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), semi-
structured 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 sub-
group 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 cross-
sectoral 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 cross-
sectoral 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 non-
academic 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, non-
academic 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 cross-
disciplinary 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 cross-
disciplinary 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 cross-
sectoral 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 decision-
making,” 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 long-
term 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 sub-
projects, 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)

Std.
Variables Mean* Deviation Minimum Maximum Missing N

Age (Years) 47.7 10.8 23 75 23 150
Work Experience (Years) 14.8 10.0 0 40 16 157
Mean Number of Ties in Respondent’s Network 11.3 9.0 1 51 0 173
Mean Friendship Score of Respondent With Ties 1.6 0.4 1 3 18 155
Mean Frequency of Respondent’s Contact With Tie 3.8 0.9 1 6 17 156
Mean Years Respondent Has Known Tie 7.9 4.7 1 27 19 154
Proportion of Ties with Whom Respondent Has Shared
Innovation Ideas (%) 62.9 34.1 0 100 0 173
Proportion of Ties With Whom Respondent Would Share
Innovation Ideas (%) 18.8 28.2 0 100 1 172
Proportion of Ties Who Would Share Innovation Ideas
With Respondent (%) 48.6 38.1 0 100 0 173
Proportion of Ties Respondent Asked For Advice (%) 56.9 34.4 0 100 0 173
Proportion of Ties Who Asked Respondent For Advice (%) 47.8 32.8 0 100 0 173
Proportion of Ties Respondent Worked With (%) 62.7 31.2 0 100 0 173

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

Table 2: Frequency Distribution of Selected Socio-Demographic Variables
(Respondent Level: N = 173)

Valid Valid
Freq Freq
Percent Percent

A: Gender 173 100% D: Highest 173 100%
Education
Male 121 69.9 Bachelors 32 18.5
Female 52 30.1 Masters 36 20.8
Total 173 100.0 PhD 71 41.0
Other 34 19.7

B: Age Recoded 150 100% E: CWN Affiliate 173 100%

20 to 29 Years 10 6.7 Yes (CWN) 94 54.3
30 to 39 Years 25 16.7 No (Outsiders) 79 45.7
40 to 49 Years 45 30.0
50 to 59 Years 51 34.0
More than 60 Years 19 12.7
Missing 23

C: Work Experience 157 100% F: Sector 173 100%

Less than 10 Years 55 35.0 Academic 70 40.5
10 to 20 Years 58 36.9 Federal Government 13 7.5
Over 20 Years 44 28.0 Provincial Government 15 8.7
Missing 16 Local Government 37 21.4
Industry 20 11.6
NGO 13 7.5
Other 5 2.9


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


Table 4: Differences Between CWN Members and Outsiders by Selected Socio-Demographics
for Respondents (Respondent Level: N = 173)

VARIABLES CWN OUTSIDE TOTAL

A: Age

20 to 29 Years -- 15% (10) 10
30 to 39 Years 14% (12) 20% (13) 25
40 to 49 Years 29% (25) 31% (20) 45
50 to 59 Years 41% (35) 25% (16) 51
More than 60 Years 15% (13) 9% (6) 19
100% (85) 100% (65) 150

B: Gender
Male 68% (64) 72% (57) 121
Female 32% (30) 28% (22) 52
100% (94) 100% (79) 173

C: Highest Level of Education

Bachelors 11% (10) 28% (22) 32
Masters 13% (12) 30% (24) 36
Doctoral 67% (63) 10% (8) 71
Other 10% (9) 32% (25) 34
100% (94) 100% (79) 173

D: Work Experience

Less than 10 Years 28% (25) 44% (30) 55
10 to 20 Years 40% (36) 32% (22) 58
More than 20 Years 32% (28) 24% (16) 44
100% (89) 100% (68) 157

E: Sector

Academic 61% (57) 16% (13) 70
Federal Government 11% (10) 4% (3) 13
Provincial Government 5% (5) 13% (10) 15
Local Government 7% (7) 38% (30) 37
Industry 8% (8) 15% (12) 20
NGO 7% (7) 8% (6) 13
Other -- 6% (5) 5
100% (94) 100% (79) 173


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 101 58.4
10 to 20 Ties 51 29.5
More than 20 Ties 21 12.1

