Predicting binary choices from probability phrase meanings

Area of Science:

• Cognitive Psychology
• Decision Science
• Probability Theory

Background:

• Understanding how individuals judge relative likelihoods is crucial for decision-making.
• Probability phrases (e.g., "likely," "unlikely") are commonly used but their subjective meanings can vary.
• Existing models often struggle to integrate the semantic understanding of probability with choice behavior.

Purpose of the Study:

• To investigate if derived scales of probability phrase meanings can predict binary choices.
• To test a memory sampling model for judging relative likelihoods.
• To determine if probability phrase meanings are measured meaningfully.

Main Methods:

• Developed scales representing the meanings of probability phrases.
• Utilized a choice model incorporating these scaled meanings.
• Assessed the model's ability to predict binary choices based on memory sampling.
• Applied sequential sampling models to refine predictions.

Main Results:

• The choice model successfully predicted outcomes for 34 out of 41 participants.
• Predictions showed a slight, consistent underestimation bias.
• Sequential sampling models significantly improved the model's predictive fit.
• Scaled probability phrase meanings served as effective proxies for memory confidence.

Conclusions:

• Derived scales of probability phrase meanings are valuable for predicting binary choices.
• The findings support a memory sampling account of relative likelihood judgments.
• The study has implications for both theoretical models of decision-making and applied contexts involving probabilistic language.