Collaborative Innovation in Governance Networks
A social network analytical approach to organizational embeddedness
and innovation in goal-driven networks
Maddisen C. Ravalin
[s2268795]
MSc Public Administration
Track: Public Management and Leadership
A thesis submitted to:
The Faculty of Governance and Global Affairs
University of Leiden | The Hague
Supervised by:
Dr. Joris van der Voet
Second reader:
Dr. Jelmer Schalk
July 15, 2019
TABLE OF CONTENTS
1. INTRODUCTION ... 3
1.1 Problem definition and research question ... 4
1.2 The network: The Hague Humanity Hub ... 6
1.3 Academic and social relevance ... 7
Academic contribution ... 7
Social relevance ... 9
1.4 Thesis structure ... 9
2. THEORY ... 10
2.1 Collaborative innovation in the public innovation literature ... 10
Phase models of innovation processes ... 12
2.2 Collaborative Innovation: Why or why not? ... 13
2.3 Social network theory and analysis ... 16
Network governance and the NAO model ... 17
2.4 Relational content and network embeddedness ... 18
Formal roles ... 18
Interaction ... 19
Transfer of material and non-material resources ... 20
3. RESEARCH DESIGN AND DATA COLLECTION ... 21
3.1 Research design and case selection ... 21
3.2 Data collection and methods ... 22
3.3 Operationalization of variables ... 23
Outcome variable – Collaborative innovation ... 23
Explanatory variables – Network embeddedness ... 24
Control variables ... 27
3.4 Reliability and validity ... 28
3.5 Analytical strategy ... 30
Integrated SNA approach ... 30
Strategy ... 30
4. RESULTS AND ANALYSIS ... 31
4.1 Descriptive statistics ... 31
4.2 Descriptive SNA and Structural Modeling ... 34
Defining centralities – Comparing measures of embeddedness ... 34
Structural models ... 35
Descriptive analysis ... 37
4.3 Linear regression analysis: Relational content ... 43
Assumptions of linear regression ... 44
Analysis ... 44
Discussion of results ... 50
4.4 Linear regression analysis: Central embeddedness ... 51
Analysis ... 52
Discussion of results ... 53
5.1 Discussion ... 54
5.2 Conclusion ... 56
Limitations and future research ... 58
Implications of the study ... 59
References ... 62
Appendices ... 65
Appendix A. – Survey questions and answer options ... 65
Appendix B. – Control variable frequency statistics ... 68
Appendix C. – Linear regression results at Phase 2 and Phase 3 ... 69
1. INTRODUCTION
Global trends unfold in highly varied cultural, economic, and political contexts (Easter & Brannen, 2016). As such, they are problematic for organizations attempting to resolve or mitigate the negative effects of societal problems. As a result, the call for innovative solutions to ‘wicked problems’ is far-reaching in public organizations (Osborne & Brown, 2011; Van der Wal, 2017) because such problems’ complexity renders them unsolvable by ordinary means (Hartley, Sørensen, & Torfing, 2013). It calls into question the effectiveness of current organizational practices (Bekkers & Tummers, 2018), which has made innovation a hot topic for organizational and management scholars across sectors over the last few decades (Damanpour, 1987). As global trends and societal problems impact everyone, innovation is a top priority for private sector businesses, societal organizations, and citizens, as well as politicians and public administrations (Bekkers & Tummers, 2018, p. 210). This trend parallels an emerging view that the public sector is no longer the only legitimate provider of public services. Consequently, innovation is increasingly seen as a collaborative process between multiple and diverse stakeholders (Torfing, Sørensen, & Røiseland, 2019).
Collaborative innovation, a theoretical type and practical strategy endorsed by Torfing (2019), has emerged from this alternative way of thinking about innovation in public governance. Torfing and others assert that multi-actor collaboration is an important driver for public innovation (Wegrich, 2019), demonstrating a break from both Classical Public Bureaucracy and New Public Management (Torfing et al., 2019). While the tradition of co-creating innovative solutions to common problems among government, businesses, and civil society organizations is strong in many societies, it is a comparatively new realm of research for public administration scholars (Torfing et al., 2019, p. 798). As a theory, it is rooted in both public and private sector innovation literature. Borins (2001) draws from both contexts to illustrate how innovation crosses boundaries by necessity, and therefore, he argues, it is uncontainable to a single unit. Its trans-boundary, ancillary nature applies at many levels of analysis, which results in coordinative mechanisms for problem solving and integrated problem solutions (Borins, 2001, pp. 64-65). This foundational idea has led to the rise the ‘milieux of innovation’ approach (Bekkers, Edelenbos, & Steijn, 2011) and more ‘open innovation’ models and perspectives (Osborne & Brown, 2011; Chesborough, 2003) premised upon the idea that innovation is not an insulated process. From these perspectives, organizations are expected to share knowledge and resources to solve problems, making them collaborate, which often occurs in networks (Bekkers & Tummers, 2018). Moreover, collaborative innovation is a process that spans organizational types, sectors, and national boundaries to create public value through collective problem-solving (Voorberg, Bekkers, & Tummers, 2015; Bekkers & Tummers, 2018).
Proponents of this approach argue that all members of society, public and private actors, citizens, and service users, have both the capacity and often the willingness to contribute to this public value creation (Torfing, 2019; Hartley et al., 2013). As such, Bekkers and Tummers (2018)
emphasize the importance of viewing public sector innovation as both stakeholder and multi-value. Diverse stakeholder values must be considered throughout the innovation process, especially when the public sector is not considered the sole generator of public value (p. 212). Individuals bring with them diverse organizational “systems of meaning” and beliefs about the way things can or should be done in multi-stakeholder collaborations and partnerships (Easter & Brannen, 2016). Balancing these often-conflicting values, to create and harness shared understanding, as well as increase trust in and responsiveness of service providers, is one of the great challenges of public innovation.
This thesis focuses on the relationship between network embeddedness and collaborative innovation outcomes of organizations in one such governance network. However, the broad social relevance of collaborative innovation as itself, an “innovation in governance” (Moore & Hartley, 2008) should not be understated. The processes under evaluation in this paper should be considered against this backdrop of a shift in governance paradigms. The network under study concretely highlights a transition away from NPM strategies and emphasis on business-like values toward diverse organizational networks and, more central to this research, fundamental changes in the way public services are resourced via those networks (Moore & Hartley, 2008).
