Sage Journals: Discover world-class research

Abstract

Guided by institutional theory, this study examines how homophily and institutional power influence the tie formation and dissolution of interorganizational collaboration networks. The analysis focuses on longitudinal network data collected from 174 international non-governmental organizations and inter-governmental organizations (IGOs) whose mission and main activities revolved around HIV/AIDS and other related health topics such as substance use, alcohol use, and smoking. This study conceptualizes collaboration as an affinity communication network and uses Separable Temporal Exponential Random Graph Modeling to investigate the effects of organizational and sector level attributes. The results reveal the important role that homophily plays in terms of geolocation and topic alignment. Furthermore, the results show IGOs’ role in driving these global partnerships. Implications on how to sustain global health alliances are discussed.

Keywords

organizational field interorganizational networks affinity networks public health separable temporal exponential random graph modeling

The popularity and importance of interorganizational alliances have grown considerably in recent years to address increased health care demand and pressure for greater efficiency. Interorganizational alliances on the global stage involve different actors, such as non-governmental organizations (NGOs) and inter-governmental organizations (IGOs), who interact for a variety of purposes (Foot et al., 2021; Wang & Yang, 2020). Internationally coordinated responses to HIV/AIDS provide a prime example of IGOs and NGOs working together to raise awareness and mobilize resources. Such collaboration has led to an effort to approach HIV/AIDS with a comprehensive strategy encompassing related issues such as substance abuse, alcohol use, other health risks, and even human rights (Kim, 2015). IGOs and NGOs across borders work together to advance health concerns by advocating for a specific cause, raising resources, providing services, and conducting research (Atouba & Shumate, 2015).

Interorganizational alliances are affinity networks—socially constructed relationships among organizational actors (Shumate & Contractor, 2013). Affinity networks capture a relationship’s positive or negative valence (i.e., friendship, ownership, association) and differ from other types of networks (such as flow networks that describe the flow of messages or information) in that they consist of more resilient ties which do not decay quickly over time (Shumate & Contractor, 2013). However, these networks can change or evolve (O’Connor & Shumate, 2018). This study focuses on alliance networks designed to generate positive outcomes in the global health realm. These affinity networks do not necessarily imply the flow, exchange, or transmission of information, and are constructed based on the agency of actors (e.g., an NGO can form positive attachments to an IGO for advocacy work).

We use the term “affinity communication networks” to describe communicated partnerships between various organizations that have formed to solve challenging problems. These partnerships are perceived as communication networks that can be identified from organizational reports on their websites or social media channels (Shumate et al., 2018). Because of the important role communication plays in constructing affinity communication networks, organizations take precautions in communicating who they partner with and the nature of their partnerships. International health organizations are important players on the global stage in coordinating resources and health solutions (Kim, 2015). The current study aims to examine what factors drive the structural patterns of affinity communication networks in the public health domain, where international health NGOs and IGOs collaborate. Specifically, the study analyzes the mechanisms underlying the formation and dissolution of these partnerships.

The existing literature has primarily focused on what drives networks to form, and few studies have examined the mechanisms of network dissolution. Our study attempts to fill this gap by unpacking whether mechanisms underlying tie formation also apply to network dissolution. Without examining both processes simultaneously (tie formation and tie dissolution), it is inadequate to draw conclusions about institutionalized structural patterns. Situated in affinity communication networks and guided by institutional theory, this study analyzes how organizations drawn by a health concern are influenced by homophily and institutional power when selecting and sustaining their partners. Findings on affinity communication networks will offer insights into why IGOs and NGOs communicate their strategic partnerships and how they sustain these relationships. The inquiry will also help us understand how the communication of the affinity networks further institutionalizes partnership patterns.

The institutional perspective on interorganizational partnerships (Lammers, 2003; Lammers & Barbour, 2006) argues that the value and regulatory environments of organizations influence what norms are adopted and acted on by organizational members. Approaching institutional theory from the lens of organizational communication, Lammers and Barbour (2006) adopted the sociological definition of institutions and emphasized a process orientation. Their approach places communication at the center of the institutionalizing process, as communicative practices are expected to create and sustain institutions (Wang & Cooper, 2022). Applying institutional theory to analyze the formation and dissolution of affinity communication networks unpacks how the institutionalization of certain ties occurs and how certain patterns tend to dominate over time.

This paper makes several contributions. First, this study contributes to institutional theory by demonstrating that interorganizational partnership patterns can be assessed as outcomes of isomorphism. Analyzing longitudinal network data from a global sample allows us to empirically test how institutional factors may have differential effects on tie formation and dissolution. The focus on longitudinal data moves beyond a static or cross-sectional view of how networks are formed and addresses the gap in the literature regarding whether mechanisms underlying tie formation may also apply to network dissolution. The findings offer theoretical implications on how and what isomorphic patterns may emerge and sustain over time. Second, the findings reveal different homophily mechanisms that influenced the partnership patterns. It shows that relational homophily was positively related to tie formation and tie dissolution. However, proximity homophily was only positively related to tie formation. Third, this study highlights the important role that IGOs play in sustaining affinity networks with other IGOs and NGOs and how their roles may evolve over time. It offers implications for sustaining affinity communication networks on the global stage to help promote public health.

Affinity Communication Networks Through an Institutional Perspective

Interorganizational partnerships are collaborative relationships that address complex social issues collectively. Institutional theory views partners as embedded in an organizational field where there is “the whole set of organizations with the potential to form or join partnerships in pursuit of common objectives in a recognized institutional space” (AbouAssi et al., 2021, p. 4). Partners belong to different organizational populations defined by organizational forms and resources (Wang et al., 2016). When analyzing affinity communication networks in the global health realm, there are at least two types of populations involved: NGOs and IGOs. The interorganizational partnerships constitute ties of commensalism (i.e., relations between similar units; e.g., NGOs collaborating with other NGOs or IGOs collaborating with other IGOs) and symbiosis (i.e., mutual interdependence between dissimilar units; e.g., NGOs forming relationships with IGOs) to assist with their survival and growth in the institutional environment (Aldrich & Ruef, 2006).

The necessity of forming interorganizational alliances to address health topics rests on the assumption that the scope and complexity of such topics rely on aggregating resources from collective actors. For example, United Nations (UN) agencies formed partnerships with health NGOs in different countries to ensure equitable access to COVID-19 treatments and vaccines. Interorganizational partnerships facilitate a global response. Other examples of affinity communication networks include health education initiatives to bridge the gap between lower-income and higher-income countries (Eichbaum et al., 2021) and to advocate for mental health (Vigo et al., 2019).

Following Shiffman (2017), we define affinity communication networks in the context of health alliances as communicated partnerships that consist of organizational partners linked by a shared concern to address a health issue. The literature has used other related terms to describe these partnerships, such as global health initiatives and global health alliances (Doyle & Patel, 2008). We consider all these relationships to be affinity communication networks. From an institutional lens, affinity communication networks are driven by various factors, including shared goals, resources, and institutional power (O’Connor & Shumate, 2018). Institutions consist of rule-like beliefs, behaviors, or practices; they tend to be fixed, enduring, formal, and independent of organizations (Lammers & Barbour, 2006). Institutions can serve as unseen constraints on organizational behaviors such as partnership formation and dissolution (AbouAssi et al., 2021).

Mechanisms Underlying the Formation of Affinity Communication Networks

The institutional perspective on organizational communication emphasizes that organizations’ value and regulatory environments influence what norms are adopted and acted on by their members (Lammers, 2003). One key concept is institutionalization, which refers to “the processes by which social processes, obligations, or actualities come to take on a rule-like status in social thought and action” (Meyer & Rowan, 1977, p. 341). Institutionalization takes place through homogenization and serial reproduction. Therefore, it has to be examined through the process perspective, which requires scholars to identify events in an observed sequence and examine patterns of interactions over time (Poole, 2013).

