Description:

This Master Prompt focuses on transforming an established RCA evidence base into a defensible causal explanation. It guides researchers from candidate causal-factor identification through cause–effect mapping, causal-structure learning, fault-propagation analysis and final root-cause localization. The central methodological principle is that RCA is fundamentally a causal problem: statistical association, anomaly magnitude, feature importance or prediction accuracy can identify suspicious variables, but these alone do not establish that a variable generated the observed failure. The selected literature explicitly distinguishes association-based approaches from causal investigation and demonstrates the value of Structural Causal Models (SCMs), causal discovery and graphical representations for recovering root-cause mechanisms from observational data.