Sage Journals: Discover world-class research

Abstract

A crisis is conceptually defined as an unexpected situation requiring an urgent solution. The global climate crisis is especially significant, threatening the stability and livability of the world. Building on this conceptual framework, this study aims to explore the dimensions of stakeholder interactions to highlight the need for collective efforts to combat the climate crisis. Twitter (now known as X) interactions between climate and environment ministries in OECD countries were analyzed in R using graph theory. The study found that the United States ministry was the most followed. As the largest network node, Canada acts as both a gateway and hub. As a central player, Canada receives and disseminates information and possesses a high potential for interaction. France is connected to influential nodes and is therefore a leader in the chain of influence. Germany is a strong center of stability and acts as a leader/bridge node. Norway, Spain, and Greece are considered to be peripheral nodes. Furthermore, the network had a medium-strength community structure divided into four parts. These results highlight that interactions among OECD countries are insufficient and that climate and environment ministries can facilitate global cooperation to combat the climate crisis. Strengthening nodal connections is a recommended policy step to translate these results to real-life and industry-specific contexts.

Plain Language Summary

Study on social media interactions between climate ministries in OECD countries

Why was the study done? The climate crisis is a major global issue, and countries need to work together to address it. The study aimed to understand how climate and environment ministries in OECD (Organization for Economic Co-operation and Development) countries interact with each other on social media, particularly Twitter. By looking at their interactions, the research wanted to see how these countries could collaborate more effectively to tackle the climate crisis. What did the researchers do? The researchers analyzed the Twitter accounts of climate and environment ministries from OECD countries. They used graph theory, a method that shows how things are connected, to see how often these countries’ ministries interact with each other on Twitter. They looked at the number of followers, the number of tweets, and who follows whom to understand the level of engagement and cooperation. What did the researchers find? The study found that the United States was the first country to create a Twitter account for its ministry and had the most followers. The United Kingdom had the highest number of tweets. However, many countries were not very active in interacting with other countries’ ministries. For example, Austria, Ireland, Luxembourg, and Türkiye had little or no interaction with other countries. The study showed that some countries were much more connected than others, with Canada and France having the highest number of interactions. What do the findings mean? The study shows that there is room for improvement in how climate ministries interact with each other on social media. The climate crisis requires collective global action, but the research suggests that interactions between ministries are currently limited. By increasing communication and cooperation through social media, these countries could work more effectively together to address the climate crisis.

Keywords

digital communication climate crises OECD countries graph theory

Introduction

Crises are part of everyday life, affecting both individuals’ business and social lives. Several contemporary crises have had massive impacts on humanity. Many have become chronic, and the necessary steps have not been taken to resolve them. The climate crisis is among the most important of these.

Many studies have been conducted on crises. Campiranon and Arcodia (2008, p. 159) examined market segmentation in times of crisis within the meetings, incentives, conventions, and exhibitions (MICE) sector in Thailand. The MICE industry is strongly affected by crises, and market segmentation plays a key role in attracting tourists to this sector. In addition, Kim et al. (2005, p. 369) discussed the negative impact of the Severe Acute Respiratory Syndrome (SARS) epidemic on the Korean hotel industry. Furthermore, Butler et al. (2016) considered the water crisis in terms of the fundamental right to access clean drinking water. Kaewkitipong et al. (2012) discussed the connection between social media and crisis management, emphasizing the importance of using social media to meet different information needs and conduct successful crisis projects.

The literature also includes many studies on the environment. For example, Sun et al. (2025) theoretically considered how digital enhancement affects the sustainable development of green public services through a cross-analysis of Structural Equation Modeling (SEM) and Artificial Neural Networks (ANNs). Di, Chen, Shi, Cai, and Liu (2024) provided guidance for policymakers aiming to stimulate sustainable consumption behaviors by through structural equation modelling and artificial intelligence analysis and synthesis methodology. Furthermore, Sun et al. (2024) presented industrial carbon reduction strategies by analyzing the synergistic contributions of digital technology applications, green supply chain management, consumer behavior, and corporate social responsibility to development.

Specific studies on climate crises have also been conducted. Bryant et al. (2015) analyzed accumulation and displacement dynamics in the Clean Development Mechanism (CDM). The study identified the CDM as a solution to the ecological crisis of climate change, securing production conditions for the fossil fuel industry and promoting new accumulation opportunities for other companies. Fraser et al. (2020) examined whether local governments’ social capital and public services improve environmental policy outcomes considering 1,741 Japanese communities from 2005 to 2017 and modeling greenhouse gas emissions in four sectors. Engstrom et al. (2020) analyzed both the COVID-19 and climate crises from a public health perspective, showing that climate policy regarding green infrastructure projects, reforestation, and the reduction of labor taxes supports economic recovery.

However, this study focuses on stakeholder interactions during the crisis phase, aiming to determine which connections between actors influence the climate crisis. Collective action among stakeholders is clearly needed in the solution phase of this crisis. Hence, the relationships and information shared between actors are important. This study considers Twitter (now known as X) interactions among climate and environment ministries of Organization for Economic Co-operation and Development (OECD) countries. The sample therefore consisted of ministries from OECD countries with Twitter accounts. As the world’s largest economies, these countries’ activities contribute the most to climate change. For this reason, this study focuses on climate change drivers that arose during the sampling phase. Existing climate change crisis literature has focused on individual actors (Becken et al., 2021; Shim, 2024) and specific social media movements (Padilla-Castillo & Rodríguez-Hernández, 2023), or on specific elements such as social media news (Hopke & Hestres, 2018) and social media post content (León et al., 2022). This study aims to address gaps in the literature by focusing on official institutions’ social media interactions and their meanings in terms of the climate crisis. In addition to providing a theoretical discussion, this study presents a cross-sectional picture and situational analysis. It explores the importance of public communication in combating the climate crisis, focusing on the interactions among high-contributing countries. It also discusses the benefits of increased public interactions in combating the climate crisis. To achieve these aims, this study employs qualitative research methods and graph theory.

Literature Review

Historically, many societies have faced crises. These crises should not be viewed only as consequences of management decisions, as they also encompass social diseases, natural disasters, etc. They can occur suddenly due to many factors. The term “crisis,” which originated in ancient Greece, refers to decisions between right and wrong, or between life and death, in the context of law, medicine, and theology. This original meaning began to change in the 17th century when the term was incorporated in politics, economics, history, and psychology. Because of its metaphorical flexibility, the term “crisis” is used in the decision-making and selection phases in almost all areas (Koselleck & Richter, 2006, p. 358).

Crises pose a threat to people and institutions, creating conditions that are inadequate (Bülbül, 2004, p. 73) and placing pressure on them to identify solutions when least expected (Balta Peltekoğlu, 2004, p. 313). More generally, a crisis can be defined as an experience that attracts collective attention, compels action, generates affect, or as an observation that generates meaning (Masco, 2017, p. 565). The term is applied to many areas of life and is associated with concepts such as uncertainty, the need for action, threats, inadequacy, time pressure, and unexpectedness.

According to Parsons (1996, pp. 26–27), crises can be divided into three categories: immediate, emerging, and ongoing (persistent). Immediate crises offer little to no warning of their approach. Emerging crises are also somewhat unpredictable. Even when an emerging crisis is predicted or identified in its early stages, acknowledging that it is a major problem is difficult. Permanent crises, on the other hand, extend over weeks, months, and years. Speculation and rumors surrounding permanent crises play important roles and are sometimes fed by the press. Gundel (2005, pp. 110–112) identifies four types of crises: conventional, unexpected, persistent, and fundamental. Conventional crises have predictable and known probabilities of impact. Dangerous engineering systems typically produce these crises, and natural disasters are rarely included in this classification. These crises are considered manageable, even if they cause significant losses. Unexpected crises are rarer than conventional ones. Their impact is greater and preparation for such crises is inadequate. Intractable crises can be predicted, but they are difficult to suppress. An example is the tragedy of the 1985 European Cup Final. Fundamental crises are the most dangerous type, as they cannot be prepared for and are initially poorly understood. Although rare, their destructive potential is high.

Crises are important turning points for both society and institutions. From an institutional perspective, crises fall within the realm of management science. Many institutions are confronted with crises for various reasons. However, crises must also be managed socially. There have been several important global crisis situations, particularly in public health. Various diseases have spread across regions (Masco, 2017, p. 565) with regional and global consequences The COVID-19 pandemic is one of the most serious examples. Others include the Ebola epidemic and antibiotic resistance.

