Collaborative Research Networks @ Cwn

NETWORK MAPPING STUDY Final Report

Prepared for the Canadian Water Network

Dimitrina Dimitrova, University of Toronto Emmanuel Koku, Drexel University Barry Wellman, University of Toronto Howard White, Drexel University

Date: June 2007

Acknowledgements We are indebted to numerous people for their support and assistance in conducting the study and preparing this report. Lee Weisser has been an invaluable member of the team. She has seen the project through from start to finish, assisting it in numerous ways. She worked as project administrator, interviewer, and editor, and in all of these capacities she excelled. Jeremy Birnholtz contributed incisive comments and ideas to the research design and the preliminary report of the study. He also conducted a number of the interviews, bringing his competence and experience to the process. June Pollard transcribed even the most difficult interviews with accuracy and speed. Kristen Mandziuk and Dolores Figueroa coded them expertly in NVivo. Kristen, in addition, spent many hours helping in the orderly wrap up of the project, sorting, verifying and cleaning records. A group of smart students assisted with numerous tasks: they transcribed interviews, entered data, searched the Internet, and hunted down articles and books. Glasha Romanovska, Lindsay Cai, Jackie D’Sa, Natalie Zinko, and Nazila Rostami were all a pleasure to work with. Few research projects receive as much support and assistance from their funding organization as this one has. We have benefited tremendously from the visionary ideas and sage advice of Don Brookes, the professionalism and engaging personality of Monica Escamilla, and the technical skills and hard work of Corban Riley. Other CWN staff including David Cotter, Bernadette Conant, and Karen Van Sickle, have also lent a hand when needed. Finally, this project was only possible because busy people working in the area of water generously shared their time and their insights with us. Working with them has been a privilege and a delight. Our sincere thanks to all!

TABLE OF CONTENTS

Acknowledgements Executive summary...................................................................................................... i Introduction..................................................................................................................... 1 Part I. Respondents ......................................................................................................... 3 1.1. Demographics: Who are the respondents?........................................................... 3 1.2. Personal networks: To whom are the respondents connected?............................ 4 1.3. Ties: Another look at the water community ........................................................ 6 Part II: Connections in the Water Network..................................................................... 9 2.1. Centrality Analysis............................................................................................... 9 2.2. Clique Analysis.................................................................................................. 13 2.3. Citation analysis................................................................................................. 21 Part III. The context of collaborative work................................................................... 27 3.1. Barriers and incentives for collaborative work.................................................. 27 3.2. Challenges on a project and strategies for overcoming them ............................ 29 3.3. Strategies for coping: team selection and independent work............................. 32 3.4. Impact of CWN on the work of academics and practitioners............................ 36 Part IV. Conclusions ..................................................................................................... 41 Appendix 1: Tables ............................................................................................................... Appendix 2: Figures.............................................................................................................. Appendix 3: Survey Data Collection and Available Data .................................................... Appendix 4: Document and Interview Data .........................................................................

APPENDICES

Appendix 1:

Tables

Appendix 2:

Figures

Appendix 3:

Survey Data collection and Available Data

Appendix 4:

Document and Interview Data

Executive summary The Network Mapping project is a social network study of the academics and practitioners working in the area of water commissioned by the Canadian Water Network of Centres of Excellence (CWN). The objectives of the study were to map the relations among the stakeholders in the area of water, describe the collaboration and knowledge exchanges among them, and examine the context in which they worked. The study included four components: a web based network survey (N=173), semistructured interviews (N=65), citation analysis of a small subgroup of academics central in the CWN (N=31), and review of documents. Several key findings emerged in the analysis of these complementary bodies of data. Socio-demographic characteristics and personal networks The survey respondents have two salient characteristics: diversity and maturity. •

Water issues are very broad, not easily captured within a discipline, and jurisdiction over water is fragmented among numerous government agencies. Hence, participants in the water network come from a range of sectors and disciplines and have different involvement in water issues. Engineers and natural scientists such as biology and earth/environmental sciences are most numerous while health, social and policy scientists are fewer. Such disciplinary and sectoral diversity provides the prerequisites for the cross-sectoral and multi-disciplinary research needed in the area.

•

At the same time, developing cross-sectoral and multidisciplinary ties strong enough to sustain collaboration is difficult. This precludes dense connections in the water network and makes water issues the playing field of experienced academics and practitioners, who have developed diverse networks. The majority of the people working in the area are mature professionals with well established networks, many of them in senior positions. In addition, despite some changes, universities tend to reward traditional work within a single discipline. This discourages junior academics building careers from doing complex collaborative research and further reinforces the maturity and seniority of the participants in the area.

Water community (whole network) Briefly put, the water network is sparsely connected yet well structured and capable of supporting multidisciplinary and cross-sectoral collaboration. •

The network has a small core of well connected central participants and a large periphery of sparsely connected participants less involved in water issues. About three quarters of these central participants are CWN members. This suggests that the agency either attracts central participants to the water network or helps its members to develop their networks and become central. Among these central participants, those

i

who actively network and reach out to others are junior academics, a few senior academics, and practitioners from various sectors. By comparison, other central participants play the role of experts who attract others. These are mostly senior academics all of whom are involved in the work of CWN. •

About two dozen participants actively work and exchange ideas with others, typically in small cliques of two to three colleagues. Some of them collaborate with colleagues from several cliques and act as bridges that connect the cliques and preclude the fragmentation of the network. Notably, people in bridge positions are academics in mid career who have already developed their networks to some extent but are still actively networking. Three quarters of the active collaborators are CWN members, confirming the key role of the agency in fostering collaboration in the area of water.

•

The composition of the cliques suggests that collaboration in the network is multidisciplinary and often cross-sectoral. Most of the work cliques include members from biology and earth/environmental sciences while the health, policy and social sciences are less represented.

•

The small size of the cliques arises from the independent work practices on the research projects. Since collaboration across disciplines, sectors, and organizations is difficult, one of the strategies to avoid problems and decrease efforts for coordination is for researchers to work independently or in small groups. Only in a very few cases do project participants work as integrated teams.

•

A second strategy to facilitate coordination and communication is for project leads to put together teams of people they know. While project teams include some newcomers recommended by other team members, researchers tend to work with a few long-term collaborators. Such teams of long-term collaborators increase commitment and decrease the efforts for developing common practices and trust — trust, commitment, and common practices have already been developed. Pre-existing ties thus facilitated the formation of teams and the work on a project.

Citation practices • Despite many multidisciplinary connections, the citation practices more closely follow disciplinary boundaries. Scientists in the water network are perceived as working in the same area and cited together. Often they are cited together with colleagues from different disciplines. This suggests that their work has multidisciplinary relevance. However, scientists in the water network do not readily see the relevance of their colleagues’ publications for their own work, rarely cite each other directly, and such direct citations more closely follow disciplinary boundaries. Such citation practices are consistent with the publication criteria of most scholarly journals which encourage working within a single discipline.

ii

CWN impact on the work of academics and practitioners • The main impact of CWN on the work of academics is networking with the right people. Academics interested in multidisciplinary and cross-sectoral research might be hard to find in a more traditional university environment and CWN helps such academics connect to each other and find partners. •

In turn, practitioners emphasized — in addition to networking — the role of the CWN as a focus of expertise in the area and as a link to academics. Even outsiders without formal partnerships with CWN turn to it when they need information.

In short, the results of the study show that CWN is successfully achieving its mission: to provide expert knowledge on critical water issues in Canada, to build scientific and human resources to address them, and to create a network of stakeholders in the area of water that serves as a catalyst for research and technology development. CWN plays a key role in holding the water network together and fostering cross-sectoral multidisciplinary collaboration: the majority of central participants and active collaborators in the area of water are CWN members. In turn, the presence of outsiders, who reach out to CWN members or already collaborate with them, suggests possibilities for creating new partnerships. Although the overall water network is sparsely knit and the core of well connected active participants is small, the ties are structured and the network is capable of supporting multidisciplinary cross-sectoral collaboration. The water network can be further improved by expanding the core, maintaining healthy balance between junior and senior academics, increasing the number of bridges, and improving the representation of health, social and policy sciences. Nonetheless, in the diverse and inherently fragmented area of water, CWN has created a viable network, established its reputation, and became a “brand name” in the area.

iii

Introduction The Canadian Water Network (CWN) was created with the mandate to support multidisciplinary research, cross-sectoral partnerships that link academics, government and industry staff, and cross-country collaboration in the area of water. CWN supports knowledge transfer and innovation by connecting researchers and practitioners across the country. Crucial to its work is a comprehensive understanding of the existing relationships among people working in the area of water that can facilitate the management and planning of CWN activities in support of the water community. In the fall of 2005, CWN hired a team of researchers to conduct a social network study of the scholars, industry partners, public sector users and regulators whose work is directly related to water. The goal of the study was to map the existing relationships among them, describe the processes of collaboration and the exchange of advice and innovative ideas, and delineate key individuals and research clusters within the network. The study addresses the following questions: • Who are the academics and practitioners in the area of water and what are their socio-demographic characteristics? • With whom are they connected? • With whom do they exchange ideas? • What is the internal structure of the network arising out of the ties among the academics and practitioners in the area of water? • Who is connected to whom? • What activities do they do together in their networks? • What is the context in which the academics and practitioners in the area of water collaborate? In other words, what are the barriers to and the challenges in collaborative research? • What are the strategies used to overcome these challenges? • What is the impact of CWN on the work of the academics and practitioners in the area of water? • How do they see the role of CWN in the area? Data collection, described below and in further detail in the Appendices, consisted of a web based national survey, semi-structured interviews, citation analysis, and a review of documents. In brief, the data collected in the study include: • 173 surveys. Among the participants are 94 academics and partners involved in CWN funded projects. They are referred to as “CWN members”. The remaining 79 respondents are academics and practitioners who are part of the water community but who have never been involved in CWN funded research. They are referred to as “outsiders.” • 65 interviews, including 56 interviews with CWN members and nine interviews with outsiders. Among all respondents, 39 were both interviewed and completed the survey. • Citation analysis results for a small group of 31 CWN members.

1

• Several dozen organizational and personal documents, including 39 research proposals, internal analyses and presentations, and several dozen resumes. Although the survey is not a representative sample of all practitioners and researchers working in the area of water, a profile of the respondents provides some understanding of the characteristics of this part of the water community. Furthermore, both the survey and the interviews include academics and practitioners working on CWN funded projects as well as individuals who are not currently involved in the work of CWN. This composition of the survey respondents reflects the overall community of people working in the area of water and enables a comparison of CWN members with other researchers and practitioners working in the area of water. The interview data, where CWN members are proportionately much more numerous, also provide opportunities for a comparison. Part I creates a profile of the participants in the area of water based on individual level data from the survey. Part II maps the internal structure of the whole network using the aggregated survey data. Part III draws on the interviews to describe how participants collaborate, thus placing the network in context and suggesting explanations for its characteristics. Throughout the report, the interpretation of the results is informed by data from several available sources.

2

Part I. Respondents The analysis below draws on both survey and interview data to describe the respondents, their networks in the area of water, and the way they use their networks for knowledge and information exchange. The discussion introduces patterns visible in the overall survey sample and then follows with comparisons between CWN members and outsiders.

1.1. Demographics: Who are the respondents? The typical respondent is a mature professional in mid-career who holds a Ph.D., works at a university, and is male. The demographic data presented show that over two thirds (69.9%) of all our respondents are men (Table 2, Section A) 1 . The mean age is 47.7 years, and the majority of the respondents are between 40 and 60 years (Table 1; Table 2, Section B). Most of the respondents have considerable work experience. On average, they have worked nearly 15 years (Table 1). More than a quarter have worked for more than 20 years (Table 2, Section C). This is a well-educated sample. All respondents who provided information on their education have a university degree, and the most common highest degree is Ph.D. Among the respondents, 40.0% hold doctoral degrees, 20.8% Master’s degrees, and 18.5% Bachelor’s degrees (Table 2, Section D). The largest group of respondents (40.5%) comes from academia, followed by government employees at various levels (37.6%), industry (11.6%), and NGOs (7.5%, see Table 2, Section F). By discipline, the largest group are engineers (Table 3). They are followed by respondents from a cluster of natural sciences: earth/environmental science/geology/ecology (Table 3). The next most sizable groups are social scientists (12.1%) and biology/microbiology (9.8%). A comparison between the CWN members and the outsiders among our respondents is presented in Table 4. As the table shows, CWN members tend to work in academia, hold doctoral degrees, be older and have longer work experience. Such differences are consistent with the focus of CWN activities and are an indication of the calibre of the people the Network works with. For instance, the majority of CWN members hold doctoral degrees and have worked over 10 years (Table 2, Sections C and D). By comparison, the majority of outsiders hold either Master’s or Bachelor’s degrees and are concentrated in the lower categories of work experience (Table 2, Sections C and D). In short, the typical CWN member among the survey respondents is an experienced academic, while the typical outsider is a slightly younger government employee. There are no significant differences between men and women although there are slightly more women among CWN members. 1

All tables are in Appendix 1.

3

1.2. Personal networks: To whom are the respondents connected? The survey data enable us to describe the water networks of the 173 respondents. The network level data characterize the network of individual respondents: they show who each respondent knows in the area of water—this group of people is referred to as their “water network”—and how they communicate and exchange information with them. Alternatively, it is possible to combine all the 1,904 ties for which respondents provide information, and discuss their overall characteristics regardless of who has provided the information. This “tie-level” data describes the entire community rather than individual respondents. When focusing on the individual respondents and their water networks, the data show that all our respondents have well-established networks in the water community and they tend to work directly with many of their ties. These network patterns are similar for CWN members and outsiders, although they are slightly more pronounced for CWN members. All respondents The typical respondent has known his network members in the area of water between five and ten years, contacts them a few times a year, works and exchanges ideas with the majority of them, and considers them acquaintances. Further, the data on work ties shows that for a large group of the respondents, the people they know in the area of water tend to be colleagues, partners, and collaborators. The respondents work directly with them, as opposed to simply knowing them or being aware of them. Almost half of the respondents (46.2%) work directly with the majority of the people in their water networks (Table 5, Section E). This group of respondents is actively working with members of their network in the area of water. Another sizable group of the respondents, 37.0%, works with fewer network members (Table 5, Section E). A relatively small group among all respondents,16.8%, works with just a fraction—less than one third—of their network members on their water issues (Table 5, Section E). These respondents know people in the area of water but work with only a fraction of them. This importance of work ties in the networks of the respondents reflects the selection of the respondents and the effort the survey requires for completion. We contacted people who are actively working and have connections in the area of water. Further, the respondents chose to describe their ties with colleagues and collaborators rather than their ties with people they simply know but do not work with directly. In other words, the survey captures the strong professional ties of the respondents. In that sense, the interesting result is not the importance of work ties but the differences among the respondents: some are actively working with their network members while others are only marginally involved with them. This suggests a diversity of the respondents which is consistent with the diversity of the stakeholders in the area of water: water issues cover a very broad content area, they are regulated under multiple jurisdictions, and concern a range of government, community, industry and academic organizations. Priorities, needs, and level of involvement in water issues of these diverse stakeholders vary.

4

A cross tabulation with sectoral data shows that the pattern of direct work with the majority of network members is common for some federal government employees and academics. By comparison, the provincial and local government staff work with fewer (between 30% and 70%) of their network members (Table 6). These are people whose main work responsibilities are in the area of water. In contrast, working with a small fraction of their network members is common for other federal employees and some industry staff (Table 6). Since federal agencies and businesses are very diverse, their involvement in the area is very different. Some federal agencies and businesses are actively working in the area, others are connected to—but do not work in—the area. The main work activities of such respondents likely require awareness and information gathering rather than direct contact with others in the water community. These three groups with different network characteristics will need and benefit from different CWN activities. CWN members Comparing CWN members and outsiders reveals only slight differences in their networks. When asked how close they are to each of their network members, both CWN members and outsiders show similar patterns. There are no differences between them in how many friends they have in their networks. Consistent with their older average age, CWN members tend to have known their water network members longer than outsiders, work with more of them, and contact each of them less frequently. For instance, many more CWN members work directly with most of their water network members compared to outsiders (Table 7, Section E). The majority of CWN members (55.0%) contact their network members a few times a year (Table 7, Section A). Outsiders, in contrast, are not concentrated in one modality of communication frequency: one third of them contact their network members a few times a year but almost as many contact their network members monthly. For a sizable group of outsiders, the average frequency of contact is weekly (Table 7, Section A). Overall, outsiders tend to contact their water network members more often. This is surprising. Because CWN members work with more of their network members compared to outsiders, they might be expected to contact them more often. Yet the opposite is true. A possible explanation is the long-term work schedules of CWN members. It is likely that their work ties are with colleagues and partners participating in CWN-funded projects, which have a relatively long duration. Further, the majority of CWN members are academics whose work also has long-term schedules. Indeed, the interview data with CWN members suggest that working with others on a project, whether CWN funded or not, does not require constant communication. Instead, project communication is concentrated in specific stages: writing the application and the reports, discussing the research design, or solving problems. Despite this burst of communication at certain stages, the average frequency of communication is not high.

