VOSS:  Understanding and Enabling Network Dynamics in Virtual Communities

Project overview

To solve the most critical intellectual and social problems, teams need to be made up of the best possible people, linked to the best possible resources. While there is growing awareness of the socio-economic consequences of team collaborations, there is little socio-technical understanding of how teams are assembled or how a given mode of assembly impacts its effectiveness.

This project seeks to address this limitation by developing a theoretical framework to understand the socio-technical dynamics shaping the assembly of teams in virtual communities. These multidimensional networks include a variety of links that exist not only among individuals, but also with documents, datasets, workflows, analytic tools, and concepts. With these new configurations in mind, this project addresses two main research questions: First, what are the socio-technical motivations that explain the assembly of teams in virtual communities? Second, to what extent do the assembly mechanisms of teams influence their effectiveness? Empirically testing such models poses formidable data collection challenges. However, this project has access to six major initiatives serving diverse scientific virtual communities including nanoscience, environmental engineering, earthquake engineering, chemical sciences, media research and tobacco research.

Increasingly researchers participate in cyberinfrastructure-enabled virtual collaboration environments where they log in to portals that provide access to other users, data, analytic tools, and documents. In these cases, the digital traces reveal information about all activities a user carried out within the portal. This project is one of first few to harness the digital traces left by large scale virtual scientific collaboration supported by information technology and demonstrats the rich opportunities to capture work processes as work is done, enabling future actors to learn from past work.

The technology available in the nanoHub community offers an infrastructure to analyze this relationship’s dynamics through iterations of applications, development teams, and user needs. With a greater understanding of these team assembly strategies and user participation practices, virtual scientific communities may better model and implement applications to support productive collaborative work. The findings can guide team assembly decision-making for stakeholder communities such as individual researchers, especially students and those not in elite institutions, leaders of virtual communities, and funding agencies.

This project is uniquely positioned to draw from a variety of relevant fields such as communication, information science, organizational behavior, and social psychology to usher a new generation of theories and methods focused on explaining an important precursor to all collaborations – the socio-technical assembly mechanisms used to generate effective teams in virtual communities. Methodologically, this project will significantly extend network analytic techniques for statistically modeling high-dimensionality multimodal networks.

Research finding highlight 1:

Size, Diversity, & Beyond: Socio-technical Dynamics of Team Composition and Thresholds of Collective Action in “Team Science”

Huang, M., Huang, Y., Margolin, D., Ognyanova, K. & Shen, C., & Contractor, N.S.

While “team science”—science conducted in team ”—is known to have higher impact than solo efforts, it is also more likely to fail. Dr. Contractor’s research team has been developing a socio-technical understanding of virtual scientific team assembly: 1) How do the assembly mechanisms of teams influence their effectiveness? 2) What are the socio-technical motivations to explain team assembly in virtual communities? The following tentative findings are based on scientometric data and digital traces of 114 teams developing software tools published on nanoHub.org, a cyberinfrastructure-enabled virtual collaboration environment.

First, this study found several significant mechanisms that account for the assembly of teams and their impact on their performance. Multi-faceted effects of team size on performance: Similar to that of traditional offline teams, team size continues to be a critical factor in virtual scientific teams as manifested by its multiple significant direct or indirect effects on various facets of software development team performance. However, due to the contextual features in virtual software development teams, the impact of team size is not consistent across all measures of team performance; its indirect effect on team performance mediated by product development activity level preliminarily verified the value of the capability of larger teams to marshal more skills, knowledge and resources to meet user needs and improve product quality. Quality building vs. relationship building: This study found only two significantly positive predictors of user satisfaction: development team members’ visibility and tool development activity level. Also, user satisfaction was not significantly related to any of the other group performance measures such as tool originality and attractiveness. Therefore, one can tentatively conclude that in the current nanoHub community, user satisfaction is driven more by perceived tool developer responsiveness and trustworthiness instead of by product quality (such as originality and user retention) per se. Alternatively, relationship building with users through building prominence of developers, responsiveness, and trustworthiness seems to be most conducive to enhancing user satisfaction.

Second, to understand patterns of contribution to nanoHub, the researchers propose that, contrary to findings in offline public goods, contribution on nanoHub will be characterized by an accelerating production function in which early investment and contributions to the site generate increasing returns in terms of future contributions by others. Such a pattern has been found in the production of tags on nanoHub. Tags assigned to resources in early stages of community development were disproportionately placed by site administrators and users affiliated with site development teams. Over time, however, tags were increasingly placed by non-administrative users.  This suggests that substantial early contributions by interested parties built towards a threshold point at which other users began to contribute. These findings further suggest that long-held assumptions about collective actions and collaborative work may have to be reconsidered in the context of online collaboration.

Research finding highlight 2:

The Effects of Diversity and Repeat Collaboration on Team Performance in a Cyberinfrastructure-supported Scientific Community

As another perspective of a socio-technical understanding of virtual scientific team assembly, Dr. Contractor’s research team has been examining the effects of diversity on team performance in cyberinfrastructure-supported cross-functional teams made up of researchers from multiple institutions. This study examines the mechanisms that account for the assembly of diverse and distributed scientific teams and their impact on team performance. It presents a moderated path model integrating team affiliation diversity, experience diversity, repeat collaboration, as well as unique characteristics of virtual scientific teams supported by ICTs. The model is tested using scientometric data and digital traces of 104 teams that were developing software tools published on nanoHub.org, a cyberinfrastructure-enabled virtual scientific collaboration environment. Results indicated that team affiliation diversity had a significantly positive effect on team innovativeness, but such a positive effect was negatively moderated by repeat collaboration among team members. The study concludes by exploring ways in which team assembly and nanoHub features might be improved to foster virtual scientific collaboration.

Research finding highlight 3:

Team Formation and Performance on Nanohub: A Network Selection Challenge in Scientific Communities

This study proposes a rationale by which researchers can use network analysis tools to improve policies and guidelines for the formation of teams within a particular community.  The study begins by defining voluntary collaborative project teams (VCPT’s) as teams where members have semi-autonomous discretion over whether to join or participate.  It is argued that, in reference to VCPT’s, those making policies or guidelines appear to face both an information disadvantage and possess an information advantage relative to individual team members stemming from differing positions regarding emergent qualities of teams.  While team members have access to team-specific information, policy-makers can use information regarding norms within the community as a whole.   It is then argued that network analysis is a useful tool for uncovering these community-wide tendencies. An example analysis is then performed using exponential random graph modeling (ERGM) of a network to uncover the underlying logics of team construction within nanoHub.org, an online community that fosters collaboration amongst nanoscientists.  Results suggest that the technique is promising, as a normative tendency which undermines team performance is uncovered.

Researchers involved:
Noshir Contractor
Yun Huang

Alina Lungeanu


The VOSS project is funded by a grant from the National Science Foundation (award number IIS-0838564).

National Science Foundation