The process of institutionalization is often captured by the notion of isomorphism, defined as bureaucratic similarity among organizational structures, cultures, and outputs (DiMaggio & Powell, 1983). The key argument behind isomorphism is that organizations tend to adopt practices from other organizations due to resource and dependency concerns, and homogenization helps them to address goal ambiguity and uncertainty and thus achieve legitimacy. Three types of isomorphism further unpack the institutionalization process: coercive pressure, normative pressure, and mimetic pressure (DiMaggio & Powell, 1983). Coercive pressure is formal and informal pressure exerted on organizations by other organizations, such as government and stakeholder groups upon which they depend; it aims to address legitimacy needs. Normative pressure is concerned with mandatory compliance derived from professionalization or expected social responsibility and comes from professional associations, self-regulation, or industry regulation. Mimetic pressure involves the need to reduce uncertainty and it is achieved via imitation.

Literature on isomorphism often assesses the outcomes of an institutional process in terms of how an organization operates and the specific practices it adopts (e.g., technology innovation). Few scholars have extended the notion of isomorphism to the context of interorganizational partnerships and conceptualized network patterns as the outcome of an institutionalization process (see exceptions in Shumate & O’Connor, 2010; Xu & Woo, 2022). One possible reason for this gap is that organizational communication scholars have argued that institutional theory alone is not equipped to explain inter-organizational connections in communication networks (Fu & Shumate, 2016). In this paper, we apply the notion of isomorphism to network contexts and demonstrate that the homogeneous selection of partners is a process of institutionalization. To explore specific mechanisms of partner selection, we return to the discussion of what drives isomorphism, i.e., how organizations adopt certain practices and patterns from other organizations to reduce goal ambiguity and uncertainty and to obtain legitimacy (Alvesson & Spicer, 2019; Baker et al., 1998; DiMaggio & Powell, 1983). We argue that homophily theory helps explain the tendency for organizations to form partnerships to reduce goal ambiguity and uncertainty via mimetic pressure (AbouAssi et al., 2021; Atouba & Shumate, 2015); while institutional power helps explain the tendency of organizations to form partnerships with certain types of actors (e.g., WHO, UN agencies) to obtain legitimacy, which takes places via coercive pressure (Atouba & Shumate, 2015; Wang & Yang, 2020). Below we explain our rationale for these ideas in more detail to unpack the process of institutionalized patterns in affinity communication networks and what ties are more likely to form and stabilize.

Homophily

One mechanism underlying the isomorphism process of interorganizational partnership is homophily. Homophily refers to the notion that partnerships are more likely to form between similar organizations (Wang et al., 2016). The homophily mechanism helps explain how isomorphism occurs at the interorganizational level and what ties tend to be formed over time. One way in which homophily influences isomorphism is through mimetic pressure. Mimetic pressure addresses the ambiguity and uncertainty organizations may face when forming partnerships. When goals are ambiguous, or the partnership environment creates uncertainty, organizations prefer forming relationships with others similar to them to reduce risks associated with such collaboration (AbouAssi et al., 2021). Once the homophilous patterns emerge, other organizations are more likely to model these patterns and eventually reinforce the effect of homophily on partnership formation.

Homophily has been examined as an important mechanism underlying tie formation. However, existing literature has shown inconsistent results in terms of the effect of homophily on partner selection (Atouba & Shumate, 2010), with some finding significant effects while others demonstrating that the effect is not always present. These inconsistent results may be attributed to how various studies have conceptualized homophily. For example, some studies (e.g., Liu & Shin, 2019) focus on geographic regions as the source of homophily, and others use shared issue priorities (e.g., Sommerfeldt et al., 2022). Existing literature also tends to focus on only one source of homophily and does not account for others. Another possible reason for inconsistent findings is that most of the literature focuses on cross-sectional network data and does not model tie formation patterns over time. To address these gaps in the literature, the current study follows the comprehensive conceptualization from Atouba and Shumate (2015) and defines homophily in terms of assortative, relational, and proximity similarities. Guided by institutional theory, this study also analyzes longitudinal data to investigate what homophilous patterns are more likely to occur in affinity communication networks and form isomorphism over time.

Assortative Homophily: Age

Assortative homophily captures the effects of actor attributes. Homophilous information about actor attributes entails perceived compatibility, trust, and predictability (Atouba & Shumate, 2015). Actors with similar attributes (e.g., age, culture, and operating systems) are more likely to view their collaboration as easier to manage (Rivera et al., 2010). Interorganizational collaboration requires conflict management and strategic coordination. The similarity in organizational attributes facilitates the logistic process of forming partnerships and reduces risks related to uncertainty (Chung et al., 2000). In particular, the literature has argued that organizational age is an important source of assortative homophily that could affect collaboration patterns in a network (Atouba & Shumate, 2015).

In an institutional environment, organizations with similar ages are more likely to experience similar market restrictions and implement similar programs and practices to survive and grow. This phenomenon is referred to as “the cohort effect” (Pfeffer, 1983). Health organizations of similar ages are more likely to form partnerships (Atouba & Shumate, 2015), as these organizations might have developed an interest in addressing similar diseases that were prominent at a given time. Furthermore, similarity in age entails that organizations are familiar with similar institutional rules and beliefs related to managing interorganizational partnerships. Given that age homophily helps reduce uncertainty and ambiguity embedded in partnership selection, we argue that organizations will model each other’s partnership patterns through mimetic pressure. This tendency leads to the emergence of isomorphic patterns driven by age homophily. Therefore, we propose our first hypothesis:

In affinity communication networks, as the distance between organizations’ founding years decreases, the probability of partnership formation increases over time.

Relational Homophily: Health Topic Area

The second type of homophily mechanism is the relational one, which indicates that homophilous information about actors’ relationships could help create or enhance legitimacy, reduce risk, and thus facilitate alliance formation (Rivera et al., 2010). Relational homophily refers to the notion that organizations operating in the same institutional environment are expected to comply with existing rules and principles to obtain or maintain legitimacy. The selection of partners based on relationship homophily buffers organizations from institutional sanctions and sustains their relationships with stakeholders. In this investigation, we focus on one institutional factor: the specific topic areas organizations aim to tackle as their mission (Guo & Acar, 2005; Sommerfeldt et al., 2022). Topic areas that organizations specialize in are defining factors for collaboration, as collaborative relationships can facilitate community building around common public concerns (Li et al., 2021; Yang, 2020). Organizations with the same health concerns are more likely to form partnerships with each other as they have a common understanding of the problem and solutions and could pool resources to collectively tackle challenges (Shiffman, 2017). Furthermore, these organizations are more likely to comply with similar standards and adopt similar practices (Li et al., 2021; O’Connor et al., 2017) and may also share common stakeholder expectations that pressure them to collaborate (Cooper & Shumate, 2012). Over time, the homophily mechanism underlying tie formation would reinforce the isomorphic patterns as more and more organizations mimic the partnership choices to reduce challenges related to collaboration uncertainty. We thus propose our second hypothesis:

In affinity communication networks, organizations focusing on the same health topics are more likely to form partnerships over time.

