EXPLORING EMERGENT NETWORKS ACROSS UNDERGRADUATE STEM EDUCATION REFORM NETWORKS: THE STEM REFORM HYDRA By Levi B ritton Shanks A DISSERTATION Submitted to Michigan State University i n partial fulfillment of the requirements for the degree of Higher, Adult, and Lifelong Education Doctor of Philosophy 2020 ii ABSTRACT EXPLORING EMERGENT NETWORKS ACROSS UNDERGRADUATE STEM EDUCATION REFORM NETWORKS: THE STEM REFORM HYDRA By Levi B ritton Shanks Over the last 30 years, formally structured inter - organizational networks have risen in popularity as a strategy for addressing large - scale and complex societal problems. Within higher education, many inter - organizational networks organize to reform undergraduate science, technology, engineering, and mathematics (STEM) education to promote the quantity, quality, and diversity of STEM graduates. While formal networks in undergraduate STEM education reform play a role in officially l inking higher education institutions, literature points to the existence of an unstructured informal network emerging from the connections established by connecting organization in reform. This sequential quan - Qual mixed methods study highlights a small, i nformal inter - organizational network of leaders in formal STEM networks. Implications underscore emergent network roles in the facilitation, maintenance, and sustainability of the formal networks, and larger impacts of this group in undergraduate STEM educ ation reform efforts. Broader impacts speak to a greater role of informal networks in innovation, organizational sense - making, and systemic change, and invites critical constituents in reform efforts to utilize emergent networks more intentionally. iii This dissertation could not have been completed without the care and support of my family . To Laura, your patience and unconditional love has bridged me across this journey . To my favorite writing partner, Ms. Figg , your even - keel , level - headedness , and boisterous purrs ha ve kept me calm during the most chaotic time s . And to Mo, whose early arrival gave me a much - needed boost of motivation to get across the finish line . I love you all. iv TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ........................ vii LIST OF FIGURES ................................ ................................ ................................ ..................... viii KEY TO ABBREVIATIONS ................................ ................................ ................................ ........ ix Chapter 1: Introduction ................................ ................................ ................................ ................... 1 Inter - Organizational Networks ................................ ................................ ................................ . 2 The Wicked Problem: Undergraduate STEM Education Reform ................................ ............ 4 Problem Statement and Research Questions ................................ ................................ ............. 6 Dissertation Organization and Conclusion ................................ ................................ ............... 8 Chapter 2: Literature Review ................................ ................................ ................................ ........ 10 Formal Collaborative Networks in Undergraduate STEM Education Reform ....................... 10 Importance of relationships. ................................ ................................ .............................. 13 Emergent Collaborati ve Networks ................................ ................................ .......................... 15 Theoretical interpretations. ................................ ................................ ............................... 16 Network functions. ................................ ................................ ................................ ............ 18 Network governance and leadership. ................................ ................................ ................ 22 Network structure. ................................ ................................ ................................ ............. 28 Existing Literature Gaps ................................ ................................ ................................ ......... 28 Conclusion ................................ ................................ ................................ .............................. 30 Chapter 3: Research Methods ................................ ................................ ................................ ....... 32 Philosophical Grounding ................................ ................................ ................................ ........ 33 Research Approach ................................ ................................ ................................ ................. 34 Participant selection ................................ ................................ ................................ .......... 37 Quantitative Methods: Social Network Analysis ................................ ................................ .... 39 Data collection ................................ ................................ ................................ .................. 40 Data cleaning ................................ ................................ ................................ .................... 41 Data analysis ................................ ................................ ................................ ..................... 42 Qualitative Methods: Semi - Structured Interviews ................................ ................................ .. 45 Data collection ................................ ................................ ................................ .................. 45 Data preparation ................................ ................................ ................................ ................ 47 Data analysis ................................ ................................ ................................ ..................... 48 Positionality ................................ ................................ ................................ ...................... 49 Conclusion ................................ ................................ ................................ .............................. 50 Chapter 4: Social Network Findings ................................ ................................ ............................. 52 Results ................................ ................................ ................................ ................................ ..... 52 Business network. ................................ ................................ ................................ ............. 55 Problems network. ................................ ................................ ................................ ............ 58 Gave Advice network ................................ ................................ ................................ ....... 61 Centralization and centrality discussion. ................................ ................................ .......... 64 v Inter - Organizational Network Properties ................................ ................................ ................ 70 Dyadic Properties ................................ ................................ ................................ .................... 73 Network Dyadic Properties Discussion ................................ ................................ .................. 77 Complexity in relationships ................................ ................................ .............................. 78 Organizationally focused dissemination ................................ ................................ ........... 79 Individually focused collaborations ................................ ................................ .................. 80 Not significant social identities ................................ ................................ ......................... 81 Dyadic modeling limitations ................................ ................................ ............................. 82 Conclusion ................................ ................................ ................................ .............................. 83 Chapter 5: Interview Findings ................................ ................................ ................................ ...... 84 Participant Profiles ................................ ................................ ................................ .................. 85 Allie. 86 Andrew. ................................ ................................ ................................ ............................. 89 Hilarie. ................................ ................................ ................................ .............................. 94 Jennifer. ................................ ................................ ................................ ............................. 98 Lindsey. ................................ ................................ ................................ ........................... 100 Luke. 103 Cross - Case Analysis: Knowledge ................................ ................................ ......................... 105 Cross - Case Analysis: Knowledge Exchanges ................................ ................................ ...... 109 Mediums of knowledge exchanges ................................ ................................ ................. 113 Cross - Case Analysis: Network Learning ................................ ................................ .............. 115 Consensus building ................................ ................................ ................................ ......... 115 Coalitions. ................................ ................................ ................................ ....................... 117 Network learning ................................ ................................ ................................ ............ 120 Cross - Case Analysis: Innovations ................................ ................................ ........................ 120 USER Emergent Network Function Discussion ................................ ................................ ... 124 Emergent network as an open system ................................ ................................ ............. 125 Emergent Network Governance and Capital Manifestations ................................ ................ 128 Cultural capital ................................ ................................ ................................ ................ 129 Organizational capital. ................................ ................................ ................................ .... 133 Intellectual capital. ................................ ................................ ................................ .......... 134 Capital, Governance, and the Emergent Network ................................ ................................ 135 Discussion ................................ ................................ ................................ ............................. 138 Conclusion ................................ ................................ ................................ ............................ 143 Chapter 6: Data Integration ................................ ................................ ................................ ......... 146 Strand One: Social Network Findings ................................ ................................ .................. 147 Strand Two: Interview Findings ................................ ................................ ........................... 148 Data Integration ................................ ................................ ................................ .................... 149 Sequential Data Integration ................................ ................................ ................................ ... 150 Peripheral actors ................................ ................................ ................................ .............. 151 Central actors ................................ ................................ ................................ .................. 153 Meso - actors ................................ ................................ ................................ ..................... 156 Sequential data integration summary ................................ ................................ .............. 158 Emphasis - Driven Data Integration ................................ ................................ ....................... 159 Emergent network function cycle ................................ ................................ ................... 160 vi Open system organization ................................ ................................ ............................... 161 Conclusion ................................ ................................ ................................ ............................ 163 Chapter 7: Implications and Conclusion ................................ ................................ ..................... 165 How Interconnected are Leaders across Formal Networks? ................................ ................. 165 How do Leaders Engage in Knowledge Diffusi on? ................................ ............................. 167 How do Leaders Engage in Network Learning? ................................ ................................ ... 169 How do Leaders across Formal Networks in USER serve as an Emergent Network? ......... 170 How does this Emergent Network affect Formal Networks in USER? ................................ 173 Theoretical Implications and Contributions ................................ ................................ .......... 174 Capital and embedded networks ................................ ................................ ..................... 175 Neo - institutionalism ................................ ................................ ................................ ........ 176 Methodological contribution. ................................ ................................ .......................... 178 Undergraduate STEM education reform ................................ ................................ ......... 179 Practical Implications and Contributions ................................ ................................ .............. 179 Undergraduate STEM education reform funders. ................................ ........................... 180 Formal network administrators. ................................ ................................ ...................... 181 Reform - minded faculty ................................ ................................ ................................ ... 182 Limitations ................................ ................................ ................................ ............................ 182 Future Research ................................ ................................ ................................ .................... 183 Conclusion ................................ ................................ ................................ ............................ 184 APPENDICES ................................ ................................ ................................ ............................ 186 Appendix A: Social Network S urvey ................................ ................................ ................... 187 Appendix B: Project Overview for Participant ................................ ................................ ..... 194 Appendix C: Sample Interview Protocol ................................ ................................ .............. 196 Appendix D: Variable Codebook ................................ ................................ .......................... 198 REFERENCES ................................ ................................ ................................ ........................... 199 vii LIST OF TABLES Table 1. Unsymmeterized Reciprocity Score for all Networks ................................ ..................... 53 Table 2 . eristics (n=17) ................................ ........................ 54 Table 3. Summary of Respondents Network Affiliations (n=15*) ................................ ................ 54 Table 4 . Network Density and Centralization Scores ................................ ................................ .. 65 Table 5 . Network of Networks Density and Centralization Scores ................................ .............. 71 Table 6 . Business Network Multiple QAP Regression Model ................................ ...................... 74 Table 7 . Problems Network Multiple QAP Regression Model ................................ ..................... 75 Table 8 . Gave Advice Network Multiple QAP Regression Model ................................ ................ 77 Table 9. Knowledge Thematic Breakdown ................................ ................................ ................. 108 Table 10. Knowledge Exchanges Thematic Breakdown ................................ ............................ 114 Table 11. Network Learning Dynamics Breakdown ................................ ................................ .. 119 Table 12. Innovations Thematic Breakdown ................................ ................................ .............. 123 Table 13. Structural Location of Participants in Networks ................................ ....................... 150 Table 14. Appendix D: Social Network Analysis Variable Codebook ..................................... 198 viii LIST OF FIGURES Figure 1 . Quan - QUAL Sequential Explanatory Mixed Methods Design ................................ ..... 36 Figure 2. Spoke a bout Network Business Sociogram ................................ ................................ ... 57 Figure 3. Correspondence Analysis of Business Network ................................ ............................ 58 Figure 4. Discussed Network Problems Sociogram ................................ ................................ ..... 59 Figure 5. Correspondence Analysis of Problems Network ................................ .......................... 61 Figure 6. Gave Advice to Others Sociogram ................................ ................................ ................ 62 Figure 7. Correspondence Analysis of Advice - Giving Network ................................ .................. 64 Figure 8. N etwork of Networks Sociogram ................................ ................................ .................. 72 Figure 9. Emergent Network Zone ................................ ................................ ............................. 139 Figure 10. Emergent Network Activation ................................ ................................ ................... 14 0 Figure 11. Consensus Building and ENZ Strategizing ................................ ............................... 141 Figure 12. Emergent Network F unction Cycle ................................ ................................ ........... 142 ix KEY TO ABBREVIATIONS ASCN: Accelerating Systemic Change Network AAC&U: Association of American Colleges and Universities AAU: Association of American Universities APLU: Association of Public Land - Grant Universities BOSE: Board on Science Education CoT: Communities of Transformation DBER: Discipline - Based Education Reform Emergent Network: Emergent Collaborative Inter - Organizational Network FIPSE: Fund for the Improvement of Postsecondary Education HHMI: Howard Hughes Medical Institute Network: Collaborative Inter - Organizational Network Formal Network: Formal Collaborative Inter - Organizationa l Network NASEM: National Academies Roundtable for Systemic Change in Undergraduate STEM Education Reform NSF: National Science Foundation PLC: Professional Learning Communities QAP: Quadratic Assignments Procedure Sloan: Alfred P. Sloan Foundati on SOTL: Scholarship on Teaching and Learning STEM: Science, Technology, Engineering, and Mathematics USER: Undergraduate Science, Technology, Engineering, and Mathematics Education Reform 1 Chapter 1: Introduction In the last 40 years, higher education institutions have become increasingly beset by complex and systemic challenges. Although higher education institutions have always mirrored societal problems (Thelin, 2011), the last four decades saw growing disillusi onment with higher education as a public good (Marginson, 2011; Singh - Kagisano, 2001; Tilak, 2008). Dramatic increases in public scrutiny through reduced appropriations (Kezar, Chambers, & Burkhardt, 2015), greater accountability (Friedman, 2007; Hillman, Tandberg, & Gross, 2014; Tandberg & Hillman, 2014), and increased perceptions of education as a private investment have progressively plagued post - secondary education since the 1990s. Compounding these public concerns were increasing distrust and political a decrease in higher education confidence by almost 10% between 2015 and 2018 (Jones, 2018). Each of these challenges required higher education institutions to respond with organizational changes (Thelin, 2011), increases in tuition (Tandberg & Hillman, 2014), and even public Cou ntless authors have identified the societal problems plaguing U.S. higher education as Hutchings & Quinney, 2015; Krause, 2012; Ramley, 2014; Watson, 2000 cked - sector challenges to which single organizations are poorly suited to address (Rittel & Weber, 1978; Weber & Khademian, 2008). Mentioned frequently with research on climate change (Hulme, 2009), U.S. healthcare reform (Council of Ac countable Physician Practices, n.d. ), public health ( Peterson, 2016), and higher education (Armstrong, 2017; Krause, 2012; Masten, 2016; Ramley, 2014), wicked problems require expertise from disparate fields to 2 cooperatively construct feasible ways to addr and often implicit challenges (Conteh, 2013; Isett et al., 2011; Rittel and Weber, 1973). This research explores one strategy employed by colleges and universities in higher education to address the wicke d problem of undergraduate science, technology, engineering, and mathematics (STEM) education reform. Inter - Organizational Networks In attending to wicked problems, institutions in varying domains rely heavily on the role of inter - organizational networks (networks). Although there is considerable discussion about definitions, these networks are generally defined as a cross - sectional group of individuals who link organizations through their relationships and share information regarding a common problem (Br yson, Crosby, & Stone, 2006; Provan, Fish, & Syndow, 2007; Weber & Khademian, 2008). Networks are a response to market failures in the face of complexity, de - legitimization, and lack of innovation posed by wicked problems (Isett et al., 2011). Networks are often seen by institutions as viable alternatives when both bureaucracies and markets fail (McGuire 2006; The benefits of networks are commonly cited as cross - fertilization among experts, organizations, and domains (Berry et al., 2004). Bringing together different perspectives provides the platform to collaboratively address the problems at - hand. Several authors highlighted the benefits of networks as access to shared resources (Bryson et al., 2006; Huxam & Vangen, 2005), shared r isk (Casebeer, Popp, & Scott, 2009), advocacy (Provan & Lemaire, 2012), increased organizational learning capacity (Bryson et al., 2009; Kenis & Provan, 2009; Weber & Khademian, 2008), innovation (Provan & Lemaire, 2012; Turrini, Cristofoli, Frosni & Nasi, 2010), and nimbleness relative to traditional organizations (Isett et al., 2011; Provan & 3 Lemaire, 2012). More importantly, the involvement of multiple actors (both individual and or a logical and organized ap proach to addressing the wicked problems (Popp et al., 2014, p. 200). Similarly, networks provide a clearer avenue for beneficial involvement from actors who may not be directly affected by the complex problems. Although inter - organizational networks are defined as a cross - sectional group of individuals linking organizations through relationships, there is considerable variation in how networks are applied in the literature. Authors bifurcate their nature among the roles of governance, structures, formatio n, and accountability into two form s: formal networks (also referred to as mandated networks) and emergent networks (commonly referred to as informal networks ) (Isett, et al., 2011; Popp et al., 2014). Although the degree to which collaboration can be form al is speculative (Chisholm, 1998; Provan & Lemaire, 2012), formal networks are These organizations typically have some form of charter, commission or bylaws, and financial resources dedicated to their common goals (McPherson, Popp, & Lindstrom, 2006). Formal networks are often commissioned by governments or other societal institutions that encourage participation through varying incentives (Popp et al., 2014). Interestingly, Carboni and Milward (2012) argued formal networks generally lack organization power or legitimacy and are not attuned to systemic problems their mandating body created them to address. These formal networks exist as a mechanism for institut - problems with little support other than a charge (Network Leadership Symposium, 2013, p. 7). Concerns with formal networks also call into question the efficacy of artificial relationships whose flourishme nt are simply left up to chance (Popp et al., 2014). 4 - growth of organizational contingencies (Isett et al., 2011, p. 162 ). In absence of a charter, these networks are based in voluntary social relationships and require considerable exchanges to establish common understandings of concepts (Rodriguez, Langley, Beland, & Denis, 2007). Provan and Lemaire (2012) argued those cri teria are necessary, but not sufficient conditions, for defining emergent networks. They contended 1) homophily in interest, 2) physical proximity, 3) heterophily in approach, 4) prior relationships, 5) legitimacy in the market, and 6) market - dependence ma nagement as six critical components of emergent networks (p. 641). Regardless of how their characteristics are subdivided, emergent networks are pervasive and often embedded within formal networks as a o ensure work is accomplished (Rodriguez et al., 2007, p. 152). These networks contribute to creating culture, shared meaning, and the network way of working which simply implies a level of collaboration that differentiates itself from traditional organiza tional work and engages participants in trusting, reciprocal relations. The Wicked Problem: Undergraduate STEM Education Reform Science, technology, engineering, and mathematics (STEM) curricular reform is among many of the wicked problems facing contemporary higher education. Increasingly, high profile organizations are calling to increase the quantity, diversity, and quality of u ndergraduate STEM education in the United States (ASEE, 2009; NAE, 2004; NSF, 1996; PCAST, 2012). With concerns of global competitiveness and meeting future labor markets, government and industry leaders have pushed for business practices like greater inst itutional accountability, performance - based funding, or other outcomes - based approaches (Friedman, 2007). 5 Despite the demands for undergraduate STEM education reform (USER), systemic change in higher education has proven difficult, if not wicked (AAU, 20 18; Austin, 2011; Fairweather, 2009; Kezar & Gehrke, 2015). Many authors (AAU, 2018; Austin, 2011; Fairweather, 2009; Kezar, Gehrke, & Bernstein - Sierra, 2018) lamented the challenges of idiosyncratic and complex systems nested within colleges and universit ies. For example, m ost change initiatives have not been able to scale up teaching innovations due to (1) a strong focus on individual classroom teaching practice (Dancy & Henderson, 2008; Fairweather, 2009; Kezar, 2011a); (2) the lack of attention to addit ional variables that could affect the adoption of teaching practices (Austin, 2011; Fairweather, 2009); and (3) the inability of research communities that study undergraduate STEM instruction to communicate and coordinate efforts (Beach, Henderson, & Finke lstein, 2012). Indeed, individual higher education institutions must contend with coordinating reform among different disciplines within their organization ; incentivizing faculty to adopt reform efforts, and successfully recruit ing and retain ing students w ithin their STEM programs. Each of these complex reform tasks must be completed within a context which does not reward reform. The systemic nature of USER, nevertheless, means a single university successfully reforming undergraduate STEM education on their campus will not successfully resolve undergraduate STEM education at all institutions. Shortcomings in systemic adoption of USER led many to push for greater collaborative approaches that emphasize interconnectivity between colleges and universities (AAU, 2014; Austin, 2011; Coalition for Reform of Undergraduate STEM Education, 2014; Elrod & Kezar, 2015; Henderson, Beach, and Finkelstein, 2011). One of the more popular responses to addressing USER concerns has been the development and use of inter - organiza tional networks to link institutions of higher education (AAU, 2018; AAU, 2015b; Bryson, Crosby, & Stone, 2006). 6 Indeed, various funding agencies (both private and public) have contributed close to a billion dollars to establish inter - organizational networ ks addressing change in STEM education (Kezar & Gehrke, 2015). These networks encourage faculty and administrators to engage in large - scale reform efforts, learn pedagogies, develop new skills, and share ideas across institutional barriers. Over the last 3 0 years, dozens of networks have begun operating in USER. Each network has different missions, memberships, governance structures, and each target s a different aspect mber institutions, while others focus on promoting better pedagogy (Fairweather, 2009; Kezar & Gehrke, 2016). Despite the heavy investment of resources, relatively little is available on networks or how they operate. Problem Statement and Research Question s Undergraduate STEM education reform is a complex, systemic, and wicked problem plaguing higher education improvement. Institutional collaborations through formal networks are one of the most pervasive strategies employed by institutions to combat the sy stemic nature of USER. Most research on inter - organizational networks comes from public administration and public health literature bases, with only a few authors taking on the mantle in higher education. Indeed, scholars across education, sociology, publi c administration, and anthropology all name the increased use of inter - organizational networks to address complex problems, and lament their assumed a place of prominence in the literature on public structures, gradually nudging bureaucracies and markets as the foremost means to organize to address complex problems, 7 etworks have moved ahead without the benefit of a well understood and easily available Although some studies are beginning to emerge on the role and efficacy of formal networks in USER (Hill, 2016; Kezar, 2014; Ke zar & Gehrke, 2016) , the existence and role of emergent networks is relatively unknown. Given the necessity of institutional collaborations in networks will better i nform funding organizations, formal networks administrators, and universities to collaborate more thoughtfully. With this imperative, I posed the following research questions and sub - questions to further explore the presence of an emergent network across f ormal networks in USER. 1. How do formal and informal leaders across formal networks in USER serve as an emergent network? 1a). How interconnected are leaders across formal networks? 1b). How do leaders engage in knowledge diffusion regarding their networks? 1c). How do leaders engage in network learning? 2. How does this emergent network affect USER formal networks? The first research question seeks to understand if an emergent network exists in USER. Each sub - question following the first question dra ws from existing public health and business literature on the structure, functions, and roles of emergent networks. Question (1a) enquires into the social structure and connectivity of leaders in the space. Unlike their formalized counterparts, emergent ne tworks do not have an organizational chart, bylaws, or charter. Instead, these amorphous groups rely on relationships or acquaintances. If members do not know one another, there is little chance they will engage in emergent network behavior. Concepts such as power and trust are linked to the sharing of organizational problems (Katz and Tushman, 1979) making it unlikely for an emergent network to exist among complete strangers. Although social 8 connectivity is foundational to establishing the existence of an emergent network , it is insufficient in determining how the leaders may serve as an inter - organizational conduit. Questions (1b) and (1c) ask more specifically about the functions of emergent networks. The former poses a question to understand both how kn owledge is shared among the formal USER networks and how leaders qualify network knowledge in the space. The latter (1c) probes how the collection of leaders engage in network learning across their affiliations. Aspects of socialization, consensus, and coa lition building are all inherent within the question. Finally question 2 begins to investigate a more tangible impact of the emergent network in USER. These may be new bylaws, or structures to aid formal networks, or new opportunities for formal networks t o utilize. Dissertation Organization and Conclusion This chapter discussed the increased reliance on inter - organizational networks by Specifically, wicked problems in education were highlighted as an impetus for the explosion of various networks in higher education. The surge and subsequent scaling of these networks consequently requires the use of more tacit emer gent networks to build capacity. The second chapter explores the literature related to inter - organizational networks in education, public policy, business, and sociology. With the intention of outlining the existing research on emergent network, the litera ture review highlights the forms of formal (formal) networks operating in the USER space, and explains scholarship on emergent network functions, structures, and governance. Following the literature review, the third chapter outlines the research design, research paradigm, methods, data management, and ethical considerations accompanying the inquiry. 9 Given emergent networks are often latent or imbedded with other social relationships, this study employed a sequential quan - Qual mixed methods research design to examine the nuances of those networks. First participants were surveyed to draw socio - metric data about the relationships of the participants. Then, some participants were interviewed about their experiences with others in the study. The fourth and fif th chapters detail the findings from both of those methods respectively, and the sixth integrates the findings from both previous strands. The final chapter concludes by directly addressing each of the research questions posed in this introduction with evi dence garnered through the research . 10 Chapter 2: Literature Review Despite wide - spread use of formal networks in undergraduate STEM education reform (USER), there is little investigation into how these organizations impact one another. Indeed, most research targets how formal networks advance their missions and affect their affiliated higher education institutions (AAU, 2014a; AAU, 2018; Gehrke & Kezar, 2016; Henderson, Beach, & Finkelstein, 2011 ; Hill, 2016 ). Whereas the previous chapter situated th e rationale for creating various collaborative inter - organizational networks, this literature review aims to further contextualize formal networks and delve deeper into emergent network functions, structures, and roles. Drawing from empirical research in p ublic administration, education, and sociology, I argue for the conceptual existence of an emergent network built through relationships and established across formal networks in the USER space. These relationships, moreover, serve USER goals through the be nefits attributed to emergent networks through the literature. The overview begins with (1) a brief discussion of different formal networks operating in USER, and (2) the impetus for formal networks in creating (in - part) emergent networks, before (3) summa rizing research on the functions, leadership, structures and benefits of emergent networks in advancing organizational change. Formal Collaborative Networks in Undergraduate STEM Education Reform For years scholars argued for USER efforts to be undertaken by faculty networks, professional learning communities, or other inter - organizational conglomerates. These various groups relied on their collective knowledge base and positional power to build co nsensus and inspire change in undergraduate STEM education reform. Indeed, many scholars have called for the use of collectivist movements to leverage systemic change in higher education (Austin, 2011; Fairweather, 2009; Kezar; 2011). As stated, these call s were answered by individuals and 11 organizations interested in improving practices in higher education. The rise and study of formal networks in USER is critical in similarly understanding the rise of emergent networks, as formal networks provide the forum for the relationship - building of otherwise disparate actors. This relationship - building created the potential rise of emergent networks. The following sections briefly outlines various formal network forms operating in USER. Drawing from previous literatu re on networks, communities of practice, and other (CoT) as a distributed group of individuals in education who employ a targeted philosophy to create new practices in undergraduate STEM education (p. 20). Kezar and Gehrke (2015) le e Gehrke, 2015, p. 19). CoTs are situated to deliver information to institutions of higher education through affiliated faculty and are typically composed of member s who share interests in reform (i.e., pedagogy, faculty incentive structures, multiculturalism, etc.). Communities of transformation are situated outside the boundaries of colleges or universities. Given this independence CoTs rely on members to share res ources, including STEM - related knowledge, inside and outside the network (Kezar & Gehrke, 2015). - Sierra (2018) expanded the ational individual professional development. Similarly, authors found CoTs driving philosophies to be a strong uniting factor in their formation and functioning (Kezar & Gehrke, 2017). Although this 12 written philosophy remained unchanged, it was continually re - interpreted to establish consensus when new members were added. This process of consensus building drove stronger personal relationships and established more posit ive and sustainable collaborative relationships (Gehrke & Kezar, 2016; Kezar, Gehrke, & Bernstein - Sierra, 2018). Similar to communities of transformation, professional learning communities (PLCs) are another network - driven initiative commonly cited in the literature as driving undergraduate STEM education reform. PLCs stress the importance of individual professional capacity - building through association with others in similar organizational positions (Astuto, Clark, Read, McGree & Fernandez, 1993). Louis, Kruse, and Bryk (1995) emphasized the importance of members positions facilitated a greater capacity for knowledge diffusion because individuals did not need to use equivalence as a precursor for involvement, PLCs share many traits with CoTs. Those participating in a PLC often have 1) a shared value or vision for their profession, 2 ) a sense of collective responsibility, 3) inclination to collaborate, and 4) a promotion for learning (Stoll, McMahon, & Thomas, 2006). In this regard, they resemble CoTs by appealing to a specific community, domain, and practice, but are generally more h ierarchically structured and member - exclusive when compared to other formal networks (DuFour, DuFour, & Eakers, 2008; Huffman & Hipp, 2001). Professional learning communities are often found in secondary education. Drawing from shared structural equivalen ce, PLCs incorporate teachers into networks to address pedagogy, policy implementation, or student concerns (Mulford & Silins, 2003; Stoll, McMahon, & 13 2001). Vesc io, Ross, and Adams (2007) found PLCs to be a driving force for pedagogical improvement in English and Humanities at a research university. Arguing instructors relied on one another to better frame open - ended questions for their classes, PLCs aided members to establish these questions and facilitated feedback. Outside of the classroom, PLCs are commonly limited to leadership and administrative roles in higher education (Bond & Lockee, 2014; Stoll et al., 2006). These communities allow for administrators of similar roles (e.g., associate vice provosts for undergraduate affairs or student affairs) to gather and discuss major issues within their domain (i.e., student success, retention, policy - changes, etc.) (DuFour, DuFour, & Eaker, 2008; Kezar & Gehrke, 2015) . In addition to the aforementioned formal networks, new research is beginning to address formal networks of higher education institutions. Rather than relying on networks of individuals, organizational change networks rely on entire organizations to leverage systemic change in higher education (ASCN, 2018). These networks utilize many of the same pri nciples of the other formal networks, however the actors within the collaborations serve as representatives, or delegates, from their higher education institutions. Although few conclusions have been reached about the degree to which network interactions b enefit their member organization s , organization change networks appear to share many properties with the other formalized networks including the role of community and domain. Importance of relationships . Formal networks are a medium for disparate actors ac ross the country to interact, share ideas, collaborate, and advocate for change. Undeniably, a formal success (i.e., as the network defines it) is determined by the ability of its members to work together (Provan & Kenis, 2008). Provan and Kenis (2008) noted the capacity to 14 formal networks (i.e., CoPs, CoTs, and PLCs) and their corresponding literatures rely heavily on individual relationship - building, and relationship - maintenance for success. Formalizing networks by creating bylaws and governing structures provides an avenue for networks to organize, apply for funding, and begin to share their strategies in USER. In many instances, however, the concepts of require increasing network size. Although formal governance structures are viewed as complementary to concepts of network trust ( Provan (Provan & Kenis, 2008, p. 237); A decline in trust and s upplementation with governance fundamentally changes the role of the network (Network Leadership Symposium, 2013). With a decrease of interpersonal trust and increase in formal governance, the network shifts away from innovation (albeit, not completely) to reproduction and diffusion (Bryson et al., 2006). Provan and Lemaire (2012) coined this stage as network maturity and sustainability and marked it as a shift of network priorities away from internal capacity building to the development of external legitim acy. Nevertheless, formal networks still require guidance, innovation, and trust in their organizations as they shift their focus to the external environment. In serving that end, individuals may lean on their relationships with others engaged in similar network - like work in the USER space. These informal relationship networks may represent a greater force for innovation within formal networks as they are organically created and are not limited by formal networks scope or mission (McLaughlin & Mitra, 2002) . Longstanding literature links organic relationships, organizational change, individual sustainability, and motivation to high - functioning emergent networks (Healey & Destafano, 15 1997; McLaughlin, & Mitra, 2002; Samoff, Sebante & Dembele, 2003). Despite this literature conceptually pointing to the existence of high - functioning informal networks in USER, the existence of multiple organizational barriers, differences in capital, and other factors prevent a wholesale adoption of the theoretical base. Simply put, are USE R reformers conference buddies who only interact at conferences? At current no research empirically indicates the interconnectedness of individual actors within the USE R . Although conceptually present, more exploration is needed to understand t he web of relationships among people in USE R before more research can begin. Emergent Collaborative Networks In contrast to formal networks, emergent networks are based in voluntary relational ties (Isett, et al., 2011). Often termed informal networks, they are organically derived, have high degrees of interconnectedness, trust, support, consensus, homophilic goals, and many heterophi lic tendencies (Isett, et al., 2011). Unfortunately, available literature detailing emergent r e is no distinct body of literature on informal networks. Consequently, there has been very little advancement of networks may be friend networks, the latter does not necessarily predict the former. Emergent networks , rather, serve as a co nduit for information within a specific domain. Dawes (1996) argued in - network information - sharing occurred tacitly as a way to reinforce valued relationships. Indeed, understanding the role of emergent networks is critical as they have a significant impac t on what, when, and with whom information is shared. In addition to information sharing, emergent networks often arise for the purpose of exploring ambiguous phenomena such as personal network capacity - building, collaborative problem solving, and 16 better n etwork service delivery to key stakeholders (Arganoff, 2007; Imperial, 2005; Provan & Milward, 2001). The following sections condenses the available