One Earth, One Humanity Versus the Virus: Examining the Global COVID-19 Social Partnership Communication Networks on Social Media

Abstract

The COVID-19 pandemic has posed severe challenges that require collaborative efforts from multi-sector organizations. Guided by an institutional theory framework that considers how both organizational fields and national level contexts affect organizations’ social partnership communication, the current study examines the COVID-19-related social partnership communication network on social media. The cross-national study using semantic network analysis and exponential random graph models (ERGMs) first maps the meaning of COVID-19 social partnership network, and then investigates the role of organizational fields and a country’s political system, economic system, educational system, and cultural system on the formation of interorganizational communication ties surrounding the relief efforts of COVID-19. Results reveal the importance of the political system—such as the presence of populist government, economic disparity, and uncertainty avoidance cultural orientation in shaping the social media-based social partnership communication network. In addition, NGOs from multiple issue areas are actively engaged in the network, whereas corporations from manufacturing and financial industries are active players.

The COVID-19 pandemic is one of the most serious global health threats in recent history. As of February 2022, this ongoing pandemic has infected more than 424 million people in over 188 countries and territories, resulting in more than 5 million deaths and triggered the worst global economic recession since the Great Depression (New York Times, 2022). Governments around the globe have scrambled to cope with this severe challenge, many woefully underprepared and deprived of the necessary resources to manage the crisis. An important way to address emerging crises like this is through social partnerships (Behrens & Helfen, 2016), which refer to the forms of “collaboration in which organizations from multiple sectors interact to achieve common ends” (Waddock, 1991, p. 481). In times of a pandemic, social partnerships can often deliver efficient relief efforts through combining corporate resources and NGOs’ issue expertise (Rim et al., 2016).

In this study, we examine how organizations from the private and voluntary sectors communicate social partnerships in response to the pandemic on social media. We argue that organizations do not communicate interorganizational relationships capriciously on publicly visible platforms, but instead strategically highlight the type of partnership that may meet stakeholder expectations or boost their reputations (Saxton et al., 2019; Shumate & O’Connor, 2010). Relationships communicated on social media allow actors to “enter the public dialogue, offer legitimacy claims, and create positive relationships with publics that can influence their institutional operating environments” (Shumate & O’Connor, 2010, p. 584). For corporations, such relationships primarily serve the goal of CSR communication and virtual signaling (Moss et al., 2015; Yang & Liu, 2020). For NGOs, such communication can raise their profile and bring public attention to the advocated issues (Rim et al., 2016).

To understand the content and structure of the COVID-19 social partnership network, we take an institutional theory framework that focuses on the roles of organizational fields and national-level contexts (Matten & Moon 2008; O’Connor & Gronewold, 2013; Yang & Liu, 2018). Specifically, we first map the meaning of this social partnership network through a semantic network analysis of relevant tweets, and then test how factors such as different countries’ political, economic, cultural, and educational contexts may influence the structure of the social partnerships network that emerges quickly in response to a global crisis. (O’Connor et al., 2017; O’Connor & Gronewold, 2013). Lessons learned from this study have implications for improving the current COVID-19 relief communication and resource mobilization, and may be applied to managing future challenges such as combating COVID-19 vaccine hesitancy and climate change.

To test our research questions and hypotheses, we collected about two million tweets from over 2,000 organizations from over 100 countries. We further extracted archive data from the World Bank and other sources and applied exponential random graph models (ERGMs) to examine the structural features and tie formation mechanisms in this global COVID-19 social partnerships communication network. Our findings identify key themes surrounding the social partnerships, illustrating the powerful influence from countries’ uncertainty avoidance culture, populist leadership in countries’ political system, unemployment, and educational level as well as industry-level factors. Theoretical and practical implications are discussed.

Communicating Social Partnerships on Social Media

Social partnerships (also known as social action systems or problem-solving networks) refer to voluntary collaborative efforts of organizations in two or more economic sectors to collaboratively solve issues or problems of mutual concerns (Behrens & Helfen, 2016). Social partnership is prevalent in the COVID-19 pandemic social responses. For example, when the COVID-19 pandemic ravaged through the US, many companies purchased basic medical supplies and personal protection gear and shipped them to hospitals and epicenters of the outbreak. In addition, many corporations worked with their NGO partners (e.g., food banks) for their issue expertise, community accesses, and facilities to deliver relief efforts (Li et al., 2021).

As corporations and NGOs engage in social partnership activities, they often utilize a variety of communication channels such as social media to share updates and inform stakeholders about their positive actions (Ihm, 2019; Saxton et al., 2019). Communication about social partnership creates a dialogue between organizations and stakeholders about what behaviors are considered socially responsible (Rim et al., 2016). Communication also allows corporations to inform stakeholders about their social performance, meet stakeholder expectations for corporate actions, and preserve competitive advantages (O’Connor & Gronewold, 2013).

Meanwhile, the social partnership network surrounding the COVID-19 pandemic also presents itself as a meaning network signaling how partner organizations make sense of the pandemic and discursively engage in the problem solution process. Such a meaning network could raise public attention to certain social issues, affect public opinions, and influence industries or societal standards for acceptable corporate behaviors (Mahoney et al., 2013). As Saxton et al. (2019) argue, the communication efforts “are designed not solely to convey performance but also to signal a firm’s CSR-related ideas, intentions, values, and goals and are sent with the aim of enhancing a firm’s reputation” (p. 359). Through communicating social partnerships, organizations can “position themselves within certain issue spaces and manage these issues” (p. 1089). Moreover, organizations can disseminate their messages through their partners’ networks and reach broader audiences.

