The Evolution of Relationships in the GRAND Network

The Evolution of Relationships in the GRAND Network

Contents

I.Introduction

Background

Purpose and data sources

II. GRAND composition changes (Appendix 1)

New members (N 63)

Composition (N 144, N 207)

New PNIs (N 13)

III. Changes in Ties: GRAND at a glance (N 83, Appendix 2)

Network Connectivity

Researchers’ Connections

IV. Changes in Cross-Boundary Ties (N 144, N 207, Appendix 3)

Cross-disciplinary Ties

Cross-provincial Ties

V. Performance (N 142, (Appendix 4)

Co-authorship networks

Effects of network position on all performance

VI. Satisfaction with GRAND (N 124, Appendix 5)

Satisfaction with GRAND

Networking opportunities

Communication procedures

VII. Conclusions

I.Introduction

Background

GRAND: The broad mandate of the Graphics, Animation and New Media (GRAND) Networks of Centres of Excellence (NCE) is research and innovation in the area of digital media. GRAND pursues both academic excellence and commercial relevance by fostering multidisciplinary, multi-institutional, and country-widenetworkof researchers and stakeholders. Funding agencies and academic literature alike consider such research networks capable ofpooling diverse expertise, addressing complex research questions, coordinating research efforts, and utilizing research facilities across the country. In addition, the active participation of industry and government stakeholders is expected to ensure the social significance and commercial viability of the research program.

Compared to other research networks, GRAND employs a unique approach to fostering synergies among its participants. Distinctive for GRAND’s approach are three components. First, GRAND leaders have taken a deliberate approach to fully utilize the networking advantages of the network.The design principles of the networkencourage extensive linkages across projects, disciplines, and institutions. For instance, as part of the terms of their participation and research funding, each member is encouraged to work in several research projects; each project must include researchers from at least three universities, several disciplines, and partners outside academia. Such requirementsare expected to integrate separate projects in a unified research program, maximize collaboration, and facilitate knowledge and technology transfer.

Second, GRAND set up a project charged with developing collaborative and management infrastructure of the network - the Media Enabled Organizational Workflow (MEOW).The mandate of MEOW is to create web-based tools for collaboration, social networking, management and financial tracking[1]. By providing digital media tools for collaboration and management, MEOW supports emerging collaboration, fosters interactions among its members, encourages information sharing, and facilitates network management. In addition, MEOW tools compile information about the primary research and management activities of the network members that can be used for analyses.

Third, monitoring of the collaborative relationships among members in GRAND is an ongoing systematic activity conducted by a separate project - the Network Assessment and Validation for Effective Leadership (NAVEL). By collecting and analyzing data about the collaborative relationships among GRAND researchers, their practices, and their experiences, NAVEL documents key aspects of the network development and functions as a self-reflection of GRAND. Particularly important is the application of Social Network Approach (SNA) – a multidisciplinary approach used to study the structure of relationships among units, be they individuals, groups, or organizations. Such ordered relationships, or networks, support the collaborative synergies among participants and, at the same time, are the outcomes of their collaboration. The SNA analyses of NAVEL make the emerging synergies in GRAND visible and actionable; these analysescan thus be used as a potential diagnostic and evaluation tool.

Although MEOW and NAVEL have different mandates, there are synergies between the two projects: NAVEL’s analyses can reveal researchers’ experiences with the tools developed by MEOW, inform design activities and thus assists the improvement of the network infrastructure by MEOW. In turn, the information collected by MEOW in the course of research and management activities such as reporting and communication complements and enhances NAVEL’s analyses.

Purpose and data sources

We earlier reported on the state of GRAND in its early stages. Hence, the purpose of the present report is to show the evolution of the collaborative ties in GRAND. Because these collaborative ties reflect the synergies among members and indicateopportunities for knowledge sharing, their evolution can serve as a measure of the successes or failures of the network. Thisanalysis examines how GRAND changed since 2010 when the network became operational. The discussion focuses on the main changes in patterns of ties, factors impacting tie development, the implications of ties for performance, and the satisfaction of GRAND members with the network. The reportwill answerthe following questions:

• HasGRAND fostered new collaborative ties?