Proximity Homophily: Geolocation

The third homophily mechanism is proximity-based, which proposes that when actors’ socioeconomic, political, and cultural environments are homophilous, they are more likely to be connected (Rivera et al., 2010). When organizations operate in a similar geographic environment, they develop familiarity and common interests confronting the same local or regional stakeholders (Atouba & Shumate, 2015). International collaboration requires the comprehension of various rules and policies. When a potential partner is in a similar geographic region, it creates familiarity and lower costs in project coordination. Geographic homophily has been examined as proximity in a broader sense, such as continents or global hemispheres, to capture organizations’ operations in similar institutional environments (Wang & Yang, 2020; Yang et al., 2020). Atouba and Shumate (2015) argue that region-ness facilitates the coordination of organized efforts in health alliances and could shape interorganizational relationships through common environmental and historical factors. For example, in the Sub-Saharan Africa region, several global health initiatives have been established to battle malaria (Lover et al., 2017). We argue that over time isomorphic patterns will emerge as organizations mimic each other’s partnership choices to collaborate with other organizations in the same region to reduce uncertainty and ambiguity. We thus propose our third hypothesis:

In affinity communication networks, organizations tend to select partners headquartered in their locale over time.

Role of IGOs in Affinity Communication Networks

Another mechanism to explain network patterns through institutional theory is the effect of institutional power to achieve legitimacy. We argue that organizations tend to form partnerships with certain actors when competing with others for material and symbolic resources (Lammer, 2003). More specifically, we argue that in the context of interorganizational alliances to promote health outcomes, coercive pressure exists to facilitate alliances with intergovernmental organizations such as the World Health Organization (WHO) and UN agencies. Normative pressure (i.e., pressure for mandatory compliance derived from professionalization) also exists to encourage alliances with legacy organizations such as professional associations. We explain below the effects of institutional power through these sources and how they may reinforce isomorphic patterns of affinity communication networks over time.

To begin with, IGOs play significant roles in connecting nation states by gathering and exchanging information, coordinating policies, and specifying basic standards (Ertürk, 2015). IGOs are “established by signature of an agreement engendering obligations between governments” (Atouba & Shumate, 2010, p. 301). In the global health realm, IGOs possess institutional power due to their global visibility, access to political resources, and ability to impose pressure on nation states (Von Borzyskowski & Vabulas, 2019). IGOs work with health organizations in several capacities, including as funders and evaluators. In certain cases, IGOs may contribute to international bargaining in negotiation activities or help to settle disputes (Ertürk, 2015). Building partnerships with IGOs allows state and non-state organizations to gain recognition and legitimacy on the global stage (Atouba & Shumate, 2015; Shumate et al., 2005). Over time, coercive pressure and the need to obtain legitimacy would result in organizations actively seeking partnerships with IGOs and show isomorphic patterns of partnership choices. Thus, we propose our fourth hypothesis:

In affinity communication networks, IGOs are more likely to be connected than what would have been expected by chance alone.

The Dissolution of Affinity Communication Networks

Existing literature on interorganizational networks has primarily focused on examining tie formation. We know little regarding the dissolution of these partnerships (but see Margolin et al., 2015; Xu et al., 2021). Institutional theory often argues that once organizations’ partnerships are formed and institutionalized, it is highly unlikely that organizations will replace existing partners with new ones due to expected benefits. This phenomenon has been conceptualized as network inertia, defined as “a persistent organizational resistance to changing interorganizational dyadic ties or difficulties that an organization faces when it attempts to dissolve old relationships and form new network ties” (Kim et al., 2006, p. 704). However, some organizational scholars have adopted a process-based perspective to argue that networks are flexible enough to dissolve (Baker et al., 1998; Kim et al., 2006). Understanding this process requires scholars to examine longitudinal data to identify whether patterns remain stabilized or institutionalized over time (Poole, 2013).

The lens of tie dissolution could offer implications for the process of institutionalization by explaining why some network ties are stable while others are transient (Baker et al., 1998). Such a lens could also help assess whether network mechanisms underlying interorganizational partnerships result in isomorphic patterns over time. Various factors might influence tie evolution. Some scholars have focused on organizational attributes and found that organizations with more distinct services (i.e., non-overlapping services) would experience greater tie churn, resulting in a higher probability of tie dissolution (Bunger & Huang, 2019). Baker et al. (1998) examined how competition, power, and institutional forces influenced tie stability. Another factor that would influence tie dissolution is fit. Once a partnership is established, ongoing interactions could reveal important information about partner fit and the quality of the relationship. Hence, the hazard of tie dissolution rises when a partner views the relationship as unfit (Clough & Piezunka, 2020).

However, the mechanisms underlying tie retention remain as “black boxes” (Clough & Piezunka, 2020). This study applies the institutional perspective to examine whether mechanisms that influence tie formation also affect the dissolution of affinity communication networks. Consider homophily, a well-studied network mechanism, as an example (Baker et al., 1998). If organizations that share common attributes are more likely to collaborate, do these types of relationships tend to persist over time? Or does the effect of homophily only take place during the tie formation process? The main reason for homophily driving network formation is that similar organizations are more familiar with their institutional environment. Thus, partnerships driven by homophily help reduce goal ambiguity and uncertainty. This can be considered with the specific example of assortative homophily. Would organizations of similar age continue to see the benefits of such peer partnerships? For organizations that work in the same health area and perceive relational homophily, would they continue to maintain their ties to conform to norms and practices in the institutional field (Kim et al., 2006)? Furthermore, would proximity homophily continue to drive tie formation over time, or would organizations adapt and venture out to explore partner resources from other global regions? All these uncertainties raise questions regarding whether isomorphic patterns driven by homophily and mimetic pressure may sustain over time when organizations consider their partnership choices. Given these ideas, we propose our first research question:

RQ1. How do assortative, relational, and proximity-based homophily affect the tie dissolution in affinity communication networks?

Another mechanism that influences tie dissolution is the role of IGOs, which can use their institutional power to obtain legitimacy. In an institutional environment, organizations rely on the resources of others to survive and gain legitimacy (Pfeffer, 1983). Therefore, partnering with organizations of institutional power helps to maintain such interdependence and manage environmental uncertainty (Wang & Yang, 2020). Once a resource is acquired from a partner, an organization may terminate the relationship to reduce dependence (Wang et al., 2016). In this situation, would IGOs’ institutional power generate constant coercive pressure for other organizations to sustain their alliance over time? Or would the relationship with IGOs terminate once partner organizations obtain legitimacy through the alliance? From these considerations, we propose the following research question:

RQ2. What role do IGOs play in the tie dissolution in affinity communication networks?

Method

Data Collection

Data were collected through the following procedures. First, using the Union of International Association’s online database, we identified 174 international health organizations whose mission and activities revolve around HIV/AIDS and other related health topics such as substance use, alcohol use, and smoking. We used keywords organizations chose in their mission statement for the sampling. The Union of International Associations (UIA) is a comprehensive source for identifying international organizations, and its annual yearbook reports information about organizations’ partnerships and attributes (such as location). In our sampling, we filtered organizations that explicitly mentioned HIV/AIDS, substance abuse, alcohol use, and smoking in their descriptive profile or in name. These subtopic areas are closely linked because high-risk behaviors such as substance use can increase the initial risk of HIV directly or indirectly (Berry & Johnson, 2018). When an organization covered multiple domains and approached public health more generally, it was coded as “general.” To further verify the coding of generalist organizations, the two authors selected a random sample of 25 organizations (20% of the total sample) and used organizations’ mission statements on their websites, Twitter bio, and Facebook public pages to code whether they could be categorized as generalists (Li et al., 2021; Yang, 2020). The Cohen’s Kappa was .94, indicating that the coding was satisfactory.

Second, we collected self-reported yearly partnership information of the sampled 174 organizations between 2005 and 2017. Data were collected at 3-year intervals, resulting in five waves: 2005, 2008, 2011, 2014, and 2017. The 3-year interval was believed to be optimal in terms of observing changes in partnership based on consultation with UIA staff and our own initial analysis. There were three types of partnerships recorded: collaboration, funding, and consultation. Network ties were constructed by combining all three types of partnerships. Third, additional data on organizational age, global region, and organizational sector were collected through the UIA yearbooks.