We therefore seek to map the meaning of the social partnership network by identifying major themes that emerge from COVID-19-related tweets from the corporate and nonprofit sectors. Twitter serves as a favorable environment for organizations to communicate social partnership efforts due to its large user base, capacity to engage stakeholders, and speed for information diffusion (Tao & Wilson, 2015). A semantic network approach enables the identification of salient concepts, their associative patterns, and the interpretative context surrounding them (Yang & Veil, 2017). We propose our first research question below:

• RQ1: What are salient themes and their associative patterns emerging from the COVID-19-related social partnership network on Twitter?

Understanding Institutional Influences on Social Partnership Communication

Other than understanding the content of COVID-19 social partnership network, the current study further explores the mechanisms that likely drive the formation of partnership ties. Previous research has suggested a divergence in tie formation patterns across organizational fields and countries (Shumate & O’Connor, 2010; Yang & Liu, 2018). In order to explain the observed difference, we draw from institutional theories as further discussed below.

Organizational Field Level Institutions and Social Partnership Communication

Institutional theory posits that organizational actions are products of institutional arrangements and social processes (Jamali & Neville, 2011). Organizations’ social partnership communication is usually shaped by complex, competing, and multi-level institutional forces (Campbell, 2007; Jamali & Karam, 2018). This study proposes that organizations’ social partnership communication related to COVID-19 may still be quite different due to the multi-dimensional influence of both national and industry-level institutional contexts.

Specifically, research has long established that organizations from the same industry fields need to adapt their behaviors and practice as a process of industry-level isomorphism (DiMaggio & Powell, 1983; O’Connor & Shumate, 2010). One of the most studied institutional contexts is the organizational field. DiMaggio and Powell (1983) noted that organizations embedded in the same organizational field tend to engage in similar practices and strategic orientations due to the mechanisms of isomorphism. These institutional pressures force organizations to modify or adapt their characteristics in compliance with the norms and rules in their operating environment (Yang et al., 2020).

Applying this notion to the social partnership communication context, scholars have argued that organizational fields can influence how organizations communicate. For instance, O’Connor and Shumate (2010) suggest that corporations’ social partnership communication varies by industry areas. Similarly, Yang and Liu (2018) found that in the context of sponsoring environmental issues, certain industries that were commonly associated with environmental hazards such as oil and gas industries, were significantly more likely to engage in social partnership building. For NGOs, their issue areas may affect the likelihood of being picked as an alliance partner. As corporations strategically select issues to support based on their core missions or stakeholders’ demand (O’Connor & Shumate, 2010), it is reasonable to expect that NGOs working in various issue areas would attract different numbers of corporate partners. To test the role of organizational fields, we propose the first research question:

• RQ2: How do organizational fields of corporations and NGOs affect the communication of interorganizational ties related to COVID-19 relief on Twitter?

In addition, the COVID-19 pandemic is truly a global challenge. Since different countries’ socio-economic contexts present different institutional environments, studying global organizations’ COVID-19 responses offers us the rare opportunity to go beyond examining organizational fields.

National Level Institutions and Social Partnership Communication

The National Business System (NBS) theory provides a broad comparative framework to explore macro and national level institutional antecedents of social partnership practices. Importantly, the NBS theory understands social partnership within the context of specific national system configurations (Jamali & Karam, 2018). It posits that the differences of social responsibility beliefs and practices across countries can be explained by the durable institutional elements of national-level institutions, such as financial institutions, governance systems, culture, and social norms (Jamali & Karam, 2018; Lundvall, 1999; Matten & Moon, 2008).

During the COVID-19 pandemic, given the extreme complexity and multi-dimensional impacts of the COVID-19 pandemic on travel, trade, and labor mobility around the globe, organizations’ decisions to engage in partnerships may be influenced by a wide variety of institutional factors. This study explores the effects of four critical aspects of institutional arrangements as prescribed by the NBS theory (Matten & Moon, 2008), including the political system, economic system, cultural system, and education, with modifications based on the COVID-19 issue context.

Political institutions

Extensive research has suggested that companies’ social partnerships are affected by political institutions (e.g., Knudsen et al., 2015; Snider et al., 2013). While previous NBS studies have established the impact of civil rights democracy on social partnerships (Aaronson, 2005; Moon, 2004; Snider et al., 2013), in this study, we also consider the impact of populism on COVID-19 social partnership communication because we observe that in countries such as the United States, United Kingdom, Brazil, and India where populist leaders led the governments in 2020, national responses to COVID-19 were especially chaotic (McKee et al., 2020; Williams et al., 2020).