1.1 Problem definition and research question
Support for collaborative innovation as a tool of increasingly networked public governance strategies is widespread (Torfing et al., 2019; Van der Wal, 2017). However, more empirical evidence is needed to discern if and how collaborative strategies drive innovation (Torfing, 2019). Social network theory provides a natural theoretical and methodological approach for studying integrated forms of leadership (Carson, Tesluk, & Marrone, 2007, p. 1220), namely, as this paper argues, collaborative networks. Contrary to other social science perspectives, it prioritizes relational information for understanding outcomes (Wasserman & Faust, 1994, p. 7), which is both theoretically and observably inherent to collaborative innovation. Network analysis entails a systematic search for patterns among social ties, as well as both the conditions and consequences of these patterns for actors embedded in them (Freeman, 2004, p. 2). Despite this intuitive link between collaborative innovation and patterns of relationships, the importance of network structure has so far been overlooked in collaborative innovation research.
This study is concerned with interorganizational relations and the network structure they give rise to as antecedents to collaborative innovation in a cross-sectoral, innovation-driven organizational network. A network, as defined here, is a group of three or more legally autonomous organizations working toward both individual and collective goals (Provan & Kenis, 2008, p. 231). Social innovation hubs and their corresponding communities are a fertile frontier for new research on collaborative innovation in goal-driven networks. These hubs have recently become a prominent
addition to the organizational ecosystems of many cities around the globe, linking together organizations and enterprises from a myriad of initiatives, governments, organizational fields, and sectors. Many offer programs and workspaces that host diverse organizations and facilitate innovative projects or entrepreneurial endeavors. One of the largest network hubs of this kind, the Impact Hub, has over one hundred locations in more than fifty countries (Impact Hub, 2019) and facilitates the exchange of knowledge and other resources amongst organizations of varying sizes, sectors, and other characteristics.
The Hague Humanity Hub (a.k.a. “The Hub” or “The Humanity Hub”), whose ego network will be analyzed in this study, follows a similar model. However, it differentiates itself in two specific ways. First, The Humanity Hub works with select organizations operating in the fields of peace, justice, development, and humanitarian action (The Hague Humanity Hub, 2019a). Second, it explicitly promotes innovation as the desired outcome of inter-organizational collaboration. Participation in the network is defined by the shared goal of innovation in these areas of practice.
In innovation research, scholars and practitioners alike recognize that some organizational factors drive innovation more than others, whilst some inhibit it (Damanpour, 1991). These factors include a variety of internal antecedents, such as organizational size, administrative capacity, and availability of slack resources. However, external communication and boundary spanning seen as critical for organizational learning are also widely acknowledged (Walker, 2014, p. 26). For example, according to Hartley and Rashman (2018), in UK public service reform, innovation is spurred by interorganizational learning, which is an inherently trans-boundary process. Other arguments for collaborative innovation state that it widens the resource pool to address shared problems (Torfing, 2019). Yet, this is no guarantee that collaborative strategies will lead to more innovation (Hartley et al., 2013, p. 827). Further, significant barriers to collaborative innovation in some contexts may inhibit the process, such as the emergence of conflicting interests among involved parties (Hartley et al., 2013). The data collected and analyzed in this paper adds to our empirical understanding of how the collaborative process unfolds in an interorganizational network setting. By analyzing how interactive organizations are in a network, and the resulting network structure(s) indicating organizations’ embeddedness in said network, we can gain some insight into the effectiveness of collaboration for the innovation process. Such insight is based on current understandings of collaborative innovation drivers and inhibitors as well as the theoretical and methodological contributions of the social network analytical (SNA) approach.
Social network theory and analysis supplies the analytical tools to quantify and analyze relationships between organizations in a network (Wasserman & Faust, 1994). Using social network analysis, empirical observations of inter-organizational behaviors in this study are considered relational characteristics of actors within the network at the organizational level of analysis. Variable indicators such as member levels at The Hub, network-related event attendance, and use of an intra-network online communication and media platform are interactive behaviors indicating relational
characteristics or content. These relational characteristics give rise to embeddedness in the network structure and the subsequent values or measures embeddedness can take (Wasserman & Faust, 1994). Using these variable indicators, quantities are measured for how much organizations, who share broad network innovation goals, engage with each other. This is to provide evidence of a relationship between this the content and quantity of relations and the occurrence of collaborative innovation throughout multiple phases of the process. The phases, which are discussed in the theoretical framework, are: idea generation, selection for piloting/testing, and implementation.
Furthermore, this paper identifies indicators of network embeddedness to divulge their relationship with collaborative innovation and aims to shed light on their potential role as driving or inhibiting antecedents of innovation. Collaborative innovation theory suggests more collaborative problem solving between and among “relevant and affected actors” (Torfing, 2019, p. 3) leads to increased levels of interorganizational interaction, granting them more access to diverse tangible and knowledge resources (Hartley et al., 2013; Torfing, 2019). When this interaction is based on interpersonal communication, innovation is expected to be linked to informal social structures via embedded resources like trust and social capital (Lewis, Ricard, & Klijn, 2018, p. 292). Proponents of collaborative innovation would argue these factors drive innovation within the network. However, there is very little empirical support for this positive relationship in the innovation literature (Lewis et al., 2018), and even less support from quantitative studies of networks. Hence, this thesis will problematize such assumptions and test them using both SNA structural modeling and linear regression methods. The results will be discussed in relation to resource dependence and network governance theories. Further, this paper explores the relationship between the central concept, network embeddedness, with its explanatory variables derived from social network theory, and the occurrence of collaborative innovation, which is analyzed in empirically distinct phases. To accomplish this, I examine the following research question:
How is the collaborative innovation of organizations related to their embeddedness in an interorganizational network?
1.2 The network: The Hague Humanity Hub
Located in the heart of the government center of the City of Peace and Justice, The Hague Humanity Hub (“The Humanity Hub” or “The Hub”) is well-situated to play host to a number of peace, justice, development, and humanitarian organizations. The Hub opened in January 2018, making it approximately a year and a half old at the time this research was conducted. It occupies three spacious floors in a building granted by the Municipality of The Hague. At first glance, The Hub is a modern yet cozy co-working space equipped with flexible workplaces, lounge areas, plenty of natural light,
and an open floor plan. A few of the offices it does contain are finely lined with glass walls that promote an atmosphere of inclusion while ensuring some degree of privacy.
