Wissen was man weiß—und was nicht: Zusammenhänge mit Entscheidungen und deren Auswirkungen in einem komplexen sozial-ökologischen System
Zusammenfassung der Projektergebnisse
The DFG research fellowship allowed me to spend 14 months at the Stockholm Resilience Centre, Stockholm, Sweden, where I worked with Prof. Maja Schlüter on the relevance of meta-cognitive information in the context of common-pool resource use. My research stay considerably extended my knowledge about, methodological skills in, research on human decisions about common-pool resources. I had proposed that metacognition, the confidence that resource users have in their knowledge of the resource regeneration, should play an important role in shaping extraction decisions. To test this assumption it was initially planned to (1) analyze existing data describing behavioral experiments where groups of resource users make repeated extraction decisions in a common-pool resource setting; and to (2) conduct cognitive modelling to computationally describe the individual-level cognitive processes involved in making these decisions. However, due to several methodological problems with the existing data, as well as the fact that my supervisor is an internationally leading expert in agent-based modeling, we jointly decided not to analyze the behavioral data. Rather, we agreed to conduct a more elaborate agent-based model that describes both individual cognitive processes, as well as interactions between group members over time. Hence, I developed AgentEx-Meta, an agent-based model that computationally describes individual and group-level processes over several resource regeneration cycles. The simulation results produced the following main finding: Groups that had access to metacognitive information arrived at more accurate collective extraction decisions compared to groups that did not have access to metacognitive information—but only when individual agents showed a myside-bias, that is, overweighted their own judgment compared to others’. This counterintuitive effect was caused by the myside-bias preserving judgment heterogeneity in the group for longer, thereby laying the foundation of groups to flexibly converge towards more accurate judgments. Without the mysidebias, in contrast, groups became entrenched in a consensus judgment prematurely. These results provide two theoretical insights. First, sharing of metacognition can deliver a social benefit for groups when solving one of the oldest collaboration problems humans are facing, the collective use of common resources. Explicit metacognition thereby provides a novel explanation for the well-known finding that communication enhances sustainable collective common-pool resource use. And second, by demonstrating how groups using biased knowledge updating can outperform the collective accuracy of groups using fully normative knowledge updating, the present results highlight how individual-level deviations from rationality can be collectively adaptive.
