Learning hub · Human Performance

Attention, memory and information processing

Everything you know about the aeroplane arrives through your senses and passes through a narrow bottleneck before you act on it. This lesson looks at that bottleneck: what you notice, how much you can hold in your head, how quickly you respond, and how a simple interruption can delete a step from a checklist.

About 15 minutes, plus the games and the quiz. Part of Human Performance.

1. The information-processing model

A common way to describe how a pilot works is as a chain of stages. Information comes in, is interpreted, is held briefly, is compared with what you know, and ends in an action. The action changes the situation, and the loop starts again.

  1. SensesEyes, ears, balance, touch
  2. PerceptionGiving the signal a meaning
  3. Working memoryWhat you hold for now
  4. Long-term memoryKnowledge and habits
  5. DecisionChoosing a response
  6. ActionHands, feet, voice

Perception is not a camera. The brain builds a picture from the signal plus what it expects to see, which is why the same engine note can be “normal” one day and “odd” the next. Working memory (also called short-term memory) holds a few items for a few seconds while you use them: a heading, a frequency, the next step. Long-term memory holds knowledge, procedures and habits for years, and it is what makes an experienced pilot fast. Attention controls what gets through the early stages. When attention is taken, the whole chain slows or breaks.

2. Attention: selective, divided and sustained

  • Selective attention is choosing one source and filtering out the rest, as when you pick your callsign out of a busy frequency. It works well, but whatever you filter out you do not notice.
  • Divided attention is sharing effort between two or more jobs. People cope when one job is practised and automatic, and badly when both need thought. Flying the aeroplane while doing mental arithmetic for a diversion is a divided-attention task.
  • Sustained attention is staying alert to something that rarely changes, such as scanning empty sky or watching a steady engine gauge.

When one task absorbs everything, attention becomes channelised: you fix on one thing and stop seeing the rest. It is sometimes called tunnel vision or fixation, although the word describes the mind, not the eyes. Typical examples are staring at a fuel-flow figure while the aeroplane drifts off heading, or trying to fix a radio problem while the altitude slowly changes. The remedy is a rule you apply deliberately: fly the aeroplane first, then look around the whole panel and outside at regular intervals.

3. How much can you hold? About four, not seven

You may meet the figure “seven plus or minus two”. It comes from a famous 1956 paper by George Miller, which noticed that people could repeat back about seven items in many tests. Later research, notably by Nelson Cowan, argued that when people cannot rehearse or regroup items, the real limit is nearer about four chunks, with estimates of three to five. Researchers still debate the exact number, and it depends on the task. For a pilot the honest summary is this: working memory is small, it fades in seconds, and you should never trust it with more than a few items.

A chunk is a unit that you already know as a whole. The digits 1 1 8 7 0 5 are six items to a stranger, but a pilot hears “118.705” as one frequency. That is why chunking works: grouping a clearance into a squawk, a pressure and a frequency turns fourteen loose digits into three or four familiar units. It is also why you write things down. A pencil and a kneeboard are working-memory extensions.

Try it: a made-up read-back appears for a few seconds. Memorise it and type it from memory. Play once with chunking off and once with it on. The numbers are invented and the game is a teaching illustration, not a measurement.

This game needs JavaScript. The idea in short: a long string of loose digits overloads working memory, while the same digits grouped into squawk, QNH and frequency are far easier to hold, and writing them down is easier still.

4. Expectation, confirmation and the radio

Because perception mixes signal and expectation, you tend to hear what you expect to hear. If you have flown the same circuit a hundred times, a clearance for runway 27 sounds like “runway 27” even when the controller said 09. This is expectation bias. Confirmation bias is its companion: once you hold a belief, such as “the weather will be fine” or “we are over the village on the chart”, you notice evidence that agrees and wave away evidence that does not.

The radio has a built-in defence called read-back and hear-back. You read back the safety-critical parts of a clearance (runway, QNH, squawk, headings, levels, frequencies where required). The controller hears your version and corrects it if it is wrong. That is only a defence if you actually say what you heard, rather than what you expected. If you were not sure, ask for the message again (“say again”). It costs a few seconds and is never a sign of weakness. Check your national and EASA rules and your instructor for exactly which items must be read back.

5. Reaction time: simple and choice

Reacting takes time, and the time depends on the job. In a lab, a healthy young adult responds to one expected signal in roughly 200 to 250 ms. This is a simple reaction: one signal, one response. Add a decision and the time grows: in a choice reaction, the time rises roughly with the logarithm of the number of possible responses (a result known as Hick’s law). Real events are slower still. A surprise you did not expect, that you must first notice, recognise and then decide on, often takes a second or more before your hands move. Tiredness, alcohol and distraction slow all of it, and the person affected is rarely the first to notice.

