This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
HMS Portland Lynx Flight
In December 2003 a Royal Navy Lynx helicopter operating at night from HMS PORTLAND crashed into the sea, all 4 crew members were sadly lost. On recovering the wreckage from the seabed air accident investigators quickly established that both the helicopters engines had been deliberately shut down in flight.
A subsequent Royal Navy Board of Inquiry concluded, The probable cause of the accident was departure from controlled flight following emergency action taken in response to an actual or perceived major aircraft malfunction. When flying at 100 feet 60 knots at night it is almost impossible to recover from a major malfunction that requires or causes the shutting down of both engines. After an extensive examination of the wreckage air accident investigators found no evident technical fault with the aircraft that could explain why the crew took the dramatic action of shutting down both engines. The actual or perceived major aircraft malfunction that might have necessitated stopping both engines in a helicopter are limited to the following:
Main transmission failure
Double engine failure
Tail rotor drive failure in flight
The aircraft did not transmit a tangible radio distress call and, as was common amongst military aircraft at the time, was not fitted with a flight data recorder or cockpit voice recorder. The technical investigation that followed the Royal Naval Board of Inquiry was tasked to look for evidence of any technical failure that might have been fundamental to the accident. Main transmission failure and engine malfunction or fire was discounted after careful examination of components leaving a tail rotor drive failure as the main focus of attention. A conventional helicopter requires a tail rotor to counteract the torque reaction between the main rotors and the fuselage. Without a functioning tail rotor the helicopter fuselage will spin uncontrollably.
Subject matter experts at the time of the accident were of the opinion that if a helicopter pilot had a tail rotor failure or control malfunction at low altitude over the sea he would have to react very quickly to prevent a catastrophe. To this end these particular helicopter crews had a set procedure for actions in the event of a tail rotor failure and these actions were rehearsed as part of every RN Lynx crew's pre flight preparation.
The means by which we learn and the way in which our memory works has been extensively researched. The purpose of this chapter is to establish the basic methods by which memory is created and subsequently recalled. A simplified explanation of the effects on memory and its recall and the potential errors produced when the memory is at fault or the recall of memory process is in error will be discussed as a potential tool for accident investigators.
The nature of accident investigation is such that when human error is present the means by which that error occurred will be thoroughly investigated; the human factors involved can be complex and open to many interpretations. By correlating the accepted theories of memory store and recall the author intends to provide accident investigators with a simple model to help establish human causal factors which might be attributable to poor or incorrect knowledge retrieval.
Memory can be defined as a process of information retention in which our experiences are stored and recovered when we need them. Popular literature proposes two distinct types of memory namely long-term memory and short-term memory or the modal model as defined by (Atkinson and Shiffrin, 1968). This model gives an explanation into how memories pass from the short-term to long-term by various mechanisms and how they can occur after a relatively short space of time.
(Hyden, 1961) explores the Biochemical Aspects of Brain Activity and provides in-depth explanations of how brain cells are activated and memory is created. At a neuronal level, changes connected with long-term memory seem to be permanent, with the formation of new synapses (junctions between cells) and protein synthesis, in contrast to short-term memory, where changes are only functional (BrainSkills, 2007).
(Atkinson and Shiffrin, 1968) propose a third memory store (actually the first in sequence) called sensory memory. The human information processing system gathers information through a number of different inputs. We then sort this information to form perceptions. The human brain has limited processing ability at higher levels so most of the incoming information is not processed immediately only pertinent information is dealt with and the remainder is retained temporarily in to buffer memory store or sensory memory. Some of this information can be recovered within a short period and processed to a limited degree despite not being heard when it was initially transmitted.
Figure 1 The model of human memory proposed by Atkinson & Shiffrin.
The memory system is further expanded through (Reason and Mycielska, 1982) who propose that the memory system is made up of four distinct stores: The Intentional Store, the Word Store, the Action Store, and the Recognition Store. There are likely to be more sub-categories present but these are unlikely to be relevant to this research.
The Intention store can be compared to short-term memory and actions stored here are subject to interference and can be easily miss-placed thereby leading to faults in the recall of memory. The comparison or subject of long-term memory can be linked to the other three stores namely Word, Action and Recognition. For something to be stored in these receptacles the language, skill or memory needs to be learnt and depending on the degree and success of that learning its ability to be recalled will be affected.