B: Mean Friendship Score / Category With Tie 155 100%

Acquaintance 123 79.4
Friend/Close Friend 32 20.6
Total 155 100.0
Missing 18

C: Mean Frequency of Contact with Tie 154 100%

Less than once a year 3 1.9
Yearly 19 12.2
A few times per year 72 46.2
Monthly 45 28.8
Weekly 12 7.7
About Daily 5 3.2
Missing 17

D: Years Respondent Has Known Tie 154 100%

Known Less than 5 Yrs 52 33.8
Known 5 to 10 Yrs 62 40.3
Known 10 Yrs or More 40 26.0
Missing 19

E: Proportion of Ties Respondent Worked With 173 100%

Respondent works with Less than 30% of Ties 29 16.8
Respondent works with between 30% and 70% of Ties 64 37.0
Respondent works with over 70% of Ties 80 46.2


Table 6: Sector of Respondents by Working Relationships (Respondent Level: N = 173)

Federal Provincial Local
Academic Industry NGO Others Total
Government Government Government

Respondent
Worked With
7% (5) 31% (4) 20% (3) 16% (6) 25% (5) 15% (2) 80% (5) 29
Less than 30%
of Ties
Respondent
Worked With 30 39% (27) 8% (1) 53% (8) 46% (17) 30% (6) 39% (5) 64
to 70% of Ties
Respondent
Worked With
54% (38) 61% (8) 27% (4) 38% (14) 45% (9) 46% (6) ------- 80
More than 70%
of Ties

100% 100% 100% 100% 100% 100% 100%
N 173
(70) (13) (15) (37) (20) (13) (5)


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 1.2% (1) 2.8% (2) 3
Yearly 16% (13) 8% (6) 19
Few Times a Year 55% (46) 36% (26) 72
Monthly 25% (21) 33% (24) 45
Weekly 2.4 % (2) 14% (10) 12
Daily 1.2% (1) 6% (4) 5
100% (84) 100% (72) 156

B: Friendship

Acquaintance 79% (66) 80% (57) 123
Friend / Close Friend 21% (18) 20% (14) 32
100% (84) 100% (71) 152

C: Years Known

Known Less than 5 Years 26% (22) 43% (30) 52
Known 5 to 10 Years 44% (37) 36% (25) 62
Known 10 Years or More 30% (25) 21% (15) 40
100% (84) 100% (89) 154

D: Number of Ties in Network
Less than 10 Ties 48% (45) 71% (56) 101
10 to 20 Ties 37% (35) 20% (16) 51
More than 20 Ties 15% (14) 9% (7) 21
100% (94) 100% (79) 173

E: Proportion of Ties Respondent Worked With

Respondent Worked With Less than 30% of Ties 11% (10) 24% (19) 29
Respondent Worked With 30 to 70% of Ties 35% (33) 39% (31) 64
Respondent Worked With More than 70% of Ties 54% (51) 37% (29) 80
100% (94) 100% (103) 173


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 13.8% (13) 25.3% (20) 19.1% (33)
Respondent shared innovation with between 30% and 70% of Ties 24.5 % (23) 19.0% (15) 22.0% (38)
Respondent shared innovation with over 70% of Ties 61.7% (58) 55.7% (44) 59.0 (102)

N 100% (94) 100% (79) 100% (173)

Section B: Proportion of Network Members Respondents Would Share Innovation With

Respondent would share innovation with Less than 30% of Ties 75.5% (71) 70.9% (56) 127
Respondent would share innovation with between 30% and 70% of Ties 14.9% (14) 13.9% (11) 25
Respondent would share innovation with over 70% of Ties 9.6% (9) 15.2% (12) 21

N 100% (94) 100% (79) 100%
(173)

Section C: Proportion of Network Members Who Would Share Innovation with Respondent

Less than 30% of Ties Would Share Innovation with Respondent 36.2% (34) 39.2% (31) 65
Between 30% and 70% of Ties Would Share Innovation with Respondent 26.6% (25) 27.8% (22) 47
Over 70% of Ties Would Share Innovation with Respondent 37.2% (35) 32.9% (26) 61

N 100% (94) 100% (79) 100%
(173)


Table 9: Percent Distribution of Ties (Tie Level: N= 1904)