literature on emergent network to first describe their (1) theoretical implications, before exploring their (2) functions, (3) implicit governance and leadership, (4) and structure. Theoretical interpretations . Several theoretical understandings aid in providing context in the network functions. Based in sociology, structural hole theory explains the origin and differ ences in social capital through position within a social network of relationships. Burt (2000) suggests all placements in a social network hold certain advantages and disadvantages. N etwork closure refers to several individuals who are highly connected and exhibit increased degrees of trust, support, and risk aversion (Burt, 2000, Coleman, 1988). A dense web of interconnected relationships lends itself to the conferring of tacit and technical information in a manner that cannot be replicated easily by indiv iduals who are acquaintances (Burt, 2000). Network closure is also associated with maintaining established group dynamics, upholding the status - quo and provides little incentives for group disruption (including innovation) (Coleman, 1988). Compared to cl osure, structural holes provided the amount of status - quo disruption that could lead to creativity and innovation (Burt, 2000; Burt & Celotto, 1992). Defined as people within a network who do not share high levels of interconnectivity with others, structur al holes flow advantages as they are often linked to other individuals who are not connected to other 17 unique individuals (Burt, 2000; Granovetter, 1973). These individuals are often source points for new information entering and exiting the network. Another theory used to interpret the function of emergent networks is Rogers (2 003) what rate new technologies are spread through a social system . Rogers (2003) argued innovation diffusion is a normally distributed curve comprised of five sequential groups: 1) innovators, 2) early adopters, 3) early majority, 4) late majority, and 5) laggards. These zones represent groups of people who adopt an inn ovation at roughly the same time. In addition to these groupings, Rogers introduced four main elements that influence the spread of an innovation : 1) the innovation itself, 2) communication channels, 3) time, and 4) a social system. In essence , an innovati on relies on individuals using their acquaintances to talk about a new technology for an innovation to spread through a system. Conceptually , one function of emergent networks is to serve as innovators and early adopters. Another social theory regarding r elationships pertains to inter - organizational boundary spanning. Tied closely with Burt (2000) structural holes, boundary spanner refers to individuals who have membership in multiple tangential organizations. Boundary spanners are individuals who connec t their organization to resources in the external environment. They use their social connections to gain valuable knowledge to support local organizational performance (Aldrich & Herker, 1977; Brion, Chauvet, Chollet, & Mothe, 2012; Katz & Tushman, 1981; L eifer & Delbecq, 1978). This action requires boundary spanners to interpret, translate, and adapt external resources to a useable form in the organization (Brion et al., 2012; Katz & Tushman, 1981). Given the import and translation of new resources, bounda ry spanner roles also include many 18 functions of organizational learning, increased social capital, and norm development (Brands, 2013; Brown & Duguid, 1991; Gherardi, Nicolini, & Odella, 1998; Hill, 2016; Phelps, Heidl, & Wadhwa, 2012; Tsai, 2002). Finall y, Weick (1969) presented the concepts of sensemaking to interpret how individuals equivocality in an enacted environment by means of interlocked behaviors embedded in co imperative and sets the base rationale for organizing as a process for removing ambiguity from the environment and uncertainty from individuals lives. The larger role of organizations thus organization itself (p. 40). Establishment of workable certainty in an organization translates to a set of mutually de fined behaviors which allow members to perceive and interact with their environment. Weick (1969) goes on to note that workable certainty is not a fixed goal, but flexes with changes in the environment and membership. In order to survive in perpetuity, process of establishing workable certainty centers on participant interactions (Weick, 1969). involvement before engaging with other members to establish common meanings. Network functions . Within the larger scholarly discussion on emergent networks, there is a strong drive to understand network functions, and differentiate them from other soci al networks (McGuire, 2006). Isolating emergent network functions allows for greater interrogation of their impacts in the social world, most notably, with the organizations in which they are affiliated. These functions are often linked to the sectors in w hich the networks belong (e.g., 19 USER), and commonly associated with the networks intended outcome (Feiock, Lee, & Park, 2012). While these functions are often broad and overlapping, Popp et al. (2014) provided three macro - functions specific to emergent ne tworks. The following sections provide an overview to the primary emergent network functions. Function one: information diffusion and knowledge exchanges . The first major function of emergent networks is information diffusion and knowledge exchange (Popp et al., 2014). Given the increased pressure for organizational performance to address larger problems, emergent networks are often seen as a mechanism for spreading ideas and practice (Hartley & Bennington, 2006). Research on knowledge sharing suggests k nowledge is not simply transferred from one context to another, but is rather continuously reviewed, taken into new settings, and redefined with that context (Hartley & Bennington, 2006). Information sharing therefore is socially mediated by network member s who help determine how it is framed, understood and mobilized within the network and their respective organizations. Similarly, emergent networks are built on trust. Trust and relationships within networks are associated with transfer of tacit knowledge (Burt, 2000; Granovetter, 1973), and are often seen as more effective at diffusing technical or esoteric information to respective groups as network members have a better understanding of the knowledge being distributed. Despite information - sharing practic es of a network, there remain barriers to appropriate diffusion. Organizations external to the network may not exhibit readiness for change or occupy a proximal zone outlined by Rogers (2003). Additionally, Huang (2015) discussed the role of power, politic s, and the awareness of the environment. Network members must be cognizant of the political impacts of their innovations in order for their outputs to take hold. 20 Function two: network learning . Closely aligned to diffusion and exchanges, network learning refers to the capacity for networks to learn how to address problems single organizations cannot (Popp et al., 2014). The concepts of network learning are drawn from literature on organizational learning, which posits the organizational gains of individual s learning is greater than the sum of learning done by its members (Stoyko, 2001). Organizational learning is associated with increased performance and efficiency (Montgomery, 1996; Nelson, Rashkind - Hood, Galvin, Essien, & Levine, 1999). Beyond organizatio definition moves network learning beyond the specific individual or organizational members to a system - level function where information is acquired and negotiated by network members until consensus of interpretation is established for further dissemination to members outside the network (Knight & Pye, 2005). Beyond consensus, network learning is also regarded by its place of origin, namely internal or external to the network. Internal network learning refers to learning generated by network members, and external originating outside the network but applied thro ugh emergent network consensus building . Eisenhardt and Martin (2000) referred to networks leveraging their individuals, their expertise, and collaboration to create new knowledge for trial and even tual more likely the network can create and disseminate more knowledge (Casebeer, Reay, Dewald, & Pablo, 2006). Similarly, a network can absorb information from m embers outside the network and assimilate, transform, and use the information for network development (Zahra & George, 21 2002). This form of learning relies on the social connections of members and organizations to access the necessary information in the env ironment. Function three: innovation . Innovation is a critical function in networks, as it is often one of the main reasons networks are formed (Keast et al., 2004; Provan & Huang, 2012). thin the environment, networks use their internal and external resources to filter information to their members and affiliated organizations (Turrini, Christofoli, Frosini, & Nasi, 2010, p. 533). Innovations within networks can either be derived or refine d from existing knowledge or generated through new tends to be a narrowing and converging process of testing, divergence, and discovery, thus setting the stage knowledge exchanges and network learning in that learning and knowledge exchange contribute to innovation (Popp et al., 2014). Factors which appear to enhance a network s capacity for innovation include: financial support (Thorgren et al., 2009), diversity of network members (Provan & Kenis, 2008), conflict - resolution oriented environment (Reay, Goodrick, Casebeer, & Hinings, 2012), formal or informal mechanisms for encoura ging participation, and transparency or trust (Human & Provan, 2000). Similar to cautions associated with information sharing and diffusion, research indicated networks consider political ramifications of their innovations. Network members are advised to c arefully consider the reception of their innovations by those outside the network, as innovation is not perceived to be relatively advantageous to the existin g process, network 22 members risk their reputation and legitimacy, and the reputation and legitimacy of those with whom they are affiliated. Network governance and leadership . Although emergent networks are primarily founded in relationships and social stru ctures, regulative structures are embedded in social interactions which guide and shape how those interactions occur (Milward & Provan, 2006). Indeed, explicit bylaws or other regulative structures generally are reserved for formal networks or the organiza you simply cannot govern network s , and that in trying to do so you will destroy everything that structures, factors such as trust and human capital all govern how individuals interact with one another (B ryson, et al., 2006; Provan & Kenis, 2008). These factors also shape the concept of leadership among members in the network. Keast, Mandell, Brown, and Woolcock (2004) stated power to their position, education, resources, or political clout which affect how those in the network interact ependence (Keast et al., 2004). Given the role of environmental influences and cultural markers effect on the network, the following sections outline various forms of human capital and how they may influence how individuals interact in emergent networks. Human capital. Broadly, human capital (capital) refers to a set of assets capable of generating a future benefit for at least some individuals within society (Bourdieu, 1987 ; Burt, 1992 ; Coleman, 1988; Ostrom, 2009). Enacted capital primarily confers vari ous manifestations, influence, or power in a social setting to an individual (Bourdieu, 1987). This influence becomes 23 a quality of an individual who can wield it to produce desire returns (Bourdieu, 1987, p. 243). Human capital is a renewable individual re source within a domain, which can only be lost by changes within the domain (Bourdieu, 1987). Social capital . Given the social space of emergent networks, social capital remains a critical component to consider in how these networks are socially shaped, an d how they communicate with organizations in the environment. Putnam (1993) defined social capital as e authors view social capital as conduits aiding someone to access influence power or other resources in the environment (Bourdieu, 1987; Burt, 2000). O ther author s argued social capital is embedded within the relationships and contribute t o building mutual trust and collective identity (Coleman, 1990; Fukayama, 1997; Putnam, 2000). This collectivist perspective articulates how individual self - interests is overridden through relationships, shared meaning and goals. Thus, social capital serves (Putnam, 2000, p. 19). Social capital is unique to other forms of capital as it inherently connects people to others, thus enhancing the role of all other human capital form s (Putnam, 1993). Regardless of the differing definitions, Burt (1992) summarizes social capital as simultaneously Relationships and relationship - building form the foundati on to emergent networks (Isett, et al., 2011). Participation and flourishment of these networks are driven by voluntary social arrangements and sustained by renewed interest and commitment to other. With regards to governance, individuals who are considere d to have high levels of social capital may hold more influence in shaping the language, norms, or behaviors simply because they may initially bridge 24 individuals together. Similarly, individuals with social connections outside the network may provide power and influence in shaping the conversations and leadership of the emergent network. Additionally, the collective emergent network could be viewed as an influential group of people due to their trust, collective identity, and collective connections to other influential people and organizations. Cultural capital. A second form of human capital concerns societal markers, to the social assets of a person that promo te social mobility and desirability (p. 244). Taking four separate but overlapping sub - power drawn from proper, positive socialization within a given society (Bourdieu, 1987, p. 245). This form is acquired over a lifetime, and unlike other forms of capital cannot be given to others. odied capital] cannot be done the embodied state , mastery of language and its relation to others within a given domain (Bourdieu, 1990). Another sub - ulation of personal property, knowledge, skills, and work (Bourdieu, 1987, p. 247). Although this may include economic assets, the objectified state chiefly refers to the cultural symbolism associated with possessing something valued by society. This may b e a priceless collection of fine art, an expensive car, or the ability to dunk a basketball. From this perspective, wealth importance is 25 derived from the cultural significance of its possession, that is, being considered wealthy by others. Finally, the 248). This type of cultural capital is comprised of power and influence granted to an individual through institutional affiliations. Another symbolic representation of powe r and influence, affiliations with education (i.e., degrees, credentials), place of employment, religious affiliations, or awards received from organization may serve as cultural markers of capital. With the emergent network, cultural capital provides unde rstand ing to the actors as they exist in the social world and offers more depth to interpreting why actors seek out one another. Whereas social capital may be conceptualized as an individual or group asset, cultural capital is distinctly situated as an ind ividual quality. Cultural markers and institutional affiliations signal other actors of the potential for a social bond to form. This form of capital can be derived from network organizational affiliations (institutionalized), previous research they cond ucted (objectified), their ability to engage in the nuanced language of USER (linguistic), or simply their charismatic and congenial disposition (embedded). Regardless of an individual cultural nables the flows of communication throughout a social network. Intellectual capital . Intellectual capital concerns the possession of content knowledge. Although originally used in business literature as a means of human resource accountancy (Nahapiet & Gho shal, 1998; Stewart, 1998), intellectual capital recently has been applied in the higher education setting (Gappa, Austin, & Trice, 2007; Khan, Arafat, & Raushan, 2019; Rowlands, 2013). Despite the increased use in higher education, few authors applied int ellectual capital as an individualized quality, choosing more to quantify the role of organizational knowledge as an intellectual asset. For the purposes of informal networks of undergraduate 26 S TEM education reformers, intellectual capital is defined as the amount of knowledge an individual possesses in a specific domain, which can be leveraged to create more knowledge or further increase understanding of an existing concept (Stewart, 1998). This contrasts from the various forms of cultural capital, as it pe rtains specifically to the asset of possessing unique or esoteric knowledge and not the cultural significance of being smart . In considering the members of the emergent networks, intellectual capital introduces the concepts of knowledge domains and the influence they bestow . As emergent networks meet, they encounter complex problems which require expertise from disparate fields a nd cross various boundaries (e.g., disciplinary, organizational, formal networks, etc.). Although individuals may have expertise in one area, they will encounter different areas in which they may not possess capital. These expertise asymmetries inform gove rnance and leadership by potentially informing strategies, language, approaches, or agenda - setting for the network. Intellectual capital can be identified both as an individual property of someone in the network, and to describe the information exchanges o ccurring. Bridging social networks to incorporate an individual with intellectual capital in a specific domain will produce social and cultural capital for network members and provide the opportunity for more intellectual capital generation. Organizationa l capital . Another form of capital speaks to the power of institutional affiliation. Organizational capital refers to the resources, power, influence, authority, communication systems, granted to organizational members for exerting power and influence both inside and outside organizational boundaries (Morgan, 1998; Scott, & Davis, 2007). Although organizational capital may be inclusive of other forms of capital (most notably, cultural capital), this form is primarily comprised of symbolic and economic power (Scott & Davis, 2007). Similarly, organizational capital is not evenly distributed to all members within an 27 organization (Fullan, 2002; Scott & Davis, 2007). Their access rather is mitigat ed by formal positions, expertise, experience, or even other forms of capital (Amey, Eddy, & Campbell, 2010). For various reasons organizational members may be more or less able to harness their organizational capital, depending on purpose alignment, environmental, or contextual factors. This distinguishes the varying le vels of power associated with specific organizational affiliation(s). Given individuals engaged in emergent inter - organizational networks inherently come from other organizations, their position, expertise, and experience within those organizations can con vey power in social networks. These organizations also often provide financial resources to networks giving another dimension to capital provided by organizational affiliation. The number of affiliated organizations actively engaging the network s arena may provide affiliated individuals with necessary economic capital to cultivate power and influence. Beyond the symbolism of cultural capital and intellectual capital, organizational capital often provides the economic driver which impacts how indi viduals act within a social network. Servant leadership . In addition to the role of capital in governing networks, several authors point to the values of servant leadership in networks (Holley, 2012; Mays & Scutchfield, 2010). Servant leadership refers t o a basic premise that leaders put the needs of their followers served (Block, 1993). Wheatley and Frieze (2011) asserted effective leaders needed to view th eir When the principles of servantship, roles of human capital, and fluid nature of interpersonal relationships are held in tandem, leadership becomes a contex tualized and amorphous entity 28 within an emergent network. In this setting different individuals exert influence and leadership in spaces where they have comparative marginal influence on others within the space. Network structure . Given relationships for m the core of emergent networks, their structures are particularly important to practitioners who seek to more purposely design formal [individuals] of emergent networks can provide information about the network structure and its ovided knowledge and information exchange is a key function of many networks, their structure is critically important to understand how exchanges are achieved. The area of emergent network structure and how it may translate to effectiveness is largely unex plored. The studies which do exist rely heavily on 2012, p. 435). Drawing from Burt (2000) and Granovetter (1973), Provan and Lemaire (2012) argued for a mix of both strong and weak ties within a network to achieve efficiency. Existing Literature Gaps Despite the synthesized review presented in this chapter, the area of collaborative networks is essentially uninvestigated. Nevertheless, t he need for more information into how networks work is critical, especially when their financial commitments and deployment in practice far exceeds a strong, supportive evidence b ase. These concerns are magnified when applied to networks in the USER space, where one estimate places total financial contributions nearing one billion dollars (Gehrke & Kezar, 2016). While the literature highlighted in this chapter begins to characteriz e emergent networks as an informal organization situated to address systemic problems, the research - base falls considerably short in three critical areas. 29 Existing research on emergent network s is primarily theoretically based, with most of the literature experience within an emergent network and how individuals activate their social capital to achieve goals. While this work is important in understanding the lived experience of an emergent network, it fails to capture any of its organizational qualities. Indeed, Popp et al. (2014) argued a formal network or any other organization. The focus on social capital also ignores the intersections of other forms of human capital. While social connections are important in establishing the structure of a network, other qualities (i.e., cultural capital, intellectual capital, and organizational capital) affect how those individuals interact. More investigation into the areas of human capital are critical to understanding how individuals cooperate and what social markers are significant in emergent network contexts. Another gap in the existing li terature relates to how few research disciplines study them. At the time of this study, emergent networks are only recognized as social phenomena in public administration, public health, and criminal justice. Authors in public administration inquired into emergent networks developed across city governments, and those in public health investigated linkages across different hospitals. Although those studies recognize the social power of emergent networks, their applicability to higher education is suspect. Wi thin higher education research there are few examples of research on formal networks and no known research on emergent networks. While the existing literature in public administration, public health, and criminal justice may help guide formal and emergent networks in higher education, more research is needed to address how networks operate in the higher education sector. 30 Despite authors bifurcating inter - organizational networks into formal and emergent groupings, no known research investigates how these two network forms interact with one another. Most authors singularly identified and examined the efficacy of formal networks or the formation of emergent networks. These two types of networks, nevertheless, undoubtedly co - exist. Indeed Isett et al. (2011) arg ued emergent networks may be the seminal place for innovations within formal networks. Recognizing this gap in the literature, the second research question sought to recognize and detail how the two forms of networks affect one another. Conclusion This chapter targeted the differing forms of formal collaborative networks as they exist within the USER space. Highlighting the contrasting amounts of structures associated with the formal networks, this review posited the role of relationship s and lifecycles in the scaling and search for external legitimacy of formal networks. As networks grow, they shift their focus from internal capacity building to production and diffusion. Following the review of the more formal networks, this chapter offe red a deeper examination of emergent networks, their functions, governance, leadership and potential structures. Given the use of networks in USER and impetus from scaling reform (AAU, 208; Kezar & Gehrke, 2015), there stands an imperative need to know m ore about how formal and emergent networks operate in tandem. How do networks establish their internal processes while shifting their focus to the external environment? What role do emergent networks play in the scaling of formal networks? How do emergent networks differ from a simple collection of friends operating in the same USER space? Are emergent networks a stronger mechanism for innovation when compared to their formal counterparts? Each of these questions is important and equally valid 31 however, this s tudy investigate s the social networks of the leaders in formal networks in order to ascertain if they work as an emergent network. 32 Chapter 3: Research Methods The extensive deployment of different networks in USER has resulted in large investments of human and financial resources. Although some literature is available to guide the committed resources, almost all understanding is derived from work in public admin istration and limited almost exclusively to the functioning of formal networks (Bryson et al., 2006; Isett et al., 2011; Popp et al., 2014). This relative dearth in research is not due to a lack of imperative as emergent networks often operate behind close d doors of formal networks or as primordial formal networks (Bryson et al., 2006). The difficulty in inter - organizational network study rather is owed to the amorphous nature of networks across all domains and disciplines (Isett et al., 2011). The fluidity is compounded when interrogating emergent networks, which often double as friendship, colleague, or peer networks (Gilchrist 2006). Popp et al. (2014) contended difficulty is not a legitimate excuse when emergent networks hold the - organizational networks outlines the research approach I took to map and interpret the functions of an emergent network operating across formal networks in undergraduate STEM education reform. As described in the opening chapter, my research questions sought to explore the existence of an emergent inter - organizational network i n USER. Specifically, I set out to explore the following research questions: 1. How do formal and informal leaders across formal networks in USER serve as an emergent network? 1a). How interconnected are leaders across formal networks? 1b). How do leaders eng age in knowledge diffusion regarding their networks? 1c). How do leaders engage in network learning? 2. How does this emergent network a ffect USER formal networks? 33 Each sub - question inquired into a different aspect of emergent networks as described in the li terature review. Question 1a) inquired into the social structures embedded in the emergent network, while questions 1b) and 1c) inquired into different articulated functions of an emergent network as described in the literature review (i.e., knowledge diff usion and networked learning). The final research question explored the impact of the emergent networks on both formal networks and USER altogether . Philosophical Grounding Given the intricacy and potential confounding nature of investigating emergent networks, I approached these research questions pragmatically. Unmarried to any particular methodology or ontology, a pragmatist research approach centers on answering research questions by the best suited tools. Powell (2001) argued The pragmatist epistemology stands in contrast to prevailing positivist and anti - positivist views of scientific discovery. Whereas positivism emphasizes the objective, law - like properties of a brute reality independent of observation, anti - positivism emphasizes the creative role of active, subjective participants, none of whom owns a privileged claim on truth. Pragmatism, on the other hand, rejects positivism, on grounds that no theory can satisfy it s demands (objectivity, falsify - ability, the crucial experiment, etc.); and rejects anti - positivism, because virtually any theory would satisfy them. (p. 884) belief, doubt, and habit (Pierce, 18 77). Social reality and knowledge are based on beliefs and habits which are constructed through processes of institutionalization, legitimation, and socialization (Berger & Luckmann, 1967; Yefimov, 2003). Berger and Luckmann (1967) argued it is impossible to understand an institution adequately without an understanding of the historical process in which it was produced. In this regard, research is always embedded exis 34 w With regards to methods of inquiry, pragmatism embraces all tools espoused by positivists and interpretivists (Creswell, 2009). Rather than selecting a methodology to guide the methods and data coll ection, pragmatism emphasize s the research problem and use s all available approaches to understand that problem (Rossman & Wilson, 1985). This approach fronts the importance of the questions posed and allows the researcher to confront the questions unbounded by the limits of one particular methodology. Paramount to this approach, however, is understanding the questions posed (Morgan, 2006) . In order to fully engage in the work, the researcher must know their questions and a multitude of different methods to know which tool may be the best to examine the phenomena. Research Approach From this perspective, I utilized a mixed - methods approach for collecting, analyzing and mixing both quantitative and qual mixing both types of data are that neither quantitative nor qualitative methods are sufficient by 97). Given the articulated difficulty of researching emergent networks, a single methodological approach is insufficient. An approach solely utilizing surveys may uncover frequencies of interactions or topical conversations among network members, but surve y data lacks the richness needed to uncover how networks operate or the experiences of those found therein. Similarly, qualitative approaches may provide data on how a network operates but provides little information beyond the individuals experience. If t he methods are used in tandem though, their 35 complementarity gives a more complete and nuanced understanding of the phenomena (Creswell & Plano - Clark, 2011; Green, Caracelli, & Graham, 1989; Johnson & Turner, 2003; Tashakkori & Teddlie, 2003). Provided this rationale, I employed a sequential explanatory quan - Qual mixed method design, comprised of two separate strands (Creswell & Plano - Clark, 2011; Creswell, Plano Clark, Guttman, & Hanson, 2003). This design was sequential as I collected quantitative data fir st before collecting any qualitative data (as opposed to collecting both simultaneously), and explanatory as I used quantitative to inform the qualitative approach. The work began with surveying participants to gather quantitatively - oriented socio - metric data, which was collected, analyzed, and used to inform an interview protocol and qualitative data collection. In this study, the quantitative data w ere employed to establish the social interconnectedness of participants in formal USER networks and help to identify the types of relationships which exist. Following the survey, interviews were conducted to further explain and contextualize the relationships beyond the social network statistics. In total , quantitative data w ere used to understand the connections of leaders, and the interviews further described those data by exploring the participants lived relational experiences in more depth. Priority in this work was centered on conducting and coding interviews, as it focused on in - de pth explanations of the quantitative socio - metric data (Creswell & Plano Clark, 2011, Creswell et al., 2003). This emphasis is signified by the capitalization of Qual in the research design description. The quantitative and qualitative strands were connec ted by using quantitatively ascertained network structure, participant placement within the network, and network homophilic/heterophilic tendencies to inform the interview protocol. Initial data integration was completed during the analysis of the quantita tive data, but prior to qualitative 36 data collection. Secondary data integration took place at the conclusion of qualitative data analysis, when quantitative findings were reinterpreted for more detailed meaning (Figure 1). Given the explanatory nature here in, the ethos driving the design focused first on understanding the emergent network dimensions before exploring the experiences of relationship - building and engaging in conversations regarding the future of USER. Figure 1 . Quan - QUAL Sequential Explanatory Mixed Methods Design 37 Participant selection . Given thousands of individuals participate in some form of USER, participant selection was a critical component in identifying which leaders could be a part of an emerge nt network across formal STEM networks. Research in public health found organizations nominate leaders in many ways including self - nomination, nomination by members of a group, nomination by formal leader s , or expert nomination (Valente & Davis, 1999; Vale nte & Pumpuang, 2007). Each of these strategies, nevertheless, w as accompanied by its own variation of selection bias (Valente & Pumpuang, 2007). Provided the bias limitation, the first step of participant selection was structured around a consensus building approach called the Delphi Method. This method required a group of pre - established nominators to confidentially recommend a group of individuals whom they believed fit preset criteria (Avella, 2016). By collecting and corroborating the disparate set of nominees, the researcher may draw a set of participants from those who are present on multiple lists. Most importantly, proponents of the Delphi Method encourage using multiple rounds of nominations to ensure nominator fidelity and reliability (Avella, 2016). For the purposes o f identifying formal and informal leaders in USER formal netwo rks, I conducted a combined Delphi process across two comprehensive USER formal networks. The first group of nominators were taken from the leadership teams of the National Academies of Science, Engineering, and Medicine Roundtable on Systemic Change in Un dergraduate STEM Education (National Roundtable). The second group of nominators were drawn from the Accelerating Systemic Change Network (ASCN). These groups served as the ideal starting point for nominators as their missions both seek to convene leaders from across STEM policy, research, and implementation communities (About ASCN, 2019; NASEM, 2018). 38 Nominator profile . For the purposes of participant selection and identification, I conducted the first round of the nomination process with six leaders with in the National Roundtable and ASCN. Three nominators represented each of the organizations. To avoid any potential bias from the start of the process, diversity was sought among the initial nominators. These individuals had served in multiple leadership c apacities stretching across multiple domains in USER (e.g., policy, funding, faculty, higher education advocacy, university administrat ion , etc.). Four of the nominators were women and two were men. Four nominators primary responsibilities were at a post - secondary institution, while 2 nominators worked for different STEM - affiliated non - profit organizations. Geographically, the nominators were located in all major regions of the United States with two nominators repr esenting the mid - Atlantic/Northeast, and one nominator located in the Central Plains, Midwest, South, and West Coast respectively. Reflecting the lack of racial diversity in the STEM fields, five of the nominators were white and one identified as Latino/a. In August 2019, the selected nominators were asked to confidentially identify 15 30 formal or informal leaders in USER networks through a Qualtrics survey (Appendix A). All nominators were provided a sample list of USER formal networks to aid in their nomination process and were advised to include individuals from additional groups if they felt it appropriate. Self - nominations were not included. Following the initial nominator submissions, their lists were provided to a second group of two reviewers for validation and consolidation. A second group was used as many of the nominated names in the initial lists included the nominators in the first round. The second group helped to avoid any conflicting interests. In the second round of review, no additional names were added or subtracted from the list 39 Participants . All six nominators returned a list of names between 19 and 30 names with a total of 128 nominees, representing over 20 different USER formal networks. Of the 128 total nominations, there were 86 un ique individuals. Nineteen of the individuals nominated appeared on three or more of the nominator lists, and nine names appear ed on five or more lists. Those nineteen individuals were validated by the second set of nominators as being preeminent leaders i n USER and were invited to participate in the study. No names were excluded by the second nominators. Provided the social network survey required participants to disclose their names to others in the study, participation consent was sought prior to the sur vey distribution. Consent was elicited to all 19 of the initial participants in mid - September 2019 with 17 consenting to participate. Quantitative Methods: Social Network Analysis After finalizing the participants for the study through the nomination process, data collection began with the social network survey . The purpose of the survey and subsequent analysis were two - fold. First, asking participants to map their social connection s among other leaders in USER networks provided an initial understanding to the interconnectedness of those leaders in formal USER networks. Social network analysis relies on metrics of centrality and cohesion as measures of brokerage, structural holes (i. e., Burt, 2000), and trust and collective identity (i.e., Borgatti, Everett, & Johnson, 2018; Putnam, 2000 ). Mapping the social components of this social network aided in identifying potential structural components (e.g., central or peripheral actors, isol ated actors, boundary spanners, etc.) of the network. Second, collecting and analyzing the social network data provided information to be included in subsequent interviews described in subsequent sections. 