Reason describes an outline for a theory of action which simply explains the known properties of mental functioning. This model provides a good framework for explaining how different memory experiences are stored and reclaimed. In Reasons model the Intention System is central to organising memory functions with the Need System providing all the biological requirements. The Need System hooks directly into the Action, Word and Recognition Stores and in the context of this study can explain the occurrence of motivated errors. With this process in place a need is processed by the Intention Store - I must remember my mother's birthday today, which would equate to a short-term memory input or long-term memory equivalent in the word store for a conversation topic, Action Store - ride the bike home or Recognition Store - remember my aircraft pre-landing checks. Some memories are recorded with stronger permanence with only slight exposure to an event whilst other occasions only persist in the memory for a brief moment. The nervous system has a capacity to record key events easily while filtering extraneous occurrences or insignificant facts (Nielson et al., 1996).
Figure 2 Basic Components of the Theory of Action -(Reason and Mycielska, 1982)
Once the requirement has been identified and processed by the Action System the appropriate Output Function is selected and feedback is provided to the input function to complete the loop. The Shunt for closed-loop operation provides a handy explanation for cognitive errors. Open loop corresponds to skilful or habitual tasks i.e. riding a bike, which we complete with little or no conscious thought. A closed loop operation equates to a task, which requires constant supervision of our action and needs concentration and attention to detail on our part. When we delegate a task to open loop cogency it is susceptible to distraction and subsequent failure of a predetermined task - I've come upstairs to collect something but now I'm here I can't remember what I came for.
Long-term memory relies on the production of new proteins within the cellular body of the brain tissue. Transmitters, receptors, and new synapse pathways will then reinforce the bond between neurons. The production of new proteins for synapse strengthening is triggered after certain signalling substances are released for example calcium within the hippocampus neurons of the brain. In the case of hippocampus cells, this release is dependent upon the expulsion of magnesium (a binding molecule) that is expelled after significant and repetitive synaptic signalling. The temporary expulsion of magnesium frees N-methyl-D-asparate receptors to release calcium in the cell, a signal that leads to gene transcription and the construction of reinforcing proteins. (Niehoff, 2005)
Short-term Memory or "Working Memory"
The ability of short term memory to recover information is somewhat limited. (Miller, 1956) contains a popular article contesting that the span of immediate short-term memory is seven plus or minus two items of information. This information can only be remembered for a few seconds and is very susceptible to the slightest distraction causing the information to be forgotten (Eysenck and Keane, 2005). As soon as the information in short-term memory is no longer thought about for between 18-20 seconds depending on which theorist you select. If the information is efficiently rehearsed the decay is stalled by neural brain activity which in term stores the information in the brain.
Further research examined the possibility that newly formed memories develop slowly enabling the content to be altered by further experience over a period of time. Studies show that arousal of the memory rather than a chemical effect produce the enhanced capability or transfer into Long Term memory (Nielson and Powless, 2007).
Long-term memory is a more permanent memory store in which information is held even though the brain is not actively accessing it. Long-term memory has an infinite capacity and will last a lifetime. A popular theory for long-term memory or permanent memory states that brain recall involves a retrieval procedure involving a process of recognition and suitability of information organised according to its association with the brains current course of action. Lose of long-term memory is associated with the brains deterioration rather than its capacity to store information.
Common thinking supports the theory that everything we ever experience creates a long-term memory store. This is unlikely as much of what we experience is never focused on or concentrated on beyond a few brief moments and probably does not result in activation of the storage process. The ability to store items in long-term memory is central to the hypothesis in this thesis and will be examined in greater detail later.
Evidence that Short-term Memory and Long-term Memory are Physically Distinct Systems
Many theorists have used the multi-store model to explain the brains memory capability. It is widely accepted that this popular explanation is an over simplification of long-term and short-term memory and that the ability of the brain to store different forms of information is a not uniform or generic (Eysenck and Keane, 2005). Popular research purports that Long-Term Memory is made up of hard wired patterns within the brain and that once established cannot be erased and that Short-Term Memory consists of neural firing in the brain that are not permanently written to the brains hard-drive. It is also widely accepted that to place something into Long-Term Memory requires consolidation or rehearsal to verify and implant the information (Pribram, 1969).
Long Term Memory is capable of storing massive amounts of information therefore an efficient memory system must be capable of shedding unnecessary information whilst still being capable of recalling relevant and important facts. (Baddeley and Patterson, 1971) suggest that this selection process happens in two ways. Firstly information is filtered through a chain of buffer stores and the second by a series of processes including rehearsal as discussed in detail by (Atkinson and Shiffrin, 1968). The information is condensed into manageable chunks of information reducing the amount of information that needs to be recalled. (Atkinson) and (Baddeley) both theorise that the information stored or recoded by the second method both reduces the information that needs to be remembered and also aligns it to known memories by association. There is obvious scope here for the corruption of information and incorrect recall of fine detail of the memory itself.
