Copyright holder: Tyndale University, 3377 Bayview Ave., Toronto, Ontario, Canada M2M 3S4 Att.: Library Director, J. William Horsey Library Copyright: This Work has been made available by the authority of the copyright owner solely for the purpose of private study and research and may not be copied or reproduced except as permitted by the copyright laws of Canada without the written authority from the copyright owner. Copyright license: Attribution-NonCommercial-NoDerivatives 4.0 International License Citation: Chapman, Mark Denis. No Longer Crying in the Wilderness: Canadian Evangelical Organizations and their Networks. Ph.D., Centre for the Study of Religion, University of Toronto, 2004. ***** Begin Content ****** TYNDALE UNIVERSITY 3377 Bayview Avenue Toronto, ON M2M 3S4 TEL: 416.226.6620 www.tyndale.ca Note: This Work has been made available by the authority of the copyright owner solely for the purpose of private study and research and may not be copied or reproduced except as permitted by the copyright laws of Canada without the written authority from the copyright owner. Chapman, Mark Denis. No Longer Crying in the Wilderness: Canadian Evangelical Organizations and their Networks. Ph.D., Centre for the Study of Religion, University of Toronto, 2004. [ Citation Page A ] NO LONGER CRYING IN THE WILDERNESS: CANADIAN EVANGELICAL ORGANIZATIONS AND THEIR NETWORKS by Mark Denis Chapman A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Centre for the Study of Religion University of Toronto © Copyright by Mark Denis Chapman (2004) [ Page i ] ABSTRACT No Longer Crying in the Wilderness: Canadian Evangelical Organizations and their Networks Mark Denis Chapman Doctor of Philosophy, 2004 Centre for the Study of Religion University of Toronto This dissertation establishes that Canadian evangelicals are no longer, if they ever were, completely isolated from each other and from the communities in which they operate. It explores how networks among evangelical churches influence the functioning of those same churches and their relationship with other organizations in the local community. It does this by examining the nature and structure of the networks evangelical churches are involved in, in one geographical location. It establishes who these evangelicals are, and how and why they interact with each other, and with other organizations in the local community. I conclude that there is significant diversity within the local Protestant religious community; that interconnections among churches can relate to organizational similarity but just as often, cross-denominational, theological, and size boundaries; that reasons for interaction are often functional, rather than based on characteristic similarity; and finally, that direct connections are relatively sparse but indirect connections are plentiful. These conclusions establish that different types of Protestant churches do not line up in two distinct camps opposite in focus in all situations. [ Page ] ii ACKNOWLEDGEMENTS Some parts of chapter 2 have been previously published as Chapman, M. D. 1999. Identifying Evangelical Organizations: a New Look At an Old Problem. Studies in Religion - Sciences Religieuses 28: 307-321. Some of the research for this dissertation was funded by the Religious Research Association’s Constant H. Jacquet Research Award. Many thanks to my advisors who were patient when the whole project took much longer than it should have. I want to thank John Simpson, my supervisor, for guiding me through the whole process, providing many invaluable insights into the sociology of religion and academia as a whole, and for helping me find work to pay for the whole thing. Without his consistent support and direction my dissertation would have been a less ambitious, less focused, and less serviceable document. Barry Wellman has provided me with good humour, practical professional advice, and excellent insight into style and editing. Without his input my thesis would be even less readable than it already is. Phyllis Airhart has been a great help making my arguments and writing intelligible to academics outside of my own little world. Her comments have made me a better writer. Jordan Moar has been an inexhaustible editor and faithful friend who never seemed to tire of listening to and critiquing my ideas. The various directors, graduate officers, and administrators at the Centre for the Study of Religion at the University of Toronto have given me constant support and encouragement. They were helpful in making me feel competent even in my most doubtful times. However, most of all I want to thank my immediate and extended family. In particular, my wife Mary and my three children. They have been very patient. They have lived with less income than they might have had in other circumstances but have rarely complained, have maintained a relentless confidence in my abilities, and have kept me human through these many years. I would not have wanted to do this without them. Finally, while I am indebted to many people for their support and input into this project the errors it contains are mine alone. [ Page ] iii TABLE OF CONTENTS 1 INTRODUCTION .... 1 1.1 Why Canadian Evangelicalism? .... 2 1.2 Why Network Analysis? .... 4 1.3 Connectivity: Preliminary Findings .... 5 1.4 Data Collection and Analysis .... 6 1.5 Chapter Summary .... 7 2 DEFINITION: WHAT IS AN EVANGELICAL? .... 8 2.1 Identifying Evangelical Organizations: A New Look at an Old Problem .... 8 2.1.1 Definitional Proliferation .... 9 2.1.2 Towards a Polythetic Definition .... 16 2.1.3 Application to Evangelicals .... 18 2.1.4 Summary .... 20 2.2 Definition in Practice .... 20 2.2.1 Initial Design .... 21 2.2.2 Scoring the Indices .... 22 2.2.3 Using the Indices to Categorize Evangelical Organizations .... 33 2.2.4 Evaluating the Final Indices .... 35 2.3 Summary .... 39 3 CONNECTION: EVANGELICALS AND THE WORLD .... 41 3.1 Perceptions of Evangelical Interactions .... 42 3.1.1 Interviews with Community Members .... 42 3.1.2 Historical Studies .... 47 3.1.3 Surveys and Sociological Studies of Evangelicals .... 51 3.1.4 What Remains to be Done .... 53 3.2 Networks in North American Protestantism .... 53 3.2.1 American Influence .... 54 3.2.2 Canadian / American Differences .... 57 3.2.3 Canadian Development .... 58 3.2.4 Networks Summary .... 63 3.3 Linking Evangelicals .... 63 3.3.1 Local Networks: Bible Colleges .... 63 3.3.2 National Networks: The Evangelical Fellowship of Canada .... 66 3.3.3 International Networks: Formal and Informal Connections .... 69 3.3.4 How Evangelicals Interact .... 72 3.4 Research Location .... 73 3.4.1 Focus .... 73 3.4.2 Why Kitchener-Waterloo? .... 74 3.4.3 Why Kitchener and Waterloo together? .... 75 3.4.4 Demographic Background .... 76 3.5 Summary .... 78 3.5.1 Considerations .... 78 3.5.2 Evangelical Networks .... 79 4 FOUNDATION: RESEARCH QUESTIONS, AND METHOD .... 81 4.1 Research Questions .... 82 4.1.1 Structural Characteristics: How do Protestant Churches Interact? .... 82 4.1.2 The Benefits and Costs of Network Relations: Why do Protestant Churches Interact? .... 84 4.1.3 Homophily and the Supply of Resources: How can we account for Network Structure? .... 85 4.1.4 Propositions Summary .... 87 4.2 Network Analysis and the Study of Interorganizational Relations .... 88 4.2.1 Why Networks? .... 88 [ Page ] iv 4.2.2 Application to Canadian Evangelicals .... 92 5 COLLECTION: STUDY DESIGN AND DATA COLLECTION .... 95 5.1 Survey Components .... 95 5.1.1 Background Interviews .... 95 5.1.2 Mail Survey .... 96 5.2 Physical Format and Deployment .... 102 5.2.1 Survey .... 102 5.2.2 Cover Letter .... 103 5.2.3 Envelopes and Stamps .... 104 5.2.4 Mail Surveys .... 104 5.3 Data Collection .... 105 5.3.1 Procedures .... 105 5.3.2 Network Data Organization .... 107 6 INVESTIGATION .... 112 6.1 Organization Characteristics and Connectivity .... 113 6.1.1 Network Involvement .... 114 6.1.2 Connections and Divisions .... 125 6.1.3 Connections with Social Services .... 137 6.1.4 Summary: How Do Religious Organizations Interact? .... 143 6.2 The Benefits and Costs of Intrachurch Network Relations .... 144 6.2.1 Costs .... 145 6.2.2 Benefits .... 150 6.2.3 Why do Religious Organizations Interact? .... 160 6.3 Homophily and the Supply of Resources .... 162 6.3.1 Organization Similarity .... 163 6.3.2 Peripheral Connections .... 171 6.3.3 City-Wide Dense Networks .... 176 6.3.4 How Can We Account for Network Structure? .... 180 6.4 Chapter Summary .... 181 7 CONCLUSIONS .... 183 7.1 What have I Shown? .... 183 7.1.1 Who are the Religious Organizations I am Studying and How did I Study Them? .... 183 7.1.2 How Religious Organizations Interact .... 185 7.1.3 Why Religious Organizations Interact .... 186 7.1.4 Describing Network Structure .... 186 7.2 Looking Ahead .... 187 7.2.1 Implications .... 187 7.2.2 Avenues for Future Research .... 189 7.3 Concluding Remarks .... 190 BIBLIOGRAPHY .... 192 APPENDICES .... 206 LIST OF TABLES Table 2.1 - Polythetic and Monothetic Definitions .... 18 Table 2.2 - Necessary Characteristics .... 25 Table 2.3 - Additive Rank/Necessary Characteristics Comparison .... 26 Table 2.4 - Rotated Factor Matrix (All Variables) .... 27 [ Page ] v Table 2.5 - Component Matrix (Necessary Characteristics) .... 29 Table 2.6 - Approximating Grenville’s “Christian Evangelicalism Scale” .... 29 Table 2.7 - Rotated Factor Matrix (Grenville Characteristics) .... 30 Table 2.8 - Remaining Variables .... 30 Table 2.9 - Rotated Factor Matrix (Grenville and Necessary Characteristics) .... 31 Table 2.10 - Rotated Factor Matrix (Remaining Variables) .... 32 Table 2.11 - Score Correlations (Spearman's rho) .... 32 Table 2.12 - Additive Totals (Mean) .... 34 Table 2.13 - Factor Totals (Mean) .... 34 Table 2.14 - Rank/Denomination Comparison .... 37 Table 3.1 - Organizations Mentioned by at Least Four Informant .... 46 Table 5.1 - Type of Network Data Collected .... 99 Table 5.2 - Data Matrices .... 109 Table 6.1 - Denomination Type of Informant Organization .... 117 Table 6.2 - Evangelicalness/Frequency in Network of Informant Organizations .... 118 Table 6.3 - Attendance * DIRECT Degree Centrality Crosstabulation .... 123 Table 6.4 - Attendance * TOTAL Degree Centrality Crosstabulation .... 124 Table 6.5 - Attendance by Denomination .... 125 Table 6.6 - Connection Type ....130 Table 6.7 - Whole Network Density .... 131 Table 6.8 - Church Attendance .... 131 Table 6.9 - Evangelicalness / DIRECT Degree Centrality .... 132 Table 6.10 - Evangelicalness / TOTAL Degree Centrality .... 132 Table 6.11 - Organizations Most Often in the Core .... 136 Table 6.12 - Centrality of Listed Organizations .... 140 Table 6.13 - Contact Type * Evangelicalness Crosstabulation .... 141 Table 6.14 - Most Listed Organizations .... 142 Table 6.15 - Type of Interaction among Directly Connected Informants .... 146 Table 6.16 - Type of Interaction for all Connections .... 147 Table 6.17 - Type of Interaction among all Informants * Organization Type Crosstabulation .... 147 Table 6.18 - Type of Connection in Reciprocated Interaction .... 148 Table 6.19 - Attendance * Tie Type Crosstabulation .... 149 Table 6.20 - Frequency of each Type of Interaction .... 153 Table 6.21 - Importance of Local Social Agencies .... 154 Table 6.22 - Associate with (ranked 1, 2, or 3) .... 154 Table 6.23 - Associate with Non-Christian Religions .... 155 Table 6.24 - Support * Evangelicalness Crosstabulation .... 158 Table 6.25 - Evangelism of Non-Christians (All Informants) .... 159 Table 6.26 - Cliques in the DIRECT matrix .... 166 Table 6.27 - Cliques in the CONNECT matrix .... 167 Table 6.28 - Contact Type * Core/Periphery Crosstabulation .... 175 Table 6.29 - Type of Interaction * Core/Periphery Crosstabulation .... 175 Table 6.30 - Average (Mean) Distance between Reachable Pairs .... 178 LIST OF GRAPHS Graph 2.1 - Additive Score Distribution .... 24 Graph 2.2 - Necessary Score Distribution .... 25 Graph 2.3 - Factor Score Distribution .... 33 [ Page ] vi LIST OF DIAGRAMS Diagram 5.1 - Matrix Dependencies .... 108 Diagram 6.1 - Direct Matrix (Evangelicalness) .... 115 Diagram 6.2 - DIRECT Matrix (Denomination Type) .... 117 Diagram 6.3 - Most and More Evangelical Churches in the DIRECT Matrix .... 119 Diagram 6.4 - CONNECT Matrix (Evangelicalness) .... 119 Diagram 6.5 - More Evangelical Organizations in the CONNECT Matrix .... 120 Diagram 6.6- DIRECT Matrix (Attendance) .... 121 Diagram 6.7 - DIRECT Matrix (Membership) .... 122 Diagram 6.8 - DIRECT Matrix (Staff) .... 122 Diagram 6.9 - TOTAL Matrix (Evangelicalness) .... 127 Diagram 6.10 - DIRECT matrix (Postal Code Similarity) .... 128 Diagram 6.11 - Illustration of 1-Step in the TOTAL matrix .... 129 Diagram 6.12 - DIRECT Matrix (Ethnicity) .... 169 LIST OF APPENDICES Appendix A - Background Interview Questions .... 206 Appendix B - Index of Evangelical Characteristics Pre-test .... 208 Appendix C - Consultation Permission Letter .... 214 Appendix D - First Contact Script .... 215 Appendix E - Survey Pre-Test Cover Letter .... 217 Appendix F - Survey Cover Letter .... 218 Appendix G - Final Survey .... 219 Appendix H - Follow-up .... 226 Appendix i - Thank-you Letter .... 229 Appendix J - Full Factor Analysis Results .... 231 [ Page ] vii 1 INTRODUCTION In this dissertation I contend that, Canadian evangelical churches are no longer, if they ever were, completely isolated from each other or from the communities in which they operate.1 There have been many studies of Canadian evangelicalism and relationships among organizations. Historians of Canadian religion have put significant energy into exploring the nature of Canadian evangelicalism (e.g., Gauvreau 1991; Stackhouse 1993; Burkinshaw 1996; Rawlyk 1997). They have studied the individuals, organizations, and relationships which make Canadian evangelicals different from other sectors of Canadian Christianity and from their American counterparts (Stackhouse 1994; Airhart 1997; Noll 1997; Reimer 1999). At the same time, sociologists have been exploring human activity in terms of network connections rather than individual characteristics (Cook 1977; Knoke and Rogers 1979; Galaskiewicz 1985; Mizruchi and Galaskiewicz 1993). I use insights from these disciplinary streams to examine the characteristics of Canadian evangelical churches and the structure of their interorganizational networks in one geographic location. These studies also provide background for developing a mail survey to collect data on evangelical churches and their networks. Furthermore, they help to interpret interaction among local religious organizations and the relationship between their characteristics and their network structure. In my dissertation, I describe how networks among evangelical churches influence the functioning of those churches and their relationship with other organizations in the local community. I establish who these evangelicals are, how they interact with each other, and how they interact with other organizations in the local community. I also discuss why they interact and suggest some reasons for these network patterns. ___________________________________ 1 Some Canadian evangelicals and fundamentalists isolated themselves from other religious groups and from the culture at large during the mid-twentieth century (e.g., William Aberhart, T. T. Shields).However, even the more extreme separatists maintained some connections to each other (Elliot 1994; Hindmarsh 1997). Furthermore, some were actively involved in their community (e.g., Aberhart’s political involvement) (Stackhouse 1993). Nonetheless, it is clear from the historical record that evangelicals were more open to involvement with non-evangelical organizations in the 19th and late 20th centuries than in the early and mid-20th century (See Rawlyk 1990; Clarke 1996). See 3.1.2 for further discussion of the degree to which evangelicals are as insular and their reputations suggest. [ Page ] 1 [ Page ] 2 My main objectives can be summarized in terms of four basic research questions: Who are the religious organizations I am studying and how are they similar? How do these religious organizations interact? Why do these religious organizations interact? and How can we account for network structure? 2 Answers to these questions provide information about Canadian evangelical networks that expands our knowledge of Canadian evangelicalism. In sum, my dissertation explores, describes, and interprets the structure of evangelical church networks in one community. In this introduction I explain why I think it is worthwhile to study Canadian evangelicalism, why I use network analysis to study it, what sort of connectivity I expected to find among Canadian evangelicals, and hint at some of the surprising aspects of my findings. I also briefly review how I collected the data, and explain the different sections of my thesis. 1.1 Why Canadian Evangelicalism? Events and individuals regularly bring evangelical involvement in North American public life to the attention of the general population.3 However, looking at isolated individuals, events, or organizations does not provide a complete picture of evangelicalism. Thus, the general public often knows little about what makes evangelicals distinctive from other Christian groups or how evangelicals are involved in the community.4 Academics have a much better understanding of evangelicals than the general public and have published extensive studies of evangelical organizations and individuals (Wells and Woodbridge 1977; Marsden et al. 1984; Sweet et al. 1984; Hunter 1987; Marsden 1987; Ellingsen 1988; Bebbington 1989; Dayton and Johnston 1991; Marsden 1991; Noll 1992; Noll et al. 1994; Rawlyk and Noll 1994). Studies of contemporary Canadian evangelicals provide data about connections centred on particular religious individuals or religious organizations and data ___________________________________ 2 That is, how can we explain the shape of the whole network that results from individual nodes connecting with individual nodes. Network structure refers to the pattern of all the connections that all individual nodes have with all other individual nodes. 3 Examples include, their bringing of hundreds of thousands of men to Washington D.C. to pray for their nation and their families, comments by Focus on the Family president James Dobson on the negative impact of cultural trends on the family, the rise of the Pentecostal Stockwell Day to the leadership of a Canadian political party, and TV preacher Jerry Falwell’s description of the political repercussions of America’s moral failure. 4 As an anecdotal example, few of the 16 individuals I interviewed were aware of the role evangelicals played in the local community. [ Page ] 3 about how evangelicals compare with other groups in society across functional, denominational, and geographical boundaries. However, many aspects of evangelical culture and their interaction with their environment remain unstudied or unpublished.5 Thus, although the interactive nature of evangelicalism is not a new observation, these relationships with other organizations have rarely been examined in terms of their influence on evangelical networks within specific geographic locations.6 Existing studies of Canadian evangelicals provide limited detail on how religious organizations in single communities co-operate together or how that co- operation affects particular organizations. Furthermore, none of these studies use the techniques of network analysis to examine evangelical organizations in a single community. Second, most studies of evangelicals originate in the United States and are focused on American evangelicals. Much can be learned about Canadian evangelicals from these studies. However, Canadian evangelical culture differs from American evangelical culture and consequently studies of evangelical Christianity in the United States cannot be applied uncritically to Canadian evangelicals. For example, Van Die (2001:4) notes that the general relationship between religion and public life in Canada is different from the American situation, in part, because of different “constitutional relationships between church and state.” More specifically, Canadian evangelicals are less aggressive in evangelism, are a smaller percentage of the population, and are less receptive to moral majority type initiatives than Americans (Simpson and MacLeod 1985; Posterski and Barker 1993; Stackhouse 1994; Reimer 1995; Corelli 1996). Finally, historical studies of Canadian evangelicals sometimes concentrate on individual organizations or figures as if they were independent from their environment and from each other.7 Alternatively, they use significant organizations or people to represent whole eras or streams of history (e.g., Noll 1992; Stackhouse 1993:268ff). While necessary to provide a manageable description of evangelical history and invaluable for providing the basic details of Canadian evangelicalism on which this study is based, neither tactic explicitly ____________________________________ 5 As late as 1994, Eskridge (1994) noted that there was a shortage of comparative works dealing with evangelicalism. 6 Some examples of community focused studies include Ammerman (1997), Noll (1994), and Yang (1998). None of these studies look at all the interactions among Protestant churches in a single community. 7 Wellman and Berkowitz (1988:2) call these sorts of studies ”aggregative analyses.” [ Page ] 4 focuses on the structure of connections among organizations or the local context in which they operate. My study collects and analyzes these kinds of data. For the purposes of this study, existing demographic surveys and historical studies of Canadian evangelicals serve the role of identifying the community and providing the background necessary to design a network survey of Canadian evangelical organizations in one community. These studies provide a starting point for my collection of data on relations among and between Canadian evangelical organizations. They give an indication of which data to look for and where to look for it. Finally, they help interpret the connections found in the data. 1.2 Why Network Analysis? My first concern with surveys and historical studies is that they rarely pursue the observation of networks among and between Canadian evangelicals in local communities to the extent that they explicitly examine the networks themselves. Existing studies of Canadian evangelicals do not fully describe the extent to which evangelical churches interact with non- evangelical organizations and social service organizations within a single community. However, historical studies do provide data with which to quantify these network connections and with which to answer structural questions relevant to the study of Canadian evangelicals. Yet, there is additional data that can be derived from looking specifically at network relations.8 Questions that can be profitably explored through an analysis of how churches interact with each other (network relations) include: Are Canadian evangelical organizations densely connected with each other? Are networks of Protestant organizations divided along theological lines? Are evangelicals less connected with social services than mainline organizations? Second, surveys and historical studies sometimes assume that the nature of a particular group is the aggregate of the personal characteristics of its individual members. Yet, organizations are characterized by their connections with each other as well as by their individual characteristics (Wellman 1988). Therefore, a study of networks among Canadian evangelicals complements surveys of their individual characteristics (e.g., Grenville 1995). ___________________________________ 8 Rawlyk’s (1996) study, Is Jesus your Personal Saviour? serves as an example of how studies of Canadian evangelicals leave room for a more specific study of relations between organizations. Rawlyk’s study provides extensive information about the attitudes, opinions and beliefs of Canadian evangelicals, but includes no information about how they relate together and very little information about how they relate to their churches. [ Page ] 5 Third, existing studies sometimes only see networks as produced by Canadian evangelicals rather than as entities that also maintain and produce Canadian evangelicalism. Thus, to get a complete picture of how evangelical churches relate to each other we must turn to an approach that explicitly examines the structure of relationships among evangelical churches not just the attributes of the individual people or organizations. One such approach is social network analysis. This sociological approach to relationship data provides the means to deal with structural questions about religious organizations (Wellman 1988; Wellman and Berkowitz 1988:31).9 For example, how are resources exchanged among organizations in Kitchener-Waterloo? Network analysis enables the identification of network structures, facilitates the identification of significant groupings, and helps to interpret the function of such groupings (Wellman 1988; Scott 1991). The units in a network system are interpreted in terms of the relationship between those units (Richardson and Wellman 1985:771). This structural manner of viewing relationships facilitates my exploration of organizational relations at the level of an entire community. Such an approach to the study of Canadian evangelicals provides significant information about how these organizations operate that cannot be provided by a study of their characteristics alone. I provide a more detailed description of the relevance of a network approach to the study of religious organizations in section 4.2. To limit such a study to a manageable size my exploration of evangelical connections begins at the level of the community, as this is where most evangelical organizations function. The most common evangelical organizations are local churches, not social services or networking organizations, therefore my study concentrates on relationships that local Protestant churches have with other religious organizations in a single geographic location. 1.3 Connectivity: Preliminary Findings The survey of organizational characteristics and my literature review address my first research question: Who are the religious organizations I am studying and how are they similar? I then use the network connectivity section of the survey to answer my second and third research question: How do these religious organizations interact? How can we account for network structure? What these three questions have in common is that they assume that ______________________________________ 9 A structural question is a question that is concerned with the relationships between organizations rather than just the characteristics of the individual organizations. [ Page ] 6 evangelical churches will interact and that that connectivity will be related to organizational similarity. Existing studies of relationships among organizations led me to expect to find similarity-based connectivity. I expected dense connections among local religious organizations with connections based on organizational similarity (e.g., theology, denomination). I anticipated dense connections because organizational similarities would lead to common goals which encouraged interaction. Furthermore, while churches are well known for various types of community involvement, I expected this involvement to be based on the characteristics of individual churches. For example, although different types of churches are involved in both evangelism and social services, the common wisdom is that evangelical churches focus their efforts primarily on evangelistic organizations and are generally isolated from the community (see chapter 4). I expected that the data would highlight this pattern both in terms of the number of connections to specific types of organizations and in terms of the type of resources moving through those network connections. My expectations were inaccurate in some respects. I did find connectivity among similar organizations—although not anywhere close to the degree that I expected. I also found that evangelicals are highly connected with their community. However, evangelical and mainline churches did not divide into the two neat camps I had anticipated. There were some evangelistic organizations that only attracted evangelical connections and there were some social services that only attracted mainline connections. Nonetheless, most organizations of both types had connections to churches across the theological spectrum. Furthermore, there appears to be little difference in the degree to which evangelical and mainline churches interact with social services. 1.4 Data Collection and Analysis Data about the structure of Canadian evangelical organizations were compiled from existing surveys (e.g., Bibby 1987; Bibby 1993; Bibby 2002) and historical studies of Canadian evangelicals. I also collected data through unstructured consultations with local community leaders who were in a position to observe local religious networks at the time of contact (academic, civic, media, evangelical, etc.).10 Additional specific historical details _____________________________________ 10 All field research was approved by a Research Ethics Board at the University of Toronto. [ Page ] 7 were drawn, as needed, from personal conversations and from the various publications of representative groups (ads, web sites, doctrinal statements, etc.). The initial list of Protestant churches was drawn from these sources and from an examination of publicly available lists of Protestant organizations.11 Once the initial study population was identified, a mail survey was used to collect quantitative data from these organizations. The survey collected data about both the structural characteristics of the organizations and their individual attributes. Organizations that had been missed in the original enumeration but which showed up in the content of returned surveys were also sent surveys. Returned surveys were entered into a database and analyzed using the statistical package SPSS (SPSS Inc. 2002) and the network analysis packages UCINET and Krackplot (Krackhart et al. 1993; Borgatti et al. 1999). 1.5 Chapter Summary In chapter 2 I use surveys and historical studies to define the term “evangelical,” and to explain how the term is used in my study. I develop a means of defining evangelical that takes into account a range of characteristics and is effective for classifying organizations for the purpose of network analysis. Building on this exercise of identification, the next chapter reviews examples of networking among evangelicals in the historical literature. My aim is to highlight some of the characteristics of Canadian evangelicals, and to demonstrate why a network analysis of these evangelicals would be profitable. This chapter ends with a brief description of the community chosen to illustrate these relationships among religious organizations. Chapter 4 builds on the context described in chapters 2 and 3 to develop some propositions about what local networks will look like and how they will operate. It also describes how network analysis can be profitably used to improve our understanding of evangelical organizations. My fifth chapter describes how these propositions were tested, describes the design of the study itself and how it was conducted. Following this description of research methods, the longer sixth chapter describes, summarizes, and analyzes the collected data. It also explores how the data contributes to explaining the propositions and discusses why the data might look the way they do. _______________________________________ 11 Such lists include the organizations involved in the Evangelical Fellowship of Canada’s effort to communicate evangelical Christianity to all Canadians by the year 2000 (Year 2000 and Beyond formerly known as Vision 2000), the Canadian Christian Sourcebook, The Shepherd’s Guide to local Christian organizations, membership lists of local church associations, and the telephone book. 2 DEFINITION: WHAT IS AN EVANGELICAL? In order to study evangelical networks effectively I need to establish who I am talking about. Thus, this section defines “evangelical” for studying Canadian evangelical organizations. In the first section of this chapter I point out problems in existing definitions of “evangelical.” I also describe an approach to defining “evangelical” that accepts some ambiguity and does not require sufficient or dominant characteristics for classification, but does provide a quantitative way to identify, classify, and compare evangelical organizations. In the process, I build an argument for the necessity of a different way of looking at the definitional task. The second part of this chapter includes a discussion of what one such approach to defining “evangelical” for studying evangelical organizations would look like. This approach to defining evangelical provides the context for subsequent discussions of how evangelicals interact with each other and with their local environment. 2.1 Identifying Evangelical Organizations: A New Look at an Old Problem Evangelicals are spotted more often in Canada than they used to be. This is not because of rapid growth or because of an effective publicity campaign but because both the popular media and academics are beginning to note their presence as an interesting element of Canadian culture (see pollster Grenville (1997), historian Rawlyk (1996) and journalist Swift (1993)). However, despite increased attention and study there is little consensus about what makes a Christian organization an evangelical organization. Identifying evangelical organizations is like recognizing family members: you can tell they belong together and can pick out some features that they have in common but you can’t quite tell how you know that they are from the same family. Likewise sociologists and historians of religion believe they know which organizations are evangelical and can pick out some features that they have in common but they have been unable to agree on exactly what it is that justifies classifying them together and what distinguishes them from other types of Christians.12 A series of sessions at the American Academy of Religion on defining ____________________________________ 12 Rosch (1978:36) explains that people can agree on categorization but disagree on boundaries. [ Page ] 8 [ Page ] 9 “evangelical” failed to produce anything near consensus—the scholars could not even agree that the term “evangelical” was useful (Dayton and Johnston 1991). The task of defining “evangelical” is confused by the diversity of evangelical organizations and is further complicated because evangelicals, by their own admission, cannot agree on what they stand for (Noll 1984; Smith 1986). Despite these ambiguities, some sort of definition is necessary to identify and describe evangelical organizations, to group like evangelical organizations together, to distinguish evangelical organizations from other types of Christian organizations, and to facilitate comparison and contrast in both historical and sociological studies. The difficulty scholars have had accomplishing these tasks is partially rooted in their reliance on a particular type of definition. According to Webster’s Ninth New Collegiate Dictionary, a definition is “a statement expressing the essential nature of something” or “the action or power of describing, explaining or making definite and clear” (Mish 1989: 334).13 Definitions of “evangelical” are trying to make the “essential nature” of evangelicals “definite and clear.” This is impossible because there are no characteristics that are shared among all evangelicals and are not shared with any other organizations. That is, there are no characteristics that in and of themselves are adequate to define “evangelical.” Such characteristics are known as “sufficient” characteristics. 2.1.1 Definitional Proliferation Scholarly approaches to defining “evangelical” commonly use historical, theological, and social-scientific methods (e.g., Hunter 1981; Marsden 1987: 58). This section will evaluate their usefulness for identifying evangelical organizations. 2.1.1.1 Historical Historical definitions classify evangelicals by connecting them to a particular stream of history (shared lineage, shared perspective on doctrine, organizational structure, classic emphases, etc.). For example, Stackhouse argues that evangelicalism is defined by the character of those Christians who belong to the broad historic stream that flows out of the sixteenth-century Protestant Reformation, down through the Puritan and Pietist channels, and into the so-called evangelical revivals of the eighteenth century. Those who (a) descend from _____________________________________ 13 Likewise, classification systems sometimes expect categories to be mutually exclusive (Bowker and Star 1999:11). However, as Bowker and Star (1999:11) explain, “no real-world working classification system that we have looked at meets these ‘simple’ requirements.” [ Page ] 10 these sources without departing from the central convictions that defined them in the first place, or who (b) later join up with the mainstream, are likewise identified as ‘evangelical.’ (1993: 7, emphasis in the original) Stackhouse (1993:7) argues that the identification of evangelicals does not “depend in the first place upon some abstract definition” but rather “the abstract definition come[s] out of the study of indubitably evangelical groups.”14 This type of definition facilitates Stackhouse’s examination of groups commonly considered evangelical. However, as an instrument to compare evangelical organizations with each other and to differentiate evangelical organizations from other Christian organizations this definition is less useful because it has some significant ambiguities. For example, how are groups who have departed from the central convictions of evangelical history (such as, “the unique authority of scripture . . . salvation through faith alone in Jesus Christ . . . concern for warm piety in the context of a disciplined life . . . the evangelism of all people,” Stackhouse 1993:7) identified? It is also unclear if an organization needs all of these characteristics to be evangelical. How much of an organization needs to have these characteristics? For example, does a church cease to be evangelical if 60% of its members do not hold to these central convictions? What if the pastor and the church’s statement of faith still hold to them? Finally, what is it about these central convictions that distinguish evangelical organizations from other Christian organizations? Stackhouse (Stackhouse 1993:3-5) himself notes that many individuals in the United, Anglican and Presbyterian denominations, which he contrasts with “evangelicals,” would share these central convictions. Stackhouse (Stackhouse 1993) is aware of many of these ambiguities and, as a result, the first chapter of his book on Canadian evangelicals explores a variety of factors that may suggest a group is evangelical. He does not coalesce these indicators into a concise definition because he recognizes that evangelicalism, if it can be described as a movement at all, is not a movement with clear boundaries or central organization and his primary concern is description not definition. Thus, Stackhouse’s definition, serving as an example of historical definitions, illustrates the need for a different type of definition for “evangelical,” as it requires careful attention to organizations individually in order to compare them. This thesis seeks a definition which can quantify “evangelicalness” to allow a less time-consuming comparison between a large number of organizations. ___________________________________ 14 See also Johnston (1991: 260ff) on “evangelical as description, not definition.” [ Page ] 11 Historical definitions can be adequate when used for a specific purpose or for looking at a particular organization at a particular place and time. However, the historical lineage and characteristics of evangelical organizations are too ambiguous for the larger task of differentiating evangelical organizations from each other and from other types of Christian organizations beyond very general categories. Nonetheless, the tendency of historical definitions to describe an organization based on a variety of different elements is an insight that will prove useful for developing a definition that may be used for quantitative purposes. 2.1.1.2 Theological Given the amount of effort put into theological clarification by Christian organizations and the amount of controversy surrounding perceived or actual transgressions of a given orthodoxy, we might expect a high degree of clarity and exactness could be obtained from a definition based on theological characteristics. Unfortunately, this is not the case. The many controversies, rather than clarifying the task of definition, have produced almost endless diversity within Christianity with the result that classification systems have often glossed over important differences within classes or important similarities between classes. For example, when about 70 churches left the United Church of Canada over the issue of homosexual ordination they insisted that they “were fully comfortable with the basic statement of faith of the United Church” (Greenshields and Robinson 1992). Theological definitions can rarely differentiate between churches that share similar statements of faith but interpret them differently. A list of theological characteristics most common among evangelicals can be helpful as an ideal type, but it provides only limited assistance in differentiating between evangelicals. An example of this within evangelical churches is the disagreement biblical literalists have over whether or not women are allowed to be elders (Balswick and Balswick 1990; Dueck et al. 1993). An additional problem is the difficulty of deciding what a defining characteristic of evangelicalism is and what degree of adherence to a given characteristic is required to be identified as an evangelical. It would be difficult to find a set of commonly understood, doctrinal characteristics that are shared by all evangelicals and are not shared with non-evangelicals. In view of the difficulty of finding characteristics that are shared exclusively by all evangelicals, some scholars have turned to classification schemes that divide evangelicals [ Page ] 12 into different types. Three common approaches are classification according to specific theological history, classification according to responses to theology and classification according to theological focus (Moberg 1977; Johnston 1991; Weber 1991). he main problem with these approaches is the artificial nature of their boundaries. In order to classify an organization, one has to decide which characteristic(s) of that organization should be given prominence since they are likely to have characteristics in more than one of the classes. Identifying representative dominant characteristics is complicated by the influence of time, leadership, and culture on dominant characteristics. Other scholars, recognizing these limits, have developed even broader classification schemes. This sometimes results in a description of evangelicalism that is almost as complicated as a description of each organization individually (e.g., Webber 1978: 32, 33). Despite such difficulties, these approaches to classification can be useful for identifying trends within evangelicalism, can help to identify sources of conflict and co- operation, and can point to shared characteristics among evangelical organizations. However, their tendency to cluster organizations together that have very little in common and to split single organizations into different classes; their dependence on the identification of sufficient or dominant characteristics; and their reliance on ideal types rather than the unique characteristics of each individual organization limits their usefulness to scholars who aim to identify and compare evangelical organizations beyond the studies for which these schemes were designed. 2.1.1.3 Social-Scientific Classifications There is a large amount of social-scientific scholarship concerned with the classification of religious organizations. A summary of three of these approaches shows that they do not overcome the deficiencies of the historical and theological definitions. Allowing religious organizations to classify themselves appears to be a solution to the definitional problem. However, a closer look reveals that self-classification is sufficiently problematic to be of limited use as a definitional tool, although it can be an important aid in identification. People are often unaware of the appropriate classification of the organizations to which they belong (Smith 1990: 35-37). Furthermore, organizations (or individuals) need not identify themselves as evangelical to be evangelical (Bebbington 1994: 367). For example, organizations that are sometimes considered evangelical (e.g., Restorationists) [ Page ] 13 often reject the label (Hughes 1991). In addition, other organizations use the term but mean something different by it. Evangelical Lutherans use the term “evangelical” according to its historical definition (i.e., gospel). 15 Finally, religious groups such as Mennonites and denominations such as the Anglican/Episcopalian Church often contain both evangelicals and non-evangelicals, sometimes in the same congregation (see Hunter 1981 for a discussion of this problem).16 Another well-represented means of classifying religious organizations is the church- sect typology. This form of classification was first developed by Ernst Troeltsch and Max Weber and has since been pursued in various forms by numerous scholars (e.g., Troeltsch 1931; Niebuhr 1957; Weber 1958; Welch 1977; Dawson 1992). Troeltsch’s objective was to distinguish between different types of Christian organizations. Troeltsch, influenced by his early 20th-century setting in Germany, divided Christian organizations into culturally connected churches and non-conformist, voluntarist sects. Evangelical organizations would be considered sects. Many scholars have taken Troeltsch’s basic idea and developed classification schemes from it. The main problem with this sort of classification scheme is that it, too, relies on sufficient characteristics. The result is a typology that only partially applies to most organizations. Organizations that share two seemingly contradictory characteristics (e.g., an organization that is both introverted and yet aims to reform society) must be classified in such a way that significant aspects of its character must be downplayed. Such a scale requires that the researcher determine which characteristics are more prevalent than others—a task that may not be possible. Thus, church-sect typologies have many of the same problems as theological classification systems.17 However, Dawson argues that “the church-sect typology was never meant to be a taxonomy” (1992: 15). He further suggests that __________________________________ 15The term “evangelical” comes from the Greek euaggelion, the koine word which is usually translated as gospel and literally means “good news” (Stackhouse 1993:6). It was not a term that first-century Christians used for themselves; rather the term first came into widespread use during the Protestant Reformation to distinguish Protestants from Roman Catholics (Nash 1987:22, 23). Initially, the term referred primarily to Lutherans but soon came to encompass all Protestants (Gerstner 1977:23). In Germany and Switzerland, the term was also used to distinguish Lutheran Protestants from those following Ulrich Zwingli and John Calvin, who were known as “Reformed” (Stackhouse 1993:6). 16 These problems have not stopped scholars from using self-identification to classify evangelicals (Smith et al. 1998) but it does keep their results from being directly comparable to other studies which use different means of classification. See also Schmalzbauer’s (1999) comments in his review of Smith et. al.’s American Evangelicalism in Sociology of Religion 60:3. 17 See Stackhouse (1993: 12-16) for further description of the inadequacies of the church-sect model as applied to evangelicals. [ Page ] 14 church-sect theory might be better understood as some kind of continuum (1992:18). This suggests that a more complex definition is needed for defining “evangelical.” A third social scientific approach to identifying evangelicals is classification by belief. Hammond and Hunter’s Index of Evangelical Beliefs can serve as an example of this definitional approach (Hammond and Hunter 1984). Their goal is to compare evangelical beliefs to worldviews in order to assess the impact of college education on evangelical beliefs. Although the beliefs Hammond and Hunter have chosen to highlight are common among evangelicals, they provide no reason to believe that they are necessary or sufficient characteristics of evangelicals. It even remains unclear what is distinctive about these beliefs that would allow them to distinguish evangelicals from Roman Catholic or Eastern Orthodox Christians (a differentiation that was unnecessary for the purposes of their study). In addition, their sample is based on a limited universe (i.e., college students) that may not be representative of the entire population. Thus, their index does not aid in identifying evangelical organizations and is an inexact tool for classifying or comparing evangelical organizations. In fairness to Hammond and Hunter, their study was not designed for the quantitative identification, classification, and comparison of evangelicals beyond a specific goal and was not designed to be used on organizations. However, an evaluation of their study’s usefulness for this purpose again illustrates the weaknesses of characteristic lists for defining “evangelical.” Nonetheless, their study has illustrated how the examination of selected characteristics may provide significant data for the examination of religious organizations. 2.1.1.4 Bebbington’s Definition Yet another approach to defining evangelicals is to combine historical, theological, and social-scientific elements into a single definition. This is David Bebbington’s approach (Bebbington 1989: 2-17). His definition of “evangelical” deserves special comment because it is currently one of the more popular definitions in use by historians of evangelicalism (see its usage in Rawlyk 1997), and because it avoids many of the common pitfalls of the other definitional schemes by emphasizing the importance of a variety of different types of characteristics. Furthermore, it highlights the main weakness of most current definitions of “evangelical,” and it is a first step towards a new approach to identifying evangelical organizations. [ Page ] 15 Christians, argues Bebbington, are evangelicals if they possess four necessary characteristics: biblicism (a reliance on the Bible as ultimate religious authority), conversionism (a stress on the new birth[/conversion]), activism (an energetic, individualistic approach to religious duties and social involvement), and crucicentrism (a focus on Christ’s redeeming work as the heart of essential Christianity) (Bebbington’s definition as interpreted by Noll et al. 1994: 6). The combination of these four characteristics is considered sufficient to identify evangelicals. This definition has become widely adopted because it accurately describes most organizations commonly considered evangelical. However, it also includes organizations commonly not considered evangelicals. For example, some Roman Catholic organizations would share these characteristics but may be rejected as being evangelical because of other characteristics such as their belief in the authority of the Pope and the role the Virgin Mary plays in a Christian’s relationship with God.18 An additional problem with Bebbington’s definition is the gap between the actual practice of evangelical organizations and their theoretical ideals. At times one or more of these four characteristics appear to be missing from organizations commonly understood to be evangelical (e.g., Christian Reformed churches, many of whom self-identify as evangelical, can act as if evangelism is not important to them). Related to this problem is the difficulty of deciding what degree of adherence to these characteristics makes an organization evangelical. For example, should organizations which believe in the verbal inspiration and inerrancy of the Bible be classified with organizations that believe that the Bible has ultimate authority in each individual’s life but is not inerrant? Bebbington’s description of evangelicals is inadequate for the quantitative examination of evangelical organizations in part because it was not designed for that purpose. Nevertheless, if the scholar’s goal is only to describe a particular type of conservative Christian then Bebbington’s definition is entirely adequate. Furthermore, Bebbington’s definition is another step towards a definition that takes into account a variety of different types of characteristics in order to identify evangelical organizations. 2.1.1.5 Summary Bebbington’s definition highlights the main deficiency running through all these attempts at definition and classification when applied to a quantitative study of organizations. _____________________________________ 18 A recent survey using this definition as a basis for defining evangelicalism did, in fact, classify some Roman Catholics as evangelical. See the usage of this survey in Rawlyk (1996). [ Page ] 16 They rely on sufficient or dominant characteristics to determine who is and who is not an evangelical. This results in ideal types that are hard to match to actual organizations. Thus, attempts to define “evangelical” beyond the narrow confines of a singular purpose have not been successful, largely because most commentators have been more interested in the description of an already identifiable organization. Furthermore, definitions designed to identify evangelical individuals may not be useful for the identification and comparison of evangelical organizations. Given these divergent definitions and the problems associated with them, the temptation is to give up and agree with Leonard Sweet that the only “indisputable facts about the evangelical tradition” are that it is important, it is understudied and it is diverse (in Johnston 1991: 252). However, these many different approaches to defining “evangelical” have highlighted some of the features a definition useful for the study of organizations should have. A definition of “evangelical” adequate for quantitative identification, classification and comparison of evangelical organizations would identify necessary characteristics, would take into account ranges of belief of a given orthodoxy in the entire population (rather than focusing on sufficient or hard-to-determine dominant characteristics), would not place firm boundaries on definitions, and would be concerned with ideals and actions. 2.1.2 Towards a Polythetic Definition 2.1.2.1 Family Resemblance19 A definitional approach known as family resemblance can be used to meet many of these requirements. The term “family resemblance” was used by Ludwig Wittgenstein to explain the complex intertwining and overlapping features of the members of a class. Wittgenstein explains that “these phenomena have no one thing in common which makes us use the same word for all, — but that they are related to one another in many different ways” (Wittgenstein 1963: 31). He goes on to use the example of games, noting that, although elements are shared between games, all the same elements do not exist in all games (e.g., compare baseball with ring-around-the-rosie). Moreover, it is not at all clear where to set the boundary between a game and a non-game (e.g., are military war games games?). This problem does not render the definition of game useless, as exact boundaries are only ____________________________________ 19 The task of categorizing the world we live in has engaged scholars interested in many different topics (Rosch and 1978; Bowker and Star 1999; Ragin 2000). [ Page ] 17 necessary for exact purposes (1963: 33). Sometimes, explains Wittgenstein, an inexact definition is the best way to describe something. Such definitions are like a thread where “th strength of the thread does not reside in the fact that some one fibre runs through its whole length, but in the overlapping of many fibres” (1963: 32). Wittgenstein’s concept of family resemblance shows us that definitions do not have to be exact to be useful and that a given class of things does not have to have necessary or sufficient characteristics to be recognized.20 Related ideas which view evangelicals as a mosaic or a network are common in the popular and the academic literature (e.g., Nash 1987; Noll et al. 1988; Stackhouse 1993; Noll et al. 1994; Mackey 1995). These ideas show that the family resemblance concept is not an entirely new idea for defining “evangelical” as much as it is a complement to existing scholarship which can be translated into a definition useful for the quantitative study of evangelical organizations. Family resemblance is a useful starting point for this task but it does not yet have enough specificity to serve as a sociologically useful definition of “evangelical.” This can be accomplished by an approach to definition drawn from biology known as a polythetic definition. 2.1.2.2 A Polythetic Definition The problem with the definitional and classification schemes that have been discussed thus far is not the information they have available to them, but rather their way of grouping that information. Most scholars who have written in this area have used what Smith (1982:2) calls a monothetic classification scheme. In a monothetic system the scholar looks for one (or more) unique defining characteristic(s) (e.g., doctrine, history, culture) without which the item would be something else. Smith (1982:5, 6) explains that although “uniqueness is an ordinary presupposition of definition and classification” there is no reason to suppose that there is one unique characteristic that distinguishes a given group from any other as there is also no reason to assume that any of the characteristics of a group are unique to that group. It is, in fact, more reasonable to assume that all identifiable groups in society share some of their characteristics with other groups in that society (although all groups do not share all characteristics). This, argues Smith, is why classification of religious groups should be based on a polythetic definition. A polythetic definition would retain “the notion of necessary but ___________________________________ 20 See related ideas in Rosch (1975) and Ragin (2000). [ Page ] 18 [abandon] the notion of sufficient criteria for admission to a class” (1982: 4). A class would be “defined as consisting of a set of properties” of which each individual of the class must possess a “large” but unspecified number of these properties and each property should be possessed by a “large number” of individuals in the class with no single property necessarily being possessed by everyone (1982:4). Table 2.1 provides a graphic representation of the differences between polythetic and monothetic classifications. Table 2.1 - Polythetic and Monothetic Definitions21 [ Please contact repository@tyndale.ca for Table 2.1 details ] Characteristics Groups 1 2 3 4 5 6 7 A A A B B B C C C D D D D F F F G G G H H Groups 1-4 form a polythetic grouping, whereas groups 6 and 7 form a monothetic grouping. Notice that no one characteristic is shared by all of the polythetic groups. Group 5, although it shares a single characteristic with groups 2-4, would not be classified with them because it does not share the majority of its characteristics with other members of the group. Characteristic D is a necessary characteristic for groups 2-4 (which is also a polythetic grouping) but it is not a sufficient characteristic as it is also shared with group 5. 2.1.3 Application to Evangelicals To define “evangelical” using a polythetic definition, a set of the most common characteristics of evangelical organizations could be drawn up. Some of these characteristics would be the “defining” characteristics highlighted in the previous sections such as conservative doctrine (e.g., the divinity of Jesus, the Bible as ultimate authority), specific history (e.g. the Protestant Reformations, the evangelical awakenings) and individual characteristics (e.g., personal spiritual experience centred on Jesus Christ, a sense of the degeneration of the moral and spiritual world, an affectionate affiliation with other __________________________________ 21 Based on a chart by Sokal and Sneath in Needham (1975: 357). [ Page ] 19 evangelicals), to name but a few possibilities. Many of these characteristics would not be shared by all organizations and none of them would be sufficient. A small number of these characteristics would be identified as necessary characteristics (possibly based on characteristics most considered necessary in existing definitions) and would be shared by all “evangelical” organizations but could also be shared with some non-evangelical organizations. An organization could be identified as “evangelical” if it has all the necessary characteristics and a large number of the other common characteristics of evangelicals. Thus, organizations could be assigned a numeric score based on the number of characteristics they had. Organizations could be classified according to this score or on their score in individual sections of the index.22 In addition to Bebbington, there are at least two scholars who have developed a definition similar to this idea. They both illustrate how a polythetic definition could be used and highlight why more work needs to be done. Lyman Kellstedt has developed a polythetic scale for the study of the political behaviour of evangelicals (Kellstedt 1989). His scale begins with minimalist criteria of Christ’s divinity, Christ as the only way to salvation, an inerrant Bible, and a commitment to spreading the Gospel and adds to them “necessary but not sufficient” beliefs (i.e., God, having a religious preference, life after death, importance of faith, consolation from religion). Although Kellstedt shows his conceptualization to be useful for differentiation between religious individuals with different political beliefs, he has also made choices which eliminate people and organizations from the classification “evangelical” unnecessarily. For example, the doctrine of inerrancy is widely misunderstood in North American culture and organizations sometimes do not use the word to avoid controversy.23 Thus, people or organizations would be excluded from evangelicalism if they do not hold to all the details of the doctrine or do not identify with the term. In addition, Kellstedt uses “get ____________________________________ 22 Smith et. al. (1998:247) have some concerns about definitions which look at religious identity as being on a continuum. They argue that the use of a continuous variable neglects complexities in religious identities, “compresses multidimensional realities into one-dimensional linear continuum, and attempts to correlate features of traditions which are incommensurate.” They argue that religious identity should be treated as a categorical variable. However, when classifying religious organizations treating religious identity as a categorical variable does not take into account the degree to which different organizations have different features. Without taking into account different aspects of an organization one cannot be certain that two organizations self-classified into the same category are the same. While multidimensional characteristics may not be commensurate, they do allow organizations to vary within the index—as they do in their actual operation. In addition, my work addresses Smith et. al.’s (1998) concerns by using more than just theological variables. That is, my definition is not unidimentional. 23 See Ammerman (1982) for a critique of the use of inerrancy as a defining feature of evangelicals. [ Page ] 20 consolation from religion” as a filter to eliminate people from consideration as evangelicals (1989:15). However, it is unclear what this is measuring. Is this measuring an individual’s current feelings or doctrinal expectations (e.g., I am supposed to, therefore, I do)? Furthermore, such a filter is of little meaning when applied to the study of organizations because organizations are not consoled. Rather, the people in them are consoled. Andrew Grenville (1995) has developed a 10-item polythetic scale for use with population surveys which has proved to be a useful tool for the study of individual evangelicals (See Rawlyk 1996). This scale has two weaknesses for the study of evangelical organizations. First, it does not take into account the existence of necessary characteristics. Grenville’s study leaves open the possibility that someone who denies the deity of Christ and someone who does not attend church regularly will be given the same score on the scale. Second, the scale is designed to identify individual evangelicals as part of a large survey and of necessity must be easily understood and relatively short. The result is a scale which, at first glance, does not appear to take into account the complexity of religious organizations. Nonetheless, in actual use it could be effectively modified to identify evangelical organizations. 2.1.4 Summary These two studies and the others highlighted in this chapter have many of the elements necessary for a new approach to defining “evangelical” for the purpose of studying evangelical organizations but also point to the need for a more complex polythetic definition that focuses specifically on organizations. The second section of this chapter builds on these definitions to produce a definition of evangelical useful for this study. 2.2 Definition in Practice The study of any religious group eventually has to deal with boundary problems; who is in and who is out. In the previous section I concluded that the most useful way of defining “evangelical” for use in the quantitative identification, classification, and comparison of religious organizations was to treat “evangelical” as referring to a class of organizations sharing a limited number of necessary characteristics and a large number of general characteristics. I argued that it is unnecessary for classification schemes to have firm boundaries and that a fluid definition can be useful for quantitative comparison. This sort of [ Page ] 21 definition avoids the main pitfall of existing definitions—the reliance on sufficient or dominant characteristics to determine who is and who is not an evangelical. In this section, I explain how I developed a polythetic index of evangelical characteristics, I describe the use of such a definition, and I evaluate its usefulness for the purposes of my study.24 2.2.1 Initial Design To define “evangelical” using a polythetic definition I used existing definitions to draw up a set of characteristics that were most commonly mentioned as characteristics of evangelicals.25 These characteristics did not have to be shared with all other evangelicals and could be shared with nonevangelical groups. An individual or an organization could then be identified as “evangelical” if it had a large number of these characteristics. After compiling this set of characteristics, similar characteristics were grouped together. Characteristics that proved to be too vague or redundant were deleted. Representative necessary characteristics were identified based on which characteristics were most often considered necessary in the definitional literature. Other characteristics usually not occurring among evangelicals were included to test for respondent accuracy and acquiescence bias (these characteristics were reverse scored). Next, denominational representatives, chosen for their knowledge of the denomination and their familiarity with the academic task, were recruited to pre-test this index.26 I called the final index the Index of Evangelical Characteristics (see Appendix G for a list of the characteristics in the index). The final edition of this index was then used in a survey of 133 churches in one community.27 As _____________________________________ 24 The index I describe could easily be adapted for use in other contexts both by changing the list of characteristics included in it and by scaling it up or down as needed. 