Telling an Incident in an Interview Without Disqualifying Yourself: the Line Between Confession and Analysis

Salle de débriefing avec tableau blanc effacé et deux chaises vides sous une lumière froide

“Tell me about a time you made a mistake.” The question comes after the technical questions, once the panel has gauged your level and is looking for something else. It is not the mistake itself that interests them: it is what you made of it, and what you are able to say about it to people who were not there.

What the panel is listening for, and it is not what you are telling

A selection interview brings together two ears that do not hear the same thing. The operational manager’s ear checks a presence: were you really in that seat, carrying that load? It picks this up from details that cannot be invented — flight phase, configuration, time available, the cue that triggered the doubt. The recruiter’s ear, or the safety manager’s, listens for something else: what you attribute what happened to, and whether that attribution holds up.

Competency-based assessment has given these two ears a shared vocabulary: situation awareness, workload management, communication, teamwork, decision-making. A panel uses it as a reading grid and looks for observable markers rather than declared qualities. Saying you are rigorous produces no marker; describing the moment a deviation was detected and what made it visible produces several.

The most common way to fail is to satisfy only one ear: a technical account, thin on decisions, leaves the second ear with nothing to work with, and an introspective account rings hollow to anyone who has actually done the job.

Confession or analysis: the line runs between intent and conditions

A confession explains an act through an intention or a trait: haste, inattention, wanting to save time. These formulas close off analysis; they offer an internal, stable — and therefore reproducible elsewhere — cause, which is exactly what an interviewer is trying to rule out. Analysis explains the same act through the conditions that made it likely: what was displayed, what the procedure required, what reality allowed, which cue was missing to see it sooner.

The useful concept is local rationality: at the moment of the action, what was done made sense given what was perceived. Reconstructing that sense is not justifying oneself, it is doing the work of an investigator. The account becomes verifiable, and the panel will test it by changing one parameter. A confession cannot be tested: it leaves only one possible follow-up question, and it is a bad one.

Symmetry matters just as much. An account that describes only the conditions slides into a cop-out: if the system did everything, the narrator did nothing. It holds together when it carries three things at once — the conditions that made the deviation likely, the personal part that was missing, and the change made since. Removing one destroys the other two.

A good answer doesn’t blame yourself or the system. It explains what happened, why it made sense at the time, and what you learned to do better next time.

A pilot describes a go-around where they missed a low-energy warning. Instead of saying “I was distracted,” they explain the cockpit noise, the checklist step that masked the cue, and the new cross-check they now use.

Speaking time compared across two accounts of an eventTwo horizontal bars compare the distribution of speaking time. In the confession account, the description of the event dominates and what follows is reduced; in the analysis account, the description is short and the aftermath takes up nearly half.confession accountcontexteventaftermathanalysis accountcontexteventdecision and aftermathshare of speaking time
What separates the two accounts is not their content but their time budget. The context takes up the same space; what shifts is the ratio between describing the event and handling what came after.

Choosing the event before writing the account

Most candidates work on the account and neglect the choice of event, even though that choice decides the outcome. Two criteria intersect: the severity of what happened and the richness of what can be drawn from it for others. A minor but dense event — a late detection that was caught in time and then analysed — is the best currency available.

A serious event does not have the same status. It is kept for a direct question, and is then told with the conclusions of the analysis carried out by the operator or the investigators, not your own. A safe, low-stakes anecdote, for its part, is seen for what it is: a polite dodge, and it is the dodge that the panel notes.

Matrix for choosing an event to tellMatrix crossing severity and transferable learning. Low severity, low learning: anecdote. Low severity, high learning: reference account. High severity, high learning: to be kept in reserve. High severity, low learning: do not open.do not openclear risk, no gainkeep in reserveif the question comes upanecdotenothing to draw from itreference accountopen it yourselfhighseveritylowlowtransferable learninghigh
The best account is not the most serious one, it is the richest one at moderate severity. The bottom-right box is the only one you should offer spontaneously; the top one is dealt with if the question comes up.