Measurement

Organizational age was calculated by subtracting an organization’s founding year from 2017 (M = 31.96, SD = 25.30). There were 14 organizations with missing data (5 from North America, 4 from Europe, 3 from South America, 2 from Africa, and 1 from the Middle East). The average age was used to replace the missing values.

Global region was measured as the macro regions in which an international organization was headquartered (Hettne, 2005). Five regions were identified: 1 = Africa and the Middle East (n = 28), 2 = Asia and Oceania (n = 22), 3 = South America (n = 6), 4 = Europe (n = 77), and 5 = North America (n = 41).

Topic area captured specific health topics an organization focused on. Five health topic areas were identified: 1 = HIV/AIDS (n = 74), 2 = alcohol (n = 14), 3 = drug (n = 45), 4 = tobacco (n = 22), and 5 = general health topics (n = 19).

Organizational sector was coded as a categorical variable: 1 = IGOs (n = 19), and 0 = NGOs (n = 155).

Analysis

Descriptive network analysis and Separable Temporal Exponential Random Graph Models (STERGM) modeling were conducted to examine partnership formation and dissolution in affinity communication networks. First, descriptive network measures were calculated to provide a general picture of the international health alliance network. Then STERGM models were used to examine all the hypotheses related to network formation and dissolution. Separable temporal exponential random graph modeling models are an extension of exponential random graph models (ERGM). The family of ERGM models allows for simultaneous estimation of structural and attribute parameters to predict whether the probability of tie formation in observed networks differs from by chance. ERGM differs from other methods (such as regression) by accounting for relationship interdependencies, i.e., the formation of ties is influenced by the presence of other ties (Pilny & Atouba, 2018; Robins et al., 2007). Standard ERGM models are useful for modeling tie formation in cross-sectional networks but are limited in their application to dynamic networks. Separable temporal exponential random graph modeling examines tie formation and dissolution as separate yet interrelated processes to estimate the effects of parameters over time (Krivitsky & Goodreau, 2019).

The first equation shown below is about tie formation, in which the expression is conditional on the tie not existing in a prior time period. The second equation describes the dissolution model, which is conditional on tie existence. In the equations, Y_ij is the random variable for the state of the node pair i, j (with realization y_ij), and y^c_ij describes the complement of y_ij, which refers to all the other dyads in the network other than the tie between i and j. The variable δ signifies the edge count of the network y, g is the vector of model statistics for network y, θ is the vector of coefficients for those statistics, and t indicates the time element of networks (Krivitsky & Goodreau, 2019). The dissolution model is also a persistence model, in which a positive estimate of a parameter would indicate a higher chance of tie persistence (p_ij) or a lower chance of tie dissolution (1 - p_ij).

The STATNET package in R was used to run the models (Handcock et al., 2019). Two indicators were assessed for model fitness. First, a model fits the data when all parameters have t values smaller than .10, indicating the standard error of each estimated parameter is within a tolerable range of the actual value of the parameter (Snijders et al., 2006). Specific parameters are significant when the values are within 1.96 standard errors of the parameters estimated by the model (p < .05; Robins et al., 2007). Second, the goodness-of-fit (GOF) measures are reported to demonstrate the adequacy of the model in providing a plausible explanation for additional network dimensions that are not included (Atouba & Shumate, 2015; Pilny & Atouba, 2018). GOF analysis compares the observed graph statistics with the values of these statistics for a large number of simulated networks based on the fitted model. Following Atouba and Shumate (2015) and Robins et al. (2007), a good fit is indicated by a convergence statistic of less than 1.0, and a convergence value can indicate a plausible model fit between 1.0 and 2.0. The model fit can also be assessed more directly through visualization of the goodness of fit (GOF) indicators. In a typical plotting, the vertical axis represents the log-odds of relative frequency, the statistics from the observed network are indicated by the solid lines, the boxplots indicate the median and interquartile ranges, and the light gray lines indicate the range in which 95% of the simulated observations fall (Hunter et al., 2008). A good fit can be concluded if the solid line primarily lies within the boxplots or gray lines.

In the formation and dissolution models, several structural parameters were added: the number of edges, isolates, reciprocity, number of transitive triples (parameter ttriple, i.e., if i->j, j->k, and then i->k), and number of cyclic triples (parameter ctriple, i.e. if i->j, j->k, and then k->i) in the network. The identification of these parameters was conducted by using a stepwise method to determine whether adding one additional parameter could result in a better model fit (Stephens et al., 2016). In the model of network formation, the following parameters were added to test the hypotheses: nodematch for issue area, nodematch for headquarters, nodefactor for IGOs, and absolute difference of age (parameter of absdiff). In the model of network dissolution, we included nodematch for the topic area, nodematch for headquarters, and nodematch for IGOs. The parameter of nodematch examines the homophily effect of a categorical variable (e.g., nodes that belong to a certain group would be more likely to form a relationship), and the parameter of nodefactor examines the degree effect of nodes with a given level of a categorical variable (e.g., IGOs are more connected in a network). The absdiff parameter examines the effect of absolute difference in a nodal attribute between two nodes on the tie formation.

The visualization of GOF plotting (see supplement material¹) demonstrates that the model tested was a good fit to the data as most of the solid lines lay within the boxplots or gray lines. To further assess the GOF, we reported the parameter values from the observed networks and simulated networks in Table 1. It indicates that the models examined for tie formation and tie dissolution fit the data well since the convergence statistics (p values) were all less than 1.0 (Pilny & Atouba, 2018).

Table 1.

Goodness of Fit for STERGM Statistics.

	Observation	Minimum	Mean	Maximum	p value
Formation model
Edges	2142.00	2642.00	2942.27	3158.00	.00
Mutual	639.00	686.00	765.65	815.00	.00
Ctriple	918.00	1086.00	1427.18	1642.00	.00
Ttriple	3671.00	4439.00	5802.72	6736.00	.00
Isolates	161.00	52.00	72.66	98.00	.00
Nodematch(issue)	1085.00	1302.00	1480.79	1609.00	.00
Nodematch(location)	1020.00	1232.00	1427.37	1573.00	.00
Nodefactor(IGOs)	2867.00	3800.00	4286.93	4672.00	.00
Absdiff(Age)	54,937.22	66,363.13	74,346.75	81,257.13	.00
Dissolution model
Edges	1302.00	1323.00	1392.58	1466.00	.00
Mutual	405.00	399.00	412.21	427.00	.08
Ctriple	521.00	471.00	494.45	515.00	.00
Ttriple	2070.00	1902.00	1972.19	2039.00	.00
Isolates	350.00	271.00	282.25	306.00	.00
Nodematch(issue)	665.00	673.00	709.43	743.00	.00
Nodematch(location)	633.00	641.00	673.99	705.00	.00
Nodefactor(IGOs)	809.00	823.00	869.66	912.00	.00
Absdiff(Age)	48.00	46.00	50.60	53.00	.04

Note. For the parameters estimated in the model, a good fit (i.e., convergence) is indicated by a convergence statistic of less than.1. Overall, the statistics table showed that both models tested were of good fit to the data.

Results

Table 2 summarizes the descriptive findings of affinity communication networks. Over the five waves examined, the number of isolates decreased, while the density slowly increased. Network descriptive measures indicated that from 2005 to 2017, the network evolved to be more connected and cohesive. Degree centrality (i.e., the average number of connections across all the nodes in a network) continued to increase over the years. Clustering coefficient (i.e., the degree to which nodes in a network tend to cluster together) also showed a consistent increase over time. Network diameter (i.e., the shortest distance between the most distant nodes in a network) and average path length (i.e., the average of the distances between all the nodes in a network) remained relatively stable over the years.

Table 2.

Network Descriptive Findings for the Five Waves.