5

In contrast, outsiders tend to be government employees, most often from municipalities (Table 4, Section E). The comments of interviewees, which touch upon the different time constraints for government and academia, suggest shorter duration and quick turnaround time for government employees compared to academics. In short, the work network characteristics of the overall sample—the large number of respondents working with the majority of their network members, five to 10-year duration of ties, majority acquaintances rather than friends—are more pronounced in the CWN sub-group than in the overall sample of respondents. On the average, academics contact their colleagues less frequently. But the average is misleading. Ties between academics vary greatly in their frequency of contact. The typical academic has few ties of frequent, intense collaboration and many less intensive ties with other academics— occasional interactions at conferences, etc. The survey data also show whether, in addition to working, respondents also exchange ideas with other participants in the water community. People often work and exchange ideas with the same colleagues, but it is not always the case. Compared to work ties, exchanging ideas is a more informal tie and at the same time requires trust. Hence, work and innovation ties may be quite different. Social psychologists often find disjunctions between what people say and what they do. That is the case with CWN. The data show an interesting dynamic: the actual and potential exchanges of ideas have different, almost opposite patterns. Table 8, Section A shows that over half of the respondents have discussed innovative ideas with a majority of their network members, suggesting a pattern of active exchanges of ideas. This is particularly common for CWN members who are mostly academics; tossing around ideas is common for them. At the same time, when asked whether they would exchange innovative ideas with others, respondents reveal a completely different pattern (Table 8, Section B): the majority of the respondents say they would share ideas with a very small proportion of their network members. They focus on obtaining ideas from a small set of other academics whom they trust as well as from grant-giving industry and government partners. Thus, when it comes to sharing ideas in the future, selectivity is the major pattern. By contrast, the respondents do not expect such selectivity on the part of their network members. They expect a sizable proportion of their network members to exchange ideas with them (Table 8, Section C). In other words, they believe they have the trust and respect of their colleagues, and they want to gather ideas—but not share them—with a wide range of network members.

1.3. Ties: Another look at the water community What do these relational characteristics mean for the community of people working in the area of water? Investigating all the ties of the survey respondents together provides a picture of the community.

6

The analysis now turns to the ties, or those 1,904 people for whom our respondents provide information in their surveys. These tie level data capture additional characteristics of the entire water community. The analysis suggests that the ties in the water community are dominated by academics and local government: these two groups are the backbone of the community. As previously discussed, the largest group of respondents is in academia. The proportion of all government employees combined (37.6%) is close but does not reach the proportion of academics (40.5%, see Table 2, Section F). Among all government respondents, those from municipalities are the most numerous (21.4%, see Table 2, Section F). The ties of the respondents are principally directed to other government staff and academics. The distribution of ties is not as strongly dominated by academics and by local government staff as the sectoral characteristics of the respondents suggest. For instance, the ties directed to academics comprise only 31.1% of the entire set of ties, even though academics are 40.5% of the respondents (Table 2, Section F; Table 9, Section A). Similarly, the ties directed to local government employees are only 16.1%, compared to a much stronger presence of such officials in the sample: 21.4% (Table 2, Section F; Table 7, Section A). Conversely, while industry employees as well as federal and provincial government staff are a smaller proportion of the sample, they comprise a larger proportion of the ties (Table 9, Section A; Table 2, Section F). These findings suggest that academics and local government staff have diverse networks that connect also to other government officials, industry practitioners, and members of non-governmental organizations (NGOs). The importance of academic and local government ties comes from their diversity as well as their sheer numbers. This is consistent with the data about who works with whom. A cross tabulation of sectoral data and the direction of ties (Table 10) more clearly shows who works with whom. Most of the ties of all respondents are within their own sector. This could only be expected: government employees work mostly with government employees, academics work mostly with academics and so on. Yet there are distinct patterns by sector. Academics are the most inward-looking group; they have ties above all with other academics, in fact half of their ties are directed to other academics (Table 10). Far behind their ties in academia are their ties to the federal government (14.0%) and industry (10.0%, see Table 10). Federal government staff is at the other end of the continuum. In fact, they are an exception: their ties with academics are more numerous than their ties with colleagues in other federal agencies, provincial or local government (Table 8). Local and provincial government, industry and NGO employees are in between academics and federal government: they work mostly with people in their own sector but are not as locked within it as academics.

7

Further, the distribution of respondents’ ties suggests that the ties of academics to other sectors are strongest in the federal government; academics are relatively weakly connected to provincial and local governments as well as NGOs. This is consistent with the ties of the federal government and municipalities. Federal government staff is connected strongly to academics. Local government, in turn, is connected to industry but to a much less extent than to academics. NGOs have the most evenly distributed ties with various sectors. However, they are well-connected to industry but weakly connected to the federal government (Table 10). Summary To summarize, academics and local government have strong presence in the area of water and ties directed to them dominate the community. Since respondents from each sector except federal government work mostly with their own sector, we can expect a fragmentation of the community along sectoral lines. When academics do work on water issues with partners outside academia, they work mostly with federal government.

8

Part II: Connections in the Water Network This section examines the connections in the water network as a whole. The discussion first examines the overall connectivity of the network and identifies the well connected members in the network, those who are most central. The analysis then turns to the internal divisions in the network drawing on “clique” analysis. The clique analysis describes the internal structure of the network, revealing sub-groups and connections between them. Finally, the discussion examines the ways researchers in the network cite the scholarly articles of colleagues. Citing other scholars can be treated as a specific type of tie. The citation analysis, therefore, provides an additional avenue to examine the connections among academics in the water network.

2.1. Centrality Analysis Network centrality, the number of connections a person has in a network, is the most common way to capture the connectivity of the overall network and the role of specific persons in it. The more connections each member of the network has, the higher the connectivity in the network. In highly connected networks, ideas travel quickly, members influence each other strongly, and resources can be mobilized easily. In turn, individuals with many connections, or with high centrality, are well-positioned to collaborate or exchange information with others. For instance, network members who are central are on communication paths that keep them in contact with others in the network; they receive information sooner than those who are less connected and benefit more from collaboration opportunities. Further, respondents may be connected to others either because they work with them or they may be connected because they exchange ideas with them. Respondents working with many collaborators may not be exchanging ideas with the same number of people, or they may be exchanging ideas with a very different set of people. The centrality of respondents is therefore calculated separately for working ties and for the ties that discuss innovative ideas. In a network, people connect to others either when they initiate a contact or when others seek them out and contact them. Thus, we can distinguish between two types of centrality. Outdegree centrality shows the extent to which a person is actively reaching out to others and initiating contacts with them. People with high outdegree centrality are the active networkers. By comparison, indegree centrality shows the extent to which other members of the network contact a particular person. People with high indegree centrality have prestige and status; they can be considered the established experts in the network. Both types of centrality reflect connectivity and are crucial for maintaining the network. Whether reaching out or responding to others, centrally located individuals hold the network together.

9

Overall connectivity The analysis suggests that the connectivity in the network is low 2 . Centrality varies depending on the particular measure (indegree or outdegree) and the types of ties examined (work ties or exchanging innovative ideas). The highest number of connections a respondent has, for instance, varies from a mean of 6 (work outdegree) to 10 (work indegree). Yet in each type of measurement, only about a dozen respondents are linked to the network by three or more ties. The network is sparse. Why is the connectivity so low? These results should be viewed in the context of the way researchers work and the diversity of the water community. All the respondents report the most important professional ties they have with people working in the area of water. Thus, the survey captures relatively strong professional ties and people have few such ties. Previous research shows, and our interview data confirm, that academics—who are almost half of the survey respondents—work closely with only a few colleagues. This is particularly visible on large research projects; even when the project team includes a dozen or more people, each project team member works closely with just a few people (see also clique analysis). These fewer but stronger ties are the type of ties captured in the survey. In contrast, the survey disregards weak professional ties. For instance, several members of a large research project have filled in the survey and listed their strong professional ties. However, they do not work closely with each other, and the analysis found no ties among them. For non-academics (over half of the sample), such strong professional ties in the area of water are likely to be even fewer. The area of water is known for the breadth of issues and the diversity of stakeholders in it. As some respondents indicate, water issues are “everybody’s concern.” The responsibility for policy and management in this area is shared by all levels of government and several agencies. A number of NGOs and industries are also involved in water issues. The points of common interest among such diverse stakeholders are likely to be few; the practitioners are therefore likely to work separately from each other. In turn, their connections to academics depend on their job and the specific needs of their organization at the moment. For many of them, their main work responsibilities are unrelated to research activities and outside of the area of water. In short, the diversity of the stakeholders fragments the community and decreases the overall connectivity. The results show that many of these non-academic respondents, especially those not involved in CWN, neither work closely together nor share ideas with others in the water network. Finally, the low connectivity in the water network is also affected by the fact that the 173 respondents who filled in the survey are just a sample of all the people working in the area of water. The analysis looks for connections among the 173 respondents who filled in the survey. Some respondents work and exchange ideas with collaborators in the area

1

Construction of the network data was derived from the survey results of personal networks in which relations with others in the water community were described. The connections of each respondent reported here are connections to the 173 people who filled in the survey. See Appendix 3 for details.

10

of water, but their collaborators have not filled in the survey. Therefore, these connections are not reflected in the analysis. In short, strong professional ties in the water network are likely to be few. The analysis captures this low connectivity. The sampling further emphasizes it. Work centrality Who are the most connected respondents in the survey and what are their connections? The discussion next examines those respondents who are better connected to others and are thus the most active in the water community. These are the respondents central to the network. The results of the survey show that the respondents most actively working with others are not always those who most actively exchange innovative ideas (Table 11). The two groups only partially overlap. Similarly, those who most actively contact others (high outdegree) are not necessarily those who are most often sought out by others (high indegree); the two groups only partially overlap. This pattern of asymmetric ties, quite common for network studies, holds true for both working ties and sharing innovative ideas. That is why it is important to examine them separately. About a dozen respondents are actively working with others in the water network, or they have high outdegree centrality (Table 11). These are the respondents who have strong interest in collaboration and are focusing their networking in the area. The interview data show that among those seeking out collaborators are several senior academics strongly involved in CWN work; two of them are project leads. At the same time, some of these active networkers are junior researchers and outsiders not who are not currently involved in CWN. In other words, the interest and focus on collaborators in the area is stronger among junior academics—those still building careers and expanding their personal networks. The presence of outsiders is particularly interesting. The person who is most central of all the active networkers is an outsider, a senior government official from the federal government; he works with six other respondents in the water network. The presence of outsiders among the active networkers is a good indication of their interest in collaborative research and the potential for developing new connections. Roughly the same number of respondents—about a dozen—are named as collaborators by at least three others members in the network (indegree). But this is a different group of people: only four of the people actively reaching to work with others are also among those most often sought by others; these four people do not have the highest indegree centrality scores. In other words, there is a low overlap between the respondents with high outdegree and those with high indegree. If the first group of respondents with high outdegree includes respondents interested in collaborative work, this second group of respondents with high indegree consists of experts with established reputations in the area who are attractive collaborators for others. There are no outsiders among them. All but

11

one are academics. About half of them lead CWN projects. The most central person (indegree), named by 10 other members in the network, is the project lead of a large CWN project. Innovation centrality The connections among respondents who exchange innovative ideas show similar patterns to the connections among those working together (See Table 11). A small group of slightly more than a dozen respondents is connected to the networks by three or more ties. Those who seek out others to share their ideas are not necessarily the recipients of such ideas; only five respondents with high outdegree also have high indegree. Notably, over half (7) of those who initiate contact to discuss innovative ideas are not academics: they are government, NGO and industry staff. Almost as many (6) are outsiders to CWN. These are people with ideas who seek out the experts in the water network. In contrast, the majority of the established experts who attract the interest and trust of others are senior academics who lead CWN projects. All are involved in CWN. Comparisons across type of ties and types of centrality Comparing work ties with ties for exchanging innovative ideas shows that the respondents who actively contact others for work are often the same people who contact others to share their innovative ideas with them. Alternatively, those who are named by others as collaborators are often the same people with whom others want to share ideas. The overlap suggests that people behave consistently across their ties. Active networkers tend to be well connected in the water network because they both work and exchange ideas with others in the water network. Similarly, high status experts attract others as both work collaborators and consultants on innovative ideas. Where differences between work and sharing innovative ideas do emerge, it points to higher proportion of non-academics and outsiders. Sharing innovative ideas, in other words, evokes diverse participants. Summary of centrality analysis In sum, the water network is only sparsely connected. Only a small group of the respondents (29) are better connected to the network, either because they initiate or attract connections by others. CWN members comprise the majority of these central network members (22) and therefore contribute most to the connectivity in the network. The respondents who hold the network together through their connections are divided into two relatively different groups. The active networkers, who are interested in collaborative work, include a sizable number of non-academics and people outside CWN. Young academics building their careers as well as some established academics are also looking for collaborators in the network. The presence of outsiders in the group—people who reach out to CWN researchers—is evidence of their interest in the work of CWN. By comparison, the second group of central people who contribute to the connectivity in the network by attracting others, or the established experts, are overwhelmingly CWN members and academics. Their centrality to others suggests the role of CWN in the water community as a focus of expertise.

12

2.2. Clique Analysis Understanding how a network functions is impossible without examining the internal structure. All networks have their own internal divisions and typically include several sub-groups of people who are closely connected. In turn, sub-groups may be connected to each other to a different degree. The number and size of the internal sub-groups as well as their connections affect the processes unfolding in the network. They determine how information travels within the network or how resources are mobilized. For example, if a network consists of many small groups that are not connected to each other, information and resources are not easily shared across the network. In such networks, information spreads slowly and resources available in the sub-groups are not pooled together. In contrast, if a network includes people who are members of more than one sub-group and thus can connect the subgroups, information and resources travel more easily across the network. Information spreads rapidly throughout such a network, jumping from one subgroup to another with the help of overlapping members. This section examines the results of analyses that identify a specific type of sub-group within the network—cliques, or groups of individuals who are closely connected. They might be working together, exchanging information and ideas, or pooling resources. In all cases, they interact directly and are more strongly connected to each other than they are to the rest of the network. Clique analysis, in other words, identifies the groups of people who are strongly connected to each other. The analysis examined two types of cliques: cliques based on working ties, in which members work closely with each other, and cliques based on exchanging innovative ideas, in which members extensively discuss their ideas. The results of the clique analysis address several questions: How many cliques of close collaborators and discussants are there in the water network? How big are they? Who works with whom in a clique? Are the existing cliques connected, i.e., are there individuals who are members of more than one clique? Finally, do the people who work closely together also exchange ideas or do people work with some colleagues but exchange ideas with others? In other words, are work cliques similar to cliques discussing innovative ideas? Work cliques The analysis found 12 small cliques, or groups of close collaborators, in the water network. Figure 1A and 1B 3 are sociograms of the work relations, i.e., they are visual representation of the ties among network members who work together (Appendix 2). The graph includes 86 respondents who work with at least one other person in the network. The analysis showed that these respondents tend to work closely with only one or two other collaborators. There are many dyads but no groups of close collaborators that are larger than three members. The small size of work cliques is consistent with the qualitative data on project practices. While projects may include numerous researchers and partners, most of them work independently from each other. Daily work is done in 3

All figures are in Appendix 2. Figure 1A and 1B both represent the same work relations; the symbols used in Figure 1A indicate the sector of the respondents while the symbols used in Figure 1B indicate whether the respondents are members of CWN or outsiders.