This Hub is more than a co-working space, however. It is a growing community of nearly eighty organizations of many shapes and sizes from the public, private, and ‘third’ sectors. Members have included various levels of government organizations, international NGOs, local and international non-profits, private businesses and startups, legal organizations, academic and research institutions, and other innovation platforms and incubators. Among other things, membership at The Hub grants access to a diverse and inclusive organizational community focused upon what they dub the ‘co-creation’ of innovative projects (The Hague Humanity Hub [LinkedIn], 2019). It is a community designed to assist in fostering and developing innovations in peace, justice, development, and humanitarian work, increase both local and global reach and impact through knowledge exchange and collaboration, and provide platforms and opportunities to promote and support this mission (The Hague Humanity Hub, 2019a). According to one prominent member, one consequence of this within The Hub community is the prevalence of conversations focused on those who stand to benefit from organizations’ work rather than revenues (The Hague Humanity Hub, 2019b).
It is from this formal platform provided by The Hub and the varied relational content of these organizations that the indicators which generate network embeddedness are derived.
1.3 Academic and social relevance
Academic contribution
The literature on collaborative innovation demands a closer work on what works in practice (Torfing, 2019). As such, this thesis contributes to the extant literature on collaborative innovation in public administration both theoretically and methodologically.
First, in response to the call made by Torfing (2019), the theoretical approach taken in this paper broadens the range of theories relevant to collaborative innovation literature. It does so by bringing social network theory into the discussion of collaborative innovation by critically analyzing network structure and the relations that give rise to it as antecedents of collaborative innovation. In public administration, collaborative innovation theory relies on the assumption that collaboration within networks is critical for public sector innovation (Torfing, 2019; Wegrich, 2019). Network collaboration is therefore assumed to be a driver for innovation at multiple levels of analysis and across sectors. In the literature, this is often attributed to another antecedent of innovation: the amount or level of interaction and communication between actors across organizational boundaries (Lewis, et al., 2018). Such ancillary behavior is implicit in any discussion of networks or collaboration, and as such, collaborative innovation. The potential contribution of social network theory to collaborative innovation theory is that a social network analytical approach assumes it is principally the
relationships created by interaction between organizations and their resulting social structures in a network that are critical for explaining organizational outcomes (Wasserman & Faust, 1994).
Yet, no studies in the collaborative innovation literature have explored whether or not the structural relationships in the network that emerge from these ancillary or networking behaviors, which require interorganizational exchange and interaction, provide any evidence of a relationship with collaborative innovation. This is the primary theoretical contribution of this thesis: to see if and how the structures that emerge from the patterns of interactions between and among organizations in a network can provide evidence of a relationship between the content of relations and the occurrence of collaborative innovation in its many phases. The innovation literature states that networks facilitate innovation though interpersonal communication by fostering the growth of social capital and trust within the network (Lewis et al., 2018). So far, however, there has been no theory development or testing of specific hypotheses in networks to see how their structural properties, like network embeddedness, are shaped by interpersonal communication and can influence the capacity of organizations to innovate through collaborative means. Moreover, integrating the literature on social networks and innovation fulfills a need to bridge these two avenues of research in a way that makes innovation scholarship visible in practice and therefore, more relevant to society (Bekkers & Tummers, 2018).
Further, the division of innovation into theoretically and empirically distinct phases in this thesis also aims to provide evidence for or against Torfing’s assertion that collaboration “enables the integration of ideas into proper solutions” (2019, p. 3). In order to obtain this empirical evidence, the process of innovation must be evaluated in multiple phases (steps or stages), rather than as a uniform process or outcome. The theoretical grounding for these phases, which will be elaborated on in Chapter 2, borrows from Meijer’s (2014) non-linear stage model. The added value of this approach lies in its empirical distinction between ideas that remain in the abstract pool of potential solutions, and which ideas are put into motion or practice.
Second, this research contributes to the public innovation literature methodologically. Related to the topic of collaborative innovation, there has so far been a predominance of theoretical research and qualitative case studies in the literature. While The Hub’s ego network was selected for its relatively self-contained ‘case-like’ quality, this thesis is designed to fill a gap in the literature where there has been an absence of large-N empirical research. Thus, the quantitative methods employed here, regression analysis of survey responses and analysis of structural network models, respond to the call made in De Vries, Bekkers, and Tummers’s (2016) meta-analysis for greater methodological variety. Specifically, the antecedents and drivers of innovation adoption are currently supported primarily by case-study evidence, and thus, would benefit from further clarification using quantitative methods and more generalizable models (Walker, 2006). In addition, De Vries et al. (2016) highlight the need for more cross-national studies and cross-sectoral studies. A quantitative analysis of activity in The Hub’s network satisfies both these criteria, given their focus on membership that is defined by
their mission and interorganizational collaboration rather than organization type, nationality, or sector. Moreover, this thesis represents a shift away from case studies towards more diversity in research designs and their corresponding methods.
Social relevance
The academic contribution of this study indicates an overlap with its relevance for society. As Coleman (1990) argues, analyzing social systems gives us the ability to critically analyze components and processes internal to the system. “An explanation based on internal analysis of a system behavior in terms of actions and orientations of lower-level units is likely to be more stable and general than an explanation which remains at the system level” (Coleman, 1990, p. 3). When applied, this theory tells us that understanding relationships between lower level units is necessary for understanding their outcomes, including the innovative performance of individual organizations highlighted in this study. Practically speaking, a shift in governance paradigms toward increasingly networked service and innovation models means that organizations are looking beyond themselves for to generate and implement innovative ideas. As stated previously, innovation hubs and their formal and informal networks are an increasingly more common institutional manifestation of this shift.
Therefore, it is important that we are able to critically assess whether or not collaborative paradigms – underpinned by collaborative strategies and their corresponding institutional forms – are actually effective for achieving innovation goals. In other words, do collaborative strategies actually help organizations innovate, when their normative perceptions as ‘good practice’ are challenged? Further, what is innovation, and can organizations determine if it is desirable without critically measuring its success or lack thereof? By seeking empirical evidence concerning the relationship between relational content of network interaction and collaborative innovation, this thesis addresses the first of these questions. By defining and measuring innovation in phases, it addresses the second. Thus, its practical implications are tied to organizations’ innovation capacity, strategies, and desirability as a whole. Finally, according to Coleman’s logic, systematically testing this relationship between collaboration and innovation via internal analysis of system relations, should provide a “more stable and general” (1990, p. 3) explanation of this phenomenon than studying the network at the system level. So, when looking for evidence of a relationship between network activity on the occurrence of organizational innovation, it is necessary to look at the behavior and outcomes of organizations themselves, rather than the network as a whole.