Crises are driven by stakeholder expectations (Coombs, 2018, p. 52). Regardless of the type of crisis, certain people or interest groups are affected by it. Activities to combat crises should consider their current characteristics, rather than relying on a general approach. It is not possible to resolve every crisis similarly. The nature and extent of crisis impacts differ, highlighting the complexity of crisis management. Particularly in turbulent environments, it is necessary to be cautious and prepare for alternative scenarios (Khodarahmi, 2009, p. 524). Many variables are important in crisis management, especially communication. The information flow established between stakeholders plays a fundamental role in crisis situations (Khodarahmi, 2009, p. 524). Crisis communication is the strategic use of words and actions to manage information and meaning during a crisis (Coombs, 2018, p. 51). In this context, communication determines the relationship networks that are utilized to manage and end crises through collective efforts.

The climate crisis is particularly pressing. The debate is no longer about whether climate change will occur but about its timing and extent (Crist, 2007, p. 29). As CO2 emissions increase, achieving the climate goals set at the Paris Meeting becomes increasingly difficult. Although the climate crisis has been foreseeable for years, CO2 levels increased significantly from 1990 to 2018 (Banister, 2019, p. 565). The 26th United Nations (UN) Climate Change Conference (COP26) has foregrounded the climate crisis, but increasing CO2 levels highlight the inadequacy of climate crisis responses.

Considering its widespread use, social media could become a significant channel for addressing the climate crisis. Social media may also play an important role in international environmental management. From the perspective of individual actors or civil society organizations, social media facilitates information gathering, interaction, and participation in policymaking. Civil society organizations and individuals have begun using social media to mobilize for environmental protection, shed light on various issues, and advocate for political reform. Government social media is important for fostering environmental awareness, international co-operation, and sharing policy developments. It bridges the divide between official environmental protection agencies and the public, facilitating the participation of non-governmental actors. Official environmental agencies are increasingly using social media, giving citizens access to information and opportunities to learn and participate at different levels in a wide range of environmental issues, including natural resource use, climate change, and biodiversity (Chang et al., 2024, p. 44).

Díaz-Díaz and Pérez-González (2016, p. 104) identified social media as an important tool for civil society because it sets the political agenda and influences how discourse is framed. While social media is primarily used for information sharing, Sinclair et al. (2017, p. 148) note its uses to initiate, sign, and send public petitions regarding employers and the local environment. Stevens et al. (2016, p. 104) define social media as a medium in which various factors interact to influence agri-food sustainability management. They emphasize that social media’s networked structure provides a public platform for information dissemination and interactive communication, as well as a decision-making resource. Nulman and Özkula (2016, pp. 11–12) demonstrate that digital tools can facilitate direct communication on environmental issues, disseminate information, generate interest in mass media, make calls to action, and drive engagement. They state that increased use of digital tools, such as social media, directly strengthens the flow of information, allows distant actors to network, increases media attention and national and international news coverage, creates public opportunities for organizing and taking action on issues or policies, and facilitates communication among supporters.

Online networks are considered to have a significant connection with policy learning. Methods such as Social Network Analysis (SNA) can present a system’s policy-learning architecture in real time by uncovering general patterns. Policy actors’ positions and network structures provide information about likelihoods of learning and policy change (Howlett et al., 2017, p. 246). Rethemeyer (2006, p. 259) argues that the Internet has transformed policy networks to allow internal and external actors to challenge those with structural power. Those with structural power then attempt to shape the Internet to maintain their positions. Examining two policy networks, they emphasize that the Internet has made networks more exclusionary as most interactions occur among electronic core members. Hence, the Internet reinforces existing patterns of authority. Considering the Indonesian biodiesel policy network, Howlett et al. (2017, p. 245) demonstrate that government officials play a significant role in policymaking and learning as the core of policymaking and knowledge transfer networks. Accordingly, those in positions of power can both increase and obstruct learning pathways for policy change. Therefore, while their practicality is debatable, online networks are an important tool for policy learning.

It is often uncertain whether online interactions lead to offline ones. Calls for physical protests can be made using social media platforms. However, individuals can also virtually participate in movements through social media, arguably reducing the cost of action. There are concerns about decreased levels of physical activism as some people only take action through digital channels; however, social media also encourages physical activism by providing information about physical protest actions (Nulman & Özkula, 2016, p. 12). In some cases, offline, and online actions may be unrelated because people behave differently in these spaces. However, as social media posts can mobilize others for offline protests, online, and offline activism are positively correlated (Greijdanus et al., 2020, p. 49). Digital activism can therefore encourage physical activism, including public participation in environmental protection initiatives and climate action across borders. Individuals can unite around the climate action cause through social media interactions and increase their leverage, thereby placing public pressure on policymakers. Indeed, YouTube was used during the COP15 protests (Askanius & Uldam, 2011), and the Fridays for Future movement utilized Twitter (Padilla-Castillo & Rodríguez-Hernández, 2023), and Instagram (Shim, 2024). Prominent climate actors like Greta Thunberg share their climate action journeys on social media (Krebs et al., 2024), and other individual climate actors have followed suit. These examples demonstrate how social media platforms and interactions can influence climate action across borders as a public sphere.

Crisis management requires close collaboration among stakeholders. The current climate crisis should also be addressed within this stakeholder framework. Di et al. (2023, p. 15) note that sharing strengths, resources, and best practices, as well technology transfer and joint research and development projects, can create synergies. Hence, those who caused the crisis must collaborate effectively with those affected by it. Arguably, the greatest responsibility for the climate crisis lies with the actors contributing to it. Their interactions constitute an important dimension of crisis management. Although some countries make larger contributions, the climate crisis is a global crisis requiring the consistent co-operation of global actors.

Methods

This study conducts a social media analysis of the climate and environment ministries of OECD countries, aiming to reveal the dimensions of stakeholder interactions to highlight the need for collective efforts to combat the climate crisis.

Population and Research Sample

While the climate and environment ministries of OECD countries form the research population, the sample consists of the climate and environment ministries of OECD founding members. This study considers the official climate and environment ministry Twitter accounts of Austria (Austrian Federal Ministry of Climate Action, Environment, Energy, Mobility, Innovation and Technology, n.d.), Belgium (Health Food Chain Safety and Environment, n.d.), Canada (Environment and Climate Change Canada, n.d.), Denmark (Ministry of Environment of Denmark, n.d.), France (Ministry of Ecological Transition and Territorial Cohesion, n.d.), Germany (Federal Ministry for the Environment Nature Conversation Nuclear Safety and Consumer Protection, n.d.), Greece (Ministry of Environment and Energy, n.d.), Iceland (Ministry of the Environment, Energy and Climate, n.d.), Italy (Ministry of Ecological Transition, n.d.), Ireland (Department of the Environment Climate and Communications, n.d.), Norway (Ministry of Climate and Environment, n.d.), Luxembourg (Ministry of the Environment Climate and Sustainable Development, n.d.), Portugal (Environment and Climate Action, n.d.), Spain (Ministry for the Ecological Transition and the Demographic Challenge, n.d.), Sweden (Ministry of the Environment, n.d.), Switzerland (Federal Department of the Environment, Transport, Energy and Communications, n.d.), the Netherlands (Ministry of Economic Affairs and Climate Policy, n.d.), the United Kingdom (Department for Business, Energy and Industrial Strategy, n.d.), the United States of America (United States Environmental Protection Agency, n.d.), and Türkiye (Ministry of Environment, Urbanization and Climate Change, n.d.).

Research Method

This study used Social Network Analysis (SNA), an application of graph theory that facilitates relationship pattern analysis among actors (Dado & Bodemer, 2017, p. 159) within social networks. SNA allows resource exchange to be examined through a relationship lens (Haythornthwaite, 1996, pp. 323, 326) between interacting social entities (Makagon et al., 2012, p. 153). This reveals the positions, characteristics, and forms of actors’ connections (Borgatti et al., 2009, p. 894). Graph theory is a branch of mathematics dealing with the arrangement of vertices and the edges between them (O’Regan, 2023). In other words, a graph G = (V, E) is a network based on a set of mathematically defined objects in the form of vertices and nodes (Chakraborty et al., 2018, p. 722). Examples include trees (connected graphs without cycles); k-regular graphs (each vertex has the same degree); and complete bipartite graphs, where vertices are divided into two sets with edges connecting every vertex in one set to all the vertices in the other (Ghanbari & Alikhani, 2023). Graph theory can be applied to social media networks; in particular, it can be applied to follower networks on Twitter (Chakraborty et al., 2018, p. 723). In this sense, graph theory can represent interactions between actors. Furthermore, graph theory has recently become more prominent and evolved (Chung, 2010, p. 726) for diverse use cases including social network modelling, social network analysis, and user analytics (Majeed & Rauf, 2020, p. 1).

Analysis

This research method was applied to official climate ministry Twitter account data collected on October 10, 2022. This study employed a qualitative approach to analyze the interactions of OECD countries in the digital environment. Qualitative research attempts to question, interpret and understand a problem in the natural environment in which it is located (Baltacı, 2019, p. 369). No keywords were used to identify the relevant countries’ accounts. Unofficial content from other accounts was not included in the study. Data were analyzed using the open-source computer programming R Studio (R Core Team, 2021). The “igraph” and “dplyr” packages were used for the network analysis (Dplyr R Package, 2025; Igraph R Package, 2022). The level of interaction among actors were represented both mathematically and visually across the network. See Appendix A for the code used in this study.