Frequency Percent Frequency Percent

A: Sector of Tie E: Years Tie Known
1 Unknown 32 1.7 0 9 .5
2 Academic 593 31.1 1 131 6.9
3 Federal Government 197 10.3 2 152 8.0
4 Provincial Government 262 13.8 3 184 9.7
5 Local Government 306 16.1 4 151 7.9
6 Industry 270 14.2 5 211 11.1
7 NGO 129 6.8 6 131 6.9
8 Other 115 6.0 7 41 2.2
8 72 3.8
9 8 .4
10 180 9.5
B: Friendship Strength 11 11 .6
Frequency Percent 12 34 1.8
1.00 Unknown 161 8.5 13 8 .4
2.00 Acquaintance 961 50.5 14 10 .5
3.00 Friend 572 30.0 15 98 5.1
4.00 Close Friend 210 11.0 16 29 1.5
Total 1904 100.0 17 12 .6
18 14 .7
19 10 .5
C: Frequency of Contact 20 84 4.4
22 5 .3
1.00 Unknown 141 7.4 23 1 .1
2.00 Less than once a year 117 6.1 24 4 .2
3.00 Yearly 122 6.4 25 21 1.1
4.00 A few times per year 658 34.6 26 4 .2
5.00 Monthly 438 23.0 28 2 .1
6.00 Weekly 260 13.7 30 13 .7
7.00 About Daily 168 8.8 33 1 .1
Total 1904 100.0 34 1 .1
35 5 .3
38 1 .1
D: Work With 40 2 .1
45 1 .1
0 No 698 36.7 50 2 .1
1 Yes 1190 62.5 Total 1643 86.3
Total 1888 99.2 System Missing 261 13.7
Missing 16 .8 Total 1904 100.0


Table 10: Sector of Respondents (Columns) By Sector of their Ties (Rows)
(Tie Level: N=1904)

SECTOR OF RESPONDENT

Federal Provincial Local
Academic Industry NGO Others Total
Government Government Government

SECTOR OF TIE

Academic 53%
26% (47) 14% (27) 10% (32) 18% (42) 11% (14) 18% (7) 593
(424)

Federal 14%
19% (34) 10% (19) 4% (12) 5% (13) 4% (5) 3% (1) 197
Government (113)

Provincial
9% (75) 16% (29) 38% (72) 11% (36) 13% (32) 10% (12) 16% (6) 262
Government

Local
6% (46) 10% (18) 14% (27) 42% (132) 26% (61) 13% (16) 16% (6) 306
Government

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% 100% 100% 100% 100% 100% 100%
N 1872
(794) (179) (188) (313) (237) (123) (38)


Table 11: Centrality Scores, CWN Membership and Sector of
Central Respondents
CWN Work with Share Innovative Ideas
ID # Sector
member OutDegree InDegree OutDegree InDegree
139 CWN Academic 3
225 Out Prov gov 5
272 Out Fed gov 6 6
146 CWN Academic 5 4 5 5
120 CWN Prov gov 4 5
122 CWN Academic 4 4 4 5
168 CWN Academic 4 3 6
232 Out Academic 4 3
244 Out Local gov 3
106 CWN NGO 3 3 3
258 CWN Academic 3 3
143 CWN Academic 3 4 6
104 CWN Academic 3
129 CWN Academic 3
156 CWN NGO 3
105 CWN Local gov 3
128 CWN Fed gov 3
155 CWN Academic 3
226 Out Industry 4
207 Out Industry 3
224 Out Industry 3
113 CWN Industry 3
Total 11 14 14 13


Table 12: Clique Membership For Work And Innovation Relations By CWN
Membership, Sector, Education And Discipline Of Clique Members
Number Number of
ID # CWN
of Work Innovation Sector Education Discipline
member
Cliques Cliques
100 0 1 CWN Industry Masters Civil, Chemical, Engineering
101 Earth / Environmental
0 1 CWN Academic PhD 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 Biology/Microbiology
209 Business/Administration/Econo
1 0 Other Local Gov Masters mics
1 0 Other Local Gov Masters Science/Geology/Ecology
120 Provincial
1 2 CWN Gov Masters Animal Biology/Zoology
122 5 7 CWN Academic PhD Biology/Microbiology
128 1 1 CWN Federal Gov Bachelors Animal Biology/Zoology
1 1 CWN Academic PhD mics
131 1 1 CWN Academic PhD Geography, Law, SocialScience
141 3 0 CWN Academic PhD Civil, Chemical, Engineeering
0 0 CWN Local Gov Masters mics
143 HumanHealth,
5 2 CWN Academic Other Medicine,Epidemiology
0 1 Other Industry PhD Science/Geology/Ecology
146 6 7 CWN Academic PhD Animal Biology/Zoology
0 1 CWN NGO PhD mics
155 HumanHealth,
0 1 CWN Academic PhD Medicine,Epidemiology
164 2 2 CWN Academic Other Other
168 1 0 CWN Academic PhD Biology/Microbiology
272 3 2 Other Federal Gov Bachelors Civil, Chemical, Engineeering