40 Data collection . In late September 2019, particip ants were sent the survey populated with the other nominated names from the nomination exercise (Appendix A). Prior to completion of the survey, each participant received specific direction on how to complete the survey and a worksheet providing additional context to the study (Appendix B). Socio - metric data w ere collected through a Qualtrics survey, distributed online to participants, with the aforementioned attachments. Participants had two weeks to complete the survey with one reminder relayed through email after one week, and a final reminder sent when 24 hou rs remained in the survey. An extension was granted to one participant to accommodate the critical need for a high response rate. In the survey, participants were asked to provide basic demographic and education information and to list any USER networks in which they are currently affiliated. Previous USER network affiliations (within the last 5 years) were collected. Following the basic survey questions, each participant was asked to denote relational connections with other members in the pre - populated lis t of participants. Socio - metric data w ere collected across three vectors loosely reflecting the literature on the functions of emergent networks. Corresponding the knowledge - sharing function of emergent networks, participants were asked to indicate with wh ich of the other participants they regularly spoke about their USER network business ( Business ) . Reflecting network learning and consensus building, participants were polled to indicate which of the other participants they periodically discussed their netw ork problems ( Problems ) or sought advice ( SoughtAdvice ) . Note, the innovations function of emergent networks was not directly interrogated in this process, as social network surveys are not situated to specifically ask about outputs or products of relation ships. Following the social network aspects of the survey, 41 participants could nominate, at most, two names of individuals they speak to regularly about USER related networks who were not included on the list. Data cleaning . Upon the conclusion of data collection, all responses were downloaded from the Qualtrics collection software and moved to fit with a Microsoft Excel file. Individual responses were reformatted from their raw responses to three unique 17x17 adjacency matric es for each network, and one 17x15 two - mode adjacency matrix to capture USER network affiliations. All participants were assigned a unique and randomized participant identification number and all categorical data were recoded to numeric identifiers. Any mi ssing demographic, education, or network responses were populated using open and publicly available information found on the internet; however, participants who selected prefer not to respond for specific questions did not have any publicly sourced informa tion included in the dataset. Following the recoding and categorical cleaning, all three networks were loaded into UCINET for further cleaning and adjustment. The Business and Problem networks were assumed to be comprised of reciprocal dyads. These networ ks were symmetrized using UCINET software. Using this function filled in any missing unreciprocated ties. Essentially, this process identified when individual, i , identified they spoke with another individual, j, but the j did not indicate the i as a conne ction. Borgatti, Everett, and Johnson (2018) indicated symmetrizing was a common aspect of data cleaning when no directionality is assumed in information flows (p. 86). Both (spoke about) Business and (discussed) Problems networks assumed multi - directional ity in the network, as one does not speak or discuss uni - directionally in a conversation. Symmetrizing the data allowed for the inclusion of consenting participants who did not respond the survey. This reduces data waste by using data other participants pr ovided about non - 42 respondents. Both the Business and Problems networks include data on all 17 consenting participants. SoughtAdvice was not assumed to be reciprocal given information asymmetry inherent in each of the dyads. Plainly, an individual may seek advice from another, but one cannot assume the latter will seek out former for advice. Given the inability to assume reciprocity in this network, only socio - metric data for the 15 contributing participants were included in this network. Although I could no t assume reciprocity in network ties, I did assume individuals who sought advice being elicited in the past tense. Provided this assumption, I transposed SoughtAdvice data to r epresent a new network. This new network was called GaveAdvice. Transposing networks of directed ties reverses the information flow and direction of the ties (Borgatti, Everett, & Johnson, 2018). I transposed this SoughtAdvice for two reasons. First, I switched the direction of ties to match the flow of information (i n this case, advice) to ease with the interpretation of centrality measures and sociograms. Second, transposing directed ties allows for more inclusion of consenting participants in the study. This translation also reduces data waste by including those who did not respond and relying on the responses of advice - seekers. Despite having 17 nodes, the GaveAdvice network has only 15 participant responses. Data analysis . Data collected from the social network survey w ere used in four strategies. First, measures understand the network density, cohesion, and to recognize the presence of subgroups. Althoug h measures of centralization are almost always used in network comparison, these measures provide insight into the cohesiveness of the emergent network in relation to their sharing of 43 formal USER network information. Importantly, centralization measures pr ovide interpretations into subgroups and core - periphery structures, which provide theoretical interpretations to how the social network functions. Second, individual respondent centrality measures were taken to understand how central members are to the eme rgent network. Compared to centralization measures, centrality refers to the structural properties of an individual person within a greater network (Borgatti, Everett, & Johnson, 2018). Measures of centrality are typically understood through the number of relationships (or ties ) an individual has within a given network and their betweenness. Betweenness is a measurement of how central of a role an actor plays in transferring information through the network. Centrality statistics have many other forms, and m ost are used to represent direct relationships or indirect relationships (Borgatti, Everett, & Johnson, 2018). Centralization and centrality statistics independently reveal little about the network beyond who talks to whom . Nevertheless, when held in tande m with known literature on social capital, the network structures begin to elicit more meaning. Concepts such as structural holes (Burt, 2000), closure (Coleman, 1990), and bridging - bonding tendencies (Putnam, 2000) provide an additional lens to understand what position within the network may translate to practice. Applying statistics to measures of centralization provides a crucial first - step to understanding the presence of an emergent network. The amount of interconnectedness among the individuals in the network can inform how (and to what degree of depth) they communicate. For preliminary social network analysis, I used the sures (p. 7). With regards to statistical measures within the network, the Goldilocks P rinciple regards normalized centralization and centrality scores below 0.30 to be low, 0.30 to 0.60 moderate, and levels above 44 0.60 to be high. These measurements can be applied to both centralization measures for the whole network (i.e., centralization an d cohesion), and individual centrality scores. The authors stressed the Goldilocks Principle is a loose interpretation of network data, and factors such as network size, context, and domains should be considered in full network analysis. In addition to ne twork specific statistics, I tabulated Quadratic Assignments Procedure ( QAP ) regressions and QAP correlations to recognize prominent factors within the network. QAP methods are calculations used specifically in social network analysis, as social networks v iolate foundational tenants of traditional regression methods (Borgatti, Everett, & Johnson, 2018). The QAP methods aid in identifying specific personal characteristics or other variables which lead to relationships forming. These methods were used to iden tify homophilic tendencies within the network, and how closed the network is to difference. Many authors highlighted the value of understanding homophily and heterophily within an emergent network. Arguing homophily (along some variable) was a driving fo rce in emergent network formation, Provan and Lemaire (2012) indicated homophily has value in the beginning stages of a network but heterophily is necessary for emergent networks to thrive. Isett et al. (2011) stated homophilic tendencies were unsustainab le and greatly limited the networks ability to innovation. Similarly, homophily is linked with network closure and increased levels of together. Finally, ind ividual respondents were charted into a map to visually represent network structure. Commonly referred to as a sociogram , this visual is a map of the different anonymized individuals and their connectedness provides a collective view of the network. Four sociograms were graphed with the data and are located in Chapter 4 . Three of the sociograms visualized the 45 individual participants and their connections to one another. The fourth sociogram incorporat e d 2 - mode of data and chart ed the interconnectedness of participants and their network affiliations. This will allow for the charting of formal network representation and connectedness. Admittedly, a sociogram does not provide much utility beyond visual clues for the researcher. D raft ing the sociogram will solely be for use in interviews with the participants. Qualitative Methods: Semi - Structured Interviews Although social network analysis is a valuable tool in mapping emergent networks, it does not fully capture the many dimension a nd contexts of relationships. In short, interpersonal relationships cannot be fully represented with a statistic or line. While helpful in establishing network boundaries and connections, social network analysis is a crucial first step in interpreting emer gent networks. Indeed, several authors argued qualitative investigation of emergent networks are critical to bounding and operationalizing the relationships (Casebeer, Huerta, o simply make attractive pictures and tell stories that could have been told without network data...the problem is that these easy presentations tend to bake in all kinds of assumptions that should always be questioned" (Frank, Kim, & Belman, 2010, p. 22 3) . In order to avoid making attractive pictures and baking assumptions, I conducted six semi - structured interviews with participants to discuss their experiences working with others identified in the study. Data collection . Following the social network data initial analysis, six participants were selected from the existing group of participants and interviewed about their experiences with other participants. All six individuals successfully completed the initial social network survey. The two participant s who did not complete the survey were not considered for interview. Criteria for selection in the interview process was based solely on composite centrality measures drawn 46 directly from the social network data. Standardized composite centrality measures w ere compiled through averaging (1) number of ties, (2) closeness, and (3) betweenness for each participant. Given that each participant represented a nominated leader in formal USER networks, diversity in these scores w as sought to understand how different aspects of the social network were experienced. Two participants with the lowest and two participants with the highest standardized composite centrality scores were invited to interview. Additionally, the two participants with the centrality scores closet s to the group averages were invited to participate in interviews. Their selection resulted in three groups of two interviewees representing high, average, and low connectivity to others within the group. Interviews began in mid - November 2019 and concluded in mid - December 2019. Each interview session lasted between 46 and 83 minutes with a unique semi - structured protocol derived, in part, by data drawn from the social network survey. A template for the interview protocol can be found in Appendix C. At the b eginning of each session, the research project was briefly described before each interviewee was asked to discuss their background in USER and affiliations to formal USER networks. Following their USER backgrounds, each interviewee was asked to elaborate on their social network responses with others in the survey. Participants were prompted to discuss how they met the individuals they signified in the survey, and how they would characterize their relationship. Following a re - introduction to othe r participants in the study, interview questions pivoted to emergent network functions derived from the literature review. Paying particular attention to how participants share information about their formal of this information, and how this information is codified to outcomes or deliverables, the interviews focused on individual relationships or 47 chains of relationships and the role these relationships played in formal network - organizational needs. Participan ts with median and high centrality composite scores were shown anonymized sociograms comprised of the social network data gathered in the first strand. These participants had their placements in the network revealed to them in the interview and were asked their initial reactions and interpretations of the social network. Those with low centrality scores had the sociograms described to them but were not physically shown any visuals. Their low centrality scores were also not revealed to them. This protocol wa s informed by literature on ethical considerations in social network analysis (Borgatti & Molina, 2003) and driven by reducing unintended harm to those who believed themselves to be more central actors. Near the end of the interviews, participants were ask ed about any additional nominations (if applicable) they added to the survey. The interviewees were specifically asked why they nominated certain individuals, and where they believed the nominated individuals belonged in the network. Data preparation . Aft er each recorded interview the audio files were transcribed verbatim within three days by the researcher. During transcription, all personal identifying information w as removed and recoded with either a participant - selected or researcher - assigned pseudonym . The pronouns used for each interviewed participant were also randomized to further protect the participant confidentiality. All references or citations to other participants in the study were reassigned to either a pseudonym (if they were interviewed) or their participant identification number (if they were not interviewed) assigned in the social network strand. Following the transcription, transcripts were further cleaned to remove all verbal fillers, pauses, technical difficulties with Zoom, researcher introductions, and other irrelevant information. 48 Removing these fillers allows for the researcher to get closer to the data (Creswell & Plano Clark, 2011). Prior to analysis, each interviewed participant was sen t a copy of their transcript and asked to c omplete a member check. Member checking is an opportunity for participants to approve particular aspects or interpretations of the data they provided (Creswell, 2009). Although some authors argue member checking is better suited for themes or initial inter pretations (Creswell, 2009; Doyle, 2007), ethical considerations of social relationships and participant indicated themself as critical gatekeeper within this group, but no other individual indicated them as critical, member checking would be, at best, counterproductive. No other notations or follow - up questions were provided by the researcher in the member check. Participants were provided one month and asked to verif y the transcripts and provide any notations, edits, or redactions. Four of the interviewees responded with no amendments, and two participants did not respond. Following the transcription and member checking, all data were loaded into NVivo for analysis. D ata analysis . Data collected through the semi - structured interviews were analyzed through thematic analysis. Thematic analysis is a flexible tool that helps to make sense of messy qualitative data (Boyatzis, 1998; Braun & Clarke, 2006, 2012; Joffe, 2012). I used Braun a nd yourself with the data, 2) generating initial codes, 3) searching for themes, 4) reviewing themes, 5) defining and naming themes, and 6) writing up the r eport. I began with the first step familiarizing the data , through transcription and subsequent reviews. Each of the semi - structured interviews were transcribed verbatim within three days of their completion. No automatic 49 transcription software was used. E ach transcription file and audio recording were read and played simultaneously in totality at least two times before coding began. Initial coding began with three thematic functions of an emergent network (knowledge diffusion, network learning, and innova tion), where evidence of emergent network behaviors was sorted into specific areas. Through the coding process, additional sub - themes were identified. These themes spanned across the initial emergent themes, and provided additional environmental factors im pacting the emergent networks functions. Each code was further contextualized and Clarke, 2006, p. 88). Once the final codes were established, I returned to original interview transcripts to verify their fit with the holistic experiences described by the participants. Positionality . Given the heavy capital investment and often contentious nature of undergraduate STEM education, this work necessitates some context and positionality on behalf of the author. As an organizational scholar, I acknowledge the pervasive power of social relationships are conduits for information, n orms, behaviors, and more tangibly, opportunities. I attribute much of my personal and professional successes to my social connections. Indeed, I met my wife through a mutual friend. As a researcher, I serve as an assistant on a National Science Foundation grant studying the role of formal USER networks. This project seeks to understand the roles, lifecycles, and challenges faced by networks comprised of organizations. My initial introduction to USER initiatives was through this project, and my interactions with those within the space is wha t spurred the development of this project. My experiences within this space led me to believe those working in USER do so amicably and with the best intentions, but often have 50 many competing responsibilities that leave formal network organizational respon sibilities to secondary tasks. As a first - generation college student from the rural southern United States, I recognize the need to positively and sustainably reform undergraduate STEM education at all institutional levels. Science education at my secon dary institution was underfunded and under - emphasized with few opportunities to engage in advanced scientific study. These experiences led to a confused and often timid reaction to all physical science and mathematics courses at subsequent education instit utions. I reject the pervasive culture that substantiates exclusivity in the sciences and firmly believe the continuation of higher education current role in the United States rests on its institutional ability to reform and advance the needs of students . Conclusion The call for more study into the existence and functions of emergent collaborative networks is as contemporary as it is necess ar y. The use of these networks as both a response to systemic change (Bryson, et al, 2006; Huxam & Vangen, 2005; Isett et al., 2011) far exceeds empirical understandings of their functions. Contemporaneous research must push beyond interpreting inter - organizational cultures, and functions of emergent networks, and uncover tenets to network success (Popp et al., 2014, p. 28). Indeed, understanding networks successes and functions may very well predict how society overcomes t he pervasive concerns of wicked problems. In serving that end, this study uncovered one emergent network which exists a longside formal networks in the USER space. Through use of a sequential, quant - QUAL mixed methods approach, this study sought to understand the interconnectedness of otherwise assumed disparate 51 actors before interrogating their relationships (or lack thereof). The social network analysis operates as a critical initial step in establishing connections of leaders and theoretically pointing (through concepts of structural holes, closure, and tie strength) to relationship functions. The semi - structured interviews provide more depth, clarity, and nuance to relationships which cannot be captured from a single cross - sectional survey. Through probing the emergent networks in USER, studying this phenomenon provides useful information to those operating within USER, and aid faculty , policy - makers, and funders to identify key places for change within the network landscape. It also uncovers potentially untapped resources within formal networks. More broadly though, this study adds to the literature by filling several articulated gaps across multiple disciplines. Through use of a requested, yet unique, approach to understanding emergent networks, this study may help practitioners, funders, and scholars truthfully begin to uncover what constitutes the network way of working . 52 Chapter 4: Social Network Findings The first strand of the mixed - method study focused on gathering socio - metric data on the nominated participants. Concentrating on the first research sub - question (1a), this strand primarily sought to establish the social relationsh ips among the participants and explore the interconnectivity of these relationships. Fifteen of the seventeen participants responded to the invitation and completed the survey. The data gathered captured individual demographic, educational and employment i nformation, and individual formal USER network memberships. Specific social network data were captured through eliciting responses for participants in three separate domains. These social networks were: 1) Spoke about Network Business ( Business ); 2) Discussed STEM Network Problems ( Problems ); and 3) Sought Advice from this person to aid my network ( SoughtAdvice ). As mentioned in Chapter 3, SoughtAdvice was transformed to GaveAdvice to reflect information flows in the network. The following chapter out lines the initial findings and responses to the social network survey. Although this chapter is primarily organized around three different networks, all three networks pertain to the same individuals in the same contexts of USER. I conclude the chapter wit h preliminary findings, initial responses to my research questions, and an evidence - based discussion on the existence of an emergent network among these participants. Results Although networks were symmetrized in the data cleaning process, reciprocity meas ures remained moderately high prior to the symmetrization. High dyad reciprocity metrics represent high level of inter - rater reliability (Borgatti, Everett, & Johnson, 2018; Loitz, Stearns, Fraser, Storey, & Spence, 2017). High reciprocity scores similarly reflect a high degree of validity with the data as dyads are providing evidence of an undirected relationship. A table of the 53 Table 1 . Unsymmeterized Reciprocity Score for all Networks Business Network Problems Network Gave Advice Network Dyad Reciprocity .6 .57 .29 unsymmeterized reciprocity statistics are found in Table 1. Note GaveAdvice was included in Table 1 to for completeness. The reciprocity for GaveAdvice is still moderate for a uni - directional network. A summary of participant characteristics is reported in Table 2 and Table 3. These data include information from the non - responsive participants. Twelve participants identified as women (71%), four men (23%), with one abstent ion (6%). A plurality of participants w as in their 50s (47%). Reflecting the lack of racial and ethnic diversity in STEM, sixteen participants indicated they were white (94%), and one participant wa s biracial (5%). About 40% of participants were employed by a four - year, public research university , and about 30% represented education non - profits or education advocacy groups. On average, respondents indicated at least 23 years working in undergraduate STEM education reform efforts. Identification of unique formal network memberships were critical t o the study, as the researched questions inquired into emergent networks across different formal networks. Then different network affiliations reported are located in Table 3. Respondents indicated 19 unique formal USER networks located across the country and funded by both public entities and private foundations. The average number of USER network memberships per individual respondent was 3.41. The Bay View Alliance (BVA) and Accelerating Systemic Change Network (ASCN) had the highest representation in the group, but the majority of networks presented only had one individual involved. 54 Table 2 . (n=17) Variable (range) Response Category n % Gender Identity Man 4 23 Woman 12 71 Prefer not to respond 1 6 Ethnicity White 16 94 Not White 1 6 Age Range 40s 4 23 50s 8 47 60s 3 18 70+ 2 12 Position Faculty 7 41 University Administrator 4 23 Non - Profit Administrator 5 28 Other 1 6 Tenure in STEM 23.4 (8 - 42) Primary Organization Affiliated Non - Profit 6 35 Regional Comprehensive College 1 6 Liberal Arts College 2 12 Research University 8 47 Number of Network Affiliations 3.4 (1 - 5) Table 3 . Summary of Respondents Network Affiliations (n=15*) STEM Network n % Accelerating Systemic Change Network 12 71 Bay View Alliance 8 47 Roundtable on Systemic Change in Undergraduate STEM Education 8 47 Network of STEM Education Centers 6 35 AAU STEM Initiative 5 29 Project Kaleidoscope 3 18 Science Education Resource Center 3 18 Coalition for Reform in Undergraduate STEM Education 2 12 POD Network: Professional and Organizational Development 2 12 Building Leadership Capacity 1 6 Center for the Integration of Research, Teaching and Learning 1 6 Cottrell Scholars 1 6 Mathematics Teacher Education Partnership 1 6 Partnership for Undergraduate Life Sciences 1 6 Student Engagement in Mathematics through Institutional Network for Active Learning 1 6 Science Education for New Civic Engagement and Responsibilities 1 6 Transforming Education: Multidimensional Evaluation of Teaching 1 3 Transforming Education, Supporting Teaching and Learning Excellence 1 3 Venture Well 1 3 55 *Individuals could list up to 5 networks in which they belong. This increased the networks represented above the number of participants. Business network. The business network results are presented in Table 4 and depicted in Figure 2. Functioning at one of the most basic levels of interactions, Business targeted a general exchange of information across the participants. This network and its cohesion statistics were meant to establish a baseline of how the disparately nominated participants interacted. Par ticipant data in the Business provided insight into how this collection of individuals operate . The associated centralization statistics demonstrate a high level of interconnectivity across multiple areas. First, overall density for the network was 75% rep resenting a high level of interconnectedness indicating 75% of all possible connections were present in the network. Degree centralization was 28% which indicated the network was not dominated by a single actor within the network. The betweenness centraliz ation (closure) of 3.07% identified few gatekeepers across this network. The statistics for components (1), compactness (88%), and average geodesic distance (1.25) all represented a close, tightly linked group of individuals. Moreover, the absence of any c liques (fragmentation) also indicated there were few places of information bottlenecks with respect to participants speaking about the business items of their networks. Taken together all of these centralization and cohesion measures depict a highly connec ted network with few central or influential individuals dominating or affecting the flow. From a whole - network perspective, the Business network operated as a cohesive grouping of individuals who frequently share or speak about the regular on - goings of the ir formal USER network business and agendas. Although density and centralization are important indicators for understanding the whole network, they provide little insight into how the nodes are distributed or affected. In short, more characteristics of the whole network should not be confused with t 56 better statistical understanding of how the node experiences the network, I measured individual node centrality and power. The average number of ties per individual in Business was 12 with a range of ties between 6 and 16 per individual. The modal number of ties was 14. Participant 15 had the highest number of ties (16), while participant 10 had the highest individual betweenness score. Participant 11 and 2 had the two lowest number of ties at 6 and 7, respectively. Individual betweenness scores also mirrored these nodes with individual 15 and 10 representing the highest (5.468) and Individual 2 with the lowest (.202). Betweenness and centrality statistics are depicted in Table 4. After reviewing the centrality and power scores for each of the individuals in the Business network, I used a correspondence analysis to map and reinterpret the distribution of individuals around four factors of centrality (e.g., degree, eigenvector, closeness, and betweenness). Traditionally, the corr analyses present a scatterplot of all individuals in the network and primary roles in the network. For my correspondence analysis, I included degree, betweenness, and closeness to represent connectedness, bridging, and proximity respectively. These types of analyses indicate trends in the primary role of individuals in the social network. For example, those w ho play a bridge in the network will have their node gravitate towards betweenness in the plot. Figure 3 shows the analysis for the Business network. Most of the actors tended to gravitate towards areas signifying connectedness. These are reflective of hi gh degree centrality and eigenvectors. Given the high levels of centralization and cohesion, this is not surprising for this network. As formal USER network items of business are 57 Figure 2 . Spoke about Network Business Sociogram 58 rarely secretive it is logical that information flows relatively freely with few gatekeepers (i.e., low congregation of nodes around betweenness). The two nodes who are nearest to betweenness Figure 3 . Correspondence Analysis of Business Network are participants 10 and 15, meaning these individuals may serve conduits for information flows. Additionally, persons 2, 12, and 11 are situated on the edges of this representation and the associated sociogram (Figure 3 ). Whil e not closely linked to many of the members of this network, these individuals may represent structural holes or places unique information enters the business network. Problems network. Moving to a more specific network than conversing network busi ness, the second network findings pertained specifically to discussing network problems. This prompted participants to think about large - scale problems their formal USER network may be facing (e.g., sustainability, financial resources, membership concerns, etc.). Problems aided in understanding how actors within the network may collaborate or cooperate. The results for the 59 Figure 4 . Discussed Network Problems Sociogram 60 Problems network are in Table 4 and depicted in Figure 4. The measures for the Problems network are very similar to those in Business . The density for this network is 69.9%, which represented another high level of interconnectivity across the participants. Degree centralization was 34% indicating a moderate level of actor dominance within the network. Betweenness centralization (4.71%) and absence of fragments indicated there were few gatekeepers in the network, as actors who served as bridges were centered within the network. Problems did not yield any cliques or subgroup, but one large compact grouping of individuals . The average degree per participant was 11.176 (similar to 12 with the business network) with a range between 4 and 16, and a mode of 10. Findings for components (1), compactness (85%), and average geodesic distance (1.3) all represented a close, tightly l inked group of individuals. Moreover, the absence of any fragments also indicated there were few places of information bottlenecks or disparate actors. In sum, the problems network is only slightly less interconnected when compared to the business network, and still highly connected. With regards to individual node centrality and power in the Problems network, Participants 15, 10, and 13 had the highest number of ties and betweenness scores. Participant 11 had the lowest number of ties (4) and was on the p eriphery of the network in both their betweenness (0) and closeness ( 28 ) . Figure 5 depicts a correspondence analysis for the problems network. This analysis shows more actors gravitating to factors of closeness. Those who do not center on closeness are be ing pulled towards both degree and eigenvectors. Again, Participants 15 and 10 tracked closer to betweenness, however their role is not as pronounced as it was in the Business network. Given the prompt for the Problems network specifically asked about with whom these individuals discuss their network problems, the nodes gathered around closeness centrality were expected; as closeness is often expressed as a time of arrival in networks, these 61 findings suggest participants are closely and quickly sharing info rmation about their USER networks to a close - knit grouping of individuals. Figure 5 . Correspondence Analysis of Problems Network Gave Advice network . The GaveAdvice network was the only network considered to be directed. This network was not symmetrized like Business and Problems . As mentioned, the process of giving advice is typically considered to be one - way, therefore, the network analysis should align with how the information flows (Borgatti, Everett, & Johnson, 2018). Despite not altering the data in the cleaning process, partici pant - indicated reciprocity remained near 30%. GaveAdvice was considerably less centralized and cohesive than the other two. Network density and indicated a low de gree of individual participants dominating network flows. Network betweenness centralization (4.71%) indicated the network gatekeepers were centered in the 62 Figure 6 . Gave Advice to Others Sociogram 63 network and did not inhibit flows through the network. Interestingly, the average geodesic distance (1.39) is like the same statistic in Business . This may indicate a similar rate of flow through the network. The average number of advisees (out - degrees) was 6.5 with a range between 3 and 10 ; t he most common number of advisees w as 9. More practically speaking, each of these statistics indicate s advice was given throughout the network by many individuals in the network. The lower number of tie s and lower betweenness indicate that these individuals may act as a bottleneck for advice. Despite cohesion and centralization measures remaining moderately high, there are demonstrable differences between this network and the other two. Namely, the exis tence of seven components of the network resulting in semi - isolated groups within the network. Fragmentation in this network (.35) was moderately high indicating the existence of multiple dyads in this network that could be wholly disconnected if one indiv idual is removed. The results for the advice network are represented in Table 4 and depicted in Figure 6. Each of the measured centrality and power scores are considerably more distributed in the GaveAdvice network. Participant 15 had the highest number of out - degree ties at 10, while participants 13, 10, and 5 each had 9 ties. More interestingly, participants 2, 8, 9, 11, 12 and 15 did not report seeking advice from anyone in the network (in - degree), indicating these individuals may rely on others outside of the GaveAdvice network for advice. Most notably, Participant 15, one of the most central actors in the other networks, is on the periphery of the Gave Advice network. This p articipant did not indicate seeking advice from any other participant in the stu dy and may indicate a person who is either highly established in USER or an individual who is only partially involved in USER efforts. 64 communication and flow. Unlike the other networks, degrees do not play a large role in the network representing smaller groups of advice seekers. The disconnected dyadic tendencies are echoed with the fragmentation findings, meaning there are collections of individuals who seek and give advice relatively regularly, but primarily at the dyadic level. This reflects how advice is often given, through one - on - one conversations among peers. Interestingly, betweenness scores in the correspondence analysis are lo w. This also explains advice - giving relationships, as participants may not consider themselves to be thoroughfares of advice simply transferring advice from one end of the network to the other. Additionally, the relative absence of betweenness may indicate that advice seekers may not seek advice from those they do not know. Centralization and centrality discussion. Centralization and centrality scores primarily aid in the interpretation of whole network structure and individual structural functions within the network. The primary goal of using these metrics was to first understand the existence of Figure 7 . Correspondence Analysis of Advice - Giving Network 65 connections across independently nominated leaders working in formal USER networks and interpret the degree to which these individuals interact. The results suggest a high degree of connectivity across several social functions, and individual members take different roles in different contexts. Additionally, this network appears to capture only a fragment of a larger emergent inter - organizational network operating in undergraduate STEM education reform. Table 4 . Network Density and Centralization Scores Density Degree Centralization (%) Node with Highest Degree Centrality Between Centra l. (%) Node with Highest Betweenness C entrality ID Ties (Normalized) ID Ties (Normalized ) Business 0.75 28.3 15, 10 16 (1) 3.07 15, 10 15.468 (.046) 5 6 15 (.938) Problems 0.699 34.2 15 16 (1) 4.71 15 7.726 (.064) 13, 10, 5 15 (.938) 10 6.267 (.052) 3 5.66 (.047) Gave Advice 0.408 24.6 16 10 (.625) 5.36 14 16.17 (.067) 5, 10, 13, 15 9 (.563) 10 14.38 (.6) Emergent inter - organizational network structure . Centralization and cohesion measures were used specifically to assess the emergent network structure. Density, average degree of ties, and degree centralization provided a basic statistic for interpreting the level o f interconnectedness. Despite the participants independent and anonymous nominations, each of the networks presented a high level of connectivity. These connections are especially evident in the Business and Problems networks. The low - stakes of the Busine ss network gave evidence of individuals freely discussing and sharing the on - goings of their individual networks, while the Problems network provides a foundation for understanding how the network may operate to tackle large - scale problems in USER . The Gav eAdvice network takes another step in understanding how this group is internally dependent. The densities of these three networks offer insights into the lived emergent network structure by indicating the participants discuss their 66 formal USER networks, th eir associated problems freely, and seek advice from those within this study. The magnitude to which a network was dominated by a small group of individuals was measured by average degree, degree centralization, and betweenness centralization. Average degree takes the mean number of ties for all members in the network, excluding any outliers who may dominate the network. Degree centralization highlights multiple factors for individuals centrality in the network, where high degre e centralization indicates a network connected by a small group of actors, the inverse is more evenly spread. High degree centralization is often associate d with generating new information (Borgatti, Everett, & Johnson, 2018). Knoke and Yang (2008) stated inter - organizational groups should strive for moderate centralizations to maximize both benefits, although they grant context always matters in network ana lysis. Average degrees fell in the high range, while degree centralization straddled moderate and moderately low. Degree centralization in the Business network seems reasonable considering the diffuseness and availability of network agendas, and the number of formal USER networks represented. The low degree centralization in Business may indicate a more widespread discussion of network business items and similar agendas across multiple formal USER networks. Similarly, the moderate level of degree centrality within the Problems network skews more to relying on trust and relationships to address network problems . This suggests a more familiar or trusting relationship across these individuals when trying to address shortcomings or challenges with individual net works. The lowest degree centralization in the GaveAdvice is sensible as this network is less dense and built upon directed information acquisition. The betweenness centralization (i.e., the degree to which a few individuals control the relationships 67 of ot her in the network) was also extremely low in all of the networks. This indicated that even in the least dense network, information could flow relatively free of gatekeepers. When taken together, density, degree centralization, and betweenness centralizati on provide insights into how this group is socially structured. In general, these individuals know each other beyond simple acquaintances, and rely on one another for cooperation, problem - solving, disseminating information and general advice to support their formal USER networks. Although the full contexts of the relat ionships remain unclear, the prompts in which the participants responded gives some definition to how the relationships function. The absence of gatekeepers within the networks point to a collection of colleagues who operate as peers to generate or (re)int erpret knowledge for the benefit of formal networks. This web of emergent networked leaders allows information and knowledge access to formal USER networks even if Dynamism of emergent network actors . Centrality scores were used to statistically describe the varying roles of individuals within each of the networks. Following the theoretical interpretations outlined in Chapter 2, both high and low centrality scores have benefits to the network. Whereas high scores may be associated with trust, relationships, and cooperation, low scores also provide the opportunity for more information or new thinking to enter the network . Participants highlighted in this section were selected by calculat ing the descriptive statistics for each of the centrality measures in each network. Individuals whose scores were consistently near one standard deviation above, below or at mean were selected for more investigation. Participant 15 is a pr i me example of a highly connected and potentially influential person in this collection. They have high degrees, eigenvectors, and low closeness scores in all three networks. This participant has the highest betweenness scores of all participants in both Business 68 and Problems , pointing to a common person who is likely involved in many conversations across formal user networks. These statistics indicate a critical core member of the network, with one notable exception. Participant 15 does not rely on advice from any particular individual in the sample. Their betweenness score and in - degree for the advice networks are both zero. This is not to posit that Participant 15 does not seek advice, but rather they likely rely on connections in other areas not represe nted in the study. The centrality scores and inferences for Participant 15 are supported by the demographic information provided by this participant . Participant 15 reported the second - longest tenure working in STEM undergraduate education reform. Particip ant 15 also indicated membership in 5 central formal user networks, and actually provided a follow up message including additional networks in which they belonged. Although none of the follow - up networks were included in the dataset, Participant 15 is high ly involved in USER and seemingly highly enthusiastic about the work. Similar to the previous individual, Participant 10 was also a highly connected and influential person in the networks studied. This person had high scores in degrees, eigenvectors, and betweenness with low scores in closeness, which indicated an individual who is likely part of many conversation across multiple networks simultaneously in multiple capacities. They likely discuss network business, problems, and giv e advice to many members of this sample. The critical difference between participant 10 and 15 is that the former is more central in the advice network. This indicates someone who is embedded more fully across networks who is giving and seeking advice from individuals who work in formal USER networks. The demographic information supports this claim, with this individual working as an administrator for a non - profit who has been working for a considerable time in undergraduate STEM reform. This person 69 likely is immersed through emplo yment to gather and diffuse information and advice through the network. Participant 13 averaged close to the mean centrality scores for each of the 9 vectors studied. This person was moderately connected to others and acted as a thoroughfare for information to certain individuals. This is particularly true for the business and advice networks. Interestingly, Participant 13 centrality scores were above the average for the problem network, indicating that while this person was loosely connected in o ther networks, they were often part of the conversations regarding network problems, or how formal networks can address these problems. These roles are somewhat supported by the demographic data provided by Participant 13. This individual highlighted they were employed by a non - profit as a staff member of a concerns with one specific network, but not necessarily disseminate those concerns to the network. Additionally , Participant 13 identified they were relatively new compared to other members working in undergraduate STEM reform. Their connections to others may still be forming as a result of their newcomer status. Participant 11 presents an individual who is only lo osely connected to the rest of the group. The degrees, eigenvectors scores and betweenness scores for this individual were low for all three of the networks studied. Despite lying in the periphery of the network, this individual was still connected to infl uential individuals within the network including P articipants 10 and 15. This person also provided advice to other influential members of the network. The demographic information further confirms and contextualizes the role of this individual in the networ k. This person indicated they served in a long - time administrative position for a formal USER network 70 but has recently retired. They also reported the longest amount of time serving in STEM undergraduate reform of any of the participants in this study. Finally, Participant 12 is another case of someone situated on the periphery of the networks. This individual scored low on all the centrality measures in the business network. Low scores in this network may indicate they are not necessarily engaged in cro ss - network discussions and may not have the opportunity to share. Interestingly, participant 12 scored closer to the mean in the other networks, indicating someone who may be referenced or consulted to aid in cross - network meaning - making or planning for pr oblem - solving. Inter - Organizational Network Properties Just as critical to the findings of the social network analysis of the individuals nominated for this research was the study of their organizational interconnectedness. While fascinating, a simple stud y of leaders working in USER only captures their personal networks. In order to investigate the existence of an inter - organizational network among formal USER networks, I compiled the participant - provided data on their network affiliations in a 2 - mode adja cency matrix. I used these data to draw centralization and cohesion measures, and centrality scores. This Network of Networks consists of formal USER networks as the nodes, and the ties represent shared memberships. Table 5 shows the centralization and coh esion scores for the user networks. Figure 8 depicts a visual representation of the reported formal USER networks in this study with each tie weighted by how many individuals share the affiliation. The Network of Networks demonstrated a moderate degree of interconnectivity with a density of about 37% and an average degree of 6.7 ties. These centralization scores indicated the formal USER networks share a considerable amount of the participants in the study, and on average are connected to almost seven other formal USER networks. The degree 71 Table 5 . Network of Networks Density and Centralization Scores Density Degree Central . (%) Node with Highest Degree Centrality Between Central . (%) Node with Highest Betweenness centrality ID Ties * (Normalized) ID Ties* (Normalized) Network of Networks 0.374 51.3% ASCN 21 (2.11) 21.22% ASCN 37.28 (0.24) BVA 29 (1.61) BVA 29.33 (0.19) NASEM 26 (1.44) PKAL 21.03 (0.14) NSEC 21 (1.16) SERC 13.03 (0.09) *Note individuals were polled to list up to 5 networks in which they belong. centralization and betweenness centralization measures are high and moderate (respectively) for a network of this size, indicating there are several nodes which dominate the network . The nodes comprising the center of the network are also represented in Table 4 as having the highest degrees and betweenness scores. Contrary to the other networks , the ties in the Network of Networks are people (not social relationships), therefore certain networks appear to convene many of the leaders working in undergraduate STEM reform networks. As many of the participants in the study indicated membership in se veral formal USER networks, they will appear in multiple ties in this network. This core of formal USER networks is easily identified in the sociographic model of the network. Figure 8 displays the Network of Networks. Each of the ties in the model are wei ghted by the number of individuals with shared memberships in bot h he darker the line the more participants in the tie) . The formal USER networks listed in Table 5 are both centrally located in the model and share several weighted lines with each other. This reflects both their high degree and betweenness centrality scores. Both the breadth and cohesiveness of formal USER networks repr esented in this study establish a suitable foundation of organizational connectedness to assert the potential existence of an emergent inter - organizational network . 