Section 2: Research Questions and Methods for Social Network Analyses

• In Event: Social Network Analysis Workshop: A Complex Systems Approach to Examining Collaboration

Abstract

The qualities of collaboration inherently draw upon complex systems concepts. The underlying structure of a collaborative group is a network, which is a defining characteristic of complex systems (Mitchell, 2009). Actors within a network can interact in dynamic ways, and the outcomes are often emergent shared goals and products. For instance, the focus of social network analysis in research and evaluation is often on publication networks, organizational characteristics, inter-organizational relationships, and other productivity measures (e.g. Honeycutt & Strong, 2012; Penuel, et al. 2007), as well as person to person relationship within human systems such as classroom or educational contexts (e.g. Part, et al., in preparation). Typical research efforts are often designed to measure anticipated outcomes at a single level (for actors or programs). While these designs may be useful for discrete outcomes; they provide limited utility for understanding socially embedded processes. In contrast, complex systems approaches using social network analyses can provide evidence for a) how ties between actors form around project activities, b) whether particular attributes of actors are related to the development of networks, or c) how patterns of dynamic change unfold and the adaptive qualities of the network structure.

At the individual level, the characteristics of the agents in a network (i.e. authorship status, gender/ethnicity, role in a collaborative group) are treated as attributes. At the group level, network configurations including sender/ receiver effects, homophily, and transitivity are operationalized as indicators of system level behavior. Data can be analyzed in many ways, such as investigating the descriptive nature of the network, or the statistical likelihood of the relationships (Lusher, et al., 2013). Research questions may address individual or network level units of analyses, dynamics within the network over time, and relations of the network to exogenous systems constraints. We discuss sampling considerations and data sources that enable SNA as a viable approach for data analyses.

Authors

• Jonathan C. Hilpert, University of Nevada Las Vegas

• Gwen C. Marchand, University of Nevada - Las Vegas