1. Introduction
Changing migratory patterns, increased urban concentration, ageing populations, and the structural transformation of economies have posed significant challenges to countries and communities worldwide. Recognising these converging trends, policy-makers have increasingly shifted their focus towards promoting greater social cohesion [
1,
2]. Often presented as a holistic and multi-dimensional concept, social cohesion is considered the glue that holds societies together and is seen as essential to address challenges and move together in a common direction. For instance, the Council of Europe [
2] defines social cohesion “as the capacity of a society to ensure the wellbeing of all its members—minimising disparities and avoiding marginalisation—to manage differences and divisions and ensure the means of achieving welfare for all members.” The concept is connected to numerous positive social outcomes in the academic literature, including economic productivity, environmental sustainability, greater social stability, increased peace, and increased physical activity [
3,
4,
5,
6,
7]. More recently, social cohesion has also been associated with an effective response to and recovery from the COVID-19 pandemic [
8,
9,
10].
Intellectually, there is a long history behind the concept of social cohesion that can be traced back to Emile Durkheim’s works in the late 19th century [
11,
12]. Since then, numerous authors from various academic fields have engaged with the concept of social cohesion, further expanding literature around the topic. As a result, the last 15 years have seen many works attempt to summarise, define, and further conceptualise the term. These efforts have led to a range of both narrower [
13,
14] and broader understandings of social cohesion [
12]. Within this body of work, what is most striking is the sheer range of (sub)dimensions associated with the conceptualisation and measurement of social cohesion. Shared values, shared experiences, civic participation, mutual help, trust in others, place identification, social networks, social order, acceptance of diversity, wellbeing, equality, and social mobility are but a handful of the dimensions considered by some to be constituent parts of social cohesion [
1,
12,
14,
15,
16]. In contrast, other authors push back against the sometimes expansive view of social cohesion, arguing that these conceptualisations confuse the core elements of social cohesion with its antecedents or consequences [
13,
17].
One of the main reasons for this expansive understanding of social cohesion is the significant amount of work done in different disciplines. This wide range of dimensions and sub-dimensions have become associated with social cohesion, as researchers have conceptualized social cohesion “based on the theoretical assumptions of their own discipline” [
11]. For example, psychology concentrates on processes within and between small groups [
11], whereas in anthropology, cultural practices and rituals are often at the centre [
18]. The high involvement of civil society and government actors further expands the perspectives present surrounding social cohesion. Over time, numerous thematic research or civil society networks have also emerged, including the International Migration, Integration, and Social Cohesion in Europe network [
19] and the Social Cohesion Hub [
20]. In turn, the scope of disciplinary and organisational contributions has created a rich but muddled research field and has made it difficult to get a sense of the structure or status of research on the topic. This creates the risk that a researcher may miss contributions and debates from other fields that could have impacted their work and supported greater theoretical development around social cohesion.
At the same time, the growing importance of social cohesion and significant work to define the concept point to a need to better map and understand this field of research. Numerous narrative reviews have been conducted on social cohesion, but these reviews have typically focused on selected disciplines, such as social science, policy [
12,
14], or psychology [
11]. Social cohesion has often been portrayed as a fluctuating quasi-concept that is bound to the assumptions of specific disciplines [
11,
21]. There has been a lack of broader, systematic approaches to map out the commonalities, connections, and differences regarding how specific disciplines engage with the subject. In turn, this can make the topic difficult to navigate and create blind spots or silos that can stifle innovation. Against this background, the following paper seeks to identify the structure of the research on social cohesion across the variety of disciplines that have tackled the subject and explore connections, silos, and key topics within this research. Though some recent research has mapped social cohesion research in relation to ideas of social vulnerability [
22], this paper aims for a much broader and expansive mapping that includes all disciplines that have investigated social cohesion, including psychology, sociology, political science, public health, and others. In particular, this study has two main objectives: (i) to identify and analyse the nature and evolution of literature related to social cohesion and (ii) to identify the thematic areas related to social cohesion research and their connection to specific disciplines. To achieve this, a bibliometric analysis of social cohesion research found through the Web of Science (WoS) was conducted. Such an approach was considered appropriate since social cohesion is a broad and large research topic that could not be easily mapped or evaluated through a manual, systematic literature review [
23]. Indeed, as Block and Fisch [
24] noted, bibliometric analysis can “structure a field and detect links between disciplines, identify topic clusters, literature gaps and academic silos, and show the most impactful authors and their research.” In the end, this work aims to give fellow scholars a roadmap to this increasingly important topic and contribute to identifying trends, gaps, and connections within the published research.
Moving forward, this paper progresses in three steps. First, the paper’s methodology is presented in more detail, especially as it concerns the search strategy and data analysis. Second, the results of the bibliometric analysis are presented, including overall publication trends, co-author analysis, and co-word analysis. Finally, these results are brought together and critically discussed to suggest future directions for social cohesion research.