Parameter	2005	2008	2011	2014	2017
Isolates	97	74	61	44	30
Network diameter	7	9	7	7	9
Density	.011	.012	.015	.017	.019
Average clustering coefficient	.098	.096	.121	.132	.137
Average path length	3.136	3.308	3.25	3.249	3.303
Number of ties	325	372	457	519	567
Average degree	1.868	2.138	2.621	2.966	3.241

Several IGOs and NGOs remained the most connected organizations in affinity communication networks over time. The most active IGOs included WHO, UN Economic and Social Council, European Commission, and UN Children’s Fund. The most active international NGOs included the International Council of Aids Service Organizations, the Global Network of People Living with HIV/AIDS (GNP+), the International Council on Alcohol and Addictions, and The International Federation of Non-Government Organizations for the Prevention of Drug and Substance Abuse. See supplement material² for network visualizations where nodes with the highest degree centrality scores were labeled with names.

In modeling tie formation, several structural tendencies were significant, including reciprocity and transitivity. The model results indicate that organizations tend to reciprocate partnership nominations (4.25, SE = .16, p < .001). Furthermore, there is a strong tendency for partners’ partners to form partnerships through transitive triple paths (.44, SE = .03, p < .001). However, the tendency toward cyclic triples was negative and significant (−.95, SE = .09, p < .001). Different results related to the transitivity effect show that organizations tend to nominate partners’ partners as partners in future years; however, partners’ partners are reluctant to return the favor. This difference makes sense given that network data were collected based on self-reporting. It indicates that organizations that communicated their partnerships on UIA surveys showed a high willingness to recognize or initiate these partnerships. Furthermore, the positive effect of isolates indicates that, given the network size, having more new organizations as nodes may lead to a higher chance of forming new partnerships.

To further unpack the results related to tie formation, we present the findings on hypothesis testing. H1 stated that the distance between organizations’ founding years would be negatively related to the probability of their forming partnerships. We modeled this by including the “absolute difference” parameter, based on organizations’ ages. The results showed that age difference did not affect tie formation. H1 was thus not supported. H2 stated that health topic homophily would result in a higher probability of tie formation. We modeled this by including the nodematch parameter based on health topic areas. The results showed a significant effect of relational homophily based on topic areas (.63, SE = .15, p < .001). H2 was supported. H3 tested region-ness in influencing tie formation. The nodematch parameter based on geographic regions was included in the modeling. A significant effect was found (.48, SE = .14, p < .001). H3 was supported. H4 stated that IGOs would be more connected in the network. We modeled this with the nodefactor parameter based on the IGO status. The results showed that being an IGO increased an organization’s connectivity in the network (.45, SE = .14, p = .001). H4 was supported.

The model of tie dissolution showed a consistent tendency for the included structural parameters to have significant effects. Reciprocal ties were more likely to persist over time (2.46, SE = .53, p < .001). More transitive triples led to tie persistence or a lower probability of tie dissolution (.28, SE = .08, p < .001). However, having more cyclic triples resulted in a higher probability of tie dissolution (−1.10, SE = .23, p < .001). The transitivity effect found in the tie dissolution model was consistent with the one from the tie formation model, indicating that partnerships with partners’ partners were more likely to sustain over time, but cyclic partnerships were less likely to remain active in future years. This finding can also be explained by the nature of the data collection, i.e., self-reported surveys.

Furthermore, the model revealed that having more isolates led to a lower probability of tie dissolution as it offered chances for forming new ties (3.40, SE = .21, p < .001). To summarize, in these data, the effects of structural parameters on tie formation and tie dissolution were consistent over time.

RQ1 examined how assortative, relational, and proximity-based homophily would influence tie dissolution in affinity communication networks. The results showed that ties based on health topic homophily tended to persist more than by chance alone (.94, SE = .30, p = .002), indicating a lower likelihood of tie dissolution. Proximity-based homophily did not affect tie dissolution. Furthermore, age had no effect either. One thing worth noting is that the cohort effect was modeled with the nodematch parameter, as the absolute difference parameter had not been incorporated in the modeling of tie dissolution (due to a non-significant effect and a worse fit to the data). The only significant effect derived from institutional isomorphism was through the homophily effect of health topics.

RQ2 examined the role that IGOs play in influencing tie dissolution in affinity communication networks. When modeling the degree effect of IGOs, it was not significant. However, we found that the homophily effect among IGOs was significant. In particular, the results showed that ties among IGOs were less likely to persist, indicating that they were more vulnerable to dissolution (−.64, SE = .29, p = .03). This finding indicates that IGOs’ partnerships are more selective and might be resource-bound over time. See Table 3 for the model summary.

Table 3.

Model Summary of STERGM.

Parameter	Formation Model		Dissolution Model
Parameter	Estimate	SE	Estimate	SE
Edges	−7.93***	.29	3.18***	.37
Mutual	4.25***	.16	2.46***	.53
Ctriple	−.95***	.09	−1.09***	.23
Ttriple	.44***	.03	.27***	.08
Isolates	1.33***	.26	3.40***	.21
H1 age (absdiff)	.004	.002
Age (nodematch)			−.85	.58
H2, RQ1 topic area (nodematch)	.63***	.15	.94**	.30
H3, RQ1 location (nodematch)	.48***	.14	−.21	.28
H4 IGOs (nodefactor)	.45**	.14
RQ2 IGO match			−.64*	.29
	AIC = 2442BIC = 2529		AIC = 454.20BIC = 504.20

Note. *p < .01. **p < .05. ***p < .001.

Discussion

International health alliances are essential in mobilizing global resources to tackle complex health issues. These partnerships function as an affinity communication network to demonstrate organizations' agency to form collaborative relationships with each other (Shumate & Contractor, 2013). Guided by the institutional perspective of organizational communication, this investigation focused on the formation and dissolution of affinity communication networks to unpack the process of institutionalization over time. It revisited the notion of isomorphism and examined how homophily and institutional power may contribute to mimetic and coercive pressures when organizations institutionalize their partnership choices. Using longitudinal network data collected from 174 health organizations, this study investigated the important roles that organizational age, topic alignment, geolocation, and IGOs played in influencing network patterns over time. It offers insights into how organizations sustain their partnerships to achieve long-term effects on the global stage (e.g., continued advocacy for patients with HIV/AIDs). We discuss the main findings below through the institutional lens.

Homophily and Isomorphic Partnership Patterns

First, we discuss how three types of homophily influence tie formation over time. This study found that assortative homophily, defined by organizational age, has no effect on tie formation or dissolution in affinity communication networks. That is, the “cohort effect” (Pfeffer, 1983) was not observed in the current study. This finding indicates that organizations do not necessarily form ties just because they share a common set of experiences and exposure to similar events within the same time interval. Although organizational age may entail organizational experience and resources (AbouAssi et al., 2021), our study shows that it is not a determining factor when deciding on partners. When a new health organization enters the global health alliance, its age should not be viewed as a barrier to forming partnerships. This finding might encourage newer organizational members to take the initiatives and join forces with legacy organizations to advocate for health courses on the global stage. The absence of age homophily in the current study also indicates that the relationship between assortative homophily and isomorphism was not established in affinity communication networks.

Second, our study found that organizations tend to form partnerships within the same sub-health area in affinity communication networks. This finding indicates the presence of relational homophily among organizations based on common interests and expertise. This result is consistent with the findings from previous studies that topic area alignment is key in partnership formation (Atouba & Shumate, 2015; Sommerfeldt et al., 2022). We also found that partnerships based on a common health topic were more likely to sustain over the long term, resulting in the formation of isomorphic partnership patterns as organizations mimic each other’s partnership choices. This pattern may arise in the public health domain because organizations tend to derive their identity from the subtopic they focus on (Li et al., 2021). Another possible explanation is that when organizations tackle the same health challenge, they are situated in a similar institutional environment which requires them to adopt similar standards and practices. Thus, familiarity may lower perceived risks related to collaboration and coordination (Shiffman, 2017). Although health organizations form affinity networks across different subtopics, it appears that long-term collaborations still occur within the same sectors, such as alcohol, drugs, tobacco, and HIV/AIDS. These findings suggest that relational homophily plays a significant role in the institutionalization process of affinity communication networks. Ties built within a particular health area are more likely to be communicated and thus emerge as institutionalized patterns over time.

