Asking who's to blame steers an investigation away from what the system could fix

What it is

The moment an investigation's goal is set, it determines what questions get asked and what information gets collected. If the goal is framed as "determine who's responsible for this accident," interviews drift unconsciously toward gathering evidence that can support blame, and the person involved instinctively goes into self-protective mode — the very details that would reveal system conditions become harder to draw out. If the goal is framed instead as "find what fixable system conditions led to this accident," the questions, the tone, and the information collected all come out entirely differently: the person is far more likely to give a candid account of what they actually faced at the time, rather than a filtered version that only says what's favorable to them.

Why it happens

The investigation's goal produces a framing effect on the interview itself. An interview aimed at assigning blame naturally carries a judgmental tone — "why did you do that," "you should have known that was wrong" — and the person's rational response to that kind of question is self-protection: simplify the account, or shift responsibility onto external factors, and the real decision process and system conditions get buried under that layer of self-defense. An interview aimed at system improvement switches to descriptive framing instead — "what did the screen actually show you at that point," "what other sources of information did you have," "if you had to do it again, what kind of cue would have helped you avoid that judgment" — and the person doesn't have to defend their own accountability while answering, making it far easier for them to fully reconstruct the real situation at the moment the decision was made. The same person, recounting the same experience, can produce information of fundamentally different quality depending on which framing the interview uses — this isn't about whether the person is being honest, it's that the interview's framing itself shapes how much they're willing to say.

Where it stops holding

Aiming at system improvement doesn't mean any individual accountability issue that surfaces during the investigation gets ignored — if evidence shows the person knowingly violated a rule or deliberately concealed something, the investigation still needs to record that fact accurately. That part should simply be routed through a separate process, handled by a different role, rather than letting the shadow of accountability hang over the entire investigation from the start and suppress the systemic information it was meant to surface.

Applying it

Separate the investigation team from the authority to decide on discipline: the people conducting interviews, gathering evidence, and writing up the system-condition analysis should not be the same people who decide whether to discipline the person involved — only then does the person have real reason to believe candor won't lead directly to a consequence against them. State explicitly at the start of the interview that this investigation's purpose is to find system conditions, not to assign blame, and let that statement be backed by the investigation team's actual independence, not just a verbal assurance. Build the interview's question list around descriptive prompts: have the person reconstruct what the interface showed at the time, what information they had on hand, what alternatives were available — rather than asking directly "why did you do this." Limit the final report's conclusion section to remediation recommendations that point at specific system conditions — interface, process, training, staffing — and treat any recommendation that names an individual for discipline as a sign the investigation has drifted from its system-improvement goal; that needs to go through a separate disciplinary process, not get folded into the system-improvement report. Verification: review the investigation's question list and final conclusions, and tally the ratio of descriptive to judgmental questions, and of systemic recommendations to individual disciplinary ones. A high share of judgmental questions or individual disciplinary recommendations shows the investigation's stated goal never actually landed on system improvement in practice.

Related

• Same group: A10.16.1 hindsight bias in accident investigation · A10.16.2 a single root-cause conclusion usually masks several contributing factors · A10.16.3 timely error reporting depends on low report cost and decoupling from punishment
• Nearby: A10.09 human reliability and blame culture · Y7.05 accident investigation and feedback loops
• Search terms: accident investigation · blame-free investigation · systemic recommendations