1.4 Thesis structure
In this chapter, an introduction to collaborative innovation as it pertains to the primary relationship of interest and research question has been provided. The subsequent chapters proceed in the following manner: Chapter 2 will introduce both collaborative innovation theory and the social network
analytical approach. It then suggests infusing collaborative innovation theory with social network theory in order to propose hypotheses about the relationship between interorganizational activity and relations in a network, their resulting network structure, and the occurrence of collaborative innovation. Chapter 3 will provide the details of the applied quantitative research design and methods for data collection, focusing on how the variables are operationalized in the research. Chapter 4 will describe and analyze the results of the integrated regression analyses and SNA. Finally, Chapter 5 will be comprised of an interpretation and discussion of the results and a conclusion outlining limitations of the research, academic and practical implications, and potential avenues for future research.
2. THEORY
This chapter will set a theoretical foundation for the empirical study. It introduces the concept and theory of collaborative innovation, which is the primary outcome variable of interest, and situates it in the broader innovation literature. It outlines the three-phase model used as a framework for understanding the dynamic and multi-stage process of innovation. Next, I introduce social network theory and the network analytical approach to discuss the lower level relational concepts that give rise network embeddedness. Then, specifically focusing on theories of network governance, I explain how network embeddedness arises in a NAO-led network. Using this, I argue why infusing collaborative innovation theory with social network theory and analysis can broaden our understanding of how network structure informs the process of collaborative innovation in an interorganizational network. Lastly, theoretical arguments are used to propose and support hypotheses that relate relational content and network embeddedness to the process of collaborative innovation.
2.1 Collaborative innovation in the public innovation literature
In an increasingly uncertain world, innovation has become the oft sought-after key to providing better public services (Van der Wal, 2017). According to Van der Wal (2017), there are at least four global ‘megatrends’ responsible for the public sector push toward innovation1. Innovation may be targeted at improving outcomes in line with public values, such as efficiency or responsiveness (Walker, 2006, p. 311). On the other hand, public sector innovation can have to do more with normative responses to policy climates pre-determined by central governments (Newman, Raine, & Skelcher, 2001). Regardless, to meet these demands, more organizations are turning toward cross-sectoral, co-creative, or collaborative networked approaches to increase their innovative capacity (Bekkers et al., 2011; Van
1 These include: increasing global connectedness in the era of social networking and big data, demands from service-users and citizens for increased transparency, public debt and pressure to do ‘more with less’, and environmental resource stress caused by climate change (Van der Wal, 2017, p. 139).
der Wal, 2017; Torfing, 2019; Torfing et al., 2019). Hence, this paper addresses innovation outcomes of organizations in collaborative networks.
To define the concept of collaborative innovation as it will be operationalized in this research, innovation itself must be understood. The term ‘innovation’ is conceptually broad enough to encompass a very wide range of definitions (Newman, 2001; Walker, 2006) and heavily debated due to its normative underpinnings (Meijer, 2014, p. 201). Most scholars agree that a definition must include both the generation of new ideas and their implementation (Damanpour, 1987; Walker, 2006; Van der Wal, 2017; Lewis et al., 2018). The logic behind this is straightforward: new ideas cannot influence performance unless they are actually used (Damanpour, 1987, p. 676; Walker, 2006, p. 313; Walker, 2014, p. 23). Innovation entails discontinuity from past or standard practices and solutions (Bekkers & Tummers, 2018). It involves a “step change” (Torfing, 2019, p. 1) that questions and transforms the way things are typically done (Lewis et al., 2018). This discontinuity has been characterized by some authors as necessarily radical, while others accommodate incremental change (Bekkers & Tummers, 2018; Hartley et al., 2013). Regardless of whether innovation is a “frame breaking” or “frame bending” process and outcome of organizational change (Van der Voet, Groeneveld, & Kuipers, 2014, p. 182), it is distinguished by a break from contemporary practices or products, and therefore, differs from continuous improvement (Lewis et al., 2018; Torfing, 2019, p. 1).
De Vries et al.’s (2016) meta-analysis uncovered that most articles on innovation in the public sector do not actually define innovation. Those that do most often use a general definition adapted from Rogers (2003). To maintain theoretical continuity with these authors, this thesis uses the following two-part definition from Walker (2006) as its point of departure for the discussion of collaborative innovation. Innovation is “a process through which new ideas, objects and practices are created, developed or reinvented and which are new and novel to the unit of adoption.” Additionally, “implementation, or actual use of an idea, has to occur in order to turn a new idea into an innovation” (Walker, 2006, p. 313).
Public sector innovation literature is ripe with typologies designed to order and conceptualize its content and processes. Exploring some of the types helps to understand many dimensions of collaborative innovation. Damanpour (1987) emphasizes the importance of distinguishing between and among various innovation types in order to understand why and how organizations adopt some and not others (pp. 675-676). As such, the emergence of collaborative innovation in the literature is tied to many interrelated types. Innovation can be service or product-oriented, by providing new goods or services to new users, old goods or services to new users, or new goods or services to old users (Hartley, 2005; Walker, 2006, p. 313). Innovations can also be process-oriented if they affect “how a service is rendered”, either administratively or technologically, impacting rules and procedures (Hartley, 2005; Walker, 2006; Walker, 2014, p. 23). Further, process innovations can occur both within and outside organizational boundaries (De Vries et al., 2016). When the process
crosses these boundaries, innovations are often considered ancillary because of the interdependencies between product or service providers and other environmental actors. Thus, ancillary innovation outcomes depend on such boundary-spanning collaboration (Walker, 2006, p. 314). The concept of governance innovations applies this idea specifically to how service-providers organize to generate and implement public goods or services. Governance innovations often result in new organizational forms and processes designed to remedy social problems (Moore & Hartley, 2008; De Vries et al., 2016) by relying on cross-sectoral partnerships and accountability sharing across organizational boundaries (Van der Wal, 2017, p. 142). Furthermore, layering these types sets a broad conceptual foundation for understanding the multidimensionality of collaborative innovation.
