Inferring conjunctive probabilities from noisy samples: evidence for the configural weighted average model

Area of Science:

• Cognitive Psychology
• Decision Science
• Probability Theory

Background:

• Everyday decision-making often involves judging the likelihood of multiple events co-occurring.
• These underlying probabilities are typically unknown and must be inferred from experience.
• Understanding the cognitive processes behind conjunctive probability judgments is crucial.

Purpose of the Study:

• To investigate how individuals judge the conjunctive probabilities of multiple events.
• To compare different cognitive models explaining these judgments using quantitative model comparison.
• To identify the most accurate model for predicting human choices in probabilistic scenarios.

Main Methods:

• Two experiments were conducted where participants chose between pairs of conjunctive events (gambles).
• Participants used limited sample information to estimate joint probabilities.
• A hierarchical Bayesian approach was used for model parameter estimation and model comparison.

Main Results:

• The configural weighted average model best described the majority of participants' choices.
• This model assumes constituent probabilities are ranked by importance, weighted, and summed.
• The configural weighted average model demonstrated superior predictive performance for participants' decisions.

Conclusions:

• The configural weighted average model offers a robust explanation for human conjunctive probability judgments.
• This cognitive modeling approach elucidates the underlying mental processes in probabilistic decision-making.
• Findings provide insights into how people estimate the likelihood of multiple events occurring together.