using only traditional data (see Table 3) despite a similar number of countries of affiliation. This was particularly evident for articles using digital methods, with 41.1% authored by researchers based in the US, 11% from the United Kingdom, and 8.2% in Germany. This suggests that digital social research in the field of DA remains the domain of privileged regions and institutions.
Given the increasing mobility of scholars worldwide, as well as the absence of information pertaining to their background in the articles under review, it was not possible to fully assess each lead author’s relationship to either GN or GS. Therefore, additional attention was given to the institutional affiliations of contributors to papers with more than one author.
Only 9.2% of the articles in the entire corpus were authored by international teams of scholars that included one or more members based outside GN countries (see Table 4). Instead, the majority were produced by either single authors (43.5%) or multiple authors from the same institution (35%). As such, collaborations between authors based in different countries were not well-represented within DA empirical research published in academic journals during this period.
Authorship of Empirical DA Articles Published Between 2011 and 2018.
Articles produced by international collaborations tended to have first authors based in US institutions (20.7%), rather than other GN countries. Analysis of the other 23 countries in which lead authors were based revealed that no country featured strongly (maximum two articles or 6.9% per country). A similar US dominance was evident in papers with more than one author based within the same country (45.1%). While the same number of articles (50) were co-authored by scholars based in other countries, no individual country featured as prominently as the US. Although this can in part be explained by the strong overall representation of US-based scholars in the field, it also demonstrates that DA scholarship is largely clustered within the US.
Discussion
Measuring GS Representation
The review provided strong evidence that the GS is underrepresented within DA research published between 2011 and 2018. Even so, the exact extent to which this is the case depends on whether the GS is conceptualized and operationalized on the basis of geographic location or geopolitical dynamics of socio-economic power or privilege, as well as the classification systems used for these.
In particular, classifying what constitutes the GS proved challenging, as its exact representation in numerical terms depends on the applied classification. To illustrate, if a conventional continental distribution is applied, continents more commonly labeled as GS regions (South America, Africa, and Asia) feature less than North America, but in some cases more than Europe and Australia. In part, this is a reflection of country distribution across traditional continental classification, where some advanced economies are situated in continents that are not typically labeled as global majority (e.g., New Zealand in Asia), or vice versa, particularly in the absence of more nuanced regional descriptions (e.g., Australasia or Oceania, cf. Table 1). It is also, in part, a result of the difficulty of weighting geographic areas by the number of countries covered in a given continent and how many of these countries are covered in the dataset; that is, the terms GN and GS are not representative of any particular geographic distribution such as population or land mass.
An underlying issue is that the GS-GN split may be understood in terms of geographic, developmental, or power-based factors. Continents and regions constitute fairly vague indicators of privilege or power. In comparison, if the GS is conceptualized based on development status (again a classification subject to differing interpretations, see Hoffmeister, 2020), geopolitical classifiers such as the United Nations or World Bank lists of development (as per Clarke, 2018; e.g., United Nations, 2023) show higher disparities. For example, in the “DAC List of ODA Recipients” (= countries and territories eligible for Official Development Assistance; Organization for Economic Co-operation and Development [OECD], 2023), ODA recipient territories are categorized by LDCs, that is, low-income, lower-middle-income, and upper-middle-income countries. Applying this classification, case studies pertaining to the least developed country category barely feature in the corpus at all.
Issues of power and privilege are also pertinent to authorship. GS scholars may well be represented within institutions in the GN, a judgment that cannot be established on the basis of a systematic review. Equally, both GS and GN scholars may be collaborating with local scholars and industry contacts, an aspect not always disclosed in publications. It also cannot be presumed that all those working at universities in the Big Four are by default privileged—whether they are citizens in or originate from these countries. Further work is therefore needed to understand the ways in which GS and semi-periphery experiences are integrated and applied in DA research. Even so, GN scholarship remains relevant in relation to digital methods, as advanced digital infrastructures, digital methods training, and software costs associated with these often lie in the domain of privileged institutions.
The ambiguities arising from the different classifications gave rise to several provocations. In particular, they suggest that an evaluation of GS representation in DA research depends on several factors: (a) what countries and regions are classified as GS (= a definitional question); (b) judgments on the relevance of marginality and power relations to understandings of the “Global South” (= an epistemological, political, and philosophical question); (c) the role that platforms and their affordances play in the visibility of certain social groups (= a question based on user behavior and platform transparency); and (d) the extent to which author affiliation (rather than ethnic origin or international experiences) represents the GS or GN.
Beyond classification, the review demonstrates that many countries across the globe are barely represented in the field of DA. The number of countries from which case studies were drawn (TD: 29; DD: 36) was relatively low given that there are 193–195 recognized states within the international system today. Furthermore, the number of countries represented in the analysis of author affiliation accounted for just over a quarter of 104 countries whose universities were included in the 2023 World University Rankings (Times Higher Education [THE], 2023). Further evidence of the unrepresentativeness of DA research to date was evident in how few movements within GS countries were analyzed in the corpus, as well as the dearth of research on platforms widely used in these contexts (e.g., Sina Weibo, WeChat, VKontakte, or Douyin).
Judgments and inferences about representation are also technically “easier” when it comes to DA research involving traditional methods rather than research drawing on digital methods, as more detail is provided in terms of sample recruitment and participant background. Regional diversity in traditional data sources (a) often relates to more easily identifiable, geographically anchored social movements, protests, and activist groups (compared with looser-tie digital networks) and (b) often involves a higher degree of informational insights into participant background—in qualitative research typically through direct participant contact; in quantitative research through targeted demographic questions to participants. While the results indicated that DA articles involving the use of digital methods were even less diverse than articles using traditional methods, it is debatable whether such claims can be fully verified given that many of the networked movements using digital methods were global in nature and not linked to specific locations.