13

small groups of close collaborators. Researchers switch from one small group to another depending on the stage of the project. A dozen such three-member work cliques exist in the networks (Table 12). Half of the respondents involved in cliques (8) are members of more than one clique; three of them are members of five or six cliques of closely related members. Because of these overlapping members, all 12 work cliques taken together include 16 people (Table 12). In short, a small number of the active collaborators in the water network (16) are involved in closely collaborating groups. Such respondents always work with two close collaborators in a group but are often involved in more than one group. These are the active collaborators in the network. The rest of the respondents might have close collaborators, but their close collaborators are either outside the water network or simply did not complete the survey. Who are the active collaborators in the network? The majority of the 16 members of the existing work cliques are CWN members (13 out of 16) and academics (11 out of 16). More than half of them are over 50 years old and have long work experience. In other words, the typical active collaborator in the water network, as captured in the survey, is a senior academic with a lot of experience who is working on a CWN project(s). However, five of the respondents who are closely working with others in the area of water, are employees from various levels of government. While academics dominate, one third of the active collaborators are government employees. What is even more interesting, three of the government employees are outsiders to CWN. One of these outsiders, a federal government employee, is involved in three work cliques. This suggests that some of the key collaborators in the water network are outside CWN and that there are still important partnerships to be built between CWN and the federal agencies. Who works closely with whom in a clique? Who works closely with whom is the next key question in understanding the network. If academics work with each other, or government employees keep to themselves, or work cliques are drawn by a single discipline, this tell us that their research is hardly crosssectoral or multidisciplinary. A closer look inside the work cliques reveals, however, that the opposite is the case. The analysis showed that work cliques in the water network cut across sectors and disciplines. In other words, collaborative research in the area of water tends to be crosssectoral and multidisciplinary. This finding is all the more significant since the work cliques include only three members. Despite this, cliques bring together collaborators with diverse backgrounds. Table 13 shows that more than half (7) of the work cliques are cross-sectoral; they include two academics and a non-academic (#1, #3, #5, #9, #10, #11, #12). The nonacademic collaborators are government employees at the three levels of government. 14

Most of the collaboration, in other words, takes place among academic and government employees. The most sought after collaborator is a federal employee: he is a member of three different cliques, each in a different part of the country. Since he is also the outsider mentioned above this result reinforces the idea that there are untapped connections between CWN and the federal government. Equally important, the composition of the work cliques is multidisciplinary, with a heavy representation of biology, followed by earth/environmental science/geology, and engineering. By contrast, there is a poor representation of social sciences (Table 13). The analysis found that none of the work cliques draws its members from a single discipline. Instead, virtually all work cliques are multidisciplinary; in half of them each member comes from a different disciplinary background. Biology is the most prominent discipline in the network. Two thirds of the cliques include at least one biologist (#3, #4; #5, #6, #7, #8, #9, #11); all of them include either a biologist or an epidemiologist. In several cliques, two of the members are biologists. As a result, in most of the work cliques, members are involved in research projects related to biological issues. There are four to five biologists who participate in the work cliques (four biologists and one microbiologist). Two of them are much more active collaborators: they are involved in five and six cliques respectively. Both of them are CWN members who have been invited to participate in a number of research projects. It is through their collaborative work that biology takes such a central place in the work cliques. The next two areas that figure prominently in the work cliques are earth/environmental science/geology (multidisciplinary by nature) and engineering. Two thirds of the cliques include earth/environmental scientists (#1, #2, #4, #5, #6, #8, #9, #12). There are five people with earth/environmental sciences backgrounds who participate in the work cliques. None of them, however, is a member of more than two work cliques; they do not contribute to the same extent as biologists to the water network. Engineering is represented in half of the work cliques (#1, #2, #3, #10, #11, #12) even though there are only two engineers. Both of them participate in several cliques, ensuring the high representation of their discipline. The social sciences are not well represented in the work cliques. Out of the 12 existing cliques, only three contain a single collaborator with a background in social sciences (#3, #10, #7). They collaborate with biologists and engineers. Each of the collaborators—a geographer and two economists—participates in a single clique. The cliques in the network are various configurations drawing on these three popular disciplines. The most common combinations include epidemiology, biology and environmental sciences, or biology with environmental sciences and engineering. For instance, one of the well-connected researchers is an engineer who appears in three cliques along with a biologist, epidemiologist, and environmental scientist.

15

In short, there is no doubt that the active collaborators in the water network, most of whom are CWN researchers, are doing cross-sectoral and multidisciplinary research. They work closely with government employees (although some of their partnerships are outside CWN) and collaborate with other academics outside their own discipline. However, the disciplines from which they draw their collaborators are limited and social sciences are underrepresented in the collaborations. Are the work cliques connected? It is important to examine the connections among cliques. Without such connections, the larger water network would be only a collection of independent groups, in which members work only among themselves and do not have common work interests. Given that the work cliques are quite small—only three members—such a situation would mean there was a very limited circulation of information and resources within the larger network. In turn, with only a few connections among the cliques, the network would be vulnerable to slipping back to a disconnected state. If only a few members collaborate in several groups, all these connections would be removed if they were to leave the network. The analysis showed that eight of the active collaborators in the network are members of more than one work clique. In other words, there is a significant overlap in the membership of the work cliques and this ensures that many of the work cliques are interconnected. This suggests that active collaborators in the network have common work interests that link them together in various configurations. Further, the active collaborators contribute to different degrees to these interconnections. Three of them are involved in as many as five or six work cliques while five other collaborators are in two or three work cliques. The remaining eight participants in work cliques are working with members of only one clique. In short, there are connections that cut across the work cliques and hold the overall water network together. People working in the area of water are thus a network and not a collection of independent groups. Yet the people contributing to these integrative connections are relatively few. Among them, an even smaller number contributes disproportionately to these connections. While for the active collaborators themselves such connections mean access to resources and information, for the network as a whole this dependence on a few key participants reveals a weakness. What brings work cliques together? How active collaborators come together to create work cliques is important not only for the understanding of the network but also for possible interventions in the network. This is not a matter that the survey can answer. However, documents and interview data provide some clues. About one third of the work cliques identified in the analysis are most likely based on CWN projects. Such work cliques consist of people who work together on the same CWN project (Cliques #2, #4, #8, #5). Most of them include the project lead, senior

16

researcher, or partners. These groups of collaborators have come together because of their CWN project. To put it differently, in these groups the membership in the work cliques is a function of the membership in CWN. A surprisingly sizable number (6) of work cliques, however, includes two academics working on the same CWN project and a government employee who is not formally listed as a partner on their project (#1, #3, #9, #10, #11, #12). It is unclear in these cases whether members work together on a CWN project or on a project funded by a different agency. In the first scenario, it is possible that the third member of the clique, the government employee, may not be listed as a partner on the CWN project for reasons of authority. The formal proposals often include high ranking contact persons from the government who do not necessarily do the everyday work. In contrast, the clique analysis captures the government employee who works on a day-to-day basis with the academics. Personnel changes in the government can also change the members of a work clique without changing the nature of the partnership and the collaborative work. Alternatively, in the second scenario, the members of the clique are working on a project unrelated to CWN. Their relationship goes beyond CWN. It is not clear which of the two scenarios corresponds to reality in each of the cliques with such composition. In both cases, however, the active collaborators demonstrate a commitment to cross-sectoral research and solid connections to the government. Finally, a third set of work cliques cuts across projects; it includes members from two CWN project teams (#6, #7, # 9). Such groups most likely work together on projects not funded by CWN. They extend their collaborative relationship across several projects. This is consistent with what the interview data reveal about the way researchers work together. Researchers are typically involved in several projects and thus work with collaborators from several formal work groups. At the same time, most researchers consciously build a group of close collaborators, and they invite them to participate in multiple projects. This is particularly true for senior researchers. Their close collaborators get involved in different configurations, and in several projects. Collaborative ties with them transfer across several formal projects. Summary To summarize, the membership in a work clique does not closely follow CWN project teams. While the majority of active collaborators in the water network are CWN members, they are not necessarily working on a CWN funded project. These results reflect the fact that research in the area of water is funded by many agencies including CWN. The collaborative ties in the water network do not all arise in CWN projects and are not entirely dependent on the work of CWN. These results are consistent with the existence of active collaborators outside of CWN. On the other hand, CWN plays a crucial role in the water network; the majority of the active collaborators in the network, and certainly all of the academics among them, are

17

CWN members. Whether CWN attracted active collaborators or, alternatively, helped its members expand their collaborative ties (interview data suggests that both processes are taking place), it has been able to link those academics who are interested in collaborative work in the area of water. Innovation Cliques Innovative ideas can be interpreted as a distinct resource in networks. The way that innovative ideas travel in a network does not necessarily follow work ties. In some cases, collaborators are experts with complementary expertise who can bring a fresh look to an issue; in others, they are close collaborators who act as sounding boards. In such cases, collaborators from a work clique not only work together but also exchange innovative ideas. Work cliques coincide with the innovation cliques. Yet there are also good reasons to expect differences between the two types of cliques. Innovative ideas are often cross-sectoral in nature. Further, people outside one’s own work group and outside one’s own discipline can bring unexpected ideas. We would expect, therefore, exchanges of innovative ideas to occur more across sectors and disciplines. How do work cliques and innovation cliques in the water network compare? Figures 2A and 2B present the innovation exchanges in the network 4 . The clique analysis found 12 small innovation cliques with three members each (Table 14). Despite the opportunities for different networks, in practice, most of the cliques coincide with work cliques (#1, #2, #3, #5, #7, #10, #11, #12). Just a third of the cliques contain one or two new members (#4, #6, #8, #9). In other words, respondents not only work with their close collaborators but also discuss their innovative ideas with them. Nonetheless, there are some interesting differences between innovation and work cliques. Who exchanges ideas with whom? Compared to work cliques, the data in Table 14 reveal that the participants in innovation cliques have fewer outsiders (2 out of 19) and more non-academics; almost half of the participants are outside academia (9 out of 19). Such changes in the background of the participants in innovation cliques can be expected; the sharing and implementation of innovative work and ideas involves collaboration between academics and non-academics, whereas collaborative research work is more limited to ties among academics. What is perhaps unexpected is that the non-academic participants in the innovative cliques are somewhat different than those in work cliques. Innovative cliques, nonacademic participants are more evenly distributed across various sectors: there are employees from the federal government (2), provincial government (1), local government 4

Figure 2A and 2B both represent the same innovative relations; the symbols used in Figure 2A indicate the sector of the respondents while the symbols used in Figure 2B indicate whether the respondents are members of CWN or outsiders.

18

(1), as well as representatives from industry (2) and NGOs (2). Notably, some of the local government participants from work cliques are not present. Thus, industry and NGO representatives not present in work cliques become members of innovation cliques. Disciplinary characteristics of the innovation cliques also slightly change compared to work cliques. Biology retains its prominence: just as in work groups, two-thirds of the cliques include a biologist and some are entirely based on biologists. However, the role of environmental sciences and engineering decreases. The number of social scientists in the group slightly increases due to the participation of more non-academics from government and NGOs. Are innovation cliques connected? Connections between such innovative cliques are especially important since such connections facilitate the spread of ideas in the larger network. Yet it is much easier to share ideas with only close collaborators. The analysis shows that the innovation cliques are connected albeit to a lesser degree compared to work groups. Out of 19 participants in innovation cliques who exchange ideas, five are involved in two cliques. An additional two are involved in five and six cliques respectively. With a few exceptions, the clique members who connect the innovation cliques are the same clique members who connect the work cliques in the network. In other words, respondents who are most actively working with collaborators from different work groups are also most actively exchanging innovative ideas with their collaborators from different groups. It is significant that the two members involved in the highest number of innovation cliques are the same two members involved in the highest number of work cliques. Since they are both biologists, they reinforce the centrality of biological sciences in the innovation network. Bridges: Who are the people connecting the cliques? The analysis of cliques showed that some respondents are members of more than one group. Two of them are also cutpoints in the data, meaning they are central to both the work and innovation networks, connecting and acting as bridges between cliques (the people identified as 146; 122, Figures 1 and 2). Such network members are known as bridges connecting otherwise disconnected groups. People in such positions act as information or resource brokers and boundary spanners within the group. They are a key to the “health” of the network; their removal could result in fragmentation of the network. Demographic information, as well as the clique analysis and centrality scores, can identify who are the important bridges and how they behave in the network. The two respondents in bridge positions are academics in mid-career. Both are biologists, the most prevalent discipline in the network. Both work on a CWN funded project although neither of them is a project lead. In other words, they are not junior researchers; they have had time to develop their professional ties in the network. Neither are they among the most senior members of the network who have less time for networking, get more easily

19

funded by various agencies, and work on many projects whose participants may be unconnected to CWN. The clique analysis also shows that the two bridges work and exchange ideas with academics and partners across the country, i.e., they have developed broad networks of distant collaborators and partners. Quite likely, in addition to their CWN project, they have research projects not funded by CWN. In turn, centrality analysis shows that they are among the few people who both reach out to others and are named by others who want to collaborate and exchange ideas with them. Their indegree and outdegree centrality scores are quite similar. Finally, they are involved not only in CWN but also in other organizations in the area of water. There is a significant overlap in the membership of the two organizations. This helps them develop and strengthen their ties to other researchers in the area. Summary Several findings emerge in the clique analysis. First, the research taking place in the network is cross-sectoral and multidisciplinary. The composition of the work and innovation cliques bears evidence to that. The finding is all the more important given the small size of the cliques. Second, CWN plays a central role in supporting these partnerships and research although it is not the only player in the area of water. The clique composition shows that it is CWN members and especially academics who dominate research collaboration. Not all of their work is on CWN projects—other funding agencies also support such complex collaborations—but CWN members are always a strong presence in these collaborative groups. Work and innovation cliques have similar memberships, although outsiders are more active in exchanging innovative ideas than in work cliques. Third, clique analysis confirms the opportunities for expansion of CWN memberships in government and industry. Outsiders from industry, federal and local government are already collaborating and exchanging ideas with CWN academics. This is consistent with the centrality analysis which shows industry and government employees actively reaching out to the networks. In this respect, the federal and provincial employees are particularly active. Notably, the analysis of personal level networks (Table 10) confirms that the largest proportion (26%) of the ties of federal employees is directed to academics in universities and not to people in their own sector. The interview data also show that some federal employees feel an affinity to academics. In short, if local government employees are already involved in research and exchanging ideas, federal employees seem to be the most important, albeit not the only, untapped resource for expanding CWN. Finally, the analysis suggests two weaknesses in addition to the overall low connectivity discussed in the previous section. Multidisciplinary research draws on a limited range of disciplines; biology dominates collaboration, followed by environmental science and

20

engineering, while social sciences are not well represented. Next, there are only two significant bridges in the network, making it vulnerable to changes.

2.3. Citation analysis Citation analysis is a technique that examines how scholars cite each other. Citing another scholar is interpreted as a specific type of tie. A group of authors citing each other is thus seen as a network of scholars who have common interests, work in the same area, and read each other’s publications. In a university environment, publications measure the value of academic work and citing is the most common recognition of this value. Citation analysis therefore captures the most significant as well as the most traditional professional tie among scholars. The discussion below presents the result of citation analysis for a small group of 31 scholars 5 . Applied to CWN members, citation analysis can show whether scholars in the network are connected by this traditional professional tie. The analysis can tell a lot about what is happening in the network by answering questions such as: do scholars in the group cite each other? In other words, does their collaboration end with the report for the project or do they continue to follow each other’s publications? Citing indicates that their publications are related and relevant for their colleagues. Who are the most often cited? In others words, who are the most visible and influential scholars? Alternatively, who cites others most often? Such scholars are familiar with the work of their colleagues and are able to link it to their own work. Citation analysis is particularly important for the CWN by revealing who cites whom in a group of scholars and thus mapping the internal structure of the group. Combined with information on the disciplinary background of scholars, this can show whether scholars from different disciplines work on common issues and find each other’s publications relevant in their own work. The analysis includes several types of measures, each of which examines a distinct citation behaviour and adds to our understanding of the network. The first measure is cocitation, or how many times any two authors in the group are cited together by anyone in any field, whether in the group or not. It is a pairwise measure. Repeatedly cocited authors are perceived to work on related issues; their work is either similar or complementary in some respect. Cocitation captures opportunities for collaboration. By comparison, intercitation shows how many times scholars from the selected group cite each other directly. The role of a scholar who frequently cites others in the group differs from that of the scholar whom others frequently cite. Scholars who cite others in the group are familiar with their work, find it relevant to their own, and recognize its

5

Citation analysis is only feasible for small groups. This analysis includes the 31scholars who are among the most central people in the network. Compared to the group of people who are central in working and exchanging ideas, this 31-member group excludes respondents outside CWN. Details of methodological issues are available in Appendix 5.

21

importance. Senior scholars citing others act as integrators of the intellectual contributions of their colleagues. In contrast, scholars who are widely cited have prestige and influence with their colleagues but may not be citing them; they may not even be familiar with their colleagues’ work at all. Such scholars are thought leaders in the group. Intercitation thus includes two separate measures. The first captures the instances in which a scholar cites others in his or her group. The second captures instances in which a scholar is cited by others in his or her group 6 . Centrality of Authors in Citation Networks: Do Scholars Cite Each Other? The scholars included in the citation analysis are all members of CWN and represent over a dozen CWN projects. Most are either biologists (10) or earth scientists (7); the rest are almost evenly distributed among chemistry, health, engineering, geography and business. The multidisciplinary composition of the group makes dense citation within the group unlikely. Most scholarly journals publish work in a single discipline and their criteria for acceptance reinforce disciplinary boundaries (Section 2). Citing the scholars you work with cannot be taken for granted: even for people working in multidisciplinary teams, publishing and citing can fold back within one’s discipline to increase the likelihood of publication. Citing, therefore, is the most rigorous test for the existence of crossdisciplinary connections. At the same time, citing across disciplines may strongly indicate the interdisciplinary nature of the work being done in the group. Do scholars cite their colleagues? The analysis suggests several salient patterns. First, scholars in the group do not cite their colleagues very often. Citations within the group are not frequent and connectivity in the citation network is low. This is consistent with the low connectivity found in the survey. To be sure, the majority of the scholars are in some measure connected to the group: they are cocited with others in it, and they cite or are cited by others in it (Table 15). However, most of the citation counts are small and some scholars neither cite nor are cited within the group. Individual scholars, though, differ considerably in terms of citation. Second, the scholars in the group are more often cited together (cocitation) than they cite each other (intercitation). Almost everyone is cocited with at least one other group member: there are relatively few zeros indicating that the scholar has not been cocited at all. This suggests that the work of the scholars in the group is perceived as having a certain measure of coherence; they are seen as working on the same issues.

