If Every Pilot Took the CogScreen, How Many Would Fail?

TL;DR: CogScreen-AE has no pass/fail line, so the honest answer is "it depends what you mean by fail." But there is a real number worth knowing. In the aviator normative sample, roughly 10% of pilots under 45 had an LRPV above 0.6, against 29.4% of pilots over 45. That threshold is not a failure mark, it is the point at which fuller neuropsychological evaluation is indicated. The norms themselves come from 584 major airline pilots, later expanded, plus 207 general aviation pilots tested at AirVenture Oshkosh and Sun 'n Fun in 2013. You are being compared to pilots, not to the general public, and that distinction cuts in a direction most people do not expect.

---

The question that will not go away

There is a thread that resurfaces on pilot forums every year or so, in slightly different words each time: if all pilots had to take the CogScreen, how many would pass?

Underneath it is a real fear, and it is not an unreasonable one. If a test is hard enough that a meaningful share of perfectly competent, currently flying pilots would score poorly, then being sent to take it starts to feel less like an assessment and more like a coin flip with your career on it.

I want to answer that question with actual numbers rather than reassurance. Some of what follows is more encouraging than the forums assume. Some of it is less.

First, the framing is wrong in a way that matters

CogScreen-AE does not have a pass/fail score. There is no line where the software prints "failed."

What it produces is a set of performance measures across its subtests, expressed against normative data, plus composite indices. The most discussed of those is the LRPV, the Logistic Regression Probability Value, which estimates the probability of cognitive dysfunction on a scale from 0 to 1. Higher means a higher estimated probability.

A neuropsychologist then interprets all of that alongside your clinical interview, your records, your history, and the rest of the battery, and writes a report. The FAA reads the report. Nobody in that chain is reading a pass/fail box, which is why pilots who ask "what score do I need" rarely get a satisfying answer.

That said, thresholds exist in practice, and the LRPV is where the conversation usually lands. We cover the mechanics in understanding your LRPV score.

Where the norms actually came from

This is the part almost nobody explains, and it is genuinely interesting.

The original normative sample was 584 pilots from Part 121 major carriers, aged 24 to 67, split between large carriers and medium or small ones. A further group of Part 121 regional pilots, aged 24 to 62, was included as the sample grew. The normative base has since expanded considerably; the test's publisher currently describes normative data on roughly 1,447 air transport pilots aged 23 to 80.

The general aviation norms are newer and have a story. In 2013 the FAA and the test's developer ran a GA norming project, collecting data at the two largest general aviation fly-ins in the United States: AirVenture Oshkosh and Sun 'n Fun. That project produced norms from 207 GA pilots, ages 17 to 86.

The GA sample was deliberately skewed toward older pilots. Contemporary coverage of the effort quoted the developer saying they were particularly interested in non-commercial older aviators. That was a sound methodological choice given the age profile of the GA population, and it matters for how you read your own results, for reasons we will get to.

Separately, large volumes of CogScreen data have been collected on Air Force pilot candidates, and smaller sets on student naval aviators and controllers, generally as baseline data rather than as the civil certification norms.

"Normed on pilots" is harder than it sounds

Here is the piece that surprises people.

Most cognitive tests you might encounter compare you to the general adult population. CogScreen-AE compares you to pilots.

Pilots are not a random sample of adults. They have been selected, trained, and repeatedly screened. As a group they perform above general population averages on exactly the kinds of abilities this test measures: processing speed, working memory, divided attention, spatial reasoning.

So being at the 30th percentile against pilot norms is a meaningfully different statement than being at the 30th percentile against the general public. The bar is set among people who are already, on average, good at this.

This is not a flaw. It is the entire point. The FAA is not asking whether you function like a typical adult, it is asking whether your cognition looks like a pilot's. But it does explain why capable people sometimes come out of the test feeling they underperformed. The comparison group is tough.

The number the forums are actually looking for

If you want a single statistic that speaks to "how many pilots would fail," this is the closest honest thing to it.

In the aviator normative sample, as reported in the aeromedical literature:

Group	Share with LRPV above 0.6
Healthy pilots under 45	about 10%
Pilots over 45	29.4%

Read that carefully, because it is easy to misread in both directions.

It does not mean 29.4% of older pilots failed. These were pilots in the normative sample, not people who were grounded. An LRPV above 0.6 is a signal that fuller neuropsychological evaluation is warranted, not a verdict.