Results

Table 1 presents the Twitter account information for the climate and environment ministries of the fifteen OECD founding countries, including the account name, its creation date, follower count, following count, and its total number of tweets. Belgium, Denmark, Iceland, Sweden, and the Netherlands did not have official ministry Twitter accounts; hence, these countries were not included in the analysis.

Table 1.

Official Climate and Environment Ministry Twitter Accounts of OECD Countries.

Country	Account name	Creation date	Followers	Following	Total tweets
Austria (ERA-LEARN)	@ERALEARN_EU	Jun 2018	744	225	1,021
Canada (Environment Canada)	@environmentca	Jan 2010	430.5 K	463	22.7 K
France (Ministères Écologie Énergie Territoires)	@Ecologie_Gouv	Sep 2011	218 K	2,356	27.6 K
Germany (Bundesumweltministerium)	@BMUV	Jul 2010	180.1 K	269	17.7 K
Greece (Υπουργείο Περιβάλλοντος και Ενέργειας)	@YpenGr	Nov 2011	15.7 K	446	4,073
Ireland (Dept. Environment, Climate, and Communications)	@Dept_ECC	Feb 2013	16.7 K	655	6,780
Italy (Ministero della Transizione Ecologica)	@MiTE_IT	Feb 2012	120.3 K	267	7,180
Luxembourg (GouvernementLU)	@gouv_lu	Jul 2014	32 K	280	5,062
Norway (Klima- og miljødep)	@kldep	Apr 2009	16.9 K	1,377	4,297
Portugal (Ambiente e Ação Climática PT)	@ambiente_pt	Mar 2016	14.4 K	490	2,994
Spain (Transición Ecológica y Reto Demográfico)	@mitecogob	Jun 2018	68.6 K	380	19.5 K
Switzerland (DETEC—UVEK—DATEC)	@UVEK_DETEC	Apr 2009	2,736	212	2,673
United Kingdom (Dept for BEIS)	@beisgovuk	Jun 2009	199.1 K	2,620	27.8 K
United States of America (U.S. EPA)	@EPA	May 2008	628.6 K	689	18.4 K
Türkiye (T.C. Çevre, Şehircilik ve İklim Dğş. Bakanlığı)<	@csbgovtr	Dec 2012	436.8 K	75	7,805

Source. Authors.

Note. Accessed on October 06, 2022.

The United States of America (USA) was the first among these countries to create an account (2008), followed by Norway, Switzerland, and the United Kingdom (UK) in 2009. Austria and Spain were the last among these countries to create an account.

Figure 1 shows that the USA has the largest following (628.6 K followers) and Austria has the smallest (744 followers).

Figure 1.

Follower counts of Official Climate and Environment Ministry Twitter accounts of OECD countries.

Figure 2 shows that the UK follows the most accounts (2,620 accounts). Türkiye, on the other hand, follows the smallest number of accounts (75 accounts) among OECD countries.

Figure 2.

Following counts of Official Twitter Accounts of Climate and Environment Ministries of OECD countries.

Among the OECD countries, the USA has posted the most tweets (27.8 K tweets). Portugal ranks second to last with 2,994 tweets, and Austria ranks last with 1,021 tweets (Figure 3).

Figure 3.

Total Tweets of Official Twitter Accounts of Climate and Environment Ministries of OECD countries.

Table 2 shows the number of Twitter users in millions among OECD countries in 2023. populations of these OECD countries. With this in mind, the ministries’ accounts are followed by substantial numbers of users. Figure 4 presents a robustness analysis comparing climate and environment ministries’ follower counts to the total number of Twitter users in their respective countries.

Table 2.

National Twitter User Counts of OECD Countries (2023).

Country name	User count (millions)	Country name	User count (millions)	Country name	User count (millions)	Country name	User count (millions)
Austria	1.1	Germany	14.1	Luxembourg	.4308	Switzerland	1.4
Belgium	1.9	Greece	.9381	Norway	1.1	Netherlands	7.5
Canada	10.4	Iceland	.1234	Portugal	1.9	UK	23.1
Denmark	.9401	Ireland	1.7	Spain	10.8	USA	95.4
France	13.7	Italy	5.5	Sweden	2.1	Türkiye	18.6

Source. WPR (n.d.).

Figure 4.

Normalized Twitter follower ratios of Climate and Environment Ministries of OECD countries.

Luxembourg (7.43%) and Canada (4.14%) had the largest normalized follower ratios in 2023, falling into the “High Engagement” group effectively using social media for environmental communication. Türkiye (2.35%), Italy (2.19%), Greece (1.67%), France (1.59%), and Germany (1.28%) constitute the “Medium Interaction” group. Although they have achieved some access relative to their populations, their potential has not yet been fully utilized. Türkiye stands out for both its absolute and normalized follower counts (see Table 2). Ireland (.98%), the UK (.86%), Portugal (.76%), the USA (.66%), Spain (.64%), Norway (.51%), Switzerland (.20%), and Austria (.07%) constitute the “Low Engagement” category, as their ministry Twitter accounts reach only a small fraction of users in their countries. This suggests that digital public communication performance is below expectations, particularly in countries with large and developed social media infrastructures (e.g., the USA and UK).

Graph theory was then used to examine the relationships between the official Twitter accounts of climate and environment ministries in the fifteen OECD countries included in this study and the Twitter accounts of equivalent institutions in other OECD countries. Institutions are numbered in alphabetical order to ensure the network diagram is readable. Figure 5 shows the diagram numbers and corresponding OECD countries.

Figure 5.

OECD countries and their corresponding diagram label numbers.

Four key centrality metrics assess the relative significance or influence of a node within its overall network structure: betweenness, closeness, degree, and eigenvector analysis (Okamoto et al., 2008). Table 3 presents the network analysis results across these metrics.

Table 3.

Centrality Analysis of Official Climate and Environment Twitter Accounts of OECD Countries.

Node	Betweenness	Closeness	Degree	Eigenvector	Role
3	6.5′	1.0	6	.87405716	Leader
12	.0	NaN	1	.21735401	Marginal
16	.0	.60	3	.65134346	Other
10	3.0	.7142857	4	.74261153	Leader
5	.5	1.0	5	1.0	Leader
6	.0	1.0	4	.87667032	Other
18	.0	.5555556	1	.21735401	Marginal
13	.0	1.0	2	.12216840	Marginal
14	.0	NaN	1	.03037993	Marginal
19	.0	NaN	4	.46090232	Other
7	.0	1.0	1	.11461375	Marginal

Note. NaN = not a number.

Centrality refers to how centrally a node (i.e., an individual in social network analysis) is located within a network. Betweenness refers to a node’s role as a bridge along the shortest path between nodes. It is used to identify influential individuals or information carriers in social networks, strategic actors in policymaking, and communication bottlenecks in institutional structures. Betweenness centrality determines mediating roles in networks. Being located at a point where other paths intersect is important, reflecting a high betweenness centrality (Zhang & Luo, 2017, p. 301). That is, betweenness centrality indicates the extent to which a node is located between other paths. Betweenness centrality also acts as a control indicator, describing the extent to which an actor controls the flow of information between others (Newman, 2005, p. 40).

In the betweenness centrality analysis, Canada has the largest node (6.5). This indicates that it is a central player, the broker through which communication between many other nodes occurs. Most information passes through Canada’s node. France (.5) and Italy (3.0) are intermediaries in the network. These large nodes play critical roles in information transfer, but they are not as central as Canada. Greece, Norway, Portugal, Spain, and the UK are peripheral actors in the network with betweenness values of zero (.0). Often having only a single connection, these counties are inefficient in carrying information to the center of the network. Directional connections indicate that the USA receives connections from many country nodes, but this does not centralize them within the information network like Canada, which is more influential. Removing Canada from this network could seriously disrupt its structure because of the network’s high structural dependency on Canada. Countries with low betweenness values, such as Spain and the UK, are typically isolated or exclusively recipients within the network. Finally, Austria, Ireland, Luxembourg, and Türkiye are not included in the network, as they do not interact with peer organizations’ Twitter accounts. The node sizes according to betweenness centrality are shown in Figure 6.

Figure 6.

Betweenness centrality analysis of Official Climate and Environment Ministry Twitter accounts in OECD countries.

Closeness centrality analysis shows how centralized or accessible the nodes are in the network (Mayer et al., 2009, p. 609), measuring the total distance between a node and other nodes. A node has a high closeness centrality value if its shortest path to other nodes is small (Zhang & Luo, 2017, p. 301). Closeness centrality can also be defined as the value calculated by solving the shortest path problem (Okamoto et al., 2008, p. 187). Figure 7 represents the closeness centrality analysis, with each node scaled according to its closeness centrality. That is, the larger the node size, the shorter its average distance to reach other nodes.

