The diversity effect in diagnostic reasoning.

Area of Science:

• Cognitive Psychology
• Decision Making
• Artificial Intelligence

Background:

• Diagnostic reasoning relies on understanding cause-effect relationships.
• The causal-diversity effect is a normative principle in diagnostic reasoning based on effect distributions.
• Psychological diversity effects may indicate the use of structured causal knowledge.

Purpose of the Study:

• To investigate the psychological causal-diversity effect in diagnostic reasoning.
• To determine if causally structured knowledge influences probability judgments of diagnoses.
• To examine the impact of causal structure and base rates on the diversity effect.

Main Methods:

• Four experiments using a quasi-medical scenario with symptom sets (effects) suggesting a diagnosis (cause).
• Participants rated diagnosis probability for proximal vs. diverse symptom sets within an acquired causal structure.
• Manipulations included causal structure, symptom position, and diagnosis base rate.

Main Results:

• The causal-diversity effect was consistently replicated across experimental conditions.
• No significant effects of causal structure or base rate variations on the diversity effect were observed.
• Computational models using causal Bayesian networks corroborated the normative justification of the diversity effect.

Conclusions:

• The observed diversity effect in diagnostic reasoning suggests participants used shortcut reasoning.
• Participants likely considered the possibilities of alternative causation when evaluating diagnoses.
• The findings highlight the interplay between cognitive heuristics and normative causal reasoning.