But it does mean the age gradient is steep and real. Roughly three times as many pilots over 45 sit above that threshold as pilots under it. If you are 55 and your LRPV comes back at 0.62, you are in a band that includes a substantial minority of your healthy peers.

This is why the published guidance is explicit that users of the test must be sensitive to the age relationship. A number that means one thing at 30 means something different at 60, and any interpretation that ignores that is a bad interpretation.

So what does a "bad" result actually trigger?

Not a denial, in most cases. It triggers more evaluation.

An elevated LRPV, a weak domain, or an inconsistent pattern is a prompt for the neuropsychologist to look harder: additional focused testing, closer record review, attention to whether something situational explains it. Fatigue, poor sleep, anxiety, and unfamiliarity with the task format all degrade performance on tests like this, and a competent evaluator knows that.

If you are staring at results you are unhappy with, what to do when your CogScreen comes back low walks through the realistic paths from there.

Does the test actually predict anything?

Before I answer that, you should know where I am standing. PilotPrep sells CogScreen-AE preparation. If you conclude from this section that the test is beyond reproach, that is good for our business. So weigh what follows accordingly, and check the sources at the bottom.

I am going to give you the argument on both sides anyway, because a reader who wants the criticism can find it in about ten seconds of searching, and a site that pretended it did not exist would deserve to lose your trust. We also do not actually have a stake in the outcome. Whether the FAA should be ordering this test is a policy question. Whether you have one scheduled is a fact about your life, and that is the problem we work on.

The case for it. CogScreen-AE was built for the FAA specifically to detect subtle cognitive change relevant to flight. Validation work during development used flight data recorder data across dozens of recorded flight parameters, and reported significant correlations between a number of CogScreen variables and flight performance. Later work by Taylor and colleagues (2000) examined the relationship between CogScreen-AE performance, flight simulator performance, and pilot age. Its design also supports repeat administration in a way most conventional neuropsychological instruments do not, which matters when the question is whether someone has changed.

The case against it, or at least against how it is used. AOPA's medical columns have argued pointedly that CogScreen has limited value outside substance and traumatic brain injury contexts, and specifically questioned its use as a gatekeeper in straightforward depression cases. Reform advocates have gone further, questioning the test's cost, its proprietary nature, and whether its use has demonstrably improved safety outcomes.

My own read, offered as a professional opinion rather than a finding: the test is on its firmest ground where it was designed to be, detecting change in cognition after a neurologic insult or in the context of substance history. It is on softer ground when it is used as a general screen in cases where the underlying question is a mood disorder rather than a cognitive one. That is a critique of application, not of the instrument.

None of which changes your situation if you have an appointment. The debate is worth understanding. It is not a strategy.

What this means for you, practically

If you are under 45 and generally healthy, the normative data suggests the odds are with you, and the largest controllable risk to your score is not your cognition. It is arriving cold to an unfamiliar interface under high stakes.

If you are over 45, know the age gradient before you see your number, so that a result in the 0.6 range reads to you as "this is a band a lot of my healthy peers are in" rather than as a catastrophe.

In either case, the variables you control are sleep, timing, and format familiarity. Not intelligence. The task formats are unusual, the instructions come quickly, and the difficulty adapts to your performance, which means correct answers are routinely followed by harder items. Pilots who have not seen that pattern often read it as failure while they are still being tested, and anxiety takes a real bite out of the measures this test is built to capture.

You cannot practice the actual CogScreen items. They are proprietary and administration controlled. Familiarization with the task formats is a different thing, and it is both permitted and sensible. The goal is not a score that overstates you. It is a score that is actually about your cognition rather than your first encounter with the interface.

Sources

• Kay, G. G. CogScreen Aeromedical Edition: Professional Manual. Psychological Assessment Resources, 1995
• CogScreen LLC, published normative data summary (air transport, general aviation, and major airline samples)
• FAA, Neurocognitive Assessment of Pilots: The FAA Perspective, presented at AsMA 2017 (normative sample composition and the 2013 GA norming project)
• AOPA, "CogScreen creator offers test at AirVenture" (2013)
• Taylor, J. L., O'Hara, R., Mumenthaler, M. S., Yesavage, J. A. "Relationship of CogScreen-AE to flight simulator performance and pilot age." Aviation, Space, and Environmental Medicine, 2000
• Published aeromedical review reporting LRPV distribution by age in the US aviator normative sample
• AOPA Flight MD columns on CogScreen use in depression cases (2022)

---

Curious where you would actually land? See what the 13 subtests measure, or try three questions per module free to see the task formats for yourself before deciding whether structured preparation makes sense for your situation.