Understanding and control of a simple dynamic system

Abstract

. His work focuses on dynamic and distributed decision making, and man-machine interface issues. AbstractThis study examined laypeople's understanding of a simple dynamic system, expressed in reasoning and strategies used by the subjects, and how it affected performance. Participants were 15 undergraduate psychology students, 4 male and 11 female; median age was 24 years, ranging from 21 to 31 years. The subjects' task was to establish equilibrium in a simple predatorand-prey system. A task analysis was performed to identify the problem structure, the vital aspects of the task, and the ideal strategies to perform the task. The subjects' actual performance was compared to these strategies. The results revealed that, even though the task was structurally simple, it was still difficult. Much of these difficulties seemed to stem from a low ability to apply indirect reasoning and thinking in terms of discrete time steps instead of in terms of continuous time. Copyright  2003 John Wiley & Sons, Ltd. Syst. Dyn. Rev. 19, 119-137, (2003 People are inevitably confronted with the task of exercising control over dynamic systems such as company production, the economy of a country, global warming, or even interpersonal relationships. The situations at hand may be described as dynamic decision problems that require a series of related decisions, in which the situation changes both by itself and in response to actions taken. In most cases there is also a certain amount of time pressure; actions have to be taken at the right moment (Brehmer and Allard 1991).The literature on dynamic decision making has primarily focused on how people deal with feedback and delays within a system (e.g. Brehmer 1995;Brehmer and Allard 1991;Diehl and Sterman, 1995;Dörner and Preußler 1990; Sterman 1989a, b). Research has been concerned more with the subjects' achievement in terms of their ability to control the system than with their understanding of the system with which they are interacting.In the area of management there is an increasing interest in enhancing decision makers' understanding of the complex and dynamic systems they are required to control. The task is formulated as one of developing systems thinking (Senge 1990). Systems thinking skills include understanding how system behavior over time is generated by interactions between its parts,

Positive transfer and negative transfer/antilearning of problem-solving skills.

“…Finally, they detected the origin of the learning phase in CDCTs with only 12 learning trials (Experiment 1, 2) and with extended practice of 40 trials (Experiment 4). This is consistent with studies showing that self-reports, like think aloud protocols, accurately track control performance (e.g., Burns & Vollmeyer, 2002;Gonzalez et al, 2003;Gonzales & Quesada, 2003;Jensen & Brehmer, 2003). Overall, these findings are particularly hard to reconcile within an implicit learning account that claims knowledge is acquired incidentally and is unavailable to consciousness.…”

Controlling uncertainty: A review of human behavior in complex dynamic environments.

“…For example, in studies of CDC tasks, judgments of self-efficacy have used as an index of agency, and been shown to have powerful effects on causal knowledge (Jensen & Brehmer, 2003;Osman, 2008a). Osman (2008a) revealed that high judgments of selfefficacy in turn led to the successful identification of causal structures in the CDC task, and to better control performance, which in turn lead to further increases in self-efficacy.…”

Why don’t well-educated adults understand accumulation? A challenge to researchers, educators, and citizens

“…We address many commonly cited reasons for poor performance in dynamic decision-making studies and stock-flow contexts (e.g., Atkins, Wood, & Rutgers, 2002;Booth Sweeney & Sterman, 2000;Jensen & Brehmer, 2003;Sterman & Booth Sweeney, 2007). Experiments 1-4 test whether information displays (Atkins et al, 2002), the cognitive burden of required calculations (Roch, Lane, & Samuelson, 2000), inadequate motivation, unfamiliar task context, or the inability to interpret or construct graphs (Ossimitz, 2002) contribute to SF failure.…”