72 Figure 8 . N etwork of Networks Sociogram 73 Dyadic Properties Moving beyond discussions of network structure and individual function conveyed in centralization and centrality, I explored the individual relationships embedded in the network. Investigating characteristics of the dyads provided insight into why individuals may have sought one another and identif ied patterns unrecognized by cohesion measures. One method used to analyze these patterns is multiple regression quadratic assignment procedure (QAP) models. This technique uses data permutations to create an expected model using ordinary least squares with a basi s by comparing each observation by the possibility of it being observed across all participants. These models identify homophilic and heterophilic tendencies within the socio - metric data. Homophiles are defined as the phenomena where individuals tend to gr avitate to those with shared identities, and heterophilies are defined as the opposite (Borgatti, Everett, & Johnson, 2018). Given the literature discussion on emergent inter - organi z ational networks use of homophily in identity and heterophily in ideas as a guide, I used QAP modeling to explore patterns of relationships in the Business, Problems, and GaveAdvice networks. Each of the different computations used all of the categorical, socio - metric, and continuous data provided by the participants to predict the presence of a tie between any two randomly selected participants. Included in each of these computations were covariates for the other two networks. Results for the Business QAP regression are displayed in Table 6. Independent variable descriptions are located in Appendix D. The results from the model show several factors predicting the presence of a tie in Business . First, the adjusted R - squared indicates a moderately high model fit, accounting for 74 almost half of the total variation in the data. Signi ficant covariates in the Business network included primary organization affiliation, similar formal USER networks memberships, and working in higher education were significant in predicting a relationship. Within the significant variables, primary organization affiliation (e.g., research university, non - profit, etc.) had an inverse correlation coefficient, indicating individuals were likely to have a t ie with an individual who did not share a similar primary organization type. A dyad was predicted by sharing a formal USER network membership with another participant. These relationships were Table 6 . Business Network Multiple QAP Regression Model R 2 Adjusted R 2 Obs Perms 0.50222 0.48274 240 2000 REGRESSION COEFFICIENTS Coefficient p - value As Large As Small Perm Avg Std Err Ethnicity 0.02434 0.72414 0.38131 0.61919 0.00034 0.07133 Age Range 0.05264 0.37031 0.18941 0.81109 0.00062 0.06035 Gender 0.05633 0.22589 0.10395 0.89655 - 0.00049 0.04857 Primary Organization - 0.13501 0.0484** 0.97401 0.02649 - 0.00386 0.07287 Position - 0.05004 0.49325 0.76512 0.23538 0.00149 0.07618 STEM tenure 0.132 0.36132 0.17441 0.82609 - 0.00151 0.15748 Same USER Network 0.04391 0.09444* 0.04948 0.95102 - 0.00132 0.02956 Gave - Advice Network 0.07157 0.09645* 0.05597 0.94453 - 0.00009 0.04542 Problems Network 0.54256 0.0005*** 0.0005 1 - 0.00253 0.07516 75 certainly, expected in Business business with one another regularly. I followed similar procedures for modeling the Problems network. This analysis switched the dependent variable ( Business ) with the covariate ( Problems ) and re - ran the model. All other covariates remained the same. This model predicts the presence of a tie between two actors who discuss major network problems (i.e., financial stability, sustainability, memberships, etc.) Results for the Problems network model are displayed in Table 7. Table 7 . Problems Network Multiple QAP Regression Model R 2 Adjusted R 2 Obs Perms 0.50259 0.48312 240 2000 REGRESSION COEFFICIENTS Coefficient p - value As Large As Small Perm Avg Std Err Ethnicity - 0.06195 0.32834 0.85657 0.14393 0.00469 0.06737 Age Range - 0.02004 0.72114 0.65717 0.34333 0.00143 0.0611 Gender - 0.02772 0.58621 0.69615 0.30435 - 0.00156 0.05302 Primary Organization 0.02518 0.72664 0.37431 0.62619 0.00028 0.0766 Position 0.06733 0.36832 0.18741 0.81309 0.00104 0.07916 STEM tenure - 0.20259 0.0909* 0.89605 0.10445 0.00187 0.17291 Same USER Network 0.07493 0.0045*** 0.003 0.9975 0.0013 0.02979 Business Network 0.61191 0.0005*** 0.0005 1 0.00148 0.08832 Gave - Advice Network 0.12217 0.01*** 0.0055 0.995 - 0.00077 0.04972 76 This model demonstrates similar findings to the business network. The adjusted R - squared accounts for a similar amount of variation in the data. Social identities, primary organization affiliation and position within the organization were all found not to be significant. Having a relationship in the other networks was significant and a predictor of a tie. Time spent working in STEM (STEM tenure) was a notable difference in the Problems network. The negative correlation again signifies heterophily within thi s network, predicting ties among dyads with different experience levels. Although this variable is only significant at the 10% interval, it is markedly different than the outputs for the Business network. This heterophily indicated the possibility for more cooperation, as formal network problem solving was not resolved for a specific group of individuals. Shared network membership also remained significant in this network, and the correlation coefficient increased indicating a larger prevalence of predictin g a tie in Problems. The final QAP regression modeled data was for the GaveAdvice network. The same process for calculating the other models was used for this regression. The results for the GaveAdvice model are shown in Table 8. These findings predict th e presence of an advice - giving relationship within the group of participants and again only track the existence of a directed relationship. As this network is unique to the other two, the regression model is distinct. Accounting for only 15% of the total v ariation, the adjusted R - square indicated there are many unspecified variables which account for an advice - seeking relationship. These factors are unobserved in this dataset. Social identities, STEM tenure and primary organization affiliation were not significant in predicting an advising relationship. The exist ence of a relationship in the Business and Problems network was significant in predicting an advising relationship, which is 77 Table 8 . Gave Advice Network Multiple QAP Regression Model R 2 Adjusted R 2 Obs Perms 0.18497 0.15307 240 2000 REGRESSION COEFFICIENTS Coefficient p - value As Large As Small Perm Avg Std Err Ethnicity 0.06257 0.44978 0.21989 0.78061 - 0.00556 0.09875 Age Range 0.09513 0.1949 0.09595 0.90455 - 0.00027 0.07911 Gender - 0.02267 0.73513 0.65117 0.34933 0.00148 0.07109 Primary Organization - 0.06195 0.49575 0.75862 0.24188 0.00346 0.09738 Position 0.16441 0.09845* 0.04698 0.95352 - 0.0023 0.10462 STEM Tenure 0.34255 0.08346* 0.04348 0.95702 - 0.00137 0.20747 Same USER Network 0.03696 0.32734 0.14643 0.85407 - 0.00234 0.0406 Business Network 0.1649 0.09895* 0.04448 0.95602 - 0.00134 0.10605 Problem Network 0.2496 0.01099*** 0.0035 0.997 - 0.00391 0.10387 logical considering the familiarity needed for an advising relationship. The role of STEM tenure increased, and its coefficient flipped from negative to positive when compared to Problems . These statistics point t o people seeking advice from peers who have with worked for a similar amount of time in reform. Institutional position also had a significant homophilic role in this network, where individuals often gave advice to those in similar position. This may be imp actful in institutional diffusion. Network Dyadic Properties Discussion The QAP modeling primarily aided in understanding individual participant preferences within the larger network discussion. Whereas centralization and centrality measures relied on 78 defi ning network structure, boundaries, and potential gatekeepers, these models begin a preliminary investigation into how the emergent network is operationalized. While the regressions are limited in their ability to quantify the lived experience of individua ls in the network, they do provide insights into group tendencies and, more importantly, what should be further probed at later stages. The following discussion highlights four findings from a cross - models comparison and underscores these findings with bot h limitations and further questions. Complexity in relationships . As social interactions drive relationships, the oversized role of the social networks predicting relationships is not surprising. Each social network used as an independent variable in the models was significant in predicting a tie and typically carried th e largest correlation coefficients within the networks. Not only do these findings indicate that the presence of a tie in one network predicts a tie another USER social network, but also relationships are one of the best indicators in the dataset. Their si gnificance moves the interpretation of the centralization measures forward by using all three networks simultaneously in a model to predict network structures based on relationships in other domains. Their significance also begins a foundational understand ing to the depth of relationships shared by individuals within the networks. As mentioned in earlier discussions, each network pertains to a different domain of formal USER networks. These models indicate the layering of the networks and that individuals r ely on their existing relationships across the different domains; specifically, the social interactions are highly correlated between Business and Problems . This points to similar people discussing network business in tandem with inherent problems with net works may occur more frequently. These relationships (and subsequent conversations) occur both within and between formal USER networks as evidenced by variable Same Network Affiliation differences in significance between Business and Problems. 79 The lower correlation among GaveAdvice and the other two surveyed networks was also expected. As advice - seeking behavior is typically not at random, its correlation is much smaller than those with whom one simply shares USER network business. Lower coeffic ients point to individuals speaking broadly about business, and problems, but only giving to advice to few within the group. More importantly, these correlations also indicate a hierarchical relationship among the dyads, where 1) individuals speak to many individuals in the business network, 2) speak to a smaller subset of the same individuals within the problems network, and 3) provide advice to an even smaller subset of those individuals. The relative correlations are also supported by the descending dens ity scores of each network outlined in the centralization discussion. Organizationally focused dissemination . Beyond the discussions of layered relationships within the network, the models demonstrate divergent foci in their respective network domains. In particular, the Business network depicted dyadic patterns consistent with organizational affiliations ; t hese patterns were not observed in the other two networks surveyed. Significant variables in Same STEM Network Affiliations and Primary Organizations indicated a pattern of information sharing across organizational boundaries. The negative coefficient with Pri mary Organization denoted a heterophilic tendency of participants to share formal USER network business with those who belonged to organizations different than their home institution. Given the diversity of organizations represented in this study, this fin ding shed light on dyads existing despite organizational boundaries and sectors (i.e., research university, liberal arts colleges etc.). These relationship patterns are especially salient considering the role of non - profit organizations in the funding, dir ection, and operation of formal networks. Note this finding only includes a select number of organization types. No community colleges or for - profit universities 80 were represented in this study. This finding suggests a low barrier for organizations which ha ve individuals highly involved in USER but sheds little light on the barriers for other institutions. The variable Same STEM Network Affiliation was also significant in Business and has implications for how this network influences its organizational affi significance indicated homophily in participant dyads when they share at least one network in common. This is a logical interpretation, as one expects the participants to discuss their network business with others within the forma l USER network. However , no two participants share the exact same formal USER network affiliations. These intra - formal USER network discussions may serve as a base point for information diffusion to other networks as members talk about their business. More investigation is needed in this specific area to fully understand how information flows through dyads within the same network to divergent formal networks. More thorough investigation is needed in this area to fully understand how information is flowing t hrough this network. The level of organizational involvement is unique to the Business network. As the prompt may express a more organizational diffusion purpose to this network. This is accentuated by the near collinearity of Position and Primary Org anization (where one would expect both variables to be either significant or not significant) yet the latter is significant while the former is not. In essence, participants involved in the study may share their formal USER network business with those in s imilar college s or universities and not with those who share a similar organizational title. This finding supports literature indicating tribalism among higher education sectors. Individually focused collaborations . Moving from the organizational themes in Business , the other two networks reflect more individual - oriented dyads. This may reflect a more 81 collaborative focus of both Problems and GaveAdvice. Whereas Business sits in a domain of basic information sharing (e.g., spoke about network business), these networks front a more intentional request for information or general collaboration. Both also require a higher level of involvement from the reciprocating individual . The different characteristics in the latter two networks portray different findings and raise additional questions. First and foremost, the null findings in the Problems network, particularly with Position, depict a group of participants who openly inte ract regardless of institution type or position held. Granted, these models do not fully account for each various human capital. The null finding supports the network learning and innovations functions outlined in the literature on emergent ne tworks. Critical to both of those functions is an open and collaborative grouping of people. Within Problems the heterophily in STEM tenure may indicate a collaborative behavior of individuals in the network as they engage with members of different experie nce levels to address problems within their networks. While this finding reinforces the concepts of collaboration within the group of participants, it contextualized the information sharing as dependent on time within STEM. This may mean individuals newer to STEM may bring fresh ideas into this group for discussion across experience levels. It suggests more experienced individuals sharing strategies or techniques for working with problems with less experienced individuals. The latter of the two explanations may also point to a certain level of socialization within this group. Given these are all nominated leaders in formal USER networks, the socialization may posit an orientation to cross - network work (and not an introduction to formal USER networks themselv es). Not significant social identities . None of the social identities collected demonstrated a capacity to predict homophily or heterophily. There are multiple interpretations to these findings. 82 The first interpretation is that social identities play no role in connecting individuals working in USER. While the outputs of the model support this interpretation, this is likely not the case. More reasonably, there is a considerable lack of diversity i n the participants nominated. Almost all the participants identified as white, and a super majority identified as women. Almost half of the sample indicated they were in the 50s. This lack of variation in the data most certainly limits the interpretation of the role of social identities in dyads. While this lack of diversity does cause disruption in the interpretation of the findings, it does reflect the lack of diversity of those working in USER. For example, c ountless articles ( e.g., Kezar & Holcombe, 20 19; Mack, 2019) point to the prevailing whiteness in USER, and Kezar and Gehrke (2015) highlighted the burden of women in undergraduate STEM education reform. Dyadic modeling limitations . The significant patterns in relationship seeking identified in eac h of the models are supported by the emergent network literature in both organizational dissemination and interpersonal collaborations. As denoted in Chapter 2, diffusion, network learning, and innovation were the major functions of these networks (Isett e t al., 2011; Popp et al, 2014). These findings, nevertheless, fall considerably short in three major areas. First, the patterns unveil nothing of human capital, influence or intentionality within the network of participants. Simply stating the patterns exi st is a critical first step, but they do not confer Second, the models do not account for information flows. Despite the directness of the social network survey p about what is shared , why is it shared, and how is it shared all remain unresolved by the models. Third, multiple independent co - variates remain unexplored in the error. Thi s is particularly salient in the GaveAdvice network, where only 16% of the total variation was explained. 83 Conclusion With regards to the research questions, the social network analysis provides some answers, and aids in the focusing of others. In conside ration of research question 1a) how interconnected are leaders across formal network, the data demonstrate a high degree of interconnectedness across multiple domains. These disparately nominated individuals know almost every other nominated individual and communicate regularly. Moreover, the density of these networks also provides insights into the emergent network structure. Based on the social network data, few gatekeepers to information exist within the network. In consideration of research questions 1b), 1c) and 2, the social network analysis demonstrates the opportunity for kn owledge diffusion, network learning, and innovations to occur, but delivers little in fully realizing these processes. Although one may assume these emergent networks are occurring, these data lack depth and clarity to many of the nuances and subtleties of emergent network activities. The following chapter discusses the findings from the qualitative inquiry into the lived roles of the participants in this study and uncovers more information on how the emergent network is operationalized in formal USER netwo rk efforts. 84 Chapter 5: Interview Findings The second strand of the study gathered and analyzed more contextualized data on the roles of the relationship web established in the social network survey. Conducted as a separate strand of the study, the interview data provide insights on how relationships were actualized among the participants and how their collective actions contributed to this group serving as an emergent interorganizational network. Although the data from the social network strand were available at the time of qualitative analysis, this chapter only speaks to qualitative findings which In designing a mixed - methods study Creswell and Plano - Clark (2011) stated one of the techniques, emphasis o address the research questions. Granting a strand emphasis in a study also acknowledges the findings as being more substantial in explaining the phenomena. In Chapter 3, the interviews strand of the research design was highlighted as the more dominant ( i.e., emphasized) aspect of the study. As the interactions and impacts of emergent networks are at the core of my research questions, the qualitative data collected in this strand were better suited to understand these phenomena. While social network desig n (the first strand of the study) was helpful in illustrating the connectivity of individuals in the network, it did not carry emphasis in the study as it is not methodologically suited to answer the majority of the research questions . As highlighted in the conclusion of Chapter 4, social network techniques were greatly limited in their ability to interrogate experiences, knowledge exchanges, and depth of relationships. Participants were selected for interview s by their standardized composite centrality measures collected in the social network strand. Two participants from those with the lowest, 85 average, and highest standardized composite centrality scores were interviewed, resulting in a total of six interviews. The following chapter outlines the findings and responses from the individual interviews and does not include many direct quotes from participants. The interview protocol asked participants to discuss their actions within a network, their network membership(s), employment, and social circles related to USER. Given the questions high degree of specificity and relatively small number of individuals working as leaders of networks, direct quotes were not an ideal method for sharing findings and presented unforeseen challenges. If quotes were inserted as spoken, they would often belie confidentiality through the participants network membership, employment, or social connections. If pseudonyms were used however, the quotes would lose their intended effect and muddle the point. In th e end, quotes were only used when redaction was not necessary or the quote was imperative to the discussion. The contents of this chapter include: (1) an overview of each interviewee; (2) larger cross - interviewee trends that support how the participants wo rk collectively; and (3) a discussion of how their collaboration serves as an emergent USER network. Participant Profiles In working with qualitative data in mixed - methods designs, Creswell and Plano - Clark (2011) stressed using participant profiles to pre serve the human aspects of the study. Arguing data - mixing often muddles or dehumanizes qualitative data, participant profiles provide audiences with a window through which to see participants (Creswell & Plano - Clark, 2011). In the spirit of preserving the human aspects of the study , I provide thematic profiles of each pseudonym and randomly generated gender pronouns. These steps were taken to further protect participant confidentiali ty. 86 Each of the profiles presents participant experiences as related to three functions of an emergent interorganizational network. Outlined in the literature review, the functions of an emergent network are centered on major actions members take in servi ng their convening purpose; these functions are knowledge diffusion, network learning, and innovation (Popp, et al., 2014). Knowledge diffusion referred to the ability of an emergent network to share information both within and outside the emergent network . Network learning was defined as a meaning - making process where emergent network members take lessons from their respective organizations to draw larger field - based perspectives. Innovations referred to outputs or contingencies the emergent network create organizations. Specific to undergraduate STEM education reform (USER) efforts, each profile throughout the network. The pro other collaborative actions and conclude with a discussion o f their role in creating new strategies or initiatives in USER. Each profile is presented alphabetically (by pseudonym with a random ly assigned gender), and the ordering is unimportant to the findings. Allie. Allie is the longest serving USER - related professional in the study. Despite not having a STEM - related degree, Allie founded an influential formal USER network. Over several decad es Allie has received high profile grants and other funding. Other participants described her as an avid consumer of formal network reports and scholarly articles in STEM education. One round one of my papers. Although Allie has remained in close connection with those working in USER, she does not actively participate in formal USER networks. 87 Knowl edge definition . As Allie has left the USER network space, most of her discussions on knowledge and transfer were either framed in the distant past or in the context of a few current interactions with professionals in USER with which she spoke. Regarding t he concepts of knowledge, Allie primarily discussed organizational outputs like grant reports, research, official national association or foundation publications, and other literature pertaining to USER. The documents contained findings she believed useful in the context of USER. This knowledge originated from large, established organizations like colleges and universities, and the National Science Foundation. She stated these documents contained valuable findings on the state of USER, formal USER network s uccesses, funding organization priorities, and new potential space (Allie). During her interview she referenced seventeen separate reports pertaining to USER dating from 1988 to 2018. Most of the documents she referenced were annual National Science Foundation (NSF) grant reports of formal networks operating in USER. Similarly, Allie spoke of formal USER network operations and outputs (e.g., research, publicat ions, press releases, grant reports) as a knowledge base. She believed her work in USER helped to shape conversations and funders were reporting back to me and say ing how often others were citing our efforts and events Aside from tangible network products, Allie also indicated knowledge generated from operating and leading a formal - nning a nationally - funded network and included logistical knowledge on how to file grant reports, 88 Allie indicated her experience in running a network was highly sought after by those in USER. According to Allie, the knowledge of how to run a formal network provided her the opportunity to interact with people across STEM disci plines. She was often approached by leaders in USER (primarily in physics, mathematics, and chemistry) for advice on how to form a network. She attributed her work with those USER leaders as critical to her success, as she learned discipline - based language , norms, and behaviors, which provided the necessary buy - in from others Knowing different STEM discipline norms and behaviors aided her to cross discipline boundaries and speak to a broader field of leaders in USER. She felt ill - equipped without the knowledge on how to speak to others and said reform work was pointless until Network learning . Allie spoke directly to how she believed other participants in the The Hobbit, le This metaphor fits with the network learning definition of sharing organizational information with the network for greater understanding of field - level problems. Filtering information out of t he network to the organization, then back to the network is a critical aspect of network learning as it provides content for the meaning - making processes of learning. Allie highlighted consensus building and coalition building as two critical aspects of ne twork learning. For consensus building Allie lamented the lack of agreement across STEM 89 academic disciplines and indicated the work of networks was to build consensus on new norms, le coming to a (Allie). In short, consensus building was necessary for the different people in USER to speak to one another. Second, Allie argued coalition building was also important to network learning, stating that learning needed passionate people with different approaches. She was adamant she group for network learning initiate change. Innovation . Allie spoke of several different products derived from her work with other f the innovations she discussed were directly linked to the formal USER network she helped to found. These innovations included additional grant proposals both related and unrelated to her USER network, various whitepapers associated with the state of USER , and collaborative workshops. Each of these different products were in association with, or inspired by, another individual within the study. The proposals written in conjunction with others in the study were typically directed to funding organizations wh herself to be a consumer, not producer of scholarship. She did not attempt to produce any original research in USER. Andrew. Andrew is a STEM faculty member working at a research institution. He received a STEM - related terminal degree and cited two other non - interviewed participants as student (Andrew). 90 Andrew is also a founder of a formal USER network and serves as a primary investigator on multiple USER grants. This participant considers himself to be heavily involved in USER. In addition to his work as a tenured faculty and researche r, he also assists with operating a Center for STEM Education improvement on his home campus. Andrew indicated a significant amount formal USER networks. Knowledge definition . Similar to Allie, Andrew primarily centered his definition of USER - related knowledge to his work with a formal network in USER. For him, knowledge was represented in three ways. First, Andrew described knowledge as STEM education expertise STEM education. He clarified by indicating expertise was related to evidence - based curricula and pedagogy, and the point of USER was to share that to share exper tise. We need some mechanism for sharing, because the disparate approach does enough to leverage change, and that expertise on how pedagogical change could occur was more difficult and precious as there was no single universal strategy. Second, Andrew described a body of knowledge related to formal networks. Again, described as expertise, this area focused on the establishing, funding, governing, and sustaining formal netw orks in USER. As a trained STEM scientist, Andrew articulated a lack of confidence in his ability to run a formal network without consulting others with experience in networks. ith multiple - connected formal USER network. Well I mean [there are several people] who are on just about everyt hing in STEM reform. They have a lot of experience doing a lot of different things in 91 understand the possibilities, how you would frame things, and that kind of stuff. (Andrew) For these particular contextual knowledge holders, Andrew stressed the importance of their multiple USER - related affiliations as a driver for his formal USER network consultations or even inclusion in his USER network. He indicated those with the contextua l knowledge in USER drastically help improve the network. He identified four individuals with whom he sought advice regarding formal networks. Andrew further clarified by indicating the relationships he had with those individuals provided a window into the operation of different formal networks and helped Andrew had with the four individuals he cited pre - existed any formal network affiliation. Andrew knew these people prior to formal networks, which he stated, drove him to contact them for advice. Third, Andrew indicated personal information was a knowledge form across some but not all individuals. This included knowing others personally, their backgrounds, hobbies, interests, families, and politics. Andrew said knowing personal information about others working in USER sharing personal relevant information helped him trust others within the space aiding in him sharing potential pitfalls or challenges within his formal USER network. Andrew also indicated several of the other participants as long - time information. Knowledge exchange . Andrew cited multiple venues and mediums where he believed people shared their knowledge with others. Not surprisingly, Andrew indicated that personal 92 relationsh ips were the single greatest asset in formal USER networks. If he was friends with someone, he would simply call them and ask his question. He quickly identified three people within the study whom he would call to ask a question. Outside of personal relati onships, Andrew underscored the role of small conferences as a place to share knowledge. He was adamant that small science education conferences played the largest role in providing a venue for people to connect and share information pertaining to USER. Th e small size of the conference was very important as its purpose often felt more targeted to Andrew. He specifically spoke of a conference that he helped organize which directly led to one of the first convenings of the formal pulled together a small conference of change strategists in higher education here in town, we had something like 40 people come with some great speakers. president, c hair, or director) were also highlighted as a venue where formal network knowledge and expertise was shared with others. These leadership roles often took the form of primary investigator, network elected official, hired network consultant, or membership o n steering committees or advisory boards. According to Andrew, sharing knowledge was the point of these rely on one another for ideas because we cannot tackle th Coalition building . When prompted about network learning, Andrew spoke more about how he worked to involve critical stakeholders in developing or growing formal networks. For Andrew, building relationships with others predated their wo rk in formal networks, but was the cornerstone to their work now. Andrew outlined how and where he met many of the other study participants at various USER - passion for STEM education formed a strong relationship in opposition to the complex problem 93 of institutional reforms. Andrew said over time, the budding conference pals became partnerships, co - PIs, and advisors with whom he could consult for ini ti atives or problems. If he needed advice for a network, he could consult some of his old conference pals for approaches to the problems Interestingly, many of the conference pals shared similar network memberships. Conference pals also could also be a benefit when applying for funding from USER gran tors. As a trained disciplinary scientist, Andrew explained he did not get much exposure to other USER - minded professionals in other disciplines. Working with his conference pals, he was able to build a case for funding an inter - disciplinary team which hel ped his chances of receiving level, and what is the national ac disciplinary areas (e.g., chemistry, biology, etc.) to increase his chances of funding he would check - in with his colleagues and invite some to join his initiative. Innovation. Andrew spoke prima rily about using innovation as a way to monetize and support existing formal networks. Focusing on network longevity (especially post - grant - funding) he spoke about ways to monetize formal network by selling services to colleges, universities, or other post secondary affiliated organizations. Included in this discussion was the possibility of selling some network products to other formal USER networks. This could be research on network effectiveness or even network organizational consulting. Each of these are The strategies discussed were presented as way to maintain existing formal networks and continue their work as non - profits advocating for USER. 94 things, what would we sell and how much would we need to sell them for in order for it to make He indic ated these were the questions he would ask members of his steering committee and other peers he sought for advice. Hilarie. Hilarie currently works for a national higher education organization and works with multiple organizations and postsecondary institu tions in USER. She described her current interfaces with many faculties, administrators, policy - makers, and higher education advocates and believes her position in the U SER landscape uniquely situates her to serve many communities advocating for change. Hilarie serves as a permanent or proxy member on several formal USER network steering committees, and often attends various formal network meetings to better understand cu rrent practices in network functions. Knowledge definition. Hilarie directly linked knowledge to specific formal network practices, organizational structures, and bylaws. Pragmatically, she discussed knowledge as a resource which could be used by networks to increase their chances of getting funding or avoid certain organizational pitfalls. In an example, Hilarie cit ed her work in drafting a formal network proposal. In creating the proposal, she borrowed bylaws, theories of change, or network governance structures from several other network proposals to build a case for why the network should be funded. Citing an esta blished funding history, Hilarie argued borrowing these practices was intentional as she needed to build a persuasive case for NSF funding. Using the precedent of previously funded projects allowed her to focus on other substantive aspects of forming a for mal network. Hilarie specifically identified three other formal USER network leaders whom she contacted in the construction of her network. These three individuals were also participants in the study. 95 Hilaire also highlighted her position at a national hig her education association gave her perspective on the status and situations of formal networks. She serves on many network leadership teams and has access to many closed - door conversations about contemporary network concerns. She felt she was aware of the successes and current shortcomings of many formal networks. She believed this form of knowledge was critical to her role in USER in two ways. First, she thought it made her a better network leader because she could draw on many different experiences to ad dress network concerns. Second, she credited her knowledge of other networks their Knowledge exchange . played a role in how she thought about sharing her expertise. Although the mission of her employer was to work with institutions of higher education, Hilarie felt she could interface with many individuals in USER. This sentiment carried over to her work with formal USER networks too, as many institutions she visited were also affiliated with USER network projects or initiatives. USER, and to share widely across all those working in USER (Hilarie). Serving as part ambassador and cheerleader, Hilarie saw herself as a conduit for sharing information about networks as she completed site visits for member organizations of her employer. Consens us building . Hilarie specifically mentioned consensus building as a necessary task for people working in USER. As a collection of cross - discipline faculty at different institution types across the country, she explained the diversity within USER led to a m ultitude 96 of definitions, processes, and ideas . In order for progress to be made in STEM reform, Hilarie indicated leaders needed to be able to understand one another. The ability of people working in USER to come to mutual agreements on those definitions w as critical to USER success, foundation part of all collaborative work in USER, and she believed it was the best strategy for moving beyond the different organiz ing silos (e.g., discipline, institutional, geographic region, etc.). Although she indicated the consensus building process permeated all aspects of USER, Hilarie emphasized the importance of consensus in formal network operations. Formal networks specific ally bring a diverse group of people together to work towards a problem, thus she contended consensus building was crucial to network successes. She bifurcated the consensus building in her discussion to intra - and inter - network consensus building. Intra - n etwork consensus building referred to meaning - making processes taken by individuals within a single network. This could be discussions and agreements made by network leaders about the state of their network efforts, critical network - organizational decision s, or future planning for the network. Intra - network consensus building consensus building where individuals negotiated personal experiences (Hilarie). She attributed it simply to being part of a network. Inter - n etwork consensus building referred to a similar process of finding mutual agreement across different formal networks. Hilarie discussed how networks try on organizational processes or language from other networks. These could be specific forms of formal ne twork governance or bylaws, grant proposal structures, written or spoken language, or strategies for network longevity. Hilarie contended formal network sharing of processes was a 97 separate form of consensus building . She believed the constant sharing of or ganizational processes had a winnowing effect on new ideas. She admitted networks were often grant - funded and unwilling to take risks on unproven organizational contingencies. It was easier for networks to adopt previously - proven processes so they could fo cus on their mission of USER. She Hilarie pointed to formal ne twork leaders as the conduit for sharing information and building - consensus. The leaders of formal networks often had more experiences with organizational processes and took those experiences from network to network. Innovation . formal networks. Hilarie cited study participants who served as primary investigators on multiple networks and stated one of her formal USER network affiliations was a product of conversations with others in this study. She described the beginnings of a formal USER network as a simple conversation: [Participant 17] had an idea for a network, so [Participant 8] and I met up with him. He was thinking he wanted to use a network structure similar t o the ones we were using. So yeah, we helped him process though that; we were there when [Participant 17] started [the network]. We helped him through the model he was trying, and I think we were on the initial grant. (Hilarie). Hilarie also mentioned tw o other proposals recently submitted to funding organizations which sought to create new formal USER networks. These proposals included various combinations of study participants and non - participant USER leaders. Hilarie spoke a lot about monetizing formal network products as a way to help financially stabilize the organization post - grant funding. She identified several other networks whose leaders were contemplating similar organizational moves and indicated the general sense that networks 98 have traditional critical point in network innovations, as funders were starting to withdraw op portunities from networks unless they could demonstrate their effectiveness. Citing a shifting resource landscape and the retirement of a specific individual at the NSF, Hilarie argued USER networks will likely need to innovate their way to solvency with a more for - profit mindset. Jennifer. Jennifer currently serves as an academic administrator at a four - year college. She has a terminal degree in a STEM - related field, and served as a faculty member, researcher, STEM center director, network leader, and fun der. This participant also is currently a co - primary investigator on a grant studying formal networks. Jennifer reported that she only belongs to one formal network, but mentioned she assisted in the formation of several other formal networks. Despite her lack of network memberships, she is often invited to attend meetings for her insights. Knowledge exchange . Despite Jennifer not belonging to many formal networks, she indicated her various roles in USER allowed her to work on the formation and further building of several networks throughout her career. The experiential aspects of working in USER for an train you on this type of work at [graduate school]. Jennifer named two distinct formal networks from which she borrowed practices, bylaws, or actually wrote that earl y on into the forming of [network 2] years later. The program was not Drawing from experience, Jennifer stated her role was not necessarily to identify knowledge within the space, but to serve as a connector of as many people as possible to increase 99 this mindset came from years working as a faculty member in close proximity to other universities and especially in her role as a formal USER funding administrator. Jennifer also indicated the role of shared formal network leadership as a catalyst for bot h knowledge generation and exchange. Advisory boards, grant meetings, formal USER network steering committees were all mentioned in her interview; however, Jennifer indicated her role was less about being present in the room and more about recommending the appropriate others to be present to help the network. When pressed further Jennifer acknowledged that she was often on various committees, but different roles throughout her career kept her from fully participating (for either conflicts of interests or si mply time restraints). Consensus building . Jennifer lamented the disparate nature of USER and called for more alignment from USER leaders. Similar to Allie and Hilaire, Jennifer argued the diversity of STEM education and complexities of USER generated comp eting ideas on how systemic change could occur. These differing ideas led to an overabundance of formal networks all focused on some different aspect of change. Although she acknowledged the variety of different networks was not necessarily a negative, the connectors attempting to build consensus. Jennifer said that was part of her role as a connector. formal network to formal network helping to share in formation across networks, create alignment across leaders and build consensus on the state of USER. Her previous positions at different organizations helped her to achieve these goals, as USER leaders were interested in her perspective on many issues faci ng formal USER networks. 