Features

August 2017 Issue




How We Forget

Humans can be creatures of habit, except when we’re interrupted and when confronting something unfamiliar.

Everyone occasionally forgets to perform some everyday task such as returning a phone call or depositing a check. But if the task is a flight-critical function, such as putting down the landing gear, surely no competent, conscientious pilot would forget, right? Well, no, even if it is a matter of life and death, pilots sometimes forget. For example, several airline catastrophes have occurred when a cockpit crew attempted to take off, forgetting to set flaps. Airline accidents have also occurred when cockpit crews forgot to lower the landing gear, turn on pitot heat, set hydraulic boost pumps to high, etc. The frequency of such oversights in the airline industry is much higher than the accident rate suggests because in most cases, the oversight is caught by a warning system in time for the crew to correct the situation. 

How should we think of the pilots who have an accident because they forgot to perform some flight-critical task? Do they lack competence? Were they not careful or conscientious enough? My NASA research team and I spent several years studying prospective memory: remembering—and sometimes forgetting—to do things we intend to do but at a later time. We concluded that all pilots are vulnerable to these memory lapses as a function of how our brains work and the specific circumstances of the flight. Experience and conscientiousness are of course essential to flight safety but are not, by themselves, adequate to prevent these memory lapses. Not surprisingly, accidents caused by these lapses are more common in personal airplanes, in part because GA pilots typically lack protections such as takeoff configuration warning systems and co-pilots.

Largely automatic

To explain why even good pilots are vulnerable to these memory lapses, I need to explain how underlying cognitive processes operate in typical flight tasks. This is only a high-level summary, but vectors to more information can be found in the box “For Further Reading” on page 6. 

Most prospective memory lapses involve routine, fairly simple tasks that we perform on every flight, which is not surprising since such tasks are the bulk of what we do in flying. When we have performed a task many times in the same manner, it becomes largely automatic; we do not have to think hard about how to do it or even when to do it, our body just executes the steps of the task when circumstances become appropriate. However, even automatic execution of a practiced task typically requires some degree of executive monitoring to make sure it plays out as expected. 

On June 11, 2013, this Cessna 172M crashed off the departure end of a runway in Louisville, Ky. The pilot was performing the fourth in a series of touch-and-go landings when he forgot to retract the flaps. All four aboard were seriously injured.

We initiate a task automatically when triggered by conditions requiring the task. A radio call from ATC prompts us to answer; the ATC call is a cue strongly associated in memory with the appropriate response. Getting ready to fly our airplane triggers us to perform a preflight inspection, a multi-step task. If this is an airplane we have flown many times and are current in, we do not have to stop and think which steps to perform—executing each step automatically triggers us to perform the next step. This is supported by the physical environment; as we move around the aircraft, we see visual cues reminding us where the next step is to be performed. Thus, automatic processing is heavily cue-driven, which is both its strength and its weakness. (Note that we should not allow ourselves to conduct a preflight in a purely automatic fashion, but rather should attend consciously to each item checked. Knowing what to do next is the automatic part.)

Automatic processing allows fluid, efficient performance and normally it is quite reliable, but it is vulnerable whenever the cueing process is disrupted. We may forget to act when the visual cues that normally trigger us to initiate a task are absent, when we are interrupted or distracted, when we intend to perform one task step out of its normal sequence or when we intend to substitute an atypical step for the habitual one.

Interruptions

Consider this scenario: You are halfway through your preflight and your passenger arrives, asking where to put his gear. You help him load up, but now you have increased vulnerability to forgetting to finish the preflight or to skip a step because the normal cueing process has been disrupted.

Let’s say you had just completed step 3 in the figure at right when your passenger walks in the hangar. Normally, completing step 3 would automatically trigger you to perform step 4, supported by the visual cues from where you are standing around the airplane. But after dealing with your passenger, the stimulus from performing step 3 is gone and you are no longer standing in a place to see the visual cues that remind you of what to do next. Obviously, if you stop to deliberately think about what you were doing and what to do next, you will remember, but it’s easy to move into your next step of a normal flow after loading the airplane instead of going back to the unfinished task.