25 Although I have paid attention to a small number of studies of evangelicals outside of North America, the majority of the studies I have used are of North American evangelicals. As such, there may be other characteristics that would be important in other geographical locations. Furthermore, although a large number of definitions have been reviewed, there are still others that have not been examined. Thus, the set of characteristics I have compiled may be missing characteristics that some definitions deem important. Although additional characteristics may add additional specificity to comparisons of groups, the comparative results would be similar as, presumably, additional evangelical characteristics would be more likely to be held by evangelicals. 26 A test of the validity of the characteristics could also be obtained through textual study and through observation of these denominations. However, the viewpoint of the outside observer could never capture all the subtleties a participant would be aware of as a result of years of involvement within a denomination. 27 Although I sent out 143 surveys, for 10 of these organizations, I was unable to find them or they no longer existed. [ Page ] 22 an additional test this index was compared with a variation of Grenville’s “Christian Evangelicalism Scale” (Grenville 1995). 2.2.2 Scoring the Indices This section describes how an evangelicalness score was derived from the collected data. Of the 133 surveys distributed to churches, 88 surveys were returned. Each of the 88 churches which returned surveys was assigned a score based on how they filled-out the index.28 I assigned two scores to organizations. The first score was based on adding up each organization’s response. A second score was based on a factor analysis of a subset of the characteristics in the survey. 2.2.2.1 Non-returned Surveys An examination of returned versus not returned surveys shows no distinctive patterns in either denomination or evangelicalness.29 Most denominations had 50-70% of their churches return surveys which approximate the 60% of surveys returned. Likewise, return percentages for various levels of evangelicalness were also in the 50-70% range. The surveys represented a good range of denominations and evangelicalness with only denominations having two or fewer local churches not returning any surveys. 2.2.2.2 The Additive Score To compute the additive score I summed each organization’s score on each of the 76 characteristics.30 This produced an index where the lowest possible score (indicating evangelicalness) was 76 and the highest possible score was 532 (indicating non- evangelicalness).31 To make the measure more intuitive (i.e., higher numbers indicate more evangelical organizations) I subtracted the raw sum from the highest possible score (532) to give more evangelical churches a higher score. I did not increase the value of necessary __________________________________ 28 Of the returned surveys 87% filled in at least 67 of the 76 possible beliefs or characteristics. Of those 76 possible beliefs or characteristics, only 2 of them were left blank more than 26% of the time. Both of those items related to the Holy Spirit. I decided to leave them in for analysis purposes because they were more likely to be understood by evangelicals and thus improve my ability to differentiate between evangelicals and non- evangelicals. 29 In this case, evangelicalness is evaluated based on denomination and on my personal judgment as an evangelicalness score cannot be calculated for unreturned surveys. 30 Non-evangelical characteristics were previously reversed (by subtracting the raw score from 8) so that marking them as less important to a church indicated a more evangelical tendency. 31 Scores lower than 76 indicated organizations that did not fully complete the survey. That is they left a large number of the questions on section three of the survey blank. [ Page ] 23 characteristics because they did not prove to be distinctive of evangelical churches. To compensate for missing characteristics, scores for missing characteristics were subtracted from the total score (so that organizations with large numbers of missing characteristics did not move artificially to the top of the ranking).32 Thus, Evangelicalness = 532 - [(Sum of the importance score of evangelical characteristics) + Sum (8 - importance score of non-evangelical characteristics)) - (number of blanks * 7] The resulting scores ranged from a high of 413 to a low of 2.33 The diversity of results supported the argument of the previous section that evangelical organizations are better understood as standing along a continuum of more to less evangelical than as being members of discrete categories (see Graph 2.1). The variation between scores increases as one moves towards the end of the scale. For ease of comparison, these scores were ranked into four categories from most (1) to least (4) evangelical. In an effort to determine if a smaller index could serve the same purpose as the larger index, I produced an alternate index which contained the sum of 10 characteristics I identified as necessary for an organization to be called an evangelical organization (based on existing definitions of evangelical, see Table 2.2). I calculated the sum of the entries on each necessary characteristic. I then subtracted that sum from 80 (to make evangelical organizations have larger scores) and subtracted 7 for each characteristic they left blank. _____________________________________ 32 To evaluate the choice to subtract missing characteristics from the score a ranking was done both for the initial score and for the final score computed by subtracting 7 * the number of items left blank on the index. Both scores were then ranked into 4 categories. A comparison of the two ranking schemes showed that 60% of the ranks were the same and 92% were within 1 ranking of the other scale. Of the 13 churches which moved between the more evangelical top half of the ranking and the less evangelical bottom half of the ranking six of those moved towards being more evangelical. The remaining seven churches all left at least 20 blank characteristics on the index. They included seven of the eight churches with at least 20 blanks on the index. The 8th church with more than 20 blanks moved from being ranked 3 to being ranked 2 indicating a very high score on other indicators of evangelicalness. Thus, subtracting the missing characteristics tends to move churches with blanks lower in the index. This is justifiable since the missing characteristics reduce the ability of the index to determine evangelicalness accurately. Churches that left a large number of blanks on the index, and thus had a lower score on the index, were not necessarily non-evangelicals. In fact, one of them (1044) appears to be very evangelical because it belongs to an exclusively evangelical denomination. However, this church appears to be an exception resulting from an easily identifiable cause (a large number of missing characteristics). 33 Excluding the score of 2 (which is an outlier as a result of leaving 75 of 76 possible characteristics blank) there is a range of 216 points between the highest and the lowest score. Thus, there is an average of 2.48 points distance between each church according to the index. This is representative of the entire index. The 7 largest gaps were 23, 12, 11, 7 and 7 points apart. [ Page ] 24 Graph 2.1 - Additive Score Distribution34 [ Please contact repository@tyndale.ca for Graph 2.1 details ] At first it appeared as if this smaller index might serve as well as the larger index as its relative range of scores was similar to the larger index. However, a closer look showed that the variation in the Necessary Characteristics Index was all near the end of the index (see Graph 2.2). Out of a total score of 70, those organizations above the 25th percentile were within 11 points of each other. Organizations below the 25th percentile were 21 points apart, not including one informant who did not answer the majority of this section. Evangelical churches are closer to the top but the differences between organizations are small (78 organizations scored at least 50 out of 70) and there are a small number of stereotypical evangelical organizations and stereotypical non-evangelical organizations throughout the index. An even more restrictive list of characteristics resulted in similar problems. _____________________________________ 34 All 88 informant organizations are represented in the graph. The 18 organizations listed on the horizontal access represent all 88 informants. [ Page ] 25 Table 2.2 - Necessary Characteristics The authority of the Bible The existence of heaven Jesus Christ as bringer of redemption Jesus Christ as deity Jesus Christ died a physical death on the cross Substitutionary atonement Salvation Justification by faith alone Evangelism of non-Christians Personal holiness Graph 2.2 - Necessary Score Distribution [ Please contact repository@tyndale.ca for Graph 2.2 details ] I had assumed that possessing the necessary characteristics would be important to differentiate evangelical from non-evangelical organizations. However, the top 74% (65 of 88) of churches surveyed rated at least 9 of the 10 necessary characteristics 1, 2, or 3 (on a scale from 1 “Important to your Church’s Identity” to 7 “Not at all Important to your Church’s Identity”). Of those same churches, 48% (42 of 88) rated all 10 necessary characteristics 1, 2, or 3. Furthermore, using my polythetic definition, churches ranked high on the necessary scale are not all evangelical (e.g., an Evangelical Lutheran church). In [ Page ] 26 addition, clearly evangelical organizations omitted some of the necessary characteristics for reasons that are not always clear from the data. That is, some clearly evangelical churches only rated 9 of the 10 necessary characteristics 1, 2, or 3 (e.g., Pentecostal, Evangelical Missionary). The use of even stricter criteria (e.g., inerrancy, premillenialism, virgin birth, salvation only through Jesus Christ, association with evangelicals) sometimes excluded groups commonly considered evangelical (e.g., the Vineyard) and sometimes included groups commonly considered not evangelical (e.g., Evangelical Lutherans). While the necessary characteristics may be important to a church’s identity, they are not all of equal importance. Thus, the necessary characteristics do not discriminate between Christian organizations as much as I anticipated (see Table 2.3). Nonetheless, an additive ranking of all 76 characteristics proved sufficient to identify churches that are more evangelical. As a result, it was not essential to refer specifically to the necessary characteristics in my analysis. Table 2.3 - Additive Rank/Necessary Characteristics Comparison [ Please contact repository@tyndale.ca for Table 2.3 details ] Additive Ranking # of Orgs in Category # of Necessary Characteristics Ranked 1, 2, or 3 10 9 8 7 or less 1 22 18 4 2 21 13 8 3 24 11 7 5 1 4 21 4 4 13 The main reason to favour the additive score over the necessary score is that there is more variation throughout the index. For this reason and because the larger index is a better measure of the whole spectrum of an organization's characteristics, the larger index was used for evaluating the evangelicalness of each organization. Nonetheless, both indices generally rank churches in the same direction and additive rankings based on individual sections of the survey generally ranked churches in the same direction. This suggested that further testing and analysis could develop an index which was just as effective as the additive score but which required fewer items to produce. [ Page ] 27 2.2.2.3 The Factor Score I chose to use principal component factor analysis to determine if a smaller subset of characteristics would be just as successful at the job of differentiating Protestant churches from each other. The large scale may be necessary for differentiating between the various different types of evangelicals. However, as discussed above, it is likely that a smaller scale would be sufficient for ranking churches into categories based on relative evangelicalness—a ranking which would be sufficient for my study. A principal component factor analysis on all 76 characteristics produced 17 factors which explained only 81% of the variance. 35 It needed six factors to explain 51% of the variance. The 17 factors contained a large number of very different variables or explained very little of the variance. The first three factors only explained 39% of the variance (see Table 2.4). In addition, 49 variables had a significant correlation with the first factor (> 0.5). Given the difficulty of interpreting this result I ran a factor analysis on the 10 characteristics I had previously identified as necessary characteristics (see Table 2.2). This factor analysis found only one factor that explained 55% of the variance (see Table 2.5) . All 10 variables were highly correlated with the first factor (> 0.5). Table 2.4 - Rotated Factor Matrix (All Variables) [ Please contact repository@tyndale.ca for Table 2.4 details ] Components Variables Orthodoxy Salvation Personal Conservative theology 0.85 0.10 0.13 Non-Christians lost 0.84 0.17 0.25 Satan is an actual spiritual being 0.83 0.19 0.16 Total depravity of humanity 0.83 0.13 -- Liberal theology -0.82 -0.23 -- Evangelism of Non-Christians 0.78 0.29 0.25 Salvation is only through Jesus Christ 0.78 0.28 -- Bible is final authority for living 0.77 0.25 0.12 Inerrant Bible 0.77 -- 0.20 Belief in absolute truth 0.77 0.22 0.27 Hell as a place of physical torment 0.74 0.17 0.10 Bible is literal word of God 0.74 -- 0.31 ________________________________________ 35 In all factor analyses I replaced missing values with the mean. A mean substitution is where you replace missing values with the mean of existing values. A mean substitution does introduce some error because it produces artificial answers to questions left blank. However, most organizations left few blanks. Furthermore, using mean substitution allowed me to produce a score for all 88 organizations. Nonetheless, the introduction of this artificial data is one more reason why I chose to use the additive score. [ Page ] 28 Components Variables Orthodoxy Salvation Personal Literal bodily resurrection of Jesus Christ from the dead 0.74 0.15 -0.18 Virgin birth 0.73 0.10 0.17 Bible is the sole authority for faith and practice 0.73 0.25 -- An optimistic view of human nature -0.71 -0.11 -- Associate with Non-Christian Religions -0.67 -0.12 0.17 Heaven 0.67 0.17 -- Associate with Mainline Churches -0.66 -- -0.15 Substitutionary atonement 0.66 0.22 -- Jesus Christ dispenses judgment 0.65 0.12 0.33 Inter-religious Dialogue -0.65 -0.15 -- Imminent second coming of Jesus Christ 0.64 0.33 0.25 Associate with Roman Catholics -0.63 -- -0.12 Salvation is necessary for all people 0.63 0.53 0.12 Salvation leads to eternal life 0.61 0.38 -- Insufficiency of human merit 0.61 0.49 -- Bible is verbally inspired 0.59 -0.21 0.12 Real historical character of all of Jesus Christ's acts 0.59 0.17 0.24 Justification by faith alone 0.59 0.44 -0.19 Authoritative Bible 0.56 0.16 -0.13 Personal holiness 0.55 0.23 0.37 Bible Colleges 0.50 -- 0.25 Divinely inspired Bible 0.49 0.45 0.18 Jesus Christ performs miracles 0.48 0.38 0.17 Bible is Normative for Christian belief and practice 0.47 -- -- Essential role of the Holy Spirit in salvation 0.46 0.42 0.33 Separation from the world 0.42 -- -- Jesus Christ brings redemption 0.35 0.71 -- Salvation leads to a changed life 0.40 0.70 0.16 Jesus Christ as mediator between God and humans 0.46 0.68 -- Salvation 0.45 0.65 -- Disciplined spiritual life 0.25 -- 0.85 Adult baptism -- -- 0.65 Personal conversion 0.55 0.25 0.62 Daily prayer 0.16 0.37 0.53 Daily Bible reading 0.27 0.39 0.51 % Variance Explained 27.15 6.93 5.38 Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. - Rotation converged in 32 iterations. - Mean substitution for missing values Absolute values <.1 displayed as -- This table only displays factors with a variance score > 5% and loadings > 0.5. The complete factor table can b found in Appendix K [ Page ] 29 Table 2.5 - Component Matrix (Necessary Characteristics) [ Please contact repository@tyndale.ca for Table 2.5 details ] Variable Component Evangelicalness Jesus Christ is Deity .83 Evangelism of Non-Christians .82 Heaven .78 Substitutionary atonement .77 Salvation .77 Justification by faith alone .72 Jesus Christ brings redemption .72 Personal holiness .71 Jesus Christ's physical death on the cross .66 Authoritative Bible .63 % Variance Explained 55 Extraction Method: Principal Component Analysis. As a point of comparison, I identified 10 characteristics which roughly approximated eight characteristics in Grenville’s “Christian Evangelicalism Scale” (1995) (see Table 2.6). This factor analysis found two factors that alone explained 62% of the variance (see Table 2.7). I tried to approximate more closely Grenville’s scale by eliminating the variables “Jesus Christ’s imminent second coming,” and “literal bodily resurrection from the dead” from the score. However, doing this had no effect. Table 2.6 - Approximating Grenville’s “Christian Evangelicalism Scale” Bible as the literal world of God Jesus Christ brings redemption Jesus Christ is deity Jesus Christ’s imminent second coming Literal bodily resurrection from the dead Physical death on the cross Opposition to secular humanism Daily Bible reading Evangelism of non-Christians Weekly church attendance [ Page ] 30 Table 2.7 - Rotated Factor Matrix (Grenville Characteristics) [ Please contact repository@tyndale.ca for Table 2.7 details ] Variables Component Action Jesus Daily Bible reading 0.90 -- Evangelism of Non-Christians 0.72 0.47 Imminent second coming of Jesus Christ 0.71 0.45 Weekly church attendance 0.70 0.19 Bible is literal word of God 0.63 0.42 Jesus Christ brings redemption 0.53 0.48 Jesus Christ is Deity 0.21 0.83 Literal bodily resurrection of Jesus Christ from the dead 0.31 0.74 Jesus Christ's physical death on the cross 0.20 0.72 Secular humanism -0.14 -0.66 % Variance Explained 31.93 30.37 Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. Rotation converged in 3 iterations. Mean substitution for missing values Absolute values <.1 displayed as -- I then tried a factor analysis on a combination of my necessary characteristics and Grenville’s characteristics. The combined 16 characteristics found four factors that explained 74% of the variance (see Table 2.9). To increase the power of this explanation I eliminated variables with the lowest extraction values (< 0.6, without mean substitution) and I removed variables which had three or more missing cases (see Table 2.8). The remaining eight variables produced two factors which still explained 69% of the variance (see Table 2.10). Table 2.8 - Remaining Variables The authority of the Bible The existence of heaven Jesus Christ as bringer of redemption Jesus Christ as deity Jesus Christ’s imminent second coming Justification by faith alone Evangelism of non-Christians Daily Bible reading [ Page ] 31 Table 2.9 - Rotated Factor Matrix (Grenville and Necessary Characteristics) [ Please contact repository@tyndale.ca for Table 2.9 details ] Variables Component Future Salvation Justification Action Bible is literal word of God 0.84 0.18 0.22 0.13 Personal holiness 0.73 0.13 0.26 0.34 Imminent second coming of Jesus Christ 0.72 0.31 0.20 0.34 Substitutionary atonement 0.64 0.56 -- 0.16 Evangelism of Non-Christians 0.64 0.20 0.46 0.40 Heaven 0.63 0.53 0.28 -- Jesus Christ is Deity 0.39 0.80 0.24 -- Jesus Christ's physical death on the cross -- 0.77 0.16 0.32 Salvation 0.27 0.61 0.24 0.34 Secular humanism -0.32 -0.50 -0.14 -- Authoritative Bible 0.28 -- 0.80 0.17 Justification by faith alone -- 0.43 0.76 0.14 Literal bodily resurrection of Jesus Christ from the dead 0.36 0.35 0.76 -- Daily Bible reading 0.48 0.12 -0.16 0.79 Weekly church attendance 0.12 0.12 0.36 0.75 Jesus Christ brings redemption 0.13 0.53 0.26 0.55 % Variance Explained 23.52 20.31 16.45 13.69 Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization Rotation converged in 14 iterations. Mean substitution for missing values Absolute values <.1 displayed as -- Two factor scores were computed from the final set of variables. The first score was computed using all the organizations and the second score excluded organizations with missing characteristics. There was little difference between the two scores. Therefore, I chose to use the first factor score because it produced scores for all 88 organizations. Every organization ranked one was given an identical score as they scored the same on all eight characteristics. The factor score produced a distribution comparable with the necessary score (see Graph 2.3). As with the additive index the distances between organization scores are relatively close at the top of the index, change gradually, and are more extreme towards the bottom of the scale.36 Thus, the factor score does not identify variation within rankings to the same degree as the additive score. In addition the two scores are highly correlated (see Table 2.11). ___________________________________ 36 The range for organizations ranked two was from -.656 to -.471, for organizations ranked three from -.457 to .358, and for organizations ranked four from .374 to 3.126. [ Page ] 32 Table 2.10 - Rotated Factor Matrix (Remaining Variables) [ Please contact repository@tyndale.ca for Table 2.10 details ] Variables Component Authority Activity Justification by faith alone 0.87 -- Jesus Christ is Deity 0.76 0.28 Authoritative Bible 0.71 0.16 Heaven 0.69 0.39 Daily Bible reading -- 0.95 Imminent second coming of Jesus Christ 0.47 0.72 Evangelism of Non-Christians 0.57 0.64 Jesus Christ brings redemption 0.47 0.57 % Variance Explained 38.51 30.11 Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. Rotation converged in 3 iterations. Mean substitution for missing values Absolute values <.1 displayed as -- Table 2.11 - Score Correlations (Spearman's rho)37 [ Please contact repository@tyndale.ca for Table 2.11 details ] Additive Score Factor Score Additive Score Correlation Coefficient 1.000 -0.768 Sig. (2-tailed) 0.000 N 88 88 Factor Score Correlation Coefficient -0.768 1.000 Sig. (2-tailed) 0.000 . N 88 88 Correlation is significant at the .01 level (2-tailed). _______________________________________ 37 The negative correlation is an artifact of the way I calculated the additive score. Characteristics rated very important are given lower values on the scale (from 1 - important to 7 - not important). Thus, the organizations which rank most of the characteristics very important actually have the lowest score. My additive score performs a calculation to give more evangelical organizations a higher score. Thus, the two scores are negatively correlated. [ Page ] 33 Graph 2.3 - Factor Score Distribution [ Please contact repository@tyndale.ca for Graph 2.3 details ] Organization ID Number (Sorted by Factor Score) I then used this factor score to divide the 88 cases into four categories as with the additive scale. Although there were two organizations with a ranking of three that appear to be definitively evangelical churches, I did not change their ranking. The first church is small, independent, and relatively free-spirited. Thus, although it clearly identifies itself with evangelicalism, it may be more on the outside than I anticipated. The second church is a large, long-established church of an evangelical denomination. Nonetheless, it did give a lower ranking to both “existence of heaven” and the importance of “evangelism of non- Christians” which suggests that some of its evangelical distinctives may have been muted over time. In both cases, these organization’s rankings will not significantly influence my analysis as my analysis is based on general trends within the community rather than specific scores on the indices of evangelicalness. 2.2.3 Using the Indices to Categorize Evangelical Organizations My polythetic approach to definition means that churches will not divide into distinct categories but will be positioned along a spectrum of evangelicalness. Organizations are not rated as evangelical or non-evangelical but rather as more likely to be evangelical or less likely to be evangelical. Nonetheless, for ease of comparison, the responding churches were [ Page ] 34 ranked into four roughly equal categories (see Table 2.12 and Table 2.13). Churches ranked one were the most evangelical churches, churches ranked two were less evangelical churches, and categories three and four were the least evangelical churches or churches which did not supply enough data to be ranked any higher. I used the mean to determine where to rank similar scores.38 This is not the only possible manner of ranking the churches but this method of ranking produced a conservative division between different categories of evangelicalness. That is, it is less likely to classify organizations as definitely evangelical (ranked one) than the sequential ranking of similar scores, the necessary score, or my judgment based primarily on denomination. Using a conservative ranking is less likely to overstate the influence of evangelicals. Any observed influence of evangelicals is likely to be significant. Thus, the use of four ranks and mean ordering was based on the interpretive value of this choice. The third column in Table 2.12 provides the percentage of organizations in each category. The final column provides the difference between the highest score in that category and the lowest score in that category. This column demonstrates that as we move from category one to category four the churches in my study become less similar. Table 2.12 - Additive Totals (Mean) [ Please contact repository@tyndale.ca for Table 2.12 details ] Category Number of Churches Percentage Point Range 1 22 25.0% 28 2 21 23.9% 25 3 24 27.3% 84 4 21 23.9% 74* * Not including the final church which had a score of 2 Table 2.13 - Factor Totals (Mean) [ Please contact repository@tyndale.ca for Table 2.13 details] Category Number of Churches Percentage Point Range 1 32 36% 0.00 2 13 15% 0.19 3 21 24% 0.81 4 22 25% 2.75 ____________________________________ 38 Ranking organizations with similar scores sequentially in the additive index placed approximately 65% of my informants in the top two categories of evangelicalness. [ Page ] 35 To provide a rough test of these rankings I compared the additive ranking and the factor score ranking to each other and to a ranking based on my personal judgment. There were substantial similarities between the three rankings. The additive scale ranked 77 organizations within one rank of my estimated rank and the factor score ranked 74 organizations within one rank of my estimated rank. In 48 cases the additive scale and the factor score scale produced exactly the same ranking. In an additional 24 cases the two scales were within one rank of each other. Score discrepancies among the remaining 16 organizations were often related to a large number of missing characteristics or low rankings on the eight characteristics used for the factor scale. This gives me confidence that both rankings are reasonable approximations of the relative evangelicalness of the different churches. 2.2.4 Evaluating the Final Indices This section reviews the usefulness of the two Indices of Evangelical Characteristics for quantitative identification, classification, and comparison of evangelicals. It also confirms the value of the polythetic approach to the definition of “evangelical” and the value provided by the format of the final indices. 2.2.4.1 Identification, Comparison, and Contrast The Indices of Evangelical Characteristics can identify and describe the relative evangelicalness of churches. The indices identify organizations as more or less evangelical both with relative scores, and through ranking organizations into four categories according to evangelicalness. The indices can also be used generally to classify different churches as more or less evangelical. Distinctions between and among organizations can be identified quantitatively and this can provide a starting point for further studies of those distinctions. For example, the indices can be used for detailed comparisons of specific organizations by looking at scores in individual sections of the additive index or by comparing scores on individual characteristics in either index. They can also be used for broad classification by ranking the scores produced by both indices. For example, the ranking of churches into four categories provides a useful way of comparing evangelicalness to network connections. The categories are distinct enough to be able to evaluate whether or not evangelicalness is a significant factor in network relations. I [ Page ] 36 expected that similar organizations would have similar network connections. The indices provide a measure of similarity to be able to evaluate this. Thus, the additive index can be used to divide groups according to doctrine or willingness to associate with certain institutions or to group them according to similar characteristics.39 The additive index aids in identifying sets of characteristics that could be used to classify organizations in different ways for different interpretive purposes. The indices facilitate comparison and contrast based on individual characteristics and on the relative value of those characteristics to individual organizations. In addition, both indices aid in explaining anomalies because they help to identify the conflicting characteristics that cause apparently similar organizations to be opponents. However, the indices cannot identify evangelical organizations per se because there is no way of deriving which characteristics are essential to being evangelical. There is no quick way to determine if there is any difference between similar scores. Therefore, the indices should be used to compare significant differences and similarities rather than shades of gray. The final indices clearly show that a large number of characteristics typically considered evangelical are shared with a large number of churches.40 Thus, small differences in the score of different organizations may not be significant.41 Furthermore, classification should take into account that the lower the score the greater the chance that differences matter and that organizations with similar scores are different in character. For example, when using the additive index to classify organizations an organization might have a low score on the doctrine section whereas a different organization might have a low score on the association section but they may both have the same total score on the index. Thus, the best way to identify and describe the evangelicalness of a church is to examine its score on the index and rate its evangelicalness relative to other churches.42 ____________________________________ 39 The 76 items on the survey were divided into five sections, which addressed doctrine, supporting institutions, associating with groups, supporting particular understandings of the world, and valuing specific practices. See Appendix G for a list of the items in each section. 40 See discussion of necessary characteristics above. 41 Although there are significant differences between churches in terms of their relative evangelicalness, the differences between individual churches in similar locations near the top of the index is often negligible. The average point difference between the top 58 churches in the additive index is less than 2 points and all 32 churches ranked 1 in the factor index shared the same factor score. 42 Another alternative would be to compare scores on the five different sections of the index. [ Page ] 37 2.2.4.2 Comparing the Indices The two polythetic indices make sense when compared with other means of classifying evangelical organizations. For example, churches universally acknowledged to be evangelical (e.g., Pentecostal Churches, Evangelical Missionary churches) generally ranked near the top of the scale and churches generally considered not evangelical (e.g., Evangelical Lutherans, United churches, etc.) generally ranked near the bottom of the scale (see Table 2.14). In addition, churches with more ambiguous denominational affiliations generally were spread throughout the index (e.g., Mennonites). Both indices provide a broad ranking that is useful for understanding differences between more and less evangelical churches and for exploring how local churches interact. To varying degrees, both indices acknowledge the diverse nature of the characteristics of religious groups but also take into account that some characteristics may be necessary to group an organization within a class. However, the two indices are useful for different purposes. The additive index allows self-definition to be a factor in identifying groups but does not rely exclusively on it. It also takes into account ideals and actions by including characteristics that emphasize both. In addition, the additive index embraces a broader range of characteristics than the factor score index. Nonetheless, the large number of characteristics in the additive index reduces its ease of use when analyzing why a certain church is classified in a certain manner. For this reason, the factor index has value as a more statistically rigorous and simpler approach to the classification of churches. However, it does not fully take into account the diversity of evangelical organizations and thus sometimes classifies churches in unusual ways. Given the equivalent usefulness of both indices, their high correlation, and different benefits, I have chosen to use the additive index for my analysis because it encompasses a broader range of characteristics, has a wider distribution in its range of scores, and because the additive score is more easily associated with scores on individual characteristics. Table 2.14 - Rank/Denomination Comparison [ Please contact repository@tyndale.ca for Table 2.14 details ] Org# Additive Rank Factor Rank Listed Denomination 1004 1 1 Baptist 1071 1 3 Baptist Convention Of Ontario And Quebec 1053 1 1 Brethren Assemblies 1060 1 1 Brethren In Christ 1087 1 1 Brethren in Christ Org# Additive Rank Factor Rank Listed Denomination 1048 1 1 Christian Reformed Church in North America 1047 1 1 Church of God in Cleveland, Tennessee 1006 1 1 Evangelical Missionary 1013 1 1 Fellowship Baptist [ Page ] 38 Org# Additive Rank Factor Rank Listed Denomination 1010 1 2 Fellowship Deaconry Of Canada (Evangelical Lutheran Background) 1029 1 1 Fellowship of Evangelical Baptist Churches in Canada 1081 1 1 Full Gospel 1038 1 1 Independent Baptist (no denomination) 1040 1 2 Mennonite 1041 1 2 North American Baptist Conference 1036 1 1 Pentecostal 1061 1 1 Pentecostal Assemblies Of Canada 1083 1 1 Pentecostal Assemblies Of Canada 1054 1 2 The Presbyterian Church In Canada 1067 1 1 Unaffiliated 1072 1 1 United Brethren Church In Canada 1052 1 1 Zion Fellowship Inc. (New York, U.S.A.) 1042 2 2 Association of Vineyard Churches Canada 1026 2 1 Baptist Convention of Ontario and Quebec 1045 2 2 Christian (Plymouth) Brethren 1064 2 1 Church of God (Anderson, Indiana) 1019 2 3 Evangelical Missionary Church 1014 2 1 Evangelical Missionary Church 1046 2 1 Evangelical Missionary Church of Canada 1080 2 1 Evangelical Missionary Church of Canada 1021 2 3 Fellowship Baptist of Evangelical churches in Canada 1050 2 2 Independent Lutheran 1005 2 3 Lutheran Church - Canada 1032 2 3 Lutheran Church - Canada 1068 2 1 Lutheran Church Canada 1051 2 2 Lutheran church-Canada 1001 2 3 Mennonite 1028 2 3 Mennonite 1069 2 2 Mennonite Conference Of Eastern Canada 1057 2 2 None independent Baptist not a denomination 1024 2 3 Not affiliated 1039 2 2 Pentecostal Assemblies of Canada 1070 2 1 Salvation Army 1002 3 1 Anglican 1079 3 3 Anglican Church of Canada 1030 3 4 Baptist Convention of Ontario and Quebec 1044 3 1 Christian & Missionary Alliance Org# Additive Rank Factor Rank Listed Denomination 1084 3 3 Christian Reformed 1085 3 2 Church of Christ In Canada & worldwide-non-denominational 1049 3 3 Congregational Christian 1034 3 1 Congregational Christian Churches in Canada 1075 3 3 Evangelical Lutheran Church In Canada 1076 3 4 Evangelical Lutheran Church In Canada 1077 3 3 Evangelical Lutheran Church In Canada 1062 3 1 Independent Lutheran 1020 3 1 Lutheran Church - Canada 1022 3 4 Mennonite 1025 3 3 Mennonite 1027 3 3 Mennonite Brethren 1086 3 3 Mennonite Church Canada 1055 3 3 Non-Denominational 1012 3 1 Pentecostal 1037 3 3 Pentecostal Assemblies of Canada 1056 3 1 Pentecostal Assemblies Of Canada 1066 3 3 Seventh Day Adventist 1016 3 4 United Church of Canada 1031 3 4 United church of Canada 1033 4 4 Anglican 1074 4 4 Anglican 1003 4 1 Apostolic Faith Church 1011 4 2 Canadian Mission board of the German Church of God of the Dominion of Canada 1009 4 4 Evangelical Lutheran Church in Canada 1018 4 3 Evangelical Lutheran Church In Canada 1035 4 4 Lutheran Church - Missouri Synod 1017 4 4 Mennonite 1043 4 4 Mennonite 1058 4 4 Mennonite 1063 4 4 Mennonite 1078 4 4 Mennonite 1073 4 4 Presbyterian 1082 4 4 Presbyterian 1015 4 4 Presbyterian Church in Canada 1065 4 3 Religious Society of Friends 1007 4 4 United Church of Canada 1008 4 4 United Church of Canada 1023 4 4 United Church of Canada 1059 4 4 United Church of Canada 1088 4 4 United Church Of Canada [ Page ] 39 2.3 Summary A polythetic definition can be an important tool in the study of the multi-faceted nature of evangelical organizations. The indices developed here are broad enough to encompass variations in theological opinion and in relations with other organizations but can be applied narrowly to differentiate groups from each other without excluding fringe groups and without relying on sufficient or dominant characteristics. A polythetic approach focuses on characteristics that can be applied to organizations, it includes a diversity of characteristics to take into account the nuances of organizations, and it measures degrees of acceptance of characteristics in response to the collective nature of organizations. Despite the utility of the Indices of Evangelical Characteristics for quantitative identification, classification and comparison of evangelicals, the indices still must be used carefully. It is not sufficient to look at the final score of different groups and to conclude, on that basis alone, that groups have irreconcilable differences. Furthermore, as the social network data show (see chapter 6), a different score on the Indices does not preclude working together. The qualitative data provided by the index is also important as a closer look at individual responses can show that there are many areas where two organizations could conceivably co-operate or that there is a plurality of different groups within a single organization.3243 Nonetheless, taking into account some important cautions and the definitions’ complexity, the Indices of Evangelical Characteristics prove to be productive for defining evangelicals for the purpose of quantitative identification, classification, and comparison of Christian organizations. Undoubtedly a polythetic definition is more complicated than standard definitions, given the large amount of material that has to be dealt with, but it carries “less risk of an arbitrary exclusion of significant features” (Needham 1975: 358). “Evangelical” as a descriptive characteristic is ambiguous and of limited use for quantitative identification, classification and comparison of organizations (Dayton 1991) but “evangelical” as a class of organizations sharing a limited number of necessary characteristics and a large number of general characteristics remains useful. What identifies ________________________________________ 32 Factors that influence the activities a group is involved in and its ability to co-operate with others include: characteristics related to culture, geography, and historical conflict that cannot be adequately dealt with in indices such as these. [ Page ] 40 organizations is not any one unique characteristic but rather a particular combination of non- unique characteristics. Furthermore, this combination of characteristics may or may not be unique. The value of this scheme of classification for my study is that it allows me to determine which local organizations are more evangelical than others. In the next chapter I explain some of the perceptions of network connectivity among evangelicals, provide some of the context of evangelical connectivity, point to suggestions of evangelical connectivity in the historical literature, and review some of the ways that evangelical organizations interact. 3 CONNECTION: EVANGELICALS AND THE WORLD This chapter provides an historical context for the state of evangelical organizational connectivity in contemporary Canada.44 It begins with a discussion of some of the limits of current perceptions of evangelical interaction. This first section provides some of the justification for a network study of Canadian evangelicals.45 The second section explores Canadian evangelicalism’s heritage. More specifically, it examines how networking among religious organizations outside of Canada provided a model for and influenced networking among religious organizations within Canada. This history provides a larger context for evangelical networking in one geographic area. A third section examines three different levels of networking among evangelicals. This section and the second section provide the background that framed my expectations of the interaction I would find among evangelicals and their community at a local level. The final section of this chapter describes the demographics of the two cities which served as the location of my research. Studying networks among all the evangelical churches in a single community builds on existing literature which describes the growth and maintenance of evangelical organizations in contemporary Canada. Furthermore, it produces some expectations regarding the type of network relationships I can expect to find in the data. Thus, this discussion illustrates the importance of understanding evangelicalism’s network context, points to evidence which suggests that the customary understanding of evangelical churches _____________________________________ 44 This section is not intended to be a history of networks among religious organizations in Canada or even of evangelicalism in Canada. There are several books which provide a good starting point for collecting more detail about the history of Canadian evangelicalism (Rawlyk 1990; Noll 1992; Stackhouse 1993; Noll et al. 1994; Rawlyk and Noll 1994; Burkinshaw 1995; Quast 1996; Rawlyk 1996; Rawlyk 1997). 45 While evangelicalism and fundamentalism are not the same thing, scholars disagree over how to draw the boundary between the two groups. In addition, the boundary between the two groups has changed over time and remains permeable. Thus, in discussions of conservative Protestantism the same organization can sometimes be classified in both groups. Furthermore, in the small world of Canadian conservative Protestantism, groups from both categories often interact. Thus, for the purposes of this chapter I have not distinguished between evangelicalism and fundamentalism. [ Page ] 41 [ Page ] 42 as isolated from their communities may be inaccurate, and notes where a more specific study of network interaction would improve our understanding of Canadian evangelicalism.46 3.1 Perceptions of Evangelical Interactions In recent years historians have filled in many of the gaps in our understanding of Canadian evangelicalism (Noll 1991; Stackhouse 1995; Burkinshaw 1996; Rawlyk 1997). However, historians have not yet published studies which deal specifically with network relationships involving all the evangelical churches in one geographic location. Furthermore, sociologists of religion have paid little attention to Canadian evangelicalism and thus we have few quantitative studies of actual connections between Canadian evangelicals. As a result, most of the information we have about such connections is historical and is centred around prominent individuals or organizations. Thus, existing studies of Canadian evangelicalism provide some vital information for studying evangelical organizations, but leave room for a more specific sociological study of these same organizations. Given the lack of quantitative data on connections among evangelical organizations, it is not surprising that there are some misconceptions about how evangelicals interact with each other. This section examines some of the perceptions local community members and scholars have of how conservative organizations generally, and evangelicals specifically, interact with their environment. It also briefly describes some of the inaccuracies of those impressions, what recent scholarship has done to correct mistaken impressions, and what still needs to be done. 3.1.1 Interviews with Community Members I interviewed community leaders in one southwestern Ontario community as an indicator of where to look for networks among local religious organizations and as a point of comparison with network data. I selected community members who were likely to have some sense of the interaction taking place in the community. The interviews were designed to ____________________________________ 46 My use of the term “network” is based on a sociological discipline known as Social Network Analysis. A basic definition of network comes from Knoke and Kuklinski (1982:12) who define it as “a specific type of relation linking a defined set of persons, objects or events.” Wasserman and Faust (1994:20) provide a more complex definition of a social network as “a finite set or sets of actors and the relation or relations defined on them.” They define actors as “discrete individual, corporate, or collective social units.” (1994:17) and relations as “the collection of ties of a specific kind among members of a group” (1994:20). Keeping in mind these definitions, I am using the term network to refer to the connections between three or more individuals or organizations (e.g., shared board membership, shared activities, shared information, etc.). For an introduction to social network analysis see Wellman (1988). [ Page ] 43 collect data about church and community interaction. That is, I was looking for formal or informal interaction between local religious organizations and between local religious organizations and the community. My informants did see evidence of network connectivity but few of them had actually witnessed these connections. Nonetheless, it was possible to discern some patterns in my informants’ understanding of local networks. Networks were understood to be theologically similar and informal. Second, the community was assumed to be Christian and concerned with social issues. Finally, my informants identified social service organizations as the most visible in the community. This section explores these perceptions in more detail. My informants identified some distinct groups of densely connected religious organizations in Kitchener-Waterloo. As I expected, my informants believed that organizations interacted according to type. Thus, mainline organizations were expected to group together, as were Catholic organizations, and, to a lesser degree, evangelical organizations. I asked specifically about interaction among evangelicals. Many of my informants assumed there was some sort of relationship between local evangelical churches, because of successful mobilization of people for special events, and evidence of regular funding of several local ministries (e.g., Oasis drop-in centre, Birthright pregnancy counselling, etc.). There were also other signs of evangelical involvement in the community (e.g., large evangelical churches, local building projects). However, few informants had any contact with evangelicals or could point to any details of this assumed organizational structure. At times evangelicals were described as sectarian (C226), less involved in discussions of faith and politics (C214), isolated (C219), not obvious (C229), and involved in the community in areas which match their moral values (C208, C234).47 Evangelicals, my informants explained, tend to connect with each other at ministers’ meetings (i.e., ministerials) in a formal manner but most connections are centred around issues (abortion, Sunday shopping, etc.) or events (March for Jesus, evangelistic crusades, Promise Keepers conferences, etc.). Permanent organizations run by evangelicals also reflected these concerns (a pregnancy crisis centre, a drop in centre for teenagers, etc.). _____________________________________ 47 Specific interviews are referenced with the letter C followed by three numbers indicating which interview I am referring to. Where a statement could potentially lead to the identification of an informant the interview number is not listed. [ Page ] 44 Thus, my informants had the impression that evangelicals were involved in the community but in a sectarian manner focused on their own agendas. Some of my informants assumed that the mainline organizations would have the largest number of organizational connections within the community whereas the evangelical organizations would have almost none (C226, C233). Those that evangelicals did have were considered to be related to their specific concerns (e.g., conservative moral issues) (C208). However, both groups were assumed to be involved in the community as individuals (C234, C238). Furthermore, there was only sparse evidence of any regular connection between mainline organizations based on specific moral or spiritual issues (body politics, evangelism, etc.). Several informants noted that it is hard to co-operate in such areas because of the diversity of viewpoints within the various denominations (C205, C233). According to the impressions of my informants, networks among religious organizations in Kitchener-Waterloo are primarily informal based on personal rather than organizational connections. My informants were sometimes aware of little involvement by any churches as churches in the community: I do not see Christians volunteering themselves at community events or at least really involving themselves in a large identifiable way. There are, no doubt, Christians individually that are offering themselves but [not] as a church to come behind and say we would like to do a lot of stuff (C219) Several informants explained that evangelicals tend to be more connected to other organizations in the community as individuals than as members of churches (C234, C238). This may explain why there was so little awareness of evangelical networks and evangelical involvement in the community.48 One example of this lack of involvement in the community is the absence of formal connections between religious and political groups. Some local mainline churches participated in demonstrations against the right-wing policies of the Progressive Conservative government of Ontario in the 1990s. However, the politicians I talked with and the offices they represented had no formal or regular connections with religious groups of any kind and very rarely heard anything from religious groups, positive or negative.49 On occasion, local ____________________________________ 48 The exception to this is the Mennonite Brethren churches which are also connected to the work of the Mennonite Central Committee. 49 This finding is, in part, a result of the focus of my questioning which was very community-centred. I may have biased the answers against descriptions of social justice initiatives which moved beyond the communities’ borders. [ Page ] 45 politicians referred positively to religious organizations that were no longer in existence or could not remember the names of local religious organizations. This lack of formal connection among organizations may come from the sense, on the part of local Christians, that Kitchener-Waterloo is a moral, even Christian community. One local religious leader explained that Kitchener-Waterloo was a fairly religious community (C218), a different informant described it as a very religious and very conservative community (C233), a political leader described the community as very Christian (C214), and a fourth informant described Kitchener-Waterloo as a community with Christian values (C222). Yet another informant argued that there was a general sense that the community was Christian-based and therefore there was little need to worry about local political culture (C235). Thus, community leaders are assumed to be functioning based on Christian principles, and therefore church leaders may feel safe to focus on internal church issues. Nonetheless, at least one informant was convinced that there had been a marked decline in Christian commitment in Kitchener-Waterloo (C232). It is also the case that some churches are not interested in political interaction. In addition to having some knowledge of who associates with whom in the community, my informants provided me with data about which organizations were the most visible in the community. Repetition of organizations by informants identified some organizations as more visible in the community than others (see Table 3.1). However, these data do not indicate what being visible says about an organization’s role in the community. It appeared that the organizations that were mentioned most often were mentioned because their contribution to the community is considered very valuable or is more widely known than that of other organizations rather than because of their influence or activities in the community. This can be illustrated by the regular mention of St. John's soup kitchen and the relative lack of awareness of the Working Centre which runs the kitchen and a number of other social service initiatives. Seven informants mentioned the soup kitchen, whereas only three informants mentioned the Working Centre. The working centre is more involved in the community, but the soup kitchen is more obvious. [ Page ] 46 Table 3.1 - Organizations Mentioned by at Least Four Informants [ Please contact repository@tyndale.ca for Table 3.1 details ] Number of Mentions Organization 7 St. John’s Soup Kitchen 6 Mennonite Central Committee 6 Kitchener-Waterloo Council of Churches 5 March for Jesus 4 House of Friendship 4 Evangelical Ministerial 4 Benton Street Baptist Church Sixteen interviews in two small cities in Ontario are not sufficient to come to broad conclusions about Canadian evangelicals but, in conjunction with national demographic and historical data, they suggest that the division of religious organizations into discreet groups is a local as well as a national phenomenon. The interview data also suggest a high degree of individualism within the evangelical community in Kitchener-Waterloo. Perceptions of isolation and individualism are not limited to these interviewees. This observation has been made by other commentators on evangelicalism. Evangelical theologian David Wells has complained that “Evangelical faith is perceived as a matter of internal fascination but abandoned as a matter of external and public relevance” (quoted in Rawlyk 1996:222). Evangelicals in Kitchener-Waterloo would appear, at first glance, to prove his point. Another example is journalist Ron Graham’s idiosyncratic exploration of Canadian religion in God’s Dominion. He describes a specific group of evangelicals in this way: “Emotions are more important than intellect, experience is more important than understanding, and there is little place for nuance or ambiguity.” (Graham 1990:313). He goes on to claim that Protestantism, in general, tends to schism: “The underlying dynamic of Protestantism in the face of feuds is to get the hell out” (Graham 1990:315). Furthermore, he traces a chain of reaction to cultural events that leads to a situation where evangelicals see their faith as “a personal and private transformation by Christ; and while it did not preclude acts of Christian charity or the sending forth of missions, it was rooted in individual sin and individual salvation.” (Graham 1990:324). He finds it surprising that “so many people, having been touched by Christ and the Holy Spirit, immediately want tougher immigration laws . . .” (Graham 1990:327-28). Ron Graham is a journalist and bases his perceptions on anecdotal evidence. Nonetheless, his book was well received by the Canadian media and reflects how some of the [ Page ] 47 Canadian population perceives Canadian evangelicals (see review excerpts from Ron Graham’s website ). The general impression of Canadian evangelicals is that they have little impact beyond their boundaries because they have little contact with any community organizations beyond their boundaries.50 To provide some balance to these subjective opinions of individuals, it is important to look at how academics describe Canadian evangelicals and their interaction with their environment. 3.1.2 Historical Studies Few books attempt to give a broad description of all of Canadian evangelicalism and how Canadian evangelicals interact with each other and their community settings. Furthermore, there is scholarship which suggests that evangelicals do not connect well with other organizations. Nonetheless, Canadian evangelicals are thriving, and plenty of recent historical scholarship details their interaction with each other and the larger environment. This section describes some of these historical studies and their limits for the study of local networks involving Canadian evangelicals. Until recently there was little commentary on Canadian evangelicals in the historical literature. Books on the history of Canada sometimes pay little attention to evangelicals or refer to them in a derogatory manner. Historian John Stackhouse (1997:447n31) explains that “In The Forked Road: Canada 1939-1957 . . . Donald Creighton pays attention to religion, but his discussion of evangelicalism is fraught with words like ‘narrow,’ ‘puritanical,’ and ‘repressive.’” Another example is Grant’s (1963) volume The Churches and the Canadian Experience, which does not refer to evangelicals by name and only deals with them tangentially in chapters on specific groups (e.g., Baptists, Sects). This reflects the relatively small numbers of evangelicals in mid-twentieth century Canada and the fact that evangelicals had not yet coalesced into an identifiable grouping distinct from the established churches.51 Nonetheless, historians were often aware of evangelicals’ existence and referred to them in passing while discussing other aspects of Canadian religion. For example, John Webster Grant (1998, first published in 1972) in his well-known discussion of Canadian churches, The Church in the Canadian Era, refers to evangelicals to note, among other ____________________________________ 50 They do have relationships with businesses to supply their physical needs, but an examination of these relationships is beyond the scope of this dissertation. 51 There continues to be active groups of evangelicals in the mainline denominations but interest in their concerns appears to be declining and they are not as visible as evangelicals in the predominantly evangelical denominations (Stackhouse 1994:382ff). [ Page ] 48 things, their efforts to provide values for the nation and their interest in higher education. However, for Grant (1998:199), evangelicals were not as affected by the societal changes of the 1960s as other Canadian churches because they saw the changes as “further proof of the long-suspected apostasy of the conventional Protestant churches.” Stackhouse (1993:207) explains that “for all of Grant’s exemplary even-handedness, one still gains little idea of the shape and significance of evangelicalism in Canada from this book.” Stackhouse goes on to explain that other recent books have also missed important aspects of Canadian evangelicalism (Handy 1976; Mol 1985; Bibby 1987). Evangelicalism was the dominant form of Protestant Christianity in Canada in the nineteenth century (Gauvreau 1991). However, by the 1940s primarily evangelical denominations represented only 7% of the population (Bibby 1987:28). Furthermore, the literature about North American evangelicals does provide evidence that evangelicals can sometimes be insular, in opposition to the larger culture, and disconnected from the larger religious community. Marsden (1984:xi) explains that: “twentieth century evangelicals all have in common a belief that faithfulness to Scripture demands resistance to many prevalent intellectual and religious currents.” Thus, American evangelicals have sometimes been considered unsophisticated and gullible (Berg 2000). Furthermore, some types of evangelicals (e.g., dispensationalists) “made a virtue out of working independent of the denominations” (Marsden 1984:xiii). In Canada, evangelicals have been described as “more committed to personal salvation than to the establishment of social justice or national righteousness” (Clifford 1977 commenting on the continued use of “His Dominion” language by evangelicals in the 1970s). More recently, Bibby and Brinkerhoff’s (1983; 1994) “circulation of the saints” thesis demonstrated that, in Calgary, only 10% of the additions to evangelical churches came from outside of the evangelical community. That is, evangelicals have not had significant success attracting non-evangelicals into their churches. Hexham (1993) argues that further research is needed before we accept this conclusion. He argues that Bibby and Brinkerhoff’s data may be flawed because it relies on church leaders for information, not church members. Hexham also argues that their data does not examine para-church groups which make converts and then pass them on to evangelical churches. Nonetheless, even if Bibby and Brinkerhoff’s data is inaccurate in regards to actual conversions it does highlight the lack of awareness of evangelicals, even within evangelical churches. [ Page ] 49 Several years after Bibby and Brinkerhoff, Stackhouse (1993) argued that the success of Canadian evangelicals would have been news to many Canadians because “there is no adequate characterization of Canadian evangelicalism in the twentieth century” (1993:6). Nor is there an adequate “estimate of its influence in Canadian life” (1993:6). Nonetheless, Stackhouse discerns a “trend among evangelicals towards greater self-consciousness and cooperation while simultaneously distancing themselves from the direction and institutions of the mainline churches” (1993:165). Stackhouse (1993:185) explains that this tendency to work with each other rather than with other types of Christians was extenuated by Canadian evangelicals’ increasing alienation from “the leadership and institutions of all three of the mainline Protestant groups” (emphasis in the original). Thus, both because of the Canadian evangelical tendency to stick together and because of the general lack of awareness of their activities, evangelicals gained a reputation for being insular and disconnected from their cultural context. As a result, scholars have sometimes been surprised to find that Canadian evangelicals are maintaining their share of the Canadian population (Rawlyk 1997:xviii). In part, scholars are surprised by thriving Canadian evangelicalism because Canadian evangelicals lack the numerical and political strength of American evangelicals. Nonetheless, evangelicals have remained a relatively stable percentage of the Canadian population since at least the 1940s (Bibby 1987:28). Furthermore, Canadian evangelicals are growing numerically, whereas other Christian groups are declining in number of adherents (Bibby and Brinkerhoff 1994). O'Toole (1996) notes growth rates as high as 75% for exclusively evangelical denominations and current estimates of the evangelical share of the Canadian population range from 7% to 16% depending on how the concept is operationalized (Bibby 1993:45; Rawlyk 1997:xviii). Yet, Stackhouse (1997:62) noted that, as late as 1997, surveys of Canadian religion have often either ignored evangelicalism, considered it confined to the private sphere or popular culture, or have thought that it was losing ground. Fortunately, within the last decade, studies of Canadian evangelicals have begun to be published in significant numbers. Scholars have sought to address the need for studies which specifically focus on Canadian evangelicalism (Rawlyk 1990; Gauvreau 1991; Noll 1991; Airhart 1992; Stackhouse 1993; Rawlyk 1994; Burkinshaw 1995; Rawlyk 1996; Rawlyk 1997). Many of these works compare Canadian evangelicals with evangelicals in other countries and compare different evangelical organizations together (Stackhouse 1994; Noll [ Page ] 50 1997; Rawlyk 1997). There are also academic treatments of Canadian Christianity and Canadian evangelicalism as a whole (Rawlyk 1990; Noll 1992; Stackhouse 1993; Rawlyk and Noll 1994; Murphy and Perin 1996) as well as studies which are more popular than scholarly (Rennie 1986; Jantz 1994; Mackey 1995). Furthermore, there is evidence that Canadian evangelicals are reconsidering public engagement (Stackhouse 1993:199). More recently, the Canadian Evangelical Experience Conference attracted dozens of scholars and provided an introductory description of the history of several aspects of Canadian evangelicalism (Stackhouse 1995; Rawlyk 1997). Several of the essays discussed the links between evangelical organizations (see Rawlyk 1997 Part Five). Unfortunately, little has been published on any aspect of Canadian evangelicalism, since the publication of Aspects of the Canadian Evangelical Experience in 1997.52 Furthermore, despite significant exceptions, much of this research concentrates primarily on individual people or specific organizations rather than on the connections between them.53 In part this is because networks among Canadian evangelicals remain largely informal (Stackhouse 1994:386). Stackhouse (1994:286) explains that: Canadian evangelicals continued to disagree about political strategies (e.g., all-or-nothing vs. piecemeal reform in the abortion struggle), about mission priorities (e.g., evangelistic projects vs. academic consultations), about ecumenical relationships (e.g., links between charismatic Protestants and Roman Catholics), and about issues facing the whole church (e.g., the public ministry of women). Such disagreements sometimes make differences more obvious than similarities. However, concluding that Canadian evangelicals do not interact would be a mistake. Recent scholarship has shown that North America evangelicals do not always live up to their stereotypes (Hexham 1993; Stackhouse 1993; Smith 2000). Evangelicals have an extensive history of interacting with each other around the world (O’Brien 1994; Piggin 1994; Stout 1994) and in Canada (Elliot 1994; Rennie 1994; Burkinshaw 1995; Hindmarsh 1997). Furthermore, there is evidence of evangelical involvement in all manner of Christian and social justice issues of the past; from Yonge Street Mission to the YMCA/WCA and from ____________________________________ 52 However, essays relevant to Canadian evangelicalism have been published in some recent collections (e.g. Lyon and Van Die (2000) and Van Die (2001)). 