2. Methodology
The bibliometric methodology encapsulates the application of quantitative techniques on bibliometric data and summarizes the bibliometric and intellectual structure of a field by analysing the relationships between different research components [
23,
24]. This data can serve to illustrate the contributions of specific disciplines, identify connections and silos, as well as identify trends and potential gaps [
23,
24]. As such, it provides both a science mapping and a performance analysis that helps establish the thematic evolution of a field of research [
22,
25].
Given this dual function, this method has increasingly been used to map out a variety of socially oriented, multidisciplinary fields, such as intellectual capital [
26], green marketing [
27], career success [
28], community resilience [
29], or agricultural policy [
30]. Likewise, as discussed above, the broad and multidisciplinary—yet still unstructured—nature of social cohesion research makes the topic suitable for this approach.
In this study, a process was established to determine the search terms, select an appropriate database, establish selection criteria for the search, select software for analysis, and analyse the results. These steps are presented in
Figure 1 below and described in more detail in the following paragraphs.
2.1. Definition of Search Terms
As described in the introduction, the goal of the present paper is to (i) identify and analyse the nature and evolution of literature related to social cohesion and (ii) to identify the thematic areas related to social cohesion research and their connection to specific disciplines.
This paper explicitly aims to explore the multidisciplinary nature of social cohesion research. However, it is also crucial to avoid including similar-sounding yet materially different concepts, such as team or group cohesion [
31]. Likewise, this study does not aim to capture cohesion as it is understood in specific fields, such as chemistry, geology, or computer science. Thus, for the purposes of this study, a single search term, “social cohesion,” was chosen to restrict results to the topic at hand.
2.2. Selection of Database
Web of Science (WoS) was selected as the database for this study. WoS is a selective, multidisciplinary, comprehensive database that covers the broad range of disciplines under investigation here [
32] and has been regularly used in bibliometric analyses elsewhere [
22,
27,
33]. In particular, WoS was chosen due to its disciplinary coverage, quality standards, and tools for data extraction/visualization.
2.3. Selection Criteria
A topical search was conducted on Web of Science using the search term “social cohesion” (title, abstract, author keywords, KeywordPlus; TS = (“social cohesion”)) on 1 November 2021. All articles, review articles, book chapters, and proceedings papers from 1994 to 2020 were included in the results. This time frame was chosen as social cohesion started becoming more widely used in literature and policy in the mid-90s [
22,
34]. The decision to expand the results beyond journal articles was taken in recognition of the significant non-journal contributions that are often cited in social cohesion texts, including from Jenson [
35] or Berger-Schmitt [
36]. Likewise, no restrictions on language were set. In short, the search term and parameters were chosen to reflect the range of research on social cohesion and the inherently broad nature of bibliometric analysis [
24]. Following the application of these criteria in WoS, bibliographic data were extracted in text (.txt) format, and no further data cleaning was performed.
2.4. Selection of Software
Two software were used to support the management and analysis of the data obtained. Microsoft Excel 2020, which is a common spreadsheet and data-visualisation programme, was used to manage data tables and generate figures related to publication trends, citation trends, top authors, most cited papers, top countries, top institutions, and the research disciplines engaged in research around social cohesion.
VOSViewer is a free software tool for constructing and visualizing bibliometric networks [
37]. This software was used to extract authorship, citation, and keyword data and perform co-citation, co-country, and co-word analysis [
37,
38].
2.5. Data Analysis
Data analysis was done in two parts. The first was a performance analysis that mapped growth patterns of publications; identified contributions made by countries, universities, and authors; and identified the most prominent journals related to social cohesion. The second component of the analysis focused on a science mapping that looked at the intellectual structure of the field through the construction of bibliometric maps [
23,
25]. In particular, here, the co-occurrences of author keywords, countries, and authors were examined.
Finally, a narrative review of keyword clusters was performed to complement this. To do this, at a minimum, titles and abstracts for the top 50 most cited papers featuring at least one of the top 5 keywords within a given cluster were reviewed. This analysis allowed for a structured summary of some of the key trends and findings within a cluster while still sensibly navigating the high amounts of content generated by bibliometric analysis.
4. Discussion
Through the use of bibliometric analysis, this paper has mapped out the structure, evolution, and key themes embedded within research related to social cohesion. Certainly, this paper is not without its limitations. The use of only one database may have excluded valuable results, especially considering that other databases (e.g., Scopus) tend to provide a greater breadth of results [
32]. Likewise, though language was not an exclusion criterion, the search term was not explicitly translated into other languages. This means that the academic contributions of non-English-speaking countries have likely been minimised. Finally, other software (e.g., Pajek, CiteNet Explorer) or analyses (co-citation analysis, three field analysis) could also expand the understanding of the research field. Nonetheless, some conclusions and potential further directions can be drawn out based on the results above.