We also found that location played an important role in forming affinity communication networks. However, proximity homophily had no effect on sustaining these networks. From the process perspective, this indicates that homophily patterns driven by geographic proximity reinforce isomorphic patterns of tie formation over time. It is consistent with the argument that organizations tend to seek partners in their region, because it reduces uncertainty and ambiguity. Organizations in the same region operate in similar institutional environments; thus, the low risk or cost may deter them from branching out to other regions (Wang & Yang, 2020). When organizations obtain benefits via proximity-based homophilous relationships, other organizations mimic the patterns. Another possible explanation is that common stakeholders or similar health concerns in the same region may have promoted such partnerships (Atouba & Shumate, 2015; Li et al., 2021).

However, as time passes, international health organizations may be more likely to balance their regional and global partnerships, thus challenging the isomorphic patterns that had been established. One possible reason why these regional partnerships may be less likely to persist is that affinity communication networks were formed around limited resources; once the resources are exhausted, these partnerships will dissolve (Lover et al., 2017). Another possible reason is that these affinity communication networks may have been formed in response to a particular health hazard a region was facing (Castillo-Chavez et al., 2015), and once the threat was under control, the partnerships were no longer needed. Given these findings, we conclude that regional partnerships are more likely to be short-lived and do not necessarily become institutionalized over time. It also suggests that affinity communication networks are flexible to change and not necessarily subject to network inertia (Kim et al., 2006).

IGOs and Isomorphic Partnership Patterns

Another mechanism underlying isomorphic partnership patterns can be attributed to the effect of institutional power and the need to obtain legitimacy. This study found that IGOs play a significant role in tie formation and tie dissolution processes in affinity communication networks. The positive degree effect of IGOs on tie formation and the negative effect of homophily ties among IGOs are worth discussing given that this study contained a small number of IGOs. Despite the small set of IGOs in our sample, the results showed that they were more likely to be connected or nominated as partners on the global stage. This emphasizes the central role of IGOs in building a global health community (Atouba & Shumate, 2010). This finding is likely due to their institutional power influencing norms surrounding partnerships and generating coercive pressure. Representing nation states, IGOs possess resources and legitimacy (Von Borzyskowski & Vabulas, 2019). Previous studies suggest that maintaining consultative status with IGOs (such as the UN and WHO) is an acknowledgment of an NGO’s expertise in a particular area, elevating the organization’s social-cultural status (Atouba & Shumate, 2015; Liu & Shin, 2019). IGOs also play an essential role in mobilizing global solidarity (Van de Pas et al., 2017). It is almost a necessity for NGOs to partner with them to be recognized on the global stage and achieve their goals (Steffek, 2013).

The results from the tie dissolution model showed that the degree effect of IGOs was no longer significant, but over time the ties among IGOs were more likely to dissolve. The small number of IGOs in the sample indicates that IGOs are more likely to reach their relational capacity, which is the maximum number of connections their resources can sustain (Monge et al., 2008). This finding indicates that IGOs may play a significant role in driving the initial growth of the global health alliance network by bringing in their peers and NGOs; however, homophilous ties among IGOs are not necessarily sustained.

This finding suggests that cross-sector collaboration between IGOs and NGOs should be communicated to sustain international health alliances. Besides the coercive pressure generated by IGOs, partnerships between NGOs and IGOs have sustained over time to demonstrate mutual benefits to both sectors. NGOs and IGOs have symbiotic relationships (Steffek, 2013). NGOs seek partnerships with IGOs to gain legitimacy, bypass a nation state that might be repressive, and seek funding from IGOs. Likewise, IGOs need cooperation from NGOs when implementing projects, monitoring compliance with international norms, acquiring expertise in formulating policies, and obtaining legitimacy in local communities. Communicating cross-sector collaborations helps address global health challenges by accounting for various stakeholders from all sectors and mobilizing comprehensive expertise and resources (Dain, 2018).

Limitations and Future Research

Several limitations warrant future research. First, our sampling was drawn from a selected list of health topic areas (e.g., HIV/AIDs and substance use). These areas are considered successful examples of international-level coordination in dealing with global health problems (Ertürk, 2015). However, the sampled organizations do not necessarily represent the landscape of global health due to the international focus of the UIA dataset. In other words, we primarily focused on IGOs and international NGOs rather than other organizations such as businesses. We call for future research to consider a more comprehensive list of organizations to replicate the analysis to increase generalizability.

Second, due to missing data, this study did not capture the influences of organizational resources (e.g., the number of staff and financial resources). Organizational resources could explain partnership patterns (AbouAssi et al., 2021; Wang et al., 2016). Resource variables can be tied to assortative homophily. It is also possible that other factors, such as organizational types (e.g., advocacy-oriented vs. service-oriented), may have played a role in influencing tie formation or dissolution. We suspect that missing data occurred as some organizations were not able to collect or report the relevant data. Additional archive data from their websites also did not generate content for coding. Future research can explore other data and examine how organizational attributes beyond age influence tie formation and dissolution.

Last, we collected data through a secondary source (i.e., the UIA annual reports). Affinity communication networks such as the global health alliances are hard to measure and often rely on organizations’ self-reports. There might be strategic reasons why some organizations are not willing to share partner information. One major limitation of using the UIA yearbook is that the self-reported ties are often coded as present or absent, simplifying the complexity interorganizational relationships may entail (Pilny & Atouba, 2018). Therefore, the dimensionality or strength of ties is not accounted for. We call for future research to collect additional network data to account for the strength of relationships among organizations and examine whether the patterns revealed in this study remain the same.

Conclusion

This study examined the determinants of partnership formation and dissolution in affinity communication networks. We collected data from an international public health community from 2005 to 2017, focusing on the IGOs and NGOs that tackled HIV/AIDS and other related health topics. The research unpacked institutionalized partnership patterns in affinity communication networks among international health organizations, and offered implications on how homophily and institutional power may influence organizations’ partner choices over time.

This study makes several contributions. First, this study contributes to institutional theory by demonstrating that interorganizational partnership patterns can be assessed as outcomes of isomorphism. Analyzing longitudinal network data from a global sample moves beyond a static or cross-sectional view of how networks are formed and addresses the gap in the literature regarding whether mechanisms underlying tie formation may also apply to network dissolution. The findings show that institutional factors have differential effects on tie formation and dissolution, providing a more nuanced view of the process of institutionalization. This study offers theoretical implications on how and what isomorphic patterns may emerge and sustain over time.

Second, this study reveals different homophily mechanisms underlying partnership patterns. Results showed that homophily helps to reduce partnership ambiguity and uncertainty, as demonstrated in the tie formation model. However, homophily does not contribute to the process of institutionalization in the long term. For example, we found that international organizations tend to form affinity communication networks with those who share common sub-interests or who are in the same region, confirming the mechanisms of relational homophily and proximity homophily. The relationships built upon topic alignment are also long-lasting. These findings offer implications on how to sustain interorganizational alliances through partner organizations’ niche focus and their shared interests in a particular health topic area.

Third, findings from this study offer important implications for how IGOs and NGOs form and sustain cross-sector alliances to build a global community that generates long-term capacity in tackling global health concerns. In particular, the research shows the critical role that IGOs play due to their perceived status and power. Our finding also unpack the necessity of symbiosis relationships in sustaining the alliance.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research,authorship,and/or publication of this article: The study was supported by the Keck School of Medicine at the University of Southern California.