Civil rights democracy and social partnership communication

NBS Theory suggests that in different political contexts, governments may undertake four types of roles in social responsibility: mandating, facilitating, partnering, and endorsing (Aaronson, 2005; Moon, 2004; Snider et al., 2013). A country’s level of democracy thus affects social partnership communication mainly for two reasons. First, in countries with fewer government restrictions, organizations have more opportunities and freedom to express their views on social and political issues. Similarly, the general publics in democratic societies tend to have more freedom to organize, express their opinions, and hold corporations accountable when necessary (Matten & Crane, 2005). In systems like these, organizations are more likely to freely engage in social partnerships and communication (Palazzo & Scherer, 2008). Second, some social responsibility activities are traditionally viewed as governmental responsibilities (Margolis & Walsh, 2003). In countries where governments hold a central role in civil lives, organizations are less expected to contribute. On the contrary, in countries with small governments, the political system allows “greater scope for corporate discretion, since the government has been less active” (Matten & Moon, 2008, p. 407). Informed by the NBS theory, we propose the following hypothesis:

• H1: A country’s level of civil rights democracy will positively affect organizations’ communication of interorganizational ties related to COVID-19 relief on Twitter.

Populist governments

NBS Theory proposes that governments could provide incentives and policies that may encourage or constrain social partnership communication (Moon, 2004). Some types of governments, such as populist governments, may discourage organizations’ contribution to COVID-19 related social partnerships and communication.

Although the definition of populism remains contentious, Mudde and Kaltwasser (2017) define populism as “a thin-centered ideology that considers society to be ultimately separated into two homogenous and antagonistic camps, ‘the pure people’ versus ‘the corrupt elite’” (p. 27). Populism has been researched since the late 1960s (Pappas, 2016). In the recent decade, the rise of populism ideologies and political parties introduced a powerful populist political movement (Chernov, 2019). Populist governance tends to place “the people” at the core of their statements, and/or are “anti-elites” (Jagers & Walgrave, 2007). Politicians with anti-elite ideologies often claim that they speak the hearts and minds of ordinary people and openly challenge established political figures, intellectuals, and scientists (Chernov, 2019).

Populism has been linked to the politicization of science, and in particular, the rejection of “truth-speaking sovereignty” of scientific elites, such as scholars and intellectuals (Mede & Schäfer, 2020). In times of a public health crisis, the disregard of scientific evidence and recommendations from public health experts can be deadly. Growing empirical evidence has suggested that during the COVID-19 pandemic, countries led by populist governments all performed poorly in their crisis response (McKee et al., 2020; Williams et al., 2020). Previous research has also indicated that populism creates challenges for achieving the best community coping strategies (Lasco & Curato, 2019). Informed by previous research, we propose that organizations from countries under populist governance in 2020 may be concerned about the political reciprocation and backlash, and therefore less willing to engage in COVID-19 related social partnership communication on Twitter, hence:

• H2: Organizations from countries currently led by a populist government are less likely to communicate interorganizational ties related to COVID-19 relief on Twitter.

Economic institutions

NBS Theory proposes that economic institutions critically shape social partnership communication (Matten & Moon, 2008). We focus on the following two aspects of the economic environment that likely shape COVID-related organizational responses: the level of economic development and economic disparity. The level of economic development could determine a country’s available resources for responding to COVID-19. Economic disparity may make certain subpopulations of a country disproportionately more vulnerable to the adversary impact of the pandemic despite the country’s abundant resources.

According to NBS, the level of economic development may influence social responsibility efforts through the following mechanisms. First, in highly developed economies, the stock market is more likely to be the primary means for corporations to pull financial resources. Stakeholders in the stock market thus create a greater expectation for corporate pro-society behaviors (Yang & Liu, 2020). Second, engaging in social partnerships requires considerable resources (Hasan & Habib, 2017). Organizations operating in a more developed economy therefore are more likely to possess the financial and institutional resources needed, particularly in a disaster-relief context. Informed by the NBS theory, we therefore propose:

• H3: A country’s level of economic development (as reflected in GDP per capita) will positively affect organizations’ communication of interorganizational ties related to COVID-19 relief on Twitter.

In addition to economic development, the societal level of economic disparity and hardship may be another factor that potentially shapes organizations’ social partnership communication. Although economic hardship may reduce organizations’ financial resources, such challenges can also drastically increase the social demand for organizations to step up and contribute (Manubens, 2009). Kavoura and Sahinidis (2015) assessed the impact of an economic crisis on a group of Greek companies’ CSR communication, and found that their post-crisis CSR practices were actually more effective than pre-crisis ones. Meanwhile, scholars posit that as a social and environmental crisis is more acutely felt in countries experiencing economic challenges, the expectation of socially responsible efforts may be greater in these economies (Chatterjee & Mitra, 2017). Taken together, for countries that experience greater economic challenges such as high unemployment rates or high inequality levels, organizations operating there may be more expected to step up and shoulder social responsibilities. We therefore propose the following hypothesis:

• H4: A country’s level of economic challenges, as measured by unemployment rate (H4a) and inequality (H4b), will positively affect organizations’ communication of interorganizational ties related to COVID-19 relief on Twitter.

Educational institutions

The NBS theory posits that education systems and the educational levels of the workforce influence social partnerships (Chapple & Moon, 2005; Matten & Moon, 2008). In this paper, we are particularly interested in how the average educational level of a country may influence organizations’ COVID-related social partnership communication. On the one hand, COVID-19 presents both safety and financial challenges to organizations’ employees. Citizens in countries where the labor force has a higher level of education are more likely to advocate for labor rights and have higher expectations for organizations to support workers (Yang & Liu, 2018). In the case of COVID-19, citizens in these countries may demand better employee protection and paid sick leaves, and request organizations to be more responsible towards communities (Matten & Moon, 2008). Additionally, citizens’ educational level is associated with prosocial and ethical consumption behaviors (Starr, 2009). An organization’s prosocial activities such as aiding local communities and donating medical supplies may have a positive influence on consumers’ perception about the organization. As COVID-19 continues to spread globally, the public may think favorably of organizations that respond actively towards COVID-19, which further motivates organizations to adopt relevant practices to mirror this trend (He & Harris, 2020). We therefore propose the following hypothesis:

• H5: A country’s average educational level will positively affect organizations’ communication of interorganizational ties related to COVID-19 relief on Twitter.

