Modeling confidence in causal judgments.

Area of Science:

• Cognitive Psychology
• Causal Inference

Background:

• Counterfactual theories suggest causal judgment relies on considering alternative possibilities.
• Recent models propose probabilistic sampling of counterfactuals to assess causal strength.
• Existing models explain normality's influence on causal judgments but not confidence.

Purpose of the Study:

• To test predictions of counterfactual sampling models regarding normality's effect on causal judgment confidence.
• To investigate how statistical normality influences confidence in causal judgments.

Main Methods:

• Conducted a large-scale causal judgment task with 3,020 participants.
• Analyzed how the normality of candidate and alternate causes affected participants' confidence.
• Utilized a counterfactual sampling model to predict observed confidence levels.

Main Results:

• Normality significantly influences people's confidence in causal judgments.
• Participants were more confident in a causal relationship when the effect was less variable across counterfactual possibilities.
• Findings align with predictions from a counterfactual sampling model.

Conclusions:

• The variability of a cause's effect in counterfactual scenarios is a key determinant of causal confidence.
• Counterfactual sampling models provide a robust framework for understanding the psychological underpinnings of causal inference and confidence.