Probabilistic Mental Models with Continuous Predictors

Area of Science:

• Cognitive Psychology
• Decision Making
• Behavioral Economics

Background:

• Gigerenzer's work focuses on psychologically plausible models of human judgment.
• Bounded rationality and one-reason decision-making are central to these models.
• Existing Probabilistic Mental Models (PMM) theory applies to binary predictors.

Purpose of the Study:

• To extend the theory of Probabilistic Mental Models (PMM) to accommodate continuous predictors.
• To develop a new PMM model that incorporates the 7 +/- 2 category limitation for judgment tasks.
• To evaluate the predictive accuracy of the extended PMM model for binary and probability judgments.

Main Methods:

• Developed an algorithm to transform continuous predictors into a limited number of categories (7 +/- 2).
• Implemented a step-down one-reason decision procedure for the transformed predictors.
• Compared the model's predictions against binary judgments and individual probability judgments.

Main Results:

• The extended PMM model successfully predicts binary judgments, performing comparably to multiple-regression models.
• The model, like existing PMM models, does not accurately predict individual participants' probability judgments.
• The 7 +/- 2 category transformation is a key feature of the extended model.

Conclusions:

• The extended PMM model offers a psychologically plausible approach for decision-making with continuous predictors.
• While effective for binary outcomes, the model's limitations in predicting probability judgments persist.
• Further research is needed to improve the prediction of subjective probability in bounded rationality models.