A candidate arrives at an entry-level aptitude battery with several dozen hours of stick time logged on a home simulator and two months of tracking exercises redone every evening. He comes out very highly ranked, joins the course, then finds himself behind everyone else the moment instrument flying begins — not because he flies badly, but because radio work, chart reading and parameter management all arrive at once. His score was not wrong. It was simply measuring something other than what his preparation had trained.
Three gateways that are systematically confused
In hallway conversation, the word “psychometric” covers three different arrangements, with neither the same commissioning body, nor the same reference criterion, nor the same consequences. Confusing them creates false expectations on both sides of the table: the candidate believes he is sitting a state examination, the recruiter believes he holds a legally binding measure of aptitude.
The first gateway is the aptitude battery administered by a training organisation, or by an operator selecting cadets. It is private and contractual. Its reference criterion is completion of the course: it was built, and then recalibrated intake after intake, to sort candidates by their probability of finishing a specific programme, within a specific timeframe, at the failure rate the organisation is prepared to absorb economically.
The second is the Class 1 medical certificate, which includes a mental health component. Here the commissioning body is the authority, the examiner is certified, and the decision is administrative: reasoned, legally binding, and subject to appeal. It ranks no one. It records fitness, unfitness, or fitness with limitations.
The third is the psychological evaluation that European air operations regulation requires the operator to arrange before a pilot begins operating in line service. Its purpose is safety, not selection: it aims to identify anything that could interfere with the safe exercise of duties, within a broader framework that also includes peer support and confidentiality safeguards. No one is scored or ranked in it, and a candidate who approaches it as a test to be passed is setting himself up to lie on the one question where lying is costly.
There are three very different tests often called “psychometric” in aviation. One predicts if you’ll finish training, one checks if you’re medically fit, and one ensures you’re safe to fly. They’re not the same, and mixing them up causes problems.
A candidate spends 60 hours practicing tracking tasks and scores 95th percentile. He passes the aptitude battery but struggles in flight school because the test didn’t measure his ability to handle multiple tasks simultaneously under real flight conditions.
| Arrangement | Who requires it | What it looks for | Effect of an unfavourable result | Retake or appeal |
|---|---|---|---|---|
| Entry aptitude battery | The training organisation, or the operator recruiting cadets | The probability of completing a given course within the planned timeframe | Application rejected, with no detailed reasoning or score disclosed | Waiting period imposed by the provider, candidate history retained |
| Class 1 medical certificate, mental health component | The authority, through an accredited centre or examiner | Fitness assessed against published criteria | Unfitness or limitation, as a reasoned administrative decision | Second opinion and appeal routes organised under the regulation |
| Psychological evaluation before line operations | The operator, in application of air operations regulation | What could interfere with the safe exercise of duties | Support, medical referral, temporary employment restriction | Ongoing arrangement, backed by peer support and confidentiality |
What a battery measures: a cost, not a knowledge
No one is trying to find out whether you can hold a heading. The aim is to find out how much it costs you to hold it, and how fast that cost falls as you repeat the task. This distinction in mechanism explains almost everything else, including the disconcerting shape of certain exercises.
Flying is a profession of competing tasks. The primary task must become automatic enough that navigation, phraseology, systems monitoring and crew management can fit into whatever attentional resource remains. That remainder is the only quantity that matters, and it cannot be observed directly: it only shows itself once it is saturated. Two candidates producing the same tracking error after three minutes are therefore not equivalent; whichever one got there while consuming less resource holds a reserve the other does not, and that reserve only shows up once a second task is added.
The second construct targeted is a rate of learning. An exercise repeated in blocks is not there to yield a better score trial after trial: it is there to measure how much the candidate improves. The slope is more informative than the level, for a very concrete reason — it holds up better against whatever the candidate may have seen beforehand. A high starting level can come from ability, targeted preparation, or a prior attempt; a strong rate of improvement on a new task is far harder to manufacture.
This is also why an aptitude battery does not look like a cockpit. Resembling the job serves no purpose if the common ground is not the mechanism being tapped: an exercise that is too realistic would mechanically favour those who have already flown, which is precisely the opposite of the goal when selecting beginners.
These tests don’t measure how well you can fly – they measure how much mental effort it takes you to learn new tasks, and how quickly you improve. The person who starts worse but improves faster is often the better candidate.
On a psychomotor tracking test with added math problems, Candidate A has 10° deviation with 3 math errors, while Candidate B has 15° deviation with 2 math errors. The test measures which candidate has more mental reserve – Candidate B, despite the worse tracking score, likely has better overall aptitude.
What training shifts, and what it does not shift
It shifts a format cost, and that cost is real. Understanding the instructions, knowing whether the exercise penalises slowness or error, knowing the travel of the stick, not wasting the first block decoding a display: all of this is noise that contaminates the measurement. Removing it improves the accuracy of the assessment, in the interest of both the candidate and the organisation. It is precisely for this reason that exercises include a trial phase, and that many organisations distribute a demonstration of their own.