(Tsang and Vidulich, 2002) demonstrates that situational awareness or mental workload can be categorised either by "black box " or "structural theory to show how psychology explains a pilot's cognitive process in the cockpit. Both theories draw heavily on existing research and provide a convenient summary of a specific memory function relating to how a pilot assimilates information provided by his senses and memory to produce a mental picture of the world around him. The following diagram illustrates the two theories of how Long-term, Short-term (working memory) and sensory input combine to produce a required output.
Figure 3 Theoretical concept of Situational Awareness (Tsang and Vidulich, 2002)
Both of these theories are applicable and make sense of the brain functions necessary to complete complicated tasks such as those being undertaken by the pilot of an aircraft.
(Beaty, 1995) postulated that learning could be grouped into three categories; 'conditioned response, rote and trial-and-error learning' more recent research (Mosier et al., 2001),(Skitka et al., 2000) clearly show that there are far more scientific and psychological approaches to this subject today. (Beaty, 1995) describes habit as a 'great flywheel'. Once set in motion our habits require little additional energy to keep them going and are hard to stop which aptly illustrates the next memory attribute to be explored.
A vast amount of literature exists about the psychology of memory bias. As is often the case with psychological subjects opinions and theories vary and for this category a number of explanations and ideas exist. The recollection of events is known to influence future decision making. People have been shown to bias remembered intentions prior to an event by their actual behaviour afterwards. There is a very strong human desire to show maturity and decisiveness when reflecting on our past decision making processes. (Pieters et al., 2006) research clearly shows actual behaviour influencing remembered intentions a similar effect has the potential to modify personal behaviour by pre loading behaviour planning with biases capable of influencing the decision making process.
(Bogacz et al., 2007) has shown that test subjects become more committed to a particular strategy as the frequency of their exposure to that strategy increases through reinforcement learning. (Kliegel and JÄGer, 2006) go on to demonstrate that people who are prone to prospective memory (PM) errors tend to use memory strategies such as imagery and rehearsal to enhance their memories. Another description of this is Habit Capture which will become a recurrent theme in this research. It is important to gain an understanding of the effect as a reference for further reading. (Freed and Remington, 1998) explain the effect simply and succinctly as follows: Habit captures occur when the model fails to allocate limited cognitive resources to retrieve task-relevant information from memory. Lacking the unretrieved information, decision mechanisms act in accordance with implicit default assumptions, resulting in error when relied upon assumptions prove incorrect. Continuous rehearsal of particular techniques is a long established technique to improve our ability to recall information from memory and has the added advantage of reducing error and susceptibility to stress or panic. Also referred to as "Overlearning" the technique implies training beyond the minimum acceptable level of performance, A commercial disadvantage to this technique is expense which provides very little benefit for normal operations(Hawkins and Orlady, 1993). Military pilots are routinely over taught routine exercises as they are less constrained by budgetary limits. The bi-product of this is an effective operational pilot taught to high levels of Crew Resource Management, handling ability, Situational Awareness and captaincy in a combat scenario becomes by default a highly competent crewmember. (Mauro et al., 2003) have developed an intuitive learning technique which employs modern psychological techniques by matching the students existing mental schema to optimise the learning process. In their research they point out that a real test of training may well be validated long after a specific course has completed by which time large chunks of the knowledge may no longer be retrievable. Again (Mauro et al., 2003) state that "things are remembered better if they are encountered multiple times in sessions that occur over a reasonable period of time instead of all at once".
Levelling and Sharpening of the memory store will also produce a bias towards a particular strategy this effect comes from a twisting of memory recollection over a period of time resulting in resilient parts of the memory being embellished or sharpened. Weaker elements lose emphasis or are levelled. Both effects become more enduring by repetition over time (Koriat and Goldsmith, 2000). With repeated exposure to the same effect a person will start to feel comfortable with the experience and is more likely to adopt the result associated with that knowledge. Recent work by (Fang et al., 2007) demonstrated this effect by comparing it to web site banner advertising. In the context of aviation a similar result can be assumed with continuous practice of an in flight drill, e.g. lowering of a helicopter collective lever in response to an abrupt noise.
The recency effect is a commonly held belief which has its own psychological definition. Put simply it describes the tendency for people to recall items from the end of a list opposed to items in the middle. This is a cognitive bias resulting from an unbalanced emphasis from a recent stimuli or rehearsal (Glanzer and Cunitz, 1966). The applicability to this research hinges on the prominence of a recently practised emergency drill and the increased or disproportionate weight it can receive when memory recall is exercised.