Cultural institutions

Finally, NBS theory suggests that national culture also shapes organizations’ decisions to engage in COVID-19 related social partnerships (Matten & Moon, 2008). Culture has long been an important factor in shaping various organizational behaviors from leadership styles (e.g., House et al., 2004) to CSR practices (Peng et al., 2014). Among the literature exploring the relationship between national culture and social partnerships, two of Hofstede’s (1980) cultural dimensions—individualism/collectivism and uncertainty avoidance—are of particular relevance in a public health crisis context (Peng et al., 2014; Ringov & Zollo, 2007). This is because responses to public health crises may require individuals to give up their personal, short-term interests to benefit collective interests. Further, public health crises often trigger considerable uncertainty.

In societies that score high on individualism, the connections between individuals are loose. People are expected to be responsible for themselves. In societies that score high on collectivism, “people from birth onwards are integrated into strong, cohesive in-groups, often extended families which continue protecting them in exchange for unquestioning loyalty” (Hofstede, 2003, p. 2). The collectivist culture emphasizes group performance and rewards the collective distribution of resources and collective action. The collectivist culture thus likely promotes greater collaborative initiatives and alliance building, as the broad societal interests are taken into organizational decision making. Waldman et al. (2006), for example, found that collectivism significantly predicted greater social responsibilities values of the top management team members among corporations.

Additionally, uncertainty avoidance describes the cultural orientation towards risk and stability. Previous research found that uncertainty avoidance significantly predicted firms’ CSR actions related to crises (Peng et al., 2014). An unprecedented crisis such as the COVID-19 pandemic introduces considerable uncertainty. In countries where members of society are comfortable with uncertainty, there may be less demand for organizational actions. By contrast, in countries where citizens are intolerant of uncertainty, the societal demand for COVID-19 related partnerships may be higher. Since no previous studies have specifically examined how individualism-collectivism culture and uncertainty avoidance may affect social partnership communication during a public health crisis, we propose a research question to guide us:

• RQ3: What is the relationship between national culture (indicated by individualism-collectivism and uncertainty avoidance) and organizations’ communication of interorganizational ties related to COVID-19 relief on Twitter?

Please see Table 3 for a list of hypotheses and research questions.

Table 3.

A Summary of Research Questions and Hypotheses Regarding the Formation of COVID-19 Social Partnerships.

Research questions addressed and hypotheses tested	Results
RQ1. What are salient themes and their associative patterns emerging from the COVID-19-related social partnership network on Twitter?	COVID-19 related social partnerships on Twitter were characterized by two themes: (1) the mobilization of instrumental support and (2) the expression of symbolic solidarity.
RQ2. How do organizational fields of corporations and NGOs affect the communication of interorganizational ties related to COVID-19 relief on Twitter?	Manufacturing and finance/insurance/real estate were more active; NGOs were generally more active than corporations.
H1. A country’s level of civil rights democracy will positively affect organizations’ communication of interorganizational ties related to COVID-19 relief on Twitter.	Not supported
H2. Organizations from countries currently led by a populist government are less likely to communicate interorganizational ties related to COVID-19 relief on Twitter.	Supported
H3. A country’s level of economic development (as reflected in GDP per capita) will positively affect organizations’ communication of interorganizational ties related to COVID-19 relief on Twitter.	Not supported
H4. A country’s level of economic challenges, as measured by unemployment rate (H4a) and inequality (H4b), will positively affect organizations’ communication of interorganizational ties related to COVID-19 relief on Twitter.	H4a supported, H4b not supported.
H5. A country’s average educational level will positively affect organizations’ communication of interorganizational ties related to COVID-19 relief on Twitter.	Supported
RQ3. What is the relationship between national culture (indicated by individualism-collectivism and uncertainty avoidance) and organizations’ communication of interorganizational ties related to COVID-19 relief on Twitter?	Individualism-collectivism dimension had little influence, whereas uncertainty avoidance was a significant predictor.

Method

Sample

To identify a representative sample of top corporations and NGOs that communicated about social partnerships on Twitter, we took the following steps. First, we started by identifying top global corporate and nonprofit organizations using the lists of (1) Global Fortune 500 corporations in 2020 and (2) the top 500 world nonprofits from the GuideStar database (guidestar.org) based on revenues. Second, we collected the most recent 3,200 tweets ¹ from these organizations’ Twitter accounts (N = 1,930,342) and only retained tweets containing a set of COVID-19 keywords. ² This step generated a total of 56,312 tweets about the COVID-19 pandemic from January 6th to June 4th, 2020. We chose to focus on this time frame because this period captured the first wave of COVID-19. At the initial stage of a crisis is when organizations reach out to build partnerships and notify the public about such relationships. Once such partnerships have been formed and stabilized, there may be less likelihood for organizations to provide continued updates.