Figure 7.

Closeness centrality analysis of Official Climate and Environment Ministry Twitter accounts in OECD countries.

Canada, France, Germany, Greece, Italy, and Portugal have high closeness centrality (all 1.0). These nodes can reach all other nodes in the network along relatively short paths. Information can spread more quickly through these nodes and their corresponding countries in the communication network. France and Germany’s node sizes and locations indicate that these countries have a high reach capacity. Small nodes represent low closeness centrality. In our study, the UK (.5555556) and Norway (NaN) have smaller nodes further from the center. This makes it more difficult to reach other countries. Spain (NaN) and the USA (NaN) have nodes that are either disconnected or have an undefined closeness value; in other words, they cannot reach all nodes in the network. Therefore, these nodes are barely visible in the graph.

If a node can reach many nodes but not vice versa, this unbalanced structure affects the closeness centrality value. High closeness is advantageous for accessing and disseminating information, that is, these nodes are ideal points for decision-support mechanisms. In our study, Canada is the most suitable hub for processes such as information dissemination, resource allocation, and strategic decision making. By contrast, nodes with low or NaN closeness values are marginal, disconnected, or weak-reach actors. Low-closeness nodes, such as those in Norway and Spain, are generally external to the network and play one-way or low-access roles.

Degree centrality relates to the total number of direct connections to other nodes in a network (Opsahl et al., 2010, p. 246; Zhang & Luo, 2017, p. 301). In the total degree centrality analysis shown in Figure 8, node size is determined by the total number of incoming and outgoing connections. Larger nodes are the most connected and active elements in the network. These nodes are influential and potential network hubs. Small nodes have few connections; that is, they are marginal and passive.

Figure 8.

Total degree centrality analysis of Official Climate and Environment Ministry Twitter accounts in OECD countries.

Canada (6) is the largest node. It predominantly has two-way relationships with other nodes. As it has many incoming and outgoing connections, it acts as a hub for both receiving and disseminating information. It is at the structural center of the network and has a high potential for interaction. France (5), Germany (4), Italy (4), the USA (4), and Switzerland (3) have medium-sized nodes. France and Germany may serve routing and relay functions, with France being connected to both Canada and Germany. While Italy and the USA are connected to Canada, they also provide connections to their own smaller networks such as Greece (1) and Portugal (2). Norway (1), the UK (1), Spain (1), and Greece (1) have small-sized nodes. In general, they are connected to a single node, mostly to the USA. Positioned at the edges of the network, they appear passive in terms of information transfer. These nodes can be considered as marginal actors within the network.

In summary, Canada is the information distribution center, the most important actor in terms of leadership or coordination, and the most effective site for initiating interventions. France and Germany served as intermediary nodes. Norway, Spain, and the UK access information later and are further away from central processes. Connecting Canada with marginal nodes such as Norway, Spain, and the UK can increase the network’s inclusiveness.

Eigenvector centrality analysis was used to measure countries’ levels of leadership or influence. This metric considers not only the number of connections a node has, but also the importance of those connections. Nodes connected to influential nodes are considered more influential than those that lack such connections. This is based on the logic that a node’s connections play a crucial role in shaping its centrality and influence (Borgatti et al., 2006). Therefore, eigenvector centrality also considers centrality measures (Maharani et al., 2014, p. 2).

France has the highest eigenvector centrality (1.0), with the largest node in Figure 9. This node is important because of its many connections and connections to influential nodes, such as Germany (.87667032) and Canada (.87405716). France is therefore at the heart of the network’s chain of influence, with Germany and Canada serving as transit points or routers for information. The Canada node is medium sized in the eigenvector centrality analysis, appearing slightly smaller in Figure 9 than in previous analyses (degree, closeness, and betweenness). This may indicate that the node lacks direct connections to highly influential actors, or that its connections are more diffusely distributed across the network. Small nodes in Norway (.21735401), the UK (.21735401), Spain (.03037993), and Greece (.11461375) have small eigenvector values; they are either connected to inactive nodes or are isolated. They are peripheral nodes that are inactive and weak in information flow.

Figure 9.

Eigenvector centrality analysis of Official Climate and Environment Ministry Twitter accounts in OECD countries.

Next, countries were classified based on their roles (i.e., “leader,”“bridge,” or “marginal”) to develop an understanding of the network’s structural characteristics and functional dynamics. This aspect of SNA considers the number of connections an actor has, as well as the connections they make, their direction, and the actor’s position within the network. This role-based approach allows for a theoretical analysis of social relations and facilitates practical policy and strategy development (Borgatti & Halgin, 2011, p. 1169; Wasserman & Faust, 1994, p. 5). “Leader” actors are generally characterized by high degree and eigenvector values (Freeman, 1978). They are at the center of a network with a high interaction capacity. “Bridge” actors have high betweenness centrality, facilitating information transfer between different sub-communities (Borgatti & Halgin, 2011). Conversely, peripheral actors with low degree and centrality values are classified as “marginal” (Hanneman & Riddle, 2005; Scott, 2017). These nodes have limited connections and low influence. The role-based classification criteria applied in this study are listed in Table 4.

Table 4.

Role-Based Classification Criteria.

Role	Condition	Literature
Leader	Degree ≥ 4; eigenvector ≥ .7	Wasserman and Faust (1994), Freeman (1978)
Bridge	Betweenness > 2 (greater more than the network average)	Borgatti and Halgin (2011), Freeman (1978)
Marginal	Degree ≤ 2, betweenness = 0, eigenvector < .25	Hanneman and Riddle (2005), Scott (2017)
Other	Nodes that do not meet any of the conditions listed above	Based on the above

Source. Authors.

Our study identifies three primary roles in the network structure: leader, marginal, and other. Canada, France, and Italy are leader nodes with high degree, closeness, and eigenvector values. They play significant roles in information dissemination within the network through direct connections with many nodes and associations with influential nodes. Norway, Portugal, Spain, the UK, and Greece are marginal actors with betweenness values of zero. They are typically connected to only one or two other nodes. Furthermore, some nodes such as Norway, Spain, and USA have undefined closeness values (NaN), indicating that they struggle to reach the wider network. Germany, Switzerland, and the USA are nodes with a medium level of connection and influence that cannot be categorized as either leader or marginal actors. Although these nodes do not play critical roles in information flow, they hold supportive positions that contribute to network integrity.

Community analysis was then performed to detect groups of countries with similar characteristics or strong interactions. A community can be conceptualized as a network of closely related entities. This structure is formed by a combination of individual entities that demonstrate reciprocity in their interactions with one another. Community members interact more frequently with each other than with those outside the group (Bedi & Sharma, 2016, p. 116). Uncovering structural information regarding the network’s internal structure or subgroups establishes an understanding of which nodes act together or exhibit common behaviors. Community analysis models how information spreads through the network, identifying which groups are isolated from each other and where their connection points are. By identifying groups that lack interaction, collaboration can be increased. Critical actors can also be identified. This study identifies countries acting as bridges between communities to highlight their strategic importance as actors within the network. Community detection algorithms, such as Louvain, Girvan-Newman, and Label Propagation, group nodes with similar connectivity (Çetin & Amrahov, 2022; Torkaman et al., 2023). These groups typically have high internal connectivity and low external connectivity, subdividing the network into more meaningful substructures. This study uses the Louvain community detection algorithm (Blondel et al., 2008). The community analysis results are shown in Table 5.

Table 5.

Community Analysis of Official Climate and Environment Ministry Twitter Accounts in OECD Countries.

Nodes	3, 12, 16, 10, 5, 6, 18, 13, 14, 19, 7
Groups	1, 1, 2, 2, 2, 2, 1, 3, 3, 3, 3
Modularity value (Q)	.3066667

The Louvain community detection algorithm divides the network of 11 nodes into 4 communities. The first community includes France, Germany, Italy, and Switzerland. These tightly connected countries have fewer connections than outside countries, forming a small group that is likely to focus on a common topic or share a role. The second group includes Greece, Portugal, Spain, and the USA. They exhibit dense connections, with the USA acting as a bridge to other communities. The third group includes Canada, Norway, and the UK. Canada is at the center of this community and is connected to other communities, as indicated by its high betweenness centrality value. The center group, which comprises Canada, Italy, and the USA, is connected to other communities. Although its boundaries are less clear, these bridge nodes facilitate information flow between communities. A community detection analysis graph is presented in Figure 10.

Figure 10.

Community detection graph of Official Climate and Environment Ministry Twitter accounts in OECD countries.