Try it: do the simple test, then the four-choice test, and compare your medians. Your results will be slower than a laboratory figure because screens, browsers and phones add delay. They are only a picture of the idea.

This test needs JavaScript. The idea in short: one expected signal gets a quick response of about a fifth of a second, while a choice between several responses takes noticeably longer.

6. Vigilance: watching when nothing happens

People are poor at watching for rare events. In experiments that go back to the 1940s, observers watching a display for faint, rare signals missed more of them as the session went on, with much of the fall occurring in the first half hour or so. This is the vigilance decrement. Cruising in good weather on a long, quiet leg, with nothing to do but look out, is exactly this situation. Boredom, warmth, a long day and monotony make it worse.

Defences are simple. Give yourself small jobs at intervals: a position fix, a fuel check, a scan of the instruments, a call to the next frequency. Keep a regular scan outside. Talk yourself through each task. Do not rely on willpower alone to stay sharp.

7. Interruptions, checklists and the missing step

Remembering to do something later, such as “set the flaps once I am on base”, is called prospective memory. It is fragile. Research from NASA Ames found that most memory failures described in airline safety reports were of this kind, and that checklists themselves can be defeated by interruptions and distractions. When something interrupts you in the middle of a job, such as a radio call, a passenger question or a traffic call, the brain’s bookmark often slips. When you return, you resume a little further along, and one step silently disappears.

  • Use a written checklist and follow it, rather than working from memory. Use the flow-then-check method your school teaches.
  • If you are interrupted, go back to a point you are sure about, or restart that checklist. Do not guess where you were.
  • Hold the checklist, or a finger on the line, until the step is finished. Move the card or mark the step to show it is done.
  • Treat a distraction on the ground, during taxi or on the approach as a warning that something may have been missed.
  • Under high workload, protect critical phases of flight: take-off, approach and landing. Say “stand by” to a non-urgent call, and fly the aeroplane first.

8. Try it: divided attention

This is a Stroop-style task. You see a colour word printed in coloured ink and must name the ink, not read the word. When the two disagree, reading gets in the way. Add the side task of counting a word and you feel attention being shared. The test is an illustration of the idea, not a measure of how you would fly.

This challenge needs JavaScript. The idea in short: when two things compete for the same attention, both get slower and less accurate, and a side task makes the effect larger.

Key points

  • Information flows from senses to perception, working memory and long-term memory, then to decision and action. Attention decides what gets through.
  • Working memory is small: the classic figure is seven plus or minus two, modern estimates are about four chunks. Chunk, write it down, and do not trust it for long.
  • Channelised attention is the tunnel: one problem fills your mind. Fly the aeroplane first and keep scanning.
  • You hear what you expect. Read back what you actually heard, and ask again if unsure.
  • Simple reactions are quick, choice and surprise reactions are slower, and fatigue and distraction slow both.
  • Vigilance falls on long, quiet tasks. Interruptions delete checklist steps: go back and re-check.

In the cockpit

  • Write down clearances and frequencies as they arrive, and read back from what you wrote.
  • Brief the radio call before you press the button, so working memory is not doing two jobs.
  • Do checklists at a pace where you can finish them. If you are interrupted, say so to yourself and go back.
  • On a long leg, set timed jobs: fix, fuel, scan, listen. Do not wait for something to happen.
  • If one problem is taking all your attention, say aloud “fly the aeroplane” and look at the attitude, heading and altitude.
  • This lesson is general education. For any question about your own health, attention or medication, ask an aeromedical examiner (AME).

Common mistakes

  • Reading back a clearance you expected, not the one you heard.
  • Resuming a checklist from where you think you were after an interruption.
  • Trusting memory for a long string of numbers because it is only a few seconds.
  • Treating a quiet cruise as a time to switch off.
  • Fixing a minor fault while the aeroplane flies itself away from the plan.

Check yourself

Five original practice questions. Answer them, press Check answers and read each explanation.

1. Which statement about working-memory capacity is the most accurate?

Miller’s seven plus or minus two is the classic figure. Later work, such as Cowan’s, suggests about four chunks, with debate about the exact number. Either way it is small and short-lived.

2. What does chunking do?

Grouping a clearance into a squawk, a QNH and a frequency turns many digits into a few familiar chunks. Writing it down is still wise.

3. You are told to expect runway 27, but the controller says 09 and you read back 27. Which effect is at work, and what is the defence?

You heard what you expected. Honest read-back lets the controller catch the error, and “say again” costs little.

4. Compared with a simple reaction, a choice reaction is usually:

A simple reaction needs no decision. In a choice reaction the time grows roughly with the logarithm of the number of options (Hick’s law).

5. A radio call interrupts you halfway through a checklist. What is the safest approach?

Interruptions make people skip a step without knowing it. Going back is cheap, and a missed step may not be.

Up next: decisions, stress, alcohol and medication. Back to the learning hub.