Analyzing An Interdisciplinary Research Network: The GRAND Approach Eleni Stroulia Computing Science Dept. University of Alberta Edmonton, AB T6G 2E8 stroulia@ualberta.ca Dimitrina Dimitrova, Barry Wellman Dept. of Sociology University of Toronto 725 Spadina Avenue, Toronto, ON M5S 2J4 dima@chass.utoronto.ca, wellman@chass.utoronto.ca Kellogg Booth Dept. of Computer Science 201-2366 Main Mall The University of British Columbia Vancouver, BC V6T 1Z4 ksbooth@cs.ubc.ca Adrian Sheppard GRAND NCE Centre for Digital Media 577 Great Northern Way Vancouver, BC V5T 1E1 adrian_sheppard@gnwc.ca 1. INTRODUCTION AND BACKGROUND Science is becoming networked. Scientific research today is the domain of large collaborative projects that are increasingly multi- institutional, multi-disciplinary, and multi-site (Olson et al., 2008; Cummings & Kiesler, 2005). Research projects, which are often distributed across several locations and even globally, have to rely on collaborative technology for data sharing, communication, remote access to facilities/equipment, and dissemination (Hey & Trefethen, 2008). The GRAND network of centres of excellence (NCE) includes over 850 researchers and trainees from technical disciplines, social sciences, humanities and the arts. They are collaborating within 34 interwoven projects focusing on all aspects of graphics, animation and digital media. Recognizing the complexity of managing this complex web of researchers and their relationships with each other, their multiple projects, funding agencies, and private and public partners, two of the 34 above-mentioned projects are “administrative” in nature. The objective of the first one, MEOW (“Media Enabled Organizational Workflow”), is to build a special-purpose web-based platform (the GRAND Forum) to support communication and collaboration among the network projects and researchers and to collect the data sought by network management as well as that required by the network’s funding agency. This data includes descriptive statistics about the network membership, its demographics, its productivity – in terms of publications, artifacts, activities and trainees, and its partnerships. The second administrative project, NAVEL (“Network Assessment and Validation for Effective Leadership”), uses social-network analysis methods to examine the network and its evolution over time, focusing on questions such as the types of relations that bind the network researchers, how these relations are influenced by the researchers’ disciplines, interests and activities, and how they evolve over time. Since its launch in 2010, the GRAND network has gone through two reporting-and-funding-allocations exercises supported by the Forum [1]. Driven primarily by the requirements of these managerial activities, the GRAND Forum today contains a wealth of information about the network members and their objectual relations, such as membership in the same project or in the same institution, collaboration on specific activities, co-publication, and co-supervision. In parallel, the NAVEL team using Limesurvey, an open-source survey software, has conducted a large-scale survey on the interaction among network members in concert with face-to-face and phone interviews (Dimitrova et al., 2012). The survey participants provided information about with whom, among all GRAND participants, they collaborate, exchange advice, network, and are friends with. In addition, the survey collected data about the use of communication media such as email, in-person meetings, phones, or instant messaging. While the forum and the survey collect different relational data, both describe ties among GRAND researchers and offer insight into their collaboration. This experience has led us to the following two observations: (a) there is a wealth of information collected in the Forum that could inform the survey design and eliminate effort duplication on the part of the MEOW/NAVEL researchers and their subjects, the GRAND community; and (b) currently available survey software is ill equipped to deal with the complex questions of interest to social-network analysis. Our observations are not particularly surprising: much of the information required by funding agencies as evidence of research progress and excellence refers to individuals and their relations, therefore any electronic repository of such reporting is bound to be relevant to social-network analysis. On the other hand, a quick review of the available survey software reveals that there is limited support for questions about social networks, where one needs to establish the respondent’s (egocentric) network(s) (potentially based on several information sources), collect data on the individuals and relations in this network, and further refine it through selection of special individuals and relations. To take advantage of the opportunity the network reporting data affords us and to address the challenge of insufficient software support, we are currently developing a new survey module on the GRAND Forum. This module is designed as an independent service that can be deployed independently or in the context of other data repositories. In the following section, we provide an overview of the most important requirements – unique to studying social networks (of researchers) – that the GRAND survey module will implement. 2. A SURVEY INSTRUMENT FOR RESEARCH NETWORKS We have identified the following objectives in our task of developing the GRAND survey module. Supporting the ingestion of social-network data for bootstrapping the survey: Based on our earlier work with