Calibrating the detail: enough to prove it, not enough to identify anyone

Technical detail buys credibility, but it buys it fast and then stops paying off. Four elements are almost always enough: the phase and configuration, the dominant constraint, the cue that raised the alarm, the time available to act. Beyond that, each further precision no longer adds probative value and starts to point to something: a flight, a base, an operator, a person.

Risk, on the other hand, does not plateau, and it is not carried by any single detail but by their intersection: an aircraft type, a period, a pair of airports and a type of event are often enough to trace the file. Anonymity is not declared: it is built by removing elements one by one.

Probative value and risk of identificationTwo curves plotted against the precision of the details given. Probative value rises quickly then plateaus; the risk of identification stays low then accelerates. A vertical band marks the window between the two.probative valueidentification riskuseful windowlevelprecision of details given
Useful detail has a ceiling, identifying detail does not. The first technical precisions buy almost all the credibility; the ones that follow start to point to a specific flight or person.

The trap. An anonymous account can still make a third party identifiable. As soon as it points to a unique function on board — the commander of that flight, the mechanic who signed off the return to service — and situates the event, that person is named without their name ever being said. What the panel then infers no longer concerns the absent person, but you.

The frame of an account that fits in two to three minutes

A usable account fits in five movements and can be timed. The setting, in two sentences, with no geography. The cue, with the moment it was perceived and what made it perceptible. The decision, with the reasoning as it stood at the time, not as it is reconstructed now. The deviation and its actual effect, without dramatising or minimising. Finally, what changed, stated as a rule applied since — and there, you stop.

Stopping is the hardest movement. The silence that follows a well-built account is a working silence: the panel is looking for its follow-up question. Filling it with one more sentence, almost always evaluative, turns the analysis back into a confession. The final rule must be verifiable: a rule you could not picture being applied tomorrow on the line is not really a rule.

The same event, told two different ways
What the panel observes Confession account Analysis account
Starting point the narrator’s state of mind the situation and its dominant constraint
Dominant grammar first person, centred on the fault the crew and the procedure, then a decision owned
Conditions invoked as an excuse linked to the decision
Other people involved identified, sometimes blamed present through their role, never identifiable
End of the account the mistake, or the sanction the rule applied since
Effect produced the panel fills the gaps with the worst case the panel tests the rule
Likely follow-up “does this happen to you often?” “and if time had run out?”

The right move. Decide before the interview where the account stops and write its last sentence word for word: it is the only part worth memorising, the rest is better left spoken naturally. Prepare the two most likely follow-up questions — frequency and the counterfactual — with a short answer for each.

Error, violation, deviation: the test the panel applies without naming it

A panel in this industry sorts things the way a safety manager does. An error is unintentional: the intent was compliant, but perception or execution failed. A violation is a deliberate deviation from a known rule, which prejudges neither its severity nor any malicious intent — many of the ones seen in daily operations are routine and functional, adopted because they appear to work. Confusing the two is noticeable.

The substitution test is the implicit tool: would another competent professional, in the same conditions and with the same information, have acted the same way? If yes, the event speaks about the system, and the account must show that; if no, it speaks about you, and it is better to say so before you are made to say it. A violation account only becomes tellable on one condition: naming what made the deviation attractive — a cumbersome procedure, a real time saving, an established local practice — then saying what has since removed that attraction.

The accounts you do not tell

An event whose analysis is not closed is not to be told, whatever its demonstrative value. An ongoing safety investigation, an open judicial proceeding, a file still moving through the operator’s process: the correct answer is short. State that the file is not closed, that you can describe what you have taken from it for your practice but not the circumstances, and offer another event instead. This answer is itself a marker.

It is also worth understanding what protection of the event report actually covers. The European framework rests on the confidentiality of the source and on the principle that a report must not be used against its author, with the reservations essentially limited to deliberate breach and manifest disregard of an obvious risk; the exact wording, its limits and its national implementation are found in the applicable text and in the operations manual, not in a paraphrase. This protection covers the reporting system; it does not follow the candidate into an interview room, where nothing said is protected by anything.