53 Burkinshaw (1996:14-15) describes of a couple of exceptions: Bruce Hindmarh’s study of a network focused in Winnipeg and Alvyn Austin’s examination of a Toronto based network (Austin 1997; Hindmarsh 1997). Neither of these studies is an examination of all the Protestant church interactions in a specific geographical area. Burkinshaw (1995) himself has written an historical study which focuses on the relationship between evangelicals across a large geographic area. [ Page ] 51 China Inland Mission to the Hospital for Sick Children (Rennie 1984:6).54 Furthermore, there is evidence of a partial return to that social responsibility in the 1980s (Stackhouse 1993:128) As historians have looked for patterns to give order to the events of Canadian history they have discovered the intense interconnectedness of Canadian evangelical individuals and organizations both with each other and with the community at large (Burkinshaw 1996:14- 15; Stackhouse 1993; Chapman 1994). Evangelical individuals are heavily involved in the local community in several selected areas (primarily evangelism and conservative moral issues: abortion, euthanasia, “family values”) (Motz 1990; Stackhouse 1993). Furthermore, concern for these issues is shared with evangelicals across the country (Stackhouse 1993; Burkinshaw 1995; Rawlyk 1997). For example, the Evangelical Fellowship of Canada’s mission statement explains that they aim to Be a public advocate of the gospel of Jesus Christ; Provide an evangelical identity which unites Canadian Christians of diverse backgrounds; Express biblical views on the critical issues of our time; Assist individuals and groups in proclaiming the gospel and advancing Christian values in our nation and around the world. (from the website of the Evangelical Fellowship of Canada, 2001) Thus, while it took a little while for evangelicals to be noticed, historians have accumulated a significant amount of evidence that evangelicals do interact with each other and do play a role in their communities. 3.1.3 Surveys and Sociological Studies of Evangelicals Some of the studies I have already referred to are more sociological than historical (e.g., Mol 1985). Sociological studies explicitly examine the larger context in which Canadian evangelicals operate.55 For example, surveys provide demographic and opinion data about religion in Canada. Such sociological studies provide information about what Canadian evangelicals look like, what they think they are doing, and what they are trying to accomplish. However, sociological studies have been less successful than historical studies at capturing the character of Canadian evangelicalism. _____________________________________ 54 In another article Rennie (1986) argues that “the pattern for evangelical social action” in nineteenth century Canada was influenced by William Wilberforce and the Clapham Sect in England. 55 The classic sociological study of Canadian religion is Clark (1948). Clark’s main concern was to trace the relationship between church and sect in Canada over time. [ Page ] 52 The most well known collector of statistics about religion in Canada is Reginald Bibby.56 His research can serve as an example of this genre. Bibby has produced several works which describe the general religious beliefs and practices of all Canadians (Bibby 1987; Bibby 1993; Bibby 2002). Bibby has also generated specific studies of Canadian denominations such as the United Church (Bibby 1994), the Anglican Church (Bibby 1986), and the evangelical Christian and Missionary Alliance church (Bibby 1999). He has also examined evangelical attitudes on evangelism for the Evangelical Fellowship of Canada (Bibby 1995). His studies aid in identifying the characteristics of Canadian evangelical organizations, determining which connections may be important when studying relationships among and between these organizations, and extending our understanding of their goals and purposes. They also provide some indication of the extent to which Canadian evangelicals actually act on those goals and purposes.57 However, Bibby’s studies, and other similar surveys, were not designed to describe evangelical organizations nor are they specifically concerned with connectivity. Thus, while surveys provide a broad picture of Canadian evangelicalism and of religion in Canada, they do not always provide the specific details necessary to classify organizations or to observe and understand their interactions. Attention to Canadian evangelicals in other types of sociological studies is almost nonexistent. A 1977 volume on religion and culture in Canada only mentions them as a group once and then only to highlight their adherence to a discredited understanding of Canadian society (Clifford 1977; Slater 1977). Evangelicals do show up in examinations of Canadian demographics, but only as denominations, and not as a recognized grouping (Kalbach and McVey 1976). Canadian evangelicals do get mentioned in sociological discussions of religion in Canada (O'Toole 1996; Noll 1998), and there are also sociological studies which explore one specific aspect of Canadian evangelicalism (Hiller 1976; Bibby and Brinkerhoff 1983; Bibby and Brinkerhoff 1994; Reimer 1995). However, there is currently no book- length sociological study of Canadian evangelicalism. Furthermore, the author of the only recent sociological exploration of Canadian evangelicals feels compelled to spend time ____________________________________ 56 Other well known sources of survey data about Canadian evangelicals include God and Society in North America: A Survey of Religion, Politics and Social Involvement in Canada and the United States (Angus Reid Group 1996) and the various publications of Census Canada (e.g., Statistics Canada, Religions in Canada 1993; Statistics Canada 2003). 57 For example, Bibby has discovered that even though 78% of evangelical leaders surveyed at a series of “Consultations on Evangelism” are attempting evangelism, only 49% actually had a plan and a budget in placeto do something about it (Bibby 1995:18). [ Page ] 53 refuting negative stereotypes and claims the Canadian academic environment is not conducive to studying evangelicals (Hexham 1993). Hexham quotes Judith Haiven describing evangelicals as “myopic, intolerant, rightwing, fascist, even anti-Semitic . . .” (Haiven [1984:22] in Hexham 1993:290). Nonetheless, Hexham goes on to explore the many ways Canadian evangelicals are concerned with social issues and are involved in social programs. Thus, sociological studies can fill in some of the details of how Canadian evangelicals operate but, provide few details about how Canadian evangelicals interact with each other and with their community in one specific geographical location. 3.1.4 What Remains to be Done Evangelicals are assumed to be isolated in some scholarly and popular perceptions, both in terms of community involvement and in terms of networking with other organizations. Nonetheless, the historical literature clearly indicates that evangelicals are, and have been, highly networked. This situation demonstrates the limits of the subjective impressions of community members and sociological understanding of evangelical interactions which see evangelicals as completely distinct from other Canadian religious groups. Thus, current studies of Canadian evangelicals could be complemented by a study which examines interactions among Canadian evangelicals in one community and compares them with the community interaction patterns of other Protestant organizations. The next two sections of this chapter review some of the influences on Canadian evangelicalism and looks at some specific examples of networks among Canadian evangelicals. These sections have three objectives. First, they provide a context for my results. Second, they illustrate some of the limits of existing studies identified in this section. Third, they highlight the need for a network analysis of Canadian evangelicals. 3.2 Networks in North American Protestantism From the beginnings of contemporary Canadian evangelicalism in eighteenth century revivalism, networks played an important role supporting and connecting its adherents in diverse locations. My interest in networks among Canadian evangelical organizations was prompted by studies of the history of Canadian evangelicalism. An examination of that history establishes why a study of networks is important for understanding Canadian evangelicals. It is important because evangelical organizations are not closed to their [ Page ] 54 environment. Rather they are heavily dependent on it. The environment of early Canadian evangelicals includes more than their geographic setting; it also includes the myriad of networks which support their worldviews, provide resources for their operations and give legitimacy to their actions. This section explores some of the historical influences on Canadian evangelicalism and briefly discusses the development of Canadian evangelicalism itself. My objective is not to provide a complete history of the development of Canadian evangelicalism, but to provide some of the context in which Canadian Protestant organizations interact. More specifically, this description provides the setting in which evangelical organizations in a single community interact. The history of interactions among Protestant Christians contributes to my assumption that there will be networks among contemporary churches. 3.2.1 American Influence Despite Canada’s continued connection to Britain (Westfall 1989; Christie 1990; Noll 1992; Murphy and Perin 1996), the development of Canadian evangelicalism was also influenced by changes taking place in evangelicalism in the United States. This section describes the development of American evangelicalism. I discuss the inception of an American Christian fundamentalist coalition, how that coalition broke apart, and how new coalitions had only limited success. I also discuss how Canadian evangelicals differ from American evangelicals. This section helps to explain the development of evangelicalism’s reputation for insularity since Canadian evangelicalism has often been understood as following the patterns of American evangelicalism. The prevalence of networks among American evangelicals illustrates why a network study of Canadian religion would be fruitful. In the early 1900’s in the United States, half a century after Canadian confederation, Princeton Seminary professors Charles Hodge and B.B. Warfield provided the name and some of the theology for a movement which came to be known as “fundamentalism” (Wells and Woodbridge 1977:11). However, cultural and theological change led to this broad-based but loosely knit coalition falling apart (Noll 1992:381ff). White Protestantism divided into two groupings that can roughly be called the liberals (or modernists) and the conservatives [ Page ] 55 (or fundamentalists).58 The liberals, for the most part, received the support of the church leadership (Marsden et al. 1984:xiii). There was an expectation that the conservatives would fade away, as this split in Protestantism reduced their influence (Marsden 1991:63). Furthermore, additional fissures occurred among the conservatives as the doctrine of separation led to some groups’ separating not only from the modernists but also from holiness and Pentecostal groups (Marsden 1991:4). Furthermore, by the 1960’s, fundamentalist came to mean separatist “and no longer included the many conservatives in mainline denominations” (Marsden 1991:3). Further conservative withdrawal from society resulted from the perceived threats of evolution, Marxism, and theological liberalism (Wells and Woodbridge 1977:12). Carpenter (1984:4) explains that Fundamentalists did recoil from the antimodernist and antievolution controversies in defeat and disarray. Some of the coalitions they had formed splintered into contending factions, and many fundamentalist leaders despaired of ever again mounting a public campaign before the return of Jesus. In spite of failures and conflicts, conservative Christianity did not disappear, and 1941 witnessed some small steps towards conservatives defending their interests in the public square. In that year, the fundamentalistic American Council of Christian Churches (ACCC) was formed (Wells and Woodbridge 1977:13). Yet, within conservative Christianity, reservations were developing about fundamentalism’s increasing reactionary, negative, and separatist approach to theology and social interaction (Marty 1977; Marsden 1991). A year after the formation of the ACCC, the National Association of Evangelicals (NAE) was formed by a group of conservatives uncomfortable with the direction that the fundamentalist movement was moving (Wells and Woodbridge 1977:13). They claimed the term “evangelical” for themselves, as it had been largely discarded in the United States. This group became known as neo-evangelicals and tried to develop yet another broad-based conservative Christian coalition (Ellingsen 1988:101). The neo-evangelicals desired to avoid some of the excesses of the fundamentalists (e.g., separation from those conservatives who associated with liberals) and wanted to restore some sense of unity to the evangelical movement. While these two groups had significant disagreements with each other and with ____________________________________ 58 The terms liberal and conservative refer primarily to these groups’ use of what was considered the “traditional” doctrine and should not be confused with the political usage of the terms. [ Page ] 56 more liberal Christians, they demonstrated the willingness of evangelicals to work with each other towards common goals.59 However, this new coalition of evangelicals was not destined to be as broad based and cohesive as the pre-1920 coalition, and by 1967 it was clear that American evangelicalism had no “unified and recognized leadership” (Marsden 1991:74). In part, this was a result of new approaches to evangelicalism generated by younger evangelicals. For example, in the 1940’s an evangelistic organization known as Youth for Christ began and chose a young preacher named Billy Graham as its resident evangelist (Marsden 1991:69). Increased interest in spirituality in the 1950’s contributed to Graham’s success allowing him a measure of independence from the rest of the movement which he used to promote a more “intellectual” variety of evangelical Christianity (Marsden 1991:73). Examples of movement towards this variety of Christianity include founding the periodical Christianity Today with Carl F. Henry and sponsoring the World Council on Evangelism in 1967 (Marsden 1991:73). The National Association of Evangelicals and Christianity Today were to become two of the main defining forces of American evangelicalism. However, as early as the 1950's, some of the fundamentalists had already had a problem with Graham's practice of including any Christian church that wanted to be involved with his crusades (Marsden 1991:73). Marsden (1991) explains that these fundamentalists eventually broke from the evangelicals over this issue. By this point, they had dropped the prefix “neo-.” Further disagreement was prompted by significant rifts in opinion over the inerrancy of the Bible (Marsden 1991:30) and over the degree of militancy and nationalism that was appropriate for an evangelical. In addition, the rapidly growing charismatic movement prompted a rethinking of some of the traditional evangelical ways of reasoning and functioning (Marsden 1991:78). Today there continues to be considerable overlap and a cautious working agreement between evangelicals and fundamentalists although, as always, there are groups that want nothing to do with others (Marsden 1991). Given this history, it is not surprising that American evangelicals have developed a reputation for being isolated and insular. If they could barely work with each other it was assumed that they would be unlikely to be able to ______________________________________ 59 Although this conflict did not have as much impact in Canada as the United States, fundamentalism did play a role in Canadian religion. Examples include leaders such as T. T. Sheilds (Wright 1990:157ff) and the related Baptist schism of Ontario (Stackhouse 1994:393). [ Page ] 57 work with non-evangelicals. Moreover, the dominance of discussions of American evangelicals in numbers and in the media in North America helps to explain why Canadian evangelicals are also assumed to be isolated and insular. 3.2.2 Canadian / American Differences American evangelicalism’s reputation for insularity has at times been applied to Canadian evangelicals because scholars have sometimes assumed that Canadian evangelicals are much the same as their American counterparts. A particularly extreme example is Ellingson’s (1988) book which lumps Canadian evangelicals with American evangelicals. Canadian evangelicals are not even distinguished in the index, whereas smaller groups of evangelicals are (e.g., Caribbean Evangelicals).60 There is considerable overlap in the characteristics, actions, and preferred resources of evangelicals in the two countries. For example, Canadians make use of American publications and evangelical leaders have moved freely between the two countries (A. B. Simpson, Oswald J. Smith, Billy Graham, etc.). Evangelical networks have been, and remain, intertwined across the Canadian / American border. However, despite being neighbours which share some characteristics, Canadian and American religion has developed along different trajectories.61 A significant difference between the Canadian and the American environment is the communitarian emphasis of Canadian society versus the individualistic emphases of American society (see for example Noll 1997:13-15). Canadian society is influenced by individualism, moderated by a statist and traditionalist heritage. Crysdale (1976203) explains that the pre-1906 ethic of “individualism” and the glorification of “personal freedom” which merged with ”economic idealism” gave way to a concern for the social structural problems of Canadian society (e.g., poverty, labour safety). He explains that churches worked together to promote their social principles. One example is the 1917 statement on social principles issued by the Social Service Council of Canada. This council included members as diverse as the Church of England and the Salvation Army (Crysdale 1976:195). Furthermore, American religious organizations were fragmenting even as Canadian religious organizations were merging, the most prominent example being the formation of the United Church of Canada in ___________________________________ 60 See also the comparison of Canadian evangelicalism with British and American evangelicalism in Stackhouse (1994). 61 For specific details of Canadian/American differences see Corelli (1996), Lipset (1990), Noll (1997), Reimer (1995; 2003), Simpson (1985), and Stackhouse (1994). [ Page ] 58 1925. Canadians tried to balance the American emphasis on religious experience with the British emphasis on order (Noll 1992:276). In a similar way, Canadian evangelicals played down differences among themselves of the kind that fragmented American evangelicalism (Stackhouse 1994). The development of networks among Canadian evangelicals reflects the distinctives of the Canadian situation. For example, the relatively small size of Canadian evangelicalism necessitates a degree of working together which was less important in the United States (Stackhouse 1994). Their small numbers make it hard for Canadian evangelical organizations to raise funds (Stackhouse 1994:398ff). As a result, they have to appeal to a broad cross- section of Canadians rather than to a specific focus group as would be possible in the United States (Stackhouse 1994). Given the differences listed here it makes sense to specifically examine the development of evangelicalism within Canada. 3.2.3 Canadian Development62 This brief review of evangelical development and interaction in the United States documents some of the historical context that has influenced Canadian evangelicalism. However, Canadian evangelicalism is not American.63 The first 150 years of Canada’s existence is the story of developing Canadian identity, the coming together and breaking apart of Protestant coalitions and the development of a Canadian evangelical identity independent from mainstream Protestantism. This section provides a review of some of the specifically Canadian factors and events that influenced Canadian evangelicals and their interaction with each other and their environment. In 1867, Canada was founded on the tenets of “peace, order and good government,” a decidedly different objective than the “life, liberty and the pursuit of happiness” set forth as ideals by their American neighbors (Noll 1992:248). The new country of Canada rapidly set about encouraging other regions to join it by building railroads to connect these regions to each other. The wide open spaces of Canada and the relatively small population (only 10,000,000 in 1900 [Noll 1992:280]) encouraged the development of networks for support and for access to resources. The forces which contributed to a growing interest in ecumenism ____________________________________ 62 Unless otherwise noted, this section is based on Grant (1972), Noll (1992), Rawlyk (1990; 1990), Rennie (1986; 1986), Stackhouse (1992; 1993; 1993), and Wright (1990). For an examination of the church in Quebec see Baum (1991) and Swift (1991). 63 While I have not discussed British influence on Canadian evangelicalism, Canadian evangelicalism is not British either. [ Page ] 59 can serve as examples of such pressure towards interaction. These early twentieth century forces which contributed to Protestant ecumenism, included low population density, shortages of clergy, the rise of trade associations . . . the existence of co-operatives favouring mutual aid, frontier anti- intellectualism which weakened arguments against union, the mood of optimism and liberalism, the growing sense of nationalism in eastern Canada and the steady expansion of Roman Catholicism which strengthened Protestant interest in a national church. (Crysdale and Wheatcroft 1976:51). However, when combined with post-war disillusionment, the same tensions which contributed to interaction, contributed to the collapse of the broad-based coalition that had held liberals and conservatives together (Noll 1992:282). When the depression hit, most Protestant churches suffered a loss of revenue and many pastors and church workers saw their salaries reduced (Wright 1990:172). Yet, conservative Christianity continued to grow in popularity because of increasingly successful radio ministries, itinerant evangelists (Wright 1990:182) and the networks created through interconnections among individuals (Rennie 1994). In addition, evangelical institutions established in the period before and between the wars continued to grow. Such organizations included Bible colleges (e.g., Prairie Bible Institute, Toronto Bible College), para-church ministries (e.g., InterVarsity Christian Fellowship), and new or imported evangelical denominations (e.g., several varieties of Pentecostalism, the Salvation Army).64 The nineteenth-century trends of liberalism and population growth and the North American rise of consumerism after World War II led to the increasing secularization (in terms of church attendance and reduced religious control of societal institutions) and religious pluralism of Canada. The result was “a common dissatisfaction with conventional expressions of Christianity” (i.e., Roman Catholicism and mainline Protestantism) (Grant 1972:177). The postwar period also ushered in an era of increased prosperity and social security for Canada. Furthermore, it saw a wave of immigration second only to that of the early part of the century. The country further asserted its own identity, distinct from Britain, and Quebec began to establish its distinct political identity within, and distinct from, Canada. Examples include the introduction of a Canadian flag devoid of British symbolism in 1965 and the Quiet Revolution in Quebec. Conservative Protestant churches increasingly identified themselves as evangelical and experienced numerical growth but also became increasingly ____________________________________ 64 See Wright (1990:168-169) for a brief description of the growth of these movements. [ Page ] 60 disaffected from general society. This general dissatisfaction was accentuated as Canada entered the recession of the late fifties. The religious scene of the late 1960s still displayed the tension between American and British characteristics from the late nineteenth century (Stackhouse 1990:215). However, churches became increasingly irrelevant to Canadians even as the mainline churches continued to try to influence Canadian culture (Grant 1972). Grant (1976:211) argued that, Canadians appeared to choose “their denominations almost at random and that comparatively few took their peculiar principles seriously.” Thus, the ferment of the 1960s reduced the importance of denominational distinctives. One result of this was the merging of like-minded religious organizations and the growth of dialogue between diverse organizations such as United and Anglican Churches (Grant 1976:211). However, not all religious organizations in Canada were interested in ecumenism. The evangelicals interpreted the forces of ecumenism as yet more evidence of the apostasy of mainline Protestant churches (Grant 1976). The evangelical Protestant churches moved more towards co-operation with other evangelicals and away from the more broad-based co- operation of the early twentieth century. The 1960s marked a shift from a more universal Protestant co-operation to a liberal / conservative split something like what had earlier taken place in the United States during the fundamentalist / modernist controversy. However, this conservative / liberal split was not as definite or as antagonistic as it had been in the American situation and a large number of churches kept a foot firmly in both liberal and conservative camps (Anglicans, some United Churches, some Baptists, some Mennonites, etc.). In addition, the cultural, theological, and socio-economic forces which contributed to ecumenism also prompted evangelicals to begin to return to the public square (Grant 1976). In the 1960s there was also a rise in awareness of conservative Protestant groups. Some of these groups appeared on the 1961 census for the first time. These groups include the Brethren in Christ, Christian Reformed, and Church of the Nazarene (Kalbach and McVey 1976). In this period evangelicals attempted to provide values for a society increasingly perceived as valueless, they became increasingly involved in education, and they consistently sent out more missionaries than the mainline Protestant churches. Of the 5,100 foreign missionaries sent from Canada in 1966, 1,700 were Roman Catholic, 700 were mainline, and 2,700 were conservative (Grant 1972:181). Stackhouse (1993:3) explains that this was the case into the 1990s. [ Page ] 61 Expo ’67 in Montreal provided evidence of more changes to Canada's religious geography. The “Christian” pavilion at Expo ’67 was supposed to represent all Christians across Canada, but it represented only some of them because evangelical Christians produced their own pavilion (Stackhouse 1993). Expo ’67 showed that Canadian evangelicals were ready to work together, even as they were parting company with the more theologically moderate views of the mainline churches. Evangelicals’ increasing interconnectedness in common actions like the Expo ’67 pavilion made them aware of enough commonalities that they could contemplate co-operating in a national association separate from the mainline churches. In 1968 the Evangelical Fellowship of Canada began as a voice for those Christians and Canadian organizations who had turned to the increasingly conspicuous evangelical Christianity. Stackhouse (1990:225ff) describes some of the changes to evangelical Christianity in Canada that took place in the 1970s and 1980s. Evangelicals continued to work together and to increase efforts to get to know each other across the country. Their institutions continued to grow, improve in quality, and increase their level of acceptance with the general population. They became increasingly united by their commitment to certain “orthodox” principles (e.g., the authority of the Bible, personal salvation) and by common distaste for cultural trends (e.g., the breakdown of the “traditional” family). They also became more active in social and political arenas and encouraged the further growth of networks connecting evangelicals throughout the country (Stackhouse 1990:225). These networks included transdenominational organizations centred around traditional evangelical concerns such as evangelism and support for conservative morality. One example of such an organization is InterVarsity Christian Fellowship (See Stackhouse 1993:90). Even groups that had traditionally been isolated or had isolated themselves began to be active participants in transdenominational evangelical organizations and schools (e.g., Pentecostals, some Mennonites, and some Baptists, [Stackhouse 1990:226-227]). Evangelicals also sponsored transdenominational schools such as Trinity Western University and Ontario Bible College (now Tyndale College), and publications such as Christian Week (Stackhouse 1990:230- 232). Furthermore, there are many other organizations which have played an important formal role linking evangelicals across the country (e.g., the China Inland Mission (Austin 1997), Bible colleges (Burkinshaw 1997), the Winnipeg Fundamentalist network (Hindmarsh 1997)). [ Page ] 62 Stackhouse (1993) regularly refers to ways in which Canadian evangelicals function like a network. Nonetheless, much significant networking among Canadian evangelicals is informal or based on personal rather than organizational networking. Interaction among Canadian evangelicals includes: formal and informal connections between people and denominations, the different denominations associated with Toronto Bible College, InterVarsity Christian Fellowship and the many denominational and missions organization connections of their staff (Stackhouse 1993: 16, 68, 89, 138, and 186ff; Stackhouse 1990:226). Stackhouse suggests that the process of evangelical organizations connecting was a process of becoming “conscious of themselves as a network of like-minded individuals and organizations” (Stackhouse 1994:385). Stackhouse (1994:387) contends that the term “fellowship” chosen by both Intervarsity Christian Fellowship and the Evangelical Fellowship of Canada is the best descriptor of how evangelicals interact because it “denotes this relationship of distinct elements united by limited, if crucial, common concerns and engaging in limited, but regular and substantial, common activity.” He explains that Canadian evangelicals can be characterized as a “largely informal network of Christians united in their central concerns but pursuing them with only limited cooperation” (Stackhouse 1994:16). Noll (1997:6) argues that there is more inter-evangelical co-operation in Canada than in the United States. More than ever, the polythetic nature of Canadian evangelicalism has become obvious. Diverse evangelical groups with different distinctives are increasingly collaborating in transdenominational endeavors. Such endeavors point to the significance of the networks that connected Canadian evangelicals—they provided assistance in accomplishing organizational goals.65 Thus, the development of Canada’s identity has been paralleled by the development of a Canadian evangelical identity. This identity is seen in the interaction among and the functioning of evangelical organizations. __________________________________ 65 Furthermore, Canadian evangelicals are increasingly co-operating with like-minded Christians in Roman Catholic churches to accomplish specific, but limited, moral and social goals (Jantz 1999) and evangelicals in mainline churches have long supported evangelical events and organizations. [ Page ] 63 3.2.4 Networks Summary Canadian evangelicalism has been influenced by American individualism. Linkages between the two countries continue to be developed and maintained.66 Although Canada shares much with its neighbour, it has also developed in distinctive ways. The picture we derive from historical studies of Canadian evangelicalism is of both increasing distinctiveness from other types of Canadian Christianity and of continued connections with Canada’s traditionally dominant churches. Thus, an examination of networks among Canadian evangelicals should find significant interaction among evangelical churches and limited interaction with other types of Christian churches and with the larger community. 3.3 Linking Evangelicals The history I have just reviewed is suggestive of interaction between evangelicals. However, this history is too general to illustrate how Canadian evangelicals have networked with other, like-minded organizations locally, nationally, and internationally. This section provides specific examples of organizations and actions which illustrate formal and informal network connectivity among evangelicals at the local, national, and international level. These examples complement the general description of the previous section. They illustrate that networks among evangelical organizations were and are a prominent feature of Canadian evangelicalism. 3.3.1 Local Networks: Bible Colleges67 Bible colleges were a ubiquitous part of evangelical life in the early part of the twentieth century and continue to play an important role in Canadian evangelicalism. On the prairies alone, twenty-seven Bible schools were founded between 1921 and 1947 (Wright 1990:164-167). Guenther (2001:21) says that as many as two hundred and forty Bible Schools have been established by evangelical Protestants in Canada. The explicit role of the Bible college was to train evangelical young people in evangelical doctrine (“the Bible as truth rather than as an academic subject”) and practice (Burkinshaw 1997:373). Burkinshaw (1997:374ff) explains that these schools provided opportunities for training to individuals who otherwise would not have access to it. However, they also served as an informal hub for _____________________________________ 66 See Bebbington (1997:43) for the example of the movement of the Toronto Blessing to England. 67 See Remus (1992) and Guenther (2001) for more details about the history of Bible colleges and Bible schools in Canada. Some authors distinguish between Bible Colleges and Bible Schools (Guenther 2001). [ Page ] 64 evangelical networks; a sort of centre for informal evangelical ecumenism (Burkinshaw 1997:374-377). This section describes how the entrepreneurial nature of these Bible colleges and their provision of services allowed them to play significant linking roles in local communities.68 I end this section with a description of some of the networks centred around Toronto Bible College (now Tyndale College). These descriptions illustrate how Toronto Bible College was networked with other Christian organizations and how the type of Christian organization it interacted with has changed over time. Guenther (2001) describes Bible schools as “a kind of invisible link binding congregations together in a common cause.” Bible colleges became network hubs because they tended to be entrepreneurial rather than bureaucratic (Burkinshaw 1997:376). Often they were started by individual pastors or business people who believed that there was a need for a Bible college or school in their community—either to provide Bible training and/or to counteract the forces of culture (e.g., Wright 1990:164). This meant that there was sometimes limited denominational control over these colleges (Burkinshaw 1997:376), and that they were free to draw their students and support from whichever constituents they chose. Another factor which contributed to the connecting of different denominations was the services these colleges provided to local churches such as preaching, music, and literature. Burkinshaw (1997:376) explains that Bible colleges provided these services, in part to advertise, but also to supply their students with practical ministry experience. The result was that churches often received more support from the local Bible school(s) than from their own denomination (Burkinshaw 1997:377). This experience further supported the already cooperative tendencies of Canadian evangelicals. One of the most significant of these Canadian Bible colleges was Toronto Bible College. Throughout its history it has served a noteworthy role in the training of future Canadian evangelical leaders and in linking evangelicals from different denominations and geographical regions. An example from Toronto Bible College’s network connections is the makeup of its board of directors. Its board has included Sudan Interior Mission’s Rowland Bingham and members of InterVarsity Christian Fellowship’s board (Stackhouse 1994:385- 386). A story which suggests this networking may have been extensive is the get-together _____________________________________ 68 While Bible colleges did play a role in connecting diverse evangelicals across denominational lines they could also play a role in the maintenance of a particular cultural identity. See Guenther (2001) for a discussion of the role Bible schools, a related educational phenomenon to Bible Colleges, played in maintaining Mennonite identity and in establishing transdenominational linkages. [ Page ] 65 held in 1946 to celebrate John McNichol’s successful 40-year career. Stackhouse (1993:53) describes the details of this event. John McNichol was principal of Toronto Bible College from 1908-1946. The celebration was attended by denominational representatives from the Salvation Army, the United Church of Canada, China Inland Mission, and the Baptist Federation of Canada. There were also representatives of Wycliffe, Trinity, Emmanuel, and Knox theological colleges and a telegram from then prime minister William Lyon Mackenzie King (Stackhouse 1993:53). This demonstrates that Toronto Bible College interacted with other Christian institutions through its board and its principal. However, this mixture of mainline and evangelical connections declined rapidly between the 1920s and the 1960s. In the 1920s the majority of students at Toronto Bible College (230 of 380) came from the Anglican, Baptist, Presbyterian and United Churches (Stackhouse 1993:68). However, by the time of the merger of the London College of Bible and Missions and Toronto Bible College to form Ontario Bible College in 1968, Anglicans, United Church members and Presbyterians made up only just over 10% of the student body (Stackhouse 1993:123).69 Although Bible colleges continue to serve a linking role among Canadians, they link a different constituency than they did fifty years ago. Bible colleges serve to illustrate the Canadian evangelical move towards an inter-evangelical and community-created network rather than the broad-based Christian networks of the past. Yet, Canadian evangelical institutions sometimes strategically link with organizations, such as universities, which do not share their worldview and have sometimes played a role in marginalizing aspects of the evangelical worldview.70 For example, Tyndale College has agreements with local secular colleges to jointly educate Tyndale’s students. These agreements serve Tyndale’s functional ______________________________________ 69 Part of the explanation for this changing clientele is that Bible college curriculum and educational focus have changed. Education qualifications went up for both students and faculty. In an effort to attract more students Bible colleges have pursued accreditation with religious and secular bodies. This effort has contributed to a movement away from a strictly Bible education to a more liberal arts based education (Burkinshaw 1997). For example, Tyndale College is now trying to obtain official University status and Canadian Bible College is moving to Calgary, in part, because it would like to start a University. Thus, Bible colleges have entered into competition with some of the more liberal institutions which once supported them. This has also meant fostering more diverse and widespread connections with organizations who may not share an evangelical college’s mandate to defend its particularities (the American Association of Bible Colleges, Universities, etc.). These developments lead Bible colleges away from a more local orientation. 70 See Burkinshaw (1995:210) for a description of the reaction of established universities to the efforts to create a Christian university in British Columbia. See also Noll (1992:365ff) for a description of the development of the secular university in the United States and how it was less concerned with the traditional values of the Christian university. [ Page ] 66 goal of equipping Christians for secular careers. That is, the connections between organizations are sometimes based on function rather than shared worldview. Bible colleges could also be used to illustrate how networking takes place on a national level. However, to make the relationship more explicit, I will illustrate national networks using an evangelical organization whose mandate is exactly this linking: the Evangelical Fellowship of Canada. 3.3.2 National Networks: The Evangelical Fellowship of Canada71 The Evangelical Fellowship of Canada developed to be a voice for evangelicals to Canadian governments and to co-ordinate common agendas. Evangelicals were not looking for organizations that would encroach on denominational prerogatives such as evangelism, and development of doctrine (Stackhouse 1993:21). Rather, this lobby/information group aimed to connect large numbers of like-minded evangelicals together for fellowship and common actions (Stackhouse 1993:166-167). This section looks at the inception of the Evangelical Fellowship of Canada and how it has expanded to connect a large number of Canadian evangelicals. Some Canadian evangelicals had been involved with the National Association of Evangelicals (NAE) in the United States since its beginning. That experience was a contributing factor in their desire to have a similar organization in Canada.72 Many Canadians who attended the NAE found it too politicized (particularly Republican), partisan militant, and militaristic.73 In 1964 the Evangelical Fellowship of Canada began as a center for evangelical thought and fellowship, as a voice for Canada's evangelical community, as a means of connecting Canadian evangelicals, and a way to increase awareness of evangelical concerns within the general population. The small group of pastors who were initially involved with it included leaders both from mainline (e.g., Presbyterian and Anglican) and uniformly _______________________________________ 71 This section is based primarily on Stackhouse (1993; 1993), Chapman (1994), and articles in the EFC's magazines Thrust, Faith Alive, and Faith Today. Interviews with individuals who have been involved with the Evangelical Fellowship of Canada have been used to fill in additional details (Stiller 1993; Redekop 1994; Rennie 1994; Zorn 1994). 72 See Stackhouse (1993) for a comparison of the National Association of Evangelicals and the Evangelical Fellowship of Canada. 73 From an interview with a former president of the Evangelical Fellowship of Canada. [ Page ] 67 evangelical groups. However, like Toronto Bible College, most of its leaders eventually came from exclusively evangelical organizations.74 The Evangelical Fellowship of Canada’s constitution was ratified in 1966 and its work officially began with a budget of just over $1000. Its main task was facilitating communication between evangelicals. The EFC became involved in organizing national conferences on issues relevant to evangelicals, took over the relief ministry Share Canada75 from the Christian and Missionary Alliance, and started a magazine initially called Thrust (later Faith Today). This assumption of duties indicates that it was already highly connected with Canadian evangelicals. The organization did not initially obtain the significant national profile in Canadian culture it desired. By 1976 the EFC's budget was only $30,000, and growth continued to be slow. This continued until 1983 when Brian C. Stiller was hired as the EFC’s first full-time Executive Director. Stiller expanded the organization by capitalizing on the emphases of the Evangelical Fellowship of Canada’s revised constitution of 1981. This revision shifted the organization’s direction from reacting to external problems, to improving the life of evangelical churches and added a new goal that emphasized communicating with government more than getting to know each other (Stackhouse 1993).76 In addition, this new constitution emphasized the development of church leadership, encouraged the promotion of unity among evangelicals, and gave the EFC a broader mandate to interact with all levels of government. With the help of this broader mandate, the Evangelical Fellowship of Canada’s income expanded from $60,000 in 1983 to over $2,000,000 in the early 1990's (Jantz 1991:16, advertisement in Faith Today Faith Today 1990:21). Faith Today reported that by 1990, 27 denominations, 750 churches, and 5,300 individuals belonged to the Evangelical Fellowship of Canada (para-church organizations and mission agencies can also be members) (advertisement in Faith Today Faith Today 1990:21). Nine years later that directly affiliated constituency had expanded to “encompasses 32 denominations and related congregations, approximately 150 missions and independent church-affiliated organizations and more than _____________________________________ 74 Interestingly the Evangelical Fellowship of Canada is now reaching out to evangelical groups within mainline churches (Walsh 2001). 75 This organization was later to become the Canadian arm of World Relief. 76 Contrast this with the editorial in the 12:1 edition of Thrust that listed the signs of a collapsing culture as seen in Tutankhamen's Egypt and then pointed to those signs in Canada (Shantz 1980). [ Page ] 68 14,000 individual members.” (from the Evangelical Fellowship of Canada’s website, 2001 http://www.efc-canada.com/) The Evangelical Fellowship of Canada is a formal expression of the informal networking of Canadian evangelicals: what already existed the Evangelical Fellowship of Canada endowed with a measure of permanence. Movement of individuals between Canadian evangelical organizations provided links between organizations that served to unite its various pieces. For example, Brian Stiller, former President of Youth for Christ and former Executive Director of the Evangelical Fellowship of Canada, is now President of the successor to Ontario Bible College, Tyndale Bible College and Seminary. Other examples of the role that the Evangelical Fellowship of Canada has played in connecting evangelicals together include its connection with diverse schools of religion (e.g., Regent College, Briercrest Bible College) and the diversity of its board (e.g., James E. Berney from InterVarsity Christian Fellowship, William MacRae of Ontario Bible College/Ontario Theological Seminary [Stackhouse 1993:171]). Nonetheless, these networks do not encompass all Canadian evangelicals. The diverse nature of Canada and Canadian evangelicals makes it difficult for the Evangelical Fellowship of Canada to represent everyone. For example, by the early 1990s, the Evangelical Fellowship of Canada represented only 1 million of approximately 3 million Canadian evangelicals. However, in the United States the National Association of Evangelicals represented only 4.6 million of an estimated 40 million American evangelicals (Stackhouse 1993). The Evangelical Fellowship of Canada calls itself a “national association of evangelicals.” However, it does not attempt to prevent members from speaking in favour of positions the Evangelical Fellowship of Canada does not hold, nor does it represent all evangelicals. Although the Evangelical Fellowship of Canada’s representatives acknowledge that what they are actually presenting is the majority position of those evangelicals who choose to join their organization, it is to the EFC’s advantage to present itself as speaking for all evangelicals. Stackhouse (1993:20) accurately points out that the EFC’s purposes would be hurt if it referred to itself as “A Partial Evangelical Fellowship of Canada.” For an excellent discussion of the mandate of the EFC as compared to its actual practice, see Stackhouse (1993). [ Page ] 69 Nevertheless, the EFC currently stands as the main national association that binds Canadian evangelicals together across denominational lines and it has become the third largest evangelical organization of this type in the world, after the American and British organizations. Its growth and continued acceptance among evangelicals indicates that the EFC is having some success maintaining its networks among its members. Its organizational structure, desire for broadening acceptance, and profuse literature are all instruments to support evangelical networks and to legitimate evangelical worldviews. Furthermore, its success also points to the increasing self-consciousness of Canadian evangelicals (Stackhouse 1993:165). Given the rough-hewn nature of connections among Canadian evangelicals seen in the development of the Evangelical Fellowship of Canada, it is not surprising that evangelical networks easily cross geographical boundaries. The very diversity of the Canadian population sometimes leads to a situation where a particular group shares more with groups outside the country than those across town. The next section explores these international networks. 3.3.3 International Networks: Formal and Informal Connections The essays in Evangelicalism (Noll et al. 1994) focus on connections among evangelicals. The editors explain that the book’s contributors have made an “effort to show connections between religious developments and surrounding political, economic, social, intellectual, or cultural conditions” (Noll et al. 1994:8). In the process of doing this they also explore how evangelicals in different locations interact. The authors compare everything from how George Whitefield conducted his ministry in different countries to differences in Scottish common sense in different countries (see Stout 1994 and Gauvreau 1994 in that volume). These studies focus on networks among evangelical organizations centred around specific organizations or individuals rather than across broad sectors of a given geographical community. Furthermore, very few of the comparative essays remark on the linkages themselves. Following the editors’ objectives, most are concerned with common themes, ideas, methods, and cultures. Nonetheless, the book does supply some useful descriptions of how international networks among organizations and individuals develop, are maintained, and benefit those who have them. While not all Canadian, these examples illustrate the extent [ Page ] 70 to which evangelicals have been and still are internationally connected. In the following paragraphs I look briefly at publication-based linkages, conscious network development, and some of the personal networks that emphasize interconnections among evangelicals. O’Brien’s (1994) chapter shows how books and newspapers provided vital links between different segments of eighteenth-century evangelical revivalism. These linkages help to explain close connections between different parts of the world which do not appear to be directly connected by specific well-known evangelists or organizations (e.g., between Scotland and New England). Such publication-based linkages assured that people throughout English-speaking evangelicalism, on both sides of the Atlantic Ocean, knew who the main figures of the movement were, and the main events that were taking place. O’Brien explains that the networks which developed to transmit these publications became important as providers of resources which were not available elsewhere. Such linkages between evangelicals were often intentionally developed. An example of conscious network development is found a chapter later. Stout (1994) explains that George Whitefield consciously developed publications and other types of networks as part of his marketing of the gospel. Stout (1994:58) notes that in London during the late 1730s, Whitefield discovered how revival could become, in effect, a consumer product that could be marketed alongside of, and in competition with, the rising swirl of goods and services feeding into an incipient “consumer revolution.” His product—the New Birth—he would market, not primarily in the churches and meetinghouses, but in the public square of the marketplace. Thus, Whitefield consciously developed “new international linkages of transportation, trade, and communications . . . built around the revival.” (Stout 1994:61). Whitefield demonstrates that networks between religious organizations and individuals were sometimes intentional and well planned. He also demonstrated that the goal of evangelicalism—to spread its particular brand of salvation—is one which can be encouraged by the secular skills of showmanship and marketing. Furthermore, he understood that networks of communication are powerful tools for the cause of revival. Stuart Piggin (1994), in a chapter on the American and British contributions to Australian evangelicalism, illustrates how such connections continued to be important in the twentieth century. He gives some specific examples of how networks can be resources which can be drawn upon to further particular organization’s or individual’s goals. For example, Billy Graham‘s 1959 success in Australia was, in part, related to his connections with [ Page ] 71 important individuals. These connections included United States presidents Dwight Eisenhower and Richard Nixon, and personal friendships with Anglican clerical leaders (Piggin 1994:299-300). A letter from Eisenhower written to be read at the crusade increased the Australian acceptance of Graham. Thus far, I have discussed non-Canadian examples. Rennie’s (1994) and Elliot’s (1994) chapters provide some excellent examples of how particular Canadian evangelicals served to connect different organizations. Rennie describes several individuals, such as P. W. Philpott and S. T. Smith, whose relationships spanned countries and linked diverse locations into interconnecting networks of evangelicals. Perhaps the most dramatic example of these types of informal networks in Canada is Oswald J. Smith, the well-known founder of the People’s Church in Toronto. Elliot (1994:369-371) explains that Smith was educated at Toronto Bible College and McCormick Theological Seminary in Chicago. In Chicago he spent time with Paul Rader’s Moody Tabernacle and was involved with a Billy Sunday campaign. On returning to Canada he worked with the Presbyterian Church, the Plymouth Brethren, the Shantymen, and the Christian and Missionary Alliance (Elliot 1994:369). He then left for the United States where he ministered at churches in Chicago and Los Angeles before returning to Canada and starting what became the People’s Church. Smith was in demand as a speaker throughout his career and during the course of his ministry preached at Moody Bible Institute, the Bible Institute of Los Angeles, Angelus Temple, Asbury Seminary, Bob Jones University, and Houghton College (Elliot 1994:370). When he died, in 1986, Billy Graham spoke at his funeral. Smith’s movement between disparate organizations attests to an international network of individuals and organizations which connected the evangelical world. These diverse connections were part of a network of relationships that allowed organizations to discern something about individuals in locations far from their main operations. These networks demonstrate that evangelical organizations collect resources from a wider area than would at first be supposed by their sometimes isolationist tendencies. Nonetheless, it appears that most of these networks developed as people slowly became aware of how individuals and organizations were similar rather than by the intentional actions of organizations (Stackhouse 1994:385). Such connections tended to be informal and personal (centred around individuals) rather than formal and organizational (centred around organizations). 72 3.3.4 How Evangelicals Interact The history of Canadian evangelicalism is, in part, the history of how diverse individuals and groups have interconnected. Many organizations have played an important role linking evangelicals across Canada (Austin 1997; Burkinshaw 1997; Hindmarsh 1997). However, connections among these evangelical organizations are often based on individuals who have common memberships or contacts in several different organizations. Formal links among evangelicals, such as those among members of the Evangelical Fellowship of Canada, remain rare. However, there is a significant network of informal links which has proved sufficient to connect like-minded Canadian evangelicals together to accomplish collective goals (Stackhouse 1994:385ff). I illustrated this using the examples of Bible colleges, the Evangelical Fellowship of Canada, and international networks among individuals and organizations. We can learn several things about networks among Canadian evangelicals from these examples. First, these networks are often informal and organic. Bible colleges networked with local communities because it was a practical way of teaching ministry skills to their students. Second, formalized networks, such as the Evangelical Fellowship of Canada, tend to exist at the level of organizations not at the level of churches in single communities. Third, Canadian evangelicals do not deserve a reputation for insularity. However, who they interact with has changed. In the early part of the twentieth century evangelicals were part of a broad- based Christian network. However, in the twenty-first century they interact primarily with other evangelicals. Nonetheless, most of the literature I have referred to pays little attention to networks among organizations in single communities. Nor do many scholars pay specific attention to the whole network out of which these specific examples arose. It is unclear from these descriptions if the networks would continue to exist if the involved individuals no longer maintained them. That is, the networks appear to be personal rather than organizational networks. However, these examples do point to the benefit of taking a closer look at actual networks of evangelical organizations in specific geographical areas. In the next section I describe a community chosen for this purpose. [ Page ] 73 3.4 Research Location This section describes the community that was chosen for this study. It explores this community’s ethnic, religious, and economic context and compares it with the rest of Canada. These data provide information about the demographic context in which local religion organizations operate and help to determine the extensibility of this thesis’ findings. It is presented, in part, to demonstrate that demographic data alone are not sufficient to explain the network structure among Protestant churches in the local area. 3.4.1 Focus I have restricted this study to examining particular aspects of one specific community. Following Rutman and Rutman (1994:36), I understand community “as simply a field of social interaction.” A “field of social interaction could refer to a geographical location” (e.g., Toronto, Willowdale, downtown), “a form of social organization that enables people to voluntarily live together” (e.g., a church, a commune, a ministerial, monasteries, and convents), or to support networks (e.g., those people or organizations who provide information, help in crisis, share holidays with) (1994:36). Wellman (1988:86) has argued that “the essence of community was its social structure and not its spatial structure.”77 Thus, the community of a particular Protestant churches is all the other organizations they interact with—whether or not those other organizations are geographically proximate. Nonetheless, to produce a manageable data set I studied organizations that are based in a specific geographical community. Specifically, my study is restricted to surveying Canadian organizations that are: active in Kitchener-Waterloo, non-profit (to focus on links between evangelical organizations and not businesses run by evangelicals), and members of the diverse family known as Canadian Protestants. The majority of organizations that meet these criteria are churches. I have excluded all organizations which are structurally separate from mainline and evangelical Protestants, despite some overlap in doctrine, activities, and membership. These organizations may include Protestant members but by common agreement of sociological and historical scholars are structurally separate from other Christian organizations.78 The ______________________________________ 77 Wellman has written extensively on communities (Wellman 1979; Wellman and Wortley 1990; Wellman and Potter 1994; Wellman and Potter 1999; Wellman 2001). 78 Local religious organizations specifically excluded from this study include Christian Science, Church of Jesus Christ of Latter Day Saints (including the reorganized church), Jehovah’s Witnesses, Orthodox (Coptic, Greek, [ Page ] 74 German heritage of Kitchener-Waterloo means that there are large numbers of Lutherans and Mennonites, two groups which contain both evangelical and non-evangelical participants. Thus, my study surveyed all the Protestant Christian organizations in Kitchener-Waterloo rather than only evangelical churches that are the focus of my study. This has the added advantage of providing data on the degree to which evangelicals cooperate with non- evangelicals and provides some comparative data against which network characteristics of evangelicals can be compared. A list of local Protestant churches was compiled from background interviews and from publicly available lists of evangelical organizations. Such lists include the organizations involved in the Evangelical Fellowship of Canada’s effort to communicate the “Gospel” to Canada by the year 2000 (known as Vision 2000), the Canadian Christian Sourcebook, a local Christian Business and Ministry Guide, membership lists of local church associations, local newspapers, and the telephone book. All churches matching my study criteria were surveyed. 3.4.2 Why Kitchener-Waterloo? Kitchener-Waterloo was chosen for this study because it has an active evangelical community, there are many different Protestant churches within its borders, and its size makes an enumeration of evangelical churches possible. Kitchener-Waterloo is not statistically similar to the Canadian average. As an example, I looked at seven of the primarily evangelical denominations represented in the community (i.e., Baptist, Brethren, Christian and Missionary Alliance, Reformed, Nazarene, Pentecostal and Salvation Army).79 According to the 1991 Canadian census, these seven denominations represent 5.3% of the Canadian population but 6.7% of the population in the Kitchener census metropolitan area. The Kitchener census metropolitan area contains 1.7% of all Canadians who affiliate with these evangelical churches, whereas only 0.9% of Canadians live in Kitchener-Waterloo. Thus the percentage of evangelicals in Kitchener-Waterloo is nearly twice as high as in the Canadian population. However, national averages are deceiving, as some areas of Canada are more evangelical than others. For example, “high commitment” evangelicals are almost _____________________________________________________________________________________________ Romanian, Serbian, Ukrainian, and Greek), Roman Catholic, Seventh-Day Adventist, Swedenborgian, Unitarian and all non-Christian religious organizations. 79 This list does not encompass all the evangelical denominations in the community. [ Page ] 75 nonexistent in Quebec and are a significantly larger percentage of the population in Saskatchewan and Manitoba than in Ontario (Angus Reid Group 1996). Furthermore, there is qualitative evidence which indicates that evangelicals in Kitchener-Waterloo share characteristics with other evangelicals in Canada. One illustration of this is that they hold demonstrations and events that are popular with evangelicals across the country. For example, in 1996 they held a well-attended evangelistic crusade featuring Billy Graham’s son Franklin, in September 1998 local evangelicals held a Promise Keepers rally, and the annual March for Jesus demonstration used to attract large numbers of people before it came to an end in 2000.80 In addition, a large number of evangelical denominations common in other parts of Canada are represented in Kitchener-Waterloo. These include varieties of Baptist, Christian and Missionary Alliance, Dutch Reformed, Evangelical Missionary, Mennonites, Pentecostal, and Salvation Army congregations. Most of these denominations are members of the Evangelical Fellowship of Canada. Kitchener-Waterloo also contains many of the service organizations popular with Canadian evangelicals: a Bible college, a crisis pregnancy centre, drop-in centres, thrift stores, and Christian bookstores, etc. 3.4.3 Why Kitchener and Waterloo together? Although Kitchener and Waterloo are actually two cities, there are good reasons to examine them together. First, the two cities run together geographically and are geographically separate from other area cities. If there was no sign to indicate the change in cities, Kitchener and Waterloo would appear as a single city. Residents often refer to the area as Kitchener-Waterloo and for many years the local newspaper was known as the Kitchener- Waterloo Record. The current federal census metropolitan area boundaries group Kitchener and Waterloo together along with Woolwich, Cambridge, and North Dumfries (The Financial Post Corporation 1998:335). Second, there are benefits which accrue to Kitchener and Waterloo collectively. For example, both cities list the two universities in Waterloo as assets. Likewise, employees of Waterloo companies can easily take advantage of lower housing costs in Kitchener. Third, the two cities have many shared resources (e.g., parks, public transit, health care) and activities (e.g., festivals, movie theaters). Waterloo has no public transit, concert ___________________________________ 80 The worldwide leaders say they stopped organizing it because they thought it had served its purpose and it was time to move on to other things, and not because the event was no longer well attended. [ Page ] 76 hall or, until recently, first-run movie theatres, in part, because all three are easily available a few minutes away in Kitchener or are provided for Waterloo by Kitchener. Finally, the relatively small size of the combined cities means that people seeking religious services often cross between the two cities. For example, someone seeking to attend a Chinese Alliance Church would have to go to Waterloo, whereas the only Spanish-speaking churches are in Kitchener. 