Third, we identified all the Twitter mention relationships from the corpus of tweets, extracting 5,857 Twitter accounts that may contain additional corporate or nonprofit accounts. We focused on mention relationships because according to Saxton et al. (2019), in comparison to retweets, mention relations are motivated by the context of communication rather than the messenger or content. Our initial assessment of the data suggested that organizations may retweet one another when they agreed with the content of tweets, but they only mentioned one another when they described certain endeavors both or multiple organizations took on together. In other words, mention relationships were most likely to capture social partnerships related to COVID-19. We then inspected these accounts’ Twitter bio descriptions to supplement the original list with additional corporations and nonprofits. Additional organizations were manually verified through Google search, and their attributes such as national origins and organizational types were collected in the verification process. The final step identified a total of 590 corporations and 1,486 global nonprofits that had at least one interorganizational tie on Twitter about the COVID-19 pandemic.

We thus constructed a network based on the mention relationships among the 2,076 organizations from 100 countries across the seven continents. A total of 2,561 tweets were retrieved that involved the 2,076 organizations and their communication about the pandemic, and this formed the text corpus for semantic network analysis.

Variables

Country’s civil rights democracy level

We focused on the civil rights aspect of a country’s democracy, using the fundamental rights dimension from the Global State Democracy Indices (2015). The dimension of fundamental rights was a composite variable consisting of three sub-areas: (1) access to justice, (2) civil liberties (e.g., freedom of expression), and (3) social rights and equality (e.g., basic welfare and social group equality). The scale ranged between 0 and 1, and a higher score indicated a higher level of civil rights (M = 0.75, SD = 0.16).

Populist government

Our data were extracted from Tony Blair Institute for Global Change (2020), a global database on populist leaders in power. This dichotomous variable measured whether or not populist leaders are currently in power. Among the 100 countries, 11 were identified as currently led by populist governments and 902 organizations in our sample were from these 11 countries.

The development levels of the financial system

A conventional measure of countries’ financial development level is GDP per capita (Harrison, & Berman, 2016). This study used the data provided by the World Bank as expressed in U.S. dollars (World Bank, 2020). We used the logarithm of GDP per capita to avoid high skewness (Pedhazur, 1997, M = 12.23, SD = 1.75).

Country’s economic inequality

Gini coefficient was used to measure a country’s economic disparity (Dorfman, 1979). The Gini coefficient of each country in 2020 was collected from the World Bank Development Indicators database (M = 0.37, SD = 0.10).

Country’s unemployment rate

The unemployment rate of each country in 2020 was collected from the World Bank Development Indicators database (M = 4.93, SD = 3.69).

Country’s educational level

The citizens’ secondary school enrollment rates as of 2020 were collected from the World Bank Development Indicators database, which were further divided by each country’s population to calculate the ratio (M = 0.07, SD = 0.03).

Individualism versus collectivism culture

We use data provided by the seventh wave of the World Values Survey (WVS, 2020). For this variable, the scale ranged between 1 and 100, with a higher value indicating a higher level of individualism (M = 69.45, SD = 0.03).

Uncertainty Avoidance Index (UAI) measures the degree to which a society could tolerate uncertainty and ambiguity. We also used the scales provided by the seventh wave of the WVS (2020), which ranged between 1 and 100, with a higher value indicating a higher level of uncertainty avoidance (M = 49.69, SD = 20.34).

Corporations’ industry fields

Using the NAICS Code, ³ we coded each corporation in 1 of the 10 industries: (1) agriculture forestry, and fishing, (2) mining, (3) construction, (4) manufacturing, (5) transportation, communications, electric, gas, and sanitary services, (6) wholesale trade, (7) retail trade, (8) finance, insurance, and real estate, (9) services, and (10) public administration.

NGOs’ issue areas

Due to the diverse geographic locations and types of nonprofits in the sample, existing typologies were unable to capture the full range of issue areas represented. Therefore, we inductively developed the following six categories: (1) humanitarian and disaster relief NGOs; (2) political and civil rights NGOs; (3) environmental NGOs; (4) health NGOs; (5) technology, education, and cultural NGOs; and (6) others.

Control Variables

Corporate financial status

Research has found that larger and financially established corporations are more likely to build alliances than smaller firms, both due to greater resources that they have and the fact that these corporations tend to be more “CSR minded” in terms of engaging in social partnerships to gain competitive advantages (O’Riordan & Fairbrass, 2008). To control corporations’ financial resources, a dummy variable was created to indicate whether an organization was one of the Global Fortune 500 corporations. About 236 corporations (11.4%) were identified from the Global Fortune 500 list 2020.

Headquarter location

To control for geographic proximity which is a widely acknowledged mechanism predicting interorganizational alliances (Atouba & Shumate, 2015), we constructed the geographic location variable by coding each organization into one of the seven regions based on its headquarter location: (1) Europe (N = 769, 37.0%); (2) Asia (N = 188, 9.1%); (3) North America (N = 795, 38.3%); (4) South and Central America (N = 79, 3.8%); (5) Oceania (N = 27, 1.3%); and (6) Africa (N = 172, 8.3%); and (7) Middle East (N = 39, 1.9%). The headquarter locations of seven organizations (0.3%) were not identifiable. Headquarter location was used to create the homophily dyadic attribute (i.e., the attribute of a relationship, rather than of an actor) in the model, where a homophily relationship existed when two organizations were located in the same geographic region.