Recent findings have shown that testing seems to slow down the rate of forgetting, as well as increasing the initial strength of a memory (Karpicke and Roediger III, 2007). Those items at the beginning of a list are the easiest to recall, followed by the items near the end of a list; items in the middle are the least likely to be remembered. People tend to recall items from the end of a list more accurately than the beginning this form of memory bias is commonly referred to as the serial position effect.
To err is human; to forgive, divine
To err is human, to forgive divine (Pope et al., 1961) articulates that we are all human and we all make mistakes so we should all try harder to forgive other people when they make mistakes.(Faulkner, 1996) suggests that human error is present in 100% of accidents and the causal factors behind them. It is sometimes convenient for safety system managers to draw a line under the human element of a failure. For instance if a nut fails causing a control linkage in an aircraft to separate and the aircraft crashes, there must have been some element of human intervention in the manufacture of the nut, which may be causal. It may however be expedient for a safety manager to list the causal factor simply as a technical failure. The same does not normally apply to human factors, as there is a natural human tendency to apportion blame either to explain or vent ones anger at something other than an inanimate object.
Freud appears to have started the ball rolling in the study of errors or the Psychology of Errors, (Freud, 1922) Reason suggests, that Freud has already taken the choicest picking? then goes on to say that Freud's studies were aimed at the unconscious mind and that errors associated with it would give an insight into hidden thoughts and actions to assist with treatment or therapy of his patients. Fascinating as many of his findings are, they do not have any direct influence on this study of human error. (Reason and Mycielska, 1982) points out that Freud did suggest that these slips give every indication of being systematic rather than random in their origin. Minor everyday slips and lapses of memory are fairly commonplace, and can increase naturally with age, when ill, or when under stress, (Reason and Mycielska, 1982) go on to express a model for The Anatomy of a Cognitive Demon this description gives a convenient explanation for the effect being examined in this thesis. Reason argues that irrespective of the other influences upon it, the activation level of a cognitive demon at any moment will reflect the number of times it has been successfully used in the past. It is safe to assume that the same effect can be obtained by the frequent rehearsal of an action even if the action is not for real.
(Reason and Mycielska, 1982) also suggest that Word, Action and Recognition stores contain cognitive demons with the Intention Store (IS) containing only fragments of conscious experience which would be visible to our consciousness for a short period short-term memory and would be highly susceptible to contamination by other thoughts. The cognitive demons remain out of sight to our consciousness and need to be searched for and retrieved by a group of retrieval demons that scour the IS and Word Store and then present images to conscious for interpretation and action by the Action System.
The Action demons residing in the Action Store send instructions to the Output Function for routine functions. Reason argues that these demons are constantly struggling for recognition and that they range from being intensively active to dormant depending on their level of recent activity. The same applies to the Word and Recognition stores. The cognitive demons have a wide range of activation: The greater the level of use or frequency of access, the more likely a particular demon is to be activated into its specific mode of action.
A more scientific explanation is given by (Bogacz et al., 2007) who has expanded work on reward learning and decision-making to show that reinforcing actions are dependant on the interval between the learning.
There is a general public view that any accident or disaster involving industry such as aviation must have an element of human error. (Strauch, 2002)discuses 'situational awareness and decision making' in it he concludes that experts and novices will have different reactions to dynamic situations based on mental models produced from previous experience. Experienced operators in a variety of complex operating systems are better able to draw conclusions from available context information about their environment and can make faster and more accurate decisions by determining subtle system interactions. Novice operators can compensate in part for inexperience through rigorous training programs but currently the old adage of "no substitute for experience" remains true. Straunch believes that the two types of decision making processes classical and naturalistic can both be influenced by memory biases effecting the quality of the decision making process.
Evidence of accidents caused by memory errors
Air accidents and disasters are frequently attributed to pilot error this is discussed in detail in the monograph by (Bennett, 2001), (Reason and Mycielska, 1982), (Hurst, 1976) to name a few, I will confine my attention to a occurrences which cite memory rehearsal as a causal factor. The difficulty of assessing past accident reports by a single causation is that they can seriously skew findings as historically an accident report will only determine a single primary cause (Braithwaite, 2001). There is a growing movement amongst the mainstay of aircraft investigative bodies to include all the significant causal factors in all occurrence reports. The disadvantage for researchers is that historical data will lack relevance and as the incidence of aircraft accidents continues to drop past reports have to be re-examined to extract all the causal events in legacy findings.