Definitions of collaborative innovation tend to draw from these various dimensions. Hartley et al. (2013) state, “A collaborative approach to innovation highlights the role of multiactor engagement in informing the understanding of the problem to be addressed, as well as in creating and implementing innovation and garnering support and ownership of the problem and the innovation” (p. 821). Thus, the authors emphasize the capacity of collaborative actors to provide useful input for innovative solutions because social problems are shared phenomena, justifying the collaborative process. They also distinguish between idea generation and implementation, indicating that innovation unfolds in various phases of a process. Alternatively, Wegrich (2019), “understands collaborative innovation as a governing arrangement where one or more public organizations engage other state or nonstate stakeholders in a collective, consensus-oriented, and deliberate decision-making process with the goal to design and implement new, creative solutions to current governance challenges” (p. 12). This definition also distinguishes between the “design” versus “implementation” stage of innovation. However, Wegrich places a stronger emphasis on the creativity of solution making that is democratically sourced from diverse stakeholders via ancillary behaviors. This definition aligns very closely with the concept of governance innovations (Hartley, 2005; Hartley et al., 2013). Further, it assumes that actors and stakeholders outside the boundaries of public service organizations are motivated to collaborate (Torfing, 2019, p. 2) and “should be activated” (Bekkers & Tummers, 2018) to contribute to public value creation through innovation. From this perspective, collaboration is the mechanism that facilitates innovation (Wegrich, 2019, p. 14).
Phase models of innovation processes
It is clear in these definitions that innovation, collaborative or otherwise, is not a unitary or uniform process. It involves multiple phases, steps, and decisions. It is widely considered that innovation is a complex, nonlinear process, whereby steps and stages are often out of prescribed order, repeated, overlapping, or skipped (Meijer, 2014; Van der Wal, 2017, p. 146). Some models are cyclical (e.g. Hartley, et al., 2013), while others depict analytical phases (e.g. Meijer, 2014 in Van der Wal, 2017 p. 146). Therefore, strict conceptual and empirical distinctions are not one-size-fits-all approaches (Meijer, 2014). Yet, to assess how collaboration may facilitate innovation, it is necessary to
distinguish between these somehow, both theoretically and empirically. Phase models do not do justice to the “chaotic” process of innovation; however, they are useful for analyzing the different dynamics (Damanpour, 1991), barriers, and roles of individuals that emerge throughout the process and over time (Meijer, 2014). Torfing (2019) argues that collaboration facilitates innovation at every stage, not just idea generation, including: idea selection, prototype testing, risk and benefit assessment, implementation, resource mobilization, and large-scale diffusion (p. 3). Nevertheless, whether collaboration is a preferable mechanism for realizing the transformation of innovative ideas into solutions (Torfing, 2019, p. 3) warrants further empirical testing. This can be done by making empirical distinctions between and among phases of the innovation process that align with their conceptual divisions.
Moreover, increased interaction may lead to increased collaboration, but not necessarily more innovation, when ideas are collaboratively generated but not used in the delivery of products or services, according to Walker’s definition. Empirical distinctions between phases ensure process indicators are of collaborative innovation, not simply collaboration. As such, for the purpose of this research, Meijer’s (2014) analytical phases provide a good starting point for distinguishing among the dynamics and decisions at play in each phase of the process. The following three phases adapted from Meijer’s stage model are used. First, ‘idea generation’ entails the creation of ideas detailing the change in practice. Second, the ‘idea selection for piloting/testing’ phase requires choosing among ideas already generated to develop further, by directing limited resources in its direction and/or testing its workability. Third, ‘implementation in product or service delivery’ encompasses any use of the innovative idea in product or service delivery, including the two remaining stages of the model, scaling-up and diffusion (Meijer, 2014, pp. 201-202).
2.2 Collaborative Innovation: Why or why not?
Rooted in the argument that collaboration can benefit innovation at every stage, there is theoretical and empirical evidence to support the notion that collaborative innovation has unique advantages when compared to hierarchical or competitive strategies, though these types are not necessarily mutually exclusive (Torfing, 2019, p. 4). This evidence relies on certain assumptions about the demand for, nature of, and barriers to innovation (Wegrich, 2019). Innovation in a bureaucratic or hierarchical context, which emphasizes bureaucratic reform (Hartley, et al., 2013), is inhibited by the typical problems caused by red tape and the inflexibility of large organizations. Competitive market-like innovation strategies, such as those emphasized by NPM, create incentives for individual organizations’ performance as opposed to collective development (Wynen et al., 2014). Further, competitive strategies may enhance pressure for innovation, but lack the follow-through to have innovations developed or implemented (Hartley et al., 2013, p. 823). Moreover, bureaucratic and
competitive alternatives emphasize closed innovation models, and by so doing, eliminate the potential benefits of “organizational, intersectoral, and open innovation” (Hartley et al. 2013).
At the forefront of the normative call for collaborative innovation in practice and as a research agenda, Torfing employs a resource-based understanding of the merits and pitfalls of such strategies. He argues that both external and internal participants should be included in the innovation process because they share knowledge and experiences of both the problems and their potential solutions. Consequently, all parties should share the responsibility to ensure problems are understood, and solutions are implementable (Torfing, 2019, p. 4). These ideas are rooted in the work of Roberts (2000) which frames the benefits of collaborative strategies in terms of resource exchange across boundaries:
In contrast to hierarchical and competitive strategies, collaborative strategies facilitate the exchange of knowledge, competences and ideas between relevant and affected actors and thus stimulate processes of mutual learning that may improve the understanding of the problem or challenge at hand and extend the range of creative ideas about how to solve it (Roberts 2000 in Torfing, 2019, p.3).
Torfing (2019, p. 4) also argues collaborative strategies are superior to alternatives, because “sustained dialogue” with external collaborators is a valuable resource. This is partially due to the fact that continued dialogue helps bring innovative ideas to fruition as solutions. This view underlines the importance of interaction in the previously outlined phases of the innovation process after the collaborative generation of ideas (Meijer, 2014; Van der Wal, 2017, p. 147). Moreover, beyond using external resources in the creative process of idea generation, improved implementation of solutions can occur when “resources are mobilized, exchanged, and coordinated and joint ownership is created through participation and dialogue” (Hartley, et al., 2013, p. 826). The importance of this dialogue over time is underscored by some authors who found that long-term networks of diverse actors produce more legitimate and implementable solutions than governments do by themselves (Dente, Bobbio, & Spada, 2005; Steelman, 2010 in Van der Wal, 2017). Consequently, it is difficult to discuss collaborative innovation at any phase without understanding how networks facilitate specific resource access.
Despite the supporting normative and positive theory, collaborative innovation is not viewed as a perfect means of creating public value, nor an infallible solution to governance issues. There are many reasons to maintain some degree of skepticism identifiable in the literature. Two, outlined here, relate to the shared resources collaborative strategies are meant to harness. First, when participating actors and organizations commandeer shared resources and steer them toward their self-interest, a tension between private individual and public value creation emerges (Hartley et al. 2013; Van der Wal, 2017). Second, prioritizing innovation above other organizational goals risks not fully utilizing existing expertise. As such, innovation should only be pursued when ‘appropriate’ and not viewed as normatively ‘good’ (Osborne & Brown, 2011, p. 1347). These examples illustrate instances in which collaboration can constrain, for some stakeholders, the pool of resources it is meant to expand.