6

These measures are similar to the centrality measures in any other type of tie and have already been used in Section 4 to show the active networkers (outdegree) and established experts (indegree).

22

Table 15 compares the number of other persons in the group that each member is cocited with (Cocite Degree), the number of persons each cites (Intercite Outdegree), and the number each is cited by (Intercite Indegree). Values above 4 are shaded. Intercitation among members is responsible for some of the cocitation they receive. As indicated by the zeros, two scholars in the group have not been cocited, and four neither cite nor are cited by their colleagues. This is consistent with the diverse disciplines from which the scholars in the group come. In sum, scholars in the group do not cite their colleagues very often. While they can see the connections between their work and the publications of a few other colleagues, they have trouble relating most of their colleagues’ publications to their own research. Cocitation Centrality: Who is most frequently cited together with other colleagues? As noted, cocitation refers to the instances in which pairs of authors are cited together by any citer in any field. Scholars with high cocitation scores usually write on similar topics and use methods in common. Almost all of the scholars in the group are cocited with at least one of their colleagues, and one is cocited with 11 others (Table 15). Such differences may be linked both to individual characteristics, such as the career trajectory of the scholar, and to contextual characteristics, such as the disciplinary composition of the group and the established practices in it. When the articles that cocite pairs of CWN authors are counted, the analysis shows that the scholars who are most often cited together tend to be biologists and earth scientists. The top scholar on the list, who is cocited with other group members in a total of 60 articles, is a biologist, works in large projects, and participates in more than one CWN project. Participating in several projects underscores the fact that his work is related and fits well with the work of his colleagues. The centrality analysis in Section 2.1 shows that he is often named by others as a collaborator and a person with whom others want to share ideas. His project participation reinforces his visibility in the group, makes him a familiar name to others, and increases the likelihood that that his work will be seen as related to the work of other colleagues. Yet, this is not the full explanation: there are other scholars who are also senior and participate in several projects but are not cocited so frequently. To understand better what is happening, the analysis needs to take into account the broader characteristics and internal structure of the networks. The discussion turns to the question of particular author pairs. Cocitation Map: Whom are they cited together with? Figure 3 is a map of the cocitation links that shows which scholars are cited together. A link between any two scholars indicates that they have been cocited in at least one article. The thickness of the link indicates the frequency of cocitation: scholars joined with heavier lines are cocited in many more articles than those with lighter lines. The heaviest

23

link is between a pair of authors who were cocited in 23 articles. Disciplinary background is indicated by colour. The map shows that the central people are cited together with scholars from several disciplines. For instance, one of the two most central people is cited together with biologists, engineers, chemists, and health scientists. This is what leads to his centrality. The second person is cited more often with people in his own discipline, but there is still disciplinary diversity. This pattern of diversity is common for many of the scholars, even though they are less frequently cocited. Figure 3 shows an internal structure in the group that further suggests patterns of cocitation across disciplinary boundaries. The cocitation map indicates that scholars tend link up in groups of three. The majority of these groups are composed of authors from different disciplinary backgrounds. In other words, their works are perceived as relevant to issues in several other disciplines and cited together. This pattern is an indication of the multidisciplinary relevance of the publications of these scholars. On the other hand, several patterns suggest the impact of disciplinary boundaries. Some of the scholars, who are cited together, include only biologists (#164, #122, #152) or only economists (129, 126, 149). Some of the biologists have very strong connections among themselves. Scientists in economics, geography, and health tend to be at the periphery of the network. In short, while there are many cross-disciplinary connections, disciplinary boundaries remain important. These patterns suggest an interesting dynamic of the disciplines in the group, best illustrated in a comparison between biology and economics. Biologists are the most numerous in the group of scholars included in the citation analysis. They dominate the work cliques (Section 2.2) and are highly visible in the network. Given this wide participation and visibility, it is easy to perceive their publications as relevant to many disciplines; biologists are therefore often cited together with scholars from other disciplines. At the same time, the sheer numbers of biologists is a temptation to cite biologists only. This accounts for the strong cocitation links between several of the biologists in the group. In contrast, social scientists are not well represented in the group included for citation analysis. They are less central in work cliques and less visible (Section 2.2). Other scholars have trouble finding the links between a social sciences discipline such as economics and others more popular in the CWN disciplines. Economists, therefore, tend to be cited together with other economists; their participation in cross-disciplinary cocitation is very low. In short, cocitation suggests that, with a couple of exceptions, all scientists are perceived as having at least minimal ties with colleagues in the group. There is a perception that scholars in the network work on common issues. Working actively in the network facilitates such perception of relevance but it is not the only factor that affects it. Individual participation in CWN projects interacts with discipline to determine who is

24

most visible and most often cited with others. Further, cocitation indicates many crossdisciplinary connections, but also that disciplinary boundaries remain important. This is particularly important for the social sciences and to some extent for the health sciences, which remain locked in their own disciplines. Intercitation Centrality: Who cites other colleagues? Intercitation occurs when scholars cite each other directly. Scholars who often cite their colleagues within this set are familiar with their publications and recognize their relevance and importance. Typically, they may be junior to these colleagues and are citing to establish the credibility of their own work. However, there are no junior scholars in the CWN group. All members are well-established in their careers and are often leading scientists in their disciplines. This, together with publication pressures to cite people within one’s own discipline, is part of the explanation why there are relatively few instances in which scholars in the group cite their colleagues, despite the many opportunities to do so. The more interesting finding is that in this group, scholars who cite their colleagues most frequently are not juniors. The person who cites group colleagues in the most articles is indisputably a senior scholar (127). He is also actively working and highly visible in CWN. He has cited colleagues in 48 articles. (By contrast, seven scholars in the group have not cited anyone at all.) The average outcitation score is 6. Notably, the next person on the list, who has cited others in 22 articles, is also a well-established scientist. These scholars are not only familiar with the work of their colleagues, but are also able to find the connections between the work of others and their own. They synthesize and integrate diverse contributions. Figure 4 shows that the scholar on the top of the list cites not only colleagues from his own discipline (biology), but also colleagues from several other disciplines. In short, he integrates the intellectual contributions of the other scholars in the group. Who is cited by colleagues? In contrast, scholars who are cited by their colleagues are visible and influential. This different role in the group is played by different people. Despite some overlap, the scholars who are cited by colleagues in the most articles are not the scholars who most often cite colleagues in articles of their own. The person most often cited by others in this group is, not surprisingly, another biologist in a senior position who is actively involved in CWN (164, Figure 4). He is highly visible among colleagues, often being named by them as a collaborator and as a person with whom they want to share ideas (Section 2.2). They cite him much more often than other members—in some 48 of their articles. (That his 48 incitations match the other leader’s 48 outcitations is a coincidence.) Although he is cited by scholars in his own discipline, scholars from different disciplines also cite him (Figure 4). In other words, they find his publications relevant in their own work—a strong recognition of his importance. (He is also the group member most highly cocited with other members.) This scholar, hence, is

25

in a position to influence his colleagues—he is a thought leader. His closest competitor in this role is cited in 17 articles. The average number of articles in which the group cites a member is 6. Thought leaders are all the more important because being cited by colleagues does not happen very often: while these scholars are leading scientists in their areas, publication criteria strengthen disciplinary boundaries and limit within-group citations. This eliminates one of the major channels for influence among scholars. Moreover, disciplinary boundaries are reinforced by organizational boundaries, which further decrease opportunities for influence. Who cites whom? Who is cited by whom? The two scholars leading the intercitation lists in sending and receiving citations are both connected to colleagues from different disciplines. Is this pattern of citing across disciplinary boundaries common for other members of the group? In other words, do intercitation patterns indicate multidisciplinary connections? Figure 4 maps the structure of the intercitation connections in the group. Direct citation is less frequent, and more scholars become disconnected from the group. The remaining connections are quite similar to the cocitation map, but they seem to follow disciplinary boundaries more closely. Direct citation happens in groups of two and three scholars and many of the groups are familiar from the cocitation analysis. The three economists remain connected among themselves at the periphery of the map. The biologists around the central person in the cocitation map are still closely connected to each other. In other words, scholars who are cited together also tend to cite each other. If they work on common problems, as evidenced in the cocitation map, they will cite each other and will have a link in the intercitation map. At the same time, the multidisciplinary nature of the connections decreases. Many of the scholars continue to have connections to other disciplines. However, compared to the cocitation map, some groups in the intercitation map have lost the member who contributed to their diversity; a few of the scholars are now connected to a single colleague from their own discipline. In other words, scholars tend to cite and be cited by colleagues in their own discipline. This holds true even when they work on common problems with colleagues outside their discipline and are cited together with them. In several cases, scholars from different disciplines are working together on a CWN project yet do not cite each other. Summary In short, when it comes to directly citing their colleagues, scholars tend to follow disciplinary boundaries. Current publication practices in most scholarly journals reinforce such preferences. This confirms the patterns mentioned earlier: while scholars in the group are perceived as working on common problems and cited together, these scholars themselves find it difficult to integrate the work of their colleagues from other disciplines in their own work.

26

Part III. The context of collaborative work To understand why the researchers and practitioners in the area of water connect to each other the way they do, it is necessary to look at the broader context of their work. In the previous sections of the report, the analysis has on occasion referred to interview and documentary data to explain results. The next sections draw heavily on interview data to discuss in more detail several themes that have emerged in the comments of interviewees. First, the analysis examines barriers and incentives for complex collaborative work of the type fostered by CWN that explain the age and status composition of CWN membership. Second, it looks at the challenges in doing collaborative work and the strategies for overcoming them. These are related to connectivity in the network, the characteristics of work cliques, and the role of personal networks in the collaboration. In sum, networking practices lead to the third theme—the impact of CWN on the work of its members and its role in the area of water. 3.1. Barriers and incentives for collaborative work Barriers for academics Existing research shows and the interview data concur that the organizational arrangements in universities are the major barrier that discourages academics to engage in cross-sectoral and multidisciplinary research. Most of the academics admit that crosssectoral and multidisciplinary collaboration is not rewarded through the formal promotion and evaluation procedures in their universities. Even universities that “pay lip service” 7 to the importance of such research rarely provide support and recognition of it. Neither are the difficulties of such work recognized. Traditionally, the evaluation of academics is based, in addition to teaching and community service, on the number of publications and the prestige of the journals in which they publish—the “publish or perish” criteria. Multidisciplinary and cross-sectoral research is not conducive for these traditional outcomes. Working with partners does not necessarily generate opportunities for scholarly publications: its outcomes might be manuals, software tools, website information—“very informative but not peer-reviewed publications”. Integrating results from empirical research across disciplines faces significant challenges: the methods, interpretation, even the scale of empirical research are different. Several researchers interviewed point to this integration of results as the major challenge they face and link it to the multidisciplinary nature of the research. When an article is produced by collaborators, publishing is less likely because multidisciplinary work is hard to assess and not many journals are willing to undertake this evaluation and thus publish such work. Finally, even when multidisciplinary work is published, such publications are not highly valued in universities. Applied journals are considered less

7

All quotations are from the interviews.

27

prestigious. These works tend to have several co-authors and some departments interpret multiple authors as inability to do independent work. The interview data therefore show that doing cross-sectoral and multidisciplinary work of the type CWN fosters is more difficult and at the same time less valued in academia. To quote the researchers themselves, “There is no real incentive” for complex collaborative work. If academics collaborate, they may be considered “tainted in some way.” Or, to put it bluntly, “individual researchers get involved in these projects at their peril.” To be sure, there is some evidence for a shift to a more positive attitude to multidisciplinary collaborative research. The establishment of new multidisciplinary programs is the best indicator for this even though multidisciplinary programs might be a result of external pressures that universities cannot ignore. In addition, two of the respondents mention that their departments or universities do not penalize or even value collaborative work although they are quick to add that this is the exception rather than the rule. Only one academic reports that continued success in partnerships carries recognition, including the intangible support and goodwill of university executives, “more latitude in decisionmaking,” and it most likely contributed to the funding for new facilities. In short, while cross-sectoral and multidisciplinary collaboration carries penalties, the rewards for it are more uncertain and intangible. The unfavourable organizational context in academia has a particularly dampening effect on the collaborative work of junior academics in the early stages of their careers. Pressures to build careers combine with the more rigorous application of traditional criteria for evaluation. Many of the senior academics volunteered comments on this issue. University criteria do not account for the value or innovativeness of research and resort to “metrics” when evaluating young academics: they count research dollars and publications. For academics with 20-25 years of experience, such criteria may not be so rigorously applied. Young academics, however, cannot expect such latitude. They often think that getting involved in multidisciplinary work is “the kiss of death.” One of the students echoed these concerns by comparing a writing a thesis based on a CWN research project to doing two Ph.D.s—one doctoral thesis satisfying university requirements and a second one meeting the needs of the partners. As a result, multidisciplinary cross-sectoral collaboration involves above all academics who have at least 20 -25 years of experience or may be close to retirement. Alternatively, collaborative projects do not include “anybody who has the pressure to build a career.” These interview data are consistent with the demographic characteristics of the CWN members who filled in the survey: about half of them are over 40 years old and have had more than 10 years of work experience. Incentives for academics Given this unfavourable organizational context and the relatively limited changes in it, one begins to wonder why academics get involved in multidisciplinary cross-sectoral research at all. The short answer is: academics, or at least a selected group of them, find intrinsic value in collaborative work, appreciate the intangible benefits of such research,

28

and look at it as an opportunity to find like-minded collaborators. Several respondents shared their interest and excitement in doing interesting research that contributes to real life outcomes. Such intellectual benefits play an important guiding role in the behaviour of academics. Some argue that CWN projects provide higher exposure compared to other agencies. In addition, CWN projects have some intangible benefits for researchers: CWN funding “is tough money to get” and might bring kudos from colleagues and university officials. Moreover, CWN funding may be a welcome increment to the funding within their own discipline. Even when they do not translate into tangible benefits, academics value the prestige such projects bring. Summary To summarize, the barriers for academics doing cross-sectoral multidisciplinary work are to a great extent structural, although any structural constraints are typically accompanied by cultural characteristics. It is the evaluation procedures in universities and publication criteria that discourage such complex collaboration. Traditional academic culture reinforces this effect. By comparison, what attracts researchers to such research are its intrinsic value and intangible benefits such as prestige. The dampening effect is particularly strong for junior researchers in the early stage of their careers. This set of barriers and incentives is consistent with the results of the citation analysis, which confirms that researchers tend to follow disciplinary boundaries when citing their other scholars. As well, it is one of the reasons for the prevalence of senior researchers in CWN. This balance in favour of senior academics may become even more pronounced in the future: as CWN develops and fine tunes its funding criteria to focus further on multidisciplinary and cross-sectoral research, it will further depart from the traditional evaluation criteria of universities that impact young academics. Paradoxically, the more successfully CWN achieves its goals to promote multidisciplinary and cross-sectoral research, the less likely it “can expect to involve young academics.” If CWN is to include junior academics, a participant suggested, its application criteria need to be more flexible and to account for the constraints faced by young academics. 3.2. Challenges on a project and strategies for overcoming them While the university barriers can prevent academics from participating in cross-sectoral and multidisciplinary research, things do not get easier once academics are committed to such research. Collaborative projects have significant coordination and communication difficulties. In the words of the respondents, they are known for their “high transaction cost” and “much frustration.” Each project has its war stories of unique difficulties and challenges. What is common for all of them, however, are the challenges that arise from the cross-sectoral and multidisciplinary nature of the projects: different work styles, the need to establish common practices, and delays. These common challenges as well as the strategies for overcoming them are the focus of this section.

29

Internal division in collaborative teams By definition, cross-sectoral multidisciplinary projects involve a mix of participants: academic and government researchers, policy and decision-makers at various levels of government and from various agencies, industry researchers and managers, as well as NGO staff. The participants, therefore, come from organizations with different priorities, timelines, funding procedures, and decision-making cycles. These constraints in turn create distinctive attitudes and work practices. Organizational and cultural differences combine to create deep divisions within collaborative projects and make their coordination difficult. The most significant divide is between academics and non-academics. Partners— especially when they are not researchers but end users—worry about the affinity of academics to pure research that may lead to impractical, hard to use deliverables. Their single most often expressed concern of partners is that academic studies sometimes sound like “tree hugging,” without real benefits. As a partner succinctly put it, there is a perception that in such studies “things are going to be studied to death and there’s going to be no real tangible benefit.” Academics, in turn, worry about the short time-lines that non-academics expect. For them, industry and governments sometimes “want things … we just can’t possibly deliver.” Project participants thus find themselves in the middle of a balancing act in which they “try to fulfill all needs.” “You never will”, an industry employee stated, “but still you try to fulfill all needs at the same time.” Alongside with this main academic/non-academic division run other divisions that are almost as prominent. Government agencies on the federal and provincial level do not work smoothly with each other. In a classical scenario for large bureaucracies, exchanging information across federal agencies or between different levels of government is slow and cumbersome. Different funding cycles between federal and provincial governments impede project coordination. On their part, local government employees are frustrated with the uncertainty that elections bring for federal government staff. For them, working on the local level has the advantage that “at least if you get people to agree to something, then it's a done deal. There are no more hoops to jump through.” Industry researchers and NGOs perceive academics as people who have trouble sharing their results and keep to themselves. The divisions among practitioners pale in comparison with divisions among academics, who split into social scientists and everyone else. As the analysis of cliques showed, in CWN—where natural sciences dominate—social scientists play a less active role in collaboration. Both sides admit social sciences are not well-understood and not well incorporated into multidisciplinary research, including CWN projects. Such divisions, running along organizational and cultural boundaries, inevitably affect the way a project functions. As a partner put it: People [in collaborative projects] come to the table with a very different background and different set of priorities and questions. I wouldn't say that's been a difficulty. It's been a challenge (Federal Government).