It shifts little beyond that, and the mechanism has a name: practice specificity. You improve on the trained task, much less on a neighbouring task where one parameter has changed — control gain, inertia, control order, presence of a disturbance, display latency. Providers deliberately vary these parameters from one session to the next. The skill acquired stays attached to the specific version that was rehearsed.
Scoring absorbs the remainder. An adaptive exercise raises difficulty until it reaches a target error rate, which turns familiarity into a mere shortcut to the true ceiling: you get there faster, you do not get there higher. And the score is almost always norm-referenced, hence relative to a population of candidates who have all gone through the same trial phase — a gain shared by everyone moves no one’s ranking.
| Score component | Shifted by preparation | Genuinely useful work | Effect of prolonged training |
|---|---|---|---|
| Understanding the instructions and the scoring logic | Yes, entirely | A few dozen minutes | Nothing more to gain: the ceiling is reached straight away |
| Familiarity with the stick, axis travel, screen layout | Yes, largely | A few short sessions | Gain that does not carry over to a different control geometry |
| Speed of applied calculation, instrument reading, orientation | Yes, moderately | A few weeks | Real progress that plateaus, and transfers poorly outside the format |
| Working-memory span and updating | Very little, and mostly on the exact task trained | Low return from the very first hours | Score that stops representing capacity during the course |
| Residual capacity under dual task | No | None | Breakdown simply postponed to the first demanding phase of the course |
| State of the day: sleep, hydration, stress management | Yes, but only the day before | One night | Probably the only high-yield lever in the final twenty-four hours |
The trap. A preparation package sold against the exact battery of a given school produces a score that is not really yours, and it is paid for twice. First right away: providers generally impose a waiting period between two attempts — typically on the order of a year — and retain the candidate’s history, so an undeclared second attempt eventually surfaces, and becomes a far more serious sincerity problem than the score itself. Then during the course: the gap between the displayed score and the actual ability is not settled at entry, it is settled at the first demanding phase, when a significant share of the funding is already committed and dropping out recovers none of it.
That leaves the state of the day, systematically underestimated. Vigilance and working memory are among the functions most sensitive to sleep debt, and they are two major components of the score. Over the course of a morning of assessment, one good night’s sleep can shift certain sub-scores more than weeks of repeated practice on a tracking task.
Why validity is modest, and why it is enough anyway
No battery predicts an individual’s trajectory with precision. The relationship between score and course completion is real, documented, and very far from deterministic: one cannot say of any single candidate that he will fail. A serious organisation never claims this, and simply rejects an application without pretending to a prognosis.
What makes the tool useful is not its precision at the individual level, it is the selection rate. When only a minority of candidates are retained, even a modest association between score and criterion noticeably shifts the completion rate of the selected intake. The practical consequence deserves to be stated plainly: a battery is good at excluding a tail of the distribution, and poor at ordering a leading pack. A few ranks of difference in the middle of a session are largely measurement noise; the gap between the top and bottom of the distribution is not.
Two statistical effects finish blurring the judgements passed on these tools. The first is range restriction: observed validity is calculated on the admitted candidates only, that is, on an already skimmed sample, which mechanically underestimates it. The second is asymmetric feedback: false positives are visible — they fail during the course, in front of everyone — while false negatives are invisible, since the organisation never learns what would have become of those it turned away.
The trap, on the recruiter’s side. A battery is validated as a composite; an isolated sub-score has markedly lower reliability and a proportionally much larger measurement error. Writing in a report that a candidate is “weak in divided attention” based on a single sub-test is commenting on noise with the authority of a number. Same family of error: reusing an entry-level battery, validated against completion of initial training, to assess a serving crew member. The criterion is not the same, nor are the norms, and the instrument is used outside its validated domain.
The right reflex. Decide the decision rule before seeing the results: what weight for the battery, what weight for the interview, and above all what to do when the two contradict each other. A combination fixed arithmetically in advance generally outperforms an impression assembled after the fact, because it cannot be swayed by good personal rapport. On the candidate’s side, the counterpart is simple: treat the format as noise to be removed, which takes a few hours; the state of the day as a real lever, which takes one night; and everything else as information you have every interest in getting right.
What no battery will ever see
A morning of exercises looks at ability at a single point in time. The course itself lasts years. What causes failure during training is not primarily aptitude: it is funding running out at the two-thirds mark, a suspended medical certificate, a life that changes, a house move, the weariness of constant evaluation. None of these factors can be measured in one morning of exercises, and all of them weigh more heavily on the completion rate than the two or three points separating two candidates on a norm-referenced scale.