Are these ties over and beyond what can be expected with the mere passing of time?

• To what extent do collaborativeties cross-boundaries, such asties across disciplines, geographic locations, and the NSERC/SSHRC divide?

Theseare the ties that are especially difficult to develop and GRAND has set out to create.

• What factors affect the development of ties?

Are new ties related to the role of members, discipline, or other individual or network level characteristics?

• What network level or individual level factors affect researchers’ performance measured by their publications, artifacts, presentations, awards, number of HQPs?

Is performance associated withthe role of members (principal investigator or not), discipline, or other individual or network level characteristics?

• How satisfied GRAND researchersare with different procedures of the network? How do researchers and students perceive the benefits of GRAND?

The data for this report are based on the following sources:

• 2012 NAVEL survey dataset (N 207), response rate 60%, or 124 respondents out of 207 GRAND members registered on the GRAND forum.
• 2010 NAVEL survey dataset (N 144), response rate 70%, or 101 respondents out of 144 GRAND members listed in GRAND administrative records.
• A third dataset (N 83) is derived from the 2012 and 2010 surveys and includes only researchers who completed both surveys.
• Performance dataset (N 142), compiled from information on publications, artifacts, presentations etc., recorded by researchers on the GRAND forum in July 2013. We use the 2010 forum data and 2012 forum data on publications or the pooled data for all the yearsin order to make co-authorship ties data comparable to the survey data which was collected at the end of 2012 and early 2013.
• Interviews with researchers and students conducted after the baseline report was submitted.

Each of the data sources provides different type of data and allows different types of analyses. The 2010 and 2012 NAVEL surveys are the major source of quantitative social networks data describing the relationships among GRAND researchers, including relationships across disciplines, geographic locations, and NSERC/SSHRC divide. In addition, the 2012 follow-up survey includes information on the satisfaction of members with different aspects of GRAND. The interviews with researchers and students show the perceived value of GRAND and the subjective experience of the members. The forum dataset is the major source of performance data containing information such as publications, artifacts, and activities. Taken together, these sources enable multifaceted analyses of GRAND evolution and, by triangulating the results, increase our confidence in the conclusions. Their advantages and caveats are discussed in subsequent sections.

This report is organized in the following way: we start with a brief description of the membership changes in GRAND, then weexamine connectivity patterns and cross-boundary ties,and finally we look for the factors affecting research performance and discussmembers’ perceptions of how GRAND functions and how it has impacted them.

II. GRAND composition changes (Appendix 1)

New members (N 63)

Between 2010 and 2012, GRAND has grown in size from 144 members to 207 members[2], a growth of 44%. The characteristics of the new members reflect the recruitment policy of GRAND – an important aspect of how the network operates. In addition, the significant influx of new members affects the composition of the network and the changes in its structure.

In terms of disciplinary background, about 40% of the growth is in two multidisciplinary areas and therefore mostly among SSHRC-funded researchers. This includes Media, Art, Design with 22% of all new members and Information, Communication & Management with19% of all new members (Appendix1, Table 1C). Over one-third of the new recruits are in Computer Sciences (32% of new members). There is a sizable group of researchers in Medicine and Health Sciences who constitute 13% of all new members.Humanities and Social Sciences, weakly represented in 2010, have had very few new members (3% and 5%, respectively) while the Engineering professions have none.

Ontarioand British Columbia, the provinces with largest number of members in 2010, generate most of the network growth. Over half of the new members come from Ontario (36, or 57%), followed by British Columbia (13, or 21%), Quebec (6, or 10%). The rest of the provinces, taken together, have only 7 new members (Appendix 1, Table 1D).