30

These differences loom particularly large in projects with many participants and new members who work across distance. They cannot be avoided even on projects with excellent collaborative relationships. Participants distinguish between the collegial relations and the team spirit on a project and difficulties of complex collaborative work in general. Difficulties that stem from the organizational differences are perceived as external, “bigger than the CWN [project]” and outside the control of the team. Not surprisingly, most respondents find it easiest to work and communicate within their own group: academics are most comfortable with other academics, engineers think “engineers are easy to work with,” and staff in different NGOs best understand each other’s constraints and objectives. Remarkably, federal researchers get along well and in some respects feel closer to academics than to other government staff. They sometimes function as a “translator” between academics and other government employees. This affinity between federal and academic researchers highlights the results of the clique analysis, which showed that federal employees actively collaborate with academics (Section 2.2). It is also consistent with the distribution of ties the survey found: most of the ties of the federal sector employees are directed to academics (Section 1.2; 2.2). But federal researchers are the exceptions rather than the rule. Because of their diverse background, the participants in a project have different work and communication practices. Working together requires setting up common practices. On many projects, participants invest time and effort to agree on how to format data, how to interpret results, or how to communicate. Without such informal agreements, the risks of miscommunication and incompatible results increase. For instance, in one of the projects, researchers were sending samples to a distant lab. They were anxious to get the delivery confirmed. It took awhile for their distant counterparts to appreciate their concern and establish the practice of confirming the delivery. Similarly, in several projects researchers initially monitored the work results of others more carefully until they felt confident the lab work was up to par, the data were properly formatted, and the numbers made sense. Similarly common for projects and probably more frustrating for participants are delays. Projects leads wait for some researchers to return emails or phone calls, submit timely reports or feedback on drafts, and to assist with administrative tasks. At reporting time, colleagues who kept their deadlines get frustrated “waiting for other people to get their stuff in.” Delays might stem from organizational procedures: in one project, getting data from government agencies required signing information-sharing agreements between government organizations. Bureaucracies have never been known for speedy information exchanges, especially if they cross organizational boundaries. In addition, delays might easily result from the constraints under which individual participants work. Typically, the participants in a collaborative research project have numerous other obligations. Senior academics work on up to 15 projects at a time. Partners, in turn, typically have job responsibilities that are entirely unrelated to the research project. Other obligations can take priority over the research. Related to these multiple obligations is the lack of time, the single most common complaint of all participants: senior academics, high-ranking government staff, and even students

31

complain that they are pressed for time. In short, both organizational and individual constraints affect the exchanges on a project and can lead to delays. Delays are often perceived as a lack of commitment and a personal working style. Hence, the highest praise for colleagues on the project is to say they can be relied on to do whatever is needed in the time frame that it is needed. Such people, researchers indicate, are often their long-term collaborators. In contrast, when a delay or communication problem is mentioned, this often happens between project participants that do not know each other. In other words, the problem is thus attributed to personal characteristics, whether lack of commitment or work overload, and the solution is seen as the selection of the right person. More often than not, the right person is the person one has worked with. Indeed, both academics and partners argue that working with people they know decreases the problems on a project. Many of the interviewees, who state they have never had any problems on their project, hasten to add that this is because project participants have known each other in advance. That is why academics and partners prefer to work with long term collaborators. Such preferences suggest one of the possible strategies for coordination; it is further discussed in the next section. Summary In short, cross-sectoral multidisciplinary research is difficult to coordinate due to organizational barriers and cultural differences. Federal employees are especially valuable participants because they function as translators between academics and government employees. The common challenges are lack of common practices and delays, which are exacerbated when the project participants do not know each other. Managing work on a research project thus becomes intertwined with managing ties and networking. How they cope with such challenges is the focus of the next section.

3.3. Strategies for coping: team selection and independent work The responsibility for the coordination on a project lies with the project lead and, in large projects, with the leaders of small sub-projects. To ensure the success of their project and cope with the coordination and communication problems, project leads use two main strategies. First, they recruit their team carefully: they typically rely above all on their long-term collaborators and partners. Second, they organize the work on a project in a way that increases the independence of individual researchers and decreases the need for coordination and communication. Putting together a team If coordination and communication on a project depend on how well participants can work as a team, bringing together a team becomes all important. Like the generals of Sun-tzu, project leads win their battles before they start. Team participants are selected as much for their expertise as for their reputation to deliver and their ability to work with the rest of the team. These are to a large extent intangible personality qualities that project leads try to capture in the phrases: people who “deliver” and who can be relied on, people

32

who are “easy to get along with” and “you can work with”, or even people “you can dance with.” Their ideal collaborators are team players who can also work independently. They are researchers who have the same “philosophy” as the rest of the team. Descriptions of selection criteria vary across individual project leads. What remains invariable is the careful selection based on reputation and a personality fit. A researcher flatly stated he does not work with people he does not like. It is not easy to find good collaborators. Once they find them, academics tend to work with their collaborators over and over again. Typically, researchers name a few core collaborators—sometimes only one, sometimes three or four—with whom they identify research opportunities, secure funding from partners, and put together teams. This becomes the basis for the most common strategy of selecting a project team. In virtually all projects, project leads put a team together by first inviting their long term collaborators with proven track records and reputations. For instance, one of the project leads sits on an advisory board in which members often toss around ideas for future research. After CWN called for proposals, these advisory board members formed the core of the project team. A number of respondents have worked together with team members on previous projects. Some respondents have “many years of intensive collaboration” with their team members. Beyond this closest circle of collaborators, the team grows by snowball recruitment and referrals. In other words, those invited first to a project recommend and bring in their own collaborators and partners from industry and government. In fact, team members are invited not only for their own expertise and personality, but also for the people they can bring to the team, or for their social capital. For practitioners, finding good academic collaborators is just as difficult and follows the same logic: they work with people they know and who have been recommended by others; such academics are a “known quantity.” Ironically, when the same strategy is applied by practitioners, it can work against academics. Academics remark on how difficult it is to find partners and involve them in projects. Meeting partners is not a function of everyday academic work and may require deliberate efforts such as going to the events where partners are, looking for introductions to key people, or visiting potential partners directly. Often, academics can meet people outside academia through participation in professional organizations and government advisory bodies. Some academics meet partners through their private consulting activities. Yet meeting partners does not ensure their participation in collaboration: academics need to gain the trust of partners and establish their reputation. As several researchers argue, real trust in the relationships with partners takes years to establish. It is “easier [to involve partners] the second time around” when researchers have worked with partners and established their reputation but the first project is always the most difficult. In other words, ties with partners are neither automatically accumulated with years of work nor are they easy to achieve. Academics need to make a conscious effort to network. In some disciplines or areas of research, fewer agencies are interested in—or

33

have resources to support—research. Regardless of discipline, academics successful in collaborative work must also have networking skills. This strategy for selecting a team and involving partners suggests that the experiences of senior and junior academics on a collaborative project are quite different. Position on the project, status, experience, and networks interact to shape these experiences. First, becoming part of a team is more difficult for junior academics who have not had the time to develop networks and establish a reputation. Junior researchers are rarely project leads who put together a team. Typically, they do not invite; they are invited into a project. For instance, a junior researcher recalls that the email with the invitation to participate in a project came almost “out of the blue.” Second, once on the team, junior researchers are likely to know fewer of their collaborators in advance. The interactions on a project for them are much more important—they need to develop a network, but at the same time such interactions are less likely. Summary The selection of team participants is a key to the successful coordination and communication of a project. One of the main strategies to achieve this is to bring longterm collaborators with proven track records and established reputations to the team. The single most common pattern describing how teams are selected is this reliance on personal networks. Partners follow the same strategy: they get involved with academics who they know or who have established reputations. Such practices show the importance of established reputations and well developed networks for individual academics. It is easy to see how academics with established reputations may attract their colleagues and why they participate in numerous collaborative projects. In other words, these practices provide a context for the patterns visible in the analysis of cliques (Section 3.2). At the same time, these practices also reveal the difficulties in developing ties and establishing reputations. They underscore the need of academics and partners to get support in networking. Such support is particularly important for junior academics as well as for the development of ties between academics and partners. Minimizing coordination and communication To ensure the success of their projects and minimize coordination and communication challenges, project leads have a second option: they can minimize coordination and communication on a project through the way they design and manage the project. Research projects vary along a continuum depending on how interdependent their tasks are. At one end of the continuum are independent projects, in which components, or subprojects, contribute to a common objective but do not require exchanges. They are like pieces of a jigsaw puzzle which only fit together at the end of the project. At the opposite end of the same continuum are integrated projects, in which components feed into each other. Such projects require constant exchanges and connections among the components. These two types of projects require different coordination and communication. Most of

34

the CWN projects gravitate to the independent type projects; integrated projects are considered much more difficult, “suicide” projects. In independent projects, participants work relatively independently from each other. Hence, coordination and communication are easier. The key to coordination is a clear division of labour. Typically, participants divide the work in the beginning of the project in a burst of meetings. However, throughout the project, they work relatively independently from others. Participants in large project may work with the members of their sub-project but do not contact the rest of the team. Project leads act as the hub of the project providing the link between the components. Communication is limited. On such projects, large formal meetings attended by all team members are very rare. They are usually limited to the beginning or to the final stage of the project when researchers present their findings to their partners. In some projects, such large formal meetings never happen and some team members never meet. By comparison, integrated projects require constant monitoring and ongoing communication. The participants depend on each other and need to exchange data and analysis and cannot rely on the project lead to connect them; they establish regular direct contacts. Such projects are more likely to organize face-to-face meetings for their whole team. In addition to discussing division of labour and work progress, such meetings also set up communication practices and give participants an opportunity to meet each other. Integrated projects require much more coordination; the project lead and all the rest of the team spend more time on project management and communication. In the eyes of some researchers, the time spent on setting up communication is time taken away from research. By contrast, the lack of face-to-face meetings in independent projects by no means jeopardizes project deliverables or impedes the achievement of project goals. No participant reports major difficulties due to lack of communication. However, the lack of meetings and communication in them limits networking. For those researchers who search for a professional community in CWN, this limits the benefits from the projects. With partners, less effort is needed to coordinate and communicate, and this work follows somewhat different patterns. Regardless of the type of the project, project leads make extensive efforts to get the end users involved. Such efforts are most visible in the beginning of the project when partners decide whether to get involved and when they have a chance to define project deliverables. For instance, academics visit partners to introduce their research, organize meetings with partners or attend community events. After this initial stage, practices vary. Since there are rarely formal procedures of reporting and updating partners, practices such as meeting partners, inviting them to scholarly meetings, and sending reports to them are at the discretion of the individual academics and partners. If there are no procedures for interim reports, partners are often out of the loop. They interact with students doing the field research, but academics tend to initiate contact mostly when they need something. When communication is so limited, it seems to partners that the partnerships seem to begin and end with the letter of support for the study. In contrast, in other projects end users are updated on a regular basis,

35

consulted on interim results, or even involved in the collection of data. At the end of such projects, the final reports are not simply “tossed at them” but there is an effort to facilitate the integration of project results in their work through presentations or training. As a partner put it, reports “are basically dead words on a page,” and it is in presentations that he and his colleagues learn and can use the results of a study. Projects that include such practices are highly valued by the end users. Summary In sum, in most projects participants work independently from each other or work with only a few other collaborators; meetings attended by the whole team are rare. Such projects are held together by their project leads and lateral connections between participants are few. Such practices decrease the risk that a delay and lack of commitment of a participant can jeopardize the project, and make coordination and communication easier. These practices fit well with the results of clique analysis which found very small cliques (Section 2.2). As well, they are consistent with the low connectivity of the network as a whole (Section 2.1) While independent projects are just as often successful in achieving their goals, they limit the opportunities for networking among project participants. For those project participants who rely on projects to develop their network, this cancels some of the benefits of the project. In some cases, such practices of limited communication extend to end users who are satisfied when they are more involved in the project.

3.4. Impact of CWN on the work of academics and practitioners It is now possible to examine how CWN affects the work of individual respondents and the area of water as a whole. CWN role in creating difficult ties Academics and practitioners perceive the networking role of CWN as the main impact of CWN on their work. For academics, CWN funding typically supports only a fraction of their research and it is “not much money” compared to other funding sources. It is valued for the opportunities it brings: such funding becomes “seed money” for complex collaborative research that is not supported by other funding agencies; it enables them to do different and conceptually interesting research or work with the best experts; and it can re-direct their research agenda. In other words, CWN funding expands the opportunities of researchers and facilitates their collaborative research. Similarly, CWN creates new opportunities for the partners. It gives partners “unique” data or “inside knowledge,” provides the efficiencies of sharing information and resources, or gives them leverage to get additional funds. CWN funding, as a partner aptly put it, is “opportunity money” that enables them to “pick up on opportunities.” However, it is networking rather than funding that makes CWN unique. One of the partners flatly states that:

36

If the help [of CWN] includes money, we would have found some other place to find a bit of money. But if the help is less tangible in terms of the value of networking and contacts and a place to talk about ideas, I think the network of CWN is not found in other places (Partner, NGO). To use the words of another partner, “networks are vital.” By creating new connections, CWN helps both academics and non-academics “to do the things they want to do” and achieve their goals. The CWN networking role is of primary importance in creating the most difficult type of ties: those that are cross-sectoral and multidisciplinary. Most of the ties among people grow naturally out of their interactions during everyday activities and events. The networking role of CWN, therefore, is in addition to these common mechanisms for developing ties; deliberate interventions by CWN complement the activities of individuals. The interview data suggest that the majority of the work ties of the respondents arise around work and study and to a lesser extent through introductions by others or during events. Interviewees met the majority of people they work with in the course of their work and studies. Their network members are: colleagues in the same department, colleagues from previous workplaces, former students, former supervisors, or former classmates who graduated with them. In other words, their working ties tend to arise in repeated interactions rather than at short-lived events such as conferences or workshops. By comparison, attending events create awareness of colleagues in the area, but such awareness rarely translates quickly into work ties. In fact, in many cases, respondents— especially senior ones—use such events to meet current network members and discuss ongoing work on a project rather than to develop new ties. Only targeted meetings such as the series of meetings that launched CWN have created ties in addition to creating awareness. The common way to develop ties—through everyday activities—is less effective for ties that are outside one’s immediate area of interest. Cross-sectional and multidisciplinary ties, as well as ties that span the divide between social and natural sciences, are among those more difficult to create. Organizational, sectoral and disciplinary boundaries decrease opportunities for interactions. Simply put, biology professors and engineers from the local government do not go to the same meetings. The lack of opportunities for interaction is always compounded by cultural differences; these impede communication, understanding, and the development of trust. The description of the challenges in project work in Section 3 attests to such difficulties. For instance, cross-sectoral partnerships are based on gaining the trust of partners; partners have to know that academics will be there not only to identify but also to solve problems. Social sciences best illustrate the lack of understanding across disciplines. Respondents comment that what social sciences can offer to larger research projects is not well understood. The contribution of social sciences to a project is often “token;” social scientists are included as an afterthought without really integrating their contributions in

37

the overall research. The reason is “not a conscious effort to de-value social sciences” on the part of natural scientists but, as a scientist suggested, simply due to the difficulties in communication and understanding. Many respondents comment on the difficulties of developing cross-sectoral and multidisciplinary ties and the time and effort they require. A common strategy researchers use to develop ties with government and industry is membership in government advisory boards, community associations, professional associations, or setting up consulting companies. Another strategy is attending events such as industry workshops, NGO initiatives, or community meetings. With the work of CWN, their networking becomes much more efficient. CWN also helps the creation of difficult ties of a different type: it helps academics find other individuals who are interested in cross-sectoral multidisciplinary research. In a more traditional university environment, finding such “kindred spirits” may not be easy. Academics involved in CWN perceive themselves as different from their more traditional colleagues: they “value something other than the things that the academic system rewards,” they “don’t fit,” they “want to make a change” as opposed to traditional academics who “are building a resume.” Getting involved in CWN helps them find people who are “trying very hard to do the same kinds of things” and join their own “academic clan.” Outside researchers looking for a way into CWN express similar feelings. In short, CWN helps academics find a community of similarly minded colleagues. CWN events and projects CWN affects networking among the stakeholders in the water community in two main ways: organizing events and funding projects. Due to the dearth of networking opportunities at work and while studying, targeted events play an important role for such cross-sectoral and multidisciplinary ties. CWN events add to the few opportunities available to interested researchers and practitioners. They also provide the added bonus of bringing together people with similar interests in collaborative research. One of the researchers recalls that he did not collaborate with even nearby scholars until they met at a CWN event and recognized a “kindred spirit.” A number of respondents forged new collaborative ties at the formative meetings of CWN, where the stated goal of creating ties spurred participants into action. Projects, compared to events, have a more complex and varied impact. Projects involve a two-year commitment and often meld into a second project. As a rule, they provide more opportunities for interaction and stronger ties than short networking events. By bringing together new people, projects expand the personal network of individual researchers. This is especially important for junior researchers and other less connected members of the water network. More importantly, working on a project facilitates the development of trust. A number of academics emphasize the difference between having met a colleague and having worked

