B06

• Input: Question, hypotheses, evidence, assumptions, indicators, uncertainties, and new observations.
• Operation: Externalize alternatives and evidence, diagnose inconsistency rather than confirming a favorite, record assumptions, define indicators, and reassess when evidence changes.
• Output: Comparative evidence matrix, confidence/rationale, indicators, and revised judgment.
• Scope conditions: Ambiguous analytic decisions with competing explanations.
• Failure conditions: Alternatives are superficial; evidence weights are invented; triggers are not monitored.
• Predicted effect: Reduced confirmation bias and earlier correction of invalid judgments.
• Measurement: Seeded-bias error, unsupported-conclusion recall, calibration, and correction latency.
• Implementation test: Independent analyst applies the matrix and responds to a seeded disconfirming observation.
• Decision rights: Analyst assesses; decision owner acts; reviewer challenges assumptions.
• Uncertainty handling: Alternatives, confidence, and missing evidence remain explicit.
• Verification: Contrarian review and indicator monitoring.
• Provenance: Evidence-to-hypothesis links and revision log.
• Implementation cost: Added analytic and review time.
• Limitations: Coverage is insufficient for Stage B. Direct: method descriptions. Inference: metrics.