3.4.4 Demographic Background81 Kitchener-Waterloo is located on highway 401 midway between Toronto and London. In 2000, the cities of Kitchener and Waterloo had a combined population of approximately 260,300 (181,700 in Kitchener and 78,600 in Waterloo). Both Kitchener and Waterloo are growing faster than Canada as a whole, although not by a significant amount. A description of the ethnic, religious, and economic demographics of Kitchener-Waterloo illustrates how the area has changed over time. 3.4.4.1 Ethnic The community has a strong German heritage as evidenced by the large number of Mennonite and Lutheran churches in the area. Of the 88 Protestant churches in Kitchener- Waterloo that returned the survey, 24 self-identify as Mennonite, Lutheran, or German. In addition, Kitchener-Waterloo claims to have the largest Oktoberfest celebration outside of Germany. Indeed, Kitchener was called “Berlin” until World War I. Nonetheless, there is a steadily growing population of area residents who have other ethnic roots. For example, after English and German the most common home language is Chinese in Waterloo, and Portuguese (followed closely by Polish) in Kitchener. However, these three languages represent a significantly smaller proportion of the population than English and German. Nonetheless, in the past several decades, the community has started to move from a predominantly German to a more ethnically diverse community. A former principal of one of the inner city schools has claimed that there were as many as 47 different languages spoken in that school. In addition, Kitchener-Waterloo continues to be one of the main areas where the federal government settles refugees. _____________________________________ 81 This data is based primarily on Kitchener's and Waterloo's web sites which are designed to attract business to their respective cities, lists of local religious organizations, interviews with community members, and on The Financial Post Corporation’s 1997/98 Canadian Markets (1998:56-57, 281, 334-335, 337-338). [ Page ] 77 3.4.4.2 Religious Kitchener-Waterloo’s religious history is rooted in its German heritage. Thus, there are large numbers of Mennonite and Lutheran churches and even some German Baptists. Evangelical-Lutheran churches are significant features of the local community despite their relative absence in the rest of the country.82 Surprisingly, some of the apparently or customarily non-German congregations have roots in some of the founding German congregations. This is true of some of the Evangelical Missionary churches and one of the local Pentecostal churches. Many of the new immigrants to Kitchener-Waterloo are Christians or have Christian roots. This ethnic diversity has added Chinese, Spanish, Romanian, Vietnamese, and other ethnically-based Christian churches to the already present German ones. The majority of the 16 community members I interviewed believed that Kitchener- Waterloo was a Christian community. Local community leaders are still predominantly Christians. Yet, several informants argued that there is less interest in organized religion than there has been in the past. One example of this is the increased use of funeral home chapels because people do not have home churches. In addition, there are an increasing number of non-Christian religions represented in the local area. Evidence for the presence of these groups includes two synagogues, several mosques, a Vietnamese Buddhist discussion group, a Sikh Gurdwara, and a Hindu temple. There are also adherents to some of the newer religious movements such as Church of Jesus Christ of Latter Day Saints (Mormons), Jehovah’s Witnesses, Scientology, and Christian Science. 3.4.4.3 Economic Kitchener and Waterloo both have a higher than average household income, a larger percentage of their population in the labour force and a lower unemployment rate than Canada as a whole. Kitchener's largest employers tend to be manufacturing industries such as Schneider’s meat products and Budd Canada which produces parts for the auto industry. In addition to manufacturing industries, Waterloo has a large number of people employed by two local universities and their spin off companies. Kitchener and Waterloo both have a higher than the national average percentage of their population in managerial occupations but ___________________________________ 82 Evangelical-Lutherans use the term “evangelical” according to its historical definition (i.e., gospel). By doctrine and practice they are usually considered to be mainline Protestants. [ Page ] 78 about the same percentage in clerical occupations. However, Waterloo has both a lower percentage of women in management and a slightly higher percentage of women in clerical positions than Kitchener and the national average. Waterloo also distinguishes itself by having a significantly higher than average household income and roughly double the percentage of the population with a university degree as compared to Kitchener and to the national average. These numbers reflect the presence of the universities in Waterloo (Maclean's 1997). The University of Waterloo, which has approximately 21,000 students, and Wilfrid Laurier University which has approximately 7,900 students. The presence of university students significantly affects the local economy both by supporting local businesses and by providing a readily available source of inexpensive labour. Although, the universities are situated in Waterloo, both cities benefit as students go to Kitchener to obtain products and services unavailable or insufficient in Waterloo such as housing, used book stores, bars, and movie theaters. 3.4.4.4 Demographics Summary Although, this brief review of demographics does not tell us anything explicit about networks among organizations, it does suggest some factors (e.g., wealth, education, ethnicity) which may be a factor in organizations interacting with each other. However, these data, in and of themselves, are not sufficient to understand how evangelical organizations relate to each other and to their community. It is not sufficient because demographically similar organizations do not necessarily interact and demographically different organizations sometimes interact. Therefore, there must be other factors besides demographic similarity that determine why organizations interact. 3.5 Summary In this chapter I have discussed some of the limits of existing descriptions of Canadian evangelical networks, some of the evidence for networks among evangelicals, how Canadian evangelicals are linked together, and the demographics of my research location. 3.5.1 Considerations Some historians have described how evangelical organizations come together for common actions and how common agendas develop (Stackhouse 1993; Burkinshaw 1995; Hindmarsh 1997). However, few sociologists have paid specific attention to networks among [ Page ] 79 evangelical organizations in a single community. Furthermore, rarely do scholars of either discipline specifically examine the connections themselves nor do they usually focus on the impact of these connections within a particular community. Studies of relationships among evangelicals typically focus on a specific organization or group of organizations rather than a community. Descriptions of the networks connecting all the evangelical organizations in an individual community are rare. Thus, we know little about the role networks play in the life of organizations in specific communities. For example, we have little quantitative data on who evangelicals are interacting with in their local areas. This study, which specifically examines network connections among evangelical organizations in a particular community, is an important contribution to our understanding of Canadian evangelicals. Focusing on a specific community provides a sense of resource input and output among the average evangelical church—not just prominent organizations or individuals. It also provides a measure of how evangelical organizations differ from other Protestant organizations in terms of connectivity. This information is not easily discerned from an examination of organizational characteristics alone because some organizational behaviours and characteristics are a product of involvement in larger systems (e.g., a network of organizations, a community, the globe) (Berkowitz 1982; Wellman 1988). Nonetheless, historical studies of Canadian evangelicals provide useful background data for a network study of Canadian evangelicals. Historical studies aid network studies in that they provide important qualitative data and give clues as to where to look for organizations, the ties that connect them, and the resources that are moving through evangelical networks. They also aid in the process of choosing the boundaries of a study of evangelicals (see chapter 2). In addition, they provide a useful point of comparison to the picture of Canadian evangelicals that is provided by a network study. Finally, it is the many allusions to networks in the historical literature that first suggested to me that an examination of Canadian evangelicalism in terms of networks would be valuable. 3.5.2 Evangelical Networks Despite the lack of specific attention to local community networks, there is much evidence of common actions and agendas in those individual communities (e.g., Billy Graham crusades, Promise Keepers stadium events, Christian bookstores, drop-in centers, letter writing campaigns). Commonalities in practice and doctrine and the ease by which [ Page ] 80 evangelicals in a particular community can be mobilized for special events are evidence that some sort of connection is being maintained between various evangelical organizations in individual communities. The historical literature indicates that at least some evangelical churches are highly connected to their local communities even in areas such as social services where the community leaders I interviewed did not expect them to operate (e.g., Hexham 1993). Furthermore, there is extensive informal networking as evidenced by the large number of common activities among evangelicals and the size of some individual networks. In addition, many of the organizations, which have been part of the formation and maintenance of Canadian evangelicalism, were run by individuals who had experience in two or more other Canadian evangelical organizations. My research builds on this literature to provide a more complete description of the networks of local churches. Canadian evangelicalism follows network patterns inherited from its heritage. One of these patterns is a significant network of informal links between organizations (Stackhouse 1994:385ff). Despite their small numbers, Canadian evangelicals have constructed an elaborate informal network of like-minded individuals and organizations. The content of evangelical networks has shifted through the twentieth century to focus primarily on exclusively evangelical organizations and then back to some limited cooperation with non- evangelicals as evangelicals’ worldview has been marginalized in the general culture. This brief exploration of the context of interaction among Canadian evangelicals demonstrates that networking is commonplace in Canadian evangelicalism. Throughout the history of Canadian evangelicalism, organizations (e.g., Bible colleges, the Evangelical Fellowship of Canada) and individuals have actively connected evangelicals in significant ways. Furthermore, there is some evidence of evangelicals having functional connections with non-evangelical organizations in their local community. 4 FOUNDATION: RESEARCH QUESTIONS, AND METHOD Canadian evangelicals are not isolated from each other, nor are they isolated from the communities in which they operate. Thus, in order to fully understand interaction between Canadian evangelicals, I need to study more than just the individual characteristics of organizations and the links surrounding single organizations. My study of networks among Canadian evangelicals complements existing studies of evangelical organizations and individuals by providing a local context which helps to explain how these individuals and organizations function and the support systems which keep them going. More specifically, the chapters that follow will develop a picture of the whole network of evangelical relations within a single community. In the process, I will provide important data on the insularity of evangelical organizations (e.g., the degree to which they rely on other organizations for resources) and present a picture of what that interaction looks like. Thus far, I have established who and what evangelicals are (see chapter 2) and have discussed the context in which Canadian churches operate (see chapter 3). This chapter builds on these discussions to develop some propositions that I use as guides for exploring my main contention: Canadian evangelicals are no longer, if they ever were, completely isolated from each other and from the communities in which they operate. I further developed specific research questions in relationship to my larger research questions: Who are the religious organizations I am studying and how are they similar? How do these religious organizations interact? Why do these religious organizations interact? and How can we account for network structure? The first section of this chapter uses these larger research questions and the context developed in chapters 2 and 3 to construct three sets of specific research questions, which I call propositions, designed to guide the data collection and analysis. To evaluate these questions I collected network data about all the Protestant churches in one local community. The second section of this chapter justifies why I have chosen to explore networks among organizations rather than just the characteristics of those organizations. This section also [ Page ] 81 [ Page ] 82 explains how the literature on interorganizational networks is useful for studying Canadian evangelicals. 4.1 Research Questions My first research question is: who are these religious organizations? Chapter 2 explained how I identified and classified Protestant churches generally, and evangelicals specifically, using polythetic survey data. Thus, this question does not need further discussion here. The next three research questions are used to develop three propositions that I have employed to guide my exploration of the survey data. My first proposition refers to how these Protestant churches interact. Proposition 2 introduces the assumption that these interactions have costs and benefits. This proposition assumes that organizational relations among Protestant churches will be similar to organizational relations observed among secular organizations.83 Proposition 3 looks at these costs and benefits as resources flowing through interorganizational networks. As organizations exchange resources with organizations this interaction changes the character of all the organizations in the network. Proposition 3 assumes that this resource exchange will lead to different networks with different characteristics and different bases of connection—even in the same geographical location. 4.1.1 Structural Characteristics: How do Protestant Churches Interact? Mizruchi and Galaskiewicz (1993:48) report that networks among individuals are a significant factor in the development of similarities among organizations. Likewise, similarities among organizations may indicate densely connected networks of individuals and organizations. DiMaggio and Powell (1983) describe some processes that make organizations similar. McPherson and Smith-Lovin (1987) argue that diversity within an organization is related to the dyadic distance between friends who are involved in that organization. The historical literature has often identified such similarity-based networks among evangelicals (see chapter 3). Differences among groups of organizations may be accounted for by the existence of more than one central, influential organization or by the existence of a variety of religious identities or types.84 Such differences and similarities suggest that the Protestant ___________________________________ 83 This assumption is defended in section three of this chapter. 84 See chapter 2, and various typologies of religious organizations (Swanson 1971) and (Welch 1977), 83 churches in the community I am studying will have multiple networks because of differences and significant network density because of similarity. Therefore, Proposition 1 argues that: (1) Protestant churches in a local community will divide into multiple, densely connected networks with some overlap among these networks. These may or may not be geographically proximate. This proposition focuses on a description of the characteristics and density of church networks in the local area. I begin by establishing who is involved in these networks and what sort of networks evangelical organizations have developed within the local community. For example, how dense are these networks? Is there more than one network in the local area? Next, I explore the relationships that connect the different religious organizations together. I examine what connects and divides these organizations. For example, do networks cohere around specific religious organizations, organizational characteristics, or ideologies? These questions all contribute to the first of my main research questions: how do Protestant churches interact? These issues are explored with the type of interaction question in section two of my mail survey (see Appendix G for a copy of the survey). This question provides data about reciprocal relations (e.g., exchange of services), multiplexy, and emotional intensity or mutual confiding (e.g., if they share staff, information or finances). The type of interaction question also provides data on which organizations are linked and how they are linked. This information can be used to determine the density of the local evangelical network and to find the central actors. A comparison of local organizations of different affiliations will help to determine if density of connections is greater among evangelicals or between evangelicals and non-evangelicals. This is facilitated by section three of my mail survey whose listing of characteristics aids in identifying evangelical and non-evangelical organizations. Section four of the survey provides data about size that may also help to explain network connections and organizational interaction. This description of the network and the basic relationships that form the network provides the context for exploring how evangelicals interact with each other and with the larger community. [ Page ] 84 4.1.2 The Benefits and Costs of Network Relations: Why do Protestant Churches Interact? Given that Protestant churches interact, I assume that there is some reason for that interaction. To discern that reason, my research explores some of the costs and benefits of network involvement. To explain interaction among organizations, much of the literature has emphasized the impact of market forces (e.g., Mizruchi and Galaskiewicz 1993). Several sociologists have argued that religion can also be understood in market terms (Bibby 1993; Iannaccone 1994). In this vein, Finke and Stark (1992) have shown that people maintain relations with Protestant churches if the benefits of doing so outweigh the costs (see also Stark and Bainbridge 1996). Such studies suggest that looking at network relations economically can provide substantial insight into how an organization operates. Thus, although there are limits to an economic model of religious action (Robertson 1992; Bruce 1993; Bryant 2000), organizations should create and maintain links with a network of organizations if the benefits of such an association outweigh the costs. These benefits and costs may be a function of network relations in situations where decisions are a result of network role and position rather than organizational attributes. Alternatively, the historical literature identifies networks as centred around a specific ideology, theology, or function which may also be related to network position. It is reasonable to assume that all of these objectives will be reflected in the data as benefits and costs. Social support is one source of costs and benefits. There is much evidence of social support among individuals. It is clear from the literature on personal ego networks that different types of ties provide different types of social support (Wellman and Wortley 1990; Wellman 1992), and there is reason to suspect that organizational connections are similar (Galaskiewicz 1979; Baker 1990). Thus, I expect that different types of ties will be prevalent in different situations. Furthermore, there is evidence that involvement in a religious institution can improve an individual’s support network (Ellison and George 1994). The role of churches as centres of community involvement and social support (Marsden 1984; Maton 1989; Wind and Lewis 1994) and the need to maintain an organization’s religious worldview (Wuthnow 1976; Stout and Buddenbaum 2001) would suggest that social support is also an important factor in relationships among religious organizations. In addition, all network [ Page ] 85 involvement requires a commitment of time and, usually, other resources as well.85 Thus, Proposition 2 argues that: (2) Networks among evangelical organizations provide the benefits of social support. This provides encouragement towards common goals which outweigh the costs of time and resource commitment which are required to maintain an organization’s links with the larger network. The survey data provides an indication of the costs of network involvement through a comparison between the resources sent and those received. A comparison of attributes, common actions, common group involvement and common identity will determine if more closely linked organizations tend to have common goals. Common goals are identified in this study primarily by common doctrines (from section three of my mail survey), common actions, common group involvement or support (from section two and three of my mail survey), and common identity (from section three). Benefits to organizations can be observed by examining type of interaction which provides data concerning resources received and by an examination of the direction of influence among organizations. The existence of multiple types of interaction gives an indication of emotional intensity and voluntariness which can be indicators of social support. Data about flow of resources (e.g., information, money, support) also provides information about the costs and benefits of network position. For example, do evangelical organizations get most of their resources from within evangelical networks? Other research questions associated with this proposition include: Are network connections reciprocated? What purpose might these networks serve? Which resources are exchanged among network members? Is movement of resources characterized by a pattern of generalized exchange (see Uehara 1990)? How can I identify social support? What evidence is there of social support? The main research question here is: Why do Protestant churches interact? 4.1.3 Homophily and the Supply of Resources: How can we account for Network Structure? Thus far, I have argued that networks reflect the similarities and differences of their members and that network interaction is driven, to some degree, by the benefits it provides. For example, people have a tendency to associate with others similar to themselves (Feld 1982), because it may provide the benefit of reinforcing their worldview (Berger 1967). I _________________________________ 85 See Galaskiewicz and Marsden (1978), Galaskiewicz and Krohn (1984) and Van de Ven et. al. (1979) for a discussion of how different resources affect networks. [ Page ] 86 expect that, in addition to influencing the whole network based, for example, on evangelicalness, these factors will also affect how evangelical organizations divide into multiple smaller networks. That is, the specific structure (e.g., how individual nodes connect with individual nodes) of the network is influenced by the same factors that contribute to the construction and maintenance of the network as a whole. The factors that influence one-on- one interaction also influence network structure. Thus, differences in organization characteristics will lead to differences in network structure. Relationships among organizations could be based on similarities or, alternatively, may simply serve a functional purpose. The location of an organization in the network is as significant as the existence of connections and the content of those connections because structural location influences resource access and movement. For example, Wright and Shuff (1995:335) found that communication among mental health care centres about AIDS/HIV improved if they were “near a care coordination site” or members “of a cohesive network of mental health centers.” This illustrates that information flow is significantly affected by an organization’s geographic and structural location within a network. Using network analysis, I expect to show that there is a high density of connections within the network and thus a short path distance between each pair of organizations. As a result, resources should be easy to exchange among most members of a given network. Furthermore, the small geographic size of the community and the degree to which active Christians are now a minority in Canada (Angus Reid Group 1996; Bibby 1987; Bibby 1993 Census Canada 1993) also leads me to expect that this community network will be almost completely connected. Given these assumptions and the evidence for organizational connections in the historical literature (see chapter 3), Proposition 3 expects to find that: (3) Network role and reasons for network involvement can be different for different organizations a) Networks and cliques will be organized primarily on the basis of organizational similarities. b) However, connections between organizations peripheral to each network and the central organization(s) of each network will be based primarily on shared group involvement and common actions. c) Nonetheless, most members of a network will be close enough to other members of their own network to share resources. This proposition, which is concerned with the relationships among organizations within the network, suggests research questions such as: How similar are closely connected [ Page ] 87 organizations? How closely connected are similar organizations (e.g., reachability)? What sorts of relationships exist between organizations in these networks? And what is the basis for interaction? In short, what is the basis for network structure? The demographic, characteristic, and size sections of the survey provide data on denominational and doctrinal similarities. Central organizations can be determined by a network analysis of the interaction among local evangelical organizations derived from section two of my mail survey (e.g., measures of centrality determine which organizations have the most access to resources). This analysis can be compared to the characteristics survey in section three to determine if organizations in close contact are similar and whether those similarities are related to doctrine and organization type. The type of interaction question identifies the different types of connections which link organizations. Peripheral actors can be identified from core/periphery procedures. Since peripheral actors are less connected to the larger network an examination of their interactions with other actors should show an emphasis on shared group involvement and common actions rather than more direct forms of interaction (e.g., shared personnel). The type of interaction question also provides data on group involvement and common action and can determine if they are more or less common in a particular network. 4.1.4 Propositions Summary In sum, I expect that there will be multiple local networks, that involvement in these networks will provide benefits for the organizations involved in them, and that reasons for network involvement will depend on organizational characteristics and network position. These three propositions provide a way of addressing the different parts of my main thesis. They guide my use of network data to explore the characteristics of local churches, how they interact, what their networks look like, and why the local interorganizational networks among Protestant churches look the way they do. The survey collects data which facilitates the confirmation or disconfirmation of my three propositions and answers my research questions. To accomplish this my study has been limited to a particular location (i.e., Kitchener-Waterloo), a particular type of organization (i.e., Protestant churches), and a particular aspect of those organizations’ contexts (i.e., their relations with other organizations). The second half of this chapter discusses the approach to network data that I used to develop the survey and interpret my findings. [ Page ] 88 4.2 Network Analysis and the Study of Interorganizational Relations My examination of relationships among organizations comes from a body of research loosely connected under the term network analysis. This body of research provides both tools and theoretical perspectives for the exploration of networks among Protestant churches. The first part of this section explores why understanding relations among churches as a network is a useful analytical tool and how a network approach to the study of organizations differs from other approaches. It also describes how existing studies of interorganizational relations are relevant to the study of religious organizations. In the second part of this section, I briefly discuss issues to take into account when applying a network approach to the study of Canadian evangelicals. 4.2.1 Why Networks? Network studies of interorganizational relations are relevant to the study of evangelical organizations in their geographic community. In this section, I first describe the benefits of network analysis for the study of interorganizational relations. I then explain how such studies are applicable to the study of relationships among Protestant churches and review some examples of relevant studies. Finally, I provide several examples of how a network approach to interorganizational relations can improve our understanding of relations among Protestant churches in Kitchener-Waterloo. 4.2.1.1 Description In past three decades, sociologists have been examining both the networks that communities of individuals are composed of (Wellman 1988) and networks within geographical communities (Smith 1997). Some of these efforts have become known as social network analysis.86 Social network analysis is based on the recognition that, social structures can be represented as networks—as sets of nodes (or social system members) and sets of ties depicting their interconnections . . . where the nodes may be persons, groups, corporate entities, clusters of ties or other institutions. The ties represent flows of resources reflecting relations of control, dependency and cooperation (Wellman and Berkowitz 1988:4; Richardson and Wellman 1985:771). __________________________________ 86 There are many good general introductions to social network ideas and methods (e.g., Berkowitz 1982; Knoke and Kuklinski 1982; Wellman and Berkowitz 1988; Scott 1991; Nohria and Eccles 1992; Wasserman and Faust 1994; Wellman 1999). [ Page ] 89 Calloway et al. (1993:378) explain that network data provide a clear way of operationalizing the environment in which organizations function, and lead to conclusions that are more than the sum total of the organizations and their attributes. Thus, they are well suited to being integrated into more established sociological theories and can be used “to estimate the interplay between social structure and organizational attributes.” Like network analysis in general, recent studies of interorganizational relations assume that the examination of structure provides more useful information about actors than a study of their individual characteristics. Wellman (1988:31) argues that “structured social relationships are a more powerful source of sociological explanation than personal attributes of system members.” Nohria (1992:6) argues, similarly, that the actions of actors are better explained by their position in a network of relations than by their attributes. Therefore, the parts of a system are interpreted in terms of the relationship between and among those parts (Richardson and Wellman 1985:771). Thus, there is a significant body of scholarship that has argued for the importance of studying network relationships in order to have a complete understanding of network members. Studies of organizational attributes provide complementary insights to those gained by studies of position (Nohria 1992:7). Attribute data are valuable for determining the attitudes and beliefs of organizations (measured as an aggregate of individual members’ beliefs) but it is hard to use such data to identify power centres, flows of resources and roles of organizations in communities because such relations are dependent on forces outside of the organizations themselves. Network analysis provides a way of studying organizations that concentrates on relationships more than attributes and thus places organizations in a larger context. Network analysis, in general, enables the identification of the structure of the network that is being examined, facilitates the identification of significant groupings and helps to interpret the function of such groupings (Wellman 1988; Scott 1991). An additional benefit of this approach is that interorganizational relations have been carefully explored in other contexts and there are many useful theoretical perspectives and analytical tools which can be applied to my research. 4.2.1.2 Application Many network analysis ideas are easily applicable to the study of interactions among [ Page ] 90 religious organizations. Network analysis has tools to follow lines of power, show consequences of connections and explain certain behaviors within a network (Phillips 1991; Brass and Burkhardt 1992; Marsden and Friedkin 1993). Network analysis can aid in identifying an organization’s structural location and in explaining what that structural location means for an individual organization. Thus, a given social structure serves to create opportunities for some organizations and to restrict opportunities for others (Burt 1992). For example, the position an organization plays in a network can influence the power that organization has over other members of the network in terms of their financial, human, and social capital (Burt 1992:31ff).87 Furthermore, differences in structural location can lead to differences in the roles played in a given community (Galaskiewicz and Krohn 1984; Brym 1988) and position in a network can influence perceived effectiveness (Phillips 1991). Finally, network analysis has studied how organizations connect individuals and how individuals connect organizations (Breiger 1988). Wellman and Berkowitz (1988:7) explain that, [I]ntergroup networks simultaneously connect persons and institutions. Two persons may be connected through an interpersonal tie. But a single person may also connect two groups when he or she is a member of both. Such joint memberships form group-to-group ties that indirectly connect all persons in each separate group. Most network studies of interorganizational relations concentrate on businesses. Many network analyses of interorganizational relationships have emphasized corporate interlocks (Mintz and Schwartz 1981; Richardson 1987) or networks between political elites (Moore 1979; Knoke 1993) as such data are often publicly available and easily accessible. However, there are many network studies which move beyond an examination of corporations or political elites in areas such as the relationship between business philanthropy and nonprofit organizations (Galaskiewicz 1985), health care providers (Walker 1992), social service agencies (Van et al. 1979), and ecological movements (Diani 1990). Such studies have not specifically examined Protestant churches but they do supply concepts and tools that have application to these organizations. Furthermore, there are studies which use network tools to examine aspects of religion in areas such as political attitudes _____________________________________ 87 Social capital is a measure of how many resources an organization has access to by virtue of its network connections. The assumption is that the resources available to an organization are contingent on the resources available to organizations that it is close to in the network. See Burt (1993:9ff) for a description of this idea in terms of relationships among individuals. [ Page ] 91 (Bienenstock et al. 1990), social resources (Ellison and George 1994), and congregational relationships (Herman 1984; Rutman and Rutman 1994). A review of this literature and of historical studies of evangelicals shows that religious organizations, like other kinds of organizations, rely on each other for resources, develop hierarchies of power, and are constrained by institutional, cultural and political structures.88 4.2.1.3 Examples Several examples from the relationships among religious organizations in Kitchener- Waterloo point to the benefit of examining structural location when studying organizational relations. Comparisons of differences in interorganizational relationships among different types of religious groups, by definition, can only be discovered through a network study. For example, a study of the attributes of the local Jewish community might suggest only minor connections with the larger religious community. However, an examination of the connections that the Jewish community has within Kitchener-Waterloo shows that it has a significant amount of interaction with the larger religious network. The most striking examples of this are a local United church and a Jewish Synagogue that share the same building and a Presbyterian church that sends their young people on a tour of a local synagogue before they become members of the church (C216).89 In addition, the local Council of Churches includes some Jewish members (C223) and Jews attend the annual prayer breakfast (C238). In response to a question about which local religious organizations have relationships with a large number of other religious and community organizations, one interviewee described the local Jewish community as influential (C214). Second, Mennonites, by history and by doctrine, are often isolated from the rest of the community. For example, in Kitchener-Waterloo there is a credit union which requires membership in a Mennonite church to join. However, my study of Kitchener-Waterloo shows that Mennonites have a large number of connections with other organizations and a degree of social capital disproportionate to their size in the community. Their social capital could not have been determined by a study of the attributes of local Mennonite organizations alone. It can only be derived from some knowledge of the community network structure in which they __________________________________ 88 See Mizruchi and Galaskiewicz (1993) for a summary of these approaches to explaining interorganizational relations. 89 Specific interviews are referenced with the letter C followed by three numbers indicating which interview I am referring to. Where a statement could potentially lead to the identification of an informant the interview number is not listed. [ Page ] 92 are embedded. Conversely, structural position can also reduce community social capital. The data show that evangelical organizations have a network among themselves that has few organizational connections to other local churches. There is little awareness of their activities in the community. This observation by itself does not explain evangelical invisibility in the community as mainline churches also do not interact with other churches but they are visible in the community. Furthermore, anyone seeking awareness of the scope of the evangelical community must reach beyond public sources, as many evangelical organizations do not publicize themselves outside of their specific location in the community. Their marketing may consist of nothing more than a sign in front of their building. Nonetheless, conversations with local evangelicals indicate high community involvement. Observation of the entire geographic community indicates that this involvement focuses on social services but includes very little interaction with other churches. Some characteristics of organizations can be more easily discerned by a study of relations among organizations than by a study of the attributes of individual organizations. This is the case because explanations of some organizational behavior depend on the structure of the entire network rather than the immediate connections of individual organizations. Such intergroup relations are a factor whenever groups interact regardless of their identity or functional area. Network analysis ideas are easily applicable to the study of Protestant churches. Thus, a network study provides data about these organizations that cannot be provided by other means. 4.2.2 Application to Canadian Evangelicals This brief discussion illustrates the basic applicability and interpretive value of network analysis for the study of religious organizations. For my purposes, network analysis is well suited to the examination of Canadian evangelical organizations. Nonetheless, Burt et al. (1994) have shown that the impact of network structure varies considerably depending on which industry an organization is in. Therefore, it is not surprising that there are differences between evangelical organizations in Canada and other types of organizations that must be kept in mind when applying insights developed in other settings. First, evangelical organizations in Canada tend to be more co-operative and less competitive than corporations. Canadian evangelical organizations often work together in [ Page ] 93 opposition to other religions, secular organizations and cultural forces rather than in competition with each other (Stackhouse 1994). One illustration of this co-operation is the freedom with which Canadian evangelicals move between evangelical denominations with only passing attention to a denomination’s distinctives (Bibby 1993:36-38). Burkinshaw (1996:15) explains that “in an ironic twist, evangelicals, who have usually eschewed official ecumenism throughout most of this century, have nonetheless practiced a pragmatic kind of ecumenism that often renders denominations almost irrelevant.” This network of cooperation can be expanded beyond evangelicalism for specific functional goals. That is, evangelicals will work with non-evangelicals to accomplish narrowly defined common goals (e.g., funding for religious schools, restriction of access to abortion). Second, evangelicals in Canada, like other voluntary associations (see Knoke and Prensky 1984:12), tend to be on the periphery of society (in terms of politics, education, culture, media, etc.) whereas businesses tend to be at the center. Although, individual evangelicals in Canada are occasionally close to the center of the power structure, they typically do not use their position to try to bring evangelical organizations closer to that center (e.g. Preston Manning, former leader of the Reform party). They also have the potential to provoke the opposite effect (e.g., Stockwell Day, former leader of the Canadian Alliance Party, has been ridiculed for his conservative religious values). This is the case, in part, because Canadians exhibit relatively low levels of religious practice (Stackhouse 1994; Reimer 1995:397). However, the Canadian evangelical sense of separateness is also a result of alienation from the mainline Protestant denominations (Stackhouse 1994:385) and of the Canadian evangelical tendency to place more emphasis on spiritual concerns, broadly defined, and less on capital accumulation. For example, most workers in Canadian evangelical organizations are not in it for the money, in part because there is very little money to be made in Canadian evangelicalism (Stackhouse 1994:398). In addition, Canada has moved further towards the functional differentiation of societal subsystems than many other parts of the world and thus religion often approaches societal influence as an outsider (Luhmann 1982; Luhmann 1984; Beyer 1994). Finally, Canadian evangelicals, like most voluntary associations, are more likely than corporations to rely on volunteers. One implication of this is that an organization has less control over volunteers than over paid workers. For example, a church member cannot always be kicked out for incompetence. Volunteers need different sorts of rewards than [ Page ] 94 employees. However, a diversity of volunteers can be a huge asset when a religious organization is attempting to influence a specific functional system or trying to obtain access to resources across diverse networks. For example, church volunteers are vital players in many social service organizations. Interaction among volunteers in social services may link churches in ways that church to church connections do not. Despite such differences between Canadian evangelicals and other groups in Canadian society the basic approaches of network analysis remain useful for the study of interorganizational relations between Canadian evangelical organizations. 5 COLLECTION: STUDY DESIGN AND DATA COLLECTION My propositions and research questions guided my collection of data about how evangelical organizations interact with each other and with their local community. I have chosen to base the data collection in Kitchener-Waterloo because of the availability of data. To collect and analyze these data I have made use of the insights of network analysis because it specifically focuses on the networks which hold the key to testing my propositions. In the previous chapter I provided a justification for exploring network connections among organizations rather than focusing exclusively on organization attributes. In light of that background, I developed a survey to collect network data about Protestant churches in Kitchener-Waterloo. This chapter provides a description of the main components of that survey, its format, and the organization of these data. 5.1 Survey Components My study has two parts: exploratory interviews for background information and a mail survey of Protestant churches. The purpose of the exploratory interviews was to get a sense of how local religious organizations interact. A second purpose was to determine which sorts of questions would provide information about these organizational interactions. The subsequent mail survey collected demographic information and data about organizational ties, organizational characteristics, and organizational identity from all the Protestant churches in Kitchener-Waterloo. 5.1.1 Background Interviews I interviewed sixteen community members to get an understanding of how aware they were of networks among religious organizations and some sense of their perception of the religious character of Kitchener-Waterloo. I interviewed evangelical and non-evangelical community members including representatives of local political, academic, and religious communities. Each interviewee was asked specific questions about their perception of the involvement of religious organizations in the community. They were also asked about their contact with evangelical organizations in the community. There was enough evidence to [ Page ] 95 [ Page ] 96 demonstrate that some organizations are more visible in the community than others and to identify some distinct groups of densely connected religious organizations. Different perceptions on how local Christian organizations are involved in the community pointed to the need for my larger survey. See section 2.1.1 for more details about these interviews. See Appendix A for a copy of the actual questions. 5.1.2 Mail Survey90 The mail survey has five sections. Survey sections were titled “section ” on the survey to avoid any bias that may result from labeling them in ways that are more specific. See Appendix G for the actual survey questions. Section one of the survey requested identifying information about the organization and the informant. Section two asked each informant to provide a list of the organizations that they had contact with during the previous year and which types of contact they had with those organizations. Section three asked each informant to indicate which characteristics are descriptive of their organization. Section four collected information about the attendance, membership, and staff size of the informant’s church. Finally, section five asked for a list of the religious organizations the informant’s church most identifies with. The back page of the survey provided room to supply comments. The following paragraphs describe each of these sections and explain the content choices made during survey development. 5.1.2.1 Section One - Demographic Information The survey begins with a page that requests identifying attribute information about the organization being surveyed and about the individual filling out the survey. Information requested from the organization included its name, denominational affiliation (if relevant), address, telephone number, fax number, email address(es) and web site(s). Information about the informant included name, informant title, date the survey was filled out, and how long the informant had been with the organization. Such information provides unique identifiers for each organization, which aided in classifying them into categories for later analysis. Organization name was requested to assure that a current and correct name was being ____________________________________ 90 Relevant methodological sources were consulted to design my study (Bassili and Scott 1996; Bishop 1987; Cartwright 1988; Chebat and Picard 1991; Clark and Schober 1992; Couper 1997; Dillman et al. 1993; 1983; Foddy 1993; Goyder 1985; Heberlein and Baumgartner 1978; Hensley 1974; Johnson et al. 1993; Lofland and Lofland 1995; Marsden and Hurlbert 1987; McPherson 1982; Nederhof 1988; Payne 1951; Rossi et al. 1983; Sheatsley 1983; Sudman 1982; Tanur 1992; Yammarino et al. 1991). [ Page ] 97 used for the organization. Organization name, telephone number, fax number, Email address and web site were important for future contact with the organization (e.g., for clarification of data collected). Denominational affiliation allows the determination of whether or not evangelical organizations group by denomination. The address of the organization is important if the informant wanted future mail and telephone calls directed to a different address than the one where the survey was received. It was also important for the distribution of a summary of research at the end of the study to those informants who requested it. Postal codes were used to determine geographical proximity. An organization’s web site was important both for collecting additional information about the organization and for gaining an understanding of the degree to which an organization uses new communications technology in their work. Knowledge of the informant's title was necessary to determine if regularly equivalent (which I will refer to as “equivalent” from now on) individuals filled out the survey. Regularly equivalent individuals have identical ties with equivalent individuals (Wasserman and Faust 1994:474). For example, Senior Pastors can be said to be in equivalent positions because they all have, in some respects, identical ties with their congregations (i.e., they have a church-granted authority over congregation members). Note that Senior Pastors in equivalent positions do not necessarily have authority over the same congregation members. Wasserman and Faust (1994:473) explain that the “notion of regular equivalence formalizes the observation that actors who occupy the same social position relate in the same ways with other actors who are themselves in the same positions.” I attempted to have the mail survey filled out by equivalent informants, in a key position in the organization. My objective was to assure comparability of response and complete information about an organization’s activities. For the purpose of my study, I assumed that Pastors were most likely to be able and available to answer the survey and to be in key structural positions. I define a key structural position as one that bridges two or more different groups. Gaps between groups of individuals have been called structural holes (Burt 1992). Informants which bridge such structural holes have access to information faster and earlier than other individuals and they have more control over that information (Burt 1992:49). Informant name and title may indicate the importance placed on the survey or which positions hold interorganizational knowledge within the organization. For example, is there a [ Page ] 98 difference between surveys filled out by secretaries versus surveys filled out by pastors. Length of time with the organization could be an important indicator of the knowledge the informant has of the organization and its connections with other organizations. 5.1.2.2 Section Two - Network Relations among Evangelical Organizations Section two is concerned with the connections Protestant churches in Kitchener- Waterloo have with other organizations based on their recall of those connections. It provides data about the degree of interaction among organizations in Kitchener-Waterloo. It also collects data that facilitates a comparison of local versus distant connections, evangelical versus non-evangelical connections, and similar versus dissimilar connections. Galaskiewicz (1979:48) notes that information about the amount of a transaction is hard to collect, as this is considered considerably more private than the existence of a transaction, which is often already publicly available. Although the inclusion of these data would have increased my ability to quantify the data, there was a danger that it would decrease response rates and make the questionnaire more difficult to fill out. Respondents were asked to list all religious organizations their church had interacted with in 1999. The period of one year was chosen to provide a comparable framework for the study of organizational interaction. This period was also chosen to restrict this study to current rather than past connections and to aid in immediate recall. The survey was distributed in late January and early February 2000. To assure that organizations answered the question in similar ways respondents were asked to report on interactions in terms of a very specific list of tie types. I asked about six different types of ties (i.e., physical support, information, events, finances, volunteers, and paid staff). These types of ties were chosen to encompass the variety of ties between evangelical organizations indicated in the historical literature (e.g., Stackhouse 1993; Burkinshaw 1995; Rawlyk 1997) and in popular Canadian evangelical magazines (Faith Today, Christian Week, Focus on the Family, etc.). Table 5.1 indicates how the types of ties collected are related to network concepts which describe the characteristics of ties. [ Page ] 99 Table 5.1 - Type of Network Data Collected [ Please contact repository@tyndale.ca for Table 5.1 details ] Ties Characteristics of Ties Type of Interaction Type of Tie Intimacy Voluntar yness Social Support Resource Flow Indirect Direct Estimated Strength Provide them with physical support (e.g., share facilities) support V V V V med Share information (e.g., upcoming events) information V V V V V low Participate with them in the same events action V V low Provide them with financial support finance V V V V V high Share volunteers personnel V V V V V med Share paid staff personnel V V V V V high Ten spaces were provided for organizations to list their contacts. A very small limit on contacts listed would not have been sufficient to allow comparison of the breadth of each organization’s network and the insularity of connections among evangelical organizations. However, requesting that churches list all the organizations they are connected to increases the work required of informants, increases the length of the survey, reduces the likelihood of a complete response, and increases the chance of non-response. Limiting the size of the survey provided the informants with some confidence that the survey would not last forever and provided me with a roughly comparable network size. Churches generally spend a small amount of time on networking and have limited resources (staff, money, etc.). Therefore, it is unlikely that many churches have more than ten contacts with other religious organizations of the specified types over a year. There are problems with limiting response spaces, as different organizations can have radically different sized networks, often dependent on their location in the network. Restricting organizations from listing all the organizations that are “significant” for their ministry could distort examinations of their centrality and influence. However, without some compelling reason to allow an unlimited number of ties there can be significant discrepancies in what is being measured (Feld 1997). For example, for a small organization, all its ties may be significant, whereas, for a large organization, only a small number of its ties may be [ Page ] 100 significant. Pre-tests showed that ten spaces were sufficient to collect information on most of the organizations informants could recall. 5.1.2.3 Section Three - The Evangelical Characteristics Index Section three of the questionnaire had two purposes. First, it asked questions about characteristics of Protestant churches in Kitchener-Waterloo. These characteristics have been used to classify and describe those organizations (see chapter 2). This classification was compared to the groupings developed from a network analysis of the data. Second, these data were used to explore differences in network structure among similar organizations. Nonetheless, attributes do not determine structures (Friedmann 1979). Consequently, similar organizations may not have connections with each other for reasons related to network organization rather than organizational characteristics (power, geography, theology, size, etc.). Thus, sections 2 and 3 of the survey complement each other. Churches often contain individuals with a variety of different beliefs. However, this study is concerned with official local church identity. Thus, answers reflect attempts by pastors to combine official doctrine with the aggregate beliefs of their congregations. Informants were provided with a list of beliefs and characteristics and asked to rate each characteristic or belief on a scale from one to seven based on how descriptive it was of their organization. This list of characteristics of religious organizations was compiled from existing definitions and classifications noting which characteristics are considered necessary while abandoning the sufficient nature of these characteristics (see chapters 2 and 3). Characteristics were chosen to prompt immediate recognition and to maximize variance. An attempt was made to balance between the need for a large number of characteristics and the need to keep the survey small to maximize response. The characteristics chosen were also restricted by the limits of my study. Thus, the characteristics collected apply to evangelicals but are not all group specific, and they are comparable across groups. 5.1.2.4 Section Four - Size Information This section collected information about different aspects of the size of local Protestant churches. Data collected included number of members, average weekly attendance, number of paid staff (full-time and part-time), and number of volunteers (full- time and part-time). [ Page ] 101 Organization size is an important variable as it can point to important distinctions among organizations such as internal differences in functioning and salience of the organization in the community (McPherson 1982:226). Different organizational structures are needed to keep large organizations functioning than are needed to keep small organizations functioning. There may also be some relationship between size data and identity data. Size data are included here rather than in section one because these are data that are more personal and require some extra effort to look up. There was a danger that people ma not have completed the survey if they encountered it right away, whereas, by the time they encounter section four on the fifth page of the survey, they have already committed to the survey and will likely complete it. 5.1.2.5 Section Five - Similar Organizations Section two asks about which organizations local Protestant churches are linked with but it does not provide any indication of how similar those churches are or how similarly they are perceived by informants. Section three provides some beliefs and characteristics that can be used to determine how similar churches are but it still does not provide data about what prompts a church to link with one organization and not another. One explanation for organizational interaction is a perceived similarity in identity and objectives. To examine identity perception section five asked informants to list religious organizations they consider similar to their church. Organizations were provided with ten spaces to list, in ranked order, organizations similar to themselves.91 It was clear from the response on this section that informants had misunderstood this question in several different ways and thus data from this question were not included in most of my analysis. The remainder of this chapter provides details about the development of the survey for those unfamiliar with survey research. Readers already familiar with survey research may want to skip straight to section 5.3.2 and then proceed with chapter 6. _____________________________________ 91 See section 5.1.2.2 for a discussion of the benefits and problems of limiting responses. [ Page ] 102 5.2 Physical Format and Deployment In addition to deciding on the content of the survey and the wording of the stimulus questions, it is important to choose a physical format and method of deployment that maximizes the likelihood that the survey will be filled out. The intent behind the decisions made in developing the survey is that people will be more likely to respond to the questionnaire if the cost of doing so is minimized and if the rewards of doing so are maximized (Dillman 1983: 361). Thus, the survey design aimed to reduce costs, and to produce a survey which appeared easy to fill out, personal, relevant, and important (Dillman 1983: 367). This section explains the formatting choices made for the survey, the cover letter, and the envelopes. It also explains why a mail survey format was chosen to collect the data for this study. 5.2.1 Survey92 Dillman (1983:362) argues that the questionnaire should be a 6.5x8.25 booklet photo- reduced from 8.5x11 on 16 lb white paper (but not less than 15 cpi (Kervin 1992: 433). However, Johnson et al (1993) found that reduced size surveys were less effective. They speculated that this might be a result of the reduced text size). Other studies have had success with 8.5 by 11 paper stapled in the corner. This is the format used in this study. The typeface should be easy to read. This study used 12 point Times as its primary font. A largely black, high contrast cover increases response rates as it is both noticeable and indicates that it is a serious survey (Nederhof 1988). LaGarce and Washburn (1995) have found that a user-friendly format (non-technical and easy to read) is more important than the use of color. Although Kervin (1992:432) argues that yellow paper may have a better chance of being noticed, colored paper is generally to be avoided to avoid confusion with junk mail (Nederhof 1988:356). The title page of the survey contained an interesting title, a neutral eye- catching illustration, and instructions to the respondent. The last page was reserved for comments. Question order usually moves from the most interesting, topic-related questions to potentially objectionable, and should end with any demographic questions. However, the first question on the survey should be universally applicable, interesting, and easy to answer. _____________________________________ 92 See Miller (1991:101, 110ff) for a breakdown of the relative benefits different techniques have for increasing rate of return. [ Page ] 103 Thus, the survey begins with demographic questions because they are not personal questions, and because they foster commitment to completing the survey better than the more demanding body of the survey. Moving from demographics to interaction with other organizations to organizational characteristics matches the criteria of universally applicable, interest, and ease of answering. The body of the survey is laid out in a black on white matrix format. It has lowercase questions in a vertical line, no overlap across pages, clear transitions, and clear visual clues to aid with the written directions. The part of the survey that lists organizations or characteristics uses periods to indicate which box goes with which organization or characteristic. Yammarino (1991) argues that a mail survey should be no more than 4 pages. The survey has 4 double-sided pages including the cover page and the comments page. 5.2.2 Cover Letter93 The cover letter had a personal appearance (e.g., addressed to specific person, signed in blue ink with ballpoint pen) (Dillman 1983:366). It used University letterhead to highlight its importance and to give it an air of authority. To make it easier to read it was only a single page, composed of short paragraphs of 1-3 sentences, and left justified. The cover letter explained the purpose (Foddy 1993:71), and the social relevance of the survey (e.g., it provided a very specific description of the benefits to the local Protestant community). It also explained why it was important that the recipient organization complete it, who should complete the survey (Dillman 1983:366), why a particular respondent was selected (Kervin 1992:429), answered initial questions,94 and let the informant know how to contact me if they had any additional questions. The cover letter assured the informants of the confidentiality of individual respondents (though not of their organization) to encourage them to answer the questions freely. It also tried to encourage self-important and fringe groups to participate and was general enough to appeal to those who do not consider themselves evangelical. It was specific about the period covered by the survey and indicated that exactness was important (Tanur 1992:88). Finally, studies have shown that even though some people do not read the _______________________________________ 93 See the text of the cover letter in Appendix F. 94 Sudman (1985) recommends that, when surveying professionals, more detail regarding the study should be given than would be given to the general public. [ Page ] 104 body of a letter, most people will read a postscript. Thus, the postscript was used to suggest completing the survey immediately. 