38

with them. If it is a positive, shared work experience, it creates trust and a stronger bond. Alternatively, a negative experience can weaken a relationship. The experience on a project thus determines the fate of the ties and the opportunities for future collaboration. This is especially common for the experienced and well-connected senior researchers. A number of them test ties and filter out those that are unsuccessful. For instance, a senior researcher indicated that after each completed project, he drops about one-third of the participants from his future plans because of their past performance. Another stated that a problem on a project happens only once, because there won’t be a second collaborative project. It is in this process of testing and filtering out ties that researchers build a core group of collaborators and partners. The process is never quite finished and the group of collaborators is never closed; researchers continue to add and subtract collaborators, taking risks with new members. Yet once trust is developed, collaborators tend to work together across multiple projects for many years—a pattern consistent with the survey finding about the long time respondents have known their network members (the mean time is 7.9 years, see Table 1). To summarize, individual members believe that the most important contribution of CWN for their work is its impact on their professional networks. CWN helps them to develop the more difficult cross-sectoral and cross-disciplinary ties. In addition, CWN helps academics working in a more traditional academic environment to find their “academic clan” of like-minded collaborators. The findings of the clique analysis about the multidisciplinary nature of the work cliques attest to the success of CWN. The data suggest a different impact of CWN on senior and junior researchers. CWN events and projects are much more valuable networking opportunities for academics that have smaller networks and have not yet cultivated ties or established their reputation. CWN projects and events help them to create new ties. In this respect, the lack of direct connections on projects and the low number of project-wide meetings, evident in Section 3.3, are worrisome. Spending two years on a project without developing solid ties with the rest of the team members is a lost opportunity. By comparison, senior, well-connected academics mobilize existing ties to put together project teams and then use project work to test their ties. As well, junior researchers network at events, while senior researchers use events to meet colleagues they already know and work with. In short, CWN helps members to both add and cull ties to gradually develop a group of core collaborators. CWN impact on the area of water CWN has the difficult task to create connections in the area of water, where the breadth of the subject and the fragmentation of the community are enormous. According to members, creating ties “has been the one big success of CWN.” As the discussion above suggests, the impact of CWN is more complex than simply adding new connections to the area. CWN gradually filters all ties to develop a small group of like-minded researchers, “those who really want to collaborate” and who have the right mindset. To be sure, the network attracts researchers with a traditional mindset as well. Yet over the years, members have seen the creation of a critical mass of collaborators.

39

While academics tend to focus on the networking contribution of CWN, practitioners emphasize the creation of a focus for expertise: “an identifiable group to go to in order to get information and advice.” Not surprisingly, some of them would like a broader and more efficient dissemination of the results of CWN research. This role of CWN is particularly evident in the comments of several government employees who are not directly involved in the work of CWN. Provincial employees who have no budget to fund academic research and formal partnerships with CWN are nonetheless linked and interact regularly with CWN researchers. Federal employees check the CWN website regularly for news and follow its work on professional meetings. For them, CWN has gained the reputation of being “plugged into the academic world” and being the place to get the latest scientific research findings. Summary In sum, the feedback from academics and partners that describes the impact of CWN on their own work and on the area of water, points to the contribution of the agency in enabling a new type of collaborative research, creating a focus of expertise, and bringing together a core group of researchers interested in collaborative research. In other words, it points to the achievement of the stated goals of the network. The data also suggest that CWN has established its reputation even among organizations that are not partners. It has gained visibility in the area—a clear recognition of its success.

40

Part IV. Conclusions This study has examined how academics and practitioners in the area of water are connected and has created maps of who is connected to whom. Such maps can show how processes of collaboration and knowledge transfer unfold in the water network and thus point to its strengths and weaknesses. In turn, understanding the overall network and how it functions is possible only if we know who the people in this network are and in what context they engage in collaborative research. Drawing on the survey, citation analysis, qualitative interviews, and organizational documents, the study can now answer four key questions: who are the participants in the water network, how are they connected to each other, whether they cite each other thus indicating the existence of an important professional ties, and what explains the connections in the network. Profile of the participants in the water network Academics and practitioners working in the area of water are a diverse community in terms of sector, level of involvement in water issues, and disciplinary background. All these factors contribute to the fragmentation of the network and set the stage of how people connect to each other. The majority of them are mature professionals working in academia and government (Table 2). The mean age is 47.7, mean work experience is 14.8 years, and they have known the people in their network for an average of 7.9 years (Table 1). CWN members are mostly academics and, compared to the rest of the participants in the water network, are principally male, well-educated, older, with longer work experience and well established networks (Table 3). These socio-demographic characteristics of the participants in the water network work are, as discussed in the last section, rooted in their institutional context and in the way they use personal ties in collaboration. Participants in the water network do not necessarily focus on water issues in their work. About half of them (54%) work with the majority of their contacts in the area of water and have work responsibilities focused on water issues. The remaining respondents, however, work with fewer of their contacts in the area; about one in six works with only a fraction of them (Table 6). In other words, a sizable group of participants in the water network is not strongly involved in it. Differences in the level of involvement, in turn, suggest different interests and needs for knowledge on water issues among the participants in the water network. The information on disciplinary background shows that the network is dominated by engineering and natural sciences such as biology and earth sciences, while social sciences are less represented. This disciplinary composition affects how the researchers in the network connect and work with each other and what kind of research they can conduct.

41

If participants in the water network do work with each other, they tend to work within their own sector. Personal network data show that people in each sector, with the exception of federal government employees, have work ties that fold back within their own sector (Table 10). This is one of the reasons for the fragmentation of the water network. Among all groups, academics are the most inward looking: they work mostly with other academics. Federal government employees are at the other end of the continuum: they work to a much lesser degree within their own sector. Yet ties do cross sectors; people from one sector work with people from a different sector. The data show that federal government employees gravitate toward academics. In fact, the interview data suggest that researchers among them feel an affinity to academics, while federal policy makers have budgets to spend on academic research. For their part, academics tend to work with federal and industry staff; industry staff tend to work with local government; and local government employees tend to work with industry staff, provincial employees and academics. In other words, researchers and practitioners in the water network do have cross-sectoral ties necessary to support the type of research and knowledge exchanges fostered by CWN. These cross-sectoral ties provide a glimpse of the connections that comprise the network and support the processes within it. The water network The overall water network arising from the ties among the academics and practitioners in the area is sparse. This is partly because the study captures relatively strong professional relationships. In other words, academics and practitioners report few ties to the water network because, even though they may know many people in the area, they work and discuss ideas with only a few of them. At the same time, the main reasons for this low connectivity are the breadth of issues and the number and diversity of the stakeholders working in the area of water. Many people are involved in water issues and they come from different sectors, organizations and disciplines. Since these stakeholders have different interests and priorities, they are not likely to work and share ideas with many others in the water network. Hence, the participants in the water network are not densely connected to each other. Just a small number (total of 29) of academics and practitioners are well connected to the rest of the network (Table 11). These are the central participants who hold the network together. The majority of them (three-quarters) are CWN members, suggesting that CWN has either attracted such central people to the network or helped them to develop their ties. In both cases, CWN members contribute most to the connectivity of the network and CWN plays a key role in maintaining it. These central participants in the water network include two very different groups of people (Table 11). The first group consists of active networkers: they reach out to the other members of the network. These are academics and practitioners who are interested in collaborative work. They work and exchange ideas with many people working in the area. They include a sizable number of non-academics as well as people outside CWN. Among the CWN members, the active networkers are young academics building their

42

careers as well as established academics who are also looking for collaborators in the network. Among the outsiders, people who actively work with others in the network are government and NGO employees. In turn, outsiders who actively exchange ideas with others in the network are government and industry employees. The presence of outsiders in the group—people who reach out to CWN researchers—is especially interesting because it is evidence of interest in the work of CWN as well as the untapped opportunities for building partnerships and expanding CWN. In contrast, the second group of central participants in the network consists of established experts, or people who attract others. These are people who are named by others as work collaborators or consultants on innovative ideas. This group mainly includes senior academics who are actively involved in CWN (Table 11). There are no outsiders among them. These results suggest that CWN has established itself as a focus of academic expertise in the community. In addition to revealing the participants with the largest number of connections holding the water network together, the analysis of the internal structure of the network also shows the participants who most actively collaborate and exchange ideas with others. Clique analysis identifies about a dozen strongly connected groups, or cliques, of active collaborators in the network who work closely together (Table 13). A similar number of strongly connected participants exchange ideas (Table 14). There is a significant overlap of the members of the cliques: about half of the active collaborators participate in more than one clique (Table 12). As a result, the overall number of active collaborators is small (total of 25). Among them, five out of six are CWN members, confirming the important role of CWN in fostering collaboration. The academics and practitioners in the area of water work closely and exchange ideas with very few other collaborators. All the cliques are the smallest possible size—three members (Table 13 and Table 14). This further contributes to the small overall number of active collaborators in the water network. The small number of active collaborators is consistent with the small number of well connected central participants; both results point to the existence a small core of key participants in the water network. The interview data links this small number of collaborators to the preferred work practices of the participants. Collaboration and exchanging ideas tends to be multidisciplinary and cross-sectoral. Even in the small cliques, collaborators come from several disciplines and most often from several sectors. All cliques are multidisciplinary, with a heavy representation of biology—the most prominent discipline in all the collaborations—and earth/environmental sciences and engineering. Health and social sciences, in contrast, are lightly represented. More than half of the cliques are also cross-sectoral. Work cliques tend to include academics and government employees (Table 13). By comparison, cliques in which people exchange ideas draw their members from several sectors (Table 14). In other words, because exchanging ideas is a more informal tie and it can more easily cross sectoral and organizational boundaries, people consult more diverse collaborators about their ideas. This composition of the cliques, just as the cross-sectoral work ties examined

43

in the previous section, suggests that the water network supports multidisciplinary and cross-sectoral collaboration. CWN plays a central role in supporting such multidisciplinary and cross-sectoral exchanges. This is visible in the affiliation of the active collaborators in the work cliques: the majority of them are CWN members. Yet there are still important work partnerships to be built. A quarter of the work clique members are outsiders. Such outsiders are already actively collaborating with CWN members although they are not yet partners of CWN. Given the tendency of researchers and partners to work with long-term collaborators on numerous projects, as discussed in the last section, such outsiders can easily become formal partners. The existence of outsiders actively collaborating with CWN members further suggests that CWN is not the only player in the water network— other agencies are funding research and creating collaboration. The clique analysis not only identifies the internal groupings of active collaborators in the network but also shows whether these groupings are connected to each other. The analysis indicates that two of the active collaborators work and exchange ideas with people in several cliques. In other words, they are bridges that link cliques and ensure that the network does not fragment into disconnected cliques. The people in bridge positions are academics actively involved in CWN. They are neither junior researchers nor among the most senior members in the network. Junior academics have not had time to develop their professional relationships. In turn, senior academics have less time for networking, get funded more easily by various agencies, and work on many projects whose participants are not connected to CWN. The bridges are academics in mid-career who focus their research on water issues. Two weaknesses are visible in the internal structure of the water network. The first is the small number of well-connected, actively collaborating participants. This suggests that the network has a small core of active researchers and practitioners with a strong interest in collaborative research on the one hand, and a large periphery of less connected, less active participants on the other hand. This makes the network vulnerable: if only a dozen participants leave, this will destroy many connections. The key role of only two bridges connecting the cliques is a similar risk to the health of the network. The network can easily fragment into numerous small isolated groups that do not exchange ideas or work with each other. The second weakness is the limited range of disciplines in the collaborative cliques and the low representation of health, policy and social sciences. While collaboration in the water network is multidisciplinary and most often cross-sectoral, it rarely extends across the divide between natural and social sciences. Developing connections in the future can look to expanding the core of the network, targeting the less represented disciplines, and continuing to build cross-sectoral partnerships.

44

Citation practices The citation analysis complements the results of the social network data by exploring one of the most traditional professional ties among scientists. Even when people work and exchange ideas together they may not necessarily cite each other or be cited together: citation is based on perceptions of relevance, common issues, complementary work and shaped by the publication criteria of scientific journals. The ties supporting work and exchanging ideas found in the survey may not be linked to citation practices at all. Overall, the citation analysis shows that citing is a relatively weak tie and, compared to work, more closely follows disciplinary boundaries. There is a difference, however, between being cited together with colleagues (cocitation) and citing or being cited by colleagues (intercitation). All the researchers examined in citation analysis are at least cited together with colleagues in the group; at least a minimal tie does exist. Overall, CWN researchers included in the citation analysis are more often cited together by others but rarely cite each other directly. This indicates that, on the one hand, they are perceived as working in the same area and doing similar or complementary work. Their membership in CWN may enhance such perception. Further, being perceived as doing complementary or similar work indirectly confirms the opportunities for collaboration among them whether such opportunities are currently utilized or not. On the other hand, the researchers themselves do not readily integrate their colleagues’ publications into their own work. This latter pattern is due to the diverse disciplinary background of the researchers in the group coupled with the publication criteria of most current journals which encourage citation within a single discipline. Such an interpretation is confirmed by the different degree of disciplinary diversity in cocitation and intercitation practices. Cocitation and intercitation practices differ not only in frequency but also in the extent of their multidisciplinary nature. When researchers are cited together, cocited authors often come from different disciplines; the more often researchers are cocited the more diverse the disciplinary background of the cocited authors. In other words, the more often researchers are cited together with colleagues, the more readily their work is perceived as having multidisciplinary relevance. However, direct citation not only tends to be less frequent but also tends to be within a single discipline. In other words, when directly citing their colleagues, scholars tend to more closely follow disciplinary boundaries in order to meet publication criteria and to increase the likelihood of publication. The researchers who are cocited more often tend to be senior academics who participate in several CWN projects and come from disciplines such as biology that dominate the network. Participating in multiple projects gives them visibility while the dominance of their discipline makes it easier for their work to be perceived as relevant for others. In short, the results of the citation analysis are consistent with the existence of the sparsely connected but nonetheless well structured and viable network found by the social network analysis. Thus, CWN researchers are perceived as working in the same area and cocited, and this in itself links them together; however, they infrequently cite

45

each other directly. Citation as the most traditional professional tie among scholars does exist although it is quite weak. Further, the analysis confirms both the presence of multidisciplinary connections and the continuing importance of disciplinary boundaries. encouraged by the existing institutional context. Context of collaboration To understand who the researchers and practitioners in the area of water are and why they connect to each other the way they do, the analysis looks at the broader context of collaborative work. As mentioned, the majority of people working in the area of water are mature professionals. Most CWN members are senior academics. Maturity and seniority are not accidental; despite some evidence for change, evaluation procedures in universities discourage multidisciplinary and cross-sectoral collaboration. Most scholarly journals reinforce such barriers. For academics in the beginning of their careers, complex collaborative research of the type fostered by CWN is “the kiss of death.” This explains the prevalence of senior researchers in the current membership of CWN: only senior academics can afford to pursue their interest in collaborative research. The incentives for them are intangible: prestige, finding “kindred spirits,” or pursuing interesting research. As the submission criteria of CWN evolve to more closely reflect its mandate of fostering multidisciplinary and cross-sectoral collaboration, they will become less appealing for junior academics. Ironically, the more successful CWN is in achieving its mandate, the more difficult it will be to attract junior academics. Unless the university and publishing environment changes significantly, the composition of CWN membership will continue to be tilted toward senior academics. Since the analysis has suggested that senior and junior academics play different yet equally important roles in building the connectivity of the network, increasing this imbalance in favour of senior academics is not beneficial to the water network. Once collaboration starts, contextual factors shape the challenges that participants face and the strategies they use to overcome them. Collaboration across organizations, sectors, and disciplines has inherent challenges. Organizational boundaries and cultural differences combine to create deep divisions within a team and to make coordination and communication particularly difficult. Such organizational constraints are compounded by personal constraints of multiple responsibilities and lack of time. Project participants typically have multiple responsibilities and commitments. Senior academics have several research projects, active professional lives, busy schedules, and little time. Work on any of their many projects is squeezed in among other projects, teaching commitments, professional activities, and occasionally, their consulting businesses. Practitioners involved in research projects are also likely to have other commitments and work under significant time pressures. The most common problem arising from these organizational and personal constraints is delays. The most common need is setting up common work and communication practices among project participants. Academics in charge of projects use two strategies to

