From symptoms to causes: diversity effects in diagnostic reasoning

Area of Science:

• Cognitive Psychology
• Medical Reasoning
• Causal Inference

Background:

• Complex causal relationships can be modeled as branching pathways, with symptoms as terminal nodes.
• Inductive reasoning often exhibits a diversity effect, where broader evidence strengthens category induction.
• The applicability of this diversity effect to causal reasoning, particularly in a medical context, remains underexplored.

Purpose of the Study:

• To investigate whether reasoning about root causes from symptoms demonstrates a diversity effect.
• To examine if this effect holds true for medical reasoning with and without background knowledge.

Main Methods:

• Conducted three studies employing reasoning tasks based on branching causal structures.
• Participants were presented with sets of symptoms and asked to infer the most likely root cause.
• Varied the 'diversity' or separation of symptoms within the causal pathway.

Main Results:

• A significant diversity effect was observed in causal reasoning from symptoms.
• The likelihood of inducing the correct root cause increased when presented with more diverse symptoms.
• This effect persisted regardless of whether participants utilized existing background knowledge.

Conclusions:

• Reasoning about root causes from symptoms aligns with the diversity principle seen in category-based induction.
• Symptom diversity is a crucial factor in accurately identifying underlying medical conditions.
• The findings suggest a generalizable cognitive mechanism for causal inference across different domains.