5.2.3 Envelopes and Stamps The envelopes were addressed to specific individuals and had a paper stamp rather than an ink stamp or mailing labels (Dillman 1983:366). This personal touch aimed to increase interest and thus response rates (Nederhof 1983). To make return envelopes easier to use all envelopes were pre-addressed and stamped. Hensley (1974) has found that response rate is improved by using different stamps on the initial mailing and on the return envelope. A pairing of a commemorative stamp and a metered stamp was found to be the most effective. This study used fancy stamps on the initial mailing and ordinary stamps on the return envelope. 5.2.4 Mail Surveys y study collected the majority of its data through a mail survey. Goyder (1985) has found that response rate to personal interviews has declined even as response maximization techniques have increased the response rate to mail questionnaires. The main predictors of response rate are number of contacts with informants and salience of topic (Goyder 1985). He concludes that if direct contact is possible an interview is probably more effective but that direct contact is increasingly hard to obtain (Goyder 1985:248). Furthermore, there are benefits to using mail surveys and questionnaires to collect interorganizational relations data. First, mail surveys are considerably cheaper, can be more widely distributed geographically . . . they are more easily repeatable; there is no interviewer effect and no problems ensuring that interviewers are adequately trained and briefed and are following instructions in the appropriate way (Cartwright 1988:172). Mail questionnaires are less expensive in part because they are less time consuming than interviews and because travel costs are more expensive than mail costs.95 Thus, mail surveys allow the examination of larger networks than other means of data collection for the same cost. Second, a mail survey places less pressure on the respondent. It takes less time than telephone or personal interviews but allows the respondent more time to answer questions if ___________________________________ 95 To conduct 150 interviews would require almost 6 weeks of continual interviewing. This assumes that interviews could be scheduled consecutively and require no more than 90 minutes for contact, travel time, interview and follow-up. [ Page ] 105 they want it. In a network survey the information desired may require some thinking or may need to be looked up, therefore it may not be immediately available in an interview. Personal contact with busy religious workers may be more difficult than contacting them by mail and the flexibility afforded them by a mail questionnaire compensates for the tendency respondents have to ignore it. The idea of committing 60 minutes of their time to an interview that is scheduled for a specific time would be more disruptive to their schedules and thus, would increase refusals and decrease the chance of contacting them directly. Third, pastors, who are the predominant leaders of evangelical organizations in Kitchener-Waterloo, are members of the population that are most likely to respond to a mail survey. That is, they are religious, tend to have some higher education, and are likely to be interested in the topic (Miller 1991:108, 111; Sudman 1983). Mail surveys are often effective in contacting professionals because they are often busy and hard to contact (Sudman 1985). Furthermore, the topic area of the survey is not sensitive and is salient to religious workers (Miller 1991:111). Mail surveys are cheaper than personal interviews, they have data collection benefits, they place less pressure on respondents, and they appeal to the audience of this survey. It is possible to get a return rate of between 65% and 77% with mail surveys (Bibby 1993:318 Dillman 1983:340). I obtained a rate of return of 66%. I sent out 143 surveys. 10 of those were not delivered because I could not find the organization or the organization no longer existed. I was able to collect 88 of the remaining 133 surveys. 5.3 Data Collection This section reviews the data collection procedures I used with the mail survey and describes how I organized the network data for analysis. Data collected in the demographic sections of the mail survey were used directly as recorded on the survey. The organization of the characteristic data of section three for analysis was described in chapter 2. 5.3.1 Procedures The data collection procedure had the following steps: [ Page ] 106 START DATE DETAILS Pre-survey - collect qualitative background Oct. 1999 - initial telephone contact (determine contact person, indicate survey will be sent, provide no opportunity to refuse). Three tries then mail blind. Jan. 2000 - deliver survey package (by mail) Feb. 2000 - follow-up by letter March 2000 - mailed a second survey April 2000 - phone-call to not returned (encourage submission of survey, or determine reason for non-response) Until June - received surveys The first step was a telephone call to each organization before the survey was mailed out. A script was used to ensure consistency. The purpose of this call was to indicate that the survey was coming and to determine to whom to address the survey. It also aimed to find someone who knows which other organizations their organization interacts with and has the time and desire to complete the survey (Kervin 1992:451). This was preferably the head pastor, but other informants were used if the church did not currently have a head pastor. The contact script was designed to find this person. No opportunity was given to refuse receipt of the survey. The second step was to deliver the package to the organization. This package included the cover letter, the survey, and a stamped return envelope. It was delivered on Tuesday, Wednesday or Thursday because these days made it more likely to be paid attention to. There were two possibilities for retrieving surveys: mail-back or personal pick-up. Personal pick-up may have provided more incentive to complete the survey. However, mail back was less costly and allowed the informants more time freedom to complete the survey. To maximize returned surveys three levels of follow up were used in the weeks following delivery of the survey: a reminder letter, a second survey, and a phone call. In the 1970’s Dillman’s method was found to produce an average response rate of 74% from mail questionnaires (Miller 1991:116). Goyder (1985) has argued that number of follow-ups is one of the key predictors of response rate. Miller reports that 20% more respondents can be expected for the initial follow-up and 12% and 10% for the subsequent two follow-ups (Miller 1991:110). Follow-up in this survey started with a letter mailed to organizations that had not returned the survey. A letter was used instead of a postcard because the cost of producing [ Page ] 107 postcards was prohibitively expensive and there is evidence of the successful use of letters rather than postcards (Nederhof 1983). The letter indicated that the survey has not yet been returned, encouraged them to contact me if they had any questions, and thanked anyone who has just sent it. Second, another survey was mailed out. Finally, a telephone call attempted to encourage the final surveys to come in and tried to determine reasons for non-response. Dillman recommends sending a third survey by registered mail, but Nederhof (1988:358) has found that telephone reminder was just as effective. Some scholars consider a third level of follow-up optional because there is usually not much additional response (but see Nederhof 1988). After doing an initial analysis of the data, I sent a letter thanking informants for their participation. I also provided a one-page summary of the data to all respondents. It took approximately four months to conduct background consultations (the 16 interviews referred to earlier). It took an additional eight months to contact each organization, distribute the survey, collect completed surveys, and follow-up on unreturned surveys. As soon as they were received data from the surveys was entered into a database and then transferred to the statistical analysis program SPSS. Some initial data exploration was done at that time. By June 2000 all the data were received and entered. 5.3.2 Network Data Organization My collected network data (from section two of the survey) was organized into 7 matrices. One 88 by 343 matrix (88 respondents by 343 unique listed organizations) was produced for each of the 6 types of interaction plus a seventh 88 by 343 matrix containing the sum of all types of interaction. My matrices only included returned surveys. A review of the 45 informants who did not return surveys shows that 14 of them are listed in the 88 returned surveys. Only one of those 45 organizations had connections to more than one of the returned surveys and this connection did not connect organizations that were not previously connected. Each cell represented the number of interactions between each pair of organizations. Each matrix was converted into smaller matrices for ease of analysis. First, an 88 by 88 matrix was formed by including only respondent to respondent connections. To produce a square matrix, zero columns were added where there were no interactions. Second, the 88 by 343 matrix was reduced to an 88 by 333 matrix by removing generic connections. These are connections described generically (local churches, missions, [ Page ] 108 social services, etc.). These 2-mode data (informant by connection) were converted into 1- mode data by counting the number of column actors that each pair of row actors are both involved with and vice versa. Each cell of this matrix represented the number of organizations with which each pair of respondents interacted. This matrix represents shared connections to the same third-party organizations not direct connections. Third, additional matrices were created by zeroing the diagonal of both of these matrices (the diagonal represents self-connections and provides a count of the number of organizational connections) and by converting the matrices to binary values. Diagram 5.1 - Matrix Dependencies [ Please contact repository@tyndale.ca for Diagram 5.1 details] Fourth, to provide a middle ground between the direct connection matrix and the shared organizational connection matrix, the 88 by 333 matrix was further reduced to eliminate interactions that may not indicate any actual contact between organizations. For example, allowing the Girl Guides to meet in your building or giving food to the food bank are shared among many organizations but are unlikely to lead to actual contact between organizations. All connections that were unlikely to lead to interaction between churches were removed from the matrix. The resulting matrix was 88 by 86. Zero columns and rows were added to this matrix to produce a square 88 by 88 matrix that was compatible with network procedures that facilitated comparison of the different matrices. This resulted in a [ Page ] 109 direct connection matrix involving all local churches and organizations that would likely lead to direct connections. The 88 by 86 matrix was also converted into 1-mode data and represents shared connections to the same organization. Table 5.2 lists each of the 21 matrices produced by this process, their size, and their characteristics for each of the seven initial matrices. This could potentially produce 147 matrices. For the purposes of responding to my hypotheses, I concentrated on the matrices that contained the sum of the seven different types of interaction. Diagram 5.1 provides a visual representation of how the different matrices are dependent on each other. Table 5.2 - Data Matrices [ Please contact repository@tyndale.ca for Table 5.2 details ] Size (based on V and final size *) Mode Data Set Diagonal Zeroed Binary 88x 343 88x 333 88x 88 333x 333 88x 141 141x 141 2 1 Columns Rows Yes No Yes No 1 * V V V V V 2 V * V V V V V 3 V * V V V V V 4 V * V V V V V 5 V * V V V V V 6 V * V V V V V 7 V * V V V V 8 V * V V V V 9 V * V V V V 10 V * V V V V 11 V * V V V V 12 V * V V V V 13 V * V V V V 14 V * V V V V 15 V * V V V V 16 * V V V V V 17 * V V V V V 18 * V V V V V 19 V * V V V V V 20 V * V V V V V 21 V * V V V V V To facilitate data analysis, my analysis concentrates on three of these matrices, which are representative of the data (matrices 5, 8, and 20). These matrices allow me to analyze [ Page ] 110 direct interaction, shared relationships with third-parties, and interactions with third-parties that may lead to direct connections. The first matrix (5) is a direct connection matrix (it shows both direction of connection and number of connections). This represents actual connections between organizations. This matrix includes only my 88 respondents. They are listed both across the top of the matrix (columns) and down its sides (rows). Each box in the matrix contains a number which represents the number of different types of connections the respondent listed at the top of the matrix has with the respondents listed down the side. This matrix is directional. That is, respondent 8 claims to have 3 different types of connections with respondent 22 without respondent 22 acknowledging those connections. The second matrix (8) is a total affiliation matrix (which shows all shared connections to secondary organizations). That is, pairs of organizations are considered connected if they have both listed interaction with the same third-party on their surveys. The total affiliation matrix began as an 88 X 333 matrix. Rows and columns contain all 88 of my informants. Rows also include 245 additional organizations my informants listed on their surveys as organizations they have interacted with in the last year. As with the direct matrix, each cell indicates the number of interactions each respondent listed across the top of the matrix has with each organization listed down the cell of the organizations. However, some analysis procedures require a square matrix. Furthermore, I wanted a means of exploring indirect connections among my respondents. Therefore, I converted this matrix into an 88 X 88 matrix. I use this matrix in my analysis and refer to it as the total matrix. Each cell of this matrix lists the number of other organizations with which the organization in the column and the row both interact. For example, if the cell representing the intersection of organization 18 and organization 22 contains the number 5 this means that organization 18 and organization 22 both interact with 5 of the same organizations. These are not necessarily direct connections between the two organizations. For example, organization 8 and 22 may both give money to the crisis pregnancy centre. This common action would count as an interaction in the total matrix. Finally, my third matrix (20) is a connect affiliation matrix. That is, pairs of organizations are considered connected if they both list connections to the same third party which are likely to lead to interaction between the original pairs. For example, if Church A and Church B both list interaction with the Evangelical Fellowship of Canada on their [ Page ] 111 surveys this interaction has the potential to lead to a direct connection between Church A and Church B. The Evangelical Fellowship of Canada specifically attempts to get churches to interact with each other. This 141 X 141 matrix shows connections that are more likely to lead to direct connections. The connect matrix is exactly like the direct matrix except third-party organizations (i.e., not respondents) which are likely to lead to connections between other organizations have been added. For example, the primary interaction between the Girl Guides and most churches is that the Girl Guides use the church building for meeting space. The objective of the Girl Guides is not to connect religious organizations with each other. Therefore, the Girl Guides were not added to this matrix. In contrast, the Evangelical Fellowship of Canada has a specific objective of helping different religious organizations to interact with each other. Therefore, it was included in this matrix. This matrix tries to capture potential connectivity between my respondents. 6 INVESTIGATION In my dissertation I have set out to establish that Canadian evangelicals are no longer, if they ever were, completely isolated from each other and from the communities in which they operate. In this vein, I have aimed to determine how networks among evangelical churches influence the functioning of those same churches and their relationships with other organizations in the local community. Thus far, chapter 2 has explained how I identify churches as evangelical. Chapter 3 discussed some of the context of evangelical church interaction with each other and with their communities. In that chapter, I explain that we have evidence that evangelical organizations have interacted with their communities. We also have detail regarding certain types of interaction between specific churches and their communities. However, we have little information about the interactions between all the evangelical churches in a single community. Furthermore, we do not know how that interaction compares with other Protestant churches or how much interaction takes place with the larger community. Thus, chapter 4 developed some propositions concerning how Protestant churches interact in a single community based on the history of evangelical interaction with their communities and on what we know about how organizations interact with local networks. Chapter 5 described the study I designed to collect these network data. This chapter builds on the background described in chapters 2 and 3 and the data collected using the survey described in chapter 5 to respond to the propositions of chapter 4. That is, it applies the interview and survey data to my propositions to describe and discuss the specifics of church interaction in one local community. My discussion in this chapter is organized as a response to the research questions presented in chapter 1 and the propositions I developed in chapter 4. First, I explore how local churches interact and the reasons for that interaction. I examine how both characteristic similarity and network shape influence connectivity. I also compare and contrast the connectivity of evangelical and non-evangelical churches. Second, I look at how the network is structured and the roles that develop from that structure. I do this through a discussion of [ Page ] 112 [ Page ] 113 the costs and benefits of network interaction and the relationship between organization similarity, resource availability, and network involvement. Finally, I explore the relationship between the characteristics discussed in the first section and the network structure discussed in the second. That is, I explore how organizational similarity affects connectivity and vice versa. In this section, I also identify prominent cliques of religious organizations, and identify which organizations play central roles in local religious networks. Each section follows a pattern. First, I reintroduce the relevant proposition, briefly summarize the degree to which my proposition proved accurate, and describe my findings in detail. I end each section with a discussion question that explores the implications of these findings for the Canadian evangelical community. 6.1 Organization Characteristics and Connectivity The following story illustrates the dichotomy I expected to find among churches in Kitchener-Waterloo. Local Community Church (not their real name) is a church plant of a much larger church in Kitchener-Waterloo. It began its life in the church basement of this larger church. It also began its life with a specific agenda. It was to be a real community church. This meant that their main goal was to be operating in the community, with other people working in the community, to meet the needs of the community. In that vein, they chose to locate their church in an area known to have a high amount of physical need and they hired a pastor that shared their concern for the community. They began their work in the community with a community fair which provided games, and food for community members free of charge. Furthermore, as part of their community outreach strategy, they also fostered links with other organizations in the local area. This included four churches of different denominations, who were geographically located in the area but not actively involved in the community, the Salvation Army, the local community centre, and the local public school. This kind of connectivity was new to the other churches in the area. None of the other churches had ever had any contact with each other before this time and continue to have fairly limited contact with each other. Some churches are highly connected to other churches and the community and other churches are a world unto themselves. Thus, I expected to find both pronounced connections and pronounced divisions among the organizations I am studying, (1) Religious organizations in a local community will divide into multiple, densely connected networks with some overlap among these networks. [ Page ] 114 These may or may not be geographically proximate. This proposition is partially confirmed by the data.. To explore this proposition I turn to my network research. It provides data on which organizations are connected together and the nature of those connections. In this section, I examine how local churches are involved in the local network, the factors that contribute to connections and divisions between them and how different churches compare in terms of interaction with social services. I come to these some of my conclusions through a visual examination of the data. The relative position of nodes in the network diagrams are based on clarity: node position has no connection to the relationship between the nodes. Different shapes and colors represent different classes of organizations (as specified in the title and legend of each diagram). Thicker lines represent more interaction between nodes. Dashed lines represent less interaction between nodes. Arrowheads indicate the direction of the relationship. If there are no arrowheads then the matrix has been symetricized. That is, connections in either direction or in both directions are considered the same. The loss of direction data occurs in some matrix manipulations. 6.1.1 Network Involvement I expected that networks among local churches would be based in part on characteristic similarity. An example from one type of local Protestant will illustrate how characteristic similarity can lead to connectivity. Every year the small town of New Hamburg, just outside of Kitchener-Waterloo, hosts the Ontario Mennonite Relief Sale (see ). At this event everything from used clothing and traditional Mennonite food, to quilts and Crokinole boards are sold or auctioned off. All merchandise and services are donated with the result that this 2-day sale has raised over $10 million for the Mennonite Central Committee since it began in 1967. The organizers estimate that over 2000 people volunteer their time and talents each year to make the sale a success. What is interesting is the diversity of churches that come together to support the Mennonite Central Committee. The Mennonite Central Committee describes itself as “a relief, service and peace agency of the North America Mennonite and Brethren in Christ Churches” (). Its Ontario head office is in Kitchener. This encompasses churches with an incredible amount of diversity. At the Ontario Relief Sale there are Mennonites from churches with a liberal understanding of their faith and churches that are clearly conservative in faith. There are [ Page ] 115 Mennonites who dress all in black and Mennonite that are indistinguishable from the rest of Ontario culture. There are Mennonites selling traditional Mennonite food (e.g., fleish piroshki, rollkuchen) and Mennonites selling spring rolls. The point here is that Mennonites that are very different in theological, cultural, and ethnic characteristics can all work to a common purpose because they share a denominational heritage. This event and other characteristics oriented groups and activities (e.g., Pastor’s groups, evangelistic campaigns) led me to believe that there would be characteristic based networks in Kitchener-Waterloo. However, I did not know if organizational similarity was a sufficient characteristic for interaction. Diagram 6.1 - Direct Matrix (Evangelicalness)96 [ Please contact repository@tyndale.ca for Diagram 6.1 details ] As I expected there are multiple interactions between local religious organizations based on specific characteristics. Furthermore, there is evidence of different types of networks among local religious organizations. That is, there is evidence of connectivity based on some of the characteristics of individual churches. To illustrate this I examine denomination, evangelicalness, and size. ____________________________________ 96 Direct connections are listed connections with other informants. [ Page ] 116 Churches of similar ethnic and denominational background tended to be more closely connected together. In the case of direct connections, churches with certain specific characteristics tended to be more prominent in the network. Table 6.1 provides a comparison of the relative number of churches of different denominational types among all my informants and among those churches that have direct connections to other informants. This table allows a determination of which type of denomination tends to have direct connections. For example, Mennonite churches are 20% of my informants but represent 30% of those churches that have direct connections with each other. Furthermore, Mennonite churches played prominent roles linking different parts of the network (see Diagram 6.2). Mennonites were four of the top ten most central churches in the Direct matrix and three of the top ten most central churches in the Total matrix.97 No other denomination type represented as many connections. Furthermore, no other denomination type is as well represented in the Direct matrix and only Lutherans have as many representatives in the top ten in the Total matrix. However, Lutherans are 15% of my informants but represent only 7% of those churches that have direct connections with each other. This finding demonstrates that some organizational characteristics lead to network connectivity. However, it appears that there is something specific about Mennonites that leads to connectivity. That is, denominational similarity did not translate into direct connectivity as strongly for any other denominational group. ______________________________________ 97 I measure centrality in two ways. First, I measure centrality as the total number of interactions each church says that they have with other organizations (out-degree). Second, I use a calculation to produce a measure of the degree to which a church is situated between other churches in the network. The logic behind this measure is that an organization between two other organizations has the potential to act as gatekeeper for resources travelling between those organizations and thus may gain prestige from or exercise power over those organizations. [ Page ] 117 Table 6.1 - Denomination Type of Informant Organizations [ Please contact repository@tyndale.ca for Table 6.1 details ] Denomination98 Direct Informants All Informants Frequency Percent Frequency Percent Not Designated 1 2.2 1 1.1 Mennonite 14 30.4 18 20.5 Pentecostal/Charismatic 6 13.0 9 10.2 Baptist 6 13.0 10 11.4 United 3 6.5 7 8.0 Other Mainline 4 8.7 10 11.4 Other Evangelical 9 19.6 20 22.7 Lutheran 3 6.5 13 14.8 Total 46 100.0 88 100.0 Diagram 6.2 - Direct Matrix (Denomination Type) [ Please contact repository@tyndale.ca for Diagram 6.2 details ] ____________________________________ 98 Churches classified as Mennonite in my analysis include all denominations with an Anabaptist heritage (Mennonite Brethren, United Brethren, etc.). Churches classified as Pentecostal/Charismatic include all churches identified as charismatic in style (including Pentecostal, Independent Charismatic, and Vineyard churches). [ Page ] 118 In a similar manner, more evangelical organizations are better represented among direct connections than they were among respondents as a whole. I ranked respondents in four roughly equal categories based on evangelicalness.99 Approximately 49% of my respondents were in the top two ranked categories of evangelicalness. However, in the Direct matrix, 58% of organizations were in the top two categories of evangelicalness (Table 6.2). More evangelical churches are more likely to have direct connections than less evangelical churches (Table 6.2).100 Furthermore, if we compare the whole networks with networks that have all less evangelical churches excluded we find that these networks remain largely connected. Diagram 6.1 shows the direct connections among all my respondents. Diagram 6.4 shows connections among my respondents that are direct or are likely to lead to direct connections. Diagram 6.3 and Diagram 6.5 are similar to Diagram 6.1 and Diagram 6.4 except churches classified as less or least evangelical are exclude from the diagram. A comparison of Diagram 6.1 with Diagram 6.3 and a comparison of Diagram 6.4 with Diagram 6.5 demonstrate that removing less and least evangelical organizations from the network does not significantly affect the connectivity of the network. I conclude that, churches that are more evangelical are the main nodes that connected the different parts of the network. That is, the local direct connection network is largely an evangelical network (27 most and more evangelical direct connections versus 19 less and least evangelical direct connections). Table 6.2 - Evangelicalness/Frequency in Network of Informant Organizations [ Please contact repository@tyndale.ca for Table 6.2 details ] Evangelicalness Direct Informants All Informants Frequency Percent Frequency Percent Most Evangelical 14 30.4 22 25.0 More Evangelical 13 28.3 21 23.9 Less Evangelical 11 23.9 24 27.3 Least Evangelical 8 17.4 21 23.9 Total 46 100.0 88 100.0 ____________________________________ 99 Evangelicalness is determined in this chapter using the additive scale developed in chapter 2. 100 I divided organizations into four categories according to evangelicalness: most, more, less, least. For ease of discussion I refer to most and more evangelical organizations as more evangelical and less and least evangelical organizations as less evangelical. [ Page ] 119 Diagram 6.3 - Most and More Evangelical Churches in the Direct Matrix [ Please contact repository@tyndale.ca for Diagram 6.3 details ] Diagram 6.4 - Connect Matrix (Evangelicalness) [ Please contact repository@tyndale.ca for Diagram 6.4 details ] [ Page ] 120 Diagram 6.5 - More Evangelical Organizations in the Connect Matrix [ Please contact repository@tyndale.ca for Diagram 6.5 details ] Another possible basis for connectivity is size. Size characteristics play a role in network interaction, insofar, as different sized churches play different roles in a network. For example, organizations 1042, 1069, 1072, which are similar on size criteria (they have 100, 110, and 122 attendees respectively), also play bridging roles in the network (see circled churches in Diagram 6.6). However, there was no other clear clustering of churches of similar sizes, nor were churches of certain sizes excluded from the network. In the data, churches of all sizes had direct connections with each other. [ Page ] 121 Diagram 6.6- Direct Matrix (Attendance) [ Please contact repository@tyndale.ca for Diagram 6.6 details ] In addition to attendance, I have two other measures of church size: membership, and paid staff. These measures provide further data about the relationship between characteristic similarity and network connectivity. A visual observation of networks highlighting these three characteristics shows that the three measures do not appear to be directly related to each other (see Diagram 6.6, Diagram 6.7, and Diagram 6.8).101 High attendance does not necessarily lead to a high membership or a large number of paid staff members. This can be explained by the different levels of importance given to membership and paid staff by different churches. However, it is clear that certain size characteristics are more common in the direct network. Churches with three or more paid staff members compose over 50% of the direct connections network. _____________________________________ 101 Since I do not have size information for non-respondent churches, I have only used the Direct matrices in this section. [ Page ] 122 Diagram 6.7 - Direct Matrix (Membership) [ Please contact repository@tyndale.ca for Diagram 6.7 details ] Diagram 6.8 - Direct Matrix (Staff) [ Please contact repository@tyndale.ca for Diagram 6.8 details ] Nonetheless, while larger churches, in all three measures, tend to have more connections they are not necessarily more central. Table 6.3 and Table 6.4 show that there is [ Page ] 123 no necessary connection between size and number of connections. For example, there were 5 churches with 6-80 people in attendance that had one connection with another informant but only 4 churches with 275-835 people in attendance that had one connection with another informant (Table 6.3). The same pattern can be seen if we look at all the connections by informants shared with other informants. Table 6.4 shows that churches with a moderate attendance (85-150) are just as likely to share connections with the same third party as churches with a large attendance (275-835). Furthermore, only two of the nine churches with weekly attendance greater than 399 people had direct connections with other informant churches. Nonetheless, small size can reduce the resources available for interaction with the rest of the community. One part-time pastor explains, Ours is a small “pioneer” work. I am employed almost full-time elsewhere. Our finances, small number of people, and my personal schedule have somewhat limited our participation in some community church-related events. With time I expect our involvement will grow. (1052) It is possible, that after churches reach a certain size, they are less reliant on resources within the local area. First, larger churches are more able to obtain their resources internally. Second, larger churches are able to obtain resources from a larger geographic area. Thus, they may choose a resource because they consider it the best resource rather than because it is the only one available. Alternatively, pastors of larger churches may have more administrative tasks and may be less directly connected with the various interactions their church has within the community. Further research would be needed to determine how to explain the actions of large churches. For my purposes, it is sufficient to note that size does play a role in network connections but, as with ethnicity, denomination, and evangelicalness, size is not sufficient to explain the overall pattern of connections within an entire network (network shape). Table 6.3 - Attendance * Direct Degree Centrality Crosstabulation [ Please contact repository@tyndale.ca for Table 6.3 details ] Direct Degree Centrality Total 0 1 2-6 6 - 80 9 5 6 20 85 - 150 9 7 7 23 Attendance 160-258 8 6 6 20 275 - 835 13 4 4 21 Total 39 22 23 84 [ Page ] 124 Table 6.4 - Attendance * Total Degree Centrality Crosstabulation [ Please contact repository@tyndale.ca for Table 6.4 details ] Total Degree Centrality Total 0 - 4 | 5 - 14 15 - 27 29 - 43 6 - 80 6 9 4 1 20 Attendance 85 - 150 6 4 5 8 23 Attendance 160-258 4 4 5 7 20 275 - 835 3 4 8 6 21 Total 19 21 22 22 84 Nonetheless, church size data do tell us useful information about a local community. For example, the size data indicates that a larger number of people are attending more evangelical churches than less evangelical churches in Kitchener-Waterloo. This parallels some national data (Stackhouse 1993:1ff). The 43 most evangelical churches have a combined attendance of 10,157 with an mean attendance of 236 people (median attendance = 170). The 45 less evangelical churches have a combined attendance of 6888, with an mean attendance of 168 people (median attendance = 130). In this section, I have shown that there was only minor evidence of multiple networks based on characteristic similarity in the data. There were multiple networks. However, they did not connect based exclusively on specific characteristics. I did find multiple, small (three to four organizations each), networks in the data that were oriented around specific characteristics. However, these networks were overshadowed by the diversity of interactions in the larger network. I found that churches of all ethnicities, denominational affiliations, evangelicalness, and sizes were involved in local networks and interacted with each other. Nonetheless, some types of organizations did interact with each other more than others. Mennonites and organizations that are more evangelical had better representation in local networks than expected from their proportion of respondents. Furthermore, most of the largest churches in town are either Mennonite or evangelical churches (see Table 6.5). [ Page ] 125 Table 6.5 - Attendance by Denomination [ Please contact repository@tyndale.ca for Table 6.5 details ] Denomination Attendance Total 6-95 100-192 200-280 300-450 500-835 Not Designated 1 1 Mennonite 6 5 3 4 18 Pentecostal/Charismatic 2 3 1 2 8 Baptist 4 1 3 2 10 United 2 1 3 6 Other Mainline 3 2 3 1 9 Other Evangelical 10 2 3 3 1 19 Lutheran 1 10 2 13 Total 24 28 13 13 6 84102 The network created by Local Community Church (described at the beginning of this section) is representative of the community. It is a small network and relatively disconnected from the rest of the community. Network members have a shared interest in the community but are diverse in size, ethnicity, evangelicalness, and denomination. The implication is that networks between churches are functional rather than structural or positional. That is, churches interact because of shared goals not exclusively because they are geographically proximate to each other, have similar characteristics, or just because they are all churches. I will further analyze this later in this chapter. 6.1.2 Connections and Divisions There are groupings of organizations based on attribute characteristics (e.g., denomination, and evangelicalness). However, these groupings were not the same for different attributes. That is, groupings in the ethnicity network were different from groupings in the denomination network, which were both different from groupings in the evangelicalness networks. Thus, the attributes of local churches point to some reasons for interaction. However, they are only part of the story. The question driving this section is, What can we know about how religious organizations connect and divide by looking at the networks as a whole? A look at the whole network reveals data not available in an examination of organization attributes or direct connections. Downtown Church (not their real name, 1004) has a large church building in a convenient location. It has been highly involved in the local community in areas ranging ______________________________________ 102 Four churches did not provide this data. [ Page ] 126 from a council of churches concerned about the health of the city to a project to develop a housing complex. Furthermore, the church has hosted many local events from training seminars for a Franklin Graham crusade, to local Choirs. On their survey they list connections with 20 different organizations including some nationally known evangelical umbrella and missions organizations (e.g., Evangelical Fellowship of Canada, Navigators, Transworld Radio) and several local social services (e.g., Oasis, Kitchener-Waterloo Crisis Pregnancy Centre). However, the only churches they list on their survey are the several ethnic congregations that meet in their building. Furthermore, no other churches list connections with Downtown Church. This illustrates that a church can be highly involved in the community without having regular direct connections with other local churches. Thus, as is obvious, some churches interact with each other and some do not. This leads to my second proposition’s assumption that there will be multiple networks among local churches. Furthermore, as we have just discussed, characteristics are not sufficient to explain all connections and divisions among churches. This section looks at the larger network for an explanation of these observed connections and divisions. I analyzed the whole network of informant connections in three ways. First, I produced network diagrams of the Direct and Connect matrices. Second, I calculated different measures of structural relationships (e.g., density, and centrality). Third, I used a grouping method to look for structural patterns. 6.1.2.1 Network Diagrams: A Visual Inspection The first conclusion that can be derived from an observation of these network diagrams based on the data is that there are densely connected networks. However, in the Direct matrix they are small and in the Total matrix there is one big network not multiple networks. A visual examination of local church networks illustrates these conclusions. Diagram 6.1 provides a graphical representation of listed direct connections between informants. There are multiple networks in this diagram—usually centred around one well- connected organization. Diagram 6.9 is a graphical representation of indirect connections among all my informants. That is, it shows connections between organizations which share a connection with the same third party. Diagram 6.10 is a depiction of the geographic proximity of directly connected churches. A visual comparison of these diagrams shows that indirect connections produce a denser network than direct connections. Second, these [ Page ] 127 diagrams show that there is overlap between networks but not all churches are involved in them. Third, geographic proximity did not make a difference for connectivity across the entire network. Nonetheless, there were a fair number of dyads whose interaction may be based on geographic proximity (see Diagram 6.10). Diagram 6.9 - Total Matrix (Evangelicalness) [ Please contact repository@tyndale.ca for Diagram 6.9 details ] These diagrams show that interactions among local religious organizations involve connections between some organizations and divisions between others. Religious organizations interact based, in part, on shared identifying characteristics or sets of characteristics (e.g., ethnic background, denomination, evangelicalness). However, shared characteristics do not necessarily lead to interaction. Nor do they explain why some organizations are more connected to the larger network than others, or why very dissimilar organizations interact. [ Page ] 128 Diagram 6.10 - Direct matrix (Postal Code Similarity) [ Please contact repository@tyndale.ca for Diagram 6.10 details ] Legend: Different shapes and colours represent different postal locations. Colors indicate the same first three digits of postal code I have already discussed what we can learn from a visual inspection of the Direct and Connect matrices. That is, some types of churches are more likely to play linking roles than other churches (e.g., Mennonites, evangelicals) and there are connections that appear to be based on characteristic similarity. Yet, other connections provided no clear evidence of characteristic-based similarity among connected churches. For example, an analysis of informant city, location of listed connections, and postal code of informants found only limited evidence of interaction based on geographical proximity. Most local churches interacted with similar and dissimilar churches. I concluded that there was no one characteristic that was responsible for all network connections. However, while local churches have few direct connections with each other they have a high number of common connections with the same third-parties. For example, a visual analysis of the data from the perspective of which organizations share a connection to the same third-party organization shows a network so dense that a visual depiction of the network is unintelligible (see Diagram 6.9). There are few visually identifiable groupings of organizations in this matrix. Nonetheless, almost every organization that has any kind of connection to anyone is reachable by anyone in the network by no greater than four steps (see Diagram 6.11) and most connections are even closer. Furthermore, only 15 organizations are isolated from the larger network and six of those listed no connections on the survey. [ Page ] 129 Diagram 6.11 - Illustration of 1-Step in the Total matrix [ Please contact repository@tyndale.ca for Diagram 6.11 details ] As an extension to a visual examination of connections we can examine what type of organization is most prevalent in the Connect matrix. Table 6.6 provides evidence that a significant number of these shared connections are with social service organizations. As a result of the large number of duplicate connections to social service organizations, even an organization with a small number of direct connections to other informants can have a large ego network. The most dramatic example of this is organization 1040, which only has one direct connection with another informant but connects with 21 other local organizations. This gives it shared connections with 43 other informants. Thus, a visual inspection of network structure demonstrates that a shared concern for local organizations (particularly social services) is a significant point of connection among almost all local churches. Churches can potentially interact through shared interests in third- parties. A visual inspection of networks among local churches shows some divisions based on characteristics and significant connections based on common connections with third- parties. Furthermore, connections and divisions appear dependent on specific purposes. For example, organizations can interact for ethnic but not denominational reasons. Finally, points of connection between some organizations can be reasons for division between others. Likewise, churches can be connected on some levels and divided on others. [ Page ] 130 Table 6.6 - Connection Type [ Please contact repository@tyndale.ca for Table 6.6 details ] Type N Percent Social Service 233 36.2 Church Related 164 25.5 Missions Related 141 21.9 Education 40 6.2 Health Care 14 2.2 Political 0 0 Other103 51 7.9 Total 643 100 6.1.2.2 Structural Relations: Density and Centrality Measures of density can be useful to interpret both the role of an individual node in the network and the characteristics of the network as a whole. Ego centred measures of density measure the number of ties a particular node has as a percentage of the total number of ties in the network (Wasserman and Faust 1994; Borgatti et al. 1999:179ff). Whole network measures of density explore what percentage of potential ties actually exist in the entire network (Wasserman and Faust 1994:101). It does this by dividing the total number of ties by the total number of possible ties. These measures are useful to gain some insight into how much interaction takes place within the network. A comparison of my three matrices shows that very few of the possible linkages actually exist in the Direct and Connect matrices. Table 6.7 shows that density in the Direct matrix is less than 1%. This is because there is little formal direct interaction between churches in Kitchener-Waterloo (see 6.3.1.1). The Total matrix is significantly more connected. In the Total matrix of shared connections, almost 20% of all possible connections among listed organizations exist. This means that even though there are very few connections among local churches these same local churches share a large number of connections with the same third-parties. Thus, churches may disagree on theological values while cooperating on social values. ____________________________________ 103 “Other” may include organizations that could be classified in one of the other categories but whose type was not clear from their name. 131 Table 6.7 - Whole Network Density Direct Connect Total 0.0067 0.0061 0.1889 Density can be affected by other characteristics of organizations. However, increasing number of connections, and thus increasing density of direct connections do not depend on church size. Table 6.19 highlights that the 48 smallest churches in the survey had 707 connections and the 36 largest churches had 741 connections. Table 6.8 shows that small churches were more likely to have direct connections than large churches. Churches of all sizes have a significant number of connections with other organizations. The benefits of network connectivity are not dependant on size. This also means that a small church can be highly connected. Nonetheless, connections may serve different purposes for different churches. For example, Local Community church (approximately 120 attendees) has a philosophical conviction that working with other local churches is better than working alone, whereas, Downtown Community church (approximately 400 attendees) appears to have no such conviction. Table 6.8 - Church Attendance [ Please contact repository@tyndale.ca for Table 6.8 details ] Attendance Direct Informants All Informants Frequency Valid Percent Frequency Valid Percent 6 - 80 11 24.4 20 23.8 85 - 150 14 31.1 23 27.4 160 - 258 12 26.7 20 23.8 275 - 835 8 17.8 21 25.0 Total104 45 100.0 84 100.0 Evangelicalness also makes a difference to network structure. It makes a difference in terms of number of connections, among direct connections but not among connections with third-parties. An examination of the different centrality scores of my informants provides another measure of how the structure of the whole network affects connectivity. I am concerned here with degree centrality. This is the total number of interactions each church has. I use this ____________________________________ 104 Four churches did not supply attendance data. [ Page ] 132 measure because it has a relationship with density as both measures look at the number of connections each node has. A node with high degree centrality increases the density of the network. able 6.9 and Table 6.10 compare evangelicalness and degree centrality. Rows indicate how evangelical an informant is. Columns indicate the number of connections an informant has with other organizations. Thus, in Table 6.9 eight of the most evangelical informants have no connections with other informants. However, one of the least evangelic organizations has three connections with other informants. Table 6.10 compares evangelicalness and connections with common third-parties. In this table, nine less evangelical organizations have zero to four shared interactions with other informants. Table 6.9 - Evangelicalness / Direct Degree Centrality [ Please contact repository@tyndale.ca for Table 6.9 details ] Evangelicalness Direct Degree Centrality Total 0 1 2 3 4 6 Most Evangelical 8 7 3 2 1 1 22 More Evangelical 8 3 5 3 2 21 Less Evangelical 13 7 3 1 24 Least Evangelical 13 5 2 1 21 Total 42 22 13 4 4 3 88 Table 6.10 - Evangelicalness / Total Degree Centrality [ Please contact repository@tyndale.ca for Table 6.10 details ] Evangelicalness Total Degree Centrality Total 0 - 4 5 — 14 15 - 27 29 - 43 Most Evangelical 5 7 5 5 22 More Evangelical 3 4 7 7 21 Less Evangelical 9 5 5 5 24 Least Evangelical 5 6 5 5 21 Total 22 22 22 22 88 Table 6.9 shows that more evangelical churches are more likely to have direct connections than less evangelical churches. That is, connections among evangelical organizations are denser than connections among other types of organizations. However, Table 6.10 shows that churches with different levels of evangelicalness have similar levels of connectivity with local third-parties. More specifically, churches that are more evangelical [ Page ] 133 are just as likely to interact with third-party organizations as less evangelical churches. This demonstrates that what directly connects and divides churches is different from what indirectly connects them to and divides them from other organizations in the local area. That is, churches are directly connected, in part, by church characteristics such as evangelicalness. However, indirect connections are related to common connections to third-parties which are more likely to be social services or missions agencies than churches. In the previous section I demonstrated that divisions and connections are not necessarily coextensive across different characteristics. Furthermore, direct connections with churches are sparse but favour more evangelical churches. However, evangelical and non- evangelical churches interact with non-church third-parties to the same degree. In this section I have shown that measures of density and centrality provide a measure of the structural relationships in the matrix and a way of comparing these relations to each other. Examining measures of structural relations supports my visual evaluation of connections and divisions. This section provides further evidence that divisions and connections are only partially dependant on specific characteristics of organizations (e.g., denomination, ethnicity). The data shows that size and evangelicalness are not necessarily related to connectivity. In these measures, the most significant differences are not between organizations with different characteristics but between different types of connections. Different types of churches have the same type of connections with third-parties like social services. The example of one of those third-parties will make this clearer. St. John’s Kitchen operates out of St. John the Evangelist Anglican Church. The kitchen was started by the Working Centre a self-help employment resource centre located in downtown Kitchener. It was started because of the increased need in the downtown core. While its largest financial contributions come from individuals, it is supported by a large number of churches from very different denominational backgrounds. Nine of my informants list connections with the kitchen (1015, 1016, 1017, 1026, 1030, 1035, 1048, 1073, 1078). The primary type of tie was financial support. However, at least one of my informants also provided physical support, shared information, participated in common events, and/or provided volunteers to the kitchen. Significantly, only five of these churches were directly connected to other informants and none of them had direct connections to each other (see Diagram 6.1). Furthermore, they are diverse in terms of evangelicalness (two were most or more [ Page ] 134 evangelical, seven were less or least evangelical) and denomination type (e.g., Christian reformed, Baptist, Lutheran, Mennonite, Presbyterian, and United). This reflects my larger findings that Kitchener-Waterloo churches do not have many direct connections with each other but significantly different churches do support the same social services and have more indirect than direct connections. 6.1.2.3 Structural Patterns In this section I have been exploring the factors that lead to connections and divisions among local Protestant churches. Thus far, I have looked at the role that similar characteristics play in network connectivity. Another way to examine connections and divisions in the whole network is to use grouping algorithms to highlight structural patterns. Cliques, distance, and centrality provide useful information about actual or potential interaction and influence. However, such measures may not tell the entire story because they focus primarily on individual nodes and the nodes that connect to them. Influence is sometimes a function of structural position. That is, to fully understand the actual or potential influence of a given node one has to take into account the entire network. For example, a secretary may have the same centrality as a pastor however if that centrality is only with other secretaries the secretary has less influence than the pastor. In this case, what matters is not centrality but who you are central with. In order to get a measure of this kind of influence social network analysis has developed a series of measures of regular equivalence. One such procedure is a core/periphery model. The categorical core/periphery algorithm in UCINET 5 provides a way to organize a matrix of organizations to determine which organizations are more closely connected (i.e., they send and receive connections) and which organizations are not well connected (i.e., they neither send nor receive connections). The most closely connected organizations are called the core, and the organizations with few or no connections are called the periphery. In some cases I have added a third category called the semi-periphery. These are organizations with have more connections than the periphery but not as many as the core. Core/periphery models are useful because they identify which organizations in my network are likely to interact and which organizations are less involved with the local church network. From this I can examine if organizations in the core, those which have more opportunity to interact, are homogeneous in their attributes. I can also examine if core and periphery organizations are [ Page ] 135 different in their attributes. This will provided data with which to examine whether individual attributes or the larger dynamics of network interaction best explain network shape. The Direct matrix allows an evaluation of core vs. periphery in terms of actual connections. That is, core organizations are informants that have more direct connections to other informants, and periphery organizations are informants which have few connections to other informants. The Total matrix allows an evaluation of core vs. periphery based on similar relations to other organizations. That is, organizations will be in the core if they share a large number of connections with third-party organizations that interact with a large number of other informants. Finally, the Connect matrix allows us to evaluate core and periphery in terms of likely relations with each other. That is, core organizations have a large number of connections both with other informants and with third-parties that are likely to lead to direct connections with other informants. These three different matrices produce different core/periphery structures, as do variations of the grouping algorithm and different states of the original matrix (e.g., valued vs. binary matrices, directed vs. symmetrical matrices) (Borgatti and Everett 2000). To account for variations in how the procedure calculates core groups I ran the UCINET 5 procedure twice for each of my three matrices using two different grouping algorithms. I considered a church a part of the core group if they were classified as part of the core at least five of the six times I ran the procedure. A core/periphery examination of the survey data found a group of 25 churches that are most often in the core (see Table 6.11). That is, these organizations send and receive most of the connections in the network. The core group is diverse in church size and in evangelicalness. However, 17 of the 25 organizations in the core ranked high in the evangelicalism scale. That is, evangelical churches are more likely to be in the core because they interact with more other organizations.105 The core was composed of both very small (27 attendees) and very large (>800 attendees) churches. This confirms my visual analysis of the networks (see section 6.1.1) and my examination of structural relations (see Table 6.8). First, size is not necessarily related to connectivity. Second, churches that are more evangelical are a significant part of the local church network and thus evangelicalness is a factor in connecting local churches. This does not necessarily mean that evangelical churches are more active in the community. However, it does indicate that more evangelical churches ___________________________________ 105 Coreness does not necessarily correlate with power or influence. [ Page ] 136 are more concerned about interaction with each other than less evangelical churches. Thus, information about events would travel faster between evangelical churches than between other types of churches and evangelical churches already have significant connections that would facilitate cooperation towards common agendas and events. Table 6.11 - Organizations Most Often in the Core106 [ Please contact repository@tyndale.ca for Table 6.11 details ] Organization Attendance Evangelicalness 1010 80 1 1013 700 1 1060 85 1 1071 170 1 1072 122 1 1081 50 1 1083 835 1 1087 200 1 1001 37 2 1024 27 2 1026 120 2 1039 375 2 1042 100 2 1046 95 2 1050 180 2 1069 110 2 1080 205 2 1022 350 3 1025 170 3 1049 25 3 1055 70 3 1009 100 4 1015 150 4 1018 125 4 1078 175 4 Characteristics that connect organizations also have the potential to divide organizations. Network diagrams and measures of structural equivalence can identify connections and divisions not evident from characteristic-centred approaches to analysis. However, an examination of network diagrams, structural relations and structural equivalence shows few patterns with the exception of the difference in function between ___________________________________ 106 Evangelicalness is measured on a scale from “1” most evangelical to “4” least evangelical. [ Page ] 137 direct and shared connections and the greater connectivity among most and more evangelical organizations as compared to less and least evangelical organizations. Significantly, indirect connections do not appear to divide structurally according to evangelicalness, whereas direct connections do to a limited degree. That is, evangelical informants share connections with third-parties with non-evangelical informants and vice versa. However, some of the direct connections are based on evangelicalness. This implies that the differences between Protestant organizations may not be where I expected them to be or that differences in one aspect of a churches relationship with others does not necessarily lead to divisions among churches in other areas. That is, divisions between churches based on theology do not necessarily translate to divisions between churches based on connections with social services. 