Diagnostic causal reasoning with verbal information.

Area of Science:

• Cognitive Psychology
• Decision Science
• Computational Modeling

Background:

• Diagnostic causal reasoning typically relies on precise quantitative information.
• Limited research exists on how qualitative, verbal expressions influence causal inferences.

Purpose of the Study:

• To investigate human diagnostic causal reasoning with qualitative, verbal information.
• To compare inferences from verbal versus numerical information presentation.
• To develop a Bayesian model for sequential diagnostic reasoning with temporal weighting.

Main Methods:

• Three experiments using a sequential diagnostic inference task with verbal information.
• Comparison of human judgments against quantitative predictions from probabilistic models.
• Development of a novel Bayesian model incorporating temporal information weighting.

Main Results:

• Inferences from verbal information closely matched those from numerical data.
• Diagnostic judgments aligned with posterior probabilities derived from prior probabilities and likelihoods.
• Individual differences in temporal evidence weighting were observed in sequential reasoning.

Conclusions:

• Qualitative verbal expressions are sufficient for accurate diagnostic causal reasoning.
• A novel Bayesian model can capture temporal dynamics in sequential judgment tasks.
• This framework facilitates computational modeling of judgment and decision-making with verbal information.