A modularity analysis was conducted to compare the network’s number of internal communities with those of a random network. Modularity is a numerical measure of a network’s community structure, indicating the extent to which a network is divided into communities. Q values greater than .3 to .4 generally indicate network modularity (Fortunato & Barthélemy, 2007, p. 37), and these values range from .3 to .7 (Sah et al., 2014, p. 4). Higher Q-values represent stronger community structure. Values less than .3 indicate weak or uncertain community structure; values between .3 and .4 are considered average; and values greater than .4 represent strong community structure. The modularity value of .3067 reported in Table 5 indicates that a moderate community structure is present in this four-category network, but that this structure is not very strong. Nodes have more connections to other nodes within their communities than would be expected in a random network; however, the structure is not strongly clustered. Hence, the community detection algorithm performed reasonably well in distinguishing whether the connections in the network were concentrated inside or outside.

Discussion

The climate crisis is one of the most pressing contemporary problems. For humanity’s survival, a livable climate with stable conditions is required. As climatic conditions are predicted to gradually worsen, collective efforts are important to encourage climate conscious perspectives. Most problems in this regard arise in the implementation of steps to combat the climate crisis. Furthermore, many actors interact with one another and contribute jointly to the climate crisis.

From this perspective, this study focuses on the interactions between actors with some extent of influence on the climate crisis. It examines the Twitter accounts of OECD countries, conducting an SNA using graph theory to uncover their interactions. According to the results, The USA was the first OECD country to create an official climate and environmental ministry Twitter account, whereas Austria and Spain lagged behind. The USA’s account had the most followers (628.6 K), the UK had the most following count (2.620). UK which also posted the largest total number of tweets (27.8 K). Austria ranked last among OECD countries, posting 1,021 tweets.

Germany is followed by Canada and Switzerland, while Germany only follows France. Norway is followed by Canada. Interestingly, Portugal follows Spain, whereas Spain does not follow Portugal. Switzerland is followed by Italy, Germany, and France. This suggests that neighboring countries are interconnected in terms of their interactions. In other words, while Switzerland does not follow Austria, Switzerland follows all of its neighboring countries. Geographic proximity significantly influences social networking (Breitmar et al., 2024, p. 1), and this result supports the idea that countries with similar climates tend to interact with each other, even if they are not neighbors sharing borders. This can also be linked to the concept of “cultural proximity,” which describes the tendency of people from similar countries or regions to interact with each other due to similarities in history, language, religion, race, etc. (Trepte, 2008, pp. 4–5). However, conversely, such differences can create “border effects,” establishing structural barriers that make it difficult to establish connections.

Canada and France have the highest levels of engagement. Although the USA is followed by four countries, it does not follow any country. Canada, on the other hand, follows four different countries, but only followed by Italy and UK. In this context, USA is the most followed country, whereas Canada has the highest number of followings. Austria, Ireland, Luxembourg, and Türkiye are not followed by other countries and also do not follow other countries themselves.

Betweenness centrality analysis identifies Canada as the largest node in the network. Hence, Canada has the highest intermediary value and communication between many nodes passes through it. France and Italy, on the other hand, play critical roles in information transfer as network intermediaries. Although their roles are not as central as that of Canada, they remain significant. To emphasize Canada’s importance in the network, its removal could significantly disrupt the network structure. Closeness centrality analysis identifies Canada, France, Germany, Greece, Italy, and Portugal as the largest nodes. They can access other nodes in the network via shorter routes, thereby allowing faster information dissemination. France and Germany stand out in terms of size and location as countries with high access capacities. In particular, Canada appears to be the most suitable hub for information dissemination, resource allocation, and strategic decision making.

Total degree centrality analysis also identifies Canada as the largest node. With predominantly two-way relationships with other network nodes, Canada has many incoming and outgoing connections. Therefore, it is a central actor in terms of both information reception and dissemination, with a high potential for interaction. This highlights Canada’s importance in terms of leadership and coordination. Eigenvector centrality analysis reveals that France has the highest value. France has many connections to influential nodes, acting as a transit point or router for information, like Germany and Canada. However, Canada, with its prominent betweenness, closeness, and total degree centrality, appears smaller in terms of eigenvector centrality. This can be explained by Canada’s lack of direct connections to influential actors; instead, its connections are dispersed throughout the network.

Taken together, these analyses demonstrate that Canada, with its very high interconnectivity, degree, and proximity, plays the role of a network gateway and hub. France, with the highest eigenvector, leads the chain of influence. Germany is a strong center of balance across all metrics and stands out as a router/bridge node. Norway, Spain, and Greece have low values on all metrics; therefore, they can be considered as marginal/peripheral nodes. Finally, community analysis divides the network into four parts. The first community consists of France, Germany, Italy, and Switzerland; the second consists of Greece, Portugal, Spain, and the USA; the third consisted of Canada, Norway, and the UK; and the fourth consists of Canada, Italy, and the USA. A modularity test revealed that these are moderate-strength communities.

Furthermore, the relevant ministries of some OECD countries do not have Twitter accounts: Belgium, Denmark, Iceland, Sweden, and the Netherlands. Although the Iceland ministry website links to a Twitter account, this link leads to the minister’s personal Twitter profile. Hence, the minister (Gudlaugur Thor Thordarson) and not the ministry itself interacted with other actors. Additionally, different countries give different names with different emphases to their ministries. Germany is the only country that emphasizes nuclear safety in its ministry name. Belgium highlights health and the food chain, France emphasizes transform regional harmony, and Ireland amplifies the communication dimension. Luxembourg underscores sustainability, whereas Switzerland’s naming encompasses transport, energy, and communication. While the Netherlands emphasizes economic relations, UK highlights business, energy, and industrial strategies, and Türkiye emphasizes urbanism.

The results of this study are similar to those in the literature, but certain differences also emerge. Limberg et al. (2024) found that establishing a dedicated climate ministry significantly reduces carbon emissions worldwide but that such ministries had no influence on implementing new climate policies. This shows that climate ministries strengthen climate action by increasing the effectiveness of existing policies rather than by introducing new ones. Ibrahim’s (2023) study of government institutions’ critical roles in combating climate change shows that these institutions could reduce health problems by implementing climate change policies, especially in areas such as air, water, and food security. Rahman (2021) evaluated the environmental sustainability performance of the Ministry of Environment, Forest, and Climate Change in Bangladesh, revealing its major problems and shortcomings in achieving its environmental sustainability goals. To overcome these difficulties, Rahman (2021) notes that institutional and legal reforms must be implemented, the capacity of the Ministry must be increased, and social support must be strengthened. Contrary to the approach taken by Limberg et al. (2024), the present study viewed climate ministries as a important actors for new policy implementation; In contrast to Ibrahim’s (2023) research, offers a general perspective instead of focusing only health issues; unlike Rahman (2021), which centers on a single ministry, this study focuses OECD countries climate ministries.

Biesbroek et al. (2010) focus on the national adaptation strategies (NAS) adopted by European Union (EU) member states in 2008. The role of NAS was found to vary across countries but constitutes a new measure of political commitment to adaptation at the national policy level. Although the strategies are similar in their topics, methods, and approaches, future policy implementations may encounter institutional challenges such as multilevel governance and policy integration. This study shows that EU member states have taken important steps towards adapting to climate change, but additional efforts are needed to further strengthen adaptation processes. Our study focuses on these interactions. Present study differs from Biesbroek et al. (2010) as focusing OECD countries climate ministries to deepen the addition efforts.

Aamodt (2018) analyzes the climate-policy roles of the Indian and Brazilian environmental ministries, arguing that environmental ministries should be important drivers of local climate policy. Historical views of environmental policy as a local or international problem, the ministry’s official role in international climate negotiations, and the development of institutional climate logic were found to affect environmental ministries’ climate policy objectives. This demonstrates the merits of common action-oriented approaches of organizations, such as the EU. Our study focuses on interactions among OECD countries as actors that are important in the climate problem phase rather than the interactions between specific actors as Aamodt (2018) did.

Di, Chen, Shi, Cai, and Zhang (2024), using ANNs and a System Generalized Method of Moments (SYS-GMM), found a win-win situation between digital empowerment and low-carbon initiatives. They present evidence that digital empowerment contributes to carbon emission reductions and that digital technology has a positive impact on promoting environmental sustainability. As in our study, Di, Chen, Shi, Cai, and Zhang (2024) in this context, they establish a link between digitalization and environmental, climate, and sustainability efforts, offering insights into how digital empowerment is central to sustainable development.

Considering this framework established by the existing literature, this study has three major findings. First, the climate crisis requires collective action, but the interaction among OECD countries is currently insufficient. Second, climate and environmental ministries can play central roles in collaboration and strategy. Third, OECD countries should be aware that they have a great responsibility to combat the climate crisis.

In summary, the level of interaction within the Twitter network was limited. As stated previously, collective effort is required to resolve the climate crisis. OECD countries have developed economies, creating the responsibility to collectively act against climate crises. These collective actions should stem from co-operation between their climate and environment ministries. In other words, countries should be aware of each other climate change efforts, and they should carry out these efforts together. Therefore, the level of interaction between countries should increase.