In terms of gender, academic seniority, and position in GRAND, the tendency is to recruit from a less represented group and mitigate old imbalances. Newcomers tend to be junior academics, over half of them are women, and most of them join GRAND in the role of Collaborating Network Investigators (CNIs), a lower status in GRAND than Principal Network Investigators (PNIs).Ahandful areAssociate and External researchers (Appendix 1, Table 1A, 1B, 1F).

In short, the characteristics of the new GRAND members suggest that the recruitment policy of the network has continued to pursue disciplinary, academic and gender diversity although regional diversity has been less apparent.

Composition (N 144, N 207)

How does the influx of these new members affect the composition of the network in terms of disciplines, locations, gender, or seniority? A comparison of the 2010 and 2012 membership characteristics suggest an overall trend to increasing disciplinary diversity, strengthening the presence of SSHRC- and CHIR-funded researchers, and mitigating some but not all of the imbalances in the2010 composition of GRAND(Appendix 1, Table 2).

• In 2012, computer scientists are still the largest group in GRAND. However, their predominance has decreased slightly: The proportion of researchers in Computer Science has declined by 4% compared to 2010. Three disciplinary groups strengthened their presence in the network. Medical and Health Sciences grew by 3 percentage points to comprise7% of all GRAND members. Medical and Health Sciences professionals are followed by researchers from interdisciplinary areas: Information, Communication and Management (2%) and Media, Art and Design (2%): these disciplinary groups have reached 15% and 17% of all GRAND members, respectively. Traditional SSHRC disciplines lag behind: the percentage of Humanities did not change while the percentage of Social Sciences decreased by 1%. Humanities researchers remain 3% and Social Scientists 7% of allGRAND members (Appendix 1, Table 2C).

• These disciplinary changes were associated withsome redistribution of GRAND researchers by funding agency.[3] Even though the disciplinary background of GRAND researchers does not map directly into their funding agency, the proportions of the researchers funded by SSHRC and CIHR have increased at the expense of researchers funded by NSERC (Appendix 1, Table 2D). The gains are more significant for SSHRC-funded researchers (5%), who comprise 39% of all the network members. CIHR-funded researchers, who grew by 2%, remain small but growing.

• Senior academics – Associate and Full Professors - comprise over half of GRAND members (62%) although their proportion has dropped.The number of Assistant Professors grew by 8%, mostly at the expense of Associate Professors whose proportion dropped by 7% (Appendix 1, Table 2B).

• The percentage of women has also grown by 5 percentage points since 2010. The number of male researchers is still twice higher (66%) than the number of female researchers (Appendix 1, Table 2A).

• In terms of geographic locations, regional diversity did not decrease but the distribution across provinces is less balanced. While the percentage of researchers from all other provinces is either lower or remained constant, the percentage of researchers from Ontario has increased by 6 percentage points. The provinces with the lowest numbers of researchers in 2010 (Nova Scotia, Saskatchewan, Manitoba) have recruited only a few researchers, but this has been sufficient for them to maintain their proportion in the network. By comparison, provinces with relatively sizable proportions of researchers in 2010 – British Columbia, Alberta, and Quebec - have a slightly diminished presence in the network (Appendix 1, Table 2A).

New PNIs(N 13)

Compared to 2010, in 2012 the proportion of GRAND researchers in a PNI rolehas decreased from 39% to 33% (Appendix 1, Table2). This is largely due to the influx of new members in a CNI role. The number of PNIs in the network has actually increased by 13 people. A total of ten GRAND membershave advanced from CNI to PNI role; three new researchers have joined the network in as PNIs. However, a much larger number – 60 – of researchers have joined the network as CNIs,thereby changing the balance in favour of CNIs.

The new PNIs are a diverse group in terms of disciplines, academic seniority, geographic location and gender. To a great extent they reflect the composition of the network. The largest numbers come fromComputer Sciences (5). Next are interdisciplinary areas such as Information, Communication & Management (3);Media, Arts & Design (2); and Humanities (2, Appendix 1, Table 3). The growing group of Medicine & Health Sciences professionals has also one new PNI, while Engineering and Social Sciences have none.