Table 13: Work Cliques by Sector and Discipline Combination
Work Clique
Sector combination Discipline combination
Members (ID #)
1: 214 Local government Environmental Science/Geology
141 Academic Civil Engineering
143 Academic Epidemiology

2: 141 Academic Civil Engineering
177 Academic Environmental Science/Geology

3: 209 Local government Economics
141 Academic Civil Engineering
168 Academic Microbiology

4: 104 Academic Environmental Science/Geology
122 Academic Biology/Microbiology
146 Academic Animal Biology/Zoology

5: 120 Provincial government Animal Biology/Zoology

6: 122 Academic Animal Biology/Zoology

131 Academic Geography

164 Academic Biology/Environmental Science

9: 128 Federal government Animal Biology/Zoology

10: 129 Academic Economics
272 Federal government Civil Engineering

11: 143 Academic Epidemiology



Table 14: Innovation Cliques by Sector and Discipline Combination
Innovation
Clique Members Sector combination Discipline combination
(ID #)
131 Academic Geography


3: 104 Academic Environmental Science/Geology

155 Academic Human
Health/Medicine/Epidemiology


6: 108 Federal government Biology/Microbiology
120 Provincial government Animal Biology/Zoology

7: 120 Provincial government Animal Biology/Zoology

8: 100 Industry Civil/Chemical Engineering
101 Academic Earth/Env. Science/Geology/Ecology
226 Industry Earth/Env. Science/Geology/Ecology

9: 105 Local government Geography/Law/Social Science
106 NGO Geography/Law/Social Science
150 NGO Business/Administration/Economics

10: 128 Federal government Animal Biology/Zoology

11: 129 Academic Economics



Table 15: CWN Citation Analysis
Centrality Scores
Cocite Intercite Intercite
ID
Degree OutDegree InDegree

101 2 0 0
104 4 2 3
121 6 0 0
122 3 5 3
124 3 0 0
126 3 1 1
500 0 0 1
127 3 7 3
501 0 0 1
129 1 0 0
130 2 2 2
139 3 1 2
140 2 2 2
502 7 4 3
141 4 3 4
143 2 0 2
144 6 1 1
146 2 2 3
149 2 3 2
152 2 1 2
161 2 1 1
503 2 4 0
162 2 2 2
164 11 6 7
166 6 1 2
168 8 2 4
504 3 2 1
505 8 2 3
177 2 2 2
179 10 3 2
191 1 5 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 = ; 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)


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 = ; 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)


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


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


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 pre-
invite 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%.


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 All survey respondents
% % %
Men 347 81.8 64 68.1 121 69.9
Women 69 16.3 30 31.9 52 30.1
N/A 8 1.9
424 100% 94 100% 173 100%
Sector
Academics 202 47.6 57 60.6 70 40.5
Fed gov 43 10.1 10 10.6 13 7.5
Prov gov 50 11.8 5 5.3 15 8.7
Local gov 40 9.4 7 7.4 37 21.4
Industry 66 15.6 8 8.5 20 11.6
NGO 16 3.8 7 7.4 13 7.5
Other 7 1.7 0 0.0 5 2.9
424 100.0 94 100.0 173 100.0

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 tie-
level 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


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)

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

9
The term “alter” refers to a member of a network.


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.


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.


Collaborative Research@CWN: Who do Scientists Network with?

More Related Content

Viewers also liked (7)

Similar to Collaborative Research@CWN: Who do Scientists Network with? (20)

Collaborative Research@CWN: Who do Scientists Network with?