Institutional solidarity, mutual visibility, and digital coordination are necessary for solving environmental problems, particularly the climate crisis. Weak interactions may fail to establish collective voices against the climate crisis in the digital arena. Voluntary collaborations should be developed to combat climate crises, in the same way that the OECD countries establish economic co-operation. Node connections should be strengthened as a policy recommendation. Adding alternative routes, creating shortcuts to shorten path lengths, expanding the communication network, and finding and strengthening key connectors, are relevant suggestions; however, strengthening internal connections is crucial to strengthen the interactions among network users with similar interests. Therefore, policies should enhance existing ties by increasing the strength, frequency, and relevance of interactions. This may require new mechanisms to increase interactions among climate and environmental ministries in OECD countries. An OECD digital environmental diplomacy strategy could be developed to encourage interactions among environmental ministries on social media. Environmental ministries can consider joint campaigns with shared hashtags. Increasing visibility through coordinated collective communication efforts on social media is recommended, including on special occasions such as World Environment Day and World Days to Combat Climate Change. Good practices can be shared among actors through explicit tagging (mentioning). Social media posts should not be limited to information sharing but should facilitate cross-platform discussions within a multifaceted digital dialogue environment through hashtags and other approaches. By increasing interactions, ministries can be informed of each other’s actions, engage in collective efforts, and establish common ground for action. Increased interactions among even a few OECD countries can exert peer pressure on other countries to follow suit. Furthermore, passive actors who feel powerless to combat climate change can gain the courage and motivation to participate in climate action. Similarly, when digital interaction increases public visibility, perceptions of inter-institutional solidarity can also increase.

This research focuses on OECD countries as the largest contributors to the climate crisis; however, this also has limitations. Other countries contribute to the climate crisis, although not to the same extent as OECD countries. Therefore, future research should be conducted with a broader sample encompassing a larger number of countries. It will then be important to assess the role and extent to which all actors contribute to the crisis. A country making a smaller contribution to the climate crisis may be more likely to participate in a solution. This example demonstrates the importance of considering other countries in these processes. Future research should also consider the quality of interactions as well as their level. A high level of interaction does not always correspond with collective action. Therefore, future studies should focus on methods that create collective action. Different methods such as sentiment analysis may have the potential to contribute to the process. Focusing on meaning, alongside positivist methods, may illuminate situations, meanings, and thoughts during emotionally charged crises, facilitating an examination of the line between action and inaction.

Footnotes

Appendix

Appendix A

The Following List Includes the Codes Used in R in the Study.

Library used	Purpose	Example code
igraph	Create a graph from edge list	B5 <- graph(c("3","12","16",…))
igraph	Plot a graph with centrality values	plot(B5, vertex.size=cls*20, …)
igraph	Calculate betweenness centrality	btw <- betweenness(B5) btw <- betweenness(B5, directed = TRUE) btw btw_df <- data.frame( Node = names(btw), Betweenness = round(btw, 2) ) print(btw_df) node_size <- scale(btw, center = FALSE)*10 + 10 node_color <- "lightblue" plot(B5, vertex.label = V(B5)$name, vertex.label.cex = 1.5, vertex.label.color = "black", vertex.size = node_size, vertex.color = node_color, edge.arrow.size = 0.5, main = "Betweenness Centrality Graph")
igraph	Calculate closeness centrality	cls <- closeness(B5, mode = "out", normalized = TRUE) closeness_df <- data.frame( Node = names(cls), Closeness = round(cls, 4) ) print(closeness_df) node_size_cls <- scale(cls, center = FALSE) * 15 + 5 plot(B5, vertex.label = V(B5)$name, vertex.label.cex = 1.5, vertex.label.color = "black", vertex.size = node_size_cls, vertex.color = "lightgreen", edge.arrow.size = 0.5, main = "Closeness Centrality Graph")
igraph	Calculate degree centrality	deg <- degree(B5, mode = "all") node_size <- scale(deg, center = FALSE) * 8 + 10 layout <- layout_with_fr(B5) plot(B5, layout = layout, vertex.label = V(B5)$name, vertex.label.cex = 1.5, vertex.label.color = "black", vertex.size = node_size, vertex.color = "lightblue", edge.arrow.size = 0.4, edge.color = "gray40", main = "Total Degree Centrality Graph")
igraph	Calculate eigenvector centrality	eig <- eigen_centrality(B5, directed = TRUE)$vector eig_df <- data.frame( Node = names(eig), Eigenvector = round(eig, 4) ) print(eig_df) node_size_eig <- scale(eig, center = FALSE) * 20 + 5 layout <- layout_with_fr(B5) plot(B5, layout = layout, vertex.label = V(B5)$name, vertex.label.cex = 1.5, vertex.label.color = "black", vertex.size = node_size_eig, vertex.color = "plum", edge.arrow.size = 0.4, edge.color = "gray50", main = "Eigenvector Centrality Graph")
dplyr	Role-based analysis	btw <- betweenness(B5) cls <- closeness(B5, normalized = TRUE) deg <- degree(B5) eig <- eigen_centrality(B5)$vector df <- data.frame( Node = names(btw), Betweenness = btw, Closeness = cls, Degree = deg, Eigenvector = eig ) med_btw <- median(df$Betweenness) med_eig <- median(df$Eigenvector) med_deg <- median(df$Degree) df <- df %>% mutate(Role = case_when( Betweenness > med_btw & Eigenvector > med_eig ~ "Leader", Betweenness > med_btw & Eigenvector <= med_eig ~ "Bridge", Degree <= med_deg & Eigenvector <= med_eig ~ "Marginal", TRUE ~ "Other" ))
igraph	Modularity calculation	B5_undir <- as.undirected(B5) comm <- cluster_louvain(B5_undir) modularity(comm)
igraph	Community detection	plot(comm, B5_undirected, vertex.label = V(B5)$name, vertex.size = 25, main = "Community Detection of the Graph") V(B5_undirected)$color <- membership(comm) plot(B5_undirected, vertex.label = V(B5)$name, vertex.color = rainbow(3)[membership(comm)], vertex.size = 30, main = "Network Graph Colored by Communities")

Acknowledgements

We would like to thank our family, colleagues, and institutions for their support during the research and writing of the article.

ORCID iDs

Esra Tuncay

Sezgin Savaş

Ethical Considerations

We wish to confirm that ethical approval was not required for the present study because the study did not involve humans, and all research was conducted to the highest possible ethical standards.

Consent to Participate

Written informed consent for publication of their details was not obtained in this study as the study did not involve living human participants.

Author Contributions

Esra Tuncay: Methodology (lead); software (lead); formal analysis (equal); review and editing (equal). Sezgin Savaş: Conceptualization (lead); formal analysis (equal); writing—original draft (lead); writing—review and editing (equal).

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 datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

References

Aamodt

(2018). Environmental ministries as climate policy drivers: Comparing Brazil and India. Journal of Environment & Development, 27(4), 355–381. https://doi.org/10.1177/1070496518791221

Askanius

Uldam

(2011). Online social media for radical politics: Climate change activism on YouTube. International Journal of Electronic Governance, 4(1/2), 69–84. https://doi.org/10.1504/IJEG.2011.041708

Austrian Federal Ministry of Climate Action, Environment, Energy, Mobility, Innovation and Technology (BMK). (n.d.). Main page. https://www.era-learn.eu/network-information/organisations/austrian-federal-ministry-of-transport-innovation

Balta Peltekoğlu

. (2004). Halkla ilişkiler nedir?BETA Yayınları.

Baltacı

(2019). Nitel araştırma süreci: Nitel bir araştırma nasıl yapılır? Ahi Evran Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, 5(2), 368–388. https://doi.org/10.31592/aeusbed.598299

Banister

(2019). The climate crisis and transport. Transport Reviews, 39(5), 565–568. https://doi.org/10.1080/01441647.2019.1637113

Becken

Friedl

Stantic

Connolly

R. M.

Chen

(2021). Climate crisis and flying: Social media analysis traces the rise of “flightshame.” Journal of Sustainable Tourism, 29(9), 1450–1469. https://doi.org/10.1080/09669582.2020.1851699

Bedi

Sharma

(2016). Community detection in social networks. WIREs Data Mining and Knowledge Discovery, 6(3), 115–135. https://doi.org/10.1002/widm.1178

Biesbroek

G. R.

Swart

R. J.

Carter

T. R.

Cowan

Henrichs

Mela

Morecroft

M. D.

Rey

(2010). Europe adapts to climate change: Comparing national adaptation strategies. Global Environmental Change, 20(3), 440–450. https://doi.org/10.1016/j.gloenvcha.2010.03.005

10.