Another scenario: Let’s imagine that most of your flying is done at non-towered airports, where you fly a typical VFR traffic pattern. Somewhat simplistically, we might represent the main steps of flying the pattern as: turn downwind, adjust throttle/speed, lower landing gear, set initial flaps, run downwind checks, turn base, clear for other aircraft on final, turn final, set final flaps and adjust throttle/speed as necessary. But let’s say that today you are approaching an unfamiliar towered airport and the controller tells you to make a long straight-in approach and to keep your speed up as much as possible. What does this do to your habit pattern? The straight-in approach removes the downwind visual cues that help trigger you to lower the landing gear. Keeping your speed up delays setting flaps, which in turn makes higher than normal demands on executive processing to manage speed on short final. In this situation, you are at increased risk of forgetting to set flaps or to lower the landing gear. 

Habit capture can occur when we intend to substitute an atypical procedural step for the habitual one but inadvertently revert to the habitual one under high workload. Let’s say that you frequently fly a standard instrument departure from your home airport that has you maintain runway heading, climb to 1000 feet, turn right to 320, etc. Today your clearance is to maintain runway heading, climb to 1000 feet, turn left to 090. You read the clearance back correctly, make your takeoff roll, transition to IMC, climb to 1000 feet, but—out of habit—turn right.

Of course, not all our intended actions are habitual. Suppose you are climbing out in IMC and ATC instructs you to report passing through 5000 feet. That won’t happen for several minutes and although it’s the simplest of tasks, it’s in fact easy to forget. To explain why, let’s look at how we do remember to act on this kind of intention when we actually succeed.

When We Remember

When we form an intention to perform a task at a later time, we cannot hold that task continuously in conscious awareness because our executive processes are busy in this case with hand-flying a somewhat complicated departure procedure. However, the intention to report at 5000 feet is not lost from our minds, but is stored in long-term memory (a vast reservoir of knowledge outside of conscious awareness until retrieved). 

Retrieving information from long-term memory (the intention to report) is a cue-based process—we notice something in the environment associated with that piece of information, triggering retrieval. (Of course, we also remember information when we conduct a deliberate search of our memory, but we must remember to do so at the proper time, which is the dilemma of prospective memory.)

Further, we must notice such a trigger cue at the moment, which in this example requires us to interleave monitoring altitude with other flight tasks. The process is probabilistic: The probability of successful retrieval is proportional to the strength of association between the cue and the stored information, and to how deeply we attend to the cue. When our workload is high, we may attend to cues less deeply, not monitor at the critical moment or inadvertently drop monitoring altogether. 

Countermeasures

The first step in reducing vulnerability to these sorts of memory lapses is to recognize that we are all inherently vulnerable, even when we work hard to be careful. Several specific techniques can help us reduce vulnerability.

When you need to remember to perform a deferred task that is not habitual, create what is called an implementation plan. Explicitly identify where and when you intend to perform that task; mentally identify exactly where you will be, what you will be doing, what cues will be in your environment when you intend to perform the task and visualize yourself performing that task. Research has shown that implementation plans can nearly double the chance of remembering to perform various kinds of tasks.

Checklists are a crucial defense against memory lapses, but be aware that checklists themselves are vulnerable to the same lapses as other tasks. If the cue that normally triggers us to start a checklist is for some reason absent, you may forget to initiate the checklist. If the checklist is interrupted, you may forget to resume it or may inadvertently skip an item.

Whenever you are interrupted—and you will be—pause before addressing the interruption to form an implementation plan. Create a reminder cue if practical. For example, if your checklist is interrupted, hold it in your hand instead of putting it down. Prominent cues are the most effective. For instance, if you have turned the fuel shutoff valve off for ground maintenance, tie a large ribbon on it as a reminder.

When multitasking, performing a procedural step out of sequence, or substituting an atypical procedural step for an habitual one, treat the situation as if it had a red caution flag. Form an implementation plan and create reminder cues. If other people are available, enlist them to help you remember.

Execute procedural tasks in a deliberate, systematic manner, which helps us notice visual cues and maintain the flow of cues and triggers that supports habit. Above all, avoid rushing, regardless of time pressure. Rushing at best saves a few seconds, and it increases our vulnerability to these and other types of errors enormously. 

Recognizing why we are all vulnerable to these sorts of memory lapses can help us develop our own reminder techniques specific to the kinds of aircraft we fly and the kinds of flying we do. I invite readers to share situations in which they have found themselves most vulnerable, countermeasures they have used and their perspectives on which things work best.

Sequence of Flight Tasks

Some flight tasks involve only one step, such as answering an ATC radio call, but many involve a sequence of steps, such as conducting a preflight inspection or flying a VFR traffic pattern. The figure at right is a generalized schematic of practiced multi-step tasks. Depending on how we define a step, each may have one or more substeps.