The new PNIs tend to come from the provinces with largest number of GRAND members: British Columbia (5) and Ontario (5), with justa single promotion among members in Alberta, Nova Scotia, and Saskatchewan. New PNIs are evenly distributed between junior (5 Assistant Professors) and senior academics (5 among Associate Professors and one Full Professor).Notably, new female PNIs (9) are double the number of new male PNIs (4).

Judging by the number of their projects, new PNIs are actively engaged in the network – all of them have more than one project. However, they tend not to have as many projects as established PNIs. While the newly promoted PNIs have 2.46 projects on average, established PNIs have 3.09 projects on average (Appendix 1, Table 4). About two-thirdsof the new PNIs (8, or 62%) have 2 projects, just a few have 3 projects (4, or31%), and only one has 4 projects. By comparison, close to half of the established PNIs have 3 projects (25 or 35%) and a similar percentage have4 or 5 projects (32, or 45%, Appendix 1, Table 4).

To summarize, GRAND has increased its disciplinary diversity and strengthenedits connections to both SSHRC and CIHR. The proportion of SSHRC-funded researchers increased because ofgrowth of the new interdisciplinary areas such as Media, Arts & Design and Information, Communication & Management, rather than the traditional SSHRC-funded disciplines such as Humanities and Social Sciences. The disciplines gaining ground in GRAND are these interdisciplinary areas as well asMedicine and Health Sciences. They increased their proportions among members and advanced new PNIs.By contrast, Social Sciences slightly decreased in proportion among members and have no new PNIs.Humanities did not increase their proportion but have advanced two new PNIs. In terms of location, the changes have increased the proportion of Ontario members and, to some extent, British Columbia.

III. Changes in Ties: GRAND at a glance(N 83, Appendix 2)

Whether naturally emerging in the course of professional interaction or deliberately designed, networks are in a constant flux.New members join the network while others leave;ongoing interactions among members potentially give rise to new relationships, dissolve old ones, or alterthe structure of ties. The changes in GRAND can be described by total number of ties in the networks (tie count); the connectivity of the network (density, or the proportion of existing ties out of all possible ties in a network of a given size); or the average number of tiesof members (degree centrality). For all three measures, the results are affected by (a)the survey missing data, and (b) network size.

There are some methodological challenges. Missing data artificially decreases the overall tie count and connectivity: For GRAND members who did not complete the survey, most ties cannotbe counted.By comparison, the growth of GRANDincreases the overall number of ties in the network but simultaneously decreases the connectivity and average number of ties in the network. As new members join GRAND, they connect with existing GRAND members thereby increasing the overall number of ties. At the same time, overall connectivity falls: New GRAND members increase the size of the network and thus the number of possible ties yet they often are less connected than long-standing members. That is why their presence decreases the overall connectivity in the network and the average number of ties for researchers.

To avoid the bias created by the larger size of GRAND in 2012 and by the presence of new members, we use a subsetof the data including 83 respondents who completed both the 2010 and 2012 surveys. This dataset reflects the pattern of ties among the same long-standing GRAND members. Yet the 83 dataset has its own caveats: it ignores the ties of these 83 long-standing members with those GRAND members who did not complete either the 2010 or the 2012 survey. Further, it includes less than half of the network members in 2012; the problems created by missing data are thus more significant.

Network Connectivity

Tie counts: Among the subset of 83 long-standing members in GRAND, not all types of ties grew. Tie counts, or the pairs of connected researchers, increased for Knowing Someone (“Know”), Receiving Advice (“Receive Advice”), and Friendship (“Friend”).The baseline tie of Know grew most strongly (Appendix 2, Table 1).In contrast, Collaborative Work (“Work”)ties slightly decreased while Gave Advice remained the same.In other words, long-standing GRAND members metnew colleagues, sought advice,anddeveloped friendships with other long-standing members.However, they began the more demanding relationship of collaborative work (with fewer colleagues.