Moreover, organizations engaging in resource-driven collaborative innovation strategies must take context, including environmental actors, into account. This is also true in networks. Formal governance networks designed to expand access to material and knowledge resources are also underpinned by critical informal social relationships (Lewis et al., 2018). Lewis et al. (2018) argue the public sector is increasingly catching on to the idea that social networks rely on informal communication among internal actors to generate “embedded resources”, such as interpersonal relationships (p. 292). Provided the barriers to innovation in the public sector, the development of softer resources within the network, for example, trust and social capital, are increasingly considered more important for innovation (Lewis, Considine, & Alexander, 2011). In combination, these new formal structures and the interorganizational relationships that support them constitute the social network, which in turn, is comprised of all actors in the network and the ties between them (Wasserman & Faust, 1994).
It is a short conceptual leap from thinking about governance networks in terms of resources to their structural relationships. In the private sector, networks have long been considered important for facilitating innovation (Lewis et al., 2018). However, some public innovation scholars argue that structural network characteristics matter for levels of collaboration and innovation as a result. According to Newman et al. (2001), “weak networks” and “strong organizational boundaries” are barriers to innovation (p. 66). Dente (2005) argues “complexity” (a measure of actors’ diversity) and “tightness” (the density of connections among them) of a network leads to greater policy innovation. More specific to this study, focusing on individual actors, or nodes, within the network rather than the network itself, sheds light on questions of how an actor’s embeddedness, or relative prominence, in a network can facilitate outcomes like innovation (Provan & Kenis, 2008, p. 232).
In sum, even when taking a more critical perspective of collaborative innovation, analyzing structural network characteristics, or the relationships between organizations within a network, has potential to inform us how they innovate. For example, when a diverse set of actors collaborate for innovation purposes, they also risk the emergence of a single dominant worldview, which could defeat the purpose of drawing on diverse perspectives and knowledge resources (Wegrich, 2019, p. 17). The composition of a network, how it is governed, who the central actors are, and the resulting links between certain actors have the potential to help explain how resources move across organizational boundaries or how dominant ideas emerge (Wasserman & Faust, 1994).
Scholarly work on collaborative innovation is new enough that it demands a closer look at whether or not collaborative strategies are effective for innovation in practice (Torfing, 2019). As discussed above, the concept of collaborative innovation is fundamentally tied to relationships, both formal and informal, that underpin intra-network resource exchange for innovation. Social network theory has been overlooked in the discussion of collaborative innovation, and yet, it offers an intuitive analytical approach to investigate the relationship between collaborative strategies and innovation. It can describe or explain how organizations relate to each other in diverse roles, interact, and exchange
resources in a network when they collaborate, and shed light on the possible effects of these relations and behaviors on innovation outcomes.
2.3 Social network theory and analysis
Despite the apparent clarity of the various definitions from the literature provided in the previous section, innovation—and by extension, collaborative innovation—is conceptually and categorically broad, making it difficult to measure. Van der Wal (2017) argues outcomes are generally challenging to quantify in public sector contexts precisely because they are multi-value. As the theory on collaborative innovation indicates, public sector services are increasingly cross-sectoral, considered more multi-stakeholder, and therefore, multi-value (Bekkers & Tummers, 2018). Qualitative outcome assessment strategies in public contexts tend to be incomparable because they are defined by and specific to particular service-providers or organization types (Van der Wal, 2017). This suggests this might also be the case in mixed sector network contexts, which have the potential to be even more multi-value. Counter to the perception that innovation outcomes are not measurable, however, this thesis argues that, given certain theoretical assumptions, it is possible to quantitatively assess innovation outcomes using analytical approaches from theories of social networks. Moreover, “in social science, the structural approach that is based on the study of interaction among social actors is called social network analysis” (Freeman, 2004, p. 2). However, as the social network analytical approach suggests, to assess or quantify organizational outcomes, we must look directly at relational characteristics of actors within the network. This approach assumes it is principally the relations created by “substantive connections” between organizational actors and their resulting structural characteristics that are critical for explaining their outcomes (Wasserman & Faust, 1994). In other words, the patterning of actors’ social ties and their embeddedness within them matters for what they do and how they do it (Freeman, 2004, p. 2).
The primary goal of the network analytical approach can be to use configurations of relationships (links) between actors (nodes) to explain the outcomes of network processes (Provan & Kenis, 2008, p. 232). Thus, structural models are used to test theories about processes and structures that are relational rather than characteristics inherent to organizations, or other network actors, per se (Wasserman & Faust, 1994, p.4-5). This strategy can be used to analyze collaborative innovation processes because of the open systems perspective that underpins this particular type of innovation. It assumes that the structures resulting from relations between actors are critical for the diffusion of information as well as the flow of resources, subsequently contributing to innovation outcomes (Wasserman & Faust, 1994). At the organizational level of analysis, how embedded an organization is within the complete network or smaller subunits of the network, has some measure of explanatory power for organizational outcomes (Provan & Kenis, 2008, p. 232).
This explanatory power of relationships is why SNA is used to study system outcomes at various levels. For example, it can be used to analyze the performance of individual organizations, other actors, or subgroups of actors within the network, as well as the performance of the network as a whole (Wasserman & Faust, 1994). As explained in Coleman’s Foundations of Social Theory (1990), system conditions are comprised of interrelated individual actor beliefs, preferences, and strategies that generate individual behaviors and outcomes at the micro-level. In the aggregate, these individual behaviors and outcomes produce system outcomes. Thus, the relationship between system conditions and system outcomes at the macro-level is mediated by these micro-level processes (Coleman, 1990).
In theory, a network analytical approach extends Coleman’s reasoning to the content, or micro-processes, of relations between actors in a network system. In the approach taken here, the relational contents are measured on a pair of actors and represent the “substantive connections”— the content of relational ties—between actors, namely how they engage with each other (Wasserman & Faust, 1994). The more substantive connections one actor has with other actors in the network, the more prominent or embedded in the network said actor is (Wasserman & Faust, 1994). While the quantity and direction of relations (a.k.a. links or ties) do inform embeddedness, pairs or subgroups of actors can be linked in a variety of ways depending on the empirical nature of their relationships. Moreover, “the content of relational ties between actors is fundamentally different in different networks” (Bodin & Crona, 2009, p. 367). In other words, the content of relationships matters for measuring embeddedness, but embeddedness is not itself a measure of relational content, but rather a measure of prominence. Thus, to test for network embeddedness, I must first enumerate the concepts from the theory on relational content that will be focused upon in this study. First, are formal roles, which predefine in some official capacity how actors can or will relate to each other via their interactions or exchanges. Second, and perhaps most obvious of these concepts, is interaction. Interaction elicits the physical presence of multiple actors in the same place at the same time and the resulting interface between actors. Third, is the transfer of material and/or non-material resources, which involves the exchange of material or communication and information resources (Wasserman & Faust, 1994, pp. 37-38).
