PIs: Prof. Dr. Arndt Bröder,
Prof. Dr. Andreas Glöckner
Involved: Dr. Marc Jekel
Institution: Universität Mannheim, Georg-August-Universität Göttingen
Abstract
A decision-maker who wants to behave ecologically rational has to adapt to various constellations of probabilistic relations in real-world environments that often provide only incomplete information, are dynamic, and are scarce in feedback. We have demonstrated people's capabilities to make good decisions by adapting to environmental structures and we have developed an extended network-based model to describe and predict behavior in situations involving such complications, the parallel-constraint-satisfaction-network model (PCS) for judgment and decision making including learning. Specifically, we have shown that people can make good choices even in situations with missing information in that they infer incomplete information adaptive to the properties of the environment. We find that people learn not only static but also dynamic probabilistic relations successfully from feedback and that they adapt to changes in the environment quicker than predicted by previous models. In extensive model-comparisons, we find that a novel learning-algorithm implemented in the PCS model can describe participants' adaptation to the environment well and better than previously suggested mechanisms (e.g., the strategy-selection-learning theory from the adaptive-toolbox framework). In contrast to alternative models, the PCS model predicts that people can even learn in situations without feedback. We currently investigate if people indeed behave in line with this prediction. Finally, based on simulations we identified learning-orders that according to PCS should lead to high vs. low performance in later decisions and we could confirm this differential prediction concerning learning-efficiency empirically. Overall, model comparisons as well as testing of unique properties provide support for the extended PCS model including learning.
Projekt-Homepage: http://coherence-based-reasoning-and-rationality.de/
Project-related Publications
Jekel, M., Glöckner, A., Bröder, A., & Maydych, V. (2014). Approximating rationality under incomplete information: Adaptive inferences for missing cue values based on cue-discrimination. Judgment and Decision Making, 9(2), 129-147.
Jekel, M., Glöckner, A., Fiedler, S., & Bröder, A. (2012). The rationality of different kinds of intuitive decision processes. Synthese, 189(1), 147-160.
Jekel, M., Fiedler, S., & Glöckner, A. (2011). Diagnostic task selection for strategy classification in judgment and decision making: Theory, validation, and implementation in R. Judgment and Decision Making, 6(8), 782-799.
More publications can be found on this page.
Involved: Dr. Marc Jekel
Institution: Universität Mannheim, Georg-August-Universität Göttingen
Abstract
A decision-maker who wants to behave ecologically rational has to adapt to various constellations of probabilistic relations in real-world environments that often provide only incomplete information, are dynamic, and are scarce in feedback. We have demonstrated people's capabilities to make good decisions by adapting to environmental structures and we have developed an extended network-based model to describe and predict behavior in situations involving such complications, the parallel-constraint-satisfaction-network model (PCS) for judgment and decision making including learning. Specifically, we have shown that people can make good choices even in situations with missing information in that they infer incomplete information adaptive to the properties of the environment. We find that people learn not only static but also dynamic probabilistic relations successfully from feedback and that they adapt to changes in the environment quicker than predicted by previous models. In extensive model-comparisons, we find that a novel learning-algorithm implemented in the PCS model can describe participants' adaptation to the environment well and better than previously suggested mechanisms (e.g., the strategy-selection-learning theory from the adaptive-toolbox framework). In contrast to alternative models, the PCS model predicts that people can even learn in situations without feedback. We currently investigate if people indeed behave in line with this prediction. Finally, based on simulations we identified learning-orders that according to PCS should lead to high vs. low performance in later decisions and we could confirm this differential prediction concerning learning-efficiency empirically. Overall, model comparisons as well as testing of unique properties provide support for the extended PCS model including learning.
Projekt-Homepage: http://coherence-based-reasoning-and-rationality.de/
Project-related Publications
Jekel, M., Glöckner, A., Bröder, A., & Maydych, V. (2014). Approximating rationality under incomplete information: Adaptive inferences for missing cue values based on cue-discrimination. Judgment and Decision Making, 9(2), 129-147.
Jekel, M., Glöckner, A., Fiedler, S., & Bröder, A. (2012). The rationality of different kinds of intuitive decision processes. Synthese, 189(1), 147-160.
Jekel, M., Fiedler, S., & Glöckner, A. (2011). Diagnostic task selection for strategy classification in judgment and decision making: Theory, validation, and implementation in R. Judgment and Decision Making, 6(8), 782-799.
More publications can be found on this page.