Analytic Procedure

To answer the first research question, we conducted a semantic network analysis of all the partnership-related tweets involving the sample organizations (N = 2,561), using AutoMap (http://casos.cs.cmu.edu/projects/automap/). After removing typical stop words such as articles, prepositions, and conjunctions, the analysis first produced a list of most frequently used terms, and then the co-occurrence network matrix containing all terms. As the full network contained over 10,000 unique words, the visualization presented (Figure 1) only showed the core network containing over 700 words with at least 10 appearances. The second component of the analysis involved using Exponential Random Graph Models (ERGMs) to test how national business factors influence the structure of the COVID-19 social partnership network. Compared to traditional linear or logistic regressions, ERGMs are better equipped to model network mechanisms because they account for the network data interdependency issue (Shumate & Palazzolo, 2010). In an ERGM model, the dependent variable is the statistical odd of tie formation between any two organizations in the sample. Using simulation techniques, the results generated from the model can inform researchers about whether certain hypothesized processes significantly influence the likelihood of tie formation. ⁴

Figure 1.

A visualization of the semantic network of COVID-19 related social partnerships on Twitter.

Note. The size of the node indicates degree centrality, and the top 30 words are presented in the visualization.

Results

A Descriptive Analysis of the Content of COVID-19 Partnership Communication on Twitter

The study identified a total of 2,076 organizations active in the COVID-19 interorganizational network, among which 71.6% (N = 1,486) were NGOs and 28.4% (N = 590) were corporations. This sample consisted of companies from 10 different industry fields, and NGOs working on 6 different issue areas. Table 1 provides a descriptive summary based on the size of organizations working in each industry field and issue area.

Table 1.

A Summary of Corporations’ Industry Types and NGOs’ Issue Areas.

By industry types (n = 10)	Number of organizations	Percentage (%)
Corporations (N = 590)
1. Agriculture, forestry, and fishing	5	0.85
2. Mining	27	4.58
3. Construction	6	1.02
4. Manufacturing	102	17.29
5. Transportation, communications, electric, gas, and sanitary service	85	14.41
6. Wholesale trade	40	6.78
7. Retail trade	53	8.98
8. Finance, insurance, and real estate	118	20.0
9. Services	153	25.93
10. Public administration	1	0.02
Sum	590	100
By NGO issues (n = 6)
NGOs (N = 1,486)
1. Humanitarian and disaster relief	426	28.67
2. Political and civil rights	387	26.04
3. Environmental	129	8.68
4. Health, including disease prevention and treatment	185	12.45
5. Technology, education, and cultural	245	16.49
6. Others	114	7.67
Sum	1,486	100

Figure 1 visualizes the semantic network based on the 2,561 tweets communicating COVID-19-related partnerships on Twitter, where the labels of top 30 most frequently-occurring terms were shown. Other than terms directly related to the pandemic, such as “coronavirus,” “COVID,” and “pandemic,” the following types of words were especially prevalent: first, the words related to the provision of relief and other resources, such as “relief,” “help,” “fund,” and “support;” second, the words expressing solidarity and community building, including “communities,” “solidarity,” and “union;” third, the words highlighting the groups that needed the most support, including “vulnerable,” “garment workers,” and “healthcare people.”

The semantic network analysis identified two overarching themes from the partnership communication, the mobilization of instrumental support, and the expression of symbolic solidarity. Tweets that highlighted the distribution of relief resources read like:

“With support from many companies and organizations, @DirectRelief today sent a major shipment of personal protective gear to Chinese hospitals at the epicenter of the #coronavirus outbreak. @FedEx, @HP.”

Meanwhile, the following tweet represented the other type that focused on expressing solidarity and symbolic support:

“We applaud @intel, @fabricatorz, and @unifiedpatents for being among the first to pledge to make their #IP available free of charge for use in ending the #COVID19 pandemic and minimizing the impact of the disease.”

One ERGM model tested the rest of the hypotheses and research questions. Table 2 presents the coefficient estimates. Hypothesis one through five tested the role of various NBS characteristics in predicting organizational alliances on Twitter.

Table 2.

Parameter Values of ERGM Model Predicting the Likelihood of Ties Formation of COVID-19 Social Media Network.

Structural parameters (for control)	Parameters	SE
Density	−8.33***	0.40
H1: Civil rights democracy	0.10	0.13
H2: Populism	−0.16***	0.03
H3: Economic development	−0.02	0.01
H4a: Unemployment rate	−0.07*	0.03
H4b: Inequality (Gini)	0.85***	0.16
H5: Education	0.01**	0.00
RQ2: Cultural orientation
Individualism-collectivism	−0.00	0.00
Uncertainty avoidance	−0.00*	0.00
RQ3a: Corporations’ industry types (N = 14)
1. Agriculture, forestry, and fishing	—	—
2. Mining	0.13	0.22
3. Construction	—	—
4. Manufacturing	0.66***	0.18
5. Transportation, communications, electric, gas, and sanitary service	0.29	0.19
6. Wholesale trade	0.09	0.21
7. Retail trade	0.11	0.20
8. Finance, insurance, and real estate	0.44*	0.19
9. Services	−0.06	0.19
10. Public administration	—	—
RQ3b: NGOs’ issue areas (N = 14)
1. Humanitarian and disaster relief	0.47**	0.18
2. Political and civil rights	0.59**	0.18
3. Environmental	0.64***	0.18
4. Health, including disease prevention and treatment	0.57**	0.18
5. Technology, education, and cultural	0.61***	0.18
6. Other	0.33	0.18
Control variables
Fortune 500 status	0.02	0.05
Geographic homophily	−0.02	0.04

Note. Industry type 1, 3, and 10 were removed from the model, as significantly fewer corporations were from these industries.