ORCID iD

Rong Wang

Notes

Author Biographies

Rong Wang (PhD,University of Southern California) is an Assistant Professor at the Department of Human and Organizational Development,Peabody College,Vanderbilt University. Her work examines how to leverage the network dynamics among organizational and individual actors to solve wicked social issues.

Jieun Shin (PhD,University of Southern California) is an assistant professor in the Department of Media Production,Management,and Technology at the University of Florida. Her research explores information diffusion on social media focusing on misinformation and news use.

References

AbouAssi

Wang

Huang

(2021). Snuggling together or exploring options? A multilevel analysis of nonprofit partnership formation and evolution in an unstable institutional context. Nonprofit and Voluntary Sector Quarterly, 50(1), 143–164. https://doi.org/10.1177/0899764020945800

Aldrich

Ruef

(2006). Organizations evolving. Sage.

Alvesson

Spicer

(2019). Neo-institutional theory and organization studies: A mid-life crisis? Organization Studies, 40(2), 199–218. https://doi.org/10.1177/0170840618772610

Atouba

Shumate

(2010). Interorganizational networking patterns among development organizations. Journal of Communication, 60(2), 293–317. https://doi.org/10.1111/j.1460- 2466.2010.01483.x

Atouba

Y. C.

Shumate

(2015). International nonprofit collaboration: Examining the role of homophily. Nonprofit and Voluntary Sector Quarterly, 44(3), 587–608. https://doi.org/10.1177/0899764014524991

Baker

W. E.

Faulkner

R. R.

Fisher

G. A.

(1998). Hazards of the market: The continuity and dissolution of interorganizational market relationships. American Sociological Review, 63(2), 147–177. https://doi.org/10.2307/2657321

Berry

M. S.

Johnson

M. W.

(2018). Does being drunk or high cause HIV sexual risk behavior? A systematic review of drug administration studies. Pharmacology Biochemistry and Behavior, 164, 125–138. https://doi.org/10.1016/j.pbb.2017.08.009

Bunger

Huang

(2019). Change in collaborative ties in a children’s mental health services network: A clique perspective. Human Service Organizations, Management, Leadership and Governance, 43(2), 74–91. https://doi.org/10.1080/23303131.2019.1606871

Castillo-Chavez

Curtiss

Daszak

Levin

S. A.

Patterson-Lomba

Perrings

Poste

Towers

(2015). Beyond Ebola: Lessons to mitigate future pandemics. The Lancet. Global Health, 3(7), e354–e355. https://doi.org/10.1016/S2214-109X(15)00068-6

10.

Chung

S. A.

Singh

Lee

(2000). Complementarity, status similarity and social capital as drivers of alliance formation. Strategic Management Journal, 21(1), 1–22. https://doi.org/10.1002/(sici)1097-0266(200001)21:1<1::aid-smj63>3.0.co;2-p. https://www.jstor.org/stable/3094116

11.

Clough

D. R.

Piezunka

(2020). Tie dissolution in market networks: A theory of vicarious performance feedback. Administrative Science Quarterly, 65(4), 972–1017. https://doi.org/10.1177/0001839219899606

12.

Cooper

K. R.

Shumate

(2012). Interorganizational collaboration explored through the bona fide network perspective. Management Communication Quarterly, 26(4), 623–654. https://doi.org/10.1177/0893318912462014

13.

Dain

(2018). Challenges facing global health networks: The NCD alliance experience: Comment on “four challenges that global health networks face”. International Journal of Health Policy and Management, 7(3), 282–285. https://doi.org/10.15171/ijhpm.2017.93

14.

DiMaggio

P. J.

Powell

W. W.

(1983). The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review, 48(2), 147–160. https://doi.org/10.2307/2095101

15.

Doyle

Patel

(2008). Civil society organisations and global health initiatives: Problems of legitimacy. Social Science and Medicine, 66(9), 1928–1938. https://doi.org/10.1016/j.socscimed.2007.12.029

16.

Eichbaum

Q. G.

Adams

L. V.

Evert

M. J.

Semali

I. A.

van Schalkwyk

S. C.

(2021). Decolonizing global health education: Rethinking institutional partnerships and approaches. Academic Medicine: Journal of the Association of American Medical Colleges, 96(3), 329–335. http://doi.org/10.1097/ACM.0000000000003473

17.

Ertürk

(2015). Intergovernmental organizations (Igos) and their roles and activities in security, economy, health and environment. Journal of International Social Research, 8(37), 333–340. https://doi.org/10.17719/jisr.20153710606

18.

Foot

Sworn

Alejano-Steele

(2021). An outcome-centered comparative analysis of counter-human trafficking coalitions in the global South. Management Communication Quarterly, 35(3), 418–444. https://doi.org/10.1177/08933189211017925

19.

J. S.

Shumate

(2016). Hyperlinks as institutionalized connective public goods for collective action online. Journal of Computer-Mediated Communication, 21(4), 298–311. https://doi.org/10.1111/jcc4.12161

20.

Guo

Acar

(2005). Understanding collaboration among nonprofit organizations: Combining resource dependency, institutional, and network perspectives. Nonprofit and Voluntary Sector Quarterly, 34(3), 340–361. https://doi.org/10.1177/0899764005275411

21.

Handcock

M. S.

Hunter

D. R.

Butts

C. T.

Goodreau

S. M.

Krivitsky

P. N.

Bender-deMoll

Morris

M. M.

(2019). Package ‘statnet’. http://mirror.psu.ac.th/pub/cran/web/packages/statnet/statnet.pdf

22.

Hunter

D. R.

Goodreau

S. M.

Handcock

M. S.

(2008). Goodness of fit of social network models. Journal of the American Statistical Association, 103(481), 248–258. https://doi.org/10.1198/016214507000000446

23.

Kim

T. Y.

Swaminathan

(2006). Framing interorganizational network change: A network inertia perspective. Academy of Management Review, 31(3), 704–720. https://doi.org/10.5465/amr.2006.21318926

24.

Kim

Y. S.

(2015). World Health Organization and early global response to HIV/AIDS: Emergence and development of international norms. Journal of International and Area Studies, 22(1), 19–40. https://www.jstor.org/stable/43490278

25.

Krivitsky

P. N.

Goodreau

S. M.

(2019). STERGM-Separable Temporal ERGMs for modeling discrete relational dynamics with statnet. Technical Report. http://statnet.csde.washington.edu/workshops/SUNBELT

26.

Lammers

J. C.

(2003). An institutional perspective on communicating corporate responsibility. Management Communication Quarterly, 16(4), 618–624. https://doi.org/10.1177/0893318902250642

27.

Lammers

J. C.

Barbour

J. B.

(2006). An institutional theory of organizational communication. Communication Theory, 16(3), 356–377. https://doi.org/10.1111/j.1468-2885.2006.00274.x

28.

Shin

Sun

Kim

H. M.

Yang

(2021). Organizational sensemaking in tough times: The ecology of NGOs’ COVID-19 issue discourse communities on social media. Computers in Human Behavior, 122, Advance online publication. https://doi.org/10.1016/j.chb.2021.106838

29.

Liu

Shin

(2019). Convergence or divergence: Exploring different mechanisms driving children’s rights organizations’ offline versus online interorganizational alliance building. Telematics and Informatics, 42, 1–12. https://doi.org/10.1016/j.tele.2019.101242

30.

Lover

A. A.

Harvard

K. E.

Lindawson

A. E.

Smith Gueye

Shretta

Gosling

Feachem

(2017). Regional initiatives for malaria elimination: Building and maintaining partnerships. Plos Medicine, 14(10), e1002401–e1002410. https://doi.org/10.1371/journal.pmed.1002401

31.

