Checklists: The Aviation Habit That Works in Any Job

Every flight at our school starts the same way: a student at the wing with a laminated card, reading items they could recite in their sleep. It looks like a crutch for beginners. It is the opposite. The most experienced pilots alive do the same thing before every flight, and surgeons copied them because it keeps saving lives. Here is where the habit came from, why it works, and how to steal it.

The airplane that was too much to fly

On October 30, 1935, Boeing's Model 299 took off from Wright Field in Ohio for an Army evaluation flight. It was the most advanced bomber ever built, the prototype of what became the B-17, and the Army officer flying it was its chief of flight testing. It climbed, stalled and crashed.

The investigation found nothing wrong with the airplane. The crew had taken off with the control surfaces still locked, a step neither pilot had done, because there were now more steps than anyone could reliably hold in their head.

The response was not more training. Air Corps pilots wrote the steps down, for takeoff, flight and landing, and read them every time. The airplane the newspapers had called too much for one man to fly went on to one of the longest service lives in aviation. Every pilot checklist since descends from that card.

Ninety years later the FAA's flight training handbook still calls checklists the foundation of pilot standardization and flight deck safety, and warns that pilots who stop taking them seriously begin to rely on memory alone. We train in factory-new airplanes with a parachute built in, and the checklist is still the first safety system a student learns.

It is not for beginners

The common misunderstanding is that checklists are for people who do not know the job yet. The evidence says the reverse. They matter most for experts doing routine work, because that is exactly when the mind wanders.

Setting	Before	After	Source
Surgery, eight hospitals in eight cities	1.5% of patients died	0.8%	Haynes et al., New England Journal of Medicine, 2009
Same study, complications	11.0%	7.0%	Haynes et al., 2009
Intensive care, central line infections, 103 ICUs	2.7 per 1,000 catheter-days	0 at three months, with a reduction of up to 66% held for 18 months	Pronovost et al., New England Journal of Medicine, 2006

The surgical checklist is 19 items. Confirm the patient's identity. Confirm the site. Has the antibiotic been given. Does everyone know each other's names. Nothing on it is something a surgical team does not know. The deaths fell by nearly half anyway, because knowing and doing are different things at the end of a long day.

The ICU checklist was five steps for inserting a line: wash hands, clean the skin, drape the patient, wear sterile gear, cover the site. Doctors knew all five. The infections came from the one that got skipped when the room was busy.

Why they work

They catch the step that gets skipped, not the step that is hard. Nobody forgets the difficult part of their job. People forget the obvious part, under pressure, on the hundredth repetition. The 1935 crew did not fail at flying a bomber. They failed to unlock the controls.
They make the quiet person speak. The FAA's airline guidance has one crew member read each item and the other confirm it out loud. That gives the junior person a scripted moment to say "not done" without having to challenge anyone. The surgical checklist's introductions do the same job.
They free up attention. Holding twelve steps in your head costs something. A pilot who reads them off a card has that attention back for the radio, the weather and the student in the left seat.
They turn a near miss into a non-event. The 2008 Spanair crash in Madrid killed 154 people because the flaps were not set for takeoff and the warning system had failed. The investigation found the crew had the check on three separate checklists and did not properly use any of them. The checklist was there. The discipline was not.

How pilots actually use them

There are two ways to run a checklist and the difference matters.

Read and do. You read an item, you do it, you read the next. This is for things you do rarely or under stress: an engine failure, a new procedure, the first time in a new airplane.
Do, then confirm. You do the job from memory and habit, in a flow that follows the layout of the cockpit, then pick up the card and check that every item is done. This is how experienced pilots fly most of the time. The FAA handbook says it plainly: checklists need not be do-lists, and the emphasis is on the check in checklist.

A few things that happen in the airplane that translate directly to any job.

Checklists live at pause points. Before engine start, before takeoff, before landing. Not during. A checklist that interrupts the work gets abandoned. One that sits at a natural break gets used.
Memory items exist, and they are few. A handful of actions for a true emergency are committed to memory and done immediately, and the checklist is read afterward to confirm. The FAA handbook calls these immediate action items. There are rarely more than a few, because that is all anyone can reliably recall under stress.
The card is short. A before-takeoff checklist in a training airplane fits on half a laminated page. The full manual is hundreds of pages. The checklist is not the manual. It is the handful of things that kill you if missed.