Organizational Fields Predictors

RQ2 explored how a corporation’s industry type and an NGO’s issue area may affect their social partnership communication. We found that the manufacturing industry (Estimates = 0.66, p < .001) and finance, insurance, and real estate industry (Estimates = 0.44, p < .05) were the two types that were significantly more active than other industry fields in building online alliances. Meanwhile, NGOs from almost all issue areas were significantly more likely to engage in COVID-19 alliance communication compared to those working in the “other” issue area (see Table 2 for details).

NBS Predictors

H1 predicted that a country’s levels of civil rights democracy positively associate with an organization’s communication of COVID-19 social partnerships, and this H1 was rejected. H2, which tested the role of populism, was supported. The results showed that organizations from countries currently under the populist government regime were significantly less likely to engage in communicating social partnerships on social media (Estimates = −0.16, p < .001).

H3 explored the role of a country’s economic development as reflected in the GDP level in predicting social partnership communication, and H3 was rejected. H4 posited that a country’s level of economic challenge, reflected in the unemployment rate and economic equality, may positively influence organizations’ engagement in COVID-19 related social partnership communication. We found that whereas the level of economic inequality was positively associated with organizational engagement (Estimates = 0.85, p < .001 (H4a supported), unemployment rate was a negative predictor (Estimates = −0.07, p < .05) (H4b rejected).

H5 tested the role of a country’s educational level. Results suggested that organizations from countries with a higher level of education were significantly more likely to engage in COVID-19 social partnership communication (Estimates = 0.01, p < .01).

RQ3 examined how national culture influenced COVID-19 alliance building. The results suggested that the individualism-collectivism dimension had little influence, whereas uncertainty avoidance was a significant predictor. Specifically, organizations from countries with high levels of uncertainty avoidance were significantly less likely to engage in COVID-19 partnership communication than others (Estimates = −0.002, p < .05).

Discussion

The COVID-19 pandemic provides a unique opportunity to examine the content and structure of interorganizational ties formed around the pandemic. In terms of the content of social partnership communication, our analysis suggests that such communication not only focuses on the mobilization of tangible resources, but it also exhibits a strong emphasis on symbolic alliances and community building. This suggests that social media-mediated communication provides unique opportunities for unity and bonding in times of social distancing and sheltering at home. Across different societies, there are shared concerns over the urgency of mobilizing support, assisting the vulnerable groups, and building global alliances. In terms of the structure of the social partnership networks, our results identify key country- and organizational field-level institutional factors that influence the level of social partnerships. Specifically, we find that populist governments, economic hardship and inequality, citizens’ educational levels, and countries’ uncertainty avoidance culture all significantly predict the level of COVID-19 social partnership communication. In addition, we find that manufacturing and financial services companies and nonprofits in most issue areas are significantly more likely to assume an active role in this global network. We discuss the theoretical and empirical implications of our findings below.

Theoretical Implications to the NBS Theory

The role of national context in a global response to COVID-19

The NBS theory proposes that national-level institutions such as political, economic, educational, and cultural systems shape the social norms and values in which organizations operate. Such systems also influence the public’s expectations for organizations’ socially responsible behaviors (Matten & Moon, 2008). In times of a global crisis such as COVID-19, organizations from different countries have behaved differently due to the divergent national systems, even when faced with a common enemy.

Specifically, while previous studies found democracy to be a salient factor influencing organizations’ social partnership communication (Campbell, 2007; Yang et al., 2020), our study shows that in the context of COVID-19, the populist government is a much more salient political factor. We find that organizations from countries led by populist leaders are significantly less likely to engage in COVID-19 partnership formation. There may be a number of reasons to explain this effect. First, when responding to a crisis at a magnitude as large as that of COVID-19, government-led, centralized resource mobilization and coordination are important. But a centralized government may run contrary to populism beliefs, and therefore populist leaders may scramble to bridge the disconnection between reality and ideological promises by using their rhetoric to discredit the pandemic and related relief efforts. Such rhetoric could discourage organizations from publicly engaging in COVID-19 relief. Second, populist leaders often attack authorities (e.g., scientists and public health officials) to appear speaking the mind of the “ordinary” people. For organizations to openly engage in COVID-19 relief efforts, they may be concerned about the backlash from those populist leaders and their followers. To our best knowledge, the current study is the first to demonstrate the impact of populism politics on social partnership communication. Since the trend of populism is on the rise (Martinelli, 2017), future research may further compare how populist governments influence other issue areas such as climate change.

The impact of economic conditions on COVID-19 social partnership communication is more complex than suggested in the original NBS theory (Matten & Moon, 2008). Our study finds that while unemployment rates could decrease organizations’ social partnerships, inequality level leads to higher tie formation rates. It is likely that these two factors illustrate the tension between organizations’ available resources and the need to meet stakeholder expectations and public demand. On the one hand, high unemployment rates may suggest that many organizations are constrained in available resources and thus have few to allocate to social causes. On the other hand, soaring inequality levels mean that many are extremely vulnerable to the threats posed by COVID-19, so the societal demand for organizational actions is likely to be high. Take the U.S. for example, although it is considered one of the richest countries, the U.S. also has a high inequality level. During the pandemic, millions lined up outside food banks for basic necessities, and countless families faced evictions due to COVID-19 related financial distress. The great economic disparity has prompted many U.S. corporations and nonprofits to step in and contribute to COVID-19 relief efforts.