Network governance and the NAO model
However, in order to understand how these various types of connections influence micro-level outcomes (e.g. organizational innovation), we have to understand how certain patterns of these relationships may facilitate or inhibit them. One way of analyzing lower (organizational) units to assess the outcomes at this level is with theories of network governance. To overcome the potential threats to collective development posed by the self-interest of individual actors (Hartley et al., 2013), networks rely on various forms of internal governance (Provan & Kenis, 2008). When networks are
goal-directed2 and their identity is shaped by those goals, a discussion of governance is necessary to understand how structure informs and affects their functioning (Provan & Kenis, 2008). An organization is considered goal-directed when it is expressly concerned with innovation outcomes and the identity of and membership in the network are dependent upon this goal (Provan and Kenis, 2008, p. 231). Provan and Kenis’s (2008) conceptualization of NAO-led networks proposes a model wherein the NAO acts as a network broker to direct organizations toward network goals. It does this to boost legitimacy in the absence of high levels of implicit trust between organizations, but in the presence of relatively high levels of goal consensus (Provan & Kenis, 2008, p. 237). In this role, the NAO is also positioned to acquire and oversee the effective and efficient use of resources.
2.4 Relational content and network embeddedness
In this research, measures of different types of relational content are expected to influence how embedded organizations are in the network. Using concepts of relational content that determine embeddedness, I propose three hypotheses concerning how the content of relations between actors, determined by organizational interaction in a network, can inform innovation.
Formal roles
Lorraine and White (1971, in Wasserman & Faust, 1994) state that formal roles should be analyzed “to understand the interrelations among relations within concrete social groups” (p. 442). Their claim is based upon the idea that formal roles “exhibit interrelations which are more regular or transparent” because they qualify relationships at a high level of abstraction (p. 442). Two important points can be taken from Lorraine and White’s reasoning, which make the use of formal roles in measuring embeddedness appropriate for this study. First, networks defined by formal membership can be considered a type of “concrete social group”. As has already been stated in this chapter, governance networks are underpinned by formal relations as well as informal ones (Lewis et al., 2018). This concept of formal roles favors the idea that formal relationships, with their corresponding resources (e.g. access to other members, material resources, legitimacy), affect how informal relationships and their accompanying resources (e.g. trust, social capital) develop. More broadly, these formal ties have implications for other types of social ties, or “secondary reciprocals” and affect how they develop (Freeman, 2004, p. 49). Second, are the regularity and transparency of this type of interrelation, which are defined by more concrete regulatory mechanisms (e.g. legal and contractual) than abstract trust relations or social capital are able to provide. Moreover, formal roles must be considered alongside informal ones in analyzing social network structures.
2 As, Moreno (1937, in Freeman, 2004) states when discussing the origin of sociometry: “ [a] decisive step was the consideration of the criterion (a need, a value, an aim, etc.) around which a particular structure develops” (p. 39). According to Freeman, the sociometric approach contained the all of the key “defining properties” of SNA, including a discussion of goal-directed networks.
The concept of formal roles is, however, quite broad and, thus, both versatile and limited in its ambiguity. Wasserman and Faust (1994) primarily discuss formal roles in terms of power and authority relations, in other words, how actors relate to each other in the instance that some actor(s) have some degree of formal authority over other actor(s) (e.g. student/teacher, doctor/patient relationships). However, the authors make it quite clear such roles based upon overt power or authority relations are only examples of formal roles, leaving its definition somewhat vague. In this research, I maintain that the concept of formal roles is indeed broad enough to be applied with some flexibility to other instances in which formal relationships are integral to “interrelations among relations” (Wasserman & Faust, 1994, p. 442) in a network structure. Moreover, it is my understanding that formal roles are utilized in this thesis in keeping with the authors’ conceptual intent. The view employed here, is that formal roles are delineated or established by the formal agreements that set the legal, contractual, partnership, etc. basis for formal governance networks. The benefit of analyzing formal roles from a global network perspective is that they describe “an association among relations that hold[s] for the entire group” (Wasserman & Faust, 1994, p. 425).
Based on this theory of formal roles, it is expected that a prominent formal role in the network will mean more opportunity for interaction and “substantive connections” with other actors in the network by virtue of their increased access to the network resources. Thus, they will have more opportunity to transfer material and knowledge resources and facilitate interorganizational learning, as well as develop more informal communication and trust-based relations (Lewis et al. 2018). A more prominent formal role may also put them in more frequent and closer proximity to the NAO, who controls the flow and distribution of many material and knowledge resources and directs them as deemed appropriate toward network-driven organizational innovation goals (Provan & Kenis, 2008). Such proximity also has the ability to offer organizations with more prominent formal roles increased legitimacy and visibility status within the network, leading them to accumulate higher levels of implicit trust and social capital (Lewis et al., 2018). Moreover, proximity may also create for efficient lines of communication between and among organizations with prominent formal roles and the NAO, which could aid innovation through sustained dialogue, mobilization of resources, and integration of innovative ideas into solutions (Torfing, 2019).
From this reasoning, the first hypothesis is derived:
H1: Having a more prominent formal role in a network will have a positive relationship with the occurrence of collaborative innovation.
Interaction
Interaction entails the presence of participating actors “in the same place at the same time” (Wasserman & Faust, 1994, p. 38). Organizations that exhibit higher levels of interactive participation in the network will facilitate the growth of more “embedded resources”, such as higher levels of trust and social capital, as a result of more informal and efficient communication (Lewis et al., 2018).