46

overcome these challenges: they put together a project team of people they know and structure the work in independent chunks. Like Sun-tzu warriors, they win their battle before it starts: their first strategy is to invite colleagues with proven track records to participate in projects. Selection is based as much on complementary expertise as on work style and personality. There are always newcomers, usually recommended by other team members, but the tendency is for researchers to work with a core group of long-term collaborators. Such a team increases commitment and decreases the effort to develop common practices—trust, commitment, and common practices have already been developed. Partners, when committing to a project, follow the same rule of going with the “known quantity.” This first strategy reveals the key role of personal ties for overcoming challenges and facilitating collaboration: ties facilitate team formation and decrease coordination and communication problems on a project. This use of personal networks further contributes to making the area of water the playing field of senior academics and practitioners who have many years of experience and established networks. The second strategy for overcoming challenges is dividing the work in such a way that it allows individual participants to work mostly independently and integrate their results only at the end of the project. This alleviates the risk of one or a few participants undermining the work of the whole team and decreases the effort to coordinate work. Researchers, especially on large projects, usually work independently or with a few colleagues; connections across the sub-projects are few. Very few projects work as integrated teams. Consistent with this strategy is the practice of organizing few meetings that include the whole team. Typically, it is the project lead that connects all team members and acts as the hub of the project. These practices fit well with the results of the survey which found low connectivity in the network and very small cliques of close collaborators. In some projects, such practices extend to partners: after intense communication at the initial stage of the project, there are few meetings to update them on the progress. However, the most enthusiastic praise comes from partners on projects in which researchers work as an integrated team and involve partners in their work. Academics perceive the main impact of CWN on their work as networking with the right people. Academics interested in multidisciplinary and cross-sectoral research might be hard to find in a more traditional university environment. Involvement in CWN connects like-minded people and helps academics to find their “academic clan.” The networking opportunities created by CWN are especially important for ties not easily created in the normal course of academic life: ties across sectors and across disciplines. This networking role of CWN is consistent with the central place that CWN members have in the water network. Given the differences between senior and junior academics in how well developed their networks are, as well as in the way they use their networks, the role of CWN varies. For junior academics in the beginning of their careers, CWN events and projects can extend personal ties. By comparison, senior researchers use projects to test ties and fine tune their group of core collaborators. In that respect, work practices raise some flags. Projects

47

on which team members work independently have lower benefits for members looking to expand their networks. Practitioners emphasize the role of CWN—in addition to networking—as a focus of expertise in the area and a link to academics. Even outsiders without formal partnerships with CWN turn to it when they need information. CWN helps them get “plugged into the academic world.” CWN, in other words, has established its reputation in the area of water. Clearly, the Canadian Water Network is more than a network in name only. It has been charged with the difficult task of creating connections in a diverse, inherently fragmented community. The existing institutional arrangements do not often facilitate its work. Yet the network mapped by this study has coherence and structure. Although the network is sparsely knit and the core of active participants is small, the ties are structured, clusters of close collaborators work together, ties cross disciplines and link academics and practitioners, and key people provide integrative bridges across the entire network. The Canadian Water Network plays a major role in creating the ties that hold the water network together and supporting cross-sectoral multidisciplinary research. In the process of creating connections, it has also gained visibility and has become a “brand name” in the area of water. In short, this report documents the networked nature of the Canadian Water Network while also suggesting some of the vulnerabilities of the network and the ways to strengthen it.

48

Appendix 1: Tables Table 1:

Summary Statistics of Selected Socio-Demographic, Relational and Knowledge Exchange Variables (Respondent Level: N=173)

Variables

Age (Years) Work Experience (Years) Mean Number of Ties in Respondent’s Network Mean Friendship Score of Respondent With Ties Mean Frequency of Respondent’s Contact With Tie Mean Years Respondent Has Known Tie Proportion of Ties with Whom Respondent Has Shared Innovation Ideas (%) Proportion of Ties With Whom Respondent Would Share Innovation Ideas (%) Proportion of Ties Who Would Share Innovation Ideas With Respondent (%) Proportion of Ties Respondent Asked For Advice (%) Proportion of Ties Who Asked Respondent For Advice (%) Proportion of Ties Respondent Worked With (%)

Mean*

Std. Deviation

Minimum

Maximum

Missing

47.7 14.8 11.3 1.6 3.8 7.9

10.8 10.0 9.0 0.4 0.9 4.7

23 0 1 1 1 1

75 40 51 3 6 27

23 16 0 18 17 19

150 157 173 155 156 154

62.9

34.1

0

100

0

173

18.8

28.2

0

100

1

172

48.6 56.9 47.8 62.7

38.1 34.4 32.8 31.2

0 0 0 0

100 100 100 100

0 0 0 0

173 173 173 173

Friendship Scores (1=Acquaintance; 2=Friend; 3=Close Friend) Frequency of Contact (1=Less than once a year; 2=Yearly; 3=A few times per year; 4=Monthly; 5=Weekly; 6=About Daily)

* Mean used (median excluded because of close similarity with the means)

Appendix 1 – Page 1

N

Table 2:

Frequency Distribution of Selected Socio-Demographic Variables (Respondent Level: N = 173) Freq

Valid Percent

Freq

173

100%

D:

Male Female Total

121 52 173

69.9 30.1 100.0

B:

150

20 to 29 Years 30 to 39 Years 40 to 49 Years 50 to 59 Years More than 60 Years Missing

C:

A:

Gender

Age Recoded

Work Experience

Less than 10 Years 10 to 20 Years Over 20 Years Missing

Valid Percent

173

100%

Bachelors Masters PhD Other

32 36 71 34

18.5 20.8 41.0 19.7

100%

E:

173

100%

10 25 45 51 19 23

6.7 16.7 30.0 34.0 12.7

Yes (CWN) No (Outsiders)

94 79

54.3 45.7

157

100%

F:

173

100%

55 58 44 16

35.0 36.9 28.0

Academic Federal Government Provincial Government Local Government Industry NGO Other

70 13 15 37 20 13 5

40.5 7.5 8.7 21.4 11.6 7.5 2.9

Appendix 1 – Page 2

Highest Education

CWN Affiliate

Sector

Table 3:

Percent Distribution of Disciplinary Affiliation of Respondents (Respondent Level: N = 173) Recoded Disciplines

Frequency

Percent

1.00 Animal Biology/Zoology

8

4.6

2.00 Biology/Microbiology

17

9.8

3.00 Business/Administration/Economics

11

6.4

4.00 Chemistry / GeoChemistry / Biochemistry

14

8.1

5.00 Civil, Chemical Engineering

39

22.5

6.00 Computer Sciences

2

1.2

7.00 Earth / Environmental Science/Geology/Ecology

34

19.7

8.00 Geography, Law, Social Sciences

21

12.1

9.00 Human Health, Medicine, Epidemiology

4

2.3

10.00 Plant Biology/Botany

4

2.3

11.00 Other

19

11.0

N

173

100.0

Appendix 1 – Page 3

Table 4:

Differences Between CWN Members and Outsiders by Selected Socio-Demographics for Respondents (Respondent Level: N = 173) VARIABLES

CWN

OUTSIDE

TOTAL

-14% (12) 29% (25) 41% (35) 15% (13) 100% (85)

15% (10) 20% (13) 31% (20) 25% (16) 9% (6) 100% (65)

10 25 45 51 19 150

68% (64) 32% (30) 100% (94)

72% (57) 28% (22) 100% (79)

121 52 173

11% (10) 13% (12) 67% (63) 10% (9) 100% (94)

28% (22) 30% (24) 10% (8) 32% (25) 100% (79)

32 36 71 34 173

28% (25) 40% (36) 32% (28) 100% (89)

44% (30) 32% (22) 24% (16) 100% (68)

55 58 44 157

61% (57) 11% (10) 5% (5) 7% (7) 8% (8) 7% (7) -100% (94)

16% (13) 4% (3) 13% (10) 38% (30) 15% (12) 8% (6) 6% (5) 100% (79)

70 13 15 37 20 13 5 173

A: Age 20 to 29 Years 30 to 39 Years 40 to 49 Years 50 to 59 Years More than 60 Years

B: Gender Male Female

C: Highest Level of Education Bachelors Masters Doctoral Other

D: Work Experience Less than 10 Years 10 to 20 Years More than 20 Years

E: Sector Academic Federal Government Provincial Government Local Government Industry NGO Other

Appendix 1 – Page 4

Table 5:

Tie Level Summaries of Selected Relational Variables for Respondents (Respondent Level: N=173) Freq

Valid Percent

A: Number of Ties in Respondent’s Network

173

100%

Less than 10 Ties 10 to 20 Ties More than 20 Ties

101 51 21

58.4 29.5 12.1

B: Mean Friendship Score / Category With Tie

155

100%

Acquaintance Friend/Close Friend Total Missing

123 32 155 18

79.4 20.6 100.0

C: Mean Frequency of Contact with Tie

154

100%

Less than once a year Yearly A few times per year Monthly Weekly About Daily Missing

3 19 72 45 12 5 17

1.9 12.2 46.2 28.8 7.7 3.2

D:

154

100%

Known Less than 5 Yrs Known 5 to 10 Yrs Known 10 Yrs or More Missing

52 62 40 19

33.8 40.3 26.0

E: Proportion of Ties Respondent Worked With

173

100%

Respondent works with Less than 30% of Ties Respondent works with between 30% and 70% of Ties Respondent works with over 70% of Ties

29 64 80

16.8 37.0 46.2

Years Respondent Has Known Tie

Appendix 1 – Page 5

Table 6:

Respondent Worked With Less than 30% of Ties Respondent Worked With 30 to 70% of Ties Respondent Worked With More than 70% of Ties

N

Sector of Respondents by Working Relationships

(Respondent Level: N = 173)

Academic

Federal Government

Provincial Government

Local Government

Industry

NGO

Others

Total

7% (5)

31% (4)

20% (3)

16% (6)

25% (5)

15% (2)

80% (5)

29

39% (27)

8% (1)

53% (8)

46% (17)

30% (6)

39% (5)

54% (38)

61% (8)

27% (4)

38% (14)

45% (9)

46% (6)

100% (37)

100% (20)

100% (13)

100% (70)

100% (13)

100% (15)

Appendix 1 – Page 6

64

-------

100% (5)

80

173

Table 7:

Differences Between CWN Members and Outsiders by Selected Relational Variables for Respondent (Respondent Level: N = 173) CWN

OUTSIDERS

TOTAL

A: Mean Frequency of Contact Less than a Year Yearly Few Times a Year Monthly Weekly Daily

1.2% (1) 16% (13) 55% (46) 25% (21) 2.4 % (2) 1.2% (1) 100% (84)

2.8% (2) 8% (6) 36% (26) 33% (24) 14% (10) 6% (4) 100% (72)

3 19 72 45 12 5 156

79% (66) 21% (18) 100% (84)

80% (57) 20% (14) 100% (71)

123 32 152

26% (22) 44% (37) 30% (25) 100% (84)

43% (30) 36% (25) 21% (15) 100% (89)

52 62 40 154

48% (45) 37% (35) 15% (14) 100% (94)

71% (56) 20% (16) 9% (7) 100% (79)

101 51 21 173

11% (10) 35% (33) 54% (51) 100% (94)

24% (19) 39% (31) 37% (29) 100% (103)

29 64 80 173

B: Friendship Acquaintance Friend / Close Friend

C: Years Known Known Less than 5 Years Known 5 to 10 Years Known 10 Years or More

D: Number of Ties in Network Less than 10 Ties 10 to 20 Ties More than 20 Ties

E: Proportion of Ties Respondent Worked With Respondent Worked With Less than 30% of Ties Respondent Worked With 30 to 70% of Ties Respondent Worked With More than 70% of Ties

Appendix 1 – Page 7

Table 8:

Sharing Innovative Ideas between Respondents and Network Members (Respondent Level: N=173)

Section A:

Proportion of Respondents Who Have Shared Innovation by Source

VARIABLES

CWN

OUTSIDERS

TOTAL

Respondent shared innovation with Less than 30% of Ties Respondent shared innovation with between 30% and 70% of Ties Respondent shared innovation with over 70% of Ties

13.8% (13) 24.5 % (23) 61.7% (58)

25.3% (20) 19.0% (15) 55.7% (44)

19.1% (33) 22.0% (38) 59.0 (102)

N

100% (94)

100% (79)

100% (173)

Section B:

Proportion of Network Members Respondents Would Share Innovation With

VARIABLES

CWN

OUTSIDERS

TOTAL

Respondent would share innovation with Less than 30% of Ties Respondent would share innovation with between 30% and 70% of Ties Respondent would share innovation with over 70% of Ties

75.5% (71) 14.9% (14) 9.6% (9)

70.9% (56) 13.9% (11) 15.2% (12)

127 25 21

N

100% (94)

100% (79)

100% (173)

Section C:

Proportion of Network Members Who Would Share Innovation with Respondent

VARIABLES

CWN

OUTSIDERS

TOTAL

Less than 30% of Ties Would Share Innovation with Respondent Between 30% and 70% of Ties Would Share Innovation with Respondent Over 70% of Ties Would Share Innovation with Respondent

36.2% (34) 26.6% (25) 37.2% (35)

39.2% (31) 27.8% (22) 32.9% (26)

65 47 61

N

100% (94)

100% (79)

100% (173)

Appendix 1 – Page 8

Table 9:

Percent Distribution of Ties (Tie Level: N= 1904) Frequency

Percent

A: Sector of Tie 1 2 3 4 5 6 7 8

Unknown Academic Federal Government Provincial Government Local Government Industry NGO Other

32 593 197 262 306 270 129 115

1.7 31.1 10.3 13.8 16.1 14.2 6.8 6.0

Frequency 161 961 572 210 1904

Percent 8.5 50.5 30.0 11.0 100.0

141 117 122 658 438 260 168 1904

7.4 6.1 6.4 34.6 23.0 13.7 8.8 100.0

698 1190 1888 16

36.7 62.5 99.2 .8

C: Frequency of Contact 1.00 Unknown 2.00 Less than once a year 3.00 Yearly 4.00 A few times per year 5.00 Monthly 6.00 Weekly 7.00 About Daily Total

D: Work With 0 No 1 Yes Total Missing

Percent

E: Years Tie Known

B: Friendship Strength 1.00 Unknown 2.00 Acquaintance 3.00 Friend 4.00 Close Friend Total

Frequency

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 22 23 24 25 26 28 30 33 34 35 38 40 45 50 Total System Missing Total

Appendix 1 – Page 9

9 131 152 184 151 211 131 41 72 8 180 11 34 8 10 98 29 12 14 10 84 5 1 4 21 4 2 13 1 1 5 1 2 1 2 1643 261 1904

.5 6.9 8.0 9.7 7.9 11.1 6.9 2.2 3.8 .4 9.5 .6 1.8 .4 .5 5.1 1.5 .6 .7 .5 4.4 .3 .1 .2 1.1 .2 .1 .7 .1 .1 .3 .1 .1 .1 .1 86.3 13.7 100.0

Table 10:

Sector of Respondents (Columns) By Sector of their Ties (Rows) (Tie Level: N=1904) SECTOR OF RESPONDENT Academic

Federal Government

Provincial Government

Local Government

Academic

53% (424)

26% (47)

14% (27)

Federal Government

14% (113)

19% (34)

Provincial Government

9% (75)

Local Government

Industry

NGO

Others

Total

10% (32)

18% (42)

11% (14)

18% (7)

593

10% (19)

4% (12)

5% (13)

4% (5)

3% (1)

197

16% (29)

38% (72)

11% (36)

13% (32)

10% (12)

16% (6)

262

6% (46)

10% (18)

14% (27)

42% (132)

26% (61)

13% (16)

16% (6)

306

Industry

10% (76)

8% (15)

16% (30)

17% (52)

31% (73)

18% (22)

5% (2)

270

NGO

6% (49)

5% (9)

3% (6)

6% (19)

3% (8)

22% (27)

29% (11)

129

Others

1% (11)

15% (27)

4% (7)

10% (30)

3% (8)

22% (27)

13% (5)

115

100% (794)

100% (179)

100% (188)

100% (313)

100% (237)

100% (123)

100% (38)

1872

SECTOR OF TIE

N

Appendix 1 – Page 10

Table 11:

ID #

139 225 272 146 120 122 168 232 244 106 258 143 131 141 177 164 101 126 104 129 144 156 105 128 155 226 207 224 113 Total

Centrality Scores, CWN Membership and Sector of Central Respondents CWN member

Sector

CWN Out Out CWN CWN CWN CWN Out Out CWN CWN CWN CWN CWN CWN CWN CWN CWN CWN CWN CWN CWN CWN CWN CWN Out Out Out CWN

Academic Prov gov Fed gov Academic Prov gov Academic Academic Academic Local gov NGO Academic Academic Academic Academic Academic Academic Academic Academic Academic Academic Academic NGO Local gov Fed gov Academic Industry Industry Industry Industry

