Extending Small Group Theory for Analysing Complex Systems Alistair Sutcliffe Centre for HCI Design, School of Informatics, University of Manchester PO Box 88, Manchester M60 1QD, UK a.g.sutcliffe@manchester.ac.uk Abstract: This paper introduces a social psychological theory – Small Groups as Complex Systems – as a contribution to the design of CSCW and CMC systems. Small Group Theory is composed of local dynamics which model the internal view of a group; global dynamics that represent whole group emergent properties; and contextual dynamics that model the influences of the group’s environment on its composition, coherence and behaviour. The potential contribution of Small Group Theory to the design of CSCW systems is investigated by model-based analysis of group members, supporting technology, and design principles motivated by the theory. Keywords: cognitive theory, modelling framework, socio-technical systems Introduction Activity Theory (Nardi, 1996) and Distributed Cognition (Hutchins, 1995) have provided insight for informing the design of collaborative systems, but do not provide a detailed model of systems. In contrast, task modelling approaches have been extended for CSCW systems (Van der Veer, Lenting & Bergevoet, 1996), while Cognitive Work Analysis (Vicente, 1999) also provides a model-based approach for design of human activity that takes social and ecological context into account. However, there has been little convergence between task modelling and theory-driven approaches in CSCW. Instead, researchers in CSCW have evolved design principles from a combination of ethnographic study and design exploration (Abowd & Mynatt, 2000; Olson & Olson, 2000). However, such design principles focus on the technology for collaborative systems and tend to ignore the need for socio- technical solutions for collaborative systems. One of the weaknesses of applying theories to HCI, e.g. Activity Theory and Distributed Cognition, is that design recommendations are indirect, i.e. they require a theory-knowledgeable expert to provide design recommendations. For example, the influence of Activity Theory analysis in design of the exemplar case study for a Coloured Petri Net tool (Bertlesen & Bødker, 2003), while plausibly explained, is not easy to generalise to other domains and examples. Perhaps, given the complexity of phenomena at the social level, it is unrealistic to expect theory to prescribe design in complex systems. This paper introduces a social-psychological theory that does account for a wide range of phenomena and investigates its power for analysing requirements for the design of socio-technical systems. The theory of Small Groups as Complex Systems (hereafter SGACS theory: Arrow, McGrath & Berdahl, 2000), is a successor to Social Dependency Theory (McGrath, 1993). The paper is structured as follows: SGACS theory is introduced and briefly explained. The use of the theory as an analytic instrument is investigated and limitations of applying it are discussed. Principles are derived from SGACS theory that might be used to influence design of collaborative systems. The paper concludes with a brief discussion of the contributions that social-psychological theories might make to interactive systems design. Small Groups as Complex Systems Theory The SGACS theory comes from a social psychological heritage which takes an eclectic approach of theory synthesis building on research into groups as information processing systems (McGrath, 1998), bringing together 13 streams of social psychology research (Arrow et al., 2000). SGACS theory limits its scope to small groups, i.e. <=20 members. The theory contains a taxonomy of groups, a timeline view of group evolution, intra-group modelling called local dynamics, whole group modelling referred to as global dynamics, and assessment of the environment in contextual dynamics. The theory is based on a set of seven propositions that govern the influences at the local, global and context levels. The theory classifies groups into task forces (single project, short duration); teams (many projects, longer duration); crews (strong role models for collaboration); and social groups (non-work goals, member-driven collaboration); see Figure 1. A set of dimensions classifies groups according to their duration, goal directedness, and mode of creation (external setup/internal motivation). SGACS theory explains how groups of different types develop over a life cycle of formation, emergence, operation, maturity, senescence; and describes qualities for successful group interaction and pathologies in structure and behaviour that can disrupt achievement of common goals and destabilise the group.