Utilizing Action Research and Modeling to Inform Planning Design and Practice

“Action research” is another way to frame the role of evaluation within complex adaptive systems. French and Bell define action research as

the process of systematically collecting research data about an ongoing system relative to some objective, goal, or need of that system; feeding these data back into the system; taking actions by altering selected variables within the system based both on the data and on hypothesis; and evaluating the results of actions by collecting more data. (1999, p. 130)

Chris Argyris and Donald Schon (1995) view action research as intervention experiments through which evaluative data are used to inform organizational practices. To structure such inquiries as interventions, a certain measure of inten- tionality in the design of the action research process must be taken (Koliba and Lathrop, 2007). Action research processes are grounded in theories of organizational and experiential learning. Action research projects may employ a wide range of research methods, including all of the methodological approaches mentioned in this chapter.

Action researchers will likely collect data that may be used by a performance management system to generate greater understanding of existing practices. This systemic evaluation leads to the development of sense making, “based on the assumption that interventions purported to enhance learning in practice should focus more on bringing out people’s natural information-seeking and learning behaviors” (Parboosingh, 2002, p. 234). By integrating action research designs and processes into their performance management system, groups of individuals within a governance network will form communities of practices designed to share information, learn, and transfer and build knowledge (Fetterman, 2002; Foth, 2006; Gajda and Koliba, 2007; Koliba and Gajda, 2009). By intentionally situating action research processes within governance network operations at this level, we believe it is possible to integrate action research practices into the performance management systems of governance networks.

The key to utilizing action research and performance management in this way is to understand how consensus among stakeholders relates to the degree of understanding and application of system dynamics (and situational awareness) to current thinking. Marjan van den Belt describes the relationship between combinations of systems thinking and consensus and the kind of action research interventions possible in Table 11.3.

We believe that this typology is helpful in distinguishing the different ways that action research and modeling can be applied to the operations of governance networks. When there is little to no consensus around goals and little degree of systems thinking, there exist very little opportunities for strategic change to be under- taken—the status quo reigns. When there is goal consensus but little adaption of systems thinking and situational awareness, good discussions are possible that may or may not lead to constructive decisions and actions. Little goal consensus, but a high degree of systems dynamics yields expert-driven models of complex adaptive systems, but the interface between experts and decision makers may be limited (as they will likely be when there is little consensus on goals).

Van den Belt suggests that mediated modeling processes may be devised when goal consensus is high and there is a widely understood appreciation of systems dynamics. She describes mediated modeling in terms of a distinct, staged process

Table 11.3 Interventions: Understanding of Systems Dynamics vs. Degree of Consensus

Source: van den Belt, Mediated Modeling: A System Dynamics Approach to Environmental Consensus Building, Island Press, Washington, DC, 2004, p. 18. Reproduced by permission of Island Press.

that mirrors the cycle of inquiry discussed in the previous chapter. Modelers work with key stakeholders to determine the critical values and variables that matter most to them. Models are constructed based on these preferences. Other forms of action research found within the organizational and experiential learning literature and applied across every policy domain make the most sense when consensus is high and system thinking prevails. Van den Belt (2004) focuses her attention to the use of STELLA models that develop by using stocks and flow and feedback factors in the design, but we believe ABMs may also be employed in this manner.