5 Whys vs Fishbone vs Fault Tree: which RCA method fits a SIF review?

A serious injury or fatality review fails when leaders reach for one root-cause tool and treat it as a universal answer. 5 Whys, Fishbone, and Fault Tree do not solve the same problem, so the wrong choice can flatten causal depth, scatter attention, or produce a diagram that looks rigorous while the barrier failure stays hidden.

Across 25+ years leading EHS in multinationals, Andreza Araujo has seen the same pattern after high-consequence events: teams want certainty fast, so they choose the tool that is easiest to run rather than the one that matches the event. As she argues in Safety Culture: From Theory to Practice, repeated decisions reveal culture, and a shallow RCA process reveals how the organization thinks under pressure.

This comparison is for incident investigators, EHS managers, and operational leaders who need a defensible method for SIF reviews. In more than 250 cultural transformation projects supported by Andreza Araujo, the faster fix was rarely the first draft of the report. It was the moment when the team stopped asking who failed and started asking which path let the failure reach the top event.

What makes an RCA method fit for a SIF review

James Reason's Swiss cheese model is useful here because it separates active failure from latent conditions. A SIF review needs the method to answer four questions: what happened first, which barrier failed, where the causal path branches, and what evidence proves the conclusion. A tool that cannot answer those questions will still create activity, but not necessarily understanding.

The wrong method usually fails in one of three ways. It stops at a human action, it organizes ideas without proving sequence, or it produces a logic tree that no one can keep alive after the meeting. That is why method choice matters before the first workshop begins.

5 Whys: fast sequence, weak on system depth

5 Whys works best when the event has one dominant path and the team already has clean evidence from the timeline. It is quick, it is easy to teach, and it helps a supervisor move from the surface event to the first relevant management decision. Used well, it can uncover missing preconditions, weak handover, or a control that was assumed rather than verified.

Used badly, it becomes a guessing game where the facilitator keeps asking why until someone says training, attitude, or discipline. That is not analysis. It is just a polite way to stop thinking.

If the timeline itself is still uncertain, start with How to Build an Incident Timeline in the First 24 Hours before you ask for the fifth why.

Fishbone: broad structure, weak causal proof

Fishbone is strongest when the team needs to organize many hypotheses from different functions. Operations, maintenance, engineering, supervision, contractor management, and training can all contribute evidence without forcing the group into a single chain too early. The diagram helps because it makes parallel causes visible.

The weakness is that visibility is not proof. A fishbone can list ten plausible causes and still leave the team unsure which branch actually mattered. That is why Andreza Araujo's A Ilusao da Conformidade, glossed in English as The Illusion of Compliance, matters here. A tidy-looking map can hide weak field evidence when the workshop rewards completeness over verification.

Use fishbone when the team needs to widen the lens, then narrow it with evidence. The companion article Operator Blame: 5 Myths That Keep RCA Shallow helps leaders avoid the trap of converting a group brainstorm into a single-worker verdict.

Fault Tree: logic-rich for top events, heavier to run

Fault tree analysis fits top events that need barrier logic and explicit sequence. It is the strongest of the three when the organization wants to test whether a fatal path could realistically reach the event, because the tree forces the team to ask which combination of failures is necessary and which barriers still hold.

The tradeoff is effort. Fault tree analysis demands discipline, good definitions, and enough technical knowledge to keep the logic honest. A sloppy tree is worse than a simple one, because it gives the impression of rigor while hiding gaps in the cut sets.

When the event touches major hazards, barrier loss, or multiple safeguards, fault tree is often the method that deserves the extra work. That is consistent with Barrier Analysis Explained: 5 Checks Before RCA, where the real question is not who slipped first but which barrier was absent, bypassed, or never reliable.

Comparison table: which method answers which question

Which method fits which case

If the event is a routine precursor with a clean sequence, use 5 Whys after the timeline is stable. If the event has many parallel contributors and the room needs a shared language, use fishbone first and verify the branches later. If the event involves a fatality, a near-fatal path, or a barrier system that must be tested against combinations, use fault tree.

In practice, the best review often combines them. The incident timeline creates order, the fishbone widens the lens, and the fault tree tests whether the feared path is actually possible. That sequence keeps the team from mistaking speed for depth.

For fatality learning, Andreza Araujo's Um Dia Para Nao Esquecer, glossed in English as A Day Not To Forget, is the book that keeps the conversation connected to consequences, not just to process. It is a better guide than any single worksheet when the organization needs to respect the weight of a SIF review.

A 60-minute sequence for leaders

Start with the incident timeline for the first ten minutes, because no method works well when the sequence is still speculative. Then spend fifteen minutes listing only verified facts, not impressions. After that, choose the method that matches the question you still cannot answer.

Use 5 Whys if the chain is narrow and the evidence is strong. Use fishbone if the room needs to widen the hypothesis set. Use fault tree if the event sits close to a fatal path and the review must test how failures combine.

Close the hour by assigning one owner to each next action, one date for field verification, and one rule for reopening the review if new evidence appears. Andreza Araujo has seen in more than 250 cultural transformation projects that investigation quality improves when the team leaves the room with ownership, not just insight.

What leaders should remember

The best RCA method is not the one that looks most sophisticated. It is the one that fits the question, the evidence, and the consequence. 5 Whys is fast, fishbone is broad, and fault tree is exacting. A mature review knows when to use each one.

As Andreza Araujo argues in Safety Culture: From Theory to Practice, culture shows up in repeated decisions. If leaders keep choosing the easiest diagram instead of the right one, the investigation will keep reproducing the same blind spots. That is why the method decision belongs at the start, not at the end.