6.1.3 Connections with Social Services An opportunity to apply these insights is provided by the question: Are Evangelicals less connected with social services than mainline organizations? Thus far, I have described who is involved in local networks between religious organizations and some of the factors that connect and divide organizations within those networks. While I did find some of the expected divisions between directly connected churches, there appears to be no difference in terms of evangelicalness between organizations which interact with local service organizations. In this section, I explore the source of the false impression that evangelicals are isolated from the community (see also section 3.1). I begin by summarizing interviewee impressions of local interaction first described in chapter 3 to provide an illustration of common perceptions of evangelical church involvement in local communities. Following that, I provide more details on the interaction that is taking place and show that much of this interaction is with service organizations. I asked specifically about interaction between evangelicals and the community in the interviews. My informants had little knowledge of the role evangelical organizations play in the community or even if they play a role at all. Non-evangelical informants rarely mentioned evangelicals in the first round of questioning about community involvement. A local Baptist church was the only evangelical organization that was mentioned regularly and the only one where any of my non-evangelical informants were specifically aware of what it [ Page ] 138 was doing in the community (C214, C216, C234, and C235).107 It appeared to one informant that this Baptist church has specifically tried to be involved in the community and to have some degree of national influence (C235). One informant thought evangelicals must be involved in the community but did not know how (C232). From my informants’ perspectives, few evangelicals have any public voice in the community. It appeared to several informants that many of the roles evangelicals do play in the community were as individuals rather than as organizations (C234, C238). One evangelical informant suggested that evangelicals prefer it that way because they are aware of the negative views many people hold of evangelism (C234). Furthermore, this same informant suggested that evangelicals were wary of organizational control over their service activities. Nonetheless, there were hints in the interview data of community interaction by local churches. Although there was little awareness of evangelical churches interacting with each other, there was evidence of mobilization to stage community events (C220). Local organizations, supported by evangelicals and mainline churches alike, were mentioned often. The single most mentioned organization in the interviews was the soup kitchen, followed closely by the Kitchener-Waterloo council of churches, and the Mennonite Central Committee. St. John’s kitchen, associated with a local Anglican church and a local social service organization (The Working Centre). The Mennonite Central Committee is particularly interesting because the most often noted aspect of its work is its efforts at disaster relief rather than the involvement the Mennonite Central Committee has in the local community. The most visible organization in the community is not focused primarily on the community. Civic leaders are proud of the fact that local community members are having an impact beyond their own community. The pride seems on the level of the pride a community feels for a business that successfully expands beyond the local area but remains a significant part of the local community. The next most mentioned organization was the House of Friendship, a local drop-in and addiction rehabilitation centre, which was started by local Mennonites and is currently run by a former Mennonite pastor. Other organizations that received several mentions were the Interfaith and the Catholic family counseling services and the church colleges affiliated _____________________________________ 107 Specific interviews are referenced with the letter C followed by three numbers indicating which interview I am referring to. Where a statement could potentially lead to the identification of an informant the interview number is not listed. [ Page ] 139 with the University of Waterloo. Most other organizations were mentioned only once. However, all of these organizations are organizations that depend on church involvement to operate. Furthermore, many of them have clear involvement by churches with both conservative and liberal theologies (e.g., the Mennonite Central Committee). Thus, my informants were aware of church involvement in the local community. However, their false impression was that evangelical churches were generally not part of that involvement. The survey data provides more exact detail about what type of organizations local Protestant churches interact with. Eight of the top ten most mentioned organizations that provide the links between responding organizations are social service organizations. The top five organizations which connect pairs of informants and the five organizations which have the highest betweenness centrality all have a social service role of some kind (see Table 6.12). Betweenness centrality measures the degree to which an organization is between other members of the network (See Freeman 1979; Bonacich 1987; Scott 1991; Brass and Burkhardt 1992 for more detail). Thus, it is a better measure of potential influence than degree centrality. Centrality can be measured both for the organizations doing the connecting and for the organizations being connected to. This procedure can also contribute to an understanding of resource flow within networks and can point to locations of power in networks. Nonetheless, degree centrality remains useful. For example, 25% of informants listed the House of Friendship and almost as many listed Oasis / Ray of Hope a poverty relief organization and an organization which runs a drop-in centre and works with troubled young people respectively. In both cases, the most common connection was giving financial resources to these organizations, followed closely by volunteers and information. This pattern held true for most of the top ten most mentioned organizations (all of which were listed at least seven times). The exceptions were organizations such as Alcoholics Anonymous and the Girl Guides whose primary connection to churches is usually that they rent space. This pattern of involvement, although less obvious, was observed when I examined all organizations which received three or more connections. Of those 53 organizations: twenty-four were social services, ten missions related organizations, eleven were church related organizations, four were educational institutions, and four were other kinds of organizations. This pattern was much the same when I looked at all the connections [ Page ] 140 my informants listed (see Table 6.6). The clear difference is that churches and church related organizations are better represented in the larger data set. However, what is important for my argument is that, in most of these cases, churches that are more evangelical are as well represented as churches that are less evangelical. Table 6.12 - Centrality of Listed Organizations108 [ Please contact repository@tyndale.ca for Table 6.12 details ] Rank Organization Degree # of Mentions Organization Between- ness # of Mentions 1 2258 House of Friendship 136 22 2258 House of Friendship 11535 22 2 2363 Oasis/Ray of Hope 121 19 2363 Oasis/Ray of Hope 7956 19 3 2218 Food Bank 90 15 2218 Food Bank 6587 15 4 2246 Habitat for Humanity 75 8 2336 MCC 3144 12 5 2219 St. John's Soup Kitchen 70 9 2246 Habitat for Humanity 2622 8 6 2181 KW Crisis Pregnancy 69 12 2181 KW Crisis Pregnancy 2420 12 7 2336 MCC 66 12 2328 Mary's Place 2168 9 8 2328 Mary's Place 59 9 2439 Teen Challenge 2102 3 9 2153 Christian Horizons 56 4 2219 St. John's Soup Kitchen 2001 9 10 2235 YWAM 53 7 2039 Kitchener Gospel Church 1714 4 Thus, there is extensive evidence of interaction with the community on issues related to social services by all types of Protestant churches. The churches that connect with those social services cross the theological spectrum. Table 6.13 provides a snapshot of how evangelicalness affects connections with specific types of organizations. My 88 informants had 643 interactions with other organizations in the community. Table 6.13 indicates the evangelicalness of the informant doing the connecting and the type of organization being connected to. Thus, 47 of the connections made by the most evangelical informants were with social service organizations. Less evangelical organizations have more interactions with social services than more evangelical organizations. However, more and most evangelical organizations still have a significant amount of interaction with social services. _____________________________________ 108 Highlighted organizations are in both lists. [ Page ] 141 Table 6.13 - Contact Type * Evangelicalness Crosstabulation [ Please contact repository@tyndale.ca for Table 6.13 details ] Contact Type Evangelicalness Total Most More Less Least Social Service Count 47 56 61 69 233 % 23.5% 34.4% 44.9% 47.9% 36.2% Missions Related Count 75 41 14 11 141 % 37.5% 25.2% 10.3% 7.6% 21.9% Church Related Count 52 51 38 23 164 % 26.0% 31.3% 27.9% 16.0% 25.5% Education Count 14 11 7 8 40 % 7.0% 6.7% 5.1% 5.6% 6.2% Health Care Count % 1 .5% 3 2.2% 10 6.9% 14 2.2% Other Count 11 4 13 23 51 % 5.5% 2.5% 9.6% 16.0% 7.9% Total Count 200 163 136 144 643 % 100.0% 100.0% 100.0% 100.0% 100.0% Few of the most connected local organizations operate outside of the local area, nor do they primarily work on those social justice issues which tend to separate churches theologically (racial reconciliation, economic oppression, etc.). Although, most of the informant-listed organizations have church connections only two of the top ten most listed organizations have any sort of explicit evangelistic objective and even those two organizations have social service objectives as well (i.e., Oasis/Ray of Hope, Youth with a Mission). Responding to poverty and personal difficulty (food banks, youth drop-in centers, etc.) are issues that sometimes transcend theological identity (see Table 6.14 and Table 6.13). Nonetheless, issues that are more controversial can divide organizations theologically. For example, the Kitchener-Waterloo Crisis Pregnancy Centre and Mary’s Place (a home for unwed mothers) appear to be complementary services. Yet, none of the organizations that support the Crisis Pregnancy Centre supports Mary’s Place and vice-versa. Furthermore, the religious organizations that support the Crisis Pregnancy Centre are primarily conservative theologically and the organizations that support Mary's Place are primarily liberal theologically (although organizations of all theological stripes did support both organizations). [ Page ] 142 Table 6.14 - Most Listed Organizations [ Please contact repository@tyndale for Table 6.14 details ] Organization Type 109 Number of Mentions Mentioned by Evangelical Churches Mainline Churches House of Friendship S 22 V V Oasis\Ray of Hope S 19 V Food Bank S 15 V V KW Crisis Pregnancy Centre S 12 V V Mennonite Central Committee S 12 V V St. John’s Soup Kitchen S 9 V V Mary’s Place S 9 V V Habitat for Humanity S 8 V V Alcoholics Anonymous S 7 V V Youth with a Mission M 7 V March for Jesus M 6 V River of Life Church C 6 V V Anselma House S 6 V V Evangelical Ministerial C 6 V KW Council of Churches C 6 V V KW Ministerial C 6 V V Emmanuel Bible College E 5 V Girl Guides S 5 V V Navigators M 5 V Why Book M 5 V Social service organizations, insofar as they need money and other resources from the churches, are dependent on those churches. In this capacity, many of these social service organizations have the potential to connect very diverse churches. Churches that disagree significantly about theology and worship style support the same social service organizations. Of the most central social service organizations, the majority of them had both evangelical ____________________________________ 109 Legend: S - Social Service, M - Missions, C - Church related, E - Education [ Page ] 143 and non-evangelical connections (see Table 6.14). The one exception, Ray of Hope, had only evangelical connections. Thus, the interviewees accurately identified the most prominent local social services but did not realize that those social services were supported by evangelical as much as by non-evangelical churches. However, the survey data provides ample evidence that evangelicals are almost as connected to social services as any other churches. 6.1.4 Summary: How Do Religious Organizations Interact? Thus far I have determined that organizational characteristics play a role in organizational interaction but do not dictate all interaction. I have also noted that there are distinct groupings of organizations. However, those groupings are based on multiple criteria. Finally, I noted that indirect connections through social services provide more potentially connectivity than direct connections among churches. The direct connections among the churches I studied are sparse. Nonetheless, my examination of direct connections found some evidence of multiple densely connected networks. Interaction based on ethnicity, denomination type, evangelicalness, and size does take place, but such characteristics of churches are not the only basis on which they interact. None of the characteristic data I collected were sufficient to explain all network connectivity. Furthermore, when churches interacted with third-parties the data did not show distinct networks with connections based on church characteristics. However, there were significant network connections with local service organizations, which have the potential to facilitate resource sharing among a large number of indirectly connected churches. These connections can help a large number of different churches accomplish common social service goals. In addition, there is evidence of network alignment based on similarity of connections. For example, churches did divide according to evangelicalness when connected with organizations that concentrated on activities directly related to evangelism or issues of the body. Yet, churches divided and connected in different ways when connecting to other types of organizations. For example, evangelicals and mainline churches interact with many of the same social service organizations. For specific groups of churches this interaction extends to working together on common events. One Pastor explained that, It seems that more and more, denominations are no longer pushing us apart. It seems there is a real interest toward majoring on the majors, ways to share the Gospel. When we had a Franklin Graham Mission two years ago it was an opportunity that seemed to show how many of us were working at the same [ Page ] 144 thing for God. Things like March for Jesus and joint New Years celebrations are doing similar work of uniting the church universal. (1046) Thus, similarity of characteristics among organizations cannot explain all network connections. There are minor, multiple networks with some overlap when interacting directly with each other. However, they are only densely connected when interacting with third- parties. What connects and divides organizations depends on whether we are exploring direct or indirect connections. Direct connections are related to organization characteristics and information flow. Indirect connections are more closely related to common interaction with third-parties (primarily social service organizations). Furthermore, social service organizations play a significant role in the density of the indirect connections network and in providing potential points of contact for diverse churches. 6.2 The Benefits and Costs of Intrachurch Network Relations Knowing what connects and divides organizations is the first step in understanding the benefits and costs of network involvement. Connections are conduits for resource exchange and divisions restrict that exchange. For organizations, benefits and costs are often related to resource exchange. My second proposition explores the pattern of resource exchange observed in networks among local religious organizations. The main research question in this section is, why do religious organizations interact? In 2002, Local Community church (not their real name) decided is was a good idea to run a second community carnival. They were already involved with a council of local churches and social services agencies. They approached that council with the idea and the council collectively decided to stage the carnival. In the end, the carnival involved several different churches and the local community centre. It include games and rides for children, food for sale inexpensively, a puppet show, and lots of free prizes. Local Community church received significant benefits from their network connections. They were able to secure a physical location for their carnival, a significant number of volunteers, and a significant number of prizes for little or no cost because they had fostered connections with other organizations in the local community. Nonetheless, those connections did have a cost. Local Community church needed to take time to explain their vision to each of the different organizations involved in the project, they had to negotiate the amount of involvement each organization had and, when one organization was accidentally left off the flyer explaining the event to attendees, they had to [ Page ] 145 produce new flyers at an additional financial cost. Furthermore, it was hard to measure if the cost of staging the carnival translated to benefits to the church in terms of community good will or increased participation in the church. Thus, for the churches involved in this carnival their network involvement led to both costs and benefits. While it is hard to measure whether the benefits out weighed the costs it is clear that Local Community church perceived that to be the case. Given that connections exist, and that maintaining connections requires resources, we can assume that religious organizations get some perceived benefit from this connectivity. My assumption is that listed connections represent intentional relationships and, although they do not include all a church’s relations, likely represent relations that are important to that church.110 These relations may, for example, provide resources or reflect the beliefs and values of the organization. Thus: (2) Networks among evangelical organizations provide the benefits of social support. This provides encouragement towards common goals, which outweigh the costs of time and resource commitment, which are required to maintain an organization’s links with the larger network. In this section, I discuss what those costs and benefits are and whether the benefits are sufficient to outweigh the costs of network connection. 6.2.1 Costs All forms of resource sharing have some cost. I did not specifically ask about the cost of network involvement. Thus, as proxies for cost I will examine three resources which demonstrate how network involvement requires time and resource commitment. These resources are time (needed, for example, to exchange information), money (financial aid), and people (volunteers). Sharing information requires time. It can have a small cost, as it may be a byproduct of other connections (e.g., local ministerial, church publications, or church organizations). Yet, intentional interaction to share information does have a cost—even if that cost is the time needed for a brief telephone conversation. For example, to connect with other organizations takes resources away from local church business. Some churches are too busy with the business of running their church to pay the cost of network connectivity: [Ours] is also a congregation in transition and just keeping momentum ____________________________________ 110 Although people often omit existing connections when filling out surveys, studies show that the connections they list reflect social structures and intense connections (Sudman 1988; Brewer and Yang 1994). [ Page ] 146 focused on calling a pastor, securing financial stability, establishing an infrastructure, has taken considerable faith and energy from a few committed lay people. (1077) These types of activities are vital to the viability of the church and to neglect them would be to threaten the church’s existence. Nonetheless, they may reduce connectivity. This church only had three connections None of these connections were with other local churches Other evidence that even sharing information has a cost is that, when asked, my informants did not claim to share information with everyone with whom they interacted. Furthermore, informants sometimes indicated that it was easier to interact with some types of organizations than others, Feel more connected to objectives of mainline churches than conservative Evangelical churches. Very open to Inter-Mennonite and ecumenical cooperation. Our church finds to difficult to connect with the conservative theology of much of evangelicalism and its emphasis on rigid forms of belief, theology, and practice. Do not have sympathy for the “rightist” political agenda of conservative evangelicalism. (1078) Nonetheless, there was a significant amount of interaction among my informants. This indicates that some churches believe that the time cost was worth the effort of network involvement. For example, over two-thirds (73%) of direct connections among informants involved the sharing of information with other informants. Table 6.15 divides the 46 direct connections my informants had with each other according to type. Table 6.16 provides the same data for all connections. Almost 51% of interaction with all other organizations involved the sharing of information of some kind. Information is shared among directly connected informants at a different level than among the network as a whole. Nonetheless, in both cases the time needed to share information was worth the cost over half the time. Table 6.15 - Type of Interaction among Directly Connected Informants [ Please contact repository@tyndale.ca for Table 6.15 details ] Type of Interaction Interaction No Interaction Total N Percent N Percent N Percent Physical Support 17 37.0 29 63.0 46 100.0 Information 34 73.9 12 26.1 46 100.0 Events 40 87.0 6 13.0 46 100.0 Finance 5 10.9 41 89.1 46 100.0 Volunteers 20 43.5 26 56.5 46 100.0 Staff 5 10.9 41 89.1 46 100.0 Total 121 43.8 155 56.2 276 100.0 [ Page ] 147 Table 6.16 - Type of Interaction for all Connections [ Please contact repository@tyndale.ca for Table 6.16 details ] Type of Interaction Interaction No Interaction Total N Percent N Percent N Percent Physical Support 207 31.9 442 68.1 649 100.0 Information 330 50.8 319 49.2 649 100.0 Events 261 40.2 388 59.8 649 100.0 Finance 380 58.6 269 41.4 649 100.0 Volunteers 230 35.4 419 64.6 649 100.0 Staff 70 10.8 579 89.2 649 100.0 Total 1478 2416 3894 While all sharing of information has some cost, that cost varies depending on what information is shared. For example, information about an already publicly advertised event may be valued differently than information about new financial resources. Furthermore, sharing of information is not the most common type of interaction in all situations. Table 6.17 provides a measure of which type of organization receives which type of interaction. This table shows that church related organizations receive the largest number of information connections. However, they receive more of the closely related event connections. The sharing of information about events has benefits for churches because successful church sponsored events can increase their church’s profile in the community and bring collective benefits to all involved churches. Nonetheless, the time required to communicate which events are happening is costly and does not guarantee a return on investment. Table 6.17 - Type of Interaction among all Informants * Organization Type Crosstabulation [ Please contact repository@tyndale.ca for Table 6.17 details ] Type of Interaction Organization Type Social Service Missions Related Church Related Education Health Care Other Physical Support 74 21 62 10 5 35 Information 95 60 115 27 5 28 Events 47 45 125 18 4 22 Finance 166 109 51 25 9 20 Volunteers 89 42 63 20 2 14 Staff 15 17 24 6 3 5 [ Page ] 148 Sharing information has a relatively small time cost. However, maintaining a long- term relationship with another organization can have a significant time cost. Only four of th connections, of all types, in the Direct matrix were reciprocated (1001:1042, 1001:1055, 1039:1071, and 1046:1080). This lack of reciprocated relationships, one measure of time, indicated an unwillingness to pay the cost of more permanent connections. No financial ties existed in reciprocated relationship among directly connected informants (see Table 6.18). Information and events were the primary points of contact. This provides further evidence that information and event ties are relatively low cost ties. Table 6.18 - Type of Connection in Reciprocated Interaction [ Please contact repository@tyndale.ca for Table 6.18 details ] Connection Physical Support Infor- mation Events Volun- teers Send Receive 1001 1042 V V V 1001 1055 V V V 1039 1071 V 1046 1080 V V 1042 1001 V V 1055 1001 V 1071 1039 V V 1080 1046 V V V V Another significant cost to network involvement is money. Financial ties were rare among informants (11% of interactions in the Direct matrix). However, 59% of all interactions with contact organizations involved finances (see Table 6.16). The main difference here is that all the informants were churches, whereas, the majority of all financial connections were with social service or missions organizations (see Table 6.17). Table 6.17 further shows that financial ties were the most common type of interaction when my informants interacted with social service, mission, or health care organizations. Indeed, there is evidence that churches play a substantial role in supporting social services (Wineburg 1990; Wineburg 1993). This highlights an important difference between network connections among churches and network connections with other types of organizations. Many churches [ Page ] 149 are self-supporting (i.e., they pay their bills through the donations of their attendees or through support from their own denomination, they do not receive government grants). When churches do give financial support to other churches it is usually because they share the same denomination and it is customarily accomplished through denominational redistribution of funds—not through direct connections between churches. Financial aid to other types of organizations is more likely to serve specific ministry goals (e.g., missions, social service). Thus, the benefits of giving to these organizations are related to organizational goals (e.g., evangelism, feeding the poor). The cost of financial ties may depend on who is receiving the aid. Financial aid, which may appear to be a high cost interaction, can actually have a lower cost than sharing information or physical resources like volunteers. I expected that, for a small struggling church any kind of giving outside of its own congregation would be significant, whereas, for a wealthy church, giving money may have almost negligible costs (Stonebraker 2003). However, a comparison of tie type by church size showed little difference between smaller and larger churches (see Table 6.19). Costs may be limited for all churches or the benefits of financial ties may outweigh the costs to a sufficient degree that it makes sense for churches of all sizes to maintain financial ties. Nonetheless, financial ties do have a cost that must be compensated for. Table 6.19 - Attendance * Tie Type Crosstabulation [ Please contact repository@tyndale.ca for Table 6.19 details ] Attendance (# Churches) Number of Ties Physical Support Infor- mation Events Finance Volunteers Staff Total # % # % # % # % # % # % # % 6 - 170(48) 92 13.01 156 22.07 142 20.08 179 25.32 100 14.14 38 5.37 707 48.8 175 - 835(36) 107 14.44 167 22.54 114 15.38 196 26.45 125 16.87 32 4.32 741 51.2 Total 199 13.74 323 22.31 256 17.68 375 25.90 225 15.54 70 4.83 1448 100 Volunteers, like financial ties, are a category of connection which appear to imply significant cost. Over 35% of all interactions and 16% of all ties involved volunteers (see Table 6.20 and Table 6.19).111 Furthermore, if an individual is volunteering their time outside of the local church that is time not available to spend on the business of the local church. However, as with information and finances, the cost can vary with church situation. I ______________________________________ 111 Each of the 649 interactions between organizations can involve multiple ties. There were 1448 ties in these 649 interactions. [ Page ] 150 expected that the cost, given access to a large number of volunteers, might be negligible to a large church whereas the cost to a small church may be significant. However, as with finances there was little difference between smaller and larger churches in the relative number of ties involving volunteers. Table 6.19 compares attendance with tie type. It shows that there is almost no size-based difference in the type of interactions that churches have with other organizations. In addition, such resources can be benefits rather than costs depending on where the organization sits in the network. A volunteer whose actions provide access to resources may more than compensate for the cost of that individual’s lack of involvement in their church. The costs of network involvement are dependent on the characteristics of individual organizations and particular network characteristics (denominational interaction, organization size, etc.). These costs include time to share information, financial costs, and the cost of sharing volunteer time. Over half of all interaction and over two-thirds of direct interaction involved the exchange of information. Surprisingly, almost 60% of all interaction involved the exchange of finances. However, only 11% of direct interaction involved finances. This indicates that different types of connections have different kinds of interaction. In this case, churches are more likely to give money to social services than other churches. Finally, only a minority of ties involved volunteers. However, the ties that did exist were almost equally distributed between different sizes of church. Thus, the costs of network connectivity, like network connectivity itself, depends on the relationship. Not all network relations are equal. However, it is clear that the costs of network involvement are not necessarily dependant on the characteristics of the churches involved. For example, indicators of cost are similar for different sized churches. In the next section, I look at how the benefits of network involvement outweigh the costs. 6.2.2 Benefits The main purpose of interorganizational networks is to exchange tangible and intangible resources of various kinds. A significant benefit to any organization is reciprocal social support and encouragement to common goals within its network. This social support and related common goals outweighs the costs of network involvement (time, money, people, etc.). [ Page ] 151 I define social support, for my purposes, as the provision of resources to accomplish organizational goals.112 In this section, I discuss how the provision of social support outweighs the costs of network involvement. I also discuss how network benefits depend on the organizations involved with the network. This is particularly an issue for churches that do not have the size or denominational structure to accomplish their goals on their own. Such organizations are reliant on their network involvement to accomplish some of their goals. In this vein, I explore the source of resources for evangelical churches. 6.2.2.1 Identifying Social Support Social support involves reciprocity of some kind. That is, organizations give to another organization because they get something from that organization. This could be actual physical resources, or less tangible resources like acceptance, and shared resources like common identity. I identify the existence of social support if some reciprocity exists, if physical resources are exchanged, or if there is some interaction based on common identity. There is an almost complete absence of reciprocity among my informants (only four instances, see arrows on the network diagrams). Only 50 of 7744 possible direct connections exist, involving only 46 different organizations. Lack of reciprocity may indicate the degree to which pastors are not aware of all the linkages they have with other organizations. Thus, differences among actor networks may have more to do with the pastors’ awareness of interconnections than with whether or not those interconnections actually exist. Linkages may be haphazard and unplanned (which may explain the lack of awareness of any sort of coordinated network despite clear links between organizations). Nonetheless, lack of awareness or intentional connection indicates the priority that pastors’ place on connecting with other religious organizations. Alternatively, lack of connectivity may be a function of the organizational structure of the church. One Pastor explains, Basically we are a church community slowly changing from an authoritative paradigm to a participatory, democratic one in which preferably most persons have a say in the decision making process. We attempt to be inclusive of all persons as well as intergenerational. (1016) Pastors of churches that have already undergone this shift would be less likely to be aware of all the connections their church has with other local organizations. ____________________________________ 112 Social support has been studied extensively in the network literature (Wellman 1992; Galaskiewicz and Wasserman 1993; Ellison and George 1994; Gulati 1995; Wellman and Gulia 1999). [ Page ] 152 Despite the lack of recorded reciprocity on the survey, there is evidence of unrecorded reciprocity. Twelve informants report that they supply physical support for another informant without that other informant listing any kind of connection with the original informant. Yet, at the very least, information must be exchanged to receive physical support. This may indicate that the issue of receiving support from other churches is not part of many church’s self-understanding. Alternatively, it may indicate relationships that have become so rote as to be invisible. Furthermore, the lack of reciprocity does not mean that resources are not traded. For example, resource exchange may be characterized by a pattern of generalized exchange (Uehara 1990). That is, organizations provide resources to other organizations in their network knowing that they will eventually receive resources from the network even if not from the same organization they helped. For example, A provides resources to B, B provides resources to C, and C provides resources to A. Churches which share a denomination may be willing to supply resources to other churches in the denomination believing that the denominational affiliation will supply resources they need when they need them or that helping another church is a contribution to common goals. In one case, a church from Mississauga came and ran the Vacation Bible School program for a church of the same denomination in Kitchener-Waterloo without requiring, or directly receiving, any resources in return. That said, there is so little recorded reciprocity that it is not possible to determine if some types of churches reciprocate support any more than other types. There may be no acknowledged reciprocity, but resources are shared and thus interconnectivity provides some benefit. One of those benefits is physical resources. The evidence for sharing physical resources is extensive. The types of interaction I asked about all provide tangible resources. Four of those ties involve the provision of physical resources (i.e., physical support, finance, volunteers, and staff). Three of these four types of ties were checked off at least 30% of the time (see Table 6.20). A financial tie existed in 58.6% of cases. Given the prevalence of financial connections their cost may not be as high as I originally expected. In a relatively well-off community, it may be easier to give an organization money than to contribute time. Staff, a resource which requires extensive co-operation to share, was shared in 10.8% of the ties. Some sharing of staff is to be expected as the Christian social services likely draw most of their staff from local Christian churches. Thus, the cost of resource exchange can also be a [ Page ] 153 benefit as network involvement can provide social support through the exchange of physical resources. Table 6.20 - Frequency of each Type of Interaction [ Please contact repository@tyndale.ca for Table 6.20 details ] Type of Interaction Included N Percent of Possibilities Physical Support 207 31.9% Information 330 50.8% Events 261 40.2% Finance 380 58.6% Volunteers 230 35.4% Staff 70 10.8% Totals 1478 38.0% Nonetheless, physical support is not shared equally across the network. Most of these resources are exchanged between churches and social service organizations not among churches (see Table 6.13). This shared support of the same social services is a form of support for the larger goals of the churches involved in the network. That is, if one church wants a particular social service to thrive and other churches also contribute resources to the social service they mutually contribute to each others goals and provide mutual benefits to compensate for the costs of network involvement. Social support is also related to how churches sustain each other’s organizational identity and goals. Thus, some of the benefits of a network are related to organizational identity and what organizations are trying to accomplish with their dispersion of supportive resources. To evaluate this I took a closer look at how important it was to my informants to support certain activities. On a scale of one (very important) to seven (not at all important) 76% of informants rated the importance of supporting local social agencies 1, 2, or 3 (see Table 6.21). In fact, at least 76% of informants supported global social agencies (77%), local missions (90%), global missions (96%), combating moral evil (85%) and combating social evil (89%) by ranking them 1, 2, or 3 on their surveys. This can be contrasted with my informants’ support for Theological Seminaries (74%), Bible Colleges (58%), Inter-religious Dialogue (41%), Universities (35%), and Pacifism (33%). [ Page ] 154 Table 6.21 - Importance of Local Social Agencies [ Please contact repository@tyndale.ca for Table 6.21 details ] Importance to Identity Frequency % More Important 1 18 22.5 2 25 31.3 3 18 22.5 4 8 10.0 5 8 10.0 6 3 3.8 Less Important 7 0 0.0 Total 80 100.0 Furthermore, associating with other religious groups was not considered very important to these churches’ identities (see Table 6.22). The 64% of my informants who associate with evangelicals is surprising. It suggests that evangelicals are not as isolated from the larger community as I first suspected. Alternatively, it may be a representation of the significant number of evangelical churches in the community. Nonetheless, more evangelical organizations were less likely to consider association with Roman Catholics, mainline churches, and Non-Christian religions as important to their identity than were less evangelical churches (see Table 6.23 as an example). Table 6.22 - Associate with (ranked 1, 2, or 3) [ Please contact repository@tyndale.ca for Table 6.22 details ] Religion Percentage Evangelicals 64 Mainline Churches 46 Roman Catholic 28 Charismatic Movement 26 Non-Christian Religions 17 [ Page ] 155 Table 6.23 - Associate with Non-Christian Religions [ Please contact repository@tyndale.ca for Table 6.23 details ] Importance to Identity Evangelical Total (N) Most More Less Least Important 1 % 66.7 33.3 100.0 (3) 2 % 100.0 100.0 (3) 3 % 14.3 42.9 42.9 100.0 (7) 4 % 44.4 22.2 33.3 100.0 (9) 5 % 36.4 27.3 36.4 100.0 (11) 6 % 22.2 22.2 44.4 11.1 100.0 (9) Least Important 7 % 52.8 30.6 13.9 2.8 100.0 (36) Total % 28.2 26.9 24.4 20.5 100.0 (78) These data indicate that providing direct social support by identifying with all other local Protestant churches is not that important for local churches. However, local Protestant churches are very closely identified with social service and missions agencies and churches may be more likely to associate with other churches which share their theological identity. This close identification with local social service and missions agencies explains the willingness of local churches to provide finances, information, and physical support to these organizations. There is extensive evidence of social support in Kitchener-Waterloo. This can be seen both in the exchange of resources and in identity based interaction both of which help churches accomplish their organization goals. Exchange of resources allow churches to get access to the information, physical resources, or physical support that makes their various programs work (e.g., access to training seminars, Sunday school curriculum, volunteers). Identity based interaction contributes to a church’s goals in a different manner. Shared identity in terms of support for common organizations helps churches determine who to work with and which other organizations are likely to share their goals and thus resources can be pooled to accomplish more together than they could as single churches. That said, the data also shows that churches are not taking advantage of the potential for vast areas of co-operation. That is, there is much common social support of social service organizations but little conscious identification with other local churches that also support [ Page ] 156 those organizations. Furthermore, that social support is confined to particular goals and particular organizations. Local churches function separately with few direct connections with each other and almost no reciprocity. The lack of direct connectivity among my informants suggests that organizational goals are centred on an informant’s own churches. The network of relations may help an organization find financial and physical resources but the organization’s goal may be to expand its own organization not necessarily to contribute to the larger network. Nonetheless, the benefits of network involvement must outweigh the costs of the resource exchange necessary to maintain the network. If this was not the case then no network would exist at all. The implication for local churches is that there are significant areas of cooperation that could be taken advantage of. This cooperation may reduce the resources each individual church needs to contribute to the project while, at the same time, accomplishing the project more effectively than any one church could by themselves. An example of this is the Franklin Graham crusade that local evangelical churches supported in 1996. No one church could have handled all the logistics or supplied all the volunteers for that event. However, with each church providing some of the volunteers, a city-wide crusade was staged that spread converts through the various churches in the city. Another implication of the data is that there are differences among my informants in terms of which organizations they identify with. The next section examines whether these differences affect the organizations evangelicals get their resources from. 6.2.2.2 Do evangelical organizations get most of their resources from within evangelical networks? Organizational characteristics, such as evangelicalness, can affect network involvement and the benefits obtained from that involvement, in that, organizations may share resources because they respect certain identity markers (e.g., evangelicalness, etc.). Identity markers are a benefit in that they reduce the cost of interaction. The expectation is that common identity makes interaction safer because values and goals are likely to be similar. One manifestation of this is the Shepherd’s Guide: Christian Business and Ministry Directory. This publication lists local churches and Christian organizations but also lists businesses ranging from real estate to jewelry. The Guide explicitly promotes itself as a way for Christians to encounter other Christians. For example, one of the ways potential advertisers are supposed to be able to determine if they want to advertise in the Shepherd’s [ Page ] 157 Guide is if they agree with the statement “I would like to cross paths with Christians more throughout the week” (Alexander and Alexander 2000:48). The assumption is that there is a benefit to having business dealings with other Christians. Such benefits might include shared values, a belief that Christians are more trustworthy, and support for fellow believers. The expectation that common religious identity reduces the cost of interaction led me to expect that most resource exchange would take place between organizations with similar identity markers. The existence of many direct connections among most and more evangelical churches in the community and the high degree of association with evangelicals also points to this conclusion (see Diagram 6.1 and Table 6.22). To evaluate this expectation I examine whether or not there is an evangelical network and where evangelical churches get their resources. More evangelical churches are more likely to have direct connections than less evangelical churches (see Diagram 6.1). Therefore, we can assume that there are greater benefits to local network involvement for more evangelical organizations than for less evangelical churches. However, there was no evidence of an exclusively evangelical network. Evangelical organizations had many connections in common with non-evangelical organizations. Furthermore, they also support social services that are actively supported by different types of organizations. My informants, both more and less evangelical, supported many of the same local religious organizations. This leads to a large number of indirect connections between churches (e.g., for information and events) which were not oriented around evangelicalness. Nonetheless, there were differences between evangelicals and non-evangelicals. For example, there is evidence of evangelical churches supporting different types of organizations than non-evangelicals. For example, evangelical organizations are more likely to support missions organizations than non-evangelical organizations (see Table 6.14). Table 6.24 shows how churches with different levels of evangelicalness favour different types of support. These data represent all the connections that my informants listed on their surveys. [ Page ] 158 Table 6.24 - Support * Evangelicalness Crosstabulation [ Please contact repository@tyndale.ca for Table 6.24 details ] Type of Support Evange icalness Most More Less Least Count Row % Count Row % Count Row % Count Row % Physical Support 37 17.9 35 16.9 71 34.3 64 30.9 Information 116 35.2 89 27.0 78 23.6 47 14.2 Events 95 36.4 68 26.1 64 24.5 34 13.0 Finance 137 36.1 88 23.2 68 17.9 87 22.9 Volunteers 73 31.7 58 25.2 49 21.3 50 21.7 Staff 26 37.1 10 14.3 19 27.1 15 21.4 Total Support (N) 484 348 349 297 With the exception of physical support, more evangelical organizations were more likely to have ties of all types (see Table 6.24). For example, 35% of all the information ties shared in the networks were shared by the most evangelical organizations. The most evangelical organizations claimed to have 283 more interactions with other organizations than the least evangelical organizations. More evangelical churches claim to be more connected to the local community than less evangelical churches. The interview data have shown that less evangelical churches have a higher profile in the community. Furthermore, the less evangelical churches tend to have a longer history than the more evangelical churches. This may mean that less evangelical churches do not have to work as hard at developing networks as more evangelical churches because they have long established connections that meet their needs. However, these data may also indicate that more evangelical churches are more involved in the community. This may be a function of the larger number of people that attend more evangelical churches. However, less evangelical churches have a greater willingness to be involved with organizations that are different from themselves (see Table 6.23). None of the most evangelical churches rated association with non-Christian religious as important to their identity. Over half of the most evangelical churches rated association with non-Christian religions as low on the scale as possible. In contrast, only 2.8% of the least evangelical churches rated association with non-Christian religions that low. These differences indicate that different types of churches have different goals. Nonetheless, it is surprising that the organizations which rated evangelism the highest do not think that associating with non- Christian religions is important to their identity. This may result from historic associations between ecumenism and liberalism. [ Page ] 159 One benefit of network connection is assistance in obtaining common goals. I assumed that churches that are more evangelical would have different goals than less evangelical churches. Table 6.25 shows how churches with different levels of evangelicalness ranked the importance of evangelicalism to their identities (“1” indicates very important and “7” indicates not important to the identity of the church). Missions and evangelicalism were more important to churches that are more evangelical. For example, the majority of more evangelical churches indicated on their surveys that evangelizing non- Christians was very important to them. Such a conviction could provide the motivation for network involvement and could restrict interaction with churches that do not share this conviction. Several of the common events staged in the local community by evangelicals reflect this common goal (e.g., March for Jesus, Franklin Graham Crusade). If we look at all my informants we find a clear relationship between evangelicalness and the importance of evangelizing non-Christians. It is not surprising that a definition of evangelicalism that includes evangelism as one of its components finds that evangelicals evangelize. Nonetheless, it is significant that evangelism remains a distinguishing feature of evangelicalism, as evangelicals have sometimes understood the retreat from evangelism as a retreat from orthodox Christianity. This particular conviction was less common among mainline churches (e.g., only one of eight responding United churches considered it representative of their churches). Table 6.25 - Evangelism of Non-Christians (All Informants) [ Please contact repository@tyndale.ca for Table 6.25 details ] Importance to Identity Evangelical Total (N) Most More Less Least Important 1 % 42.9 38.8 16.3 2.0 100.0 (49) 2 % 10.0 20.0 50.0 20.0 100.0 (10) 3 % 87.5 12.5 100.0 (8) 4 % 50.0 50.0 100.0 (6) 5 % 14.3 85.7 100.0 (7) 6 % 100.0 100.0 (5) Least Important 7 % 100.0 100.0 (2) Total % 25.3 24.1 27.6 23.0 100.0 (87) [ Page ] 160 There is evidence for evangelicals sharing resources with each other. However, there is also evidence for evangelicals interacting with and sharing resources with non-evangelical organizations. The important factor is not sharing resources but what those resources are supporting. Thus, evangelicals interact with other evangelicals for evangelism purposes but interact with non-evangelicals to accomplish social service goals. The implication here is that interaction among churches is not based entirely on theological divisions. Relationships between churches are more functional. Some of my informants made the functional nature of some of their interactions explicit, I am more interested in a dynamic ecumenism than an institutional ecumenism. Our congregation is gaining a good idea of why we exist, and we’re interested in joining forces with other groups who share similar elements of our mission (1027). Thus, churches are willing to interact (or support common institutions) in some areas but not others based on their particular goals not just based on their organizational identity. 6.2.3 Why do Religious Organizations Interact? The presence of network connections implies that the benefits of social support and encouragement to common goals outweigh the costs of network involvement. In this section, I described the costs and benefits of network involvement. I found that there were significant costs to maintaining links within the network. These costs are related to time (e.g., information flow) and resource commitment (e.g., financial aid, volunteers) . Nonetheless, I also found that these costs were not the same for all organizations. For example, the type of resources shared depended on whether or not a connection was a connection with another church or with a service organization. Information was the resource most commonly shared among churches. Whereas, finance was the resource most commonly shared in the network as a whole. Thus, the costs of an interaction depend on the type of connection. The cost of interaction with a church was most often information but the cost of interacting with a service organization was most often money. The rarity of direct connections among churches indicated that direct connections have higher costs than other types of connections. Benefits of interaction include the provision of social support for other churches and for local social service organizations. Direct connections between churches primarily provided support through the sharing of information and participation in common events. A smaller number of churches provided physical support or volunteers to another church. [ Page ] 161 However, direct connections to other local organizations, which are primarily service organizations, often include significant financial support, information sharing, and participation in common events. Different types of social support flow through the network depending on the organizations involved. Thus, different organizations receive different benefits from network involvement. In addition to evidence of social support, there is also evidence of common goals such as common involvement with similar organizations. Common goals can also be represented by type of connections. My examination of type of connection showed that churches have clear boundaries based on evangelicalness (e.g., association with different religious groups) but connections with service organizations transcend most of these boundaries. Nonetheless, divisions based on theology can be quite explicit: I would just like to say that we are Historical Baptists as opposed to Protestant Baptists. That is that we did not come out of the Catholic church. True Bible- believing Baptists were around before there ever was a Catholic Church. (1038) Furthermore, there is evidence of network differences based on what kind of organization is being interacted with. For example, mainline churches are more comfortable associating with non-Christian religions than are evangelical churches. Mainline churches are more likely to make statements such as, All these vary between individuals in the congregation and vary over time. [Our Church] is committed to be welcoming irregardless of these differences, and want to hear many different perspectives and views. (1058) Thus, the existence of barriers between theologically different churches and the preference for working with organizations which share a church’s worldview points to the perceived need to have common goals for direct connections. Nonetheless, churches with very different worldviews have common goals related to social services. Thus, functional considerations can be more important than theological considerations when connectivity does not threaten a church’s worldview. Network connections do provide support for common goals but these goals vary according to which organizations are being connected. I conclude that, local churches and religious organizations interact to share resources and to support their goals. Furthermore, there are costs and benefits associated with these interactions (e.g., financial costs, social support benefits). However, the extent of interaction among local religious organizations demonstrates that the benefits of interaction outweigh the costs. Nonetheless, even relatively low cost connections (e.g., information and events) are [ Page ] 162 not shared with the entire network and costs and benefits are relative to what kind of organization is being connected to. Religious organizations interact because the benefits of doing so outweigh the costs. Those benefits are as likely to be based on functional goals as on identity maintenance. This diversity in interaction is a reflection of the diversity of evangelicalism itself. Evangelicalism is no one thing with no one worldview. However, evangelicals do have some common beliefs and values. Furthermore, they are a minority in the Canadian environment. Thus, they recognize the need to work together on some common goals despite having significant differences in other areas. Furthermore, because of the differences within the larger community of Protestantism, some of this interaction takes place outside of the circle of evangelicalism. Evangelical churches are already familiar with working with other organizations that are not similar to them in all respects. This willingness to band together for specific functional goals while maintaining significant differences suggest that evangelicalism can adapt to changing environments to improve its chance for survival. 6.3 Homophily and the Supply of Resources The costs and benefits of network involvement help explain why organizations interact. My third proposition concentrates on how this network involvement relates to organizational characteristics. In particular it is concerned with homophily and the supply of resources. That is, it is concerned with the degree to which similar organizations act the same and group together and how that grouping affects resource exchange. An examination of homophily and resource exchange provides a response to the question: how can we account for network structure? The network I found in the data is informal. It is not primarily a network that organizations join consciously for specific reasons. It exists by virtue of the various interactions that organizations have in their regular functioning. Thus, reasons for network involvement are related to organizational similarity, common goals, and resource exchange (e.g., material and social), as these characteristics and circumstances provide the main basis for interaction. Given these reasons for network involvement, the relatively small size of the community, and the degree to which Protestant organizations have the same concerns, I expected that the local network would be highly connected and that the distance between [ Page ] 163 various organizations would be small. I also expected to find a collection of densely connected networks organized around denominational and doctrinal similarities and organization type. Thus, (3) Network role and reasons for network involvement can be different for different organizations a) Networks and cliques will be organized primarily on the basis of organizational similarities. b) However, connections between organizations peripheral to each network and the central organization(s) of each network will be based primarily on shared group involvement and common actions. c) Nonetheless, most members of a network will be close enough to other members of each network to share resources This section looks at each of the three expectations of this proposition in sequence: networks organized around organizational similarity, peripheral organizations connected by shared group involvement and common actions, and the existence of a city-wide, relatively dense network for resource sharing. I found that different organizations had different network roles and involvement. Cliques did sometimes relate to organizational similarity but, just as often, had more functional purposes. Peripheral and central organizations were not significantly different in how they interacted with the network and most members were only close enough to share resources if we take into account common interactions with third-parties. 6.3.1 Organization Similarity Local Community Church moved into the neighbourhood several decades after Established Church (not their real names). Local Community Church conceived of itself as a community church and so it set out to build bridges with social services organizations and other churches already established in the community. As a result of this effort, Local Community Church and Established Church began to have a lot of contact with each other. Their pastors talked regularly, they had joint services and seminars, they ran a community carnival and a Vacation Bible School together, and Local Community Church sometimes used Established Church’s building for their congregational meetings. The two churches were similar in theology. However, they came from different denominations, had different ideas about how to interact with the community, and, had different internal authority structures. While both churches would say that contact was positive, there is only limited evidence that they have become more similar over their years of contact. In fact, Local [ Page ] 164 Community Church’s emphasis on working with the community forced Established Church to rethink its relationship with the community and define it in a somewhat different way than Local Community Church. At the same time, Local Community Church’s community involvement has opened up opportunities for community involvement that Established Church may not have pursued on its own. This example highlights the ambiguity of influence across organizational connections. Local Community Church and Established Church share some significant characteristics and have developed in similar ways. However, they have also maintained some significant differences and developed their ministries in different directions. This section looks at the survey data on how closely connected different organizations are to try to determine how connectivity and similarity are related. I expected that networks of relations would be organized around organizational similarities such as denomination, doctrine, and organization type. My previous two propositions have already dealt with some aspects of this expectation and have found that, although such relationships do exist, they are not universal nor are they the only basis of interaction. Observed network relations had some similarities. However, single instances of organization similarity, by themselves, are not sufficient to explain all network connections. This section describes how organizational similarity affects network relations. Like the classification of Protestant churches discussed in chapter 2, what leads to network connectivity are a specific set of non-unique circumstances and characteristics that come together to encourage organizational interaction. These interactions may have more to do with the functional value of network involvement than with the specific characteristics of the organizations themselves. That is, organizations do not interact only because they are similar. Organizations interact because to do so provides them with some sort of functional benefit. As discussed in section 6.2.2, these benefits can be physical benefits like information and money or they can be less empirical benefits like development of identity or “storing up treasures in heaven.” Thus, organizational similarity is only one factor in organizational interaction—and perhaps not the most important one. To examine the role organizational similarity plays in organizational interaction in more detail, I examine whether or not closely connected organizations are similar and whether or not similar organizations are closely connected. [ Page ] 165 6.3.1.1 How similar are closely connected organizations? I have three measures of the closeness of connections among organizations in the local community: interviewee impressions, network diagrams and related clique models, and models of regular equivalence. The data do show that closely connected organizations may be similar. However, there is little evidence that close connections are necessarily based on organizational similarity. The interviewees’ perception was that the mainline organizations formed a clique and to a lesser degree so did the evangelical organizations. Several political leaders expressed the desire that non-Christian religious groups be more integrated into the activities of the religious community. Non-Christian religious groups were not perceived to be involved in these cliques (with the notable exception of the Jewish community, which appeared to be part of the mainline clique). According to the interviews, there was some overlap between these cliques, most notably in the area of meeting the physical needs of the community’s poor. This overlap had its most obvious physical expression in the organization known as the Core Area Churches, which was formed to address the problems of the Kitchener downtown core from a religious perspective. It included representatives of a variety of Christian denominations co-operating with the city towards common goals. This group eventually disbanded when the health of the downtown improved and its raison d’etre disappeared. Thus, the interviewees’ perception was that close connections implied organizational similarity. A look at my network data can test these perceptions. Organizations that appear to group according to similarity can be observed in my network diagrams. For example, there were close connections between Pentecostals, Mennonite, and mainline churches (see Diagram 6.2). Nonetheless, not all churches of those types were part of those networks, and churches that did not share those denominational affiliations were part of the networks (see 6.1.1). Churches often had multiple points of connection with churches from different denominations, even as they did not necessarily interact with other churches from their own denomination. For example, Mennonite churches (1072) interacted with Pentecostal churches (1039). One Lutheran Church (1050) has direct connections with a Baptist, a Pentecostal, and a Mennonite church, but only one connection with another Lutheran church (1010). However, in this case, most of these churches share a high level of evangelicalness. Evangelicalness can also be a reason for interaction. My evangelicalness network diagrams show some interaction based on evangelicalness, but also [ Page ] 166 significant interaction between organizations which do not share the same level of evangelicalness (see Diagram 6.1 and Diagram 6.4). Thus, there is no evidence from my network diagrams that closely connected organizations are necessarily similar. Table 6.26 - Cliques in the Direct matrix [ Please contact repository@tyndale.ca for Table 6.26 details ] Clique Organizations 1 1039 1072 1083 2 1039 1081 1083 3 1013 1046 1080 4 1001 1042 1055 1069 A related measure of network connectivity are clique models. The clique model I am using looks for groupings of at least three organizations that are all connected to each other. Wasserman and Faust (1994:254) supply us with the mathematical definition that a clique is “a maximal complete subgraph of three or more nodes.” That is, a clique is a group of at least three organizations that are all connected to each other with no other organizations being connected to all of them (Wasserman and Faust 1994:254). The data can identify some distinct groups of densely connected religious organizations in Kitchener-Waterloo. However, I do not require a rigid adherence to Wasserman and Faust’s definition in my analysis. Clique models can help identify groupings of organizations missed in a visual inspection. Ignoring the direction of ties, I found four cliques in the Direct matrix (see Table 6.26). These cliques were similar with respect to evangelicalness but not denomination type. There were seven aditional cliques in the Connect matrix (see Table 6.27). I do not have evangelicalness or denomination types for non-informants. However, the informants that are involved in these cliques can be similar in evangelicalness and denomination type (1081- 1083) or different (1045-1081). A less rigid definition of clique produced more cliques but, given the variety of connections in the network, would not be likely to be any more homogeneous than the cliques discussed here. If we use the same procedure to identify cliques based on which churches share interaction with the same third party, as in the Total matrix, we find 136 cliques. One of these cliques has 21 members. Thus, clique models provide little evidence that closely connected organizations are necessarily similar. [ Page ] 167 Table 6.27 - Cliques in the Connect matrix [ Please contact repository@tyndale.ca for Table 6.27 details ] Cliques Organizations 1 1039 1072 1083 2 1039 1081 1083 3 1013 1046 1080 4 1001 1042 1055 1069 5 1081 1083 1285 6 1081 1083 1299 7 1041 1050 1144 8 1045 1081 1199 9 1055 1069 1109 10 1069 1078 1296 11 1039 1081 1199 The two models of equivalence I used produced similar results (see 6.1.2.3). Organizations grouped based on connections were not necessarily similar in the characteristics I measured. This was the case when I examined organizational grouping based on which organizations send and receive the most ties (i.e., core/periphery models) and when I examined organizational grouping based on which organizations shared the same role position in the network (i.e., blockmodels). Core/periphery models provide a measure of which organizations are most closely connected together. That is, the core both sends and receives a large number of connections with other informants in the core. Blockmodels measure which organizations have similar roles in the network. That is, which organizations relate to other organizations in similar ways. These measures provide two different ways of grouping organizations based on their connections. In both cases there was no evidence that organizational similarity was the main factor which organized those groupings. A significant implication of this finding for researchers is that it cannot be assumed that organizations are connected together primarily because they are similar. Organizations do have to have some similarities, such as common goals, in order to make interaction worthwhile. However, organizations do not have to be similar in all respects. My informant have clearly demonstrated that they are willing to interact with organizations because they share some common goal. Most commonly, in the data, that common goal is helping the underprivileged. [ Page ] 168 These various procedures demonstrate that closely connected organizations were not necessarily similar. More specifically, close connections between organizations did not imply similarity of evangelicalness or denomination type. Nonetheless, there were some similarities between closely connected organizations. Thus, it is worthwhile to study the question from the other direction. 