Blondel

V. D.

Guillaume

J. L.

Lambiotte

Lefebvre

(2008). Fast unfolding of communities in large networks. Journal of Statistical Mechanics: Theory and Experiment, 2008(10), Article 10008. https://doi.org/10.1088/1742-5468/2008/10/P10008

11.

Borgatti

S. P.

Carley

K. M.

Krackhardt

(2006). On the robustness of centrality measures under conditions of imperfect data. Social Networks, 28(2), 124–136. https://doi.org/10.1016/j.socnet.2005.05.001

12.

Borgatti

S. P.

Halgin

D. S.

(2011). On network theory. Organization Science, 22(5), 1168–1181. https://doi.org/10.1287/orsc.1100.0641

13.

Borgatti

S. P.

Mehra

Brass

D. J.

Labianca

(2009). Network analysis in the social sciences. Science, 323(5916), 892–895. https://doi.org/10.1126/science.1165821

14.

Breitmar

Harding

M. C.

Zhang

(2024). Analysis of proximity informed user behavior in a global online social network. arXiv preprint arXiv:2404.18979. https://doi.org/10.48550/arXiv.2404.18979

15.

Bryant

Dabhi

Böhm

(2015). ‘Fixing’ the climate crisis: Capital, states, and carbon offsetting in India. Environment and Planning A, 47(10), 2047–2063. https://doi.org/10.1068/a130213p

16.

Bülbül

(2004). Halkla ilişkiler. Nobel Yayınları.

17.

Butler

L. J.

Scammell

M. K.

Benson

E. B.

(2016). The Flint, Michigan, water crisis: A case study in regulatory failure and environmental injustice. Environmental Justice, 9(4), 93–97. https://doi.org/10.1089/env.2016.0014

18.

Campiranon

Arcodia

(2008). Market segmentation in time of crisis: A case study of the MICE sector in Thailand. Journal of Travel & Tourism Marketing, 23(2–4), 151–161. https://doi.org/10.1300/J073v23n02_12

19.

Çetin

Amrahov

Ş. E.

(2022). A new network-based community detection algorithm for disjoint communities. Turkish Journal of Electrical Engineering & Computer Sciences, 30(6), 2190–2205. https://doi.org/10.55730/1300-0632.3933

20.

Chakraborty

Dutta

Mondal

Nath

(2018). Application of graph theory in social media. International Journal of Computer Sciences and Engineering, 6(10), 722–729. https://doi.org/10.26438/ijcse/v6i10.722729

21.

Chang

Liu

(2024). The role of government social media in enhancing environmental governance. China Economic Journal, 17(1), 40–55. https://doi.org/10.1080/17538963.2023.2300865

22.

Chung

(2010). Graph theory in the information age. Notices of the AMS, 57(6), 726–732.

23.

Coombs

W. T.

(2018). Crisis communication: The best evidence from research. In Miller

C. C.

Gephart

Helgesson

K. S.

Gephart

R. P.

(Eds.), The Routledge companion to risk, crisis and emergency management (pp. 51–66). Routledge.

24.

Crist

(2007). Beyond the climate crisis: A critique of climate change discourse. Telos, 141, 29–55.

25.

Dado

Bodemer

(2017). A review of methodological applications of social network analysis in computer-supported collaborative learning. Educational Research Review, 22, 159–180. https://doi.org/10.1016/j.edurev.2017.08.005

26.

Department of the Environment Climate and Communications. (n.d.). Main page. https://www.gov.ie/en/organisation/department-of-the-environment-climate-and-communications/

27.

Department for Business, Energy and Industrial Strategy. (n.d.). Main page. https://www.government.nl/ministries/ministry-of-economic-affairs-and-climate-policy

28.

Chen

Shi

Cai

Liu

(2024). Analysing the impact of coupled domestic demand dynamics of green and low-carbon consumption in the market based on SEM-ANN. Journal of Retailing and Consumer Services, 79, Article 103856. https://doi.org/10.1016/j.jretconser.2024.103856

29.

Chen

Shi

Cai

Zhang

(2024). Digital empowerment and win-win co-operation for green and low-carbon industrial development: Analysis of regional differences based on GMM-ANN intelligence models. Journal of Cleaner Production, 445, Article 141332. https://doi.org/10.1016/j.jclepro.2024.141332

30.

Chen

Zhang

Shi

Cai

Liu

(2023). Regional unevenness and synergy of carbon emission reduction in China’s green low-carbon circular economy. Journal of Cleaner Production, 420, Article 138436. https://doi.org/10.1016/j.jclepro.2023.138436

31.

Díaz-Díaz

Pérez-González

(2016). Implementation of social media concepts for e-government: Case study of a social media tool for value co-creation and citizen participation. Journal of Organizational and End User Computing, 28(3), 104–121. https://doi.org/10.4018/JOEUC.2016070107

32.

Dplyr R Package. (2025). Igraph. https://igraph.org/r/

33.

Engström

Gars

Jaakkola

(2020). What policies address both the Coronavirus crisis and the climate crisis?. Environ Resource Economics, 76, 789–810. https://doi.org/10.1007/s10640-020-00451-y

34.

Environment and Climate Action. (n.d.). Main page. https://www.portugal.gov.pt/pt/gc23/area-de-governo/ambiente-e-acao-climatica

35.

Environment and Climate Change Canada. (n.d.). Main page. https://www.canada.ca/en/environment-climate-change.html

36.

Federal Department of the Environment, Transport, Energy and Communications. (n.d.). Main page. https://www.uvek.admin.ch/uvek/en/home.html

37.

Federal Ministry for the Environment Nature Conversation Nuclear Safety and Consumer Protection. (n.d.). Main page. https://www.bmuv.de/en/

38.

Fortunato

Barthélemy

(2007). Resolution limit in community detection. Proceedings of the National Academy of Sciences of the United States of America, 104(1), 36–41. https://doi.org/10.1073/pnas.0605965104

39.

Fraser

Cunningham

Bancroft

Hunt

Lee

Nasongo

(2020). Climate crisis at city hall: How Japanese communities mobilize to eliminate emissions. Environmental Innovation and Societal Transitions, 37, 361–380. https://doi.org/10.1016/j.eist.2020.09.006

40.

Freeman

L. C.

(1978). Centrality in social networks conceptual clarification. Social Networks, 1(3), 215–239. https://doi.org/10.1016/0378-8733(78)90021-7

41.

Ghanbari

Alikhani

(2023). A graph related to the Euler ø function. Mathematical Gazette, 107(569), 263–272. https://doi.org/10.1017/mag.2023.57

42.

Greijdanus

de Matos Fernandes

C. A.

Turner-Zwinkels

Honari

Roos

C. A.

Rosenbusch

Postmes

(2020). The psychology of online activism and social movements: Relations between online and offline collective action. Current Opinion in Psychology, 35, 49–54. https://doi.org/10.1016/j.copsyc.2020.03.003

43.

Gundel

(2005). Towards a new typology of crises. Journal of Contingencies & Crisis Management, 13(3), 106–115. https://doi.org/10.1111/j.1468-5973.2005.00465.x

44.

Hanneman

R. A.

Riddle

(2005). Introduction to social network methods. University of California, Riverside.

45.

Haythornthwaite

(1996). Social network analysis: An approach and technique for the study of information exchange. Library & Information Science Research, 18(4), 323–342. https://doi.org/10.1016/S0740-8188(96)90003-1

46.

Health Food Chain Safety and Environment. (n.d.). Main page. https://www.belgium.be/en/economy/sustainable_development/sustainable_development

47.

Hopke

J. E.

Hestres

L. E.

(2018). Visualizing the Paris climate talks on Twitter: Media and climate stakeholder visual social media during COP21. Social Media + Society, 4(3), 1–15. https://doi.org/10.1177/2056305118782687

48.

Howlett

Mukherjee

Koppenjan

(2017). Policy learning and policy networks in theory and practice: The role of policy brokers in the Indonesian biodiesel policy network. Policy and Society, 36(2), 233–250. https://doi.org/10.1080/14494035.2017.1321230

49.

Ibrahim

(2023). Climate change and health hazards: Mitigation roles of public sectors (ministry, department and agencies). In Leal Filho

Vidal

D. G.

Dinis

M. A. P.

(Eds.), Climate change and health hazards: Addressing hazards to human and environmental health from a changing climate (pp. 363–379). Springer Nature.

50.

Igraph R Package. (2021). Igraph. https://igraph.org/r/

51.

Kaewkitipong

Chen

Ractham

(2012). Lessons learned from the use of social media in combating a crisis: A case study of 2011 Thailand flooding disaster [Conference session]. Thirty Third International Conference on Information Systems, Orlando, (pp. 1–17).

52.

Khodarahmi

(2009). Crisis management. Disaster Prevention and Management, 18(5), 523–528. https://doi.org/10.1108/09653560911003714

53.

