Bayesian reasoning without numbers: Are proportions more natural than probabilities?

Area of Science:

• Cognitive Psychology
• Decision Science
• Bayesian Cognition

Background:

• Bayesian reasoning involves updating beliefs with new evidence.
• Processing single-event probabilities hinders accurate Bayesian reasoning, leading to errors like prior neglect or averaging.
• Proportions (relative frequencies) are mathematically equivalent to probabilities but may align better with intuitive mental representations.

Purpose of the Study:

• To investigate whether framing Bayesian reasoning tasks using proportions versus single-event probabilities affects how individuals combine information.
• To determine if proportions facilitate more accurate Bayesian judgments compared to probabilities.

Main Methods:

• Three experiments used bar charts without numbers to assess conceptual understanding of Bayesian reasoning.
• Tasks were verbally framed in terms of either proportions or single-event probabilities.
• Participant responses were analyzed to compare information combination strategies and judgment accuracy between framing conditions.

Main Results:

• Framing did not significantly affect whether participants combined prior and likelihood information.
• Proportion framing led to more accurate Bayesian judgments compared to probability framing.
• Proportion framing reduced the prevalence of an averaging strategy, which was common with probability framing.

Conclusions:

• The representation of information (proportions vs. probabilities) significantly impacts the quality of Bayesian reasoning, not just the engagement with information.
• Proportions offer a more natural and effective mental framework for combining evidence in Bayesian tasks.
• These findings suggest that leveraging ratio sense through proportional representations can enhance cognitive strategies for probabilistic reasoning.