IT'S HUMAN NATURE!
By Loukia D. Loukopoulos
Based
on a forthcoming book on Concurrent Task Demands of Routine Flight Operations
by Loukia Loukopoulos, Key Dismukes, & Immanuel Barshi.
"As we pushed back, I noticed
a lengthy line of aircraft waiting for takeoff at our anticipated runway, which
was just a short distance behind us. I made a decision to just taxi the short
distance on one engine (engine #2). This break in our normal flow was
distractive enough that I didn't call for flaps. Ground Control then assigned
us a different, distant runway with more complicated than normal instructions.
... Still anticipating a wait at the end, I continued taxiing on one engine.
During the taxi, we continually evaluated the heavy rain showers we would
encounter on our departure... We stopped at the end of the parallel and Ground
sent us to Tower. Tower told us to pull up (#1 in line). We started the second
engine and with rollback, I started moving immediately to #1 (in line), fearing
delay might make ATC change their mind about us being next. Again, the break in
flow resulted in not calling for takeoff flaps. We continued scanning the
weather as we moved ahead and turned, running the checklist. The combination of
doing these things resulted in passing through the flaps item on the checklist
without confirming their position. We were cleared for departure and as I
pushed the throttle up we got 2 chirps from the takeoff warning horn."
(the crew aborted the takeoff
after the warning sytem alerted them to the fact that the flaps had not been
set for takeoff as required).
Aviation Safety Reporting System -
incident report # 519061
(edited narrative)
Why would a highly experienced airline captain forget to
ask the first officer to set flaps to takeoff position, a simple but crucial
command he must have routinely communicated thousands of times previously, as
required by standard operating procedures? And why would an experienced first
officer not notice that the captain did not call for him to execute this
crucial step he must have performed on every previous flight? Is this an
extraordinary, rare case of bad coincidences? Or is there something more to it?
Studies of ASRS incident reports similar to this one
(e.g., Dismukes, Young, and Sumwalt, 1998; Loukopoulos, Dismukes, & Barshi,
2001, 2003), and data from safety audits (e.g., Helmreich, Klinect, &
Wilhelm, 2001) point to the sobering fact that errors are an inevitable aspect
of routine, passenger-carrying flights. Fortunately most of these errors are
either inconsequential or are detected before harm is done, although crew error
is still widely cited as the most common "cause" of aircraft
accidents. Before you assume these errors to be manifestations of pilot
"carelessness," "complacency," or "inattention,"
however, consider that a growing body of scientific research indicates that
many errors of skilled experts are ... well, human nature! The source of these
errors? The context within which pilots operate and the multiple, concurrent,
challenging, and unpredictable demands placed on human cognitive processes. Our
research on cognition in the cockpit shows that such demands are an
under-recognized threat that renders even the most expert, well-trained,
conscientious pilots vulnerable to errors of forgetting to perform familiar and
habitual tasks. If we want to truly understand the source of a seemingly
careless omission, such as forgetting to set the flaps for takeoff, one must
look deeply at basic human cognitive skills and their interaction with the
challenges of the routine operating environment.
To this end I will analyze
the incident in the ASRS report above in terms of the cognitive demands of the
crew's multiple, unfolding tasks. But first here is some background readers who
are not pilots will need:
In an
ideal, perfect world, air carrier flight operations manuals instruct crews to
accomplish a series of actions comprising a large number of procedural steps in
a well-defined sequence during the taxi operation. Here I assume a 2-pilot
crew, a 2-engine aircraft and a flaps 5 takeoff, and for brevity only present a
high level description of actions:
After both engines have
been started at the ramp area, the captain calls "flaps 5, taxi clearance."
In response, the first officer positions the flap lever, visually confirms the
flap gauge indicating 5, and contacts the ground controller to request taxi
clearance (permission to taxi and exact routing instructions). When the
clearance has been granted, the captain releases the brakes and steers the
aircraft (with the flaps set to 5) along the instructed taxi route. The first
officer monitors the aircraft movement on the taxiways, performs procedures to
prepare the aircraft for takeoff and executes the Taxi checklist. One of the
items on this checklist requires the crew to check that the flaps are in the
takeoff position (in this case, 5 degrees). During the taxi phase, the pilots
remain in radio communication with the Ground controller. As they approach the
runway, they switch radio frequencies and expect permission for takeoff from
the Tower controller. As soon as this latter clearance has been issued, the
crew conducts a short Pretakeoff checklist and takes position on the runway.