Interaction outside ordinary work settings rely on this type of social networking behavior, because it depends on less formal mechanisms and face-to-face interaction that can facilitate trust. Partnerships that develop between interacting organizations who seek complementary resources from each other are able to generate ideas for innovative solutions and take them through the process toward implementation with fewer inhibiting formal procedures (Hartley et al., 2013, p. 821). Continued interaction, resulting in “sustained dialogue” (Torfing, 2019) over time will grow these “embedded resources” and are more likely to take innovation further than the idea generation phase in the innovation process. Additionally, when individuals of different organizations and skillsets meet in networking settings, trust relations facilitate future interorganizational and trans-boundary learning. Interaction with a NAO or prominent members in the network can lend legitimacy to the members with whom they interact. This is because the NAO, and those close to it, are often widely perceived as legitimate by the network it administrates according to network governance theory (Provan & Kenis, 2008). Further, with whom and how often organizations interact will have an effect on their collaborative innovation outcomes. Thus, the expectation is:
H2: Higher levels of interaction will have a positive relationship with the occurrence of collaborative innovation.
Transfer of material and non-material resources
Communication platforms that connect the NAO to other organizations in the network facilitate the role of NAO as a broker to link members with complementary interests (Provan & Kenis, 2008). Communication from the central node in the network is considered one efficient way of reaching more peripheral actors. The NAO’s brokering role enables a more efficient exchange of complementary resources to develop those complementary interests, limiting interference from other middle-man organizations. Increased access to external material and knowledge resources generated by higher levels of resource transfer between organizations will help improve implementation of solutions beyond the idea generation stage (Torfing, 2019, p. 3). This is because “participation and dialogue” help actors create “joint ownership” of innovative ideas (Hartley, et al., 2013, p. 826). Further, non-material resource exchange may help mobilize further vital material resources, such as financial support or human capital, toward implementation (Bekkers & Tummers, 2018). On the other hand, communication from the NAO could provide organizations with the means to resolve conflict and engage in “collective and mutually supportive action” (Provan & Kenis, 2008 p. 231) toward innovation goals. Thus, I propose the third hypothesis:
H3: Higher levels of resource (material and non-material) exchange will have a positive relationship with the occurrence of collaborative innovation.
Each of these hypotheses presupposes that the more active an organizational actor is in network participation, the more embedded they are in its structure. Network embeddedness is then, the central
concept this thesis seeks to relate to innovation outcomes. It provides us with the three explanatory variables used to hypothesize and test the relationship between network embeddedness and innovation outcomes. The next chapter will provide and explain the methodological choices made in this research to test these hypotheses and analyze the accompanying data collected from the network.
3. RESEARCH DESIGN AND DATA COLLECTION
This chapter discusses the research methodology undertaken for this thesis. First, it will describe and explain the research design and selection of cases. Second, it will discuss the methods employed for data collection. Third, it will describe how the variables that were tested were operationalized. Finally, I will assess the research in terms of reliability as well as internal and external validity.
3.1 Research design and case selection
The empirical study presented here employs an X-centered, large-N research design. The objective of this research is to investigate the relationship between organizations’ embeddedness in the network structure (X)—resulting from the relational content of organizations—and collaborative innovation outcomes (Y). Large-N designs strategically allow researchers to bring together quantitative data from a large number of individual cases (i.e. organizations) by testing only a few variables from each case (Toshkov, 2016, p. 201). To depart from the predominance of qualitative and normative research in the collaborative innovation literature, which is primarily supported by case studies (Torfing, 2019), this research will quantitatively test assumptions about the relationship between organizational embeddedness in a network structure and collaborative innovation. More specifically, to accomplish this, it will use statistical regression and social network analytical methods to analyze this relationship. Subsequently, the assumptions about resource dependency and network governance outlined in Chapter 2 will be evaluated against this innovation-driven empirical network of organizations within the peace, justice, development, and humanitarian fields.
A large-N design was an appropriate selection for this research because working with the network’s administrative organization presented an opportunity to potentially reach and test multiple variables on every member organization in their ego network. Simultaneously, this approach fills a gap in the research for large-N studies. Moreover, with a larger sample, it is possible to obtain a better understanding of the association between embeddedness in network structure and the desired outcome, collaborative innovation, than with small-N samples (Toshkov, 2016, p. 205). Associations alone cannot reveal causality, so, this thesis seeks to provide evidence of the primary relationship of interest, but it does not presume to reveal causality of those relationships given the limitations of the study. However, the added benefit of an SNA approach lies in introducing models and previously tested assumptions from social network theory to statistically derived associations between the explanatory and outcome variables (Toshkov, 2016, p. 205). Looking at the patterning of social ties
gives a more comprehensive picture of the relationship of interest by elucidating what empirical phenomena give rise to them and what are the consequences (Freeman 2004, p. 2).
The individual cases that comprise this large-N study are organizations. Each of these organizations is a member of an ego network coordinated and administrated by The Humanity Hub. An ego network is made up of a “focal actor, termed ego” and its alters who have ties to the focal actor, as well as the measurements on the ties that link any actors to each other (Wasserman & Faust, 1994, p. 42). Only members of The Hub’s ego network are investigated in this research study. In other words, no organizations outside The Hub’s formal network membership or indirectly associated were included in the sampling. These cases were selected for several reasons. First, because as previously stated, organizational innovation is a prescribed desired outcome of membership in the network. Although actual reasons for joining such a network are likely more complex for both individual actors and organizations, The Hub actively promotes collaborative innovation in the form of “co-creation” projects (The Hague Humanity Hub [LinkedIn], 2019). Therefore, it is assumed that member organizations are, at least to some degree, goal-oriented toward innovation as an outcome of network collaboration. Second, the cases represent a wide array of organizational types and sectors. Such organizational diversity is tied to assumptions about resource pooling and interorganizational learning in the collaborative innovation literature that will be critically evaluated. Third, I test and analyze all variables at the organizational level because, while interactive or networking behaviors that indicate interorganizational collaboration are carried out by individuals, these individuals do so on behalf of organizational goals. Fourth, pertaining to both the large-N design and organizational level of analysis, enough data could be translated into a partial reconstruction of the network. However, because it was established that the outcome variable should be measured at the organizational level, this model would only be informative about the impact of network structure on organizational innovation if the explanatory variables are also measured at the organizational level.
3.2 Data collection and methods
The data for this research was obtained from a survey that was distributed by the researcher with the assistance of The Hub. In this way the methods for data acquisition support the methodological reasoning behind studying an ego network. The survey was designed and published by the researcher using Qualtrics. It was initially posted on Slack: the internal platform The Hub uses to communicate with its members. In the attempt to gain a high response rate, it was distributed with an email notification to the seventy-eight member organizations who have at least one active email address on the Slack platform. When the initial two postings on Slack did not stimulate a sufficient number of responses, the researcher contacted organizations directly via publicly available individual emails, LinkedIn messages, and general website email addresses and contact forms. Given the nature of the questions on the survey and the fact they aimed at measuring participation within The Hub’s network,
