Academic journal article Public Administration Quarterly

Connecting the Dots: Applying Complexity Theory, Knowledge Management and Social Network Analysis to Policy Implementation

Academic journal article Public Administration Quarterly

Connecting the Dots: Applying Complexity Theory, Knowledge Management and Social Network Analysis to Policy Implementation

Article excerpt


It has been more than 25 years since the publication of Lipsky's Street Level Bureaucracy (1980), and while his work is cited repeatedly in studies of policy implementation, little theoretical work has been done to deepen our understanding of the fundamental mechanisms that lead to the emergence of behaviors resulting in implementation success or failure. This article brings together complexity theory as a language for understanding, and social network analysis (SNA) as a method for examining policy implementation, while extending the work of (Mischen, 2007) that treats policy implementation as the outcome of knowledge management.

The "Dots"

Several authors have made important connections between complexity theory and network analysis (Carroll & Burton, 2000; Costa, Rodrigues, Travieso, & Villas Boas, 2007), network analysis and policy implementation (Choi & Brower, 2006), policy implementation and knowledge management (Mischen, 2007; Sandfort, 1999), and knowledge management and complexity theory (Bardzki & Reid, 2004; McElroy, 2003; Ruggles & Little, 2000; Tasaka, 2002), but there has been no attempt to integrate all four concepts. Additionally, while complexity theory has been applied to the public sector (Elliot & Kiel, 1999; Morcol, 2002; Rhodes & MacKechnie, 2003) there are no studies that apply complexity theory specifically to policy implementation. Likewise, network analysis has been used extensively to study intraorganizational behavior, but has not been linked to the study of knowledge management in particular.

Successful knowledge management is critical to successful policy implementation. Choo (1998) describes a "knowing organization" as one that is able to successfully manage the sensemaking, knowledge creation and decision making processes of an organization. As noted by both Lipsky (1980) and Pressman and Wildavsky (1984), implementation is the outcome of a decision-making process. In a street-level bureaucracy, large numbers of frontline workers make decisions concurrently concerning clients.

While "first generation" knowledge management (KM) studies focused largely on the role of information technology, more recent KM scholars have argued that knowledge develops and is shared by complex adaptive systems (Bardzki & Reid, 2004; McElroy, 2003; Ruggles & Little, 2000; Tasaka, 2002) and should be integrated with theories of organizational learning (Easterby-Smith & Lyles, 2003).

A complex adaptive system (CAS) is one in which a large number of moderately connected and interdependent agents co-evolve when they find themselves far-from-equilibrium. For new structures and order to be created, the system must be pushed away from an equilibrium condition, otherwise changes will be temporary and as the system will revert to its stable state (Mitleton-Kelly, 2003). Through feedback loops, these agents self-organize and create behavioral "paths" within a limited space of possibilities. What emerges is a pattern of behavior that is influenced greatly by the historicity and locality of the system. In other words, agents adapt to their environments by learning, which is a social process. McElroy (2003) argues that complexity theory provides the missing theory of how cognition happens in social systems, which has been lacking from both knowledge management and organizational learning theory.

Organizations and interorganizational networks are social systems because they involve interactions between agents. Various systems--everything from neural networks to ant colonies--have been studied from a complexity theory perspective (Kauffman, 1995; Waldrop, 1992). Social networks differ from other biological networks and processes (Newman & Park, 2003) because social networks involve a dense level of historicity: humans remember the past. Therefore, the ability of individuals to adapt their behavior as a result of learning from the past allows adaptations in human systems to occur at a much quicker pace than other systems (Holland, 1995). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.