100 Innovation. Jennifer framed most of her innovation responses by referencing her formal USER networks which she had either served as a founding member or advised colleagues in its formation. This list did not include any formal networks she had funded in her administrator capacity. She also gave an in - depth story about how another study participant approached her in the founding of a formal nvolved and interested in leveraging [STEM] funding. We both realized that both being scientists, the conversations began to look too science - y (Jennifer). Jennifer argued she and another participant created a formal USER network to re - center students in the conversation of USER to make it more accessible. Similar to Allie, Jennifer also mentioned site visits, consulting trips, or other workshops in which she presented. Many of these interactions were driven through connections she had made within the netw ork. Although Jennifer downplayed her role as a major funder in USER, she acknowledged some of her invitations may have come from that role. Some of these invitations were from people within this group wanting to share information about the work happening in other parts of USER. Lindsey. Lindsey is a professional working at a national higher education organization. Although not directly involved in the leadership of networks, she serves on multiple steering committees, advisory boards, or consulting groups for many formal USER networks. Sh e has worked collaboratively with faculty and administrators at colleges and universities to publish scholarship on the state and roles of formal USER networks, and as a result she is often invited to attend various USER - related meetings across the country as a function of her employer. Knowledge exchanges . influenced her perspective on knowledge exchange. Lindsey stated she was on countless steering 101 committees and at least four USER - affilia ted advisory boards, attributing her membership primarily to a function of her employment with the national association. With regards to the nature of knowledge in formal USER networks, Lindsey argued her definition of knowledge was irrelevant, deferring t o others especially faculty members. Instead, Lindsey viewed her role similar to what Jennifer articulated. She thought herself to be a social resource or a people - connecter. When asked how information flows through leaders in USER, Lindsey stated it likel y occurred in both formal and informal settings surrounding steering committee meetings. en they walk out the lent itself to credibility. By being in the room, conversations among leaders spurred specialized knowledge which leads to being asked to part icipate in other specialized conversations. Lindsey also discussed the importance of sharing personal information to build trust and multidimensional relationships which were much more durable in times of formal network turmoil. She spoke of every single p articipant in the studying saying, them, I know their families, I know about their lives in pretty significant details. I mean this applies to some more than others, but I could tell you the ages of their kids, I can tell you where their kids go to school. Who is recently divorced or back in a relationship. I can tell you about their personal lives outside of work, who has had a death recently, who are feeling pressures or regret at their jobs and This level of personal knowledge aided, in what Lindsey called, the emotional support of networks, and increased the flows of information especially in times of crisis. Having the contex tual relationship knowledge increased the likelihood that Lindsey would share knowledge more quickly as she felt as though she was helping a friend in need. 102 Network learning . Speaking directly about network learning, Lindsey highlighted multiple steering c ommittee meetings she believed underscored how decisions were made in USER. Although she mentioned the steering committees typically had highly structured agendas and schedules, the discussions pertaining to formal experiences in other formal USER networks. During discussions, members of the committee would discuss different approaches their formal USER network took to address the agenda topic, and members would add amendments or reject the items discussed entirely. The topics of the steering committee agendas served as a prompt for members to begin to draw their collective experience and share what strategies worked. As an example, Lindsey spoke about her experience working with an unspecified network in charging me mbership dues. In this discussion, she emphasized how different individuals talked about their formal USER network members from resource - scarce universities. Innovat ion. Lindsey did not identify specific outcomes, outputs, innovations, or products coming from the group of participants. Although she stated additional networks were a product of conversations among some participants, she related most of the innovating to activities within formal networks. Lindsey contended new formal networks were the product of gaps in USER. Lindsey also referenced individuals who were not participants in the study as strong contributors to USER. All of these individuals were STEM facu lty conducting STEM education - based research at large public research universities around the United States. After listing work each of nationally, but are doing v ery strong, solid work in the space...I think they may be more known 103 herself. As she connects individuals to others, she may not participate in additional conversa tions or products from the social networks. Luke. Luke is a STEM faculty member at a research institution. He worked on several funded grants in USER and is an active participant in several formal USER networks. He explained that most of his work in STEM reform occurred at the institutional level or withi n his geographic region. Only recently had Luke entered USER at the national level and become involved in nation - wide formal networks. Luke felt close with many people working in the has several pending grant proposals to begin new formal USER networks at the time of data collection, Luke had not been given a response on the status of his proposals. Knowledge exchange . people w orking in USER outside of the study. To him, knowledge was defined as simply knowing whom to ask what. This spoke to a certain level of understanding of key players and language in USER. Luke also cited the role of steering committees and advisory boards a s a mechanism for becoming acquainted with others in USER and providing contexts for whom he could consult. Luke primarily talked about social connections established at these meetings as a feedback loop for his goals and thoughts for transformation in USE R. Beyond the personal connections provided by the formal USER network steering committees or advisory boards, Luke openly shared disdain for their function stating there was - tanki usually just sit around and spitball discussions on how we can improve networks, never mind 104 open disconnect with others in the study, he did state the origin of a new formal USER network currently in the proposal stage. He attributed the forming of this network to conversations between him and others at a fo rmal grant meeting for a different network several years prior to his interview. Coalition building . When prompted on his thoughts of coalition building among the participants in USER, Luke primarily discussed how participants organically built partnership s in (Luke). Although the reasons for coalition building varied, Luke speculated funding was the primary driver of coalition building. He argued most reform strategies created needed funding to get started. Thus, reformers would seek coalitions with others who could increase their chances of receiving funding. This could take the form of someone who is socially connected to funders, had received funding in the past, or could appeal to the funding priorities of grantors (e.g., interdisciplinarity, diversity, equi ty, and inclusion, etc.). Innovation. Along with several interviewed participants, Luke highlighted how many individuals in the study collaborated together to form different formal USER networks. He indicated many study participants entered USER from va rious organizations and backgrounds, and the diversity of experience led to the proliferation of new networks to fit different niches. Educational Researchers (DBER) all pulling different directions and coming up with new 105 both outlining his progression through multiple NSF grants, and his exasperation with systemic change in USER. Cross - Case Analysis: Knowledge A cross case analysis of emergent network findings provide several insights in how knowledge is defined within the space. Understanding how the group of participants defined and exchanged knowledge provides evidence to the existence of emergent network functions, and more importantly it provides greater context to the role of connections among the participants beyond them simply knowing one another. First and foremost, knowledge took many forms, and was dependent on the context in which it was needed. Whereas one may seek knowledge on how to structure or build a USER network, they may value knowledge on forming a network more than someone who has led a network for several years. Similarly, one who views themselves as a connector of people in USER may value and define personal friendships in that space differently than one who is required to be in the space because of their employment. This value - context is a driver for what constitutes knowledge, how valuable it is at th at time, and helps to account for the varying definitions of knowledge within the study. Knowledge, as defined by the participants in the interview, primarily falls into five overlapping themes: (1) formal network expertise; (2) formal network outputs; (3 ) disciplinary background; (4) USER domain; and (5) personal information. Table 9 contains each of the sub - themes, definitions, and examples for knowledge. Fundamentally, the most common recognition of knowledge within the group of participants was forma l network expertise. Participants widely valued any information that could aid their formal USER network affiliations. This knowledge concerned many of the functional and organizational needs of beginning and sustaining a formal USER network and included s teps on how to get initial funding, how to grow a formal network, questions of financial 106 sustainability, leadership concerns, and formal USER network partnerships. Knowledge in this form was critically valued by those who helped to begin a formal USER netw ork, as many networks were founded at the same time and were experiencing similar problems (Lindsey). This definition of knowledge was the most commonly cited by participants. Relatedly, formal network outputs were another highly valued form of knowledge. This form differentiated from formal network expertise as it related to the specific purpose and function of formal networks. Knowledge in this theme was more about formal network products and services and not directly associated with how to run a network . Defined as initiatives, projects, reports, conferences, forums, or other activities held by a formal network, this form of knowledge was critical to leaders working in USER as it gave actualized definition to the purpose of each network. Knowledge about the on - goings of other networks kept leaders current on the state of USER and provided potential activities for them to use within their own networks. For example, if [ formal network A] convened a successful forum [ formal network B] may host a forum to rep Disciplinary background represented another area of valued knowledge with participants. Defined as knowledge of specific discipline - based cultural and social norms, this form speaks to the interdisciplinarity of U SER. USER exists at the intersections of many different fields and disciplines. Aside from each faculty in STEM - related fields, USER includes education faculty, university staff and administrators, higher education associations, and funders. Each of these groups carry specific and unique languages, backgrounds and cultural norms associated with their domains. Importantly, this theme does not necessarily include discipline content (i.e., ecologists knowing civil engineering). Seeking out discipline backgroun d knowledge was crucial for many leaders seeking to build networks or other initiatives, as it allowed them to speak to 107 wider audiences and gave them the necessary cultural capital to mobilize leaders across discipline silos (Allie, Hilarie, Jennifer). Thi s knowledge was crucial for participants who did not hold STEM - related degrees and allowed for participants to engage with a wider variety of individuals within the space. Background knowledge also was a critical component in operating formal networks. Re lated to discipline backgrounds, USER domain knowledge refers to knowledge gained from learning language and a set of behaviors associated with those working in USER. Primarily expressed from those who had not been working in USER for an extended period, s everal participants articulated the value of using appropriate language or formatting on grant proposals, or awareness of individual and organizational actors in USER. With regards to USER - specific acronyms (e.g., NSF, INCLUDES, story in USER work. This process included learning how to act at various USER meetings or conferences, and some jokingly referen ced perceived missteps in their few years working in USER (Hilaire, Lindsey, & Luke). The fifth theme for knowledge was personal information. This constituted any information about individuals in the study that was not directly linked to formal networks or USER more broadly. Personal information included knowledge most commonly found with social friendships. People described knowing other hobbies, families, work - life, dietary preferences, and other personal information. This information appeared to modify existing relationships beyond simply discussing USER business. Individuals in this study who knew personal information about the others may have been more likely to share information about the other knowledge themes than with people they did not know as well. 108 Table 9 . Knowledge Thematic Breakdown Knowledge: Information or other expertise as defined by participants and valued by its scarcity Categories Definition Examples Formal Network Expertise Knowledge and experience working with formal USER networks. Knowledge on how to: 1. Get network funding 2. Grow a network. 3. Manage a network. Formal Network Outputs Knowledge and information derived from the actions of networks. Knowledge related to: 1. Network grant reports. 2. Network press releases. 3. Network conferences and forums. Disciplinary Background Language, knowledge, credentials, and cultural understandings associated with working in a specific academic discipline. USER Domain Knowledge Process by which new entrants into USER learn the language, norms, stakeholders, and behaviors associated with working in USER. Knowledge of: 1. STEM Acronyms 2. Formal network names. 3. Influential funders in USER. Personal Information Knowledge of other participants personal lives. Knowledge of other participants 1. Family members. 2. Recent personal occurrence. 3. Dietary need s. 109 Cross - Case Analysis: Knowledge Exchanges Although defining knowledge in the formal USER space provides some insights about this group of participants, it does little to address questions of why and how USER knowledge moves through the social networks. Given that information diffusion and knowledg e exchanges are foundational aspects of an emergent network (Isett, et al., 2011; Popp, et al., 2014), the motivations moving information around the social network must be explored further. Defined as the movement of knowledge bases throughout a network, i nformation diffusion and knowledge exchanges refer to the actions taken to share or seek information in a network, albeit differently. Diffusion of information is an organizational term used to refer to network - wide or organizational saturation of knowledg e (Rogers, 2003). Typically paired with a time - component information diffusion is more closely related to network analysis. Knowledge exchanges on the other hand, refer to individual interactions where two people share knowledge or information (Hartlet & Bennington, 2006). These exchanges include the time, place, and manner or the exchanges, and also the motivations for exchange (Huang, 2015). As the social network chapter discussed aspects of information diffusion, this section will more focus on aspects of the knowledge exchanges. At the conclusion of this section, Table 10 depicts a brief summary of the knowledge exchange discussion. The first question in approaching knowledge exchanges may be: what knowledge is exchanged? In this study, I drew upon the bodies of knowledge defined by the participants in the previous section. Those groups were (1) formal network expertise, (2) formal network outputs, (3) disciplinary background, (4) USER socialization, and (5) personal information. With regards to the exch ange of the knowledge previously outlined, the participants identified four knowledge exchange themes. These themes were to: (1) aid internal formal networks; (2) seek input from 110 peers; (3) contribute network successes to USER; and (4) USER socialization. Interestingly, the the knowledge section) as an area of knowledge exchange. Whereas people may have sought knowledge to aid a formal network, getting to kno w someone did not appear to be a primary reason for people exchanging information. This is likely attributed to the participants knowing one another quite well at the time of the interviews. Knowledge exchanges to aid internal formal network processes rel ated directly to the operation, sustainability, and overall stability of USER networks. In general, many participants cited their lack of expertise in network management, advice from peers, or the need for precedence as other motivating factors in knowled ge exchanges. Most participants described needing to access knowledge within their social network directly following a successful grant. Hilarie and Andrew described a sense of overwhelming aimlessness when starting a USER network and feeling like they wer so or overwhelmed ebbed and flowed throughout the course of the formal network. Several participants discussed si milar feelings when their grant cycles neared completion. At this point, leaders described consulting with others in USER about how their formal networks could exist without grant - funding. Participants discussed seeking out other leaders who had been assoc iated with older networks which made similar transitions. Allie, who made a network - funding transition, specifically commented on how many consulting visits she had tended to in her formal network operations in times of uncertainty from individuals within the USER. 111 Participants also indicated a strong desire to seek the input from others in this study. - (Lindsey), these knowledge exchanges were more about validating or refutin g reflections on observations made in USER. This knowledge exchange was particularly prevalent when the consulting others in their social networks in the seminal p hases of beginning a formal network. In these conversations they acknowledged the gaps in current USER efforts, then sought input on how their proposed networks could fulfill the needs. Conversation on USER - related observation also included other forms of participant - defined knowledge (i.e., internal formal network operations, etc.) the fundamental purposes differed for sharing the information. Although most participants contended seeking input was positive for those working in USER, Jennifer argued the see king input from others of network and initiatives working in USER, Jennifer felt many of the conversations were opportunities to appear to be seeking input. In actuality she said these interact ions were more to spread new ideas in USER and gain support from the community for upcoming initiatives or proposals. Overall participants all indicated leaders in USER engaged in knowledge exchanges for the purposes of seeking input or consensus from othe rs, and they all acknowledged how relationally - connected many leaders were in these exchanges. Showcasing formal network successes was a third area of knowledge exchanges. This form of exchanges is defined by USER leaders using multiple mediums to update o r share information about the work their formal network is successfully doing. This knowledge exchange could be in a formal setting like conferences, forums, or meeting, and could also 112 networks were doing. These sentiments were echoed by Jennifer and Lindsey, who visualized nifer articulated she thought sharing information across networks and leaders were her primary function in USER. Other participants explained many formal network events were designed for the purposes of network knowledge exchanges. Andrew and Luke stated they hosted (and often attended) network forums or In addition to sharing network successes, most participants expressed a desire to share formal network shortcomin participants discussed the necessity of seeking this knowledge to avoid similar mistakes or pitfalls. In contrast to network successes though, network failures were typically shared in personal of fline (Allie). Knowledge exchanges related to sharing network shortcomings re lied strongly on the personal relationships and interactions of USER leaders. Outside of knowledge exchanges on formal networks, individuals sought knowledge to acclimate to USER. As mentioned in the knowledge section, USER is a space with domain specifi c language, norms, and behaviors. The desire to seek USER domain - specific knowledge mirrors many aspects of a USER - centered socialization. Described as a process of learning values, skills, attitude, norms, knowledge, and language, Merton (1957) argued pro per socialization was necessary for admission, acceptance, and social mobility in a given group. Participants recounted multiple instances of trying to learn USER language and norms. Hilaire and Lindsey both recalled taking actions to learn specific langu age as a function of their 113 - the - talk, or no one will Participants discussed seeking others at the as a language which had proven to be funded by grantors in years past. Multiple participants (Andrew, Lindsey, & Luke). In addition to these aspects of socialization, these individuals also s with others in USER. Mediums of knowledge exchanges . The place where knowledge exchanges occurred varied for each of the participants. Although participants cited many communication mediums (i.e., telephone calls, e - mails, in - person conversations, etc.) , in - person meetings were mentioned by all. Occurring at conferences, workshops, or other collective USER functions, these physical spaces provided a forum for casual conversations on contemporaneous events with formal USER networks. Andrew, Lindsey, and L uke all highlighted conferences where they sought other participants for their input on USER - related topics. More in - depth conversations among the participants occurred in formal network steering committees, grant advisory boards, national association meet ings, or other small - group gatherings. Lindsey and Jennifer both cited the disc ussing them. These meetings allowed multiple participants to regularly interface with one another and discuss topics in more depth when compared to other mediums. 114 Table 10 . Knowledge Exchanges Thematic Breakdown Knowledge Exchange: Motivations and settings for diffusing USER knowledge across the emergent network. Categories Definition Examples Aid Internal Formal Network Functions Seeking knowledge to build processes in a current or future formal network with the intention of addressing network problems. Seeking knowledge on: 1. Network processes and structure. 2. How networks were funded. 3. How networks grew. Seek Input from Peers Individuals consulting with peers about observations they made regarding formal network actions or other USER - related efforts. Consulting others about: 1. Shortcomings in USER efforts. 2. Shortcomings in their formal networks. 3. Effective strategies from other domains Contribute to USER Success Sharing positive and negative network outputs for leaders in the community to replicate or avoid. Looking for knowledge in: 1. Personal communication 2. Network grant reports. 3. Network press releases. 4. Network conferences and forums. USER Socialization Seeking information related specifically to the USER domain for the purposes of learning to operate more effectively in the space. Seeking knowledge related to: 1. Current USER efforts. 2. Existing formal Networks 3. USER - specific language 115 Cross - Case Analysis: Network Learning While the discussion of knowledge and knowledge exchanges highlight how an emergent network in USER values and shares information, it does little to unveil how the group of participants collaborates with one another. Simply highlighting a collection of ind ividuals which shares knowledge does not indicate an emergent network; in fact, it makes it indistinguishable from any other network (Isett, et al., 2011). In order to distinguish an emergent network from a simple collection of leaders in USER, network le arning (the second emergent network function) showcases how an emergent network operates organizationally (Popp et al., 2014). Defined as a a process of refl ection and meaning - making involving individuals who represent organizations. The process involves individuals bringing information from their respective organizations for - level discussion and meaning - ed, & Huang, 2012). As highlighted in the literature review, network learning is also a process built on a foundation of relationships. Without the interperson al cornerstone of trust and fairness, network learning cannot occur (Leach, Weibe, Vince, Siddili, & Calanni, 2014). In reviewing the cases participants discussed three themes related to trust - building, fairness and learning. The themes were (1) consensus building, (2) coalition building, and (3) network learning. Table 11 displays the thematic breakdown for network learning and interpersonal dynamics. Consensus building . consensus building requires a group to reflect on prior experiences and observations for the purposes of establishing agreement (Bezdeck, Spillman, & Spillman, 1978). These agreements provide mutual definitions for individuals to ground conversations and actions and shape future work and 116 hav e the potential to become ubiquitous within a domain. While consensus does not necessarily imply unanimity, it does establish a generalized set of understanding for a group to base their interactions. With regards to consensus building in USER, participant s generally sought to build consensus in several areas related in USER. In many cases participants discussed observations on USER landscape or sought input from others on problems within their formal networks. Participants also discussed reaching out to ot hers for input and meaning - making for building a observations and were often looking for clarity or a thought - partner. Several participants discussed the necessity o f seeking consensus and acting as a sense - making partner. The most prominent example of consensus building occurred in formal network steering committee meetings. In these meetings, participants discussed agenda items and built consensus based on how they is to discuss problems coming from [mandated] networks. The agendas are, more or less, set up in dicated funding was an agenda item at a recent steering committee meeting. In that meeting, leaders in USER convened to discuss how other networks in which they were involved confronted funding concerns. They combined their previous approaches with the cur rent context served as a consult for USER - specific knowledge. Drawing from her long career, she indicated people often came to her in need of historical contex t required to make sense of current network phenomena. The role of consensus with the participants demonstrated another level of complexity and interdependence within this group. The capacity and motivations to consult with one another 117 depict a sense of c ollectiveness as the participants seek to make meaning from others in similar situations. As mentioned, steering committees or advisory boards played a large role in the g committees or advisory boards allowed the involved individuals to aid a particular network, but more than the aid, the meetings also served as a space for USER leaders to build approaches to USER problems more generally. The consensus or shared meaning t aken by the leaders from these meetings were used to both frame conversations for their networks but also used in subsequent meetings as established foundational understandings. Thus, each of these meetings built on one another, offering opportunities for these leaders scaffolding to continue to filter new information and build consensus. Coalitions. The second theme regarded how participants sought partnerships in network - related actions. Coalitions are defined as a collection of individuals who share a co mmon interest who agree to work together to achieve a common goal (Wolff & Berkowitz, 2000). The process of negotiating the agreement to work together is commonly cited as coalition building and often includes aspects of consensus building as the parties i nvolved define their intentions, aims, and goals (Kaye & Wolff, 2002). Different forms of coalitions were cited by all participants in different capacities. First, Andrew and Jennifer discussed building coalitions as a step in forming a formal network. In their discussions they highlighted the importance of getting critical stakeholders on - board with the new initiative or network. Andrew and Jennifer, nevertheless, differed on who should be recruited to build a coalition. Jennifer explained building part nerships with organizational actors such as funders or universities and contended that funders sought partners to move USER efforts forward. These coalitions could drive convergence in a disparate field of 118 n example of a funder pushing for common funders set parameters fo r individuals to present opportunities for coalitions. Andrew, on the other hand, insisted on building coalitions within formal networks. He wanted individuals who were reform - interested but otherwise unengaged in USER efforts. Underscoring the importance of persuading stakeholders who were not currently involved in the USER, he cited this group as a source of more knowledge and expertise. Building coalitions was a strategy for Andrew in the facilitation of formal network management and survival. In a importance of leaders building coalitions across her self - described silos in USER. These silos were informal divisional boundaries that segregated reformers in veins of USER wor k such as research silos, discipline - based silos, and formal across them to be successful. But to talk across the groups, you have to have a buy - 119 Table 11 . Network Learning Dynamics Breakdown Network Learning: Interpersonal processes for achieving system level learning. Sub - theme Definition Examples Consensus Sharing knowledge or observations within USER with others for the purposes of reaching agreement. Employed to create a foundational understanding and operate as a collective. Conversations among leaders to: 1. Understand USER environment 2. Aid formal network 3. Align formal networks. Coalitions Shared interest partnerships created across USER to increase the success of an individually articulated USER goals. 1. Receiving a grant from a funder. 2. Seeking others for a project of initiative. 3. Building relationships across the different disciplines in USE R. Network learning Sharing USER - related organizational knowledge (e.g., formal networks) within emergent network to gain perspective on system - level occurrences. 120 - you. Network learning . Although a few participants mentioned situations of shared organizational experience and meaning - making, there were little data to suggest the system - wide learning. One participant suggested systemic change in higher education was the goal in USER, but further mentioned that aim is too ambiguous for operationalization and measurement (Luke). Popp et al. (2014) stated emergent network learning must have an emergent organizational objective with which to channel the shared knowledge and measure growth. Despite sharing experiences, building consensus and coalitions, much of these efforts were for short - term and specific ends (i.e., how to get funding) without further organizational implications or feedback loops. There was no consensus expressed on how to achieve systemic change in higher education. There was little evidence of a formalized togetherness expressed by those interviewed. Cros s - Case Analysis: Innovations The final cross - case analysis concerns outputs of the emergent network. While knowledge exchanges and network learning are critical functions of an emergent network, they tell us little about how the network interacts with the environment. The final emergent network function (innovations) highlights how an emergent network affects the domain in which it is situated. Innovation refers to the collaborative capacity for a network to operationalize its knowledge and learning to chan ge its organization or system (Provan & Huang, 2012). Keast et so what in their interviews were: (1) new formal networks; (2) proposals; (3) formal network governa nce 121 structures; and (4) formal network sustainability strategies. Table 12 depicts the thematic breakdown for innovations at the conclusion of this section. By far, the most cited innovation referenced by participants was their collective capacity to gener ate new formal networks. Participants cited changes in undergraduate STEM education on college and university campuses, faculty concerns, and changes in the funding environment as drivers for creating new formal USER networks. Several individuals stated th e sprawl of new networks was necessary as systemic change was difficult. Creating new networks allowed people in USER to engage new individuals in reform efforts and evolve with the universities (Jennifer & Lindsey). The new formal networks spawned by part icipants typically targeted unengaged actors (university faculty and administrators, education policy - makers, etc.) in USER, and in some cases were in direct response to new funding opportunities. In particular, two interviewed participants and another un - interviewed participant had recently submitted two proposals to begin a new formal network. One of those proposals aimed to directly engage faculty on educational research and came from conversation among these individuals at various conferences or worksho ps. In connection to new formal network generation, most of the participants indicated their collective capacity to create funding proposals for various initiatives. Different from new formal USER network proposals, these proposals were broader and included funding for specific projects, workshops, conferences, information depots, or toolkits. Proposals generated by participants were almost entirely directed to the NSF or FIPSE and sought funding for formal network business initiatives. For example, one formal network was looking to incor porate more initiatives in diversity, equity, and inclusion, so some participants discussed applying for an ADVANCE grant to be managed or attached to their formal network. Most participants indicated 122 their submissions were in response to open calls for pr oposals from major funders and were in conjunction with formal network needs. Some participants stated their networks needed financial sustainability and grants provided the necessary funding to support their efforts. Others argued these proposals allowed their formal networks to align with changing priorities in USER. Multiple participants cited different proposals for diversity, equity, and inclusion initiatives offered by funders. Third, participants indicated various governance strategies for for mal USER networks . formal networks. network to be used in another because of her expe rience with its success in a different area. Hilarie cited the research action cluster (RAC) model as an organizational form she used in developing her network. In her example, she specifically talked about consulting with members in [a formal network] and their experience with a RAC. Similarly, Luke indicated using a networked improvement community (NIC) in his proposal borrowing the thought from two other formal networks. There were divergent motivations for borrowing and adapting these proc esses. Some articulated experience and proven success with structures imported to development. For these individuals, they wanted to focus on other aspects of the pro posal or network and argued that using a previously funded network model increased their chances of funding through precedence. Others also granted that they were not experts in organizational design and simply needed a starting point for their formal netw orks to debate. This borrowing of governance structures was not a simple copying of the governance from one network to another, 123 Table 12 . Innovations Thematic Breakdown Innovations: The capacity for a network to operationalize the knowledge and learning to change their organization or system. Sub - theme Definition New Formal Networks The creation of new formal networks in USER as a result of emergent network interactions. Other Proposals The creation of other, non - formal network, initiatives, programs, toolkits or workshops as a result of emergent network interactions. Formal Network Governance Structures The creation or adaptation of new formal network governance or leadership structures to assist with formal network development. Formal Network Sustainability Strategies The creation or adaptation of organizational stability strategies to assist with the long - term viability of formal networks. 124 knowledge related to the governance structure; these conversations included an adaptation contexts. Through these conversations, adapted governance structures were cus tomized and created to fit a new formal Fourth, formal network sustainability strategies were cited as a strategy - in - development for many participants. Almost all of the participants interviewed expressed anxiety over the sustainability of their formal network. This sustainability was mostly linked to the fina ncial health of a formal network and its connection to a grant - funding cycle. Most of these strategies sought to decouple the formal network from the funding cycle and assumed major organizational changes for the formal user networks. These strategies were to: (1) transition formal networks to independent non - profits; (2) link the formal network to an existing college or university; or (3) link the formal network to an existing higher education association. Although these strategies were highlighted by part icipants, they were viewed as in development as most of the participants were wrestling with issues of sustainability. USER Emergent Network Function Discussion Independently each of the emergent network functions discussed by participants provide unique insights into how this collection of individuals operate in USER. They seek USER - specific information about formal network operation or other phenomena in USER, and transfer knowledge in steering committees or advisory board meetings. These meetings also a llow for consensus and coalition building to occur and often generate some initiative or output for networks or USER strategies as a whole. An emergent network, however, is more than a group of individuals working in tandem. Emergent networks are organizat ions (Isett et al., 2011; Popp et al., 2014), complete with governance, power, and processes . Although recognizing the 125 emergent network - like functions (i.e., knowledge diffusions, consensus and coalition building, innovation) are important in understanding some process - oriented aspects of the emergent network, they do not address aspects of power or governance. The following discussion builds on the emergent network functions established in the previous sections comparing the data to existing literature on human capital and organizational theory. Emergent network as an open system . Interpreting the emergent network functions as a cycle presents an opportunity to view this network as a structured organization using systems theory. A collection of organization al theories on structuration of processes, termed systems theory, posit organization as amalgamations of social structures that provide a schema for those within the organization to act (Scott & Davis, 2003; Bertalanffy, 1956). A subset of systems theory, known as open systems theory, contend an organization socio - behavioral structure are built through the consensus of loosely affiliated parts (Ashby, 1968; Glassman, 1973). In this theory, traditional organization boundaries are porous and ill - defined. Inst ead an organization is bounded by those who partake in institutionalized processes and behaviors, with individuals entering and exiting the organization fluidly (Scott & Davis, 2003). The processes and behaviors which constitute the organizational body rel y on reciprocal social ties built upon interdependence and information flows (Burton & Obel, 2004). Boulding (1956) likened open system - govern the body but many di fferent heads could encounter information and act. These semi - autonomous decisions were informed through consensus and coalitions built across the many - - how, and meani 126 As the theory suggests, participants in the emergent network are not constrained by any formal organizational boundary. Instead, they are all loosely affiliated through mutual social relationships and shared memberships on committees or boards scattered throughout USER. Using the organism analogy, this organization of individuals is often activated by observations in - ho w and shortcomings in USER efforts. The emergent network functions (i.e., knowledge exchanges, network learning, and innovation) occur within the emergent network, b ut do not need to incorporate all members simultaneously. The leaders do not serve as a single monolith within the network, but rather act on new information or in the face of shortcomings in USER with consulting only a few other leaders. This mirrors the multi - cephalous metaphor presented by Boulding (1956) by anchoring individuals in social processes but allowing them to act semi - independently. Using an open systems approach to the emergent network in USER presents several opportunities to frame the functions. First and foremost is the importance of the environment. In without undergraduate STEM education reform or the resources created by the funders of USER. Moreover, without environmental problems in USER, there would be litt le need to convene and strategize remedies to the environment. In short, without the critical role of the environment, the network (and its functions) would simply not be relevant. Setting aside the role of the environment, the necessity of interdependenc e is critical to an open system organization. Exploring each function of this network demonstrated the need for 127 interdependence. Indeed, knowledge exchanges, consensus and coalition building cannot be accomplished in a vacuum. They require others with whic h to exchange or cooperate. Even innovation (which does not assume interdependence) was assumed to be a cooperative process by participants. The interdependence exhibited by the participants in serving these functions demonstrate an organizational bond whi ch links them together in contrast to the environment. Finally, interpreting the emergent network as an open system provides perspective on ill - defined notions of network learning. Many of the participant s struggled to define or highlight instances of network learning occurring in the emergent network, instead underscoring moments of coalition or consensus building as a means of sense - making. With only loose affiliations tying the participants together, the re are no centralized systems in which to channel any systematic organizational learning, much less network learning. Instead, feedback loops are disaggregated and spread across the organizations. While these loops still provide sense - making opportunities, they do not rise to the degree of system - level learning commonly associated with network learning. While the open systems theory provides much needed context for understanding how this emergent network operates in USER, it does fall short in several crit ical areas. First, the theory unveils little about the origination of the function cycle. Although some scholars may point to the processes originating in the environment (Scott & Davis, 2007), more exploration is needed. Additionally, while the open syste m theory recognizes a porous organizational border, it does not detail conditions for open inclusion or exclusion in the network (i.e., who is in or out and why). The following section begins an exploration in human capital and explains how manifestations of capital affect interactions among participants in this network. 