First and foremost, the findings here illustrate the significant growth and attention given to the concept of social cohesion. In fact, the last five years of this analysis featured more than 50% of the retained publications and 57% of citations. This alone confirms the increasing recognition of social cohesion as an integral component of social progress, be it in relation to health, economic development, peace, or any other number of areas. Indeed, part of the growth in social cohesion research can likely be attributed to greater academic and political recognition of the multi-dimensional nature of social development. As MacFadden and colleagues noted, social progress is no longer solely explained by economic factors but also by a range of political and social factors [
22]. Relatedly, many of the most productive countries in terms of social cohesion research have developed policies and related programmes focusing explicitly on social cohesion, including in Australia [
97,
98], the United Kingdom [
71,
99], Canada [
35], and Europe [
2,
100]. These policies and associated funding may have also spurred the growth in research on social cohesion.
The recognition of the many socio-political factors behind social progress also manifests itself in the content of the publications, many of which explore the complex and intertwined realities of social cohesion, inequality, identity, education, health, and wellbeing. Likewise, the disciplines, keywords, and thematic areas investigated by the retained publications strengthen the contention that social cohesion is a highly multidisciplinary research topic [
11]. The multidisciplinary nature of social cohesion may be even greater than initially expected. Schiefer and van der Noll [
14] argued that academic “discourse takes predominantly place within and between the disciplines of Sociology, Political Science, and Psychology.” However, the literature identified here is significantly broader and prominently includes contributions from public health, education, economics and management. Overall, 197 distinct categories (Web of Science Categories) were identified, and other major disciplines, such as history, philosophy, or anthropology, also feature in the top 30.
The analysis of citations and keywords can also help draw some conclusions about the disciplinary nature of social cohesion research. Though there is this wide range of disciplines associated with the topic, research broadly coalesces around a few central themes. One research cluster seeks to explain how structural factors, such as inequality, economic development, or education, affect social cohesion at different geographic levels. A second cluster zooms in to the individual level and investigates how facets of identity, for example, at the ethnic, religious, or economic level, mediate social cohesion. Finally, a last cluster of research explores how (perceived) social cohesion affects subjective and objective measures of health and wellbeing. The structure of these different clusters provides a first mapping of research on social cohesion and has some valuable theoretical implications. In their review of social science research on social cohesion, Schiefer and van der Noll [
14] contended that inequality is an antecedent of social cohesion, whereas wellbeing is a result of greater social cohesion. The thematic clusters identified here and the content within the included publications certainly reinforce this. Numerous papers suggest that inequality and segregation are critical drivers of reduced social cohesion. Likewise, extensive literature in the health field ties higher social cohesion to any number of positive health or quality of life outcomes.
Yet, despite the connections between inequality, health, wellbeing, and social cohesion suggested here, there is an evident lack of disciplinary cooperation in social cohesion research. Though health-focused researchers regularly collaborate, social scientists appear much more isolated from each other, and there is little overlap between health and social scientists. One potential consequence of this disciplinary segregation is continued confusion and overlap between distinct concepts. As Van der Meer and Tolsma argued, there has been a “use of broad concepts, such as social capital, social cohesion and social trust, to denote widely different empirical phenomena” [
83]. This is reflected in the publications included here, especially in the health cluster, where certain concepts, such as trust or social capital, are presented as equivalent to social cohesion. Greater interaction and collaboration across disciplinary boundaries could contribute to further theoretical and conceptual precision and help dispel the notion that social cohesion is merely a “quasi-concept” [
21]. At present, however, it appears that research on social cohesion is very compartmentalized, and this may stifle the collaboration and innovation required to tackle multifaceted issues, such as social cohesion [
101].
Finally, the results here can provide some general future directions for research. At the keyword level, a number of terms are directly related to individuals or narrowly defined groups and their identities (e.g., immigrants, youth, race). Policy around social cohesion has been criticized for putting the responsibility for greater social cohesion at the feet of individuals already facing various forms of social insecurity [
102,
103]. As Lynch and colleagues [
42] noted, there has been limited discussion on how “focusing on what materially and politically disenfranchised communities can do for themselves may be akin to victim blaming.” Almost two decades later, this point was echoed by Nixon, who observed that many policy and research responses focus exclusively on marginalised groups and fail to address underlying structural issues [
104]. The top keywords identified here suggest that social cohesion research may also be somewhat guilty of this “victim-blaming.” Terms related to government or policy (e.g., policy, welfare state) are in the bottom third of occurrences amongst the 90 retained keywords. Moreover, other terms referring to groups in positions of power or privilege (e.g., businesses, wealthy people) do not appear. Though there have certainly been numerous authors who have analysed the role and impact of policy or government action on social cohesion, these results suggest that there is ample room for more empirical and theoretical work in this area. Likewise, there is fertile ground to explore how people in positions of relative privilege experience and influence social cohesion.