Kim

S. S.

Chun

Lee

(2005). The effects of SARS on the Korean hotel industry and measures to overcome the crisis: A case study of six Korean five-star hotels. Asia Pacific Journal of Tourism Research, 10(4), 369–377. https://doi.org/10.1080/10941660500363694

54.

Koselleck

Richter

M. W.

(2006). Crisis. Journal of the History of Ideas, 67(2), 357–400. https://doi.org/10.1353/jhi.2006.0013

55.

Krebs

C. L.

Loizzo

Barnett

C. P.

Baker

(2024). Climate change cyber activism: A visual communication content analysis of youth activist Greta Thunberg’s Instagram. International Journal of Arts, Humanities and Social Sciences, 5(1), 1–10. https://doi.org/10.56734/ijahss.v5n1a1

56.

León

Negredo

Erviti

M. C.

(2022). Social Engagement with climate change: Principles for effective visual representation on social media. Climate Policy, 22(8), 976–992. https://doi.org/10.1080/14693062.2022.2077292

57.

Limberg

Steinebach

Nyrup

(2024). Dedicated climate ministries help to reduce carbon emissions. npj Climate Action, 3(1), 1–8. https://doi.org/10.1038/s44168-024-00147-9

58.

Maharani

Adiwijaya

A. A.

Gozali

A. A.

(2014, October 23–24). Degree centrality and eigenvector centrality in twitter [Conference session]. 8th International conference on telecommunication systems services and applications, (TSSA), Bali, Indonesia, pp. 1–5. IEEE.

59.

Majeed

Rauf

(2020). Graph theory: A comprehensive survey about graph theory applications in computer science and social networks. Inventions, 5(1), Article 10. https://doi.org/10.3390/inventions5010010

60.

Makagon

M. M.

McCowan

Mench

J. A.

(2012). How can social network analysis contribute to social behavior research in applied ethology? Applied Animal Behaviour Science, 138(3–4), 152–161. https://doi.org/10.1016/j.applanim.2012.02.003

61.

Masco

(2017). The crisis in crisis. Current Anthropology, 58(Suppl. 15), S65–S76. https://doi.org/10.1086/688695

62.

Mayer

F. S.

Frantz

C. M.

Bruehlman-Senecal

Dolliver

(2009). Why is nature beneficial? The role of connectedness to nature. Environment and Behavior, 41(5), 607–643. https://doi.org/10.1177/0013916508319745

63.

Ministry of Economic Affairs and Climate Policy. (n.d.). Main page. https://www.government.nl/ministries/ministry-of-economic-affairs-and-climate-policy

64.

Ministry of Environment Climate and Sustainable Development. (n.d.). Main page. https://mecdd.gouvernement.lu/en.html

65.

Ministry for the Ecological Transition and the Demographic Challenge. (n.d.). Main page. https://energia.gob.es/en-US/Paginas/index.aspx

66.

Ministry of Climate and Environment. (n.d.). Main page. https://www.regjeringen.no/en/dep/kld/id668/

67.

Ministry of Ecological Transition. (n.d.). Main page. https://www.epa.gov/

68.

Ministry of Ecological Transition and Territorial Cohesion. (n.d.). Main page. https://www.ecologie.gouv.fr/

69.

Ministry of Environment Urbanization and Climate Change. (n.d.). Main page. https://www.csb.gov.tr/

70.

Ministry of Environment and Energy. (n.d.). Main page. https://ypen.gov.gr/

71.

Ministry of Environment of Denmark. (n.d.). Main page. https://en.mim.dk/

72.

Ministry of the Environment. (n.d.). Main page. https://www.government.se/government-of-sweden/ministry-of-the-environment/

73.

Ministry of the Environment, Energy, and Climate. (n.d.). Main page. https://www.government.is/ministries/ministry-of-the-environment-energy-and-climate/

74.

Newman

M. E. J.

(2005). A measure of betweenness centrality based on random walks. Social Networks, 27(1), 39–54. https://doi.org/10.1016/j.socnet.2004.11.009

75.

Nulman

Özkula

S. M.

(2016). Environmental nongovernmental organizations’ digital media practices toward environmental sustainability and implications for informational governance. Current Opinion in Environmental Sustainability, 18, 10–16. https://doi.org/10.1016/j.cosust.2015.04.004

76.

Okamoto

Chen

X. Y.

(2008). Ranking of closeness centrality for large-scale social networks. In International workshop on frontiers in algorithmics (pp. 186–195). Springer.

77.

Opsahl

Agneessens

Skvoretz

(2010). Node centrality in weighted networks: Generalizing degree and shortest paths. Social Networks, 32(3), 245–251. https://doi.org/10.1016/j.socnet.2010.03.006

78.

O’Regan

(2023). Graph theory. In O'Regan

(Ed.), Matematical foundations of software engineering (pp. 119–130). Springer.

79.

Padilla-Castillo

Rodríguez-Hernández

(2023). International youth movements for climate change: The #FridaysForFuture case on Twitter. Sustainability, 15(1), Article 268. https://doi.org/10.3390/su15010268

80.

Parsons

(1996). Crisis management. Career Development International, 1(5), 26–28. https://doi.org/10.1108/13620439610130614

81.

R Core Team. (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing.

82.

Rahman

M. M.

(2021). Assessing the progress and pitfalls of the Ministry of Environment, Forest, and Climate Change in achieving SDGs in Bangladesh. Bangladesh Journal of Public Administration, 29(2), 140–158. https://doi.org/10.36609/bjpa.v29i2.228

83.

Rethemeyer

R. K.

(2006). Policymaking in the age of internet: Is the internet tending to make policy networks more or less inclusive? Journal of Public Administration Research and Theory, 17(2), 259–284. https://doi.org/10.1093/jopart/mul001

84.

Sah

Singh

L. O.

Clauset

Bansal

(2014). Exploring community structure in biological networks with random graphs. BMC Bioinformatics, 15(1), Article 220. https://doi.org/10.1186/1471-2105-15-220

85.

Scott

(2017). Social network analysis (4th ed.). Sage Publications.

86.

Shim

(2024). Personalising climate change—How activists from Fridays for future visualise climate action on Instagram. Humanities and Social Sciences Communications, 11(1), 1–9. https://doi.org/10.1057/s41599-024-03556-7

87.

Sinclair

A. J.

Peirson-Smith

T. J.

Boerchers

(2017). Environmental assessments in the Internet age: The role of e-governance and social media in creating platforms for meaningful participation. Impact Assessment and Project Appraisal, 35(2), 148–157. https://doi.org/10.1080/14615517.2016.1251697

88.

Stevens

T. M.

Aarts

Termeer

C. J. A. M.

Dewulf

(2016). Social media as a new playing field for the governance of agro-food sustainability. Current Opinion in Environmental Sustainability, 18, 99–106. https://doi.org/10.1016/j.cosust.2015.11.010

89.

Sun

Shi

Irfan

(2025). Digitally empowered green public services in environmentally vulnerable areas: Insights from SEM-ANN analysis. Journal of Retailing and Consumer Services, 84, Article 104216. https://doi.org/10.1016/j.jretconser.2024.104216

90.

Sun

Shi

Zhang

(2024). Study on the development path of low-carbon retail clusters empowered by digital empowerment. Journal of Retailing and Consumer Services, 81, Article 104006. https://doi.org/10.1016/j.jretconser.2024.104006

91.

Torkaman

Badie

Salajegheh

Bokaei

M. H.

Ardestani

S. F. F.

(2023). A four-stage algorithm for community detection based on Label propagation and game theory in social networks. AI, 4(1), 255–269. https://doi.org/10.3390/ai4010011

92.

Trepte

(2008). Cultural proximity in TV entertainment: An eight-country study on the relationship of nationality and the evaluation of U.S. prime-time fiction. Communications, 33(1), 1–25. https://doi.org/10.1515/COMMUN.2008.001

93.

United States Environmental Protection Agency. (n.d.). Main page. https://www.epa.gov/

94.

Wasserman

Faust

(1994). Social network analysis: Methods and applications. Cambridge University Press.

95.

WPR. (n.d.). Twitter/X users by country. https://worldpopulationreview.com/country-rankings/twitter-users-by-country?

96.

Zhang

Luo

(2017). Degree centrality, betweenness centrality, and closeness centrality in social network. In 2nd International conference on modelling, simulation and applied mathematics (MSAM2017) (pp. 300–303). Atlantis Press.

Twitter Analysis of Collective Action of OECD Countries Against Climate Crises

Abstract

Plain Language Summary

Keywords

Introduction

Literature Review

Methods

Population and Research Sample

Research Method

Analysis

Results

Discussion

Footnotes

Appendix

Acknowledgements

ORCID iDs

Ethical Considerations

Consent to Participate

Author Contributions

Funding

Declaration of Conflicting Interests

Data Availability Statement

References