128 Emergent Network Governance and Capital Manifestations Although there is not much research on the how networks operate, the existing scholarship articulates the importance of formalized net worked governance (Kezar, Gehrke, and Bernstein - institutions, structures of authority, and collaboration to allocate resources to coordinate the joint action across the networ formal networks use written charters, constitutions, or bylaws as binding documents which support formal network stability and effectiveness (Milward & Provan, 2006). Unlike the formal networks however, emergent networks do not h ave formalized bylaws, governance, or leadership hierarchy with which to govern how members interact with each other and the network. Indeed, the lack of organizing documents are, in part, what make emergent networks difficult to study (Isett, et al., 201 1). In emergent networks, items typically outlined in formal network charters are amorphous and undefined. Whereas a formal network may simply list the member organizations of the network, an emergent network membership is embedded in individual relationsh ips, with no outright Although emergent networks do not have written structures governing how individual s t in social settings (Provan & LeMaire, 2012, p. 45). In lieu of written bylaws, different forms of human capital influence how people interact in emergent networks. Human capital refers to an nerating benefit for the individual (Becker, 1962). Human capital generally confers influence, power, trustworthiness, or some other socially desirable quality. Without written bylaws aspects of human capital guide how interactions occur. Although not uniq ue to emergent networks, finding commonality in how 129 capital in USER is valued and used provides insights to how a more latent governance structure manifests in emergent networks. Interviewed participants discussed several forms of human capital as they exi sted in USER: namely, cultural, organizational, and intellectual capital. Each form of capital bestowed some form of trust, influence, or credibility to those who possessed desirable forms of capital by those within the space. As a note, the impacts and ma nifestations of social capital were outlined primarily in the social network analysis and findings chapter. Cultural capital . Cultural capital refers to the social assets of a person that promote their social mobility and desirability (Bourdieu, 1977). Ori ginally outlined with multiple sub - forms, cultural capital generally represents power through symbols and affiliations. T hese affiliations bestows influence or tr ust if the organization is viewed favorably by a social group. Ostrom (2019) used the example of celebrities entering politics by leveraging influence granted to them from television, professional athletic organizations, or military affiliations. More spec ific to USER, participants referenced the influences of cultural capital in negotiating social interactions - forms of cultural capital as a framework, I present findings on how cultural capital manifests in the eme rgent network. First, the institutionalized state (Bourdieu, 1977) was easily identified by participants. During the interviews, individuals referred to other participants by their affiliations with colleges and universities, higher education associations, government entities, or funders. These affiliations appeared to help participants remember and contextualize the relationship they had name, their role and col lege or university, and then any funding organizations with which they 130 working on some really interesting stuff in [the teaching and learning center]. He s tarted [ formal network] with [non - her introductory comment, Lindsey listed 5 organizations with which the participant was tied and served to credential the person in question. Altho ugh simply listing organizational affiliations does not necessarily bestow cultural capital or inherently indicate power in a social setting, two types of organizations were consistently referenced as desirable: funders and higher education associations. A lthough there is some overlap across the two categories, the most commonly cited funding organizations were the National Science Foundation, Howard Hughes Medical Institute (HHMI), and the Alfred P. Sloan Foundation. Jennifer explained the role of funders - cultural gave credibility reputation. Their grant funding and affiliation provided a shortcut for individuals to vet a The second grouping of organizations were higher education associations ( Associations). Unlike the relatively few funders cited, participants identified numerous different Associations. The groupings were typically non - profit organizations with either a mission to address USER or some organizational imitative to address STEM ed ucation. Some Associations served as political advocates or centralized confederated organizing bodies. The most commonly cited groups were the Association of American Universities (AAU), Association of Public Land - Grant Universities (APLU), the National A cademies on Sciences, Engineering, and Medicine (and its corollary the Board on Science Education), and the American Association for the Advancement of Science 131 (AAAS). Different than the symbolic power associated with grants, affiliations and power linked to the associations provided a more collective or representative power linked with college and university legitimacy. Jennifer discussed the symbolic power of those affiliated with the AAU, ctually successful; the reason I think it is so valuable is if AAU says that teaching is important, everybody else is going to get on - worked with AAU or the National Ac institutionalized state sub - form of cultural capital is prevalent in USER, as funders and associations bestow influence on those affiliated with their work. A second form of cultural capital, termed the objectified state , refers to power or influence an individual receives for their domain - defined desirable qualities (Bourdieu, 1977). Ostrom (2019) indicated there were clear in - groups and out - groups when considering the objectified state, and these qualities could be earned or unearned (e.g., education or physical appearance). Within the group of participants, an individual with a STEM - related terminal degree held more influence than those without a STEM - related degree. Clear distinctions were made when participants discussed others with terminal STEM degrees and those without STEM degrees working in USER. Giving more power to those who had a STEM degree, the shared experience of being in the sciences carried the symbolic weight of being in the know or having a similar educational background. This distinction was made clear by a participant (Person A) who dismissed another participant (Person B) by stating Person B was 132 contextualized their contributions as someone who does not carry as much credibility in the USER. Knowing the USER lands cape, language, critical players, and organizational actors represents a third form of cultural capital. Bourdieu (1977) defined linguistic capital as distinct gi ven domain. This form of capital is the knowledge about how to talk about the landscape, critical players, and organizational actors within the domain. For example, Ostrom (2019) talked reflected on his first time on a park basketball court where he refere nced the National Basketball Association by its full name. Although he knew of the organization, he did not know the colloquial term NBA and was ridiculed for it. Turning back to the study, several participants discussed similar phenomena occurring (with l ess ridicule). Luke highlighted his socialization process by arguing how USER is full of acronyms, a random assortment of funders, and influential people. He mentioned how any misstep in understanding the players in USER could marginalize an individual as an outsider. You have to know of Carmon [pseudonym], she gives the marching orders, but you also need to know AAAS and that is different than BOSE inside of National Academies, which have pieces dedicated to STEM reform, but also your institution may be a ffiliated with AAC&U, or APLU or AAU, or both, and they - off STEM affairs. (Luke) In discussing USER social ization, none of the participants cited particular education strategies (i.e., evidence - based instruction, multi - cultural pedagogy, etc.) pushed by many of the formal networks. The socialization focused almost entirely on learning the language to use in re ference to organizations, and individual actors within the space. 133 Organizational capital. Opposed to the symbolic significance of organization affiliation with cultural capital, this organizational capital speaks more to pragmatic aspects of controlling r esources (i.e., financial, physical, geographical, etc.). The ability to access and use resources to pursue a goal falls within this category as they can exhibit power and influence (Morgan, 1998; Scott, & Davis, 2007). Participants discussed organization al capital primarily through affiliations with USER funders, formal USER networks, and higher education associations. In addition to the cultural capital provided through a successful grant, the grantor - grantee relationships provided the avenue for indivi duals within USER to complete their work. Most commonly, funding relationships provided large sums of money to individuals working in USER and physical space to either host meetings, conferences, or events. Many participants discussed how this capital help ed them establish a formal network or initiating a study of USER. Allie recalled working with the Fund for the Improvement of Post - Secondary Education (FIPSE) at the doing the work of reform, Luke viewed funding relationships as a cycle for social mobility beset with exclusionary barriers. Linking organizational capital to cultural capita funding is half [of] the equation; once you do something with it, then you start getting invited to tural capital in the USER. The symbiotic relationship between organizational and cultural capital was also echoed by Hilar i e whose recently funded proposals were occupying more of her time than she intended. In her example, she felt ethically and professio 134 funding relationship, nevertheless, provided the power and resources for an individual to create other forms of human capital in undergraduate STEM education reform. Affiliation with a national higher education association demonstrated another form of organizational capital. As mentioned, national higher education associations (Associations) are a diverse group of non - profit organizations, primarily located in Washingt on, D.C. Associations may be funded through government agencies (e.g., National Academies, AAAS) or through college and university membership dues (e.g., AAU, APLU, AAC&U), and act as a convening or advocacy space for institutions or individuals. These ass ociations typically carry significance in USER due to their proximity to funding or policy - making. Every participant in the study held some formal affiliation with an Association. Some individuals were employed by an Association, while others were on commi ttees, task - forces, or other association initiatives. While affiliation with Associations provided cultural capital (as discussed in the cultural capital sections), they also lent meeting spaces, convening power, and financial resources to their affiliates . Intellectual capital. Participants also discussed knowledge as a form of capital. Often referred to as intellectual capital, this form concerns the asset of possessing unique and pertinent knowledge, experience, or other scarce information. Stewart (19 98) described intellectual capital as domain - specific knowledge could be leveraged to generate other forms of capital (i.e., social capital). Individuals with domain - specific knowledge hold capital in their ability to use and apply it. Intellectual capital aligns with much of the discussion on knowledge, knowledge exchanges, and may be a driver of knowledge transfer and coalition building. The participants discussed intellectual knowledge primarily as those who were sought to aid formal networks. Allie spo ke about seeking outliers and provocateurs in USER for 135 assumptions. To her, the provocateurs were able to see STEM differently than those already engrained in the cultura often sought individuals with different perspectives to work with in formal networks. Jennifer s to were considered others, or non - participants, by those interviewed. Intellectual capital then, was sought by participants seeking novel information for formal networks or Associations. Incidentally, intellectual capital was leveraged by individuals to gain access to the emergent regard, intellectual capital was used by those external to the emergent network to generate social capital and gain access to conversations occurring in the emergent network. As an example, provided independently by two participants, Lindsey discussed a closed - door committee at an Associ ation, where committee - review, interviews, and approvals. In a separate interview, Luke stated he was invited to join the e. Although the his participation in a regional USER conference a few weeks before the call. He mentioned in his sessions he asked interrogative questions, leaders during the conference. Capital, Governance, and the Emergent Network In sum, four forms of human capital impact how individuals within the emergent network interact. The previous sections outlined how cultural, organizational, and intellectual capital all bestow power and influence on individuals who hold capital in those areas. Social capital creates 136 a fourth form of human capital. Although not highlighted in this section, social capital was explored heavily in the 4th chapter. Within social capital theories, individual power and influence is derived from structural holes (Burt, 2000), boundary spanning (Katz & Tushman,1980), and weak ties (Granovetter, 1977). Essentially, individuals gain influence by connecting individuals who do not know one another. To revisit the findings from Chapter 4, individuals within the study were closely linked to one another and exhibited many aspects associated with network closure (i.e., high degrees of trust, free flow of information, etc.) (Burt, 2000). While a highly connected network does have many network - wide benefits, it does not provide much leverage for an individual to wield connecting power. As most of the individuals know one another, social capital does not appear to necessarily serve as a form of power or influence. In short, everyone knows one another in the network, therefore no one can serve as a gatekeeper of information within the network. Instead, power and influence primarily comes from (as Jennifer stated) organizational affiliation and funding. Returning to the discussion of how human capital serves as an informal set of governing bylaws, several findings emerge. Inclusion in the emergent network is connected to cultural capital. Being connected to higher education associations or funding organizations (institutionalized state), having appropriate STEM - related credentials (objectified state), and speaking the appropriate language (linguistic) applied to all of the participants interviewe d. More than simple inclusion or exclusion, organizational affiliations were how individuals gained access to the most highlighted knowledge exchanges: steering committees and advisory boards. Without participation within these groups, individuals in the emergent network would not be able to engage in the primary functions outlined in the previous sections. 137 Similarly, cultural and organizational capital provided the basis for aspects of emergent network leadership. Discussing networks in healthcare, Metzge r, Alexander, and Weiner (2005) defined network leadership as an ability to create a vision for the objectives of a network, yet still provide flexibility and consensus for those working within the network. In their discussion, the authors stressed the imp ortance of multiple - stakeholder input in determining how the vision was achieved, but the overarching goals were determined by leaders. As Luke discussed, a formal position within a funder or higher education association provides power to shape STEM reform to shape how reform work is conducted or setting initiatives in reform. Indeed, three of the individuals interviewed had recently held executive positions with in a funding organization or higher education association. As in the discussion of the multi - cephalous, open system emergent network, aspects of human capital inform how the many heads of the emergent network interact with one another and explains some of the semi - autonomous behaviors. Other than providing the aforementioned knowledge exchanging spaces, organizational affiliations lent prestige to influence how consensus could be built and leveraged funding to shape innovations of the emergent network. Al lie and Hilarie both discussed how they constructed proposals to align with higher education association and funder priorities. While these actions had a pragmatic grounding (i.e., get funding to continue USER work), they also allowed them to gain greater influence (through affiliations) in USER. With regards to coalition building, individuals appeared within the network partnered with one another to offset any perceived deficiencies their own human capital. Many of the examples encircled know - how (intellec tual capital) in USER. Luke referenced his inclusion in a formal network began when someone noticed his novel ideas at a conference and 138 wanted to include him. Andrew discussed how he started a formal network but had no idea on how to operate the group. In his example, he sought others in the emergent network to draw on their experience and help with his project. In this regard, the multi - cephalous emergent network put some of their proverbial heads together to achieve a goal of an organizational affiliation . Discussion Considering the findings outlined in this chapter, several emergent network functions among participants and informal governance structures arise from the analyses. Whereas functions describe the collaborative actions of those within the group , the different forms of capital construct an informal governance structure underpinning and informing how individuals interact with one another, and most importantly, establish an in and out group. In this section, I present a model of the emergent networ k function cycle using the findings gleaned in the earlier sections of this chapter. I construct the function cycle by outlining different aspects of the emergent network before presenting the model in its entirety. First, the emergent network zone (ENZ) is an open area which includes all individuals operating as leaders in USER. Those within the ENZ are individuals who are highly involved in USER - work and possess the necessary funder or association organizational affiliation (cultural capital), lan guage (cultural capital), resources (organizational capital), and relationships (social capital) to participate. Depicted in Figure 9, the ENZ creates a fluid boundary for leaders who operate in an emergent network capacity (i.e., exchanging knowledge, n etwork learning, innovation). Within the ENZ, interactions are influenced by operationalized human capital, and individuals interact at various USER functions. On the periphery of the ENZ boundary are others working in USER who either 139 are interested or disinterested in participating in emergent network activities. The area external to the ENZ also represents the organizational environment in which the ENZ interacts. Individuals within the ENZ interact with one another at various points throughout the year but are activated when a stimulus arises from the USER environment. As outlined by participants, this could be an issue with formal network operation, shortcoming in existing USER strategies, or an opportunity to access more funding. As the stimulus enters and spreads across individuals in the emergent network various processes begin simultaneously. When individuals encounter the stim ulus, they compare it with their domain - specific knowledge and Figure 9 . Emergent Network Zone 140 experiences and confer with others in the ENZ. This may include some or all of the participants but does not necessarily imp licate every individual within the ENZ. Those implicated individuals contextualize the stimulus through conversations and compare it with established domain - specific knowledge. This knowledge may target amorphous concerns arising from the need for systemic change in higher education or specific needs of formal USER networks. The value of the domain specific knowledge is dependent on its scarcity and the credibility of the knowledge - holder. This knowledge ranges from domain specific language (e.g., theory of change) and the alphabet soup of acronyms within the field, to the intricacies of transitioning a formal network from grant - funding to a more stable financial structure. These processes of knowledge exchanges among these leaders occur across USER at vario us meetings, steering committees, advisory Figure 10 . Emergent Network Activation 141 boards, or other small conferences that host high profile individuals in USER, which are only organizational affili ations and funding dollars. Simultaneous to knowledge exchanges, those individuals engage in consensus building. During this process, similar definitions, approaches, language, and understandings are negotiated (Figure 11). Conferring and sharing with each other illuminate either new opportunities for grants, proposals or potential holes in the emergent network zone. If those engaged in the emergent network zone believe they have enough credibility and content - knowledge, they may move forward with a proposed strategy to address a shor tcoming in USER. This may be a funding proposal for a new formal network, additional funding or processes for an existing formal network, or other initiatives. If additional insights, knowledge, or credibility is needed by those engaged in the consensus bu ilding they locate someone with the necessary credibility to Figure 11 . Consensus Building and ENZ Strategizing 142 add to the project. This person may be within the emergent network zone but may also be outside the group. Individuals outside of the emergent network zone who are recruited to join a project oft en undergo a socialization process, where they are brought up to speed with prior consensus established in previous conversations. Although these newly recruited individuals may be asked to partake in certain aspects of innovating, they generally lack the capital to renegotiate any of the previously agreed upon definitions or processes within the network. Following the consensus and coalition building in the emergent network, individuals within the ENZ create projects, initiatives, or other strategies. In many cases these new creations are new formal networks, strategies to improve formal networks, or publ ications from formal networks. In all instances, the innovations drew upon perceived shortcomings in USER and sought to share new approaches or innovations to address those shortcomings. The innovation aspect of the cycle is also the point in the emergent network cycle where grant funding may be Figure 12 . Emergent Network Function Cycle 143 requested and awarded. If the innovation receives a grant, the cycle continues with the innovation diffusing within and out of the ENZ to other STEM reformers. If the innovation requires a grant and is not funded, t he cycle begins anew. Individuals in the ENZ share their unfunded projects with others and build consensus and coalitions to increase their chances of receiving funding. After innovations are implemented in USER, all the information associated with the inn ovations creation and impact are diffused through social connections to the rest of the emergent network zone (Figure 12). In total, the emergent network within USER is a collection of individuals who are affiliated with higher education associations and f unding organizations and collaborate to generate strategies to help reform undergraduate STEM education. Their processes follow a cyclical pattern which originates from observed concerns in the USER environment and follows USER grant funding cycles. Notice , network learning is not included in this model. Despite it serving as a research - based function of emergent networks, few participants indicated any collective or organizational meaning making from the function cycle. Indeed, several participants (Hilari e, Lindsey, and Luke) all indicated their efforts were typically future - focused and considered what progress should be made with future endeavors. Based on the nature of organization networked - itiative or the emergent network. Conclusion The analyses presented here offer compelling evidence to the individual relationships and the functional capacities o f the web of relationships. Overall, the nominated USER leaders acted as a collaborative interorganizational emergent network through actualizing many of the 144 processes associated with interorganizational network activities. First, various manifestations o f human capital create an informal barrier among this group of leaders creating an in - group and out - group dynamic. Those in the in - group share information, build consensus, and engage the out - group through coalition building activities, before creating pro ducts, interventions or other networks to address challenges in USER efforts. Regarding the proposed research questions, the qualitative strand of this study targeted the following questions: 1b). How do leaders engage in knowledge diffusion regarding the ir networks? 1c). How do leaders engage in network learning? 2. How does this emergent network affect USER formal networks? The findings in this chapter argued for the knowledge within the emergent network zone to be composed of: (1) formal network expertise; (2) formal network outputs; (3) disciplinary background and expertise; (4) USER socialization; and (5) personal - relation al information. Knowledge was activated and moved throughout the social network at different formal events related to formal networks (e.g., advisory boards, steering committees, etc.) to address a targeted need within undergraduate STEM education reform. This chapter also discussed the absence of emergent network learning by any scalable metric. Although some individuals cited network learning as an aspect of their work in USER, their examples either directly pertained to their work in formal networks or l acked the specificity to constitute inter - organizational network learning. Nevertheless, many of the examples provided by participants did provide insights into how this group negotiated consensus and rationale for coalition building in USER. The effects of the emergent networks on formal USER networks is also layered. Many of the participants are founders, leaders, or advisors of formal networks. The emergent network 145 serves as a knowledge base and support structure for those in the leadership of formal n etworks and provides a group to strategize solutions for problems in formal networks and, at times, offer emotional support for those dealing with the problem. Additionally, the emergent network serves as a potential birthplace for new formal USER networks . In the concluding chapter, I integrate the findings from both strands of the study to present a holistic view of the emergent network structure, functions, and bylaws. I summarize the study before discussing its resonance with the literature base on emer gent networks and conclude with future research into the roles of collaborative interorganizational emergent networks in higher education. 146 Chapter 6: Data Integration The intention of this study was to further clarify the existence and role of an emergent network within undergraduate STEM education reform. Given the ambiguous and latent nature - methods design adopted a pragmatic methodological investigation (Creswell & Plano Clark, 2011, p. 2). Employing a sequential quan - qual research design provided multiple vantage points to begin decoding the actions of those within the network, and how the network operates as an organizational unit. Followi ng a multi - round nomination process, 17 participants completed a social network survey, and six individuals from the sample were subsequently interviewed. In accordance with mixed - method design, each method was selected to address a proposed research quest ion. 1. How do formal and informal leaders across formal networks in USER serve as an emergent network? 1a). How interconnected are leaders across formal networks? 1b). How do leaders engage in knowledge diffusion regarding their networks? 1c). How do leader s engage in network learning? 2. How does this emergent network a ffect USER formal networks? this final data chapter completes the intended research design by blending a nd reinterpreting the placement in the social network (establish ed in the social network analysis) and adds to the role of the undergraduate STEM education reform (USER) environment highlighted in interview research questions, before highlighting future research opportunities. 147 Strand One: Social Network Findings The first strand of this study employed quantitatively - focused social network analysis to map social structures embedded in the emergent network. Data collection inclu ded gathering - metric data on participant relationships (i.e., who knew whom). This strand used centralization and cohesion measurements to explore network - wide properties, cen trality metrics to measure individual participant characteristics, and Quadratic Assignment Procedure (QAP) regression modeling to identify potential patterns in network relationships. Individuals within the network were also highly interconnected with no structural holes and few information bottlenecks. Cohesion metrics showed one dense, reciprocal, unfragmented network, with three structural layers: a center, middle, and periphery. Also, a composed sociogram network - of - networks displayed these individuals represented nearly 20 different formal networks or other committees operating in USER. This eliminated the possibility of all the participants coming from a single formal network, university, or other organization. evealed complexity to each relationship as individuals generally interacted with the same people across different USER network concerns. QAP modeling yielded similar findings. Finding heterophily in organizational sectors and homophily in prior relationshi ps indicated individuals interacted across organizational boundaries and used existing social relationships to problem solve issues in USER. All findings in the first strand of the study indicated an open, densely connected social network of highly connect ed leaders in USER. More than simply connected, the findings from the social network strand quantitatively indicated complexity and depth to many relationships, meaning these relationships were more than simple acquaintances. 148 Strand Two: Interview Findings The second strand drew participants from the social network survey to conduct semi - structured interviews. Interviewed participants were selected based on their placement in the network. Two participants from the center, middle, and periphery of the networ k were individually interviewed in one - hour video conferences. Findings from the interviews presented more in - depth explorations on how the emergent network operates, the functions it serves in USER, and its relationship with the environment. After analyzi ng participants responses, definitions emerged for USER knowledge, knowledge exchanges, network learning, and innovation, and the emergent network was identified as an open system s organization operating as a multi - headed , semi - autonomous organism. Much like a hydra, individuals in the emergent network operated independently or in coalitions to generate innovations or other USER outcomes. Despite some autonomy, each of the metaphorical heads participated in a similar cycle of exchanging USER - related knowledge, building consensus and coalitions before proposing innovations. Following the assertion of the emergent network as an open system , environmental factors were explored as a means to give shape to aspects of power, infl uence, and informal emergent network governance. These findings yielded affiliations to funding organizations or higher education associations and financial resources provided power and influence within the (Jennifer) . The strand concluded with a conceptual model of an emergent network operating in USER. This model incorporated a porous border which interplayed with the environment and included the open system functi on cycle for those within the emergent network. 149 Data Integration In following the intended mixed methods research design, the final analysis of the study combined data and reinterpreted findings from both prior strands. Integrating the findings is critical in mixed methods designs as the process adds robustness to otherwise disconnected findings and creates a more comprehensive analysis of the phenomena in question (Creswell & Plano Clark, 2011). While integrating data from different forms of inquiry may ta ke many different forms (Creswell & Plano Clark, 2011), mixed methods designs have two ways of from the second strand to reinterpret the first (Creswell & Pl ano Clark, 2011, p. 71). This process methods deigns refers to a st The second form of data integration, therefore, uses the un - 20 11, p. 86). Considering this study, I present findings from both forms of data integration. Sequential data integration findings are presented first. This integration and discussion concern the findings from the social network analysis strand. Given new findings from the interviews, more context and understanding were added to the original evidence presented at the conclusion of the social network analysis. An emphasis - driven data integration follows the sequential data integration and highlights how the social network strand aids in the understanding of the open system , emergent network function cycle. 150 Sequential Data Integration As this integration form takes findings from the interview strand of study and folds them into findings from the social network study, this section primarily focuses on network structure and interconnectedness. Using centralization and cohesion metrics, reinterpreted findings are presented first at the network - wide structural level. This provides a more nuanced understanding of the network str ucture and individuals placement within the network. After concluding the network discussion, reinterpreted QAP findings aid in understanding dyadic level interactions. At the network structural level, participants were arranged by their average centrality scores into three categories based on their location in the network. Their locations were determined by averaging nine statistics spread across three networks, which yielded three structural zones. Although traditional network analysis bifurcates network structure by central and peripheral actors, a substantive break in actor betweenness measures necessitated a third Table 13 . Structural Location of Participants in Networks Structural Location ID Average Degree Average Closeness Average Betweenness Peripheral 11 (Allie) 4.67 31.00 0.00 2 5.67 29.67 0.10 8 8.33 27.33 0.14 12 (Luke) 8.00 27.33 0.24 9 8.00 27.33 0.70 Meso 17 8.67 28.00 1.67 6 (Andrew) 8.67 27.33 1.59 1 9.00 27.33 2.61 13 (Hilarie) 8.33 27.67 5.96 4 12.00 26.33 1.47 16 11.67 24.33 2.12 Central 15 (Jennifer) 13.67 21.67 4.40 7 11.67 24.33 4.77 5 9.67 23.00 4.91 3 12.00 24.00 4.54 14 12.33 23.67 5.07 10 (Lindsey) 13.33 22.67 8.71 151 structural location. This third, middle group, was termed the meso - level as the actors fell between the central and peripheral actors. Table 10 displays the structural location of participants and their structural location in the social network (derived f rom their average degree, closeness, and betweenness measurements across Business , Problems , and GaveAdvice ), and the interviewed participants are marked by their pseudonym. Each subsequent section discusses the properties of their structural location. Peripheral actors . Individuals on the peripheral of the emergent network were those who held the lowest average betweenness scores among participants. The term peripheral comes from the visual plotting in a sociogram where these actors are visual ly on the outside edges of a given network (Borgatti, Everett, & Johnson, 2018). This placement is always relative to others in a study. Peripheral actors are not viewed as internal network information brokers or internal information gatekeepers, and typic ally have less influence in social networks as they simply know less people within the network. Interestingly, Granovetter (1977) and Burt (2000) argued individuals on the periphery (or with weak ties) are uniquely situated to bring new information into a network. This placement gives peripheral actors more power to connect others in the network to information and people outside of the network, yet few individuals within the network with which to communicate. iphery encompassed five people including interviewed participants Allie and Luke (Table 10). These participants had the lowest average centrality metrics across Business , Problems , and GaveAdvice . Most notably, were their averages in betweenness. Network b etweenness is a statistic which describes how connected or between an individual is relative to others (Borgatti, Everett, & Johnson, 2018). All of the individuals in the periphery had the lowest betweenness and generally low degree and high closeness metr ics. 152 Given insights provided by Allie and Luke interviews, several interpretations emerge regarding the structural role of the periphery of this network. First, both interviewed participants indicated a time component of their work in USER. Allie continually as other hand, described himself as recen - formal network efforts. The large variation in time spent working in STEM was reflected by all participants in the periphery. Those five participants indicated working in STEM between 45 years (the longest) and the 8 years (shortest). The temporal component of the periphery adds complexity to the working understanding of its role and begins to describe an entry point to the emergent network. In his example of a cold - cal - funded project affiliated with a national association. By joining the grant, Luke became affi liated with two influential organizations in USER. This example provides two findings for the emergent network. First, the process reflects an avenue for individuals to enter the emergent network. Leveraging intellectual capital and knowledge related to th e chemistry field, Luke was able to affiliate with influential organizations and enter the emergent network. Second, his inclusion demonstrated a capacity of aspec ts of the emergent network function cycle and the hydra reaching into the environment so that its internal processes may be sustained. Actors who were retired or nearing retirement also occupied this space. Two of the participants in the periphery reported over 35 years of working in USER. Although they had 153 limited connections to the larger group, these older participants were tied to the most central actors, and not connected to those in other network locations. The two participants with the longest time in USER were in the periphery and were not connected to one another. The retired actors generally held an emeritus title affiliated wit h a formal network and generally did not participate in many of the knowledge exchanges. Instead, these individuals used personal communication (e.g., telephone and email) to stay connected with their friends. These relationships served as an information c onduit for both individuals involved. The retired actors provided advice on formal USER networks, context to historical outputs of networks, and were generally friends with the people with whom they were connected. Those with whom the peripheral actors wer e connected provided updates to the state of USER, and in one instance gave advice on formal network financial sustainability (i.e., formal knowledge expertise). times pan of individuals working in USER combined with the entrance and exiting dynamic presented by Luke and Allie point to a transient space of coming - and - going. Members within the periphery may be ramping - up their work in USER having recently entered the emer gent network zone, or scaling - back because they are retiring from their USER - work or fulltime employment. Central actors . In contrast to the peripheral, central individuals in the emergent network were those who held the highest averages of centrality sco res among participants. Given their high connection to other individuals, the term central refers to both the visual representation within a social network (i.e., in the center), and the critical role these actors play in facilitating network operations (i .e., communication flows, connecter) (Borgatti, Everett, & Johnson, 2018; Wasserman & Faust, 1994). Central actors in social networks are typically associated with high trust, influence, power, or popularity relative to others structural zones. In contrast to those in the 154 periphery, central actors are viewed to be efficient in communication, knowledge exchanges, and diffusion of information (as they are connected to many others), but these actors are often viewed as insular or otherwise unable to reach beyo nd the boundaries of the existing network(s) (Burt, 2000; Hawe, Webster, & Shiell, 2004). In this study, there were six central actors including interviewed participants Jennifer and Lindsey (Table 10). These participants all held the highest averages in t heir centrality measurements. Table 10 displays a clear break in betweenness metrics between central actors and other locations. With the exception of Participant 5, central actors also had more connections (degrees) to others, which indicated they simply knew many more people across Business, Problems, and GaveAdvice. All individuals in this location were administrators of high - profile higher education organizations, and most served in some capacity of formal USER network leadership (i.e., executive direct or, founder, president, etc.). Over the last 10 years these actors received 39 different grants from the NSF for over $ 2.75 million. This figure does not include 2 central actors because they were ineligible to receive funding due to employment conflicts of interest. Five of the central actors either work or previously worked in Washington, D.C. This group represents the core of the emergent network and epitomizes the emergent network function cycle. Their affiliations as administrators in high - profile org anizations combined with their demonstrated ability to get funding provides the organizational and cultural capital for them to serve as connectors in building consensus across the emergent network zone. This group is also split into two groups: funders an d high - interconnectedness demonstrates their power in the network and their ability to participate in all functions of the emergent network. The funders also participate in a different function of the emergent network - fu 155 emergent network function as it represents a cross - over between knowledge sharing, innovation proposals, and funding. Importantly, their roles as funders provide an official avenue f or them to continually review new ideas in the form of funding proposals. These roles ideally situate them for identifying potentially new people and ideas for USER and potential inclusion in the emergent network. High profile individuals share many simil ar qualities to the funders, with one notable exception: they are not funders. They have similar forms of influential organizational and cultural capital, which may come from their employer, numerous and sizable grants from the National Science Foundation, or both. These individuals participate in many functions of USER and are present on many steering committees, advisory boards, or other small gatherings outlined in Chapter 5. They are in the leadership or founders of at least one formal USER network. Hig h profile individuals participate in similar capacities in the emergent network but have different access to new knowledge and participants. Their primary organization roles (i.e., higher education associations or expansive, high profile networks) provide this the center of the emergent network with access to new information that may help formal USER networks or the larger USER efforts. Each of these findings present s Many of the central actors are (or have been) employed by high - profile organizations and funders. Aside from providing cultural and organization capital through association, their positions inherently co nnect them with many others working in USER. In fact, Lindsey stated it was inherently a condition of her employment. These connections allow the core to more actively interact with the environment than the formal definition of central actors indicate. Thi s also presents a new interpretation for the role of social capital. Whereas the discussion on capital manifestation in the interview strand dismissed social capital as a grantor of influence, the 156 through grant proposals provides power by knowing the USER landscape and potentially perpetuating their placement in the center of the emergent network. Central actors and their organizational affiliations aid in explaining the composition of the emergent network function cycle outlined in the interview chapter. The cycle of knowledge exchange, consensus building , and innovation mirrors actions individuals take in preparing a grant proposal. This cycle also keeps those who connect individuals and funders in the center of the emergent network as they often are those deciding who may get funding. Finally, central actor analyses also provide geographic hub of emergent network. As many of the central actors (and network - connectors) work in Washington, D.C., most of the remaining network members are tied to the city. This may translate to many knowledge exchanges occurring in Washington, D.C. at meetings, conferences or other forums. In sum, the integration of two strands demonstrates the critical role higher educ ation associations and funders play in the functioning of the emergent network. They are central. Meso - actors . As the name implies, the meso - structural location lies between the peripheral and central actor network regions. Individuals in this group had t heir average centrality scores fall within one standard deviation of the overall centrality mean. Traits in this area are less defined by literature on networks as there can be myriad reasons on why people are located there (Wasserman & Faust, 1994). Indee d, most bodies of research tend to dichotomize network structure into central and peripheral as it makes for easy comparisons (Borgatti, Everett, & Johnson, 2018). The distinction between meso - and other actors in this study is attributed to a major break between the peripheral and meso actors indicate a potential different role among these individuals in the network. 