Why the outcome changes interpretation
Once harm is known, earlier signals may appear obvious. Investigators can forget that those signals coexisted with ambiguity, competing priorities or normal outcomes. Maritime and Coastguard Agency guidance describes this tendency to view past events as more predictable than they originally were.
During root cause analysis, the bias may lead to claims that someone should have known what would happen without establishing which information reached them, how it was presented or what opportunity they had to interpret it and act.
Reconstruct available information
Place observations, communications and decisions in chronological order. At each point, distinguish information recorded somewhere from information a person could access, information they actually received and facts discovered only later during the investigation process.
For example, a defect identified during a subsequent inspection cannot automatically be treated as a visible warning to an operator before the accident. Investigators need to examine indications, access, viewing conditions and communications that existed at the relevant time.
Interview without supplying the answer
Questions about what someone observed, what they expected and which alternatives they considered help reconstruct reasoning. Questions that assume the outcome was obvious may encourage defensive responses or reconstructions that do not reflect the person’s original experience.
Systemic approaches such as CAST examine responsibilities, information, context and constraints. Understanding why an action appeared reasonable at the time can reveal inadequate controls. This does not mean every decision was correct or replace any separate legal assessment that may be required.
Conclusions that support learning
A useful report separates facts, inferences and uncertainties and links recommendations to demonstrated mechanisms. Instead of merely calling for greater attention, it may identify difficult information, ambiguous alarms, conflicting objectives or insufficient resources affecting the decision.
Independent review can help detect statements of inevitability that the evidence does not support. It is also useful to examine similar decisions that ended well: the final outcome alone does not prove that a decision was safe or unsafe when it was made. Preserve conflicting accounts where they remain unresolved rather than rewriting the chronology into an artificially clear sequence.