Even so, common entry points in digital methods research were hashtags or other tag-based networks, meaning that software-based approaches are by design more likely to capture visibility-oriented contents (at least where searches are tag-based), and are therefore less likely to capture alternative spaces of resistance where marginalized groups may be using code language or other practices of obfuscation. While this raises questions as to whether tag-based approaches allow for an adequate capturing of marginalized groups, it is also questionable whether regional diversity in digital datasets can be accurately judged on the basis of country, rather than, for example, the platform, platform subspace, community forum, language, and/or the demographics that inhabit these. Although the same may apply online, it requires closer collaboration with individual spaces and their users. Very often this insight into participant background is missing from social media data gathered solely via digital methods. It also suggests that the notion of geopolitical regional boundaries in digital spaces may need to be reconsidered given their complexity.
Northern Visibilities
These findings raise questions about the value of using overly simplistic dichotomies such as GN versus GS in a field where researchers are often unable to establish the background of participants and case studies that move beyond national borders. Regardless of classification, results indicate that DA research (a) revolves around the study of groups and movements based in GN or global majority countries (above all the Big Four), (b) relies predominantly on data from platforms that are popular within these contexts (e.g., Facebook and Twitter), and (c) tends to focus on highly visible practices (e.g., hashtag activism). We define these skewed representations as Northern Visibilities. This term highlights a range of injustices in GS representation stemming not only from methodological choices, but also the underlying socio-economic and political conditions that privilege GN scholars and institutions in communication research, for example, the dominance of GN media practices and platforms, Western publishing conventions, and paywalls in publishing and software acquisition. These render them more visible than their colleagues in the GS.
In doing so, this review follows in the tradition of de-westernization scholarship that critiques Western dominance in literature. Despite the advent of digital methods, the gaze in this field remains firmly on activism within the GN. There is no decolonial “turn” or shift in focus within DA research as a result of digital methods being more freely (or cheaply) available. Indeed, it could be argued that recent changes in how data are licensed by platforms such as Twitter (now known as X) and Facebook are likely to consolidate the hegemony of wealthy institutions in the US and Western Europe within the field. Overall, the study suggested that digital methods perpetuate Northern Visibilities within DA research rather than provide opportunities for a greater focus on the GS from researchers based in these regions.
Conclusion
Reconsidering DA Research
The findings of this review suggest that researchers may need to rethink (a) where (e.g., which platform spaces) activists from the GS and semi-periphery are active; (b) in what ways digital social research and particularly digital methods may be employed to not perpetuate Northern Visibilities; (c) gaps in DA research, above all in relation to “unheard” groups; as well as (d) what these skewed representations mean for inclusive research practice. Consequently, more DA research should also be conducted in situ, that is, from within the studied locations (here, the GS and semi-periphery), using traditional qualitative or quantitative methods as well as situated digital social research.
Limitations
A number of factors could not be considered within the scope of the review. They include the language of analyzed data and researchers’ ties to local networks. Knowledge of the latter in particular would carry relevance for judgments on whether Northern paradigms are applied in localized contexts. This review also could not adjust for certain effects that contributed to underrepresentation, such as publication conventions, difficulties in situating or retrieving subaltern or non-anglophone data, or the specific reasons (such as the availability of training and funding for data acquisition) for the dominant use of digital methods in the GN. Unfortunately, it was not within the scope of this review to determine the primary reason for these distributions. Other limitations relate to the data coding in that some data carried ambiguity for coding. Here, other aspects were taken into consideration for an evaluation of GS presence.
Future Research
Building on the results presented above, future scholarship should explore whether efforts to de-westernize empirical DA research in recent years have been successful (or not). This is particularly pertinent given that new platforms such as TikTok have become increasingly integral to DA; there has also been a wave of protest events in the intervening period, including anti-COVID-19 lockdown marches and the mass street demonstrations following the murder of George Floyd in May 2020. Future research should also empirically explore the backgrounds, motivations, and challenges faced by authors of empirical DA research. Interviews and surveys should probe some of the key findings from this systematic review, such as their relationship to the GS, challenges faced when trying to publish work on case studies from outside the GN (e.g., peer review, language issues), and the resources made available to them by their respective institutions to support them.
We wish to conclude by proposing a future research agenda for DA researchers which redresses this imbalance in favor of DA research from and within the GS. Specifically, researchers should consider how research designs and reporting can be adapted toward engaging with activism on the ground, contextualizing data, and gaining insights into what types of demographics occupy specific platform spaces. Such research would be more inclusive of hitherto marginalized activist voices as well as better capture how DA operates in contexts outside the GN. Above all, future research would benefit from (a) more reflexivity on regional representation in research methodologies, (b) more insights and granular data on the specific contexts of data collected via digital methods and how this relates to populations and/or regions, and (c) a stronger focus on working with communities.
Footnotes
Acknowledgements
The authors would like to thank Jenny Hayes, the reviewers, and participants at international conferences such as the ECPR and AoIR general conferences for their helpful comments.
Author’s Note
This research was conducted at the University of Trento (Italy) where the lead author was based at the time.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: S.M.Ö. is supported by funding from the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement number 101027274.