The impact of education is consistent with the NBS theory’s prediction and that of previous research (Matten & Moon, 2008; Yang et al., 2020; Yang & Liu, 2020). We find organizations from countries with higher secondary education enrollment rates tend to be more active in forming COVID-19 social partnerships. It is likely that in countries with more educated citizens, the public place considerable value on organizations’ socially responsible behaviors.

Finally, we identify a significant impact of countries’ uncertainty avoidance culture on organizations’ social partnership communication. We find that organizations from cultures exhibiting high levels of uncertainty avoidance are more likely to step in and act in response to COVID-19. It is likely that in countries with low uncertainty tolerance, publics and organizations are greatly troubled by COVID-19 related uncertainties and thus urgently demand relief actions. This finding highlights the importance of culture for future efforts to mobilize organizational resources and address wicked problems such as climate change or structural inequality. In countries with low uncertainty tolerance, communication practitioners may develop messages that highlight the uncertainties associated with those challenges, which may lead to favorable public opinions.

Nuances in organizational fields

Our analysis identified the nuanced effects associated with different organizational fields. Overall, around the world, the NGO sector is more active and eager to build interorganizational ties around COVID-19 relief efforts than corporations. Our analysis suggests that in most issue areas, including humanitarian and disaster relief, political and civil rights, environmental protection, health, technology, education, and culture, NGOs are actively forming COVID-19 relief communication ties. This is not only a testimony of the multifaceted nature of challenges associated with the pandemic, but also illustrates the unique institutional norms characterizing the nonprofit sector. Compared to their private counterparts, nonprofit organizations are more expected to form collaborative relationships and contribute to the solution of wicked social problems (Mosley, 2014). Overall, such a finding illustrates the leading roles of NGOs in issue advocacy and their great values as partners in social partnerships.

In addition, for corporations, our analysis shows that manufacturing and financial services companies are significantly more likely to form COVID-19 interorganizational ties than other industry areas. This may be explained by two reasons. First, both manufacturing and financial service industries are consumer-facing. An active engagement in disaster relief may be directly driven by shareholder expectations (Madsen & Rodgers, 2015). Second, compared to other industries, such as tourism and retail that have been facing great distress from the COVID-19 pandemic, these two industries are less compromised financially. Corporations from these industries are thus better resourced to assist with crisis relief.

Practical Implications

Our findings provide several practical implications. First, the fact that national-level institutional factors influence the communication of social partnership suggests that national context needs to be taken into consideration when involving both corporations and NGOs in crisis responses. In a society led by populist government, for example, more proactive effort is needed to normalize and promote social partnerships. Second, as corporations from different industrial fields show varying levels of partnership participation, from the standpoint of mobilizing corporate contribution, it would be cost-effective for nonprofits or policy makers to target corporations from certain industries than others. Last but not the least, given the active role of NGOs in social partnerships, crisis managers are advised to design campaigns that center on nonprofits and leverage their social connections to reach a wider scope of resources.

Limitations and Future Research

Our research has a number of limitations that can be addressed through future research. First, our data were collected on social media. Therefore, any countries where social media such as Twitter is banned, are excluded from the study. The current findings thus cannot be generalized to authoritarian regimes like China or North Korea. To supplement social media data, future work may collect additional data from corporate reports and third-party reports to replicate our study and verify the robustness of these findings using additional data. Second, as part of the nonprofit sample is identified from GuideStar, the nonprofits included in the current study may disproportionately represent U.S. based large nonprofits. Inevitably, this leaves out small and local nonprofits that may be equally involved in the pandemic. Last but not least, COVID-19 is far from the only common issue that countries from around the world face. Future studies may collect data on social partnership networks across multiple issue fields, such as climate change and structural inequality, and see if the variables identified in our study remain significant. It is important to understand the interplay between institutional variables and specific issue context, as such findings may help us develop better theoretical frameworks to account for the impact of complex institutional contexts.

Conclusion

In their fight against the common enemy, organizations from different countries demonstrated different COVID-19 social partnership communication patterns. Understanding how national- and organizational field-level institutions affect organizations’ social partnership communication allows organizations to design and implement country and industry-specific campaigns to raise public support and strengthen partnerships. In response to COVID-19 and other challenges such as climate change, more efforts are needed to build and expand national and international collaboration networks around shared concerns to fight for a better future.

Author Biographies

Wenlin Liu is an Assistant Professor of Strategic Communication at the University of Houston. Her research focuses on interorganizational alliance, disasters, and social media.

Aimei Yang is an Associate Professor of Public Relations at the University of Southern California.

Jieun Shin is an Assistant Professor of Communication at the University of Florida.

Yiqi Li is an Assistant Professor of Information Studies at the Syracuse University.

Jingyi Sun is an Assistant Professor of Business at the Stevens Institute of Technology.

Yan Qu is a Postdoc Associate at the University of Minnesota, Twin City.

Lichen Zhen is a Doctoral Student at the University of Southern California.

Hye Min Kim is an Assistant Professor of Communication at the University of Massachusetts Boston.

Chuqing Dong is an Assistant Professor of Public Relations at the Michigan State University.