6.3.1.2 How closely connected are similar organizations? I have already discussed connectivity based on evangelicalness and denomination type (see 6.1.2). I discovered that, while there were groupings of similar organizations, these organizations also have significant connections with dissimilar organizations. A prime example is local pastor’s associations (ministerials). These are groupings of pastors with similar perspectives on certain issues who get together to support each other and discuss those issues. However, they do not necessarily interact outside of the Pastor’s association. One other place to look for similarity that might provide close connections is ethnicity. There was evidence of ethnicity-based connections in the data (see Diagram 6.12). Organizations oriented around ethnicity may be less likely to be involved in the larger network for both internal (e.g., language barriers) and external (e.g., prejudice) reasons. However, when an ethnic denomination begins to expand by attracting members from the larger community (e.g., the Mennonite-Brethren) it becomes less insular in its relations with other organizations. Thus, some of these connections were closely related to denomination in addition to ethnicity (Compare Diagram 6.12 with Diagram 6.2). Nonetheless, churches that belong to the same denomination can be radically different ethnically. For example, a local Vietnamese and an English-speaking church share a building and yet, have little interaction with each other beyond what is necessary to share the facility. Furthermore, neither of them have much interaction with the Chinese church, of the same denomination, ten minutes down the road. In contrast, there are explicit connections between the three Vietnamese churches in town. Thus, denominations can officially connect organizations without prompting them to work together. In this instance, similar ethnic affiliation proved to be stronger at keeping churches together than was denominational affiliation. While there is evidence of connections that may be based on evangelicalness, denominational or ethnic affiliation, most of those organizations had similar connections with organizations not based on denomination or ethnicity. Thus, similar organizations are not [ Page ] 169 necessarily closely connected. Therefore, there must be a basis of interaction besides similarity. I deal with this question in the next section. Diagram 6.12 - Direct Matrix (Ethnicity) [ Please contact repository@tyndale.ca for Diagram 6.12 details ] 6.3.1.3 What is the basis for interaction? Closely connected organizations were not always as similar as expected. Nor are similar organizations necessarily closely connected. Churches do not necessarily group with other churches solely based on organizational similarity. While similarity is not the only basis on which organizations interact some characteristics do lead to certain kinds of connectivity. Evangelicalness, for example, does play a role in which organizations are the most central. Thus, while similarity can be a basis for network connectivity, it is not the only basis for such connectivity, nor do organizations have to be similar in all respect to interact. What then is the basis for interaction? The primary reason for interaction appears to be practical or functional. For example, organizational interactions based on similarity are based on which similarities best match what the organization is trying to do. Thus, a church may interact with church of a different denomination that shares its ethnicity rather than churches of the same denomination which do not. Likewise, connection with geographically proximate organizations is related to [ Page ] 170 church goals not just proximity. For example, Local Community Church has very specific connections with social services that are geographically proximate to where it operates (the Salvation Army, a community centre, a local public school, etc.) because those connections have the functional benefit of furthering its agenda of attracting church members from the geographic community of which it is a part. The need for a functional basis for many types of connectivity means that close connections do not necessarily lead to similarity. The common assumption is that contact will spread characteristics. While this is the case in some circumstances (e.g., the common concern for certain political issues among evangelicals), it is not necessarily the case. The data show that very different churches can both support the same service organization without becoming like each other. For example, churches which say that the belief in an inerrant Bible is important to their identity (1051, 1013) and churches which say it is not (1009, 1060) can both support the same food bank because they share a common value: it is important to help the poor. The implication of this is that churches can share common connections based on common values without having all values and characteristics in common and without becoming more like each other. Nonetheless, it is possible that if churches were aware of the degree to which they support the same service organizations they might be more willing to work directly with each other. This diversity of interaction among local churches is a reflection of the diversity of the churches themselves. I will let the words of one pastor conclude this section by describing some of the functional considerations that provide a basis for interaction among local organizations, Theologically our congregation is more liberal than most in our synod, but still falls easily within the parameters of . . . orthodoxy. We have a large Roman Catholic and United Church subculture in our congregation; Eucharistic piety, belief in the power of prayer, a highly-developed social conscience, interest in music and the arts, and sensitivity to [churches with a different denomination than our own] and [the] non-Christian faith community are distinguishing characteristics of our church. The majority of our members are families with parents in their 30’s and 40’s. We have almost 50 children in Sunday School, about 20 active in the Jr./Sr. Youth groups. we always have a seminarian placed (10hrs/wk) . . . interest in . . . Campus Ministry. We have just voted to begin a[n] . . . expansion of our facilities. We have a huge property, and a . . . community garden which we manage and support. (1009)113 ___________________________________ 113 Some identifying details have been changed to protect the identity of this informant. [ Page ] 171 Many and diverse functional concerns can lead to many and diverse interactions with other community organizations. 6.3.2 Peripheral Connections Organizational interaction often has a functional rather than a characteristic oriented basis. One area which can bring together functional and characteristic oriented concerns is common goals (as represented here by shared group involvement and common actions). I expected that common goals would be a common point of connection between central and peripheral organizations. Peripheral organizations are organizations that are not highly connected with the rest of the network. For example, they might only be part of the network in that they receive or send one tie from another network member. In so far as peripheral organizations intentionally connect with the network, there must be some reason for peripheral organizations to interact with a larger network, despite not being in a position to control or direct most resources. The interviews provided evidence that churches grouped around common goals. Therefore, I proposed that these common goals could be seen in common group involvement and common actions. I argue in this part of Proposition 3 that, connections between organizations peripheral to each network and the central organizations of each network will be based primarily on shared group involvement and common actions. The difference between central and peripheral groups is relative to the ties they have with the rest of the network. Therefore, to a certain degree, what constitutes centrality is arbitrary. For example, sending and receiving two or more ties could be central or sending and receiving three or more ties might be needed to make an organization central. The choice of what constitutes a central versus a peripheral organization depends on what the research is trying to explore. Thus, I need to establish what the difference is between central and peripheral organizations based on what I am trying to find out. In view of this, I identify centrality in two different ways. An organization can be considered central if it has a large number of connections (degree centrality—it both sends and receives a large number of connections), and if its position in the network is between a large number of other organizations (betweenness centrality). In both cases, central organizations have the potential to exercise power over peripheral organizations because they are more likely to control access to resources. In this section, I explore the way that peripheral organizations connect with the larger [ Page ] 172 network. First, I examine the evidence for shared group involvement and common actions in the interview data. Then I examine evidence for this in the survey data. I look at shared group involvement, involvement with the local community, and interactions with social services to determine if there are any centre/periphery differences among connected organizations. More specifically, I explore whether or not there are centre/periphery differences based on evangelicalness. The interviewees explained that, they were not aware of actual connections between churches. However, they were aware of organizations and events that implied some form of inter-church cooperation. They did not recognize that inter-church cooperation as representing actual connections because the connections did not necessarily happen between churches but among churches meeting for a common task or to support a common event. The two areas of cooperation mentioned most often were social services and evangelism. The interviewees did not comment on centre / periphery differences. However, they did perceive differences between mainline and evangelical churches. They reported that, mainline churches tended to group around poverty, ecumenical, and social justice issues (C208, C226) whereas the Evangelicals tended to group around evangelism (e.g., Franklin Graham Crusade) or conservative moral issues (e.g., Pro-life) (C208, C234, C238). Furthermore, their impressions were that, when evangelicals did enter the social justice arena, they tended to do it by themselves (e.g., Oasis, C233). Nonetheless, even in the interview data there is evidence of both evangelical and mainline churches involved in social services and evangelism (e.g., mainline Protestants at Promise Keepers, Evangelicals volunteering at St. John’s Soup Kitchen). Other churches appeared to sit between the evangelical / mainline divide. Mennonites, for example, appeared to straddle both worlds; as if they have a conservative moral conscience and a liberal social conscience. The interviewees provided evidence of local churches which had both common group involvement and common actions (e.g., the core area churches—social services, the Franklin Graham crusade-evangelism). However, this anecdotal description of church interaction does not tell us how these connections differ between central and peripheral organizations except that, the interviewees thought that evangelicals were peripheral to social service oriented networks and mainline churches were peripheral to evangelism oriented networks. Furthermore, the connections mentioned by the interviewees were sometimes not mentioned at all on the surveys I received. For example, neither the core area churches nor [ Page ] 173 the Franklin Graham crusade were current activities at the time of the survey and thus did not show up on the survey. The survey data can be used to evaluate the interviewees’ impressions. The interviewees’ general impression that interaction between churches and the community is oriented around social services and evangelism proved accurate. Survey respondents listed 342 unique organizations representing many different types of organizations on the surveys (churches, missions agencies, educational institutions, social service agencies, etc.). Connections to social services were the most common connections listed (233 connections). These were followed by churches (164 connections) and then missions oriented organization (141 connections) (see Table 6.6). To illustrate interaction with these organizations, this section concentrates on the 20 organizations that were mentioned by five or more respondents (see Table 6.14).114 The most surprising finding is the number of organizations that were mentioned by both evangelical and mainline churches. Thirteen of these twenty organizations were mentioned by at least one mainline church and at least one evangelical church. The seven organizations with the most connections to respondent churches were all involved in social services of some kind. Eleven of the top twenty were involved in social services. Of the remaining nine organizations: four were missions related organizations, four were churches or organizations of churches and one was an educational institution. Not surprisingly, the mission organizations were only mentioned by evangelical respondents. Thus, there is no evidence that evangelical organizations are peripheral to all social service oriented organizations. However, there is evidence that interaction with social services is a higher priority for mainline organizations than for evangelical organizations (see Table 6.28). Mainline organizations do not put as much time into evangelism therefore they may have more resources for social services. These data demonstrate that churches of all kinds are extensively involved with the local community, and that there is no clear centre/periphery difference between evangelical and mainline churches in regards to interaction with social services. _____________________________________ 114 This actually encompasses the top 23 most mentioned categories. However, three of those categories were “Miscellaneous Churches,” “Miscellaneous charities,” and “Miscellaneous Missions.” These three categories have been removed from this discussion because the organizations referred to cannot be precisely identified. [ Page ] 174 Similar patterns emerge if we look at all the organizations receiving connections. A significant number of the organizations receiving connections were social service organizations. For example, 36% of all connections were to social service organizations. Of the ten most mentioned connections, seven of them are social service organizations (see Table 6.14). Twenty-four of the 49 organizations receiving three or more connections were social services. An additional 14 were mission organizations or education institutions. Eight of those top ten organizations also have high betweenness centrality which suggests the possibility of exercising power over the organizations they connect (see Table 6.12). The most mentioned organization was the House of Friendship (an organization dedicated to poverty relief) with 22 mentions, followed closely by the youth focused social service organization, Ray of Hope, with 19 mentions. The food bank, the crisis pregnancy centre, the Mennonite Central Committee, St. John’s Soup Kitchen, a safe home for women, and Habitat for Humanity were all in the top ten mentioned organizations. Moreover, the House of Friendship, Ray of Hope, and the food bank are the most likely to play connecting roles between other organizations (see Table 6.12). The same pattern of social service organizations being the most often mentioned organizations exists in the interview and the survey data. These service organizations play the middle role between different churches. Furthermore, the data indicates that evangelicals are just as involved in social services as mainline churches. Social services connect with both central and peripheral organizations. However, we still do not know if there are differences in how central and peripheral churches interact with the network other than the definitional difference that central churches have more interaction with other members of the network than peripheral churches. To evaluate whether or not there are differences between central and periphery organizations I examined differences in their connections in the data. Table 6.28 and Table 6.29 examine differences between how core and periphery churches interact with different types of organizations and whether or not core and periphery churches have different types of ties. By definition, churches are in the core if they send and receive more ties than other organizations. That is, they have high degree centrality. The data show that core organizations are generally more involved with different types of contacts and have more of each type of tie. For example, 44% of event ties are with core organizations as compared to 25% with periphery organizations (see Table 6.29). However, in several cases Periphery [ Page ] 175 organizations have more connections than Core organizations. This is the case for interactions with social services and for physical support and financial ties. There are core/periphery differences in how organizations participate in common actions. However, those differences are not always in favour of the core. Peripheral to the network is not necessarily peripheral to the community. For example, a church may prioritize interaction with the community more than interaction with other churches. This may explain the tendency of mainline churches to be at the periphery of the local church network. Table 6.28 - Contact Type * Core/Periphery Crosstabulation [ Please contact repository@tyndale.ca for Table 6.28 details ] Contact Type Core/Periphery (%) Total (N) No Connections Core Semi- periphery Periphery Social Service 3.9% 31.3% 24.5% 40.3% 100.0 (233) Missions Related 2.1% 34.8% 24.1% 39.0% 100.0 (141) Church Related 2.4% 56.7% 23.2% 17.7% 100.0 (164) Education 15.0% 25.0% 12.5% 47.5% 100.0 (40) Health Care 42.9% 35.7% 21.4% 100.0 (14) Other 13.7% 15.7% 29.4% 41.2% 100.0 (51) Total 4.5 (29) 37.2 (239) 24.0 (154) 34.4 (221) 100.0 (643) Table 6.29 - Type of Interaction * Core/Periphery Crosstabulation [ Please contact repository@tyndale.ca for Table 6.29 details ] Tie Type Core/Periphery (%) Total (N) No Connections Core Semi- periphery Periphery Physical Support 7.2% 26.6% 33.8% 32.4% 100.0 (207) Information 5.8% 36.7% 31.5% 26.1% 100.0 (330) Events 5.4% 44.1% 26.1% 24.5% 100.0 (261) Finance 5.5% 30.3% 27.9% 36.3% 100.0 (380) Volunteers 5.7% 35.2% 30.0% 29.1% 100.0 (230) Staff 5.7% 41.4% 35.7% 17.1% 100.0 (70) Total 5.1 (33) 36.8 (239) 23.7 (154) 34.4 (223) 100.0 (649) The data presented in this section demonstrate that organizations that are part of the network look very similar in terms of how they interact with other organizations. Organizations that are extremely peripheral to the network, by definition, do not interact with the network at all (thus, there is no interaction to explore). The difference between more [ Page ] 176 central and more peripheral churches is more closely related to the network itself than the characteristics of individual churches. That is, organizations are more or less central because of the structure of the entire network not because they have certain characteristics. Certain characteristics may explain why individual churches interact with individual churches. However, knowing the characteristics of a church does not tell one whether they are part of the local network, nor, if they are connected, what role they play in that network. For example, organizations with the same theology may not interact with each other. Knowing that an organization is evangelical and ethnically German does not tell us whether that organization has a large number of local interactions or very few. By contrast, information about relative location in the network (e.g., core/periphery) does tell something about network involvement and role. For example, core churches are more connected with other churches but not necessarily more connected with social services than periphery organizations. Core churches have access to resources of all kinds from the network. Whereas, periphery churches must seek some of those same resources elsewhere. Core churches are connected to social services, churches, and missions organizations in that order. Furthermore, there is no clear relationship between core / periphery differences and evangelical / mainline differences. What this tells us is that, shared group involvement is only one factor in how peripheral and central organizations interact with each other. Considerations related to the practical necessities of running a church play a considerable role in who connects with whom. Having examined one type of local organization (organizations peripheral to the network) I now return to a consideration of the whole network. 6.3.3 City-Wide Dense Networks The last part of Proposition 3 states that, most members of a network will be close enough to other members of each network to share significant resources. I expected that this community network would be highly connected. That is, I expected that most local churches would have a large number of connections with other churches in the local area. Therefore, all the resources available through the network would be, theoretically, available to all members of the network. Nonetheless, I expected that the more central members of the network would be more likely to take advantage of them. In this section, I explore my expectation that most members of the network would be close to each other. That is, I [ Page ] 177 expected that most members of the network would be directly connected or could easily make direct connections with other members of the network through common third-parties. In some ways, different network members are not noticeably different in resource usage. They have similar proportions of different ties to the organizations with which they interact. They have relatively few connections with other local churches. They share connections with many other local organizations (particularly social services). The plenitude of common connections provides reason to believe that most members of the network will share resources with each other. My expectation was that common connections to third-parties could connect previously unconnected churches in two ways. First, churches could physically meet or become aware of each other through common involvement with the third-party. Second, third-parties could consciously bring together churches they connect with for their own purposes. I will first review the interviewees’ impressions of network density. That is, how closely network members are connected to each other. Following that, I examine evidence from the survey data. Then, having explained that my expectations were not entirely accurate, I discuss how resources are actually exchanged in the local network. I have described these data elsewhere. Therefore, this section is a summary of previous discussions reorganized to emphasize a different point. The interviewees were not aware of a significant amount of interaction among local churches. They could point to some evidence of interaction (food bank, evangelistic efforts, etc.) and they were aware of some formal networks. For example, they mentioned local ministerials (regular meetings of pastors), and a group called the core area churches (which was no longer operating). However, when I asked specifically about networks among evangelicals, even the evangelical informants could not identify significant interaction. Furthermore, they were of the general impression that there was little interaction among local churches. Thus, according to the interviewees, many network members are not close enough to each other to share resources of any kind on a regular basis. The survey data confirms some of this story. I have already established that churches are rarely directly connected to each other. However, I have also established that there are significant points of connection centred around local social service organizations. An examination of whole network density shows that very few of the possible connections exist [ Page ] 178 in the Direct and Connect matrices (see Table 6.7). Yet, the mean distance between organizations that are actually part of the network is relatively small (see Table 6.30). Average density between reachable pairs is a measure of how many ties there are between various members of the network. For example, if two churches are directly connected then the distance between them is one. If they are not directly connected but share a common connection with a third-party then the distance between them is two. Distance measures the shortest path between each pair that is connected to the network. To compute average distance the shortest distances between each pair in the network are added together and then divided by the total number of pairs to produce a mean. Such a measure for the Connect matrix provides the average distance between potential direct connections. Since the Total matrix is already a measure of common connections between third-parties, the average distance score is a measure of distance between churches assuming a common connection leads to a direct connection. In practical terms it means that the distance between churches in the Total matrix starts at a distance of two ties. Therefore, the listed average distance is only the average distance if all shared connections become direct connects. In addition, I have assumed that interaction is symmetrical. That is, I assume that if A interacts with B, then B also interacts with A—whether or not B claims to interact with A. Table 6.30 - Average (Mean) Distance between Reachable Pairs Direct Connect Total 3.695 2.314 1.872 This measure is useful because it gives a gauge of connectivity within the entire network. We can determine from this measure that directly connected organizations on average require almost four ties to connect to other members of the network. A look at a visual representation of the network demonstrates why this is the case (see Diagram 6.1). There are a large number of churches on the periphery of the network that are connected together through a smaller number of bridge organizations which connect different parts of the network. Practically this means that resources are more accessible near the centre of the network. A distance of four or five from another church means that there is little chance for interaction in most circumstances. However, some resources easily travel long distances [ Page ] 179 (e.g., information). Thus, it is not surprising that churches which do not directly connect or give each other money can cooperate on a common community event (e.g., the March for Jesus, the Franklin Graham crusade). Nonetheless, given the large number of churches that are not directly connected to the network (this procedure only examines connected churches), it is quite likely that many churches are unaware of the existence of some of the other churches in the local area even if they share significant similarities with each other. However, direct connections do not tell the entire story. Many of these churches interact with the same social service organizations. Resource exchange between churches and other types of religious organizations is plentiful and, particularly in the case of social services, can serve as a forum where resources are shared towards a common goal. Thus, although direct connections between churches are of limited use for trading resources across the entire network, in theory, mutual connections to a third party, that is not a church significantly reduce distance between churches. For example, the Connect number shows that the distance between churches may not be as large as they initially appear. In the Connect matrix I added organizations listed by my informants that are likely to lead to direct connections. These additional organizations reduce the distance between my informants. For example, organization 2299 is a charismatic/Pentecostal church located just outside of Kitchener-Waterloo. Three informants indicated on their surveys that they share information and participate in the same events as organization 2299 (1024, 1081, 1083). Churches 1081 and 1083 were already linked. However, the connection between 1024 and 2299 significantly reduces the distance between the two sides of the local network (see Diagram 6.4). Furthermore, common participation in the same events suggests that there is a significant potential for direct connections between church 1024 and churches 1081 and 1083. Thus, resources may be more easily exchanged than suggested by the averaged distance of the Direct matrix. If we assume that all connections to third-parties will lead to direct connects than the distance is even smaller. The average church can get to any other church in the network through less than two third-parties. Many of these common connections will never lead to direct connections but others will. However, this finding does not tell us if these connections through third-parties allow the trading of resources beyond those necessary to fulfill the function of that third party. The data refutes the proposition that all types of resources can easily travel anywhere in the network (see Table 6.6). However, it is clear that some types of resources are easily [ Page ] 180 exchanged through out the system. Yet, regular resource exchange is obstructed by the lack of direct connections between local churches. 6.3.4 How Can We Account for Network Structure? Network role and reasons for network involvement are different for different organizations. However, I did not find the differences in network role and reasons for network involvement I was expecting. There is not necessarily a relationship between network position and characteristic similarity. Nonetheless, some types of similarity (e.g., evangelical churches are more likely to have direct connections) and common connections to the same service organizations (primarily social services and missions organizations) do play a role in network structure. Furthermore, some characteristics (e.g., evangelicalness, denomination) were related to interaction in some situations. However, local networks were not homogeneous nor did they include all churches with a particular characteristic. That is, similar churches are not necessarily closely connected. Nor are closely connected organizations necessarily similar. Furthermore, many churches are not homogeneous internally. One Pastor explains that, My impressions of the congregation are strictly from my experience and broadly given. The fact is this church contains a wide gamut of religious beliefs from those bordering on charismatic to radically liberal. For the most part folks here are in the middle theologically. (1015) This between and within church diversity leads to network structure being partially dependent on functional needs. Churches interact with other churches and with social service organizations based on the function those organizations play in the life of the church. For this reason, interactions with social services and missions organizations are much more common than interactions with other churches. Thus, there is evidence of shared group involvement and common actions but not necessarily between churches. While churches could easily exchange resources with each other through the extended network of churches (i.e., common connections with third-parties), churches are much more closely connected to other local organizations. Church characteristics do not explain why churches were part of the core or the periphery. This is better explained by network connections. For example, peripheral organizations are on the periphery because they have different functional goals than organizations more closely connected to the network, not because they necessarily had different characteristics than other organizations in the network. [ Page ] 181 The interview and survey data point to the general lack of direct connections between churches. This lack of direct connections restricts the ability of churches from different parts of the network to exchange resources. However, the survey data also show that resources are exchanged outside of this direct network and beyond the local area. A church with few local connections is not necessarily a church without access to resources. Yet, resources are not necessarily exchanged in kind, nor are they necessarily encompassed by the physical resources about which I asked. Furthermore, there is a significant amount of connectivity with local service organizations that could facilitate resource exchange. This resource exchange may be a higher priority for some churches than interaction with other churches. Therefore, the lack of direct connections does not necessarily mean no interactions are taking place. To synopsize, some types of similarities are important for group structure, and shared group involvement and common actions shape the network, but resource exchange between churches is primarily based on functional considerations such as common support of the same service organization. Network role and reasons for network involvement are different for different organizations. However, these differences are based on position in the network, as much as they are on the characteristics of organizations or centre/periphery distinctions. 6.4 Chapter Summary This chapter has examined my research questions in light of the survey data. I examined the connections and divisions among local churches, the costs and benefits of network relations, and how organization similarity affected organizational interactions. First, churches did not necessarily connect and divide according to their individual characteristics. Furthermore, there is evidence of both interaction among local religious organizations and church involvement in the local community. Churches hold common events, support common service organizations, and sometimes self-identify with specific local communities (geographic and demographic). Nonetheless, in Kitchener-Waterloo, community leaders are aware of little interaction among churches nor are they aware of much interaction between churches and the community. Local churches, themselves, claim to have little interaction with each other. I did find some differences in direct connections that appear related to the characteristics of individual churches. However, the data show little difference between [ Page ] 182 evangelical and mainline churches in how they interact with the larger network. While there are clear differences between evangelical churches and other types of churches, those differences are not manifest in how they interact with the local community. That is, by most measures evangelical churches are as involved with the local community, including social services, as any other type of Protestant church in the Kitchener-Waterloo area. I found differences between direct and indirect connections. Direct connections were more likely to be based on organizational characteristics than indirect characteristics. Indirect connections were related to common interaction with third-parties, usually social services. Second, this chapter argued that the benefits of network interaction outweigh the costs. The benefits of social support and common goals outweigh the cost of time and resource commitment necessary to maintain the network. However, the costs and benefits of network interaction are different for different organizations. Thus, churches were more likely to bear the costs of interacting with a social service than with another church. Furthermore, an examination of the costs and benefits of interaction demonstrated that the reasons for interaction are often based on the function of that interaction (e.g., access to resources) not on the characteristics of the organizations interacting. Third, this chapter explained that similarity did not necessarily lead to interaction. For example, there is little direct interaction between local Protestant churches. Nonetheless, there is some interaction. Furthermore, local churches share a significant number of connections with the same third-parties. Some of this interaction is among churches with similar characteristics (e.g., evangelicalness, ethnicity). However, much of it is based the functional benefit of interaction. For example, the benefit of interacting with the Food bank outweighs the potential cost of working with another church which does not share your theology. Furthermore, the benefits of interaction accrue to organizations with different characteristics and which are in different locations in the larger network. The implication of these findings is that evangelicals are not isolated from the larger community network because network connectivity is based on function rather than organizational characteristics. In the concluding chapter, I review these findings, explore some possible explanations of my findings, and explain what they might mean for evangelicals across the country. 7 CONCLUSIONS Canadian evangelicals are no longer, if they ever were, completely isolated from each other and from the communities in which they operate. In order to defend this claim I looked at interactions involving Protestant churches in one local community. I explored why and how these churches interacted and the reasons for that interaction. In the process, I defended the use of network analysis to study Canadian evangelicals, described the construction of a survey used to collect network data, pointed out evidence of connectivity, and explored reasons for that connectivity. In short, I have explored, described, and interpreted the structure of evangelical church networks in one community. In this conclusion, I summarize the results of my study, provide an explanation for some of the data, and look ahead to how my research could be extended. 7.1 What have I Shown? This section summarizes my results by looking at the questions that guided the data collection and analysis. I review my discussion of who “evangelicals” are, how they are similar, how and why they interact, and provide a description of the structure of that interaction. 7.1.1 Who are the Religious Organizations I am Studying and How did I Study Them? Organizations that are typically called “evangelical” are diverse and multifaceted and, as a result, cannot be completely examined using a definition which relies on necessary and sufficient characteristics. Therefore, I developed a polythetic definition that did not require sufficient characteristics. I tested this definition and then used it to develop an index of evangelical characteristics for classifying Protestant churches according to evangelicalness. The diversity that led to the need for a new definition of “evangelical” is a result of competing influences on Canadian Christian religious identity. These influences include British stoicism, American charisma, and indigenous Canadian developments (e.g., Canada is a multicultural nation of immigrants). Chapter 3 reviewed some of these influences. It also explained that historical studies of Canadian evangelicals could benefit from a study of the [ Page ] 183 [ Page ] 184 structure of Canadian evangelical connectivity in a single community. Canadian evangelicals are not isolated from each other nor are they isolated from the communities in which they operate. Thus, in order to fully understand interaction between Canadian evangelicals I needed to study more than just the individual characteristics of organizations and the links surrounding single organizations. The use of network analysis to study Canadian evangelicals complements existing studies of evangelical organizations and individuals. It does this by describing a local network context that helps to explain how local religious organizations function. It also describes aspects of the support systems that keep them going. More specifically, I used network analysis to develop a picture of the whole network of Protestant church interactions within a single community (i.e., Kitchener-Waterloo). In the process, I collected data on the insularity of evangelical organizations (e.g., the degree to which they rely on other organizations for resources) and the characteristics of their networks. Network analysis also helped examine the nature of interactions in Kitchener-Waterloo. The prominence of particular types of organizations in that interaction provide data concerning which areas of interaction are the most common among local Protestant churches. To collect these data, I developed a survey to explore network relations among evangelical organizations in Kitchener-Waterloo. This survey collected data about how various local churches were related to each other and the impact these relationships had on their functioning. I also explored the costs and benefits of network involvement and what kind of relationships actually existed between churches. The historical literature suggested, and my research showed, that the Canadian Protestant church environment is more complex than a simple division between evangelistic-minded conservatives and social service minded liberals. In chapter 6, I analyzed the survey data in light of the insights of the historical and network analysis literature. I looked for similarities among organizations that could explain the network connectivity implied by the historical literature and found in my network data. I found that, while organizations did interact based on specific characteristics (e.g., denomination, ethnicity, evangelicalness), specific characteristics could not predict connectivity. For example, there were small networks composed of churches that were all evangelical. However, many of those churches also interacted with non-evangelical churches. Furthermore, some evangelical churches were not connected to the main evangelical [ Page ] 185 networks at all. Thus, knowing that a church is evangelical does not provide enough data to predict the type of organizations that church interacts with. In the remainder of this section I expand on this summary of the interaction I found among local organizations. 7.1.2 How Religious Organizations Interact My first proposition expected multiple, closely connected networks within the local community. I discovered that while similar churches interact, similarity (e.g., ethnicity, denomination, evangelicalness, and geographic proximity) was not the only basis for interaction, nor can it explain all network connections. Churches interact based on similarity but they also interact with churches that are not at all like themselves. For example, different organizations could agree that the same service organization was worth supporting. Thus, similar organizations did not necessarily interact, and interaction did not necessarily imply similarity. Nonetheless, it was possible to identify divisions related to theology, style, and denomination. Evangelicals, for example, interact with each other to a greater degree than mainline churches, but those interactions are still sparse and do not appear to be exclusively related to evangelicalness. There was little evidence of multiple, densely-connected networks based only on church characteristics. As I just described, while evangelicalness plays a dominant role in direct interactions, these interactions are not exclusively evangelical. Furthermore, evangelicals are almost as highly involved with service organizations as churches without an evangelical identity. In fact, local networks are only densely connected when interacting with service organizations not when churches interact directly with other churches. Thus, the degree to which organizations interact is dependent on the sort of organizations with whom they are interacting. There are clear connections between individual churches and smaller networks of like- minded churches, but it is unlikely that any one church has a sense of the entire network. Furthermore, the type of organization affects connectivity. For example, service organizations are more accessible than churches. Data about church characteristics do not explain why some organizations cross these characteristic-based boundaries. However, my network data indicate that crossing boundaries may be influenced by functional considerations such as the common belief in meeting the needs of the poor. [ Page ] 186 7.1.3 Why Religious Organizations Interact The interviewees pointed to examples of events that required the cooperation of a large number of evangelical churches. The survey data show that local churches have many common connections with social service organizations and missions organizations. These interactions point to the costs and benefits of network involvement. Proposition 2 argued that organizations interact because the benefits outweigh the costs. There are significant costs to network involvement including such resources as time, money, and people. These costs are different for direct connections with churches as compared to direct connections with service organizations. Nonetheless, any interaction implies that the benefits of that interaction, such as support through the sharing of information and common goals manifest in participation in common events, outweigh the costs of network involvement (i.e., time, money, people). Thus, while local churches may have very few direct or permanent connections, they are aware of each other sufficiently to cooperate towards common goals on a temporary or ad hoc basis. As with costs, different benefits are gained from interaction with different types of organizations. Benefits of interaction include social support and movement towards shared goals. However, different types of interaction provide different benefits. Interaction with churches primarily provides information, whereas interaction with service organizations acts to fulfill some aspect of the Christian mandate to help the needy (social services are the most interacted with organizations). Thus, the most common type of interaction with social services was finance. Furthermore, different churches interact with other local organizations to different degrees and share different types of resources. Time and resource commitment costs are outweighed by the benefits of social support and shared goals. Nonetheless, costs and benefits are different for different members of the network and different members of the network play different roles in resource exchange 7.1.4 Describing Network Structure Proposition 3 expected that organizations in different locations in the network and organizations with different characteristics would play different roles in the network. I discovered that the location of a church in the network defined whether it was part of the core, not its characteristics, or the type of interaction it had with other organizations. Core organizations both send and receive connections. Periphery organizations rarely send or [ Page ] 187 receive connections. For example, there was little difference in how core churches and periphery churches interacted with the rest of the network. Periphery churches were less connected to the network. However, the connections they did have were similar to those of the core churches. However, there was a significant difference between direct and indirect network connections. For example, there are very few direct connections between churches. Therefore, most members of the Direct matrix were not close enough to share resources. However, common connections to other local organizations were plentiful. In theory, such connections could connect churches that are otherwise far apart in the network. Direct connections do identify loose groupings based on theology or denomination. More surprisingly, churches of all varieties interact with the same social service organizations. As noted earlier, those social service organizations are more likely to receive connections than any other type of organization. Aspects of network structure such as network role and reasons for network involvement are different for different organizations. The way in which those roles and reasons are worked out depends on individual churches and their network connections. Churches interact based on similarity but they also interact with churches that are not at all like themselves. This interaction is often based on functional considerations such as supporting organizations which accomplish some aspect of their goals. There is evidence that resources may be easily transferable within the network but this depends on how easily resources flow through shared connections. That is, while organizations have few direct connections they have many shared connections with third-parties. Explicit attention to these shared connections could reduce the costs of accomplishing collective goals. For example, if two churches realize that they were both supporting the same social service they could coordinate their fund raising activities to avoid raising funding from the same people at the same time. 7.2 Looking Ahead 7.2.1 Implications The data from this study contribute to current understandings of how evangelical organizations function by providing information about how evangelical churches interact with each other and with other organizations in a local community. Existing studies show [ Page ] 188 interconnections between evangelicals from the perspective of individual organizations or particular networks. I demonstrated that a complete understanding of the features of Canadian evangelical organizations is only possible in the context of an understanding of the structural environment in which they operate. I also explored relationships between evangelicals and other types of Protestant organizations. I used a single region as a “social laboratory wherein we can observe social structure and social processes in detail.” (Galaskiewicz 1979:38). Network analysis has been applied to voluntary associations but only rarely to religious organizations. Using it to study Canadian evangelicals provides a structural context for historical studies of evangelicals. Furthermore, my research demonstrates that a network approach to the study of religious organizations can find patterns in a large amount of data that cannot be easily seen by concentrating on the individual characteristics of organizations alone. The data also shows that studies of structural relations among religious organizations within local communities are important for showing the details of how religious organizations actually work and relate to each other. Evangelicals have been thought of as isolated and insular. However, the data shows that they are highly involved in their local communities.115 Surprisingly, evangelicals are just as involved with local social service organizations as churches with a less conservative theology. My study also confirms Luhmann’s observation that different social systems are not necessarily coextensive (Beyer 1994:64). Interaction networks involving face-to-face contact among individuals are not necessarily the same as organizational systems (e.g., churches). That is, churches and their connections are not just a function of interaction among church members from different organizations. Church members can interact without connecting their organizations. Furthermore, organizations which may be connected through interaction among individuals can find that those connections take on a life of their own, independent of face-to-face interaction. Thus, interconnection among churches happens at several levels. At the level of the organization there is more concern for boundaries (who is in and who is out). Therefore, there are few direct connections among churches. However, the interaction system is not necessarily constrained by these boundaries, and thus contains a wide range of organizations. As a result, I found that very theologically different churches interact with the same social _____________________________________ 115 This provides confirmation of the impressions of the historical literature and the interviewees. [ Page ] 189 service organizations. A further implication of this finding is that interaction with a social service organization is not directly translated into part of the boundary that separates one church from another. 7.2.2 Avenues for Future Research The data point to some useful avenues for future research. A study exploring national evangelical networks within and emanating from Canada could confirm the conclusions of this study. It could also provide a structural framework on which historical studies could be hung so that they could be understood in relation to the larger story of global evangelical interaction. Second, although this study described the structure of a current evangelical network, it did not explain its origins and could only guess at how it is maintained. A history of the development of the evangelical community and its networks would fill in this gap. A description of changes in the global environment may also explain the drive to intensify identity and co-operate with different others at the same time. The fact that this network is maintained points to a third area for future research: the relationship between networks among evangelicals and Canadian culture. How does the network of evangelical organizations help evangelicals relate to and resist the forces of Canadian culture? Is it the network or the worldview that aids in resisting the culture? What exactly is the relationship between Canadian evangelicalism and Canadian culture? Does Canadian evangelicalism have an effect on Canadian culture? Fourth, this study could profitably be extended through comparisons of the findings in Kitchener-Waterloo with related studies in other communities (e.g., Ammerman et al. 1997; Smith 1997). It would also benefit from a more detailed study of Kitchener-Waterloo to determine exactly what evangelicals are doing and whether the limited number of direct connections is intentional or incidental. Likewise, a comparison of my findings with Canadian demographic data on other communities across the country may help to situate my findings in the national environment. Finally, historical studies suggest evangelicals once played a more prominent role in Canadian society (Westfall 1989; Noll 1992:246). A useful study would explore what precipitated this change from the perspective of a single community. For example, some sociologists have recently suggested that secularization should be understood as a decline in [ Page ] 190 religious authority (Chaves 1994; Yamane 1997). Is the tendency of evangelicals to connect as individuals, not as organizations, a reflection of declining confidence in religious authority and, as a result, declining societal influence? 7.3 Concluding Remarks My network analysis of local Protestant churches is sufficient to demonstrate that evangelical churches are not isolated from each other, nor from the communities in which they operate. The data leads to four important conclusions: • There is significant diversity within the local Protestant religious community (e.g., denomination, church size, evangelicalness, etc.). • Interconnections, which are primarily internal to Kitchener-Waterloo, can relate to organizational similarity but just as often, cross-denominational, theological, and size boundaries. • Reasons for interaction are often functional, rather than based on characteristic similarity. • Direct connections are relatively sparse but indirect connections—primarily through a small number of service organizations—are plentiful. The first of these conclusions is not terribly surprising. However, the last three conclusions indicate that Canadian Protestant organizations do not interact or divide exactly as popular impressions suggest. Organizations of different theological stripes share some commonalties both in terms of networks and in terms of characteristics. What is most striking about these conclusions is the degree to which evangelical organizations function like non- evangelical organizations. Evangelical organizations may favour those they are closest to and become most like those they are closest to but they can move in more than one direction at once. They build up boundaries based on theology and build bridges based on social concern at the same time.116 Furthermore, it is clear that Canadian evangelicals are no longer, if they ever were, completely isolated from each other and from the communities in which they operate. ______________________________________ 116 Smith et al. (1998:36ff) note that American evangelicals are highly involved in the same kind of social action I observed in Kitchener-Waterloo even if their individualism keeps them from understanding the structural causes of the problems they are working to address (Smith et al. 1998:178ff). [ Page ] 191 Finally, the lack of a clear division between how “evangelical” churches act as compared to “mainline” churches suggests that the problem may be with those categories themselves. As evangelicals are best understood polythetically, so too are Canadian Christians in general. Contemporary Canadian Christians do not line up in two distinct camps opposite in focus in all situations. Rather, Canadian Christians divide and interact in different manners depending on why they are interacting. As my propositions were based on the assumption of a distinctive evangelical network that was primarily internally connected they only half captured the networks that I actually found. For the location of my research, the similarities in network interaction are more significant than the differences and network characteristics are more important predictors of interaction than organizational characteristics. BIBLIOGRAPHY Airhart, P. D. 1992. Serving the Present Age: Revivalism, Progressivism, and the Methodist Tradition in Canada. Kingston: McGill-Queen’s University Press. Airhart, P. D. 1997. “As Canadian as Possible Under the Circumstances:” Reflections on the Study of Protestantism in North America. In New Directions in American Religious History, edited by D. G. Hart, and H. S. Stout. 116-140. New York: Oxford University Press. Alexander, D., and S. Alexander, eds. 2000. The Shepherd’s Guide: Christian Business and Ministry Directory. Brampton: Dean and Shelley Alexander. Ammerman, N. T. 1982. Operationalizing Evangelicalism: an Amendment. Sociological- Analysis 43: 170-171. Ammerman, N. T. et al. 1997. Congregation & Community. New Brunswick: Rutgers University Press. Angus Reid Group. 1996. God and Society in North America : a Survey of Religion, Politics and Social Involvement in Canada and the United States. Toronto: Angus Reid Group. Austin, A. J. 1997. The Transplanted Mission: the China Inland Mission and Canadian Evangelicalism. In Aspects of the Canadian Evangelical Experience, edited by G. A. Rawlyk. 351-368. Montreal: McGill-Queen's University Press. Baker, W. E. 1990. Market Networks and Corporate Behaviour. American Journal of Sociology 96: 589-625. Balswick, J., and J. Balswick. 1990. Adam and Eve in America. Christianity Today Bassili, J. N., and B. S. Scott. 1996. Response Latency as a Signal to Question Problems in Survey Research. Public Opinion Quarterly 60: 390-399. Baum, G. 1991. The Church in Quebec. Ottawa: Novalis. Bebbington, D. W. 1989. Evangelicalism in Modern Britiain: a History From the 1730s to the 1980s. London: Unwin Hyman Ltd. Bebbington, D. W. 1994. Evangelicalism in Its Settings: the British and American Movements Since 1940. In Evangelicalism: Comparative Studies of Popular Protestantism in North America, the British Isles, and Beyond, 1700-1990, edited by M. A. Noll, D. W. Bebbington, G. A. Rawlyk. 365-388. New York: Oxford University Press. Bebbington, D. W. 1997. Canadian Evangelicalism: a View From Britain. In Aspects of the Canadian Evangelical Experience, edited by G. A. Rawlyk. 38-54. Montreal: McGill- Queen's University Press. Berg, T. C. 2000. “Too Good to be True”: the New Era Foundation Scandal and Its Implications. In More Money, More Ministry: Money and Evangelicals in Recent North America History, edited by L. Eskridge, and M. A. Noll. 374-395. Grand Rapids: Eerdmans. Berger, P. L. 1967. The Sacred Canopy: Elements of a Sociological Theory of Religion. New York: Doubleday. [ Page ] 192 [ Page ] 193 Berkowitz, S. D. 1982. An Introduction to Structural Analysis: the Network Approach to Social Research. Toronto: Butterworths. Beyer, P. 1994. Religion and Globalization. London, UK: Sage. Bibby, R. W. 1986. Anglitrends: A Profile and Prognosis. Toronto: Anglican Diocese of Toronto. Bibby, R. W. 1987. Fragmented Gods: the Poverty and Potential of Religion in Canada. Toronto: Irwin. Bibby, R. W. 1993. Unknown Gods: the Ongoing Story of Religion in Canada. Toronto: Stoddart. Bibby, R. W. 1994. Unitrends. Toronto: Department of Stewardship, United Church of Canada. Bibby, R. W. 1995. Evangeltrends. Waterloo, Ontario: Vision 2000 Canada. Bibby, R. W. 1999. The Alliance Future Survey Summary Report. Willowdale: The Christian and Missionary Alliance in Canada. Bibby, R. W., and M. B. Brinkerhoff. 1983. Circulation of the Saints Revisited: a Longitudinal Look At Conservative Church Growth. Journal for the Scientific Study of Religion 22: 253-262. Bibby, R. W., and M. B. Brinkerhoff. 1994. Circulation of the Saints 1966-1990: New Data, New Reflections. Journal for the Scientific Study of Religion 33: 273-280. Bibby, R. W., 1943. 2002. Restless Gods : the Renaissance of Religion in Canada. Toronto: Stoddart. Bienenstock, E. J. et al. 1990. The Effect of Network Density and Homogeneity on Attitude Polarization. Social Networks 12: 153-172. Bishop, G. F. 1987. Experiments With the Middle Response Alternative in Survey Questions. Public Opinion Quarterly 51: 220-232. Bonacich, P. 1987. Power and Centrality: a Family of Measures. American Journal of Sociology 92: 1170-1182. Borgatti, S. P., and M. G. Everett. 2000. Models of Core/Periphery Structures. Social Networks 21: 375-395 Borgatti, S. P. et al. 1999. Ucinet 5 Version 1.00. Nantik: Analytic Technologies. Bowker, G. C., and S. L. Star. 1999. Sorting Things Out : Classification and Its Consequences. Cambridge: MIT Press. Brass, D. J., and M. E. Burkhardt. 1992. Centrality and Power in Organizations. In Networks and Organizations: Structure, Form and Action, edited by N. Nohria, and R. G. Eccles. 191-215. Boston, Mass.: Harvard Business School Press. Breiger, R. L. 1988. The Duality of Persons and Groups. In Social Structures a Network Approach, edited by B. Wellman, and S. D. Berkowitz. 83-98. Cambridge: Cambridge University Press. Brewer, D. D., and B. L. Yang,. 1994. Patterns in the Recall of Persons in a Religious Community. Social Networks 16: 347-379. Bruce, S. 1993. Religion and Rational Choice: a Critique of Economic Explanations of Religious Behaviour. Sociology of Religion 54: 193-205. Bryant, J. M. 2000. Cost-Benefit Accounting and the Piety Business: is Homo Religiosus, At Bottom, a Homo Economicus. Method & Theory in the Study of Religion. [ Page ] 194 Brym, R. J. 1988. Structural Location and Ideological Divergence: Jewish Marxist Intellectuals in Turn-of-the Century Russia. In Social Structures: A Network Approach, edited by B. Wellman, and S. D. Berkowitz. 359-379. Cambridge: Cambridge University Press. Burkinshaw, R. K. 1995. Pilgrims in Lotus Land: Conservative Protestantism in British Columbia 1917-1981. Montreal: McGill-Queen's University Press. Burkinshaw, R. K. 1996. Aspects of Canadian Evangelical Historiography. Studies in Religion - Sciences Religieuses 25: 7-20. Burkinshaw, R. K. 1997. Evangelical Bible Colleges in Twentieth-Century Canada. In Aspects of the Canadian Evangelical Experience, edited by G. A. Rawlyk. 369-385. Montreal: McGill-Queen's University Press. Burt, R. S. 1992. Structural Holes Check Date. Cambridge: Harvard University Press. Burt, R. S. 1993. The Social Structure of Competition. In Explorations in Economic Sociology, edited by R. Swedberg. 65-103. New York: Russell Sage. Burt, R. S. et al. 1994. Contingent Organization as a Network Theory: the Culture- Performance Contingency Function. Acta Sociologica 37: 345-370. Calloway, M. et al. 1993. Accuracy and Reliability of Self-Reported Data in Interorganizational Networks. Social Networks 15: 377-398. Carpenter, J. A. 1984. From Fundamentalism to the New Evangelical Coalition. In Evangelicalism and modern America, edited by G. M. Marsden. 3-16. Grand Rapids, MI: Eerdmans. Cartwright, A. 1988. Interviews Or Postal Questionnaires? Comparisons of Data About Women's Experiences With Maternity Services. Milbank Quarterly 66: 172-189. Chapman, M. D. 1994. Rebuilding the Broken Wall: The Evangelical Fellowship of Canada and Canadian Evangelicals. Waterloo: Wilfrid Laurier University. Chaves, M. 1994. Secularization as Declining Religious Authority. Social Forces 72: 749- 774. Chebat, J. C., and J. Picard. 1991. Does Prenotification Increase Response Rates in Mail Surveys? a Self-Perception Approach. Journal of Social Psychology 131: 477-481. Christie, N. 1990. In These Time of Democratic Rage and Delusion”: Popular Religion and the Challenge to the Established Order, 1760-1815. In The Canadian Protestant Experience, edited by G. G. Rawlyk. 9-47. Burlington: Welch Publishing Company. Clark, H. H., and M. F. Schober. 1992. Asking Questions and Influencing Answers. In Questions About Questions: Inquiries into the Cognitive Bases of Surveys, edited by J. M. Tanur. 15-48. New York: Russell Sage Foundation. Clark, S. D. 1948. Church and Sect in Canada. Toronto: University of Toronto Press. Clarke, B. 1996. English-Speaking Canada From 1854. In A Concise History of Christianity in Canada, edited by T. Murphy, and R. Perin. 261-359. Toronto: Oxford. Clifford, N. K. 1977. His Dominion: A Vision in Crisis. In Religion and Culture in Canada/Religion et Culture au Canada, edited by P. Slater. 24-41. Waterloo: Corporation Canadienne des Sciences Religieuses/Canadian Corporation for Studies in Religion. Cook, K. 1977. Exchange and Power in Networks of Interorganizational Relations. Sociological Quarterly 18: 62-82. Corelli, R. 1996. How Very Different We Are. Maclean's 36-40. [ Page ] 195 Couper, M. P. 1997. Survey Introductions and Data Quality. Public Opinion Quarterly 61: 317-338. Crysdale, S. 1976. Social Awakening Among Protestants, 1872 to 1918. In Religion in Canadian Society, edited by S. Crysdale, and L. Wheatcroft. 