157 The meso - region was comprised of six individuals and included interviewed participants Andrew and Hilarie. With the exception of one actor, the meso - region was comprised entirely of faculty members. All faculty members in this space are on active NSF grants affiliated with formal USER networks. The non - faculty membe r in this location (Hilaire or P articipant 13) works as an administrator at a higher education association. The distinction in the meso - betweenness scores indicate greater communication and interconnectedness among these leaders and the central ac tors and less communication with those in the peripheral. be driven by a need for more formal USER knowledge. Both participants discussed the desire to access and imple ment new information to build or otherwise aid formal networks in which they were affiliated. Hilarie highlighted the importance of reaching out to peers to borrow ideas for the sake of precedent in funding proposals, and Andrew stressed his reliance on pe ers in negotiating the future of a formal network he led. For more in - depth discussion of their experience, please reference the interview chapter. The combination of mostly faculty formal network leaders in the meso - region presents an interesting case. In some respects, this region may include many established leaders in USER actively working and sharing information throughout their social networks to improve their USER efforts. When compared to the peripheral, those in the meso may be those with more capi tal (i.e., more funding and organizational affiliations). These individuals are likely doing the formal network work in running reform efforts within their formal networks. Compared to the central actors, those in the meso are those not employed or otherwise directly tied to higher education associations or funders. Meso - actors' comparatively lower human capital provides less 158 influence in the space, but still allows them to connect to one another on issues concerning formal network operations. Sequential data integration summary . Blending and analyzing the findings from the interview strand provide greater insights into the different structural locations established by the component related to their membership. Some individuals are relatively new to the emergent network, while others are long established and seeking to take a step back from their time in USER. Central actors, on the other hand, are individuals with a lot of cultural and organization capital. These individuals typically have employment or consulting experience with large national higher education funders or associat ions and serve in a self - capacity (Jennifer, Lindsey). Finally, the meso location appears to primarily be comprised of faculty members who are active on funded - grants and delivering institutional reform or services directly to institu tions of higher education. Applying these refined findings also aids in understanding how the emergent network operates from a process standpoint and gives insights into where specific knowledge exists in the network, and in part, the content of dyadic kn owledge exchanges. For example, knowledge exchanges between two central actors may pertain to how to better connect with others (e.g., individuals or institutions), while a knowledge exchange between a central actor and a meso - actor may pertain more to ope rationalizing a grant. This area needs more exploration, but these initial findings demonstrate a second level to the processes established in the emergent network cycle (as discussed in the interview chapter). 159 Emphasis - Driven Data Integration While the previous sections on sequential data integration add greater context to the social network analysis, the findings are limited by the integration process. Inherently, the sequential data integration process folds the interview findings into the quantitative social network data. This foregrounds the first set of findings and uses the second to complement and complicate the first. When using a quantitative method as the first strand, this has the effect of constraining qualitative data by limiting the stories participants tell (Creswell & Plano - Clark, 2011). In this study, the findings from the qualitative analyses were limited to explaining aspects of the social network structure presented in the first strand. While integrating in this way presented a fruitful analysis of the social network structure, it did little to explain other aspects of the emergent network studied. In contrast to sequential integration, emphasis - based integration completes a similar process but instead applies the un - findings to the emphasized. This process effectively flips the data integration by allowing the research to integrate the findings in a way that most fits the intended research design (Creswell & Plano - Clark, 2011). As mentioned, the interview (second) st rand of inquiry was the strand that carried emphasis as it was better suited to address many of the intended research questions. This discussion, therefore, is structured thematically by findings in the interview strand and uses social network findings to complement and complicate meaning derived from the data. The emphasis - driven integration and subsequent analysis relied heavily on findings gleaned in the QAP regression modeling discussion outlined in the social network strand. This predictive modeling i s a non - traditional regression method used to identify relational patterns embedded in a social network. In general, this process identifies when homophily (dyads share a 160 particular trait) or heterophily (do not share a trait) is common throughout the netw ork. With respect to the social network of the participants in this study, four specific findings were discussed in the social networks strand: a) complexity in relationships; b) organizationally focused dissemination; c) individually focused collaboration ; and, d) not statistically significant social identities. A more thorough explanation of these findings may be found in Chapter 4. Emergent network function cycle . The emergent network cycle outlined in the interview chapter is a continuous process beginn ing with an observed problem or shortcoming originating in USER efforts. Following the observation some leaders within USER begin a process of applying their knowledge, seeking new knowledge, finding consensus, and building coalitions. Once a proposed solu tion is agreed upon, the individuals act. Depictions of the emergent network function cycle are located in the interview strand findings. The QAP findings both reinforce and provide more context to how the emergent network function cycle occurs. The unders tanding that dyads exist across employment organizational boundaries as highlighted in the organization focused dissemination complement the notions of knowledge exchanges in USER. This is particularly salient when considering the heterophily in an individ mentioned they participate in many different USER related meetings which bring together diverse individuals from disparate post - secondary institutions, disciplinary fields, and se ctors (e.g., colleges, associations, funders, formal networks, etc.). The inter - organizational nature and boundary spanning aspects are furthered again by the structural density of the network. Each of the supporting findings gleaned from the social networ k analysis point to a multitude of organizational identities represented within the emergent network. These findings better capture 161 knowledge exchanges within the emergent network as the exchanges come from multiple cross - sector organizations. Secondly, th e QAP findings identified individually focused collaborations as a behavior of the network. The collaborations identified by the analysis support both concepts of consensus building and coalition - building in the emergent network function cycle. Briefly, th e QAP finding indicated homophilic tendencies based on time working in USER and employment position, and homophilic tendencies when discussing formal network problems. The distinction in these two behaviors point to participants building relationships differently dependent on the knowledge needed. In cases of discussing network problems, participants looked for those with different experiences. Hilarie and Andrew both highlighted this behavior when discussing their work with formal networks, underscoring the need to bring in new ideas or expertise to help their network achieve goals of receiving funding or financial sustainability. When it came to advice seeking though, the behavior appeared to be limited to those with similar experience. The act of seeking someone with similar experience for advice reflects some aspects of the consensus building processes outlined in the function cycle. As individuals en countered different problems in the USER landscape, they consulted others with similar experiences to clarify or make sense of the phenomena. Andrew spoke of this when he highlighted how his network was looking for financial sustainability opportunities. I n general, individually focused collaborations provides a bit more context to the traits of individuals engaging in coalition and consensus building activities, but more exploration is certainly needed in this area. Open system organization . Moving beyond the discussion on the function cycle, the second strand posited the collection of participants operated as an open system organization. Arguing the participants were bound to the emergent network cycle and vested human capital, 162 each person acted semi - auto nomously with the environment in their USER efforts. The goal of this emergent network - organization is to improve undergraduate STEM education. This organization was likened to a multi - cephalous organism, able to work on multiple and unrelated projects at one time. When placed in this context, three of the social network findings add more perspective to how this organization operates. First, the finding outlining complexity in relationships demonstrates the familiarity and interdependency of its members in serving the goals of reform. The interconnectedness and dependency are also supported by the network centralization and fragmentation metrics. As mentioned in the social network strand, complexity in relationship extends beyond simple acquaintances and ind icates the participants rely on one another for problem - solving and advice. These relationships bounded by the goal operationalized by the emergent network cycle all indicate a group of individuals working as a unit. Individually focused collaborations als o provide perspective on how the network may operationalize the emergent network cycle. Aside from indicating that individuals may work collaboratively, it showed small groups were the preferred method of working in USER. Indeed, no participant indicated a single instance where all participants were working together or jointly. Jennifer bemoaned this in her interview arguing the need for a centralized coordinating system for the efforts. Allie likewise agreed complaining that there were too many silos in US ER efforts, and people were talking past one another. Although the silos may exist in other parts of USER, the social network findings indicate that collaborations and knowledge exchanges were central to the functioning of the emergent network. These behav iors point again to the STEM hydra. Individual choose to exchange knowledge, ideas, and occasionally work together, but are semi - autonomous in how they operationalize the knowledge gained. The heads of the hydra though, are bound to the processes establish ed by central actors and high - profile USER organizations. 163 Despite many of the social network findings supporting the open system organization, some questions still remain. Namely, other than the significant co - variates listed in the QAP, what characteristics or conditions are necessary to drive collaboration among participants? Does the emergent network ever act in unison toward a specific goal more nuanced than USER? Similar questions exist regarding the how the emergent network cycle formed. Conclusion This chapter completed the research design by integrating the findings from the two previous strands of inquiry. Although the social network analysis demonstrated the interconnectedness and social structures of the network, it alone could not unveil the valued aspect of human capital or how power was distributed throughout the network. The social network analysis, fu rthermore, could not elucidate the contexts of the relationships beyond the survey prompts. Through the sequential data integration, findings from the interviews provided data to further explain these relationships, and their contact frequencies. Indeed, m any of the participants are more than simple acquaintances, and the structure of the emergent network indicates power is centered with funders in Washington, D.C. The emphasis - driven integration similarly combined data by using the interview data as a lens to re - interpret the QAP findings in the social network strand. Findings in knowledge and knowledge exchanges gave greater depth to concepts of knowledge transfer outlined in the regression models by providing both 1) what knowledge was exchanged , and 2) w here and when the knowledge was exchanged. Following the data integrations, I concluded with a discussion on how the behavior of the emergent network aligns with aspects of open systems organizational theory. I contended each of the participants acted sem i - autonomously to serve their network and USER goals but were bound 164 to one another through social structures. These social structures mirrored the emergent network behaviors and capital outlined in the interview strand. In the final chapter, I revisit each of my research questions, provide evidence to address them and offer both theoretical and practical implications for scholars and practitioners alike. 165 Chapter 7: Implications and Conclusion Returning to the research questions established at the onset of the study, the questions sought to establish the existence and role of an emergent inter - organizational network in USER. The first question identified individual behaviors indicative of an emergent network, and each of the sub - questions investigated diffe rent literature - supported functions of an emergent network. The second research question targeted impacts of this group on other USER efforts. Specifically, the research questions asked: 1. How do formal and informal leaders across formal networks in USER se rve as an emergent network? 1a). How interconnected are leaders across formal networks? 1b). How do leaders engage in knowledge diffusion regarding their networks? 1c). How do leaders engage in network learning? 2. How does this emergent network a ffect USER formal networks? Following multiple rounds of data collection, analysis, and integration, the findings for this dissertation are a bit overwhelming. Indeed, Creswell and Plano - Clark (2011) argued mixed - methods design often complicate and obscure findings. In order to better clarify the evidence elicited by the research, I provide data to address each question directly. The following sections consider a specific question (or sub - question) and couple the question with evidence. I begin by outlining the sub - questions. How I nterconnected are L eaders a cross F ormal N etworks? In attending to the first sub - research question, data in the social network analysis provided ample evidence to address questions regarding leader interconnectedness. The results fr om a social network survey revealed a high degree of interconnectivity across all participants. These findings included high network density statistics, no fragmentation, and low betweenness centralization measures. The centralization measures were consist ent across all three of the social 166 networks investigated (i.e., Business, Problems, and GaveAdvice ). On average, the participants were directly linked to 70% of the others in the study. Each of these statistics point to a highly connected group of individu als. Translating the statistics beyond the numbers and percentages, the social network survey indicated the participants knew most of the other participants reasonably well. Most participants communicated with one another regularly, and often saw one anot her semi - regularly in various meetings in Washington, D.C. These relationships ranged from acquaintances to trusting friendships, and in many cases, participants knew details about others personal lives. With regards to USER - related work, there were few st ructural bottlenecks for USER - related information to get trapped. The lack of network fragments in the denoted USER knowledge freely flowing throughout the network. In short, almost everyone was in - the - know with regards to USER work and operating from a si milar understanding on the current state of USER. A potential explanation for the high degree of connectedness among the participants would be that all of them belong to the same formal network. If true, this formal network would explain why all participan ts knew one another and calls into question the existence of an emergent network across USER. In fact, the findings outlined in the dissertation could all be formal network. To address the potential pitfall of dismissing the findings, all participants were surveyed on their network memberships. Although some USER formal networks were more prevalent than others, no single network could be linked to all participants. Instead th e 17 participants provided almost 20 different formal networks, which were populated into a social network. This Network - of - Networks (discussed in Chapter 4) was less densely connected than the other social networks highlighted (i.e., Business, Problems, G aveAdvice ). The high degree of 167 interconnectedness across participants combined with the low degree of interconnectedness in the Network - of - Networks pointed to a collection of leaders who represented their organization. How do L eaders E ngage in K nowledge D iffusion? Concerning the second sub - question, findings gleaned from the semi - structured interviews provided insights into what the participants considered to be knowledge, why knowledge is moved across the network, and when it is shared. Partici pants generally defined knowledge as the possession of critical and scarce information that may aid USER efforts. Knowledge fell into five different domains: (1) formal network expertise; (2) formal network outputs; (3) disciplinary background; (4) USER so cialization; and (5) personal information. These different knowledge domains were valued by all participants regardless of their structural location in the network, time spent working in USER, or other discernable variable s . The first two knowledge domains (i.e., formal network expertise and formal network outputs) focused on the work of the formal networks and relied on individual participants sharing knowledge through their experience working with formal networks. This may have been using previous governa nce structures or network programs. The latter three domains were more individually focused. Participants discussed seeking individuals who held knowledge about a discipline or highlighting information they encountered through working in USER for an extend ed period of time. Drawing again from the interview data, knowledge diffusion occurred through personal exchanges among participants in various mediums. These actions were termed knowledge exchanges and were signified by participants discussing any of the knowledge domains. Knowledge exchanges occurred primarily in three venues. Featured most prominently in this study were conversations held in formal, closed - door meetings pertaining to USER grant - work, most often cited as either steering committee or advi sory board meetings. In these meetings, 168 various leaders from across USER came together to discuss concerns with networks and at - large USER issues and were often comprised of different combinations of participants. The second venue identified were small gra nt - funded USER conferences. These venues provided an opportunity for the leaders to update one another on their network affiliations and loosely discuss strategies for aiding formal networks. Third, participants discussed personal communications like email and telephone calls. This was mentioned less frequently than the other methods and presupposed a level of familiarity. Findings from the social network strand also support a high degree of knowledge exchange and a general ease of information diffusion. As mentioned, the participants produced an interconnected social network. Burt (2000) associated high degrees of interconnectivity with Essentially, everyone k nows one another and has the ability to stay updated on what others in the network are doing. Coleman (1990) also argued closure facilitates greater information diffusion across social networks because individuals trust one another and view knowledge shari ng as individuals know personal information about others as it may signify a social relationship beyond the formal ties of organizational membership. In summary, th e social network findings provide a conceptual understanding of the social structures in which knowledge flows through the network. This indicates the mechanisms of how knowledge is diffused. The network of participants is dense, and information flows freely across the members through their social connections. Findings from the interviews provide more dimension to what knowledge is moved, and when or where the information is operationalized. 169 In that discussion, the five knowledge domains are moved through the network and during various meetings of different USER leaders throughout the year. How do L eaders E ngage in N etwork L earning? With regards to emergent network learni ng, the findings were more nuanced than for the other sub - questions. On the surface, few interviewees articulated aspects of network learning beyond simply wanting to improve undergraduate STEM education. Indeed, the organized aspects of network learning m ay be more inherent in formal networks, which may have a more explicit mission focused on network learning and improvement (Popp, et al., 2014). The social network findings also provided little evidence o f network learning. Although the findings indicate a social structure for network learning to occur, there was no socio - metric statistic to indicate active feedback loops. In response to these findings, the role of emergent network learning in USER remains unclear; however, the lack of formal organizing str uctures and amorphous nature of the open system organization point to the actions occurring dyadically, not systemically. In lieu of network learning, participants indicated other social organizing activities which occurred within the network. Interviewees frequently discussed their roles in building coalitions and consensus with others operating in USER. Members of the emergent network sought input from others to make - sense of the phenomena occurring in USER efforts and formal networks. This often took pla ce through conversation with others at conferences, steering committees, or other emergent network knowledge exchanges. Participants also discussed building coalitions as an aspect of their work. Coalitions were primarily sought by participants in the inn ovations the emergent network put forth. Most prominently, this was featured when emergent network 170 members were co - authoring a grant proposal or initiative to address some problem perceived in the undergraduate STEM education landscape. How do L eaders acr oss F ormal N etworks in USER serve as an E mergent N etwork? Recalling the opening sections of this dissertation, the study sought to identify and explore the existence of a collaborative emergent inter - organizational network operating in undergraduate STEM intention. In serving this question, I define each component of the collaborative emergent inter - organizational network and provide evidence which support or negate its existence among the participants. together to provide a public good, service, or value when a single institution is unable to create p. 157). The collaborative nature of those working in USER fit within the collaborative network definition, as they often enter networks as willing independent actors trying to help their institutions address the systemic problem of USER. Indeed, many part icipants discussed the collaborative work completed in the service of formal as an administrator, and Hilarie talked about her work with higher education associations stating, rmation to our university their views the collaborative work was two - fold. First, they worked as a group of reformers assisting colleges and universities to implement better undergraduate STEM education practices. Second, they worked as a group of leaders in formal USER networks assisting in better practices 171 w ithin the networks. While these collaborations targeted different audiences with different messages, they still fit with the collaborative network definition as these independent individuals worked together to help different organizations. Second, inter - or ganizational simply refers to an entity comprised of actors representing different organizations. Literature indicated inter - organizational networks were a preferred icked participants represent multiple organizations. First, no two participants were employed by the same institution. In fact, no sector of higher education (i .e., research university, liberal arts college, etc.) comprised the majority of participants. Table 2 in the social network analysis chapter highlights the employing organization breakdown. Secondly, each participant indicated they represented multiple for mal inter - organizational networks. Participants identified 19 different formal networks in which they belonged, thus adding a second level of inter - organizational boundaries to the study. Table 3, Table 5, and Figure 8 in the Social Network Chapter display the relevant formal network inter - organizational statistics. The third term, emergent network , is more difficult to define and identify. The ambiguous and amorphous nature of social relationships often muddle or complicate individuals acting as friends, c olleagues, confidants, or in a networked - capacity. Although available research on emergent networks is relatively limited, several authors provide different frameworks to identify their existence in social networks. Isett et al. (2011) described emergent n etworks as loose needs (p. 162). Popp et al. (2014) indicted emergent networks were informal collections of individuals who respond to large systemic problems an d address organizational needs. Several 172 authors provided approaches to identify emergent networks, but many lamented the difficulty of 2007, p. 325). Popp et al. (201 4) pushed a strategy of identifying emergent networks by how the social network behaves. They argued recognizing emergent network functions (i.e., knowledge diffusion, network learning, and innovations) within an inter - organizational social network would i ndicate that an emergent networks existence. These functions underpinned the sub - questions, whose findings were established in their respective sections. As a brief review, the leaders work as an open system s organization structured with norms established by organizational affiliations (cultural capital). These social structures were termed the emergent network function cycle and are depicted in Figure 12 in the interview chapter. In brief, the leaders in USE R continually share information about observed problems in USER, their USER initiatives, formal networks, or opportunities with one another. These knowledge exchanges happen in many different venues, but primarily occur in steering committees, advisory boa rds, or other small USER meetings throughout a given year. The knowledge exchanges provide a space for leaders to share information, wrestle with consensus, and strategize paths forward for both formal networks, and USER in general. Throughout their time w orking in USER and with knowledge exchanges, the leaders work in tandem to confront their concerns, and build a common approach or understanding to address their concerns. Once new approaches, strategies, or even new formal networks are implemented, they b ecome a new form of knowledge embedded in the social networks of the emergent network and may be activated to address future problems. In sum, this collection of leaders serves as a collaborative group of individuals working across many sectors of higher e ducation and higher education advocacy to offer services to 173 reform undergraduate STEM education. They act as an open, loosely connected, organization who share knowledge related to USER, lean on one another for meaning - making, and co - construct strategies o r innovations for pushing USER forward. The evidence provided in the findings herein support the existence of a collaborative inter - organizational emergent network operating across leaders of formal networks in undergraduate STEM reform, who perpetually cy cle through knowledge sharing and innovating. How does this E mergent N etwork affect F ormal N etworks in USER ? The final research question concerns the impact of the emergent network. Despite highlighting the existence and role of an emergent network in the previous sections, the effect of this network is not explicitly articulated. Results from the interviews presented several outputs which both directly and indirectly affect formal USER networks. The outputs were: (1) new formal networks; (2) proposals; (3 ) formal network governance structures; and (4) formal network sustainability strategies. Directly affecting formal USER networks, formal governance structures and sustainability strategies presented ways formal networks could more effectively organize to grow and operate. New formal networks and proposals were additional outputs from these participants that created new opportunities for inv olvement and brought more individuals into USER. Beyond the scope of research questions, evidence suggests the emergent network function cycle operates differently at different levels of the emergent network. At the periphery of the social network, individ uals are either entering or exiting the network. These individuals rely more on their social capital as a means of holding influence and seek to grow their organizational and cultural capital through USER socialization, new formal networks, or other grants . If they are exiting, they remain close to specific influential central actors with high 174 organizational and cultural capital. Central actors are primarily funders or actors working at high profile institutions. These individuals leverage their cultural ca pital, organizational capital, social capital to attract new knowledge to the space and connect otherwise disparate actors into the emergent network functions. Central actors are often ex - officio attendees of the various and numerous meeting venues for the emergent network. People in the meso - level of the social network are primarily faculty involved in the leadership of formal networks. These individuals confer with their central - actor funders and a few of the peripheral actors for more information on how to properly organize and operate their networks. Despite the interconnectivity, different capital manifestations in the interviews demonstrated a form of hierarchy within the network. Based primarily on organizational affiliations and funding dollars, larg e funding organizations bestow credibility through legitimizing proposals in the space. Funders serving as the central actors supports this point. Organizational capital, in the form of position affiliation and funding, has a large role in how individuals in the network operate. The resources from organizations carry influence in allowing different leaders to pursue their proposed ideas. Cultural capital, primarily in the form of linguistic and symbolic capital, similarly carry weight through organizational affiliation by lending organizational legitimacy to those affiliated with the organization. In sum, the participants in this study serve as a collaborative inter - organizational emergent network. Despite originally establishing the bounds of the study as an emergent network across formal networks in USER, this emergent network operates as a group across USER. Theoretical Implications and Contributions Aside from highlighting the dearth of research, few empirical research studies mention, much less explore , emergent networks and their function (Isett, et al., 2011; Popp, et al., 2014). 175 Whereas some works target the network structures (Ahuja, Soda, & Zaheer, 2012; Provan & LeMaire, 2012; Uzzi, 1997), many of these sought to map emergent networks through unbo unded social network analysis across multiple domains. To date, this is the only known study applying multiple methods to interrogate the structural and enacted functions of an emergent network in higher education. As such, its findings are a first step to wards establishing a clearer picture of how emergent inter - organizational networks operate. The results from this research show many different layered processes embedded within the informal - relational structures of emergent network. They also suggest the n ecessity of complex relations to form an emergent network. More importantly, this work adds to the existing knowledge of emergent inter - organizational networks in four fundamental ways. Capital and embedded networks . First, both strands of the study highli ghted the roles of social, cultural, organizational, and intellectual capital in the emergent network. In short, an socialization and social capital) impact how individuals engage in the network and its functions. These forms of capital govern how individuals interact, who seeks whom within a network, and why. In part, various forms of capital also implicate the structure of the network. As noted in the data int egration sections, forms of intellectual, symbolic - cultural, and organizational capital inform aspects of the peripheral, meso, and central actors of the network. Objectively, concepts of capital seem reasonable, as individual reputations and organizationa l legitimacy are not left at the proverbial door when networks convene. Most work on inter - organizational networks focuses on organizational issues, cost - benefit analysis, reasons for convening a network, or how to effectively manage a network from formation to sunset. Future 176 studies should further explore the roles of capital deployment, particularly in emergent networks, where formal bylaws do not govern individual engagements. Further, this study departs from the literature in highlighting an inter - organizational network spanning inter - organizational networks. Tangential to the conversation on the role of capital, this study introduces the concepts of multiple organizational affiliations. Whereas traditional inter - organizational network explorations studied formal networks (Isett et al., 2011; Popp et al., 2014), this resear ch begins to uncover a network embedded within inter - organizational USER networks. Situating a study in this fashion provides insights into the role of individual capital, but also may suggest formal network self - perpetuation or disparate bureaucratization hosted in the informal relationships of lead participants, where individual knowledge brokers form relationships with power - holders to perpetuate their capital. This finding is certainly supported by the finding of these leaders producing more networks an d proposals. Neo - institutionalism . Tenets from this work also add to the literature on neo - institutional theory of organizations (DiMaggio & Powell, 1983). This theory posits environmental peer pressure (also referred to as isomorphism) pushes individua l organizations to become origins in three forms of isomorphic behavior: coercive, normative and mimetic. Although any isomorphic behavior can push conformity, all three can be present in a particular organization. In this framework, organizations seek environmental legitimacy as it bestows the power for an organization to achieve its goals (DiMaggio & Powell, 1983). There are several aspects of neo - institutional t heory apparent in the emergent network findings. Several participants highlighted the knowledge sharing across the emergent network to 177 aid in the founding or development of formal networks. The primary driver cited for this action was to gain resources fro m funders. This mimicking of formal networks for greater network sustainability reflects mimetic isomorphism occurring across the emergent network. Similarly, many of the value structures, knowledge forms, and disciplinary backgrounds inform formal network structures. Many of the formal netw orks cited in this study either began or continue to maintain a grant primary - investigator form of leadership, and reflect legitimate normative behaviors associated with structure in the STEM disciplines. Finally, Jennifer captured the final form of coerci ve isomorphism when discussing the role of funding organizations in USER. She se their power by awarding funding to certain primary - investigators or networks. Funding these actors pressures those in the resource environment to act similarly or risk funding. l networks participants (emergent and mandated) seek legitimacy in their organizations through funding to achieve their goals. In this sense, cross - organizational legitimacy. The prevalence of iso morphism, or the convergence of behaviors and practices, in USER seemingly calls into question the innovative rationale for creating inter - organization networks. If formal networks are conforming to legitimate and normalized behaviors, their capacity to th ink outside the legitimacy box may be inhibited. Nevertheless, organizational affiliations and dollars may provide the necessary social leverage for individuals and organizations to effectuate change in USER. Indeed, the literature is undecided on how isom orphism and innovation intersect, so future research in USER should further investigate the effects of isomorphism on reform efforts. Given the exploratory nature of this study, future 178 research should further probe manifestations of capital, legitimacy and the roles of isomorphic behavior in USER efforts. Methodological contribution. Recognizing the nuances of relationships was critical to the findings in this study. The methodological approach used provides more robust findings for emergent networks. Prio r network research either employed social network analysis to define experiences in networks (Provan & LeMaire, 2012). Although each of those approaches provid ed some insights into the functions and roles of networks, they were generally limited by their methodology in that they only investigated small pieces of an emergent network and were unable to more fully describe the whole. Whereas social network analysis could inform interconnectedness (Ahuja, Soda, & Zaheer, 2012; Uzzi, 1997), actor selection (Daly, 2018), or actor influence (Daly, 2010), the approach could not inform the content of information flows, or how capital is actualized in a network. Social net work analysis is also limited by the dynamic phenomena, but those experiences may lack resonance across the social network. Plainly, people experience and define re lationships differently. Using a combination of nominations, social network analysis and interviews to interpret, analyze, and reinterpret the data provided an abundance of findings. Investigating emergent networks with a mixed methods design provided mult iple avenues to avoid the shortcomings of previous research. but also provided reliability and triangulation measures beyond those married to specific methodologies. For example, the interconnectedness of actors found in the social network analysis was con firmed and expanded upon in the qualitative strand of the study. Instead of relying on traditional balance or robustness tests for validity with the social network data, I 179 asked the participants to further detail their responses. Similarly, the social netw ork analysis provided a more complete picture of actor connectedness beyond what any individual could explain or draw. Future research into emergent networks should rely, and improve upon, the mixed methods approach used in this study. Undergraduate STEM e ducation reform . In terms of USER - specific implications, this study offers insights into the deployment and use of formal networks operating in undergraduate STEM reform. As mentioned in the introduction, inter - organizational formal networks are a popular organizational strategy used by USER funders to leverage change in higher education. larship on their role and functions considerably lags behind their use (Isett et al., 2011). Although scholarship on formal networks has begun to better develop our understanding (Gehrke & Kezar, 2016; Kezar, 2014), this study provides insights into the so cial mechanisms that drive, shape, and support formal networks operating in USER. Knowledge of the existence and role of the emergent network should inform funders, faculty, and other affiliated individuals on how they limit, perpetuate, or substantiate e fforts in social relationships before any formal actions occur. These findings may also aid actors in USER better identify their processes and information flows for better inclusion in emergent network discussion. Practical Implications and Contributions T he conclusions from this study also have tangible implications for those working to reform undergraduate education. Given inter - organizational networks implicate a cross - section of organizations and sectors, the findings from this study resonate with a bro ad range of stakeholders. As stated, formal networks are strategies used by stakeholders in higher education to improve undergraduate STEM education. Although not strictly organized or funded, emergent 180 networks co - exist with the formal and offer benefits t o those working in reform. The emergent network acts as an open - system organization which leverages change through several actions. These functions promote alignment across USER efforts by engaging USER leaders in (1) knowledge exchanges, (2) consensus and coalition building opportunities, and (3) innovations. Emergent network functions transcend traditional organizational and formal network boundaries and allow those involved to communicate across USER regardless of sector or organizational memberships. Th e following sections highlight implications for how various stakeholders may be more intentional in recognizing and incorporating the emergent network into the USER efforts. Undergraduate STEM education reform funders. Given the findings of this study, th ere are several considerations for organizations which provide USER funding. Much of the funding efforts in USER have been dedicated to creating and sustaining formal networks. While these efforts are important to synchronizing and diffusing reform efforts to member campuses, little attention has been given to the informal relationships generated through these formalized networks. Indeed, the findings from this study support the co - existence of an informal, emergent network which operates alongside, and sup ports, the operations of their formal counterparts. Funders should be more cognizant of this social phenomenon, however somewhat paradoxically, they should not move to formalize the emergent as a formal network. The transient nature of individuals in the network combined with the organic relationships found therein limit the ability to mandate that type of collaboration. Instead, funders should seek opportunities to enhance the functions of the emergent network. For example, the emergent network function cycle begins, and is sustained by, continual knowledge exchanges. Funders should continue to foster opportunities to bring together small groups of leaders in USER through roundtables, small 181 conferences, or task forces. Funders should also replicate conven ings similar to the grant advisory board meetings or formal network steering committees comprised of formal network leaders to discuss current concerns in USER efforts and strategize potential strategies. These convenings should be small, but with a pre - se t agenda to focus participants on specific aspects of change efforts. - grantee relationships. Grant funded projects (or even formal networks) have the potential to build inter - personal relationships with participants and generate more USER innovations . Indeed, the creation of a new formal network may be the result of an emergent network operating across USER. Second, connections with funding organizations provide human capital to grantees. This capital allows grantees to interface with others in USER m ore regularly. Given this phenomenon, funders have the power to bring individuals to the proverbial table and participate in the emergent network functions. Formal network administrators. Individual formal network leaders, staff, and administrators also may utilize the findings of this study. Several participants directly cited one role of the emergent network was to strategize and support formal network improvement. Support took the form o f assisting formal networks with financial sustainability, leadership transitions, organizational transitions, or formal network governance concerns. The emergent network offers a wealth of knowledge, consensus, and innovations to aid formal networks. Form al network leaders should be aware of the emergent network and the potential benefits they offer to network stability. In actuality, many formal network leaders may participate in the emergent network already but may not label the interactions as such. For mal network administrators should use the conclusions of this study in two distinct manners. First, recognize individuals in their formal 182 network who may have connections to the emergent network, and utilize their connections to aid their formal network. S econd, formal network leaders should encourage participation, consensus, and interpersonal development of those within their network. For larger networks, this may mean creating opportunities for the network to feel smaller to participants (e.g., break - out groups, roundtables, committees, etc.). Developing these relationships within a formal network will lead to more network stability, but the relationships generated may endure to be part of a future USER emergent network. Reform - minded faculty . Individual faculty members also have several factors to consider in light of this study. The emergent network is an amalgamation of reform - minded faculty who have actively engaged in developing relationships with others working in USER. Faculty who want to become inv olved in the emergent network should seek to fully engage with their formal network affiliations, build relationships with those involved and assist with formal network operational concerns. Gaining experience in formal network operations may lead to indiv iduals in the emergent network seeking to incorporate that specific knowledge into their processes. Limitations Although careful attention was given to the research design, data collection, and analysis, several limitations were inherent in the process of conducting this study. The first potential limitation concerns the nominations process. Although multi - round nominations processes are viewed as a mechanism for avoiding selection bias in participants, they are not free of limitations or confounding bias es (Avella, 2016; Henderson, 2018). Delphi M ethod limitations broadly fit within three critical themes: (1) lack of clear guidelines; (2) lack of continued participation; and (3) no evidence of reliability (Avella, 2016). Attending to the lack of proper g uidelines, the 183 nominators were provided specific instructions on how they should interpret the questions posed. The second limitation was not of concern, as 100% of the nominators participated in the second round of the process. The third limitation does p resent an area of concern. Despite using the Delphi M ethod, few strategies exist to mitigate nominators from conferring with others on whom they should nominate. This may bias some of the reliability and validity of those nominated to participate. Although some cross talk may occur, it would be remiss not to acknowledge some bias with the process. As with any study in a relatively explored arena, this work begins inquiry into emergent networks and to inform more robust analyzes in the future. Another limita tion concerns social network response rates. Whereas more traditional quantitative analyses maintain a lower threshold for responses, the interdependent relationships embedded in social network requires near 100% participation (Wasserman & Faust, 1993). Ni neteen initial nominees met the threshold for inclusion in the study, but only seventeen elected to participate. The two non - consenting participants represent a potential hole in the social network. Of the 17 consenting participants, their response rates f or the social network survey was near 90 percent. Multiple strategies were employed to identify non - respondents' place in the network(s) including using symmetrized data and transposed adjacency matrices (Borgatti, Everett, & Johnson, 2018; Morrie & Deckro , 2013). F uture R esearch The findings in this study confirm the existence of a collaborative emergent inter - organizational network operating in undergraduate STEM education reform. These findings indicate a highly interconnected group of individuals who carry power in the form of funder affiliations and grant dollars. They confer in various closed - door forums about the state of USER, and often create additional networks or other initiatives. Despite the abundance of 184 evidence indicating an emergent network, several themes were not included in this work as they did not address the research questions or were not fully defined through the interview data. More work should be conducted to identify additional individuals in the emergent network zo ne. Interviewed participants indicated between 30 50% of formal USER network leaders were represented in this study. Subsequent studies should include or incorporate their social connections and experiences as they pertain to emergent network functions. T his exploration offers several additional routes for further inquiry. Exiting the emergent network is not addressed in this study. Emergent network boundaries are undoubtedly fluid, with individuals entering and exiting regularly. As this study focused on individuals within the emergent network, it is limited to those who were considered leaders at the time of the study. Questions about the process, conditions, and experience of exiting an emergent network are still outstanding. Research interrogating this phenomenon will almost certainly uncover more information on the role of human capital deployment in the emergent network (i.e., what happens when an individual is no longer grant - funded or otherwise unaffiliated), and the intentionality of the exiting pro cess. More in - depth work must further interrogate themes identified in Chapter 5, particularly USER socialization and consensus building . Although this study identified the processes in the emergent network, it fails to provide greater depth or context to the processes. Conclusion Inter - organizational networks continue to be a strategy employed by organizations across Although the benefits and challenges associated with the formalized networks are becoming clearer, research on emergent networks, embedded in social relationships, remains sparse and limited to specific domains. Just as formal networks exist across multiple disciplines, emergent 185 networks also exist, alb eit more tacitly. Indeed, emergent networks very well may be the original inter - organizational network prior to any formalized and funded effort. The findings and subsequent implications highlighted in this study indicate emergent networks as a guiding han d, in many ways shaping reform efforts in undergraduate STEM education reform. The emergent network functions to produce and share knowledge, establish consensus and meaning, before driving new approaches to reform efforts. Despite the abundance of informa tion uncovered from this study, more research must be conducted to fully uncover the network way of working . 