NASA studied pilots taking 26-minute naps during long flights and found their performance improved by 34 percent and alertness by 54 percent, evidence that in aviation, rest could be a safety tool, not a weakness

In the early 1990s, a group of NASA researchers arranged for twenty-one long-haul pilots on transpacific Boeing 747 routes to do something most airlines discouraged. They let some of them sleep. The full study was published by NASA Ames Research Center as a technical memorandum titled Crew Factors in Flight Operations 9: Effects of Planned Cockpit Rest on Crew Performance and Alertness in Long-Haul Operations, and it became one of the most cited studies in aviation fatigue research.

The version that circulates online is compact: a 26-minute nap improved performance by 34 per cent and alertness by 54 per cent. Those figures are the ones usually cited from a 1995 summary paper by Rosekind and colleagues in the Journal of Sleep Research, “Alertness management: strategic naps in operational settings,” rather than from the detailed memorandum, whose own summary describes the effect in words rather than percentages. What gets lost in the compression is almost everything that makes the numbers meaningful.

What follows is a reading of one study and the policy history around it, not advice on how anyone should manage their own sleep. And it is worth being clear at the outset: this is one study, small and specific, not a settled consensus about rest in general.

What the study actually measured

The work was led by Mark Rosekind, then at NASA Ames, with co-authors including R. Curtis Graeber, David F. Dinges of the University of Pennsylvania, and Linda J. Connell. The pilots flew regularly scheduled overnight trans-Pacific legs in three-person crews with no relief pilot available.

Twelve of them formed the Rest Group. During the low-workload cruise phase, each was given a planned 40-minute window to rest in the seat while a colleague flew. The other nine formed the No-Rest Group, given the same 40-minute block but told to carry on with normal duties. Throughout, the researchers recorded brain-wave and eye-movement activity, tracked reaction time on a vigilance task, and logged subjective alertness.

The rest-group pilots slept on 93 per cent of the opportunities offered. They fell asleep in an average of 5.6 minutes and slept for an average of 25.8 minutes, which is where the famous 26 comes from.

The 34 and the 54 attach to different measures. The 34 per cent refers to reaction-time and vigilance performance, tested with a psychomotor task. The 54 per cent, often written as “up to 54 per cent,” refers to physiological alertness derived from the brain-wave recordings, essentially a reduction in the brief involuntary lapses, or microsleeps, that show up on an EEG when someone is fighting to stay awake. Rosekind and colleagues report these figures in the 1995 summary paper. The underlying memorandum is more restrained: it states that the nap was associated with improved physiological alertness and performance, and that the benefit carried through the descent and landing, when it mattered most.

What the numbers do and do not say

The comparison is the part most summaries drop. These are not gains over a rested, sharp baseline. They are improvements relative to the No-Rest Group, a set of tired pilots getting steadily more tired across a long overnight leg. The napping crews were not turned into superhumans. They were held closer to functional while the others slid.

The authors are also candid about the limits of what a nap buys. In their reading, the cockpit rest did not reduce the pilots’ cumulative sleep debt and did not change how they slept during the layover afterwards. A 26-minute nap is a countermeasure for one demanding stretch of duty, not a repayment on chronic sleep loss.

None of this makes the finding trivial. A short, planned rest measurably reduced dangerous lapses during the highest-stakes phase of a flight, and it did so with minimal disruption and no safety problems the researchers could identify. That is a useful result. It is a narrower one than “napping makes you a third more productive.”

Why the cockpit does not translate cleanly to the cubicle

Most people meet this study through a productivity article that quietly swaps the pilots for office workers. The leap is shorter in the headline than in reality.

The population was unusually specific: professional aircrew, deep into an overnight long-haul rotation across multiple time zones, in a job where a lapse of attention has consequences a spreadsheet error does not. Their fatigue was severe and circadian in origin. A 26-minute nap has a large effect precisely because there was so much impairment to claw back.

An adequately rested person at their desk in the early afternoon is not in that hole and will not see the same swing. The study says a well-timed nap rescues a fatigued professional during a critical task. It does not say a nap adds 34 per cent to anyone’s ordinary working day, and it was never designed to test that claim.

Where the finding actually landed

The more interesting afterlife of this research is regulatory, and it splits along a line that still holds three decades later.

The evidence that short in-seat rest improves late-flight alertness helped shape controlled-rest procedures now permitted for some or all operators in Europe, Canada, Australia, New Zealand, China, Israel and elsewhere. EASA guidance allows a controlled rest period of roughly 45 minutes plus recovery time, with the other pilot monitoring and cabin crew checking in. The framing across these regimes is consistent: controlled rest is an in-flight response to unanticipated fatigue, used within a defined policy, not a scheduling tool and not a substitute for proper rostered sleep. The Flight Safety Foundation’s resource for operators lays out that logic in detail.

The United States is the notable holdout. The Federal Aviation Administration considered controlled rest during its 2010 flight-and-duty rulemaking and left it out of the final rules. In its own guidance on aviation fatigue, the FAA acknowledges that even naps of 25 to 30 minutes restore alertness and performance, yet it permits in-flight napping only when an augmented crew allows a pilot to rest away from the controls, not in the seat on a standard two-pilot flight. Same research base, different regulatory conclusion.

That divergence is the real story hiding inside the tidy statistic. A finding can be well established scientifically and still fail to become policy in a given jurisdiction, held up by liability concerns, union positions, and institutional caution rather than by any dispute over the data.

Rest as a monitored procedure

There is a cultural reading of this study that treats it as vindication against the idea that rest is for the soft. We would be careful with that. What the aviation world did was more specific than a mood shift about laziness. It took rest and turned it into a defined procedure, with a fixed window, a monitoring pilot, a length short enough to avoid the grogginess of deeper sleep, and a recovery buffer before landing.

The transferable idea is not “naps are good.” It is that fatigue is a predictable operational hazard, and that a structured, bounded rest can be built into a demanding task as a countermeasure. Whether that generalises to a hospital, a long-haul truck cab, or a control room is a reasonable question. It is not one this study, on twenty-one pilots, answers on its own.

For anyone tempted to treat the 26 minutes as a magic number, the guidance from fatigue researchers is duller and more honest: the right length, timing and depth of a nap vary by person and situation, and a nap that runs too long can leave someone worse off on waking than before they lay down.

What we keep coming back to is how much a single clean number can carry, and how little of the original work travels with it. The pilots, the transpacific legs, the descent where the benefit showed up, the sleep debt that stayed unpaid: all of it thins out into “34 per cent.”

The open question now is whether the American position moves. NASA’s fatigue group has continued studying controlled rest in real operations in recent years, and the gap between what the FAA concedes about napping and what it permits in the seat is the kind of thing that tends to get revisited after the next fatigue-related incident report, rather than before it.