191-206. Toronto: Macmillan. Crysdale, S., and L. Wheatcroft, eds. 1976. Religion in Canadian Society. Toronto: Macmillan. Dawson, L. L. 1992. Church/Sect Theory: Getting it Straight. North American Religion 1: 5- 28. Dayton, D. W. 1991. Some Doubts About the Usefulness of the Category "Evangelical". In The Variety of American Evangelicalism, edited by D. W. Dayton, and R. K. Johnston. 245-251. Knoxville: The University of Tennessee Press. Dayton, D. W., and R. K. Johnston, eds. 1991. The Variety of American Evangelicalism. Knoxville: The University of Tennessee Press. Diani, M. 1990. The Network Structure of the Italian Ecology Movement. Social Science Information 29: 5-31. Dillman, D. A. 1983. Mail and Other Self-Administered Questionnaires. In Handbook of Survey Research, edited by P. H. Rossi J. D. Wright, A. B. Anderson. 359-377. New York: Academic Press. Dillman, D. A. et al. 1993. Effects of Questionnaire Length, Respondent-Friendly Design, and a Difficult Question on Response Rates for Occupant-Addressed Census Mail Surveys. Public Opinion Quarterly 57: 289-304. DiMaggio, P. J., and W. W. Powell. 1983. The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields. American Sociological Review 48: 147-160. Dueck, L. et al. 1993. Women in Ministry. Faith Today November/December: 18-33. Ellingsen, M. 1988. The Evangelical Movement: Growth, Impact, Controversy, Dialog. Minneapolis: Augsburg Publishing House. Elliot, D. R. 1994. Knowing No Borders: Canadian Contributions to American Fundamentalism. In Amazing Grace: Evangelicalism in Australia, Britain, Canada and the United States, edited by G. A. Rawlyk, and M. A. Noll. 349-374. Montreal: McGill-Queen's University Press. Ellison, C. G., and L. K. George. 1994. Religious Involvement, Social Ties, and Social Support in a Southeastern Community. Journal for the Scientific Study of Religion 33: 46-61. Eskridge, L. 1994. Bibliographic Afterword: an Introductory Guide to the Literature of Comparative Evangelical History. In Amazing Grace: Evangelicalism in Australia, Britain, Canada and the United States, edited by G. A. Rawlyk, and M. A. Noll. 401- 409. Montreal: McGill-Queen's University Press. Faith Today. 1990. Advertisment. Faith Today January/February: 21. Feld, S. L. 1982. Social Structural Determinants of Similarity Among Associates. American Sociological Review 47: 797-801. Feld, S. L. 1997. Elicitation Problems, Cont. Comment on SOCNET Mailing List. The Financial Post Corporation. 1998. 1997/98 Canadian Markets, 71st Edition. Toronto: Maclean-Hunter. [ Page ] 196 Finke, R., and R. Stark. 1992. The Churching of America 1776-1990: Winners and Losers in Our Religious Economy. New Brunswick, NJ: Rutgers University Press. Foddy, W. 1993. Constructing Questions for Interviews and Questionnaires: Theory and Practice in Social Research. Cambridge: Cambridge University Press. Freeman, L. 1979. Centrality in Social Networks: Conceptual Clarification. Social Networks 1: 215-239. Friedmann, H. 1979. Are Distributions Really Structures? A Critique of the Methodology of Max Weber. Connections: Bulletin of the International Network for Socal Network Analysis 2: 72-80. Galaskiewicz, J. 1979. Exchange Networks and Community Politics. Beverly Hills: Sage. Galaskiewicz, J. 1979. The Structure of Community Organizational Networks. Social Forces 57: 1346-1364. Galaskiewicz, J. 1985. Interorganizational Relations. Annual Review of Sociology 11: 281- 304 Galaskiewicz, J. 1985. Social Organization of an Urban Grants Economy: a Study of Business Philanthropy and Nonprofit Organizations. Orlando: Academic Press. Galaskiewicz, J., and K. R. Krohn. 1984. Positions, Roles, and Dependencies in a Community Interorganization System. Sociological Quarterly 25: 527-550. Galaskiewicz, J., and P. V. Marsden. 1978. Interorganizational Resource Networks: Formal Patterns of Overlap. Social Science Research 7: 89-107. Galaskiewicz, J., and S. Wasserman. 1993. Social Network Analysis: Concepts, Methodology, and Directions for the 1990s. Sociological Methods and Research 22: 3-22. Gauvreau, M. 1991. The Evangelical Century: College and Creed in English Canada From the Great Revival to the Great Depression. Montreal: McGill-Queen’s University Press. Gauvreau, M. 1994. The Empire of Evangelicalism: Varieties of Common Sense in Scotland, Canada and the United States. In Evangelicalism: Comparative Studies of Popular Protestantism in North America, the British Isles, and Beyond, 1700-1990, edited by M. A. Noll, D. W. Bebbington, G. A. Rawlyk. 219-252. New York: Oxford University Press. Gerstner, J. H. 1977. Theological Boundaries: the Reformed Perspective. In The Evangelicals: What They Believe. Who They Are, Where They Are Changing, 21-37. Grand Rapids: Baker. Goyder, J. 1985. Face-to-Face Interviews and Mailed Questionnaires: the Net Difference in Response Rate. Public Opinion Quarterly 49: 234-252. Graham, R. 1990. God’s Dominion: a Sceptic’s Quest. Toronto: McClelland & Stewart. Grant, J. W. 1963. The Churches and the Canadian Experience: a Faith and Order Study of the Christian Tradition. Toronto: Ryerson Press. Grant, J. W. 1972. The Church in the Canadian Era: the First Century of Confederation. Toronto: McGraw-Hill Ryerson. Grant, J. W. 1976. A Decade of Ferment: Canadian Churches in the 1960s. In Religion in Canadian Society, edited by S. Crysdale, and L. Wheatcroft. 207-218. Toronto: Macmillan. Grant, J. W. 1998. The Church in the Canadian Era: the First Century of Confederation, Updated and Expanded Edition. Vancouver: Regent College Publishing. [ Page ] 197 Greenshields, M., and T. Robinson. 1992. A Modern Exodus and a New Church: an Interview With Leaders of the Congregational Christian Churches in Canada. North American Religion 1: 29-52. Grenville, A. S. 1995. Counting the Sheep: Using the Christian Evangelicalism Scale to Identify Evangelicals. Toronto: Angus Reid Group. Grenville, A. S. 1995. Development of the Christian Evangelicalism Scale. Toronto: Angus Reid Group. Grenville, A. S. 1997. The Awakened and the Spirit-Moved: the Religious Experiences of Canadian Evangelicals in the 1990s. In Aspects of the Canadian Evangelical Experience, edited by G. A. Rawlyk. 417-431. Montreal: McGill-Queen's University Press. Guenther, B. L. 2001. Training for Service: the Bible School Movement in Western Canada, 1909-1960. Montreal: McGill University. Guenther, B. L. 2001. “Monuments to God’s Faithfulness”: Mennonite Brethren Bible Schools in Western Canada, 1913-1960. Direction 30(1): 21-32. Gulati, R. 1995. Social Structure and Alliance Formation Patterns: a Longitudinal Analysis. Administrative Science Quarterly 40: 619-652. Hammond, P. E., and J. D. Hunter. 1984. On Maintaining Plausibility: the Worldview of Evangelical College Students. Journal for the Scientific Study of Religion 23: 221- 238. Handy, R. T. 1976. A History of the Churches in the United States and Canada. New York: Oxford University Press. Heberlein, T. A., and R. Baumgartner,. 1978. Factors Affecting Response Rates to Mailed Questionnaires: a Quantitative Analysis of the Published Literature. American Sociological Review 43: 447-462. Hensley, W. E. 1974. Increasing Response Rate By Choice of Postage Stamps. Public Opinion Quarterly 38: 280-283. Herman, N. J. 1984. Conflict in the Church: a Social Network Analysis of an Anglican Congregation. Journal for the Scientific Study of Religion 23: 60-74. Hexham, I. 1993. Canadian Evangelicals: Facing the Critics. In The Sociology of Religion: A Canadian Focus, edited by W. E. Hewitt. 289-302. Toronto: Butterworths. Hiller, H. H. 1976. Alberta and the Bible Belt Stereotype. In Religion in Canadian Society, edited by S. Crysdale, and L. Wheatcroft. 372-383. Toronto: Macmillan. Hindmarsh, D. B. 1997. The Winnipeg Fundamentalist Network, 1910-1940: the Roots of Transdenominational Evangelicalism in Manitoba and Saskatchewan. In Aspects of the Canadian Evangelical Experience, edited by G. A. Rawlyk. 303-319. Montreal: McGill-Queen's University Press. Hughes, R. T. 1991. Are Restorationists Evangelicals? In The Variety of American Evangelicalism, edited by D. W. Dayton, and R. K. Johnston. 109-134. Knoxville: University of Tennessee Press. Hunter, J. D. 1981. Operationalizing Evangelicalism: a Review, Critique and Proposal. Sociological Analysis 42: 363-372. Hunter, J. D. 1987. Evangelicalism: the Coming Generation. Chicago: University of Chicago Press. Iannaccone, L. R. 1994. Why Strict Churches Are Strong. American Journal of Sociology 99: 1180-1211. [ Page ] 198 Jantz, H. 1991. Evangelicals in the Canada of the 90s. Ecumenism 101: 13-16. Jantz, H. 1994. Canada’s Evangelical Face. Christianity Today 38: 41. Jantz, H. 1999. Keeping Company With One Another. Faith Today May/June: 20-28, 31. Johnson, T. P. et al. 1993. Dimensions of Mail Questionnaires and Response Quality. Sociological Focus 26: 271-274. Johnston, R. K. 1991. American Evangelicalism: an Extended Family. In The Variety of American Evangelicalism, edited by D. W. Dayton, and R. K. Johnston. 252-272. Knoxville: University of Tennessee Press. Kalbach, W. E., and W. W. McVey. 1976. Religion Composition of the Canadian Population. In Religion in Canadian Society, edited by S. Crysdale, and L. Wheatcroft. 221-240. Toronto: Macmillan. Kellstedt, L. A. 1989. The Meaning and Measurement of Evangelicalism: Problems and Prospects. In Religion and Political Behaviour in the United States, edited by T. G. Jelen. 3-22. New York: Praeger. Kervin, J. 1992. Methods for Business Research. Orlando: Academic Press Knoke, D. 1993. Networks of Elite Structure and Decision Making. Sociological Methods & Research 22: 23-45. Knoke, D., and J. Kuklinski. 1982. Network Analysis. Beverly Hills: Sage. Knoke, D., and D. Prensky. 1984. What Relevance Do Organization Theories Have for Voluntary Associations? Social Science Quarterly 65: 3-20. Knoke, D., and D. L. Rogers. 1979. A Blockmodel Analysis of Interorganizational Networks. Sociology and Social Research 64: 28-52. Krackhart, D. et al. 1993. Krackplot: a Picture's Worth a Thousand Words. Connections: Bulletin of the International Network for Socal Network Analysis 16: 37-47. LaGarce, R., and J. Washburn. 1995. An Investigation Into the Effects of Questionnaire Format and Color Variations on Mail Survey Response Rates. Journal of Technical Writing and Communication 25: 57-70. Lipset, S. M. 1990. Continental Divide: the Values and Institutions of the United States and Canada. New York, NY: Routledge. Lofland, J., and L. H. Lofland. 1995. Analyzing Social Settings: a Guide to Qualitative Observation and Analysis. Belmont: Wadsworth. Luhmann, N. 1982. The Differentiation of Society. New York: Columbia University Press. Luhmann, N. 1984. Religious Dogmatics and the Evolution of Societies. Lewiston, New York: Edwin Mellen Press. Lyon, D., and D. Van, Marguerite, eds. 2000. Rethinking Church, State, and Modernity: Canada Between Europe and America. Toronto: University of Toronto Press. Mackey, L. 1995. These Evangelical Churches of Ours. Winfield, BC: Wood Lake Books. Maclean's. 1997. Maclean's Universities 1997. Maclean's 42-45. Marsden, G. M. 1984. Introduction: the Evangelical Denomination. In Evangelicalism and Modern America, edited by G. M. Marsden. vii - xix. Grand Rapids, MI: Eerdmans. Marsden, G. M., ed. 1984. Evangelicalism and Modern America. Grand Rapids, MI: Eerdmans. Marsden, G. M. 1987. The Evangelical Denomination. In Piety & Politics: Evangelicals and Fundamentalists Confront the World, edited by J. N. Richard, and M. Cromartie. 55- 68. Lanham, MD: Ethics and Public Policy Center. [ Page ] 199 Marsden, G. M. 1991. Fundamentalism and American Evangelicalism. In The Variety of American Evangelicalism, edited by D. W. Dayton, and R. K. Johnston. 22-35. Knoxville: University of Tennessee Press. Marsden, G. M. 1991. Understanding Fundamentalism and Evangelicalism. Grand Rapids: Eerdmans. Marsden, P. V., and N. E. Friedkin. 1993. Network Studies of Social Influence. Sociological Methods & Research 22: 127-151. Marsden, P. V., and J. S. Hurlbert. 1987. Small Networks and Selectivity Bias in the Analysis of Survey Network Data. Social Networks 9: 333-349. Marty, M. E. 1977. Tensions Within Contemporary Evangelicalism: a Critical Appraisal. In The Evangelicals: What They Believe. Who They Are, Where They Are Changing, edited by D. F. Wells, and J. D. Woodbridge. 190-208. Grand Rapids, MI: Baker. Maton, K. I. 1989. Community Settings as Buffers of Life Stress? Highly Supportive Churches, Mutual Help Groups, and Senior Centers. American Journal of Community Psychology 17: 203-232. McPherson, J. M. 1982. Hypernetwork Sampling: Duality and Differentiation Among Voluntary Organizations. Social Networks 3: 225-249. McPherson, J. M., and L. Smith-Lovin. 1987. Homophily in Voluntary Associations. American Sociological Review 52: 370-379. Miller, D. C. 1991. Handbook of Research Design and Social Measurement, Fifth Edition. Newbury Park, CA: Sage. Mintz, B., and M. Schwartz. 1981. Interlocking Directorates and Interest Group Formation. American Sociological Review 46: 851-869. Mish, F. C., ed. 1989. Webster's Ninth New Collegiate Dictionary. Markham, ON: Thomas Allan & Son Limited. Mizruchi, M. S., and J. Galaskiewicz. 1993. Networks of Interorganizational Relations. Sociological Methods & Research 22: 46-70. Moberg, D. O. 1977. Fundamentalists and Evangelicals in Society. In The Evangelicals: What They Believe. Who They Are, Where They Are Changing, edited by D. F. Wells, and J. D. Woodbridge. 163-189. Grand Rapids, MI: Baker. Mol, H. 1985. Faith and Fragility: Religion and Identity in Canada. Burlington: Trinity. Moore, G. 1979. The Structure of a National Elite Network. American Sociological Review 44: 673-692. Motz, A., ed. 1990. Reclaiming a Nation: the Challenge of Re-Evangelizing Canada By the Year 2000. Richmond, British Columbia: Church Leadership Library. Murphy, T., and R. Perin, eds. 1996. A Concise History of Christianity in Canada. Toronto: Oxford. Nash, R. H. 1987. Evangelicals in America: Who They Are, What They Believe. Nashville: Abingdon Press. Nederhof, A. J. 1983. Effects of Repetition and Consistency of Personalization Treatments on Response Rate in Mail Surveys. Social Science Research 12: 1-9. Nederhof, A. J. 1988. Effects of a Final Telephone Reminder and Questionnaire Cover Design in Mail Surveys. Social Science Research 17: 353-361. Needham, R. 1975. Polythetic Classification: Convergence and Consequences. Man 10: 349- 369. [ Page ] 200 Niebuhr, H. R. 1957. The Social Sources of Denominationalism. Cleveland and New York: World. No Author. 2001. The Web Site of the Evangelical Fellowship of Canada. http://www.efc- canada.com/. Nohria, N. 1992. Introduction: is a Network Perspective a Useful Way of Studying Organizations? In Networks and Organizations: Structure, Form and Action, edited by N. Nohria, and R. G. Eccles. 1-22. Boston, Mass.: Harvard Business School Press. Nohria, N., and R. G. Eccles, eds. 1992. Networks and Organizations: Structure, Form and Action. Boston MA.: Harvard Business School Press. Noll, M. A. 1991. Christianity in Canada: Good Books At Last. Fides et Historia 23: 80-104. Noll, M. A. 1998. Religion in Canada and the United States : Comparisons From an Important Survey Featuring the Place of Evangelical Christianity. Crux 34: 13-25. Noll, M. A. 1984. Evangelicals and the Study of the Bible. In Evangelicalism and modern America, edited by G. M. Marsden. 103 - 121. Grand Rapids, MI: Eerdmans. Noll, M. A. 1992. A History of Christianity in the United States and Canada. Grand Rapids, MI: Eerdmans. Noll, M. A. 1997. Canadian Evangelicalism: a View From the United States. In Aspects of the Canadian Evangelical Experience, edited by G. A. Rawlyk. 3-20. Montreal: McGill-Queen's University Press. Noll, M. A. D. W. Bebbington, G. A. Rawlyk, eds. 1994. Evangelicalism: Comparative Studies of Popular Protestantism in North America, the British Isles, and Beyond, 1700-1990. New York: Oxford University Press. Noll, M. A., and D. F. Wells, eds. 1988. Christian Faith & Practice in the Modern World: Theology From an Evangelical Point of View. Grand Rapids, MI: Eerdmans. O’Brien, S. 1994. Eighteenth-Century Publishing Networks in the First Years of Transatlantic Evangelicalism. In Evangelicalism: Comparative Studies of Popular Protestantism in North America, the British Isles, and Beyond, 1700-1990, edited by M. A. Noll D. W. Bebbington, G. A. Rawlyk. 38-57. New York: Oxford University Press. O'Toole, R. 1996. Religion in Canada: Its Development and Contemporary Situation. Social Compass 43: 119-134. Payne, S. L. 1951. The Art of Asking Questions. Princeton: Princeton University Press. Phillips, S. D. 1991. Meaning and Structure in Social Movements: Mapping the Network of National Canadian Women's Organizations. Canadian Journal of Political Science 24: 755-782. Piggin, S. 1994. The American and British Contributions to Evangelicalism in Australia. In Evangelicalism: Comparative Studies of Popular Protestantism in North America, the British Isles, and Beyond, 1700-1990, edited by M. A. Noll D. W. Bebbington, G. A. Rawlyk. 290-309. New York: Oxford University Press. Posterski, D. C., and I. Barker. 1993. Where’s a Good Church? Winfield, B.C.: Wood Lake Books. Quast, K., and J. Vissers, eds. 1996. Studies in Canadian Evangelical Renewal: Essays in Honour of Ian S. Rennie. Markham: Faith Today Publications. Ragin, C. C. 2000. Fuzzy-Set Social Science. Chicago: University of Chicago Press. Rawlyk, G. A., ed. 1997. Aspects of the Canadian Evangelical Experience. Montreal: McGill-Queen's University Press. [ Page ] 201 Rawlyk, G. A., ed. 1990. The Canadian Protestant Experience 1760-1990. Burlington, ON: Welch Publishing Company. Rawlyk, G. A. 1990. Politics, Religion, and the Canadian Experience: a Preliminary Probe. In Religion and American Politics: From the Colonial Period to the 1980s, edited by M. A. Noll. 253-280. New York: Oxford University Press. Rawlyk, G. A. 1994. The Canada Fire: Radical Evangelicalism in British North America 1775-1812. Montreal: McGill-Queen's University Press. Rawlyk, G. A. 1996. Is Jesus Your Personal Saviour? in Search of Canadian Evangelicalism in the 1990s. Montreal: McGill-Queen's University Press. Rawlyk, G. A., and M. A. Noll, eds. 1994. Amazing Grace: Evangelicalism in Australia, Britain, Canada and the United States. Montreal: McGill-Queen's University Press. Redekop, J. 1994. Interview. Personal Communication. Reimer, S. H. 1995. A Look At Cultural Effects on Religiosity: A Comparison Between the United States and Canada. Journal for the Scientific Study of Religion 34: 445-457. Reimer, S. H. 1999. A Generic Evangelicalism?: Comparing Evangelical Subcultures in Canada and the USA. Personal Communication. Reimer, S. H. 2003. Evangelicals and the Continental Divide : The Conservative Protestant Subculture in Canada and the United States. Montreal: McGill Queen's University Press. Remus, H. 1992. Bible Colleges. In Religious Studies in Ontario: A State-of-the-Art Review, edited by H. Remus W. C. James, D. Fraikin. 269-286. Waterloo: Wilfrid Laurier University Press for the Canadian Corporation for Studies in Religion/Corporation Canadienne des Sciences Religieuses. Rennie, I. 1984. Gratitude for the Past. Evangelical Recorder 90: 6-11. Rennie, I. 1986. Part Three: Was Canada Ever Christian? Faith Today July/August: 36-37. Rennie, I. 1986. Part Two: Was Canada Ever Christian? Faith Today May/June: 36-37. Rennie, I. 1994. Interview. Personal Communication. Rennie, I. 1994. Fundamentalism and the Varieties of North Atlantic Evangelicalism. In Evangelicalism: Comparative Studies of Popular Protestantism in North America, the British Isles, and Beyond, 1700-1990, edited by M. A. Noll, D. W. Bebbington, G. A. Rawlyk. 333-350. New York: Oxford University Press. Richardson, R. J. 1987. Directorship Interlocks and Corporate Profitability. Administrative Science Quarterly 32: 367-386. Richardson, R. J., and B. Wellman. 1985. Structural Analysis: the State of the Canadian Art. Canadian Review of Sociology and Anthropology 771-793. Robertson, R. 1992. The Economization of Religion? Reflections on the Promise and Limitations of the Economic Approach. Social Compass 39: 147-157. Rosch, E. 1978. Principles of Categorization. In Cognition and categorization, edited by E. Rosch, and B. B. Lloyd. 27-48. Hillsdale, N.J.: L. Erlbaum Associates. Rosch, E., and C. B. Mervis. 1975. Family Resemblances: Studies in the Internal Structure of Categories. Cognitive Psychology 7: 573-605. Rossi, P. H. J. D. Wright, A. B. Anderson, eds. 1983. Handbook of Survey Research. Orlando: Academic Press. Rutman, D. B., and A. H. Rutman. 1994. Small Worlds, Large Questions: Explorations in Early American Social History, 1600-1850. Charlottesville: University Press of Virginia. [ Page ] 202 Schmalzbauer, J. 1999. American Evangelicalism: Embattled and Thriving. Reviewed in Sociology of Religion 60: 336-337. Scott, J. 1991. Centrality and Centralization. In Social Network Analysis: A Handbook, 85- 102. London: Sage. Scott, J. 1991. Social Network Analysis: a Handbook. London: Sage. Shantz, J. A. 1980. Tut-Mania. Thrust 12: 6-7. Sheatsley, P. B. 1983. Questionnaire Construction and Item Wording. In Handbook of Survey Research, edited by P. H. Rossi J. D. Wright, A. B. Anderson. 195-230. New York: Academic Press. Simpson, J. H., and H. MacLeod. 1985. The Politics of Morality in Canada. In Religious Movements: Genesis, Exodus and Numbers, edited by R. Stark. 221-240. New York: Paragon House. Slater, P., ed. 1977. Religion and Culture in Canada/Religion Et Culture Au Canada. Waterloo: Corporation Canadienne des Sciences Religieuses/Canadian Corporation for Studies in Religion. Smith, C. et al. 1998. American Evangelicalism: Embattled and Thriving. Chicago: University of Chicago Press. Smith, C. 2000. Christian America? What Evangelicals Really Want. Berkeley: University of California Press. Smith, G. 1997. Mapping the Voluntary and Community Sector: a Network Analysis in One London Borough. Paper presented at the National Council for Voluntary Organizations Research Conference, London. Smith, J. Z. 1982. Imagining Religion: From Babylon to Jonestown. Chicago: University of Chicago Press. Smith, T. L. 1986. The Evangelical Kaleidoscope and the Call to Christian Unity. Christian Scholar's Review 25: 125-140. Smith, T. L. 1987. Evangelical Christianity and American Culture. In A time to speak: The Evangelical-Jewish Encounter, edited by A. Rudin, and M. Wilson. 58-75. Grand Rapids, MI: Eerdmans. Smith, T. W. 1990. Classifying Protestant Denominations. Review of Religious Research 31: 225-245. SPSS Inc. 2002. SPSS 11 for Mac OS X. Chicago, IL: SPSS Inc.. Stackhouse, J., John G. 1990. The Protestant Experience in Canada Since 1945. In The Canadian Protestant Experience, edited by G. G. Rawlyk. 198-239. Burlington: Welch Publishing Company. Stackhouse, J., John G. 1992. Whose Dominion? Christianity and Canadian Culture Historically Considered. Crux 27: 29-35. Stackhouse, J., John G. 1993. Canadian Evangelicalism in the Twentieth Century: an Introduction to Its Character. Toronto: University of Toronto Press. Stackhouse, J., John G. 1993. The National Association of Evangelicals, the Evangelical Fellowship of Canada, and the Limits of Evangelical Cooperation. Unpublished Manuscript Stackhouse, J., John G. 1994. More Than a Hyphen: Twentieth-Century Canadian Evangelicalism in Anglo-American Context. In Amazing Grace: Evangelicalism in Australia, Britain, Canada and the United States, edited by G. A. Rawlyk, and M. A. Noll. 375-400. Montreal: McGill-Queen's University Press. [ Page ] 203 Stackhouse, J., John G. 1995. The Historiography of Canadian Evangelicalism: a Time to Reflect. Church History 64: 627-634. Stackhouse, J., John G. 1997. Who Whom?”: Evangelicalism and Canadian Society. In Aspects of the Canadian Evangelical Experience, edited by G. A. Rawlyk. 55-70. Montreal: McGill-Queen's University Press. Stark, R., and W. S. Bainbridge. 1996. A Theory of Religion. New Brunswick, NJ: Rutgers University Press. Statistics Canada. 1993. Religions in Canada. Ottawa: Industry, Science and Technology Canada. Statistics Canada. 2003. Religions in Canada. Ottawa: Industry, Science and Technology Canada. Stiller, B. 1993. Interview. Personal Communication. Stonebraker, R. J. 2003. Allocating Local Church Funds to Benevolence: the Impact of Congregational Size. Review of Religious Research 45: 48-58. Stout, D. A., and J. M. Buddenbaum eds. 2001. Religion and Popular Culture : Studies on the Interaction of Worldviews. Ames: Iowa State University Press. Stout, H. S. 1994. George Whitefield in Three Countries. In Evangelicalism: Comparative Studies of Popular Protestantism in North America, the British Isles, and Beyond, 1700-1990, edited by M. A. Noll, D. W. Bebbington, G. A. Rawlyk. 58-72. New York: Oxford University Press. Sudman, S. 1982. Estimating Response to Follow-Ups in Mail Surveys. Public Opinion Quarterly 46: 582-584. Sudman, S. 1983. Applied Sampling. In Handbook of Survey Research, edited by P. H. Rossi J. D. Wright, A. B. Anderson. 145-194. New York: Academic Press. Sudman, S. 1985. Mail Surveys of Reluctant Professionals. Evaluation Review 9: 349-360. Sudman, S. 1988. Experiments in Measuring Neighbor and Relative Social Networks. Social Networks 10: 93-108. Swanson, G. E. 1971. An Organizational Analysis of Collectivities. American Sociological Review 36: Sweet, L. I., ed. 1984. The Evangelical Tradition in America. Macon, GA: Mercer University Press. Swift, A. 1991. A Changing Quebec. Faith Today January/February: 18-24. Swift, A. 1993. A Nation of Private Christians? Faith Today July/August: 20-28. Tanur, J. M., ed. 1992. Questions About Questions: Inquiries Into the Cognitive Bases of Surveys. New York: Russell Sage Foundation. Troeltsch, E. 1931. The Social Teachings of the Christian Churches. London: George Allen and Unwin. Uehara, E. 1990. Dual Exchange Theory, Social Networks, and Informal Social Support. American Journal of Sociology 96: 521-557. Van de Ven, A. H. et al. 1979. Coordination Patterns Within an Interorganizational Network. Human Relations 32: 19-36. Van Die, M., ed. 2001. Religion and Public Life in Canada: Historical and Comparative Perspectives. University of Toronto Press. Walker, R. 1992. Inter-Organizational Linkages as Mediating Structures in Community Health. Health Promotion International 7: 257-264. Walsh, G. 2001. Friends in Other Places. Faith Today July/August: 7. [ Page ] 204 Wasserman, S., and K. Faust. 1994. Social Network Analysis: Methods and Applications. Cambridge: Cambridge University Press. Webber, R. E. 1978. Common Roots: a Call to Evangelical Maturity. Grand Rapids, MI: Zondervan. Weber, M. 1958. The Protestant Ethic and the Spirit of Capitalism. New York: Charles Scribner's Sons. Weber, T. P. 1991. Premillennialism and the Branches of Evangelicalism. In The Variety of American Evangelicalism, edited by D. W. Dayton, and R. K. Johnston. 5-21. Knoxville: University of Tennessee Press. Welch, M. R. 1977. Analyzing Religious Sects: an Empirical Examination of Wilson's Sect Typology. Journal for the Scientific Study of Religion 16: 125-141. Wellman, B. 1979. The Community Question: the Intimate Networks of East Yorkers. American Journal of Sociology 84: 1201-1231. Wellman, B. 1988. Structural Analysis: From Method and Metaphor to Theory and Substance. In Social Structures: A Network Approach, edited by B. Wellman, and S. D. Berkowitz. 19-61. Cambridge: Cambridge University Press. Wellman, B. 1988. The Community Question Re-Evaluated. In Power, Community and the City, edited by M. P. Smith. 81-107. New Brunswick, NJ: Transaction. Wellman, B. 1992. Which Types of Ties and Networks Provide What Kinds of Social Support? Advances in Group Processes 9: 207-235. Wellman, B., ed. 1999. Networks in the Global Village. Boulder, CO: Westview Press. Wellman, B. 2001. Physical Place and Cyber Place: the Rise of Personalized Networking. International Journal of Urban and Regional Research 25, 2: 227-52 . Wellman, B., and S. D. Berkowitz. 1988. Introduction: Studying Social Structures. In Social Structures a Network Approach, edited by B. Wellman, and S. D. Berkowitz. 1-14. Cambridge: Cambridge University Press. Wellman, B., and S. D. Berkowitz, eds. 1988. Social Structures: a Network Approach. Cambridge: Cambridge University Press. Wellman, B., and M. Gulia. 1999. The Network Basis of Social Support: a Network is More Than the Sum of Its Ties. In Networks in the Global Village, edited by B. Wellman. 83-118. Boulder, CO: Westview Press. Wellman, B., ed. 1999. The Elements of Personal Communities. Networks in the Global Village. Boulder, CO: Westview Press. Wellman, B., and S. Potter,. 1994. The Elements of Personal Community Networks. In American Sociological Association, Wellman, B., and S. Wortley. 1990. Different Strokes From Different Folks: Community Ties and Social Support. American Journal of Sociology 96: 558-588. Wells, D. F. 1987. No Offense: I Am an Evangelical: a Search for Self-Definition [response By T Askew, Pp 41-44]. In A time to speak: The Evangelical-Jewish Encounter, edited by A. Rudin, and M. Wilson. 20-40. Grand Rapids, MI: Eerdmans. Wells, D. F. 1994. On Being Evangelical: Some Theological Differences and Similarities. In Evangelicalism: Comparative Studies of Popular Protestantism in North America, the British Isles, and Beyond, 1700-1990, edited by M. A. Noll, D. W. Bebbington, G. A. Rawlyk. 389-410. New York: Oxford University Press. Wells, D. F., and J. D. Woodbridge, eds. 1977. The Evangelicals: What They Believe. Who They Are, Where They Are Changing. Grand Rapids, MI: Baker. [ Page ] 205 Westfall, W. 1989. Two Worlds: the Protestant Cultures of Nineteenth Century Ontario. Montreal: McGill-Queen's University Press. Wind, J. P., and J. W. Lewis. 1994. American Congregations, Volume 1: Portraits of Twelve Religious Communitie. Chicago: University of Chicago Press. Wineburg, R. J. 1990. A Community Study of the Ways Religious Congregations Support Individuals and the Local Human Services Network. Journal of Applied Social Sciences 15: 51-74. Wineburg, R. J. 1993. Social Policy, Community Service Development, and Religious Organizations. Nonprofit Management and Leadership 3: 283-297. Wittgenstein, L. 1963. Philosophical Investigations. Oxford: Basil Blackwell & Mott, Ltd.. Wright, E. R., and I. M. Shuff. 1995. Specifying the Integration of Mental Health and Primary Health Care Services for Persons With HIV/AIDS: the Indiana Integration of Care Project. Social Networks 17: 319-340. Wright, R. A. 1990. The Canadian Protestant Tradition 1914-1945. In The Canadian Protestant Experience, edited by G. G. Rawlyk. 139-197. Burlington: Welch Publishing Company Wuthnow, R. 1976. The Consciousness Reformation. Berkeley: University of California Press. Yamane, D. 1997. Secularization on Trial: in Defense of a Neosecularization Paradigm. Journal for the Scientific Study of Religion 36: 109-122. Yammarino, F. J. et al. 1991. Understanding Mail Survey Response Behavior: a Meta- Analysis. Public Opinion Quarterly 55: 613-639. Yang, F. 1998. Chinese Conversion to Evangelical Christianity: the Importance of Social and Cultural Contexts. Sociology of Religion 59: 237-257. Zorn, K. 1994. Interview. Personal Communication. APPENDICES Appendix A - Background Interview Questions To assure that the interview questions were clear, unambiguous, relevant, interesting, non-offensive, and easy to understand they were tested and refined through four levels of pre-tests. They were first tested on colleagues to assure that none of the proper conventions of academia were overlooked. Second, they were tested on relatives and non-academic friends in an attempt to determine the intelligibility of the survey to those outside of higher education. Third, they were tested on topic experts who helped to verify that the interview examined what it was supposed to be examining. Method experts were given the final pre- test to assure that the survey collected the sort of data it was supposed to be collecting. Nonetheless, at the time of the interview, questions were shaped by a given informants’ particular knowledge and were used primarily as guidelines for the direction of questioning rather than as a rigid script. This project was designed to collect qualitative data rather than to obtain quantitatively comparative data. Nonetheless, the testing and implementation of the data collection was sufficient to ensure that the informants had a common understanding of the questions and that the data was comparable. Questions: Religious Organizations: 1. In your view, which religious organizations in Kitchener-Waterloo have relationship with a large number of other local religious and community organizations? 2. a. Do local religious organizations co-operate with each other and with other organizations in Kitchener-Waterloo? b. In what ways? c. Is this co-operation extensive or limited to a few times a year? 3. a. From your perspective, which religious organizations are most involved in the community? b. Which religious organizations play similar roles in the community (e.g., coordinating social services, organizing community events, etc.)? c. Which religious organizations are most prominent? [ Page ] 206 [ Page ] 207 Conservative/Evangelical Protestants: 4. To what degree are conservative/evangelical Protestant religious organizations involved in the community? 5 a. Do conservative/evangelical Protestant organizations co-operate with each other? b. In what ways? Finding Other Organizations: 6. Other than the phone book, do you have any suggestions for where I should look for lists of religious organizations in Kitchener-Waterloo? 7. Are there other people who have knowledge of the local religious community that you recommend I consult? [ Page ] 208 Appendix B - Index of Evangelical Characteristics Pre-test To test this index I chose representatives from four Christian denominations that, according to consensus, have different degrees of “evangelicalness.” The Christian & Missionary Alliance in Canada (CMAC) and the Pentecostal Assemblies of Canada (PAOC) are two fast-growing, self-identified evangelical churches which have different styles of worship and practice. The Mennonite Conference of Eastern Canada (MCEC) rejects the label “evangelical” but does have some characteristics considered typical of evangelicals. Finally, the Evangelical Lutheran Church in Canada (ELCIC) is by all measures distinct from evangelicalism. A tentative confirmation of the utility of the Index of Evangelical Characteristics took place when the PAOC and the CMAC had a significant number of characteristics and the ELCIC had very few. The MCEC fell somewhere between the other two. Characteristics were compiled primarily from historical descriptions of evangelicalism (Gerstner 1977; Moberg 1977; Smith 1982; Marsden 1984; Noll 1984; Hunter 1987; Marsden 1987; Nash 1987; Smith 1987; Wells 1987; Ellingsen 1988; Noll et al. 1988; Dayton 1991; Johnston 1991; Marsden 1991; Marsden 1991; Weber 1991; Stackhouse 1993; Bebbington 1994; Noll et al. 1994; Wells 1994; Mackey 1995) The initial pre-test showed that the idea of an index had potential. However, a four page survey was too long to be included in a larger survey, the question needed more detail to clarify how to fill out the survey, the division of responses into two categories was too limited to capture nuances in group characteristics and the category “qualified” was too confusing to produce results without a verbal explanation. In response to these and other related concerns I removed the historical section as it was unable to provide useful distinctions between organizations. I then removed ambiguous or repetitive characteristics in other sections. I also clarified the wording of items on the survey about which the initial pre- testers were unclear. To further clarify the index I divided it into smaller categories and completely rewrote the introduction question and the headings to focus on the importance of the belief or characteristic to the church's identity. To get a more refined idea of the relevance of certain characteristics to organizations I changed the scale to a 5 point rating scale. Finally, I formatted the index to fit on one double-sided page (see Appendix G). [ Page ] 209 This revised survey was then submitted to an additional group of pre-testers as part of pre-test of a larger survey. Based on these pre-tests, beliefs or characteristics were lengthened to more exactly specify what I was asking about or were otherwise clarified, the question was again reworded, the scale was expanded to 7 points and the layout of the survey was changed to be easier to fill out. Index of Evangelical Characteristics Pre-test Results [ Please contact repository@tyndale.ca for chart details ] Outline Necessary Characteristics (-) Negatively correlated characteristics Characteristic Denomination PAOC CMAC MCEC ELCIC Historical Roots Anabaptist 2 2 Arminian 1 1 Awakenings of the 18th and 19th century 2 1 2 Calvinist 1 Charismatic 1 Confessional 1 2 Creedalism 2 Experiential 2 1 Holiness 2 2 Methodist 1 Peace churches Pentecostalism 2 2 Pietism 1 Puritan Reformation 1 2 Total Political Tendencies 15 6 2 11 Change the world (-) 1 Closer relations with evangelicals than with other groups 2 2 1 Conservatism 2 2 1 Freedom from political constraint 2 2 2 2 Individualism 2 2 1 Infiltrate the world 1 1 Interconnection with evangelicals 1 2 1 Interdenominational dialogue (-) 2 2 Interreligious dialogue (-) 1 2 Little regard for the institutional church 1 Opposition to “liberal” theology 2 2 [ Page ] 210 Characteristic Denomination PAOC CMAC MCEC ELCIC Opposition to “modernist” theology 2 2 Opposition to “relativism” 2 2 2 2 Opposition to “secular humanism” 2 2 1 Opposition to “theological and social decay” 2 2 1 Optimistic view of human nature 1 Pacifism (-) 2 Relationship with the “world” (-) 2 Resistance to modern intellectual & religious currents 2 2 1 Alignment with evangelicals against common enemies 2 Sense of being an outsider 2 1 Separate from the world 2 2 1 Social concern 2 2 2 Strong sense of community 2 2 2 2 Strongly held views 2 1 2 Transdenominational identity 2 2 1 Total 35 26 15 8 Doctrines God 4 4 4 4 Holy Spirit 4 4 4 4 Baptism of the Holy Spirit 2 Holy Spirit a gift of God 2 2 2 2 Second work of the Holy Spirit 2 Jesus Christ 4 4 4 4 Brings redemption 4 4 4 4 Death on the cross 4 4 4 4 Deity 4 4 4 4 Dispenser of judgment 2 2 2 2 Doer of miracles 2 2 2 2 Heart of Christianity 2 2 2 2 Imminent second coming 4 4 4 Incarnation 4 4 2 4 Literal bodily resurrection 2 2 2 2 Mediator between God and humans 2 2 1 2 Real historical character of all his acts 2 2 2 1 Substitutionary atonement 4 4 4 2 Virgin birth 2 2 1 2 Salvation 2 2 1 2 Born again 2 2 1 Changed life 2 2 1 2 Conversion 4 4 2 2 Essential role of the Holy Spirit 2 2 1 2 Insufficiency of human merit 2 2 1 2 [ Page ] 211 Characteristic Denomination PAOC CMAC MCEC ELCIC Justification by faith 4 4 4 4 Leads to eternal life 2 2 2 2 Moral and spiritual rebirth 2 2 1 1 Only through Jesus Christ 2 2 1 2 Personal 2 2 1 2 Regeneration 2 2 2 Rooted in history 2 2 2 2 Salvation by faith alone 2 2 1 2 Sanctification 2 2 1 2 Spiritual renewal 2 2 1 1 Trinity | 2 2 2 2 The Bible 4 4 4 4 Authoritative 4 4 4 4 Centrality of the “the gospel” 2 2 2 2 Divinely inspired 2 2 2 1 Final authority 2 2 1 2 Inerrancy 2 2 1 Infallibility 2 2 1 Normative for Christian belief and practice 2 2 2 2 Scripture justifies itself 2 2 1 Sole authority for faith and practice 2 1 2 Supernaturalist understanding of the Bible 2 1 Ultimate religious authority 2 1 2 Verbal inspiration 2 2 1 General Belief in absolute Truth 2 2 1 Denial of the authority of tradition 2 1 Depravity of Humanity 2 1 1 1 Dispensationalism 2 Non-Christians lost 2 2 1 1 Priesthood of all believers 2 2 2 2 The Church 2 2 2 2 Total 126 119 95 92 Practices Communal Baptism 2 2 2 2 Bible study 2 2 2 2 Communion 2 2 2 2 Evangelism 4 4 4 4 Lay participation 4 4 4 4 Missions 2 2 2 2 Preaching 2 2 2 2 [ Page ] 212 Characteristic Denomination PAOC CMAC MCEC ELCIC | Bible based - the "word" 2 2 1 2 ; ; Evangelistic 2 2 1 Energetic 2 2 1 2 ; Worship 2 2 2 2 Wide Hymnody 2 2 2 2 Lifestyle Importance of a disciplined life 1 2 2 2 Personal holiness 4 4 2 Prayer 2 2 1 2 Spiritual transformation 2 2 1 Social concern 2 2 2 2 Energetic 2 2 2 Reformation of social evil (-) 2 2 Total 41 40 33 32 Institutions Education encouraged 2 2 2 Bible colleges 2 2 1 2 Liberal arts colleges (-) 2 2 Seminaries 1 2 2 2 Sunday school 2 2 2 2 Universities (-) 2 2 Missions Support Billy Graham 1 1 Denominational outreach 2 2 1 2 Global 2 2 1 2 Local 2 2 1 2 Para-church 1 Information Music 2 2 2 Publishing 2 2 2 2 Social agencies 2 2 2 | Canadian 2 2 | Global 2 2 2 2 Local 2 2 General Cooperation with Evangelicals 1 2 1 Transdenominational (-) 1 2 World-wide 2 2 1 2 Total 21 25 19 23 Other Important Characteristics (added by respondents) Extensive use of choruses 2 Sacramental 2 [ Page ] 213 Characteristic Denomination PAOC CMAC MCEC ELCIC Ordained Ministry of Word and Sacrament 2 Episcopal Church Structure 2 Historic Liturgy 2 Total 2 0 0 8 Grand Total 238 216 164 166 Percentage of Possible Characteristics (Score) 86% 78% 59% 60% Distance from next highest group 8% 19% -1% Adjusted Grand Total* 223 210 162 155 Adjusted Percentage (Score)* 90% 85% 65% 63% Adjusted Distance* 5% 19% 3% * Not including Historical Roots [ Page ] 214 Appendix C - Consultation Permission Letter Dear «Contact Name», I appreciate your interest in my research. Its goal of is to get a better understanding of how evangelical organizations work together. The data will be collected through a survey of religious organizations in Kitchener-Waterloo. I am interested in your understanding of the religious community in Kitchener- Waterloo to provide a background for the survey and to help me determine which religious organizations I should survey. I would like about 30 minutes of your time to ask you a few questions about the local religious community. Your name and any identifying information about you will be kept confidential. All information that I collect will be kept in a locked filling cabinet or in a password protected computer file. If you have any questions about my research after the interview please contact me at (9 am - 5 pm, Monday or Friday) or anytime at Thank you, Mark Chapman University of Toronto Please sign your name here to indicate that you understand the purpose of my research, that you are aware of your right to refuse to answer my questions in full or in part, and that you give me permission to use the information you provide me in my research. Name Date [ Page ] 215 Appendix D - First Contact Script Answer Script Hello, my name is Mark Chapman. I am a doctoral student at the University of Toronto conducting a study of churches in the Kitchener-Waterloo area. I am conducting a brief survey and would like to verify that I have accurate names and address information for << organization name>> If its organizational identity is uncertain: a. Can you tell me what kind of organization <> is? Is it a church? If it is a church: a. The name I have for your head pastor is <>. Is this right? or b. Can you tell me the name of the head pastor at <>? and a. The address I have for <> is <
>. Is this right? or b. Can you tell me where <> is located? If they ask for more details: The goal of this survey is to get a better understanding of how Protestant organizations work together. All local Protestant churches will receive a survey and specific results will be kept confidential. The usefulness of the information collected by this study increases as more local organizations participate. Potential benefits of this study for the local religious community include: (1) an improved understanding of how religious organizations in Kitchener- Waterloo do or do not work together (2) aid in identifying similar organizations which will help religious organizations find the resources they are looking for [ Page ] 216 (3) help to identify duplication of activities within the community so that duplication can be eliminated or organizations with common goals can work together. (4) identification of areas where there are currently no religious organizations working. A summary of the results of the survey will be supplied, on request, to any organization which completes the survey. The questionnaire is simple to fill out and will take between 30 and 60 minutes of your time. It should be filled out by the head minister of <> to assure that the information collected is comparable. [ Page ] 217 Appendix E - Survey Pre-Test Cover Letter , Thanks for helping me with the survey. I have included two copies of the survey. They are identical in content but have slightly different formats. I have taped them together so that you can see the two alternatives. The final survey will be printed double sided. I would appreciate your comments on which format you think is more conducive to encouraging response to the survey. I am trying to get some idea how the survey might be perceived by a pastor. I also welcome any comments other colleagues might have on the survey. Thanks again, Mark [ Page ] 218 Appendix F - Survey Cover Letter Letterhead «Current Date» Dear «Primary Name», I am a doctoral student (Ph.D.) at the University of Toronto living in Kitchener-Waterloo. My doctoral thesis is on relationships among Protestant churches. It will be based on data gathered from pastors using the enclosed questionnaire. I know that Pastors are very busy people and so I have tried to design the survey to be easy to fill out. I would be most grateful if you would take the time to fill it out. For your convenience a self-addressed stamped envelope has been provided for returning the survey. If it is easier for you please feel free to complete the final two pages of the survey on a computer or typewriter. I am asking that the head ministers of local Protestant churches complete the survey. This is to assure that the information collected is comparable. Each completed survey will increase the usefulness of the study’s findings. A summary of the aggregate results of the survey and an interpretive discussion of these results will be supplied to any church which completes the survey. No specific churches or individuals will be identified and the contents of individual surveys will be kept STRICTLY confidential. If you have any questions about the survey please contact me at (Tuesday to Thursday) or anytime at / . For further information you can also contact Dr. John Simpson, my doctoral supervisor, at or at It would be very helpful to me if you could return the survey as soon as possible within the next two weeks. Thank you for your time. Sincerely Yours, Mark Chapman University of Toronto P.S. If you have time you may find it convenient to fill out the survey right now. [ Page ] 219 Appendix G - Final Survey Religious Connections Survey [ Please contact repository@tyndale.ca for diagram details ] NOTE: Formatting has been adjusted to fit in this document [ Page ] 220 Section One Please complete section one with the applicable information regarding your church and yourself: Church Name: Denominational Affiliation: Church Address: Street: City: Postal Code: Phone Number: Fax Number: Email: Website Filled Out By: Title: Date Filled Out: Time with Church: Year(s) Month(s) [ Page ] 221 Section Two Thinking back over 1999, please list all the religious organizations (e.g., charities, churches, denominations, ministerials, social agencies, etc.) your church has interacted with, in the listed ways, since January 1999. Mark an X in the subsequent columns to indicate the type of interaction which took place. Please cover local and non-local organizations including organizations associated with your own denomination if relevant . [ Please contact repository@tyndale.ca for chart details ] Organization and City Type of Interaction with other Organizations (As many as apply) Provide them with physical support (e.g., share facilities) Share information (e.g., upcoming events) Participate with them in the same events Provide them with financial support Share volunteers Share paid staff 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. [ Page ] 222 Section Three Within Christian circles there is wide variation in the beliefs and characteristics that form part of a church’s identity. Please circle a number from 1 to 7, where 1 indicates a belief or characteristic that is Very Important to your particular local church’s identity and 7 indicates a belief or characteristic that is Not at all Important to your particular local church’s identity. An important belief or characteristic is one that has noticeably affected the operation of your church. The following list beliefs and characteristics has been organized by subject and then placed in alphabetical order. [ Please contact repository@tyndale.ca for chart details ] Belief or Characteristic Important to your Church’s Identity Very Not at all Important Important Believing the following doctrines: Bible Authoritative 1 2 3 4 5 6 7 Divinely inspired 1 2 3 4 5 6 7 Final authority for living 1 2 3 4 5 6 7 Inerrant 1 2 3 4 5 6 7 Literal word of God 1 2 3 4 5 6 7 Normative for Christian belief and practice 1 2 3 4 5 6 7 Sole authority for faith and practice 1 2 3 4 5 6 7 Verbally inspired 1 2 3 4 5 6 7 Holy Spirit 1 2 3 4 5 6 7 Essential role of Holy Spirit in salvation 1 2 3 4 5 6 7 Second work of the Holy Spirit 1 2 3 4 5 6 7 General Belief in absolute truth 1 2 3 4 5 6 7 Dispensationalism 1 2 3 4 5 6 7 Heaven 1 2 3 4 5 6 7 Hell as a place of physical torment 1 2 3 4 5 6 7 Life after death 1 2 3 4 5 6 7 Non-Christians lost 1 2 3 4 5 6 7 Prayer has positive effects 1 2 3 4 5 6 7 Premillenialism 1 2 3 4 5 6 7 Satan is an actual spiritual being 1 2 3 4 5 6 7 Total depravity of humanity 1 2 3 4 5 6 7 Jesus Christ Brings redemption 1 2 3 4 5 6 7 Is Deity 1 2 3 4 5 6 7 Dispenses judgment 1 2 3 4 5 6 7 Imminent Second Coming 1 2 3 4 5 6 7 Literal bodily resurrection from the dead 1 2 3 4 5 6 7 Mediator between God and humans 1 2 3 4 5 6 7 Performs miracles 1 2 3 4 5 6 7 Physical death on the cross 1 2 3 4 5 6 7 Real historical character of all his acts 1 2 3 4 5 6 7 Substitutionary atonement 1 2 3 4 5 6 7 Virgin birth 1 2 3 4 5 6 7 [ Page ] 223 Belief or Characteristic Important to your Church’s Identity Very Important Not at all Important Salvation 1 2 3 4 5 6 7 Insufficiency of human merit 1 2 3 4 5 6 7 Justification by faith alone 1 2 3 4 5 6 7 Leads to a changed life 1 2 3 4 5 6 7 Leads to eternal life 1 2 3 4 5 6 7 Moral and spiritual rebirth 1 2 3 4 5 6 7 Necessary for all people 1 2 3 4 5 6 7 Only through Jesus Christ 1 2 3 4 5 6 7 Personal conversion 1 2 3 4 5 6 7 Supporting the following institutions: Bible colleges 1 2 3 4 5 6 7 Billy Graham ministries 1 2 3 4 5 6 7 Global missions 1 2 3 4 5 6 7 Global social agencies 1 2 3 4 5 6 7 Local missions 1 2 3 4 5 6 7 Local social agencies 1 2 3 4 5 6 7 Theological seminaries 1 2 3 4 5 6 7 Universities 1 2 3 4 5 6 7 Associating with the following groups: Roman Catholics 1 2 3 4 5 6 7 Charismatic movement 1 2 3 4 5 6 7 Evangelicals 1 2 3 4 5 6 7 Mainline churches 1 2 3 4 5 6 7 Non-Christian religions 1 2 3 4 5 6 7 Supporting: An optimistic view of human nature 1 2 3 4 5 6 7 Combating moral evil 1 2 3 4 5 6 7 Combating social evil 1 2 3 4 5 6 7 Conservative theology 1 2 3 4 5 6 7 Freedom from political/legal control 1 2 3 4 5 6 7 Individualism 1 2 3 4 5 6 7 Inter-religious dialogue 1 2 3 4 5 6 7 Liberal theology 1 2 3 4 5 6 7 Pacifism 1 2 3 4 5 6 7 Relativism 1 2 3 4 5 6 7 Secular humanism 1 2 3 4 5 6 7 Separation from the world 1 2 3 4 5 6 7 Valuing the following practices: Adult baptism 1 2 3 4 5 6 7 Counseling 1 2 3 4 5 6 7 Daily Bible reading 1 2 3 4 5 6 7 Daily prayer 1 2 3 4 5 6 7 Disciplined spiritual life 1 2 3 4 5 6 7 Evangelism of non-Christians 1 2 3 4 5 6 7 Personal holiness 1 2 3 4 5 6 7 Preaching 1 2 3 4 5 6 7 Regular Bible study 1 2 3 4 5 6 7 Weekly church attendance 1 2 3 4 5 6 7 [ Page ] 224 Section Four To assist in comparing organizations could you please provide the following information about your church: Number of Members? Average Weekly Attendance? Number of Paid Staff? Full-time Part-time Number of Unpaid Staff? Full-time Part-time Section Five Please list the religious organizations (e.g., charities, churches, denominations, ministerials, social agencies, etc.) which are the same or very similar to your church in terms of their identifying characteristics and ministry objectives. Please list these organizations in order, from most similar to least similar to your church. Listed organizations do not have to be organizations with which you have any contact. Organization and City 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. [ Page ] 225 Comments: Please use this section to provide additional information you think is relevant to this study, to address questions you have about the survey, or to expand upon your responses to the survey. Thank You! [ Page ] 226 Appendix H - Follow-up Week 2 - Reminder Letter Letterhead «Current Date» Dear «Primary Name», Two weeks ago, as part of my doctoral research at the University of Toronto, I mailed you a survey which asked questions about relationships among churches. Although I have not yet received a completed questionnaire from «Prefix»«Organization Name», your completed questionnaire and this letter may have crossed in the mail. If this is the case, I apologize for taking more of your time and thank you for making space in your busy schedule to assist me in my research. I know that pastors are very busy people. However, I would be very grateful if you could spare some time for the survey. The more surveys I receive, the more useful my research becomes and the more potential benefit my research has for the local religious community. A summary of the aggregate results of the survey and an interpretive discussion of these results will be supplied to any church which completes the survey. No specific churches or individuals will be identified and the contents of individual surveys will be kept STRICTLY confidential. If you have any questions or concerns about the survey please contact me at (Wed. to Thurs.) or anytime at For further information you can also contact Dr. John Simpson, my doctoral supervisor, at or at Thank you for your time. Sincerely Yours, Mark Chapman University of Toronto [ Page ] 227 Week 4 - Second Survey Cover Letter Letterhead «Current Date» Dear «Primary Name», One month ago, as part of my doctoral research at the University of Toronto, I mailed you a survey which asked questions about relationships among churches. Although I have not yet received a completed questionnaire from «Prefix»«Organization Name», your completed questionnaire and this letter may have crossed in the mail. If this is the case, I apologize for taking more of your time and thank you for making space in your busy schedule to assist me in my research. I know that pastors are very busy people. However, I would be very grateful if you could spare some time for the survey. The more surveys I receive, the more useful my research becomes and the more potential benefit my research has for the local religious community. For your convenience, I have enclosed a second questionnaire in case the original questionnaire was lost in the mail or misplaced. A summary of the aggregate results of the survey and an interpretive discussion of these results will be supplied to any church which completes the survey. No specific churches or individuals will be identified and the contents of individual surveys will be kept STRICTLY confidential. If you have any questions or concerns about the survey please contact me at, (Wed. to Thurs.) or anytime at . For further information you can also contact Dr. John Simpson, my doctoral supervisor, at or at. Thank you for your time. Sincerely Yours, Mark Chapman University of Toronto P.S. If you have time you may find it convenient to fill out the survey right now. [ Page ] 228 Week 6 - Phone Call Script Hello, can I please speak with <>? Hi, my name is Mark Chapman. I am a doctoral student at the University of Toronto. I sent you the survey on relations between _____________________ and other local Protestant churches. I have not received a survey from you and I was wondering if you had any questions about it that I could answer. Thank-you very much I look forward to receiving your complete survey or Would it be possible to let me know your reasons for choosing not to complete the survey? This will help me determine if the survey is flawed or biased towards certain types of churches. Final Contact Phone Call Script Hello, can I please speak with __________________? Hi my name is Mark Chapman I sent you the survey on relations between _________________ and other organizations in Kitchener-Waterloo. I have not yet received a survey from you and I was wondering if you had any questions about it that I could answer. Thank-you very much I look forward to receiving your complete survey or It would be useful to me if you could let me know your reasons for choosing not to complete the survey. This will help me determine if the survey is flawed or biased towards certain types of organizations. [ Page ] 229 Appendix I - Thank-you Letter Letterhead «Current Date» Dear «Primary Name», I am a doctoral student (Ph.D.) at the University of Toronto living in Kitchener-Waterloo. My doctoral thesis is on relationships among Protestant churches. It is based on a mail questionnaire sent to pastors in Kitchener-Waterloo. Almost a year ago I asked you to fill out a survey to assist me in my research. I appreciated the time you took to fill out the survey. Please find enclosed a summary and interpretive discussion of the aggregate results of the survey. If you have any questions about the findings of the survey please contact me at ( Thank you again for your time. Sincerely Yours, Mark Chapman University of Toronto [ Page ] 230 Summary of Results (numbers are rounded) Surveys Returned: 88 of 133 Total number of connections: 647 Total number of unique connections: 342 Total number of connections in Kitchener-Waterloo: 400 Number of direct connections between churches: 50 of a possible 7656 Count of type of organization connected to: Social Services: 232 Church Related: 167 Missions Related: 143 Education: 40 Health Care: 14 Other/Unknown: 51 Percentage of time a particular type of interaction was checked off: Finances 59% Information 51% Events 40% Volunteers 35% Physical Support 32% Staff 11% Average Church Membership: 326 Average Church Attendance: 203 [ Page ] 231 Appendix J - Full Factor Analysis Results [ Please contact repository@tyndale.ca or Appendix J table details ] Rotated Factor Analysis (All Characteristics) Variables Orthodoxy Salvation Personal Education Associate Action Deity/Death End Times Church Freedom Secular Second Work Missions Rebirth Life after Death Positive 17 Prayer Conservative theology 0.85 0.10 0.13 -- -- 0.12 -- -- -- 0.15 -- -- 0.23 -- -- -- -- Non-Christians lost 0.84 0.17 0.25 -0.10 0.11 0.11 0.10 -- -- -- -- 0.11 0.14 0.11 -- -- -- Satan is an actual spiritual being 0.83 0.19 0.16 -- 0.11 -- -- -- -- -- -- -- -- 0.21 0.11 -- 0.13 Total depravity of humanity 0.83 0.13 -- -- 0.11 0.10 -- -- 0.19 -- -- -- 0.14 -- 0.21 -- Liberal theology -0.82 -0.23 -- 0.15 -- -- -- -- -- -0.15 -0.31 -- -- -- -- -- -- Evangelism of Non- Christians 0.78 0.29 0.25 0.16 0.11 -- -- -- 0.14 0.11 -- 0.10 -- -- -- -- Salvation is only through Jesus Christ 0.78 0.28 -- -- -- -- -- -- -- 0.10 0.13 0.16 -- -- -- -- -- Bible is final authority for living 0.77 0.25 0.12 -0.13 -- 0.15 -- -- -- -- -- 0.12 0.17 -- -- 0.18 Inerrant Bible 0.77 -- 0.20 -0.23 -- 0.13 0.12 -- -- 0.10 -- 0.18 0.22 -- 0.12 -- -- Belief in absolute truth 0.77 0.22 0.27 -0.13 0.14 -- 0.13 -- -- -- -- -- 0.13 -- -- 0.16 -- Hell as a place of physical torment 0.74 0.17 0.10 -- -- 0.13 -- 0.25 -- -- -- 0.11 0.14 -- 0.17 -- -- Bible is literal word of God 0.74 -- 0.31 -0.19 -- -- 0.17 0.17 -- -- -- 0.19 -- -0.12 0.12 -- 0.11 Literal bodily resurrection of Jesus Christ from the dead 0.74 0.15 -0.18 -- 0.20 0.20 0.19 -- -- -- 0.16 -0.17 -- -- -- -- -- Virgin birth 0.73 0.10 0.17 -- 0.15 0.19 0.32 -- -- -- 0.26 -- -- 0.14 -- -- -0.14 Bible is the sole authority for faith and practice 0.73 0.25 -- -- -- 0.15 -- 0.10 -- 0.36 -- -- 0.20 -0.12 -0.10 0.12 -0.13 An optimistic view of human nature -0.71 -0.11 -- -- 0.14 -- -- -- 0.21 -- -0.26 0.10 -- -- -0.12 -- -- Associate with Non- Christian Religions -0.67 -0.12 0.17 0.27 0.27 -- 0.16 -0.12 -- -0.22 -0.14 -- 0.17 -- -0.21 -- -- Heaven 0.67 0.17 -- -- -- -- 0.31 0.14 -- 0.23 0.14 0.14 -- -- 0.35 0.21 -0.11 Associate with Mainline Churches -0.66 -- -0.15 0.24 0.47 -- -- -- -0.10 -0.16 -- 0.10 -- -- -- 0.18 -0.13 Substitutionary atonement 0.66 0.22 -- -- -- 0.13 0.35 -- -- -- 0.24 0.23 -- -0.13 -- -0.20 Jesus Christ dispenses judgment 0.65 0.12 0.33 0.11 -- 0.16 0.44 -- -0.15 -- -- -- -- 0.13 -- 0.15 0.14 [ Page ] 232 Component Variables Orthodoxy Salvation Personal Education Associate Action Deity/Death End Times Church Freedom Secular Second Work Missions Rebirth Life after Death Positive Prayer 17 Inter-religious Dialogue -0.65 -0.15 -- -- 0.26 -- 0.17 -0.22 0.14 -- -0.12 0.11 -- -0.11 -- -- -- Imminent second coming of Jesus Christ 0.64 0.33 0.25 -- -- -- 0.28 0.18 -- 0.14 -- 0.28 -0.15 -- -- -- 0.12 Associate with Roman Catholics -0.63 -- -0.12 0.21 0.53 -- 0.18 -0.13 -- -- -- 0.12 -- -- -0.15 -- -- Salvation is necessary for all people 0.63 0.53 0.12 -- -- 0.19 0.14 -- -- 0.11 0.12 0.13 0.12 0.18 -- -- 0.13 Salvation leads to eternal life 0.61 0.38 -- -- -- -- 0.18 0.16 -- -- 0.15 0.32 -- 0.15 -- -0.23 0.26 Insufficiency of human merit 0.61 0.49 -- -- -- -- 0.24 -- -- 0.21 0.13 0.13 -- -- 0.15 -0.15 Bible is verbally inspired 0.59 -0.21 0.12 -0.20 -- 0.21 0.15 0.16 0.19 -- -- 0.15 0.18 -0.25 0.14 0.19 -0.14 Real historical character of all of Jesus Christ's acts 0.59 0.17 0.24 -- -- 0.14 0.48 -- -0.14 0.19 -- -- -0.18 -- -- -- 0.21 Justification by faith alone 0.59 0.44 -0.19 -- 0.12 -- 0.14 -- -- -- -- -0.34 -- 0.17 -- 0.23 0.11 Authoritative Bible 0.56 0.16 -0.13 -- 0.24 0.22 -- -- -- 0.20 -- -0.18 -- -- -- -- 0.54 Personal holiness 0.55 0.23 0.37 -- 0.13 -- 0.12 0.25 0.16 0.22 -- 0.14 -- -- 0.20 -0.13 0.21 Bible Colleges 0.50 -- 0.25 0.14 0.17 -- -0.13 0.21 -- 0.46 -0.11 -- -- 0.33 -- 0.14 -- Divinely inspired Bible 0.49 0.45 0.18 -- -- 0.12 -- -- -0.12 0.41 0.16 0.35 -- -- 0.12 -- 0.13 Jesus Christ performs miracles 0.48 0.38 0.17 -- -- -- 0.30 -- -- -- -- 0.28 0.13 0.15 0.13 -0.14 0.23 Bible is Normative for Christian belief and practice 0.47 -- -- -0.10 -- -- 0.18 -- -- 0.46 0.18 0.18 0.41 -- -- 0.15 0.14 Essential role of the Holy Spirit in salvation 0.46 0.42 0.33 0.10 -0.17 -0.18 0.12 0.27 -- 0.20 -- -0.15 -0.16 -- 0.13 -- Separation from the world 0.42 -- -- -0.30 0.15 0.17 -- 0.39 0.12 0.41 -- 0.11 -- 0.15 -- -- -- Jesus Christ brings redemption 0.35 0.71 -- -- -- 0.16 -- -- -- -- 0.15 -- 0.16 0.20 -- -- -- Salvation leads to a changed life 0.40 0.70 0.16 -- -- 0.23 -0.15 -- -- 0.11 -- -- -- 0.19 -- -- 0.20 Jesus Christ as mediator between God and humans 0.46 0.68 -- -- -- -- 0.28 -- 0.12 0.11 -- -- -- -- 0.11 0.11 -0.10 Salvation 0.45 0.65 -- -- 0.12 0.12 0.30 -- -- -- -- -- -- -- -- -- Disciplined spiritual life 0.25 -- 0.85 — 0.13 -- 0.10 -- -- -- -- -- -- 0.11 -- -- Adult baptism -- -- 0.65 -0.12 0.19 -- 0.18 0.24 -- 0.22 -0.18 -0.11 -0.19 0.25 -- -0.15 -- Personal conversion 0.55 0.25 0.62 -0.13 -- -- -- -- -- -- -- 0.11 -- 0.24 -- -- -- Daily prayer 0.16 0.37 0.53 — -- 0.20 -- 0.20 0.28 0.12 -- -- 0.11 -0.29 -- 0.28 -0.11 Daily Bible reading 0.27 0.39 0.51 0.13 -- 0.30 -- 0.11 0.25 -- 0.38 0.21 -0.20 -- -- -- [ Page ] 233 Component Variables Orthodoxy Salvation Personal Education Associate Action Deity/Death End Times Church Freedom Secular Second Work Missions Rebirth Life after Death Positive Prayer 17 Theological Seminaries -- -- -0.16 0.82 -- 0.20 -- -0.12 0.18 -- -- -- -- -- -- -- -- Universities -0.20 -- 0.14 0.73 -- 0.17 -- -- -- -- -0.11 0.16 -- -0.13 -0.17 -- 0.22 Global Social Agencies -0.29 -- -- 0.72 -- -0.20 -- -0.21 0.22 0.15 -- -0.16 0.11 0.18 -- -- Local Social Agencies -0.34 -- -- 0.63 -- -- -- -- -- -- -- -0.21 0.19 -0.11 0.32 -- -- Associate with Charismatic Movement Associate with Evangelicals -- 0.34 -- 0.13 0.13 0.14 — — 0.86 0.69 — 0.13 -- -- 0.10 0.24 -- 0.14 -- 0.14 -- -- 0.12 -- -- -- -- 0.15 -- -- -- -- 0.15 -0.11 Billy Graham ministries 0.37 -- 0.20 0.19 0.37 0.16 -- 0.19 0.15 0.35 -- -- -- 0.24 -- -- -0.36 Preaching 0.25 0.25 -- -- -- 0.73 0.17 -- -- -- 0.19 -- -- 0.18 0.15 -- -- Weekly church attendance 0.37 0.20 0.31 0.31 0.13 0.67 -- 0.13 -- -- -- -- -- -- -- -- -- Regular Bible study 0.45 0.22 0.38 -- -- 0.48 0.14 0.14 0.18 0.17 -- -- -- -- -- 0.13 -- Jesus Christ is Deity 0.52 0.33 -0.19 -- -- — 0.64 -- -- -- 0.14 0.11 -- -- 0.14 -- -- Jesus Christ's physical death on the cross 0.21 0.36 0.13 -- -- 0.24 0.50 -- 0.13 -- 0.29 -0.14 0.18 0.19 0.19 0.20 -- Dispensationalism 0.13 -- 0.12 -0.12 -- -- -- 0.86 0.11 -- -- -- -- -0.14 0.10 -- -- Premillenialism 0.34 -- 0.18 -- 0.11 -- -- 0.79 -- -- -- -- 0.20 -- -- -- -- Combating Moral Evil 0.24 0.14 -- -- 0.12 -- 0.16 -- 0.83 -- -- -- -- -- -- -- 0.11 Combating Social Evil -0.11 -- -- 0.24 -- -- -0.16 -- 0.80 -- -- -- -- -- -- 0.18 Counseling -0.11 -- 0.12 0.38 -- 0.36 -- -- 0.45 -- -0.26 -- -- 0.16 -- -0.16 Freedom from political/legal control 0.27 0.20 0.11 -- -- -- -- -- -- 0.71 -0.15 0.15 -- -- -- -- Secular humanism -0.36 -0.15 -- -- -- -- -- -- -- -- -0.78 -- -- -- -- -- -- Relativism -0.41 -0.12 0.12 0.14 -- -- -0.15 -- -- -- -0.71 0.13 -- -- -- -- -- Pacifism -0.40 -- 0.12 0.13 -- -0.22 -0.17 -- -0.16 0.31 -0.45 -- -0.14 -- -- 0.13 0.23 Second work of the Holy Spirit Holy Spirit 0.12 0.27 0.12 0.10 -- 0.13 -- 0.24 0.25 -- -- -0.20 -- 0.14 0.14 -- -- -- -0.10 -- -- 0.31 0.72 0.49 -- 0.17 -- 0.31 -- 0.38 0.12 -- -- -- Global Missions 0.27 -- -0.13 0.15 -- -- -- 0.27 0.14 -- -- -- 0.76 -- -- -- -- Local Missions 0.13 -- 0.26 0.33 -- -- -- 0.12 -0.13 0.26 -- -- 0.55 -- -- 0.38 -- Individualism 0.16 0.19 0.27 -0.34 0.25 0.19 0.14 -0.12 -- -- -0.19 -0.17 0.42 -0.12 0.20 -- -0.20 Moral and spiritual rebirth 0.20 0.30 0.14 -- 0.12 0.15 -- -- -- -- -- 0.10 -- 0.74 — 0.15 -- Life after deal 0.25 0.10 -- -- -- 0.15 -- -- -- -- -- -- -- -- 0.83 0.17 -- Prayer has positive effects -- 0.12 0.10 -- -- -- -- -- 0.31 -- -- 0.12 -- 0.18 0.25 0.75 | -- [ Page ] 234 Component Variables Orthodoxy Salvation Personal Education Associate Action Deity/Death End Church Times Freedom Secular Second Work Missions Rebirth Life after Death Positive Prayer 17 % Variance Explained 27.15 6.93 5.38 4.63 3.48 3.33 3.31 3.26 3.25 3.19 3.17 2.85 2.73 2.45 2.34 2.00 1.80 Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. Rotation converged in 32 iterations. Mean substitution for missing values Absolute values <.1 displayed as -- ***** This is the end of the e-text. This e-text was brought to you by Tyndale University, J. William Horsey Library - Tyndale Digital Collections *****