Model-Based Explanation of Feedback Effects in Syllogistic Reasoning

Area of Science:

• Cognitive Science
• Psychology
• Computational Modeling

Background:

• Human syllogistic inference has been extensively modeled.
• Model parameters are typically used for theory evaluation, not hypothesis generation.
• The effect of feedback on reasoning is a known phenomenon.

Purpose of the Study:

• To apply computational models to understand the feedback effect in syllogistic inference.
• To derive novel hypotheses from model parameters.
• To test these derived hypotheses experimentally.

Main Methods:

• Applied three state-of-the-art models: probability heuristics model (PHM), mReasoner, and TransSet.
• Utilized data from reasoning experiments with feedback.
• Conducted a new experiment to test hypotheses derived from model parameters.

Main Results:

• Replicated and confirmed the robustness of the feedback effect.
• Demonstrated the utility of model parameters for generating new research hypotheses.
• Provided novel explanations for the feedback effect grounded in existing theories.

Conclusions:

• Computational models can extend beyond their original scope to generate testable hypotheses.
• Model parameters offer a rich source for understanding cognitive phenomena.
• Feedback significantly impacts human syllogistic reasoning, with explanations derived from computational models.