Naming one root cause usually hides several factors that combined to cause the accident

What it is

When an investigation report reduces an accident to a single "root cause," that conclusion is usually incomplete on its own terms. Real accidents almost always happen because several contributing factors act together at the same moment — a flaw in an interface design, a gap in training, understaffing at a particular time, an ambiguously worded procedure. No single one of these would have been enough to cause the accident on its own; it took their coming together to produce the outcome. Stopping an investigation at "we found a cause" flattens what was actually a multi-factor situation into a single-cause narrative. The factors left out don't disappear just because the report doesn't mention them — they stay in place, waiting to line up with the next contingent factor that comes along.

Why it happens

A "single root cause" narrative is appealing because it delivers a clean conclusion that points directly at one concrete fix — repair this one thing and the problem is solved — which is far simpler than dealing with three or four mutually independent contributing factors at once, and an investigation naturally tends to stop as soon as it finds the first explanation that sounds plausible. But the mechanics of how accidents actually happen make this simplification distort the picture: if a single factor on its own were enough to cause the accident, that class of accident would very likely already be happening frequently — and serious accidents are, in fact, typically rare. That rarity is itself evidence that several independent factors need to line up at once for it to happen. Explaining a phenomenon that's fundamentally a multi-factor coincidence with a single-cause narrative produces a fix that only blocks one of the pathways; the accident can still recur through another pathway that was never addressed.

Studying it

One way to test whether an accident is genuinely single-cause or multi-cause is to list, one by one, every other contributing condition that came up during the investigation but never made it into the final "root cause" conclusion, and check whether each of those, appearing on its own, had already left a lighter trace in historical records — a near-miss, a minor slip. If these conditions had already left independent traces before, that shows they're long-standing contributing factors that only escalated into an accident this time because they happened to coincide. Methodological caveat: the number of contributing factors found is itself sensitive to how deep the investigation goes — a more thorough investigation typically surfaces more of them — so judging "single cause versus multi-cause" can't rest on how many causes a report happens to list; it also requires checking whether the investigation process itself systematically pursued every link in the chain, rather than stopping at the first explanation that fit.

Where it stops holding

Pointing out the limits of a single-cause narrative doesn't mean an investigation should list contributing factors without limit and without prioritizing them — if every factor is listed as equally important with no distinction between what's directly fixable and what's merely background context, the report loses its ability to guide any remediation just the same. What this principle asks for is "don't stop after finding the first cause," not "don't rank the factors" — the two aren't in tension: after laying out multiple contributing factors side by side, it's still necessary to judge which ones are the highest-priority levers to act on right now.

Applying it

Build an explicit rule into the accident investigation process: once a plausible cause is initially found, the investigation isn't allowed to close immediately — it must push at least two more rounds of "if this were the only cause, why hasn't a similar operation caused a problem at other times," until it can identify the remaining contributing conditions that set this accident apart from other similar situations. Present the final report's findings as a list of every identified contributing factor, each tagged with its own remediation action, rather than compressing the conclusion into a single sentence: "the root cause is..." Verification: review a batch of past investigation reports and tally the average number of contributing factors listed in the "cause" field. If most reports list only one, and similar accidents have happened before, the investigation process has a tendency to converge too early, and continued questioning needs to be made mandatory in the process.

Related

• Same group: A10.16.1 hindsight bias in accident investigation · A10.16.3 timely error reporting depends on report cost being low enough and decoupled from punishment · A10.16.4 the point of an investigation is to find improvable system conditions, not to pin down who to blame
• Nearby: A10.07 Swiss cheese model
• Search terms: root cause analysis · contributing factors · multi-causal accident