186 APPENDICES 187 Appendix A: Social Network Survey Start of Block: Introduction Greetings Participant , Thank you again for offering your time to this project. As previously mentioned, this study explores the interconnectedness of highly involved leaders working in undergraduate STEM education reform networks . The following survey will prompt you to provide some basic demographic information. These demographic questions are optional but your contributions will greatly add to this study's richness. Following the demographic questions, the survey asks you to indicate your level of acquaintance with each of the other nominated leaders in undergraduate STEM education reform networks. All of your submissions will be kept anonymous . If you have any questions or hesitations, please contact Levi Shanks at shanksle@msu.edu at your earliest convenience. Cordially, L evi What is your preferred gender identity? o Man o Woman o Preferred Gender Identity ________________________________________________ o Prefer not to respond 188 Please indicate your age range. o Under 30 years old o 30 - 39 years old o 40 - 49 years old o 50 - 59 years old o 60 - 69 years old o Over 70 years old o Prefer not to respond 189 With which racial or ethnic group do you identify? (Choose all that apply) Native American or Alaskan Native African American East Asian or Asian Americ an Caribbean / West Indian Indian (Subcontinent) or Indian American Latinx or Hispanic Middle Eastern Multiracial Native Hawaiian or Pacific Islander White Prefer not to respond From what institution did you attain your terminal or most recent degree? ________________________________________________________________ 190 Please indicate the institution type in which you are currently employed. o Two - year Community/Technical College o Four - year Private Primarily Undergraduate College or University o Four - year Regional Comprehensive College or University o Four - year Public Research University o Four - year Private Research University o I do not work at a college or university. End of Block: Introduction Start of Block: Block 1 Are you currently involved with any undergraduate STEM Education Reform Networks? If you are unsure, please check the sample list below. o Yes o No Please indicate between 1 and 5 Undergraduate STEM education reform networks in which you are currently involved, or have been involved with in the last 5 years. The order is not important. ( NOTE: If you are involved with more than 5 networks, please list the 5 in which you are most involved). o Network 1 ________________________________________________ o Network 2 ________________________________________________ o Network 3 ________________________________________________ o Network 4 ________________________________________________ o Network 5 ____________________ ____________________________ 191 Suggested Undergraduate STEM Education Reform Networks to consider: (NOTE: Your responses are not limited by this list, it is provided simply to aid your brainstorming.) Accelerating Systemic Change in STEM Higher Education (ASCN) Association of American Universities (AAU) STEM Network Initiative The Bay View Alliance (BVA) BEACON BioQuest Center for the Integration of Research, Teaching, and Learning (CIRTL) National Academies Roundtable on Systemic Change in Under graduate STEM Education Network of STEM Education Centers (NSEC) Partnership for Undergraduate Life Sciences (PULSE) Process Oriented Guided Inquiry Learning (POGIL) Project Professional and Organizational Development Network in Higher Education (POD) Project Kaleidoscope (PKAL) Science Education for New Civic Engagement and Responsibilities (SENCER) Science Education Resource Center (SERC) End of Block: Block 1 Cons idering the columns below, please indicate the individuals with whom you interact regarding undergraduate STEM education reform network related business. If you have not interacted with that person, please leave their selections blank. (NOTE: These indivi duals do NOT necessarily need to belong to the same network(s) as you). Spoke about STEM network business (e.g., network initiatives, functions, or agenda items) Discussed STEM network problems (e.g., funding, sustainability, network scaling). Sought advice from this person to aid my network. Participant Redacted Participant Redacted Participant Redacted 192 Participant Redacted Participant Redacted Participant Redacted Participant Redacted Participant Redacted Participant Redacted Participant Redacted Participant Redacted Participant Redacted Participant Redacted Participant Redacted Participant Redacted Participant Redacted Participant Redacted End of Block: Block 2 Start of Block: Block 3 Considering the prompts from the previous question, are there any leaders in undergraduate STEM education reform networks with whom you regularly interact that were NOT listed ? o Yes o No 193 Please nominate 1 - 2 person(s) you believe should be included in this list. o Nominee 1 ________________________________________________ o Nominee 2 ________________________________________________ End of Block: Block 3 194 Appendix B: Project Overview for Participant Research Project Summary This research study proposes to investigate the c onnections among individual leaders operating in undergraduate STEM education reform faculty networks. Although many different networks exist in STEM reform, research and literature review point to a collection of informal relationships which share network information across field of networks. This exploratory investigation seeks to establish the existence of these social relationships and their purposes in serving undergraduate STEM education reform networks goals, needs, and functions. Participant Commi tment & Considerations Participation in this study is divided into two data collection phases. Total time estimated for participation in this study is about one hour. The details of each data collection phase are listed below. As part of this study, parti cipants will be identified to one another. The names of each willing participant will be populated into a survey that allows others to indicate their level of acquaintance with one another. Although names of participants will be shared, individual responses in the survey will not be shared. These p ractices are in accordance with common social network research and with prior approvals of the Michigan State University IRB. Data Collection 1: Social Network Survey (anticipated participant completion time: 10 minutes) After all participants in the s tudy are established, the research will begin by mapping the social connections among members engaged in undergraduate STEM education reform. You will be sent a survey through Qualtrics which asks you to provide your name, basic demographic information, an d your STEM network affiliations. In addition to those prompts, you will also be asked to indicate your level of social interaction with other participants within the study. The purpose of the social network survey is to establish relationships among netwo rk leaders and conceptually interpret how these relationships function. Data Collection 2: Semi - Structured Interviews (anticipated participant completion time: 45 minutes) Following the social network survey, participants may be contacted to set up an interview to discuss their responses. These interviews will focus on: 1) Relationships indicated in the social network survey 2) Context and nature of the relationships estab lished in the survey 3) Perceptions of whole network responses through visuals 4) Additional thoughts from participants on the study 195 Additional Questions? Thank you again for your generosity in participation. I look forward to learning together about t he role of networks in advancing improvement in undergraduate STEM education. If you have any additional questions, please do not hesitate to reach out with any questions or concerns as they arise. Contact information is listed in the body of the email. 196 Appendix C: Sample Interview Protocol Interviewee: Date: Interviewer: Social Network Analysis Data Current Institution: Total time in STEM Reform: Current Network Involvement: Centralization Centrality Scores QAP Probe Density : .75/.69/.24 Degree (in) : 14 / 14 / 6 (11) Advice Homophily: STEM Tenure Reciprocity : 1 / 1 / .29 Between: 2 / 5.5 / 5 Business Heterophily: Inst. Type : 12 / 11.17 / 6.52 Closeness : 18/ 18 / 38 Problem Homophily: Same Network Introduction Hello and thank you again for your participation in my study on informal networks in STEM education reform! As my previous correspondence indicates, I am investigating the interconnectedness of people working across multiple STEM initiatives. Do you have any questions about the study so far? You were sent the consent form with your invitation to participate, at the beginning of the survey, and I want to remind you of it again today. If you consent, I would like to continue our conversation with a recording of our discussion. Do you consent? I have planned for this interview to last no longer than one hour. During this time, I have several questions regarding your work in STEM education reform, and those with whom you regularly work. I also have some questions drawn directly from the survey you took last month. 197 We have quite a bit to cover! If our time starts to run a bit short, I may interrupt you in order to make sure we complete the protocol. Foundations In reviewing some of the names of peop le in the study, please tell me a bit more about the people you know. Where did you meet? How did you meet? How often do you interact? Which STEM networks do you believe to be on the forefront of undergraduate education reform? Information Sharing How often do you talk about STEM networks? What about the networks do you discuss? How (or in what location) do these discussions take place? Is there anyone on the list with whom you regularly seek advice? I f so, can you provide an example? Netw ork Learning How have you collaborated with others in this list to improve your STEM networks? Which STEM networks Network Impact How have your discussions with others in this study impacted your network? How have your discussions with others in this study aided your network through difficult transitions? Conclusion Thank you! I anticipate having a fully - transcribed copy of this interview in about a week. Would you like a copy of our conversation? In the interest of interviewer fidelity, I may reach out to you in a few weeks to clarify or confirm some of the aspects of our conversation. Would that be alright? What is the best method for contacting you? Thank you again and have a great day! 198 Appendix D: Variable Codebook Table 14. Appendix D: Social Network Analysis Variable Codebook Variable Label Variable Description Age Range Participant age range (binned by decade) Business Network Individual participant responses in the "Spoke about Network Business" prompt. Ethnicity Participant preferred ethnicity Gave Advice Network Individual inversed participant responses in the "Sought Advice" prompt. Gender Participants preferred gender identity Position Categorical variable denoting institutional position Primary Organization Participant employer Problems Network Individual participant responses in the "Discussed Network Problems" prompt. Same USER Network Binary variable denoting if participant shares a USER Network Membership with another. STEM Tenure Continuous variable for years spent working in USER. 199 REFERENCES 200 REFERENCES Accelerating Systemic Change Network. (2018). About ASCN: Vision. Retrieved from https://ascnhighered.org/ASCN/about.html Ahuja, G., Soda, G., & Zaheer, A. (2012). The genesis and dynamics of organizational networks. Organization Science, 23 (2), 434 448. Aldrich, H. & Herker, D. (1977). Boundary spanning roles and organization structure. The Academy of Management Review, 2 , 217 230. Amey, M. J., Eddy, P. L., & Campbell, T. G. (2010). Crossing boundaries creating community college partnerships to promote ed ucational transitions. Community College Review, 37 (4), 333 - 347. Arganoff, R. (2007). Managing within networks: Adding value to public organizations . Washington, D.C: Georgetown University Press. Armstrong, D. (2017). Wicked problems in special and inclusi ve education. Journal of Research in Special Education Needs. 17 (4) 229 - 236. Association of American Universities (2015a). Undergraduate STEM education initiative: About. Retrieved from https://stemedhub.org/groups/aau/about Association of American Univers ities (2015b). AAU STEM network. Retrieved from https://stemedhub.org/groups/aau/stem_network Association of American Universities (2018). Scaling improvement in STEM learning environments: The strategies role of a national organization. Retrieved from htt ps://www.aau.edu/strategic - role - national - organization Association of American Universities. (2014). Framework for systemic change in undergraduate STEM teaching and learning. Retrieved from https://stemedhub.org/groups/aau/File:AAU_Framework_030114.pdfAAU 2018 Astuto, T. A., Clark, D. L., Read, A. M., McGree, K., & Fernandez, L. (1993). Challenges to dominant assumptions controlling educational reform. Andover, MA: Regional Laboratory for the Educational Improvement of the Northeast and Islands. Austin, A. E. (2011). Promoting evidence - based change in undergraduate science education: A paper commissioned by the National Academies National Research Council Board on Science Education. Washington, D.C.: National Academies. Avella, J. R. (2016). Delphi panels: R esearch design, procedures, advantages, and challenges. International Journal of Doctoral Studies, 11 (1), 305 321. 201 Bakker, R.M., Raab, J., & Milward, H.B. (2012). A preliminary theory of dark network resilience. Journal of Policy Analysis and Management, 3 1(1), 33 62. Beach, A. L., Henderson, C., & Finkelstein, N. (2012). Facilitating change in undergraduate STEM education. Change: The Magazine of Higher Learning, 44 (6), 52 59. Berger, P. L., & Luckman, T. (1967). The social construction of reality: A treat ise in the sociology of knowledge . Garden City, NY: Random House. Berry, F.S., Brower, R.S., Choi, S.O., Goa, W.X., Jang, H., Kwon, M.J., & Word, J. (2004). Three traditions of network research: What the public management research agenda can learn from oth er research communities. Public Administration Review, 64 (1), 539 52. Block, P. (1993). Stewardship: Choosing service over self - interest . San Francisco: Berrett - Koehler. Bond, M. A., & Lockee, B. B. (2014 ). Building virtual communities of practice for dist ance educators . New York, NY: Springer. Borgatti, S. P., & Molina, J. L. (2003). Ethical and strategic issues in organizational social network analysis. The Journal of Applied Behavioral Science, 39 (3), 337 349. Borgatti, S. P., Everett, M. G., & Johnson, J. C. (2018). Analyzing social networks . Los Angeles: Sage. Bourdieu, P. (1977). Outline of a Theory of Practice (Cambridge Studies in Social and Cultural Anthropology) (R. Nice, Trans.). Cambridge: Cambridge University Press. Bourdieu, P. (1987). What mak es a social class? On the theoretical and practical existence of groups. Berkeley Journal of Sociology, 32 , 1 17. Bourdieu, P. (1990). The logic of practice. Stanford university press. Boyatzis, R. E. (1998). Transforming qualitative information: Thematic analysis and code development . Los Angeles: Sage. Brands, R. A. (2013). Cognitive social structures in social network research: A review. Journal of Organizational Behavior, 34 (S1), S82 S103. Braun, V., & Clarke, V. (2006). Using thematic analysis in psych ology. Qualitative Research in Psychology, 3 (2), 77 101. Braun, V., & Clarke, V. (2012). Thematic analysis. In H. Cooper (ed.), APA Handbook of Research Methods in Psychology: Vol. 2. Research Designs (pp. 57 71). Washington, D.C.: American Psychological A ssociation. 202 Brion, S., Chauvet, V., Chollet, B., & Mothe, C. (2012). Project leaders as boundary spanners: Relational antecedents and performance outcomes. International Journal of Project Management, 30 , 708 722. Brown, J. S., & Duguid, P. (1991). Organizational learning and communities of practice: Toward a unified view of working, learning, and innovation. Organizational Science, 2 (1), 40 57. Bryson, J. M., Crosby, B. C., & Stone, M. M. (2006). The design and imp lementation of cross - sector collaborations: Propositions from the literature. Public Administration Review, 66 (SUPPL. 1), 44 55. Burt, R. S. (2000). The network structure of social capital. Research in organizational behavior, 22 , 345 423. Burt, R. S., & Celotto, N. (1992). The network structure of management roles in a large matrix firm. Evaluation and Program Planning, 15 (3), 303 326. Burt, R.S. (1992). Structural holes . Cambridge, MA: Harvard University Press. Carboni, J. L., & Milward, H. B. (2012). Go vernance, privatization, and systemic risk in the disarticulated state. Public Administration Review, 72 (S1), 36 S44 Casebeer, A. L., Reay, T., Dewald, J., & Pablo, A. (2010). Knowing through doing: Unleashing latent dynamic capabilities in a public sector context. In Walshe, K., Harvey, G., & Jas, P. (Eds.). Connecting knowledge and performance in public services (251 275). Cambridge: Cambridge University Press. Casebeer, A., Huerta, T., & VanderPlaat, M. (2006). Re - navigating the research and practice ter rain of health services networks. Healthcare Papers, 7 (2), 68 72. Casebeer, A., Popp, J., & Scott, C. (2009). Positively deviant networks: What are they and why do we need them? Journal of Health Organization and Management, 23 (6), 610 626 Chisholm, R.F. ( 1996). On the meaning of networks. G roup and Organization Management 21 (2), 216 236. Coalition for Reform of Undergraduate STEM Education (2014). Achieving systemic change: A sourcebook for advancing and funding undergraduate STEM education. Washington, D. C.: Association of American Colleges and Universities. Coleman, J. (1988). Social capital in the creation of human capital. The American Journal of Sociology, 94, S95 S120 Coleman, J. (1990). Foundations of social theory . Cambridge, MA: Harvard University Press. Conteh, C. (2013). Strategic Inter - Organizational Cooperation in Complex Environments. Public Management Review, 15 (4), 501 521. 203 reform and change. R etrieved from http://accountablecaredoctors.org/health - care - reform/a - wicked - problem - healthcare - system - reform - and - change/ Creswell, J. W. (2009). Research Design: Qualitative, Quantitative, and Mixed Methods Approaches (3rd ed.). Thousand Oaks, CA: Sage Pub lication Creswell, J. W., & Plano, C. V. L. (2011). Designing and conducting mixed methods research. Los Angeles: SAGE Publications Creswell, J. W., Plano Clark, V. L., Gutmann, M. L. & Hanson, W. E. (2003). Advanced mixed methods research designs. In A. T ashakkori and C. Teddlie (Eds), Handbook on mixed methods in the behavioral and social sciences (pp. 209 - 240). Thousand Oaks, CA: Sage Publications Dancy, M. H., & Henderson, C. (2008). Barriers and promises in STEM reform. A paper commissioned by the Nati onal Academies National Research Council. Washington, D.C.: National Research Council. Davies, B.D. & Bansel P. (2007). Neoliberalism and education. International Journal of Qualitative Studies in Education, 20( 3), 247 - 259 Dawes, S.S. (1996). Interagency information sharing: Expected benefits, manageable risks. Journal of Policy Analysis and Management 66 , 377 394 DuFour, R., DuFour, R., & Eaker, R. (2008). Revisiting professional learning community at work: New insights for improving schools . Bloomington, IN: Solution - Tree. Eisenhardt, K. M., & Martin, J. A. (2000). Dynamic capabilities: What are they? Strategic Management Journal, 21 (10), 1105 1121 Elrod, S., & Kezar, A. (2015). Increasing student success in stem: A guide to systemic institutional change a workbook for campus teams. A Keck/PKAL Project at the Association of American Colleges & Universities (AAC&U). Retrieved from https://www.aacu.org/pkal/educationframework \ Fairweather, J. (2009). Linking evidence and prom ising practices in science, technology, engineering, and mathematics (STEM) undergraduate education. Washington, D.C.: Board of Science Education, National Research Council, The National Academies Feiock, R.C., Lee, I.W., & Park, H.J. (2012). Administrator networks: Selecting partners to reduce risk in economic development. Public Administration Review, 72 (S1), S58 S68. Frank, K.A., Kim, C.M., & Belman, D. (2010) Utility theory, social networks, and teacher decision ma Daly (Ed.) Social Network Theory and Educational Change . Cambridge, MA: Harvard Education Press. 204 Friedman, T.L. (2007). The world is flat: A brief history of the twenty - first century . New York: Picador. Fukayama, F. (1997). The Tanner lectures on human values. Delivered May, University of Oxford Fullan, M. (2006). Turnaround leadership. San Francisco, CA: Jossey - Bass. Gappa, J. M., Austin, A. E., & Trice, A. G. (20 07). strategic imperative. San Francisco, CA: Jossey - Bass. Gehrke, S. & Kezar, A. (2016). STEM reform outcomes through communities of transformation. The Magazine of Higher Learning, 48 (1), 30 38. Gherardi, S., N icolini, D., & Odella, F. (1998). Toward a social understanding of how people learn in organizations: The notion of situated curriculum. Management Learning, 29 (3), 273 297 H ealthcare Papers, 7 (2): 28 31 Granovetter, M. (1973). The strength of weak ties . American Journal of Sociology, 78 (6), 1360 1380. Green, J. C., Caracelli, V. J., and Graham, W. F. (1989). Toward a conceptual framework for mixed - method evaluation designs. E ducational Evaluation and Policy Analysis, 11 (3), 255 274. Hartley, J., & Bennington, J. (2006). Copy and paste, or graft and transplant? Knowledge sharing through inter - organizational networks. Public Money & Management, 26 (2), 101. Healy, F. H., & DeStef ano, J. (1997). Education reform support: A framework for scaling up school reform. Policy Paper Series (Research Triangle Park, NC: Research Triangle Institute. Hejnova, J. (2010). Beyond dark and bright: Towards a more holistic understanding of inter - gro up networks. Public Administration, 88 (3), 741 763. Henderson, C., Beach, A., & Finkelstein, N. (2011). Facilitating change in undergraduate STEM instructional practices: An analytic review of the literature. Journal of Research in Science Teaching, 48 (8), 952 984 Henisez, W.J., Holburn, G.L.F., & Zepler, B.A. (2005). Deinstitutionalization and institutional replacement: State - centered and neo - liberal models in the global electricity supply industry. World Bank Policy Research Working Paper No. 3690. Retrie ved from https://papers.ssrn.com/sol3/papers.cfm?abstract_id=801446 205 Hill, L. (2016) Advancing Improvement in STEM undergraduate reform through multi - institutional networks: The role of formal boundary spanners. (Dissertation). Michigan State University. Ea st Lansing, MI. United States of America. performance funding in higher education. Educational Evaluation and Policy Analysis, 37 (4), 501 519. Hillman, N. W., Tandberg, D. A., & Gross, J. P. K. (2014). Performance funding in higher education: Do financial incentives impact college completions? The Journal of Higher Education, 85 (6), 826 857. https://doi.org/10.1353/jhe.2014.0031 Hirschman, D., & Berrey, E. (2017). The partial deinstitutionalization of affirmative action in US higher education, 1988 to 2014. Sociological Science , 4 , 449 - 468. Hogan, B., Carrasco, J.A., & Wellman, B. (2007). Visualizing personal networks: Working with participant - aid ed sociograms. Field Methods, 19 (2), 116 144. Holley, J. (2012). Network weaver handbook. A guide to transformational networks . Athens, Ontario: Network Weaver Publishing. Hutchings, M., & Quinney, A. (2015). The Flipped Classroom, Disruptive Pedagogies, Enabling Technologies and Wicked Problems: Responding to" The Bomb in the Basement". Electronic Journal of e - learning , 13 (2), 106 - 119. Network. Public Admin istration Review, 74 (5), 587 598. Huffman, J. B., & Hipp, K. A. (2001). Creating communities of learners: The interaction of shared leadership, shared vision, and supportive conditions. International Journal of Educational Reform, 10 (3), 272 - 281. Hulme, M. (2009). Why we disagree about climate change: Understanding controversy, inaction and opportunity. Cambridge University Press. Human, S.E., & Provan, K.G. (2000). Legitimacy building in the evolution of small - firm multilateral networks: A comparative stud y of success and demise. Administrative Science Quarterly, 45 (2), 327 365. Huxham, C., & Vangen, S. (2005). Managing to collaborate: The theory and practice of collaborative advantage. New York: Routledge Imperial, M.T. (2005). Using collaboration as a gov ernance strategy. Administration and Society, 37 , 281 320. Isett, K. R., Mergel, I.A., LeRoux, K., Mischen, P. A., & Rethemeyer, R. K. (2011). Networks in public administration scholarship: Understanding where we are and where we need to go. Journal of Public Administration Research and Theory, 21 (SUPPL . 1), i157 i173 206 educational leadership in higher education: A mixed methods study. Research in Higher Education, 48 (1) 93 135 James, W. (2000) What pragmatism means, In J. J. Stuhr (Ed.) Pragmatism and the Classical American Philosophy: Essential Readings and Interpretive Essays (2nd ed), pp. 193 202 New York: Oxford University Press J.C. Bezdek, B. Spillman, R. Spillman. (1978). A fuzzy relation space for group decision theory, Fuzzy Sets and Systems. 1 (4) 255 268. Joffe, H. (2012). Thematic analysis. In D. Harper and A.R. Thompson (eds), Qualitative research methods in mental health and psychotherapy: A guide for students and practitioners (pp. 209 23). San Francisco, CA: John Wiley & Sons Johnson, B., & Turner, L. A. (2003). Data collection strategies in mixed methods research. In: Tashakkori, A., and Teddlie, C. (eds.), Handbook on mixed methods in the behavioral and social science s. Thousand Oaks, CA: Sage Publications. Jones, J.M. (2018, October). Confidence in higher education down since 2015. Gallup. Retrieved from https://news.gallup.com/opinion/gallup/242441/confidence - higher - education - down - 2015.aspx Katz, R., & Tushman, M. (1981). An investigation into the managerial roles and career paths of gatekeepers and project supervisors in a major R & D facility. R & D Management, 11 (3), 103 110. Kaye, G. & Wolff, T. (2002). From the ground up : a workbook on coalition building and community development . Amherst , MA: AHEC/Community Partners. Keast, R., Mandell, M.P., Brown, K., & Woolcock, G. (2004). Network structures: Working differently and changing expectations. Public Administration Review, 64 (3), 363 371. Kezar, A. (2011a). What is the best way to achieve b roader reach of improved practices in higher education? Innovation in Higher Education, 36 , 235 247. Kezar, A. (2014). Higher education change and social networks: A review of research. The Journal of Higher Education, 85 (1), 91 125 Kezar, A. (2014). Highe r Education Change and Social Networks: A Review of Research. The Journal of Higher Education, 85 (1), 91 125. https://doi.org/10.1353/jhe.2014.0003 Kezar, A. J., Chambers, T. C., & Burkhardt, J. C. (2005). Higher education for the public good: Emerging voi ces from a national movement . San Francisco: Jossey - Bass. Kezar, A., & Gehrke, S. (2015). Communities of transformation and their work scaling STEM reform. Los Angeles, CA. Retrieved from https://pullias.usc.edu/wp - content/uploads/2016/01/communities - of - tr ans.pdf 207 Kezar, A., Gehrke, S., & Bernstein - Sierra, S. (2018). Communities of transformation: creating changes to deeply entrenched issues. The Journal of Higher Education, 89 (6), 832 864. Kezar, A., & Maxey, D. (2012). Missing from the institutional data p faculty. New Directions for Institutional Research , 2012 (155), 47 - 65. Khan, A. M., Arafat, M. Y., Raushan, M. A., Saleem, I., & Dviwedi, A. K. (2019, April). Does Intellectual Capital affect the Venture Creation?. International Con ference on Sustainable Management. Knaub, A.V., Henderson, C. & Quardokus - Fisher, K. (2019). Finding the leaders: an examination of social network analysis and leadership identification in STEM education change. International Journal of STEM Education, 5 (26), 1 14 Knight, L. (2002). Network learning: Exploring by interorganizational networks. Human Relations, 55 (4), 427. Knight, L., & Pye, A. (2005). Network learning: An empirically - derived model of learning by groups of organizations. Human Relations, 58 (3), 369 392. Krause, K. L. (2012). Addressing the wicked problem of quality in higher education: Theoretical approaches and implications. Higher Education Research & Development , 31 (3), 285 297. Kuhn, R.L. (1985). Frontiers in creative and innovative man agement. Cambridge, MA: Ballinger. Leifer, R., & Delbecq, A. (1978). Organizational/environmental interchange: A model of boundary spanning activity. Academy of Management Review, 3 (1), 40 50. Louis, K. S., Kruse, S. D., & Bryk, A. S. (1995). Professionali sm and community: What is it and why is it important in urban schools. Professionalism and community: Perspectives on reforming urban schools , 3 22. Maclean, R., & Pavlova, M. (2013). Vocationalization of secondary and higher education: pathways to the world of work. Revisiting global trends in TVET: Reflections on theory and practice , 40 . Marginson, S. (2011). Higher education and public good. Higher Education Quarterly, 65 (4), 411 433. Masten, M. (2016, September). Professional development: A wicked problem in education. Retrieved from http://inservice.ascd.org/professional - development - a - wicked - problem - in - education/ Mays, G.P., & Scutchfield, F.D. (2010). Improving public health system performance through multiorganizational partnerships. Preve nting Chronic Disease, 7 (6), A116. 208 McCarty, C., Molina, J. L., Aguilar, C., & Rota, L. (2007). A comparison of social network mapping and personal network visualization. Field Methods, 19 (2), 145 162. McGuire, M. (2006). Collaborative public management: As sessing what we know and how we know it. Public Administration Review, 66 (1), 33 43. McLaughlin, M. W., & Talbert, J. E. (2001 ). Professional communities and the work of high school teaching. University of Chicago Press. McLaughlin, M., & Mitra, D. (2002). Theory based change and change based theory: Going deeper, going broader. Journal of Educational Change, 2 , 301 323 Mezirow, J. (1991). Transformative dimensions of adult learning . San Francisco, CA: Jossey - Bass Milward, H. B., Provan, K.G., Fish, A., Ise tt, K. R., & Huang, K. (2010). Governance and collaboration: An evolutionary study of two mental health networks. Journal of Public Administration Research and Theory, 20 (SUPPL . 1), i125 i141 Milward, H.B., Kenis, P., & Raab, J. (2006). Introduction: Towa rds the study of network control. International Public Management Journal, 9 (3), 203 208. Montgomery, J.R. (1996). Goal - based learning: Accelerating performance change . St. Charles, Ill: Anderson Worldwide. Morgan, D. L. (1998). Practical strategies for co mbining qualitative and quantitative methods: Applications to health research. Qualitative Health Research, 8 (3), 362 376. Morgan, D. L. (2006). Paradigms lost and pragmatism regained: Methodological implications of combining qualitative and quantitative m ethods. Journal of Mixed Methods Research, 1 , 48 76. Morris, JF, & Deckro, RF. (2013). SNA data difficulties with dark networks. Behavioral Science with Terrorism Political Aggression, 5 (2), 70 93 Mulford, B., & Silins, H. (2003). Leadership for organizati onal learning and improved student outcomes What do we know? Cambridge Journal of Education, 33 (2), 175 195. Nahapiet, J., & Ghoshal, S. (1998). Social capital, intellectual capital, and the organizational advantage. Academy of Management Review, 23 (2), 24 2 266. National Academies of Sciences Engineering Medicine. (2018) Roundtable on systemic change in undergraduate STEM education. Retrieved from https://sites.nationalacademies.org/DBASSE/BOSE/systemic_change_in_Undergrad_ST EM_Education/index.htm 209 Nelson, J . C., Raskind - Hood, C., Galvin, V. G., Essien, J.D., & Levine, L.M. (1999). Positioning for partnerships: Assessing public health agency readiness. American Journal of Preventive Medicine, 16 , 103 117. Networks Leadership Symposium (2013). If networks are the answer, what is the question: Rethinking networks. Symposium Proceedings, Victoria, BC. (unpublished report) of public management in networks. Public Admin istration Review, 64 (6), 681 Olssen, M. & Peters, M.A. (2005). Neoliberalism, higher education and the knowledge economy: from the free market to knowledge capitalism, Journal of Education Policy 20 (3), 313 345. Osborne, D., & Gaebler. (1992). Reinventing government: How the entrepreneurial spirit is transforming the public sector. Reading, MA: Plume. Ostrom, E. (2009). What is social capital? Social capital: Reaching out, reaching in Peirce, C. S. (1877). The fixation of belief. Retrieved from http://www. peirce.org/writings/p107.html Phelps, C., Heidl, R., & Wadhwa, A. (2012). Knowledge, networks, and knowledge networks: A review and research agenda. Journal of management, 38 (4), 1115 1166. Popp, J.K., Milward, H.B., MacKean, G., Casebeer, A., & Lindstrom, R. (2014). Inter - organizational networks: A review of the literature to inform practice. IBM Center for the Business of Government: Washington, D.C. Powell, T. C. (2001) Competitive advantage: logical and philosophical considerations. Strategic Management Journal, 22 (9), 875 88. Provan, K. G., & Lemaire, R. H. (2012). Core concepts and key ideas for understanding public sector organizational networks: Using research to inform scholarship and practice. Public Administration Review, 72 (5), 638 648. Provan, K., & Huang, K. (2012). Resources tangibility and the evolution of a publicly funded health and human services network. Public Administration Review, 72 (3), 366 375. Provan, K., & Kenis, P. (2008). Modes of network governance: Structure, managemen t, and effectiveness. Journal of Public Administration Research and Theory, 18 (2), 229 Provan, K., Fish, A., & Sydow, J. (2007). Interorganizational networks at the network level: A review of the empirical literature on whole networks . Journal of Managemen t, 33 (3), 479 Putnam, R. (1993). The prosperous community: social capital and public life. The American Prospect, 13( 4) 241 262. 210 Culture and Politics (pp. 223 - 234). New York: Palgr ave Macmillan. Ramaley, J. A. (2014). The changing role of higher education: Learning to deal with wicked problems. Journal of Higher Education Outreach and Engagement , 18 (3), 7 22. Ramon, S. (2008). Neoliberalism and its implications for mental health in the UK. International Journal of Law and Psychiatry, 31 (2), 116 125. Reay, T., Goodrick, E., Casebeer, A., & Hinings, C.R. (2012). Legitimizing new practices in primary health care. Health Care Management Review, 38 (1), 9 - 19. Rittel, H.W.J., & Webber, M. (1973). Dilemmas in general theory of planning. Policy Sciences, 4, 155 - 169 Rodríguez, C., Langley, A., Béland, F., & Denis, J. (2007). Governance, power, and formal collaboration in an inter - organizational network. Administration & Society, 39 (2), 150 Rogers, E. (2003). Diffusion of Innovations . New York, NY: Simon and Schuster Rossman, G. B., & Wilson, B. L. (1985). Numbers and words: Combining quantitative and qualitative methods in a single large - scale evaluation study. Evaluation Review, 9 , 627 - 643 Rowlands, J. (2013). Academic boards: Less intellectual and more academic capital in higher education governance? Studies in Higher Education, 38 (9), 1274 1289. Samoff, J., Sebatane, E. & Dembele, M. (2003). Scaling - up by focusing down: Creating space to expand education reform. Paper presented at the Biennial meeting of the Association for the development of education Africa in Arusha, Tanzania, October 7 11 Scott, W. R., & Davis, G. F. (2007). Organizations and organizing: Rational, natural and open systems perspectives. Upper Saddle River, NJ: Pearson - Prentice Hall. Singh - Kagisano. (2001). Higher education for the public good: Views from the South . Trentham Books: London, United Kingdom . Stewart, T. E. (1998): Intellectual Capital: The New Wealth of Organizations . New York, NY: Bantham Double Day Publishing Group. Stoll, L., Bolam, R., McMahon, A., Wallace, M., & Thomas, S. (2006). Professional learning communities: A review of the liter ature. Journal of Educational Change, 7 (4), 221 258. Stoyko, P. (2001). The learning journey: A guide to achieving excellence. Ottawa, ONT: Canadian Center for Management Development Roundtable on Learning Organization. Tashakkori, A., & Teddlie, C. (2003) . Handbook on mixed methods in the behavioral and social sciences . Thousand Oaks, CA: Sage Publications 211 Thelin, J.R. (2011). A History of American Higher Education . Baltimore, MD: The Johns Hopkins University Press Thorgren, S., Wincent, J., Ortqvist, D. ( 2009). Designing interorganizational networks for innovation: An empirical examination of network configuration, formation, and governance. Journal of Engineering and Technology Management, 26 (3), 148 166. Tilak, B.G. (2008) Transition from higher educatio n as a public good to higher education as a private good: the saga of Indian experience. Journal of Asian Public Policy, 1 (2), 220 234. competition, and intra - organizational knowledge sharing. Organization Science, 13 (2), 179 190 Turrini, A., Christofoli, D., Frosini, F., & Nasi, G. (2010). Ne twork literature about determinants of network effectiveness. Public Administration, 88 (2), 528 550. Valente, T. W., & Davis, R. L. (1999). Accelerating the diffusion of innovations using opinion leaders. The Annals of the American Academy of Political and Social Science, 566 (1), 55 - 67. Valente, TW, & Pumpuang, P. (2007). Identifying opinion leaders to promote behavior change. Health Education Behavior, 34 (6), 881 896 Vescio, V., Ross, D., & Adams, A. (2007). A review of research on the impact of profession al learning communities on teaching practice and student learning. Teaching and Teacher Education, 24 , 80 91. Warren, C., & Foxx, V. (2011). Accountability, Bureaucratic Bloat, and Federal Funding of Higher Education. Academe , 97 (4), 52 Wasserman, S., & Faust, K. (1994). Social network analysis: Methods and applications . Cambridge, England: Cambridge Publishing. Higher education quarterly , 54 (1), 5 - 21. Weber, E. P., & Khademian, A. M. (2008). Wicked problems, knowledge challenges, and collaborative capacity builders in network settings. Public Administration Review, 68 (2), 334 349 Weick, K.E. (1969). The Social Psychology of Organizing. New York, NY: Mc Graw Hill. Williamson, T. W., Hughes, E. S., & Head, P. L. (2018). An Exploration of Administrative Bloat in American Higher Education. Planning for Higher Education , 46 (2), 15 - 22. 212 Wheatly, M. & Frieze, D. (2011). Walk out, walk on: A learning journey into communities daring to live the future now. San Francisco, CA: Berrett - Koehler Publishing. Wolff, T., & Berkowitz Bill. (2000). The Spirit of the Coalition . American Public Health Association. Williams, G. (2011). Some thoughts on higher education, Browne and coalition policy. London Review of Higher Education, 9 (1) 1 4. e European Association for Evolutionary Political Economy, Maastricht, 7 10 November Zahra, S. A., & George, G. (2002). Absorptive capacity: A review, reconceptualisation and extension. Academy of Management Review, 27 (2), 185 203. Zywicki, T. J., & Koopma n, C. (2017). The changing of the guard: The political economy of administrative bloat in American higher education. George Mason Law & Economics Research Paper , (17 - 12).