The trap. The follow-up question that destroys a good account is almost never technical: it is “and how often does this happen to you?” An honest but bare estimate turns an event into a personality trait, and a trait travels with you into the next job. The sustainable answer distinguishes the class of error, common to everyone, from its detection, and describes what makes that detection reliable in your practice today.

What the target position changes about the account

For a pilot position, the expected account is a threat-and-error-management account: detection, decision, trajectory control, task allocation; the panel listens for where the attention was and how the crew was used. In the cabin, the event is more often relational — a passenger, a coordination that was missed, an instruction applied too late — and what is assessed is holding a limit and escalating the information.

For a certifying staff position, the account touches the signature: the panel expects an event where the return to service was delayed or refused, with the documentation that settled it and the pressure that was applied. For a flight safety position, the personal account gives way to a handling account — the analysis of an event reported by a third party, the measure adopted, the feedback given to the reporter. The confession trap shifts there: it becomes the trap of confiding about a previous operator.

Building a reserve, and the zone that stays grey

Three events are enough, calibrated to different severities, written out once and then spoken aloud and timed. Have a peer listen and ask the two follow-up questions; discussions among practitioners on forums.fl330.com are a good place to test an account before a panel does. The training courses that include self-analysis build the same discipline.

A man in a suit gestures while speaking at a meeting with others around a table.
Frame incidents as learning opportunities in interviews to showcase growth.
  • What the operations manual says about disclosing information from the reporting system outside that system
  • The exact status of each of your events, closed or not
  • And, depending on the State, the existence of professional record-sharing schemes consulted by a future employer

The grey zone will not disappear. No text sets the threshold beyond which an anonymised account becomes identifiable again: it depends on how well known the event is and on the size of the community. The only way to handle this uncertainty is to move it — build the account so that its value rests on the reasoning rather than on the event. A panel that cannot identify the event and still retains the rule has received exactly what it was looking for.

This text offers a method of expression, not a standard. In case of divergence, the applicable regulations and the operator’s approved documentation prevail over what is written here.

Questions fréquentes

Can you answer that you have never made a mistake?

No, and the answer is read as a lack of self-analysis rather than as a strong performance. An experienced professional who cannot find any event to tell signals either that they do not detect their own deviations, or that they refuse to talk about them, and both hypotheses are unfavourable. It is better to offer an event of low severity but rich in lessons, prepared in advance.

Can you talk about an event that was the subject of an occurrence report?

Yes, for what concerns your own practice and what you drew from it; no, for what identifies a third party, a flight or an operator. The protection organised by the occurrence-reporting framework targets the reporting system and the use an employer might make of it; it does not extend to what is said spontaneously in an interview. If the file is not closed, say so and offer another event instead.

What should you answer to “and how often does this happen to you?”

Distinguish the class of error from its detection. The class of error — a configuration omission, a reading that was too quick, a forgotten item on resumption — is common to everyone, and denying it convinces no one. What should be described is the mechanism that makes it visible early in your current practice, and the moment it kicks in.

Should you give the aircraft type, the date or the operator's name?

The aircraft type and the flight phase serve credibility and can be given without difficulty. The exact date, the pair of airports, the registration and the operator’s name add no demonstrative value and, once cross-referenced, often make it possible to trace the file. Remove them as soon as you write the account, not when you tell it.

How should you handle an event that led to a sanction?

Mention it soberly as a fact, without contesting it in front of the panel, then spend most of the time on what has changed in your practice since. Contesting a sanction in an interview reopens a conflict in front of people who have no way of settling it. Where a professional record-sharing scheme exists, staying silent about the episode costs more than raising it yourself.

Does a personal event account work for a flight safety position?

It works as an opener, but it is not enough. The panel mainly expects a handling account: an event reported by a third party, the analysis carried out, the measure adopted, the feedback given to the reporter, and the check that the measure produced its effect. Discretion about the previous operator is judged there just as severely as the quality of the analysis.

Commentaires

Leave a Reply

Découvrez l’écosystème FL330

FL330, c’est toute l’aviation professionnelle — explorez nos autres espaces.