How to write one that gets used

Atul Gawande's team at Harvard and Dan Boorman, who writes checklists for Boeing, published a one-page checklist for writing checklists. Its tests, slightly compressed:

Does it have a clear pause point where it is meant to be used, and does that match a natural break in the work?
Is every item a critical step that is in real danger of being missed, and not already caught some other way? If not, cut it. Gawande calls these killer items.
Fewer than ten items per pause point. Beyond that, people start skipping.
Does it fit on one page, in plain language, in the order the work actually happens?
Has it been tested by the people who will use it, in the real setting, and revised?

Then two rules from the cockpit that the medical versions kept:

Read it aloud, and have someone else confirm if there is someone else. A checklist read silently to yourself is better than nothing and worse than that.
Decide which kind it is. Read-and-do for the rare and the frightening. Do-then-confirm for the daily. Label it, so nobody treats a confirm list as instructions.

Where this works outside a cockpit

Shipping software. Before deploy: tests green, migration reviewed, rollback known, someone watching the dashboard. Five items, read at the pause point, has saved more weekends than any tool.
Closing a shift. Doors, alarms, cash, fridge temperatures, the one light that gets left on. Half a page by the back door.
Sending anything to a client. Right recipient, right attachment, numbers checked by someone who did not produce them, price matches the quote.
Leaving the house for a trip. Passport, chargers, medication, stove off, a neighbor with a key. The reason everyone forgets the charger is that packing has no pause point. Give it one.
Starting a lesson, a meeting, a procedure. Who is here, what is the goal, what would make us stop.

None of these need software. The 1935 card was paper. Ours still is.

When a checklist is the wrong tool

For judgment. A checklist cannot tell you whether to take off into that weather. It can only make sure you looked at it. Problems that need thinking need a decision, and a list of questions to ask yourself is a different document from a list of steps to confirm.
For things you do once. Writing a checklist costs an hour. It pays back on the tenth use, not the first.
When it gets long. A 40-item checklist is a procedure manual wearing a costume, and it will be skimmed. The FAA's airline guidance is blunt about the other failure too: a checklist is of no value if the crew is not committed to using it. Length is the fastest way to lose that commitment.
As a substitute for knowing the work. Pilots learn the flow first and the card confirms it. Reading steps you do not understand gets the step done and the meaning missed, which is how people tick "controls free and correct" without moving them.
When it replaces the conversation. The surgical checklist works partly because it forces a team to talk. A checklist done alone, in silence, to satisfy an audit, has kept the form and lost the function.

Try the real thing

If you want to see what a well-used checklist feels like in the hands of someone who does it every day, a discovery flight starts with one. You will walk around the airplane with an instructor and a card, and by the end of the walk you will understand why the most experienced pilots in the world still read theirs.

Questions

Why do pilots use checklists if they know the procedures?

Because knowing and doing come apart under routine and pressure. The 1935 crash that produced the first checklist was flown by the Army officer in charge of flight testing. Checklists catch the obvious step that gets skipped on the hundredth repetition, and the FAA's handbook warns that pilots who rely on memory alone become complacent.

Do checklists really improve outcomes?

In surgery, a 19-item checklist across eight hospitals cut deaths from 1.5% to 0.8% and complications from 11% to 7%. In 103 intensive care units a five-step line-insertion checklist took the infection rate from 2.7 per 1,000 catheter-days to zero within three months, a reduction held for 18 months.

How long should a checklist be?

Fewer than ten items per pause point, on one page, in the order the work happens. Longer lists get skimmed. The before-takeoff checklist in a training airplane fits on half a page.

What is the difference between a read-do and a do-confirm checklist?

Read-do means you read each item and perform it, used for rare or high-stress tasks like an emergency. Do-confirm means you perform the work from habit, then read the list to confirm nothing was missed, which is how experienced pilots fly most of the time. Decide which kind yours is and label it.

Where did the checklist come from?

From the crash of Boeing's Model 299, the B-17 prototype, at Wright Field on October 30, 1935. The crew took off with the controls locked. Rather than concluding the airplane was too complex to fly, Air Corps pilots wrote down the steps for each phase of flight and read them every time.

Sources

Last reviewed October 2026. The medical figures are from the published trials and have not changed; the FAA handbook and advisory circular are revised from time to time, so check the current edition.