Another thing almost entirely escapes the written exercise: the capacity to be corrected. In line service, a career hinges on accepting a debriefing, on what one does with an unpleasant and accurate remark. The group exercise and the assessment flight catch a little of this, the structured interview a little more, the battery none at all.
As for personality questionnaires, their status needs to be stated clearly: they are transparent. An informed adult can see what is expected and answer accordingly, and designers have known this for a long time. They therefore serve to generate follow-up interview questions and to spot internal inconsistencies or contradictions with the candidate’s own account, not to filter. Nor, incidentally, is any stable “pilot profile” known: the traits that emerge are broad, common to many demanding professions, and far too non-specific to settle an application.
What to check before paying for preparation
Four questions should be put to the organisation before any spending, and they should be put in writing. Is the battery a straight pass/fail cut-off, or weighted alongside the interview and the file? What is the waiting period in the event of an unfavourable result, and does it run from the session or from the decision? Is the result usable for another application, or must everything be redone elsewhere? Is any debriefing planned, even a summary one? An organisation that answers all four precisely has thought about its own tool; an organisation that dodges them is already telling you something.
The grey zone, however, will not close. Norms are not published, the raw score is not disclosed, no appeal route in the administrative sense of the term exists, and the candidate cannot audit the instrument that rejects him. A subtler drift compounds this: the criterion itself moves. A battery validated against completion of a conventional course keeps being used for career paths whose degree of automation, phase structure and pace of checks have changed. Validation is never acquired once and for all, and it is the kind of subject practitioners discuss more readily among themselves than in brochures — the exchanges on the FL330 forum give a more faithful picture of it than presentation pages, and the detailed course paths on formations.fl330.com show where intakes really lose their students.
For anyone preparing an application today, one last thing deserves to be laid out plainly. An unfavourable result is not a career verdict: it says that an organisation, with its own criterion and its own selection rate, prefers not to commit its course to this particular file. Other routes exist, slower and more gradual, where the same person builds the reserve that was missing. A score obtained by polishing a familiar exercise, on the other hand, serves no one: it does not buy ability, it buys a delay before the question resurfaces, at the exact moment it costs the most.
In the event of any discrepancy between the foregoing and an actual case, the applicable regulation, the organisation’s contractual terms and the operator’s approved documentation prevail in all circumstances.
Questions fréquentes
Can you retake an aptitude battery after a failure?
Yes, but rarely straight away. Providers generally impose a waiting period between two attempts, precisely so that the familiarisation effect does not contaminate the measurement, and they retain candidates’ history. Declaring a prior attempt up front is therefore the only workable attitude: concealment is discovered by simple cross-checking of files, and a sincerity problem weighs far more heavily than an average score. Use the interval to work on what genuinely trains, starting with applied calculation and orientation.
Does a low score mean you will never fly?
No, and the instrument does not claim that. A battery estimates the probability of completing a given course, within a given timeframe, against a criterion set by the organisation; it is far more reliable at excluding a tail of the distribution than at ordering candidates who are close together. The same person can build, through a slower route, the reserve that was missing on the day of the exercise. What is costly, on the other hand, is bypassing the signal through targeted training, because the question then resurfaces during training.
Can personality questionnaires be “answered well”?
They are transparent, and the designers know it. An informed adult has little trouble identifying the expected answer, which is why they almost never serve as a stand-alone filter. Their real use is to generate follow-up interview questions and to spot internal inconsistencies, or contradictions between the questionnaire and the candidate’s own account. An answer that is too smoothed-over across every scale is itself a signal, and it steers the interview toward the points that were being masked.
Is the psychological evaluation required before line operations an aptitude test?
No, its logic is different. It falls under air operations regulation, it is carried out by the operator, and its purpose is to identify what could interfere with the safe exercise of duties. It produces neither ranking nor score, and sits within a broader system that includes peer support and confidentiality safeguards. Approaching it as a test to be passed leads to concealing precisely what the arrangement is meant to allow one to address.
How much time should be usefully spent on preparation?
A few hours to neutralise the format: instructions, scoring logic, getting used to a stick, one or two demonstration sessions. A few weeks, at most, for applied calculation, instrument reading and spatial orientation, which genuinely improve and go on to serve during training. Beyond that, returns collapse and risk increases: the score stops representing the ability that will be called on during the course. Relative to the time invested, the night before the assessment often remains the single most cost-effective lever in the whole preparation.
Does training on a home simulator improve the psychomotor score?
Partially, and less than one hopes. The gain stays largely attached to the trained configuration: control gain, inertia, control order and display latency differ from one device to another, and providers deliberately vary these parameters. Adaptive scoring adds to this, turning familiarity into a mere acceleration toward the real ceiling rather than a higher ceiling. The most solid benefit of those hours lies elsewhere: they tell you, the candidate, how much you actually enjoy the task.

Leave a Reply
You must be logged in to post a comment.