The captain pushes the throttles forward and the aircraft proceeds to take off.
Our analysis of this
incident appears in plain face text after bold face text from the ASRS report:
"As we pushed back, I
noticed a lengthy line of aircraft waiting for takeoff at our anticipated
runway, which was just a short distance behind us.
A
captain typically listens to the Ground controller issue instructions to other
aircraft to stay informed about the conditions at the airport at the time of
pushback. The act of seeking and integrating information in order to maintain
situational awareness, however, increases the amount of workload and diverts
some of the captain's attention away from the main, ongoing task (monitor the
pushback, start the engines). I made a decision to just taxi the
short distance on one engine (engine #2). Taxiing on two engines is the default, well-practiced,
habitual procedure, however some air carriers allow captains to taxi on one
engine when concerned about fuel conservation. Taxiing on one engine requires deferring
starting the second engine later during taxi (i.e., out of the normal
sequence), and also requires that the Taxi checklist be deferred until the
second engine has been started (again, out of the normal sequence). Deferring
an activity requires having to remember to do it later. Pilots sometimes create
cues to remind them to perform a deferred action at the appropriate
time. When this action is important and highly habitual, however, they often do
not attempt to create cues because they assume it would be impossible to
forget. What they fail to recognize is that, under normal circumstances, they
remember to perform habitual actions because those actions are by definition
imbedded in a sequence of other actions they have practiced many times in the
same order. In the normal situation they do not have to think about what to do
next: performing one step of the sequence automatically retrieves the
next step out of memory. But when a habitual action is deferred out of its
normal sequence, the previous step no longer helps trigger retrieval of
the next step in the sequence from memory. This break in our normal
flow was distractive enough that I didn't call for flaps. The
deliberate rearrangement of activities (i.e., start the second engine later
than usual) disrupts the chain of events and actions that normally prompt the
captain to remember to call for flaps. Also at this point the captain appears
to be already thinking ahead, mentally noting the upcoming, long queue of
aircraft and planning a single engine taxi to save fuel. Although thinking
ahead is highly desirable, it competes for his attention, making it less likely
that he will notice that he has not automatically called for flaps at the usual
point. Having two crewmembers that back each other up is an important safeguard
against pilot error, so why does the first officer not notice that the captain
has not called for flaps? Again we suspect automaticity (executing highly
practiced skills with a minimum of conscious supervision), which is highly
reliable when actions are performed in normal sequence but much less reliable
when they are performed out of sequence or when normal cues are absent. Very
likely the first officer has come to unconsciously rely on the captain to call
for flaps to automatically trigger this habitual action. Lacking this habitual
trigger, the first officer does not retrieve this action from memory and does
not recognize that he has not performed the action. And perhaps the first
officer's attention is also diverted by the single engine taxi procedure.
Ground Control then assigned
us a different, distant runway with more complicated than normal
instructions... Changes
in taxi routing are not uncommon at a busy airport and in changing weather.
When they do occur, however, they require substantial attention from the crew
to work out the unplanned taxi route while in motion, often by referring to
charts if the route is complicated. Still
anticipating a wait at the end, I continued taxiing on one engine. The controller's unexpected instruction diverts yet
more of the two pilots' attention away from the ongoing task. The captain,
still thinking ahead of the long queue of aircraft ahead and focusing on the
repercussions on fuel consumption, considers the implications of the new
instructions only with respect to his prior decision to taxi on one engine.
This effectively draws his attention even further from the normal flow of
activities and thus reduces his opportunity to mentally review his actions and
perhaps to realize that he has forgotten to call for the flaps.
During the taxi, we
continually evaluated the heavy rain showers we would encounter on our
departure. We stopped at the end of the parallel and Ground sent us to Tower.