Work with OutDegree

Share Innovative Ideas

InDegree

OutDegree

InDegree

5 6 5 5 4

5

3 6 5 4 4 4 4 3 3 3 3 3

4 4 3

5 6

3 3 3 4

3 6

3 10 9 7 5 5 3 3 3 3

8 8 8 4 5

4 3 3 3 4 3 3

11

14

Appendix 1 – Page 11

14

3 13

Table 12:

Clique Membership For Work And Innovation Relations By CWN Membership, Sector, Education And Discipline Of Clique Members

ID #

Number of Work Cliques

Number of Innovation Cliques

100

0

CWN member

Sector

Education

Discipline

1

CWN

Industry

Masters

0

1

CWN

Academic

PhD

Civil, Chemical, Engineering Earth / Environmental Science/Geology/Ecology

104

1

1

CWN

Academic

PhD

Animal Biology/Zoology

105

0

1

CWN

Local Gov

Bachelors

Geography, Law, Social Science

106

0

1

CWN

NGO

Masters

Geography, Law, Social Science

108

0

1

CWN

Federal Gov

PhD

1

0

Other

Local Gov

Masters

1

0

Other

Masters

1 5 1

2 7 1

CWN CWN CWN

Local Gov Provincial Gov Academic Federal Gov

Biology/Microbiology Business/Administration/Econo mics Earth / Environmental Science/Geology/Ecology

1 1 3

1 1 0

CWN CWN CWN

Academic Academic Academic

PhD PhD PhD

0

0

CWN

Local Gov

Masters

5

2

CWN

Academic

Other

2

1

CWN

Academic

PhD

0 6

1 7

Other CWN

Industry Academic

PhD PhD

0

1

CWN

NGO

PhD

0 2 1

1 2 0

CWN CWN CWN

Academic Academic Academic

PhD Other PhD

2 3

0 2

CWN Other

Academic Federal Gov

PhD Bachelors

101

209 214 120 122 128 129 131 141 142

Masters PhD Bachelors

143 144 226 146 150 155 164 168 177 272

Appendix 1 – Page 12

Animal Biology/Zoology Biology/Microbiology Animal Biology/Zoology Business/Administration/Econo mics Geography, Law, SocialScience Civil, Chemical, Engineeering Business/Administration/Econo mics HumanHealth, Medicine,Epidemiology Earth / Environmental Science/Geology/Ecology Earth / Environmental Science/Geology/Ecology Animal Biology/Zoology Business/Administration/Econo mics HumanHealth, Medicine,Epidemiology Other Biology/Microbiology Earth / Environmental Science/Geology/Ecology Civil, Chemical, Engineeering

Table 13:

Work Cliques by Sector and Discipline Combination

Work Clique Members (ID #)

Sector combination

1: 214 141 143

Local government Academic Academic

Environmental Science/Geology Civil Engineering Epidemiology

2: 141 143 177

Academic Academic Academic

Civil Engineering Epidemiology Environmental Science/Geology

3: 209 141 168

Local government Academic Academic

Economics Civil Engineering Microbiology

4: 104 122 146

Academic Academic Academic

Environmental Science/Geology Biology/Microbiology Animal Biology/Zoology

5: 120 122 144

Provincial government Academic Academic

Animal Biology/Zoology Animal Biology/Zoology Environmental Science/Geology

6: 122 144 146

Academic Academic Academic

Animal Biology/Zoology Environmental Science/Geology Animal Biology/Zoology

7: 122 131 146

Academic Academic Academic

Animal Biology/Zoology Geography Animal Biology/Zoology

8: 122 146 164

Academic Academic Academic

Animal Biology/Zoology Animal Biology/Zoology Biology/Environmental Science

9: 128 146 164

Federal government Academic Academic

Animal Biology/Zoology Animal Biology/Zoology Biology/Environmental Science

10: 129 143 272

Academic Academic Federal government

Economics Epidemiology Civil Engineering

11: 143 146 272

Academic Academic Federal government

Epidemiology Animal Biology/Zoology Civil Engineering

12: 143 177 272

Academic Academic Federal government

Epidemiology Environmental Science/Geology Civil Engineering

Appendix 1 – Page 13

Discipline combination

Table 14:

Innovation Cliques by Sector and Discipline Combination

Innovation Clique Members (ID #)

Sector combination

Discipline combination

1: 122 131 146

Academic Academic Academic

Animal Biology/Zoology Geography Animal Biology/Zoology

2: 122 144 146

Academic Academic Academic

Animal Biology/Zoology Environmental Science/Geology Animal Biology/Zoology

3: 104 122 146

Academic Academic Academic

Environmental Science/Geology Animal Biology/Zoology Animal Biology/Zoology

4: 122 146 155

Academic Academic Academic

Animal Biology/Zoology Animal Biology/Zoology Human Health/Medicine/Epidemiology

5: 122 146 164

Academic Academic Academic

Animal Biology/Zoology Animal Biology/Zoology Biology/Environmental Science

6: 108 120 122

Federal government Provincial government Academic

Biology/Microbiology Animal Biology/Zoology Animal Biology/Zoology

7: 120 122 144

Provincial government Academic Academic

Animal Biology/Zoology Animal Biology/Zoology Environmental Science/Geology

8: 100 101 226

Industry Academic Industry

Civil/Chemical Engineering Earth/Env. Science/Geology/Ecology Earth/Env. Science/Geology/Ecology

9: 105 106 150

Local government NGO NGO

Geography/Law/Social Science Geography/Law/Social Science Business/Administration/Economics

10: 128 146 164

Federal government Academic Academic

Animal Biology/Zoology Animal Biology/Zoology Biology/Environmental Science

11: 129 143 272

Academic Academic Federal government

Economics Epidemiology Civil Engineering

12: 143 146 272

Academic Academic Federal government

Epidemiology Animal Biology/Zoology Civil Engineering

Appendix 1 – Page 14

Table 15:

CWN Citation Analysis Centrality Scores

ID

Cocite Degree

101 104 121 122 124 126 500 127 501 129 130 139 140 502 141 143 144 146 149 152 161 503 162 164 166 168 504 505 177 179 191

2 4 6 3 3 3 0 3 0 1 2 3 2 7 4 2 6 2 2 2 2 2 2 11 6 8 3 8 2 10 1

Intercite OutDegree 0 2 0 5 0 1 0 7 0 0 2 1 2 4 3 0 1 2 3 1 1 4 2 6 1 2 2 2 2 3 5

Appendix 1 – Page 15

Intercite InDegree 0 3 0 3 0 1 1 3 1 0 2 2 2 3 4 2 1 3 2 2 1 0 2 7 2 4 1 3 2 2 5

Appendix 2: Figures Figure 1A: Work Relations by Sector by Bridges/Cutpoints

Legend Cutpoints = Size (Larger Nodes such as #176 and #164 are Cutpoints / Bridges) Shapes = Sector :

Academic=

;

Federal=

;

Provincial =

Appendix 2 – Page 1

; Local =

; Industry =

; NGO =

; Other =

Figure 1B: Work Relations by Source by Bridges/Cutpoints

Legend Shapes = Data Source (Squares = CWN; Circles = Other). Cutpoints = Size (Larger Nodes such as #122 and #141 are Cutpoints / Bridges)

Appendix 2 – Page 2

Figure 2A: Innovation Relations by Sector by Bridges/Cutpoints

Legend Cutpoints = Size (Larger Nodes such as #257 and #216 are Cutpoints / Bridges) Shapes = Sector :

Academic=

;

Federal=

;

Provincial =

; Local =

Appendix 2 – Page 3

; Industry =

; NGO =

; Other =

Figure 2B: Innovation Relations by Source by Bridges/Cutpoints

Legend Shapes = Data Source (Squares = CWN; Circles = Other). Cutpoints = Size (Larger Nodes such as #257 and #216 are Cutpoints / Bridges)

Appendix 2 – Page 4

Figure 3: Co-citation Map of Sub-Set of Central CWN Members

Appendix 2 – Page 5

Figure 4: Inter-Citation Map of Sub-Set of Central CWN Members

Appendix 2 – Page 6

Appendix 3: Survey Data Collection and Available Data Data Collection: The survey was launched in July 2006. The invitation to participate in the online survey was initially sent to CWN members. 8 . The majority of the respondents received a preinvite by email and a written invitation later. Depending on the available contact information, some respondents were contacted only by mail and others only by email. About a dozen mailing addresses were incorrect. In the course of the study, six new people were added to the list of respondents. In addition, two weeks after the initial launch, a letter of invitation was mailed by the Ontario Water Works Association (OWWA) to their 677 members. Some people learned about the survey during the fourth and fifth Pollution Probe workshops (Guelph on June 9, 2006 and in Moncton on October 4–6, 2006) or during the Annual CWN conference in Montreal (November 20-23, 2006). In some cases, the information about the survey was spread by word of mouth. It is impossible to estimate how many people learned about the survey. Respondents: The survey used sociometric (social network) techniques to collect information from respondents about the people they interact with on a regular basis about water-related issues and the type of relations they have with each of these ties. The average completion time varied with the number of ties a specific respondent described but, on average, the survey took about 30 minutes to complete. As with most social network studies, obtaining such detailed information is difficult. Most CWN members, while supportive of the study, are senior researchers and managers with little time to spare. Hence very few completed the survey within the first month. The research team and CWN staff worked to increase participation through reminders by email and phone. A first email reminder was sent by the researchers on August 6; a second email reminder was sent by CWN in the second week of September. CWN staff made phone calls to key respondents at the end of October. The Internet café at the CWN Annual Conference provided another opportunity to fill in the survey. The researchers made the last reminders by phone and individual emails at the end of November. As of January 13, 2007, 173 respondents had completed the survey. Among them 94 are working on CWN-funded projects and are considered members of CWN. The remaining 79 respondents are outsiders. They include 46 OWWA members and 33 respondents who are neither on the CWN nor on the OWWA lists. CWN members who completed the survey Of all the CWN members, who are those who have completed the survey? Using the CWN membership database, we can compare the gender and sector of all people who work on CWN funded projects to the gender and sector of CWN members who completed the survey. Among the survey respondents, women are over-represented—they comprise 16.3% of the CWN database and 31.9% in the survey, confirming a pattern known from previous studies—women 8

Our list of CWN members contained 424 people. It included the names from CWN database, to which we added the names of people, typically non-academics, mentioned as investigators or partners in the research proposals submitted to CWN. The contact information we found for these additional names was not always complete or accurate. In all, 380 CWN members received an invitation. The response rate for CWN members is 24.7%.

Appendix 3 – Page 1

usually have a higher response rate. As well, academics are significantly over-represented among the CWN respondents of the survey—they comprise 48.2% of the CWN databases and 60.6% among all CWN members who completed the survey. In contrast, industry participants are significantly under-represented (15.6% in the CWN databases compared to 8.5% among CWN survey respondents. Provincial and local government staff, as well as NGO staff, are also underrepresented, albeit to a lower extent. This under-representation reflects the difficulties of tracking and contacting non-academics. The sub-sample of the CWN survey respondents, therefore, deviates from the overall CWN membership. While it does include representatives of the various sectors comprising CWN, the characteristics and practices of CWN respondents in the study will be influenced by the higher proportion of academics. Distributions by Source, Gender and Sector Gender CWN database CWN survey respondents % % 347 81.8 64 68.1 Men 69 16.3 30 31.9 Women 8 1.9 N/A 424 100% 94 100% Sector 202 47.6 57 60.6 Academics 43 10.1 10 10.6 Fed gov 50 11.8 5 5.3 Prov gov 40 9.4 7 7.4 Local gov 66 15.6 8 8.5 Industry 16 3.8 7 7.4 NGO 7 1.7 0 0.0 Other 424 100.0 94 100.0

All survey respondents % 121 69.9 52 30.1 173

100%

70 13 15 37 20 13 5 173

40.5 7.5 8.7 21.4 11.6 7.5 2.9 100.0

Available Data: The survey collected both demographic data and information about the respondents’ networks. The survey data consists of two sub-datasets, the respondent level and tielevel data. Respondent Level Data: These data provide information about each of our 173 respondents. They contain: •

Socio-Demographic Information o Year of birth o Gender o Primary Affiliation (Organizations, University, etc.) o Position o Highest Degree o Discipline (in which degree is earned) o Work Experience with water issues o Whether they are familiar with CWN o Extent to which CWN has affected their work in the area of water

•

Tie-Level Summaries o Size of each respondent’s advice, work and innovation networks o Mean frequency of contact between respondent and alters

Appendix 3 – Page 2

o

Network composition: proportion of network members that respondents give advice to, receive advice from, work with, discuss innovations with, etc.

Network Data: We have the following information on each dyad (or respondent-tie link): • Ego-Alter 9 Information - Who the Respondent: o Gave advice to o Received advice from o Worked with o Discussed innovative ideas with o Would like to discuss innovation ideas with o Thinks would share innovation ideas o Would like to meet (potential network) o Would like to collaborate with (potential network) Also, we have data on: o Relational closeness between respondent-alter o Frequency of communication between respondent-alter o Relational tenure (Years Known) •

9

Alter attributes o University/Organization Affiliation o Department Affiliation o Sector (industry, government, academia, others)

The term “alter” refers to a member of a network.

Appendix 3 – Page 3

Appendix 4: Document and Interview Data Documents The initial steps in the data collection involved collecting and analyzing organizational and individual-level documents. The team analyzed 39 research proposals submitted to CWN in December of 2004 and 2006 and later funded. Additional documents analyzed include several internal CWN reports/presentations, and curricula vitae of researchers involved in CWN work. These research proposals proved to be an excellent source of information about who participates in CWN funded projects, how work is divided, how coordination is planned and implemented, how communication technology is used to further knowledge transfer, and how the expertise of individuals enhances the whole community of knowledge about water issues in Canada. The examination of the CWN research proposals enabled the researchers to expand the list of who participates in CWN projects, since many of the partners on these projects are not recorded in CWN’s database. This expanded list was later used for sending survey invitations and selecting interview respondents. Basic information about the research projects (program, amount of funding, number of researchers, number of partners and HQP, participants) was entered in an Excel file, and provided an overview of CWN funded research and enabled comparison by project. This information facilitated the selection of interview respondents. In addition, the rich background information about the research projects enabled triangulation of different sources as well as complemented the information from the interviews. The CWN internal documents were used in preparing the survey and interview protocols. Personal documents added to the interview information and expanded the data available for individual participants. Interviews Data Collection: Interviewing was done at several events involving water issues such as the CWN Researcher Retreat in November 2005, the Fourth and Fifth Pollution Probe workshops in June and October 2005, and the CWN Annual Conference in November 2006. In addition, many interviews were conducted throughout the summer and fall of 2006. Sampling: In selecting interviewees, the goal was to include participants with diverse experiences and practices in collaborative research and knowledge transfer. The sampling targeted - CWN participants from different disciplines and research areas - CWN participants from different sectors - CWN participants in different project roles - outsiders The majority of the participants (45) were selected in a two-stage procedure: 1. Selection of research projects. The projects were drawn from the three program areas of CWN. Additionally, efforts were made to select projects in which project leads are located in different geographic areas. Such selection ensured content and regional diversity.

Appendix 4 – Page 1

2. Selection of individuals involved in the same project. The main concern was to include the project lead, academic researchers, and partners. Such selection enabled comparing the experience of academics and partners and a better understanding of collaboration across sectors. Methodologically, the advantage of this strategy was the triangulation of the information from different respondents. The rest of the respondents were selected based on pragmatic considerations. Senior researchers and participants involved more closely with CWN were approached because of their knowledge and understanding of CWN’s role. The few students interviewed provided information on project practices; typically their supervisors suggested such interviews. Outsiders were contacted at events in the area of water (Pollution Probe workshops) and asked to comment on their interest in CWN and broadly on the role of partnerships in their work. Their comments provide additional insight into partnerships and opportunities for attracting new members. Interviewees: The interview dataset consists of 64 interviews, including: • Academics working on CWN funded projects - 33 • Partners from government, industry and NGOs involved on CWN funded projects 18 • Students working on CWN funded projects - 5 • Outsiders from academia (2), government (5), industry (1) and NGO (1), not currently involved in CWN projects – 9 Among all respondents, 39 people completed both the survey and the interviews. Some of these respondents have also been included in the citation analysis. For the respondents who did both the survey and the interview, quantitative data on network characteristics can be combined with qualitative contextual and background data on respondents’ experiences. Available Data: The interview data include: • Background information on the interviewees, their positions and, for academics, their research projects • Involvement in CWN events and CWN-funded projects or in water issues • Coordination and communication on research projects • The role of collaborative projects in their work • Challenges, benefits and disadvantages of collaborative work In the course of the interviewing, it became clear that many interviewees answered questions intended for key informants. Hence, the subsequent interviews with long-term and active members of CWN elicited comments about the role of CWN previously intended for key informants. • Impact of CWN on their work and in the field of water as well as recommendations for future changes.

Appendix 4 – Page 2