Margolin

D. B.

Shen

Lee

Weber

M. S.

Fulk

Monge

(2015). Normative influences on network structure in the evolution of the children’s rights NGO network, 1977-2004. Communication Research, 42(1), 30–59. https://doi.org/10.1177/0093650212463731

32.

Meyer

J. W.

Rowan

(1977). Institutionalized organizations: Formal structure as myth and ceremony. American Journal of Sociology, 83(2), 340–363. https://doi.org/10.1086/226550

33.

Monge

Heiss

B. M.

Margolin

D. B.

(2008). Communication network evolution in organizational communities. Communication Theory, 18(4), 449–477. https://doi.org/10.1111/j.1468-2885.2008.00330.x

34.

O’Connor

Parcha

J. M.

Tulibaski

K. L. G.

(2017). The institutionalization of corporate social responsibility communication: An intra-industry comparison of MNCs’ and SMEs’ CSR reports. Management Communication Quarterly, 31(4), 503–532. https://doi.org/10.1177/0893318917704512

35.

O’Connor

Shumate

(2018). A multidimensional network approach to strategic communication. International Journal of Strategic Communication, 12(4), 399–416. https://doi.org/10.1080/1553118X.2018.1452242

36.

Pfeffer

(1983). Organizational demography. Research in Organizational Behavior, 5, 299–357.

37.

Pilny

Atouba

(2018). Modeling valued organizational communication networks using exponential random graph models. Management Communication Quarterly, 32(2), 250–264. https://doi.org/10.1177/0893318917737179

38.

Rivera

M. T.

Soderstrom

S. B.

Uzzi

(2010). Dynamics of dyads in social networks: Assortative, relational, and proximity mechanisms. Annual Review of Sociology, 36(1), 91–115. https://doi:10.1146/annurev.soc.34.040507.134743

39.

Robins

Pattison

Kalish

Lusher

(2007). An introduction to exponential random graph (p*) models for social networks. Social Networks, 29(2), 173–191. https://doi.org/10.1016/j.socnet.2006.08.002

40.

Scott Poole

(2013). On the study of process in communication research. Annals of the International Communication Association, 36(1), 371–409. https://doi.org/10.1080/23808985.2013.11679140

41.

Shiffman

(2017). Four challenges that global health networks face. International Journal of Health Policy and Management, 6(4), 183–189. https://doi.org/10.15171/ijhpm.2017.14

42.

Shumate

Contractor

N. S.

(2013). Emergence of multidimensional social networks. In Putnam

L. L.

Mumby

D. K.

(Eds.), The SAGE handbook of organizational communication (pp. 449–474). Sage.

43.

Shumate

Fulk

Monge

(2005). Predictors of the international HIV–AIDS INGO network over time. Human Communication Research, 31(4), 482–510. https://doi.org/10.1111/j.1468-2958.2005.tb00880.x

44.

Shumate

Hsieh

Y. P.

O’Connor

(2018). A nonprofit perspective on business– nonprofit partnerships: Extending the symbiotic sustainability model. Business and Society, 57(7), 1337–1373. https://doi.org/10.1177/0007650316645051

45.

Shumate

O’Connor

(2010). Corporate reporting of cross-sector alliances: The portfolio of NGO partners communicated on corporate websites. Communication Monographs, 77(2), 207–230. https://doi.org/10.1080/03637751003758201

46.

Snijders

T. A. B.

Pattison

P. E.

Robins

G. L.

Handcock

M. S.

(2006). New specifications for exponential random graph models. Sociological Methodology, 36(1), 99–153. https://doi.org/10.1111/j.1467-9531.2006.00176.x

47.

Sommerfeldt

E. J.

Saffer

A. J.

Luoma-aho

(2022). Civil society networks and Malaysian government reform: Considering issue homophily in interorganizational relationships. Journal of Communication, 72(2), 264–296. https://doi.org/10.1093/joc/jqac001

48.

Steffek

(2013). Explaining cooperation between IGOs and NGOs – push factors, pull factors, and the policy cycle. Review of International Studies, 39(4), 993–1013. https://doi.org/10.1017/S0260210512000393

49.

Stephens

Chen

Butler

J. S.

(2016). Bubbling up the good ideas: A two-mode network analysis of an intra-organizational idea challenge. Journal of Computer-Mediated Communication, 21(3), 210–229. https://doi.org/10.1111/jcc4.12158

50.

Van de Pas

Hill

P. S.

Hammonds

Ooms

Forman

Waris

Brolan

C. E.

McKee

Sridhar

(2017). Global health governance in the sustainable development goals: Is it grounded in the right to health? Global Challenges, 1(1), 47–60. https://doi.org/10.1002/gch2.1022

51.

Vigo

D. V.

Patel

Becker

Bloom

Yip

Raviola

Saxena

Kleinman

(2019). A partnership for transforming mental health globally. The Lancet. Psychiatry, 6(4), 350–356. https://doi.org/10.1016/S2215-0366(18)30434-6

52.

Von Borzyskowski

Vabulas

(2019). Credible commitments? Explaining IGO suspensions to sanction political backsliding. International Studies Quarterly, 63(1), 139–152. https://doi.org/10.1093/isq/sqy051

53.

Wang

Cooper

K. R.

(2022). Corporate social responsibility in emerging social issues:(Non) institutionalized practices in response to the global refugee crisis. Journal of Communication Management, 26(1), 98–114. https://doi.org/10.1108/JCOM-04-2021-0042

54.

Wang

Tanjasiri

S. P.

Palmer

Valente

T. W.

(2016). Network structure, multiplexity, and evolution as influences on community‐based participatory interventions. Journal of Community Psychology, 44(6), 781–798. https://doi.org/10.1002/jcop.21801

55.

Wang

Yang

(2020). The structure and evolution of the international human rights network: Unpacking the influences of countries’ contextual factors and network configurations. International Interactions, 46(6), 922–952. https://doi.org/10.1080/03050629.2020.1814769

56.

Woo

(2022). Key players in corporate social responsibility (CSR) institutionalization: An analysis of multinational companies’ interorganizational positioning via CSR reports. Management Communication Quarterly. Advance online publication. https://doi.org/10.1177/08933189221095770

57.

Fulk

Monge

(2021). Ecological influences on network tie dissolution in the evolution of affiliation networks in the international communication association, 2009– 2015. Human Communication Research, 47(1), 75–103. https://doi.org/10.1093/hcr/hqaa009

58.

Yang

(2020). The issue niche theory of nongovernmental and nonprofit organizations’ interorganizational network ecology. Communication Theory, 30(1), 41–63. https://doi.org/10.1093/ct/qtz014

59.

Yang

Liu

Wang

(2020). Cross‐sector alliances in the global refugee crisis: An institutional theory approach. Business Ethics: A European Review, 29(3), 646–660. https://doi.org/10.1111/beer.12288

Determinants of Alliance Formation and Dissolution Among International Health Organizations: The Influence of Homophily and Institutional Power in Affinity Communication Networks

Abstract

Keywords

Affinity Communication Networks Through an Institutional Perspective

Mechanisms Underlying the Formation of Affinity Communication Networks

Homophily

Assortative Homophily: Age

Relational Homophily: Health Topic Area

Proximity Homophily: Geolocation

Role of IGOs in Affinity Communication Networks

The Dissolution of Affinity Communication Networks

Method

Data Collection

Measurement

Analysis

Results

Discussion

Homophily and Isomorphic Partnership Patterns

IGOs and Isomorphic Partnership Patterns

Limitations and Future Research

Conclusion

Footnotes

Declaration of Conflicting Interests

Funding

ORCID iD

Notes

Author Biographies

References