Tower told us to pull up (#1 in line). We started the second engine and with
rollback, I started moving immediately to #1, fearing delay might make ATC
change their mind about us being next. Again, the break in flow resulted in not
calling for takeoff flaps. Crews
are always looking out for factors that may affect the flight. In this
instance, keeping track of evolving weather conditions is yet one more factor
that requires the crew's attention. Both pilots must interleave thinking
about weather implications with other, ongoing activities. The crew's workload
is further compounded when the Tower controller asks for the aircraft to be
pulled up to the runway. The deferred action of starting the second engine is
suddenly brought to the foreground, as it must be performed immediately- -the
aircraft cannot be brought to the runway until both engines are running and
ready. The crew starts the second engine, aware of the fact that any delay on
their part can easily translate into a much longer delay if they lose their
place in the queue. Juggling the almost opposing priorities of taking the time
to carefully execute the critical activity of starting a jet engine and rushing
to retain position in line is challenging. In general, managing multiple, concurrent,
cognitively complex task demands is challenging and limits the amount of
attention devoted to any single task. More insidiously, when attention is
switched back and forth among several tasks, one task may drop out of awareness
altogether. As a result, neither pilot is prompted to think to check for an
action (setting the flaps) that they normally would have accomplished much
earlier and would not normally check at this later stage of the operation.
We continued scanning the
weather as we moved ahead and turned, running the checklist. The combination of
doing these things resulted in passing through the flaps item on the checklist
without confirming their position. Starting the second engine, because of its normal
association with running the Taxi checklist, acts as a trigger reminding
the crew of this second deferred activity, and they promptly perform the Taxi
checklist. Their attention, however, is already taxed by the multiple,
concurrent activities and the element of time-pressure. When visual
verification of the flaps position is called for by the Taxi checklist, neither
pilot notices that the flaps remain retracted. Expectation may play a
role here: because the flaps have normally been set considerably earlier, the
crew probably thinks they are set and are less likely to notice the discrepancy
if they do not take time to perform the check in a deliberate fashion that
requires mental effort. Checklists are a major safeguard, but because items
being checked are almost always in the desired state, pilots' responses to the
checklist items tend to become automatic (e.g., First Officer:
"Flaps"? Captain: Flaps 5") and may become de-coupled from the
action of visually verifying status of items checked. Note that this
automatization and de-coupling of the response happens without pilots
awareness, and effort is required to maintain the proper habit of checking
status of items in a controlled manner at a deliberate pace.
We were cleared for departure
and as I pushed the throttle up we got 2 chirps from the takeoff warning
horn." When the Captain
pushes the throttles forward to begin the takeoff roll, the warning system
indicates the inappropriate flaps setting and prevents an impending disaster.
I recognize that my account
is speculative and that in the interest of brevity I have only sketched out the complex cognitive
processes involved. However our
research group believes that analyzing these seemingly simple, preventable,
errors in light of the complexities of the real operating environment and the
cognitive skills and vulnerabilities of human operators helps open up a new
perspective on pilot error. The cognitive complexities at play in the cockpit
situation (e.g., deferring actions, maintaining and remembering intentions,
managing concurrent task demands, interrupting and rescheduling activities) are
issues we are actively researching. Using this forum we hope to stimulate
discussion and interaction between the operational community and the scientific
community; from this interaction may grow deeper insight and perhaps ways to
improve safety.
Pilots: what similar
situations have you encountered? Do you have personal techniques to reduce your
vulnerability to errors of omission? How effective and practical to use do you
find them?
Scientists: Does this
characterization of cognitive underpinnings of errors of omission make sense to
you? Can you suggest other aspects that should be considered?
Others: Have you
observed similar phenomena in other areas of human performance?
References
Helmreich,
R.L., Klinect, J.R., & Wilhelm, J.A. (2001). System safety and threat and
error management: The line operations safety audit (LOSA). In R. Jensen (Ed.), Proceedings of the 11th International
Symposium on Aviation Psychology. Columbus, OH: Ohio State University.
Dismukes, R. K., Young, G. E., & Sumwalt, R. L. (1998). Cockpit interruptions and distractions: Effective management requires a careful balancing act. ASRS Directline, 10, 4-9.
Loukopoulos, L.D., Dismukes, R. K., & Barshi, I. (2001). Cockpit interruptions and distractions: A line observation study. In R. Jensen (Ed.), Proceedings of the 11th International Symposium on Aviation Psychology. Columbus, OH: Ohio State University.
Loukopoulos, L.D., Dismukes, R. K., & Barshi, I. (2003). Concurrent task demands in the cockpit: Challenges and vulnerabilites in routine flight operations. In R. Jensen (Ed.), Proceedings of the 12th International Symposium on Aviation Psychology (pp. 737-742). Dayton, OH: The Wright State University.
If you would
like to comment on this article,
please click here .
We invite your thoughtful feedback and comments, and we encourage you to
submit opinions, anecdotes, personal experiences, and ideas that relate to this
article. The author of each article will compile a summary of the responses
received and post it before the next hot topic is presented.