The paradox of helicopter emergency training is discussed by (De Voogt and Van Doorn, 2007) where they identify that emergency procedures, particularly in helicopters, are not based on the frequency of recorded occurrences. There research showed that a number of helicopter emergencies are regularly practised when the occurrence rate is extremely rare and yet other emergencies are ignored even though they still happen.
A similar phenomenon is discussed by (Besnard and Cacitti, 2005) where they argue that an operator faced with a new piece of equipment or situation will revert to a previously learnt interaction modes and action. The assumption being that human cognition will strive to achieve a shortcut and effect savings in memory resources. The outcome can and frequently does result in an accident because the operator defaults to a familiar situation in a false belief that it is the correct response to the emergency. In fact the erroneous response now becomes a causal factor in the accident.
(Mosier et al., 2001) discuss the phenomenon described as "phantom-memory" when studying pilots coping with a suspect engine fire during simulator training. The majority of crews tested shut down an engine with a spurious engine fire indication despite their instinctive feeling that there was no fire present! The simulation was conducted in a modern airliner with high levels of automation in the cockpit.......
The categorisation of 'repeated rehearsal' as a causal factor is not listed in any research on accident causal factors. The CAA's Global Fatal Accident Review 1980-1996 (CAA, 1998) list's 'Omission of action/inappropriate action' as being the primary causal factor of 19.7% of fatal accidents. In the same report "Analysis of circumstantial factors" list 'Training Inadequate' in 40 (6.4%) of fatal accidents.
Boeing 747-4Q8, G-VHOT incident December 2006 at London Heathrow Airport the counter to this memory effect was described in (AAIB, 2008). Here accident investigators identified that when knowledge was not routinely rehearsed it would become 'dormant' and not easily recalled from memory. During take-off both stick shakers began to operate continuously shortly before the aircraft reached V1. The subsequent fault diagnosis required the crew to carry out one procedure with reference to the aircrafts Quick Reference Handbook (QRH) and another set of actions from memory. On this occasion because the recall procedure was not routinely practised the actions where not completed fully.
In an accident involving a Lockheed T-33, ex military fighter, the (AAIB, 2007b) report implies that the pilot transferred a skill set from a Boeing 747-400 using a typical pitch attitude on take-off as described in the Boeing Flight Crew Training Manual of 15 degrees nose whilst taking off in the T-33. The recommended nose up attitude in the T-33 is 5-6 degrees. Because the pilot was in-experienced on the aircraft type he reverted to his routinely practised pitch attitude from the 747. The aircraft settled back onto the runway and crashed in the overshoot.
During a visual approach to Sumburgh Airport, Scotland a Dornier 328-100 aircraft inadvertently flew in close proximity to the ground. Although the crew where alerted to the terrain by on board equipment the aircraft commander did not respond to the 'PULL UP' warning it generated.(AAIB, 2007a) The aircraft subsequently landed safely and the ensuing investigation highlighted a number of 'organisational, training and human factor issues' which were instrumental in the crew's failure to respond to the warnings. The associated causal factor to this research involved the training for the updated Ground Proximity Warning System (GPWS) which was new to both crew members. The crew had not been trained in the enhanced functions of the new system and both defaulted to previous training in a less sophisticated GPWS.
A classic example of an accident caused by 'negative-habit transfer' was that of an Royal Australian Air Force CT4 Airtrainer which crash landed 2nm south of RAAF Base Laverton following a simulated engine failure after takeoff (Cross, 2008). The accident happened during the students check ride prior to his second solo trip. During climb out from the runway the instructor retarded the throttle of the single engine trainer to simulate an engine failure after takeoff (EFATO). When the student opened the throttle to overshoot the engine failed to respond. The instructor took control of the aircraft and faced with very few options manoeuvred the aircraft to avoid trees, power lines and a farm house before landing heavily in a field sliding for 40 metres before hitting a fence and coming to rest. The uninjured crew escaped from the aircraft whilst fuel from ruptured fuel tanks spilt onto the ground. The 'main cause' of the accident became very apparent early in the investigation as both crew members confirmed that during the EFATO, the student inadvertently switched both ignition switches off. The student was reported as being 'particularly conscientious' with his pre-flight preparation, he would frequently use the CT4 cockpit mock-up to practise his checklist and procedures. Consequently he had become accustomed to actually manipulating switches and levers when practising emergency drills. When placed under pressure whilst practising the EFATO with his instructor his rehearsed action of actually performing the drill as opposed to simulating it came to the fore and he shutdown the engine. This behaviour happens when a well-learned and practised behaviour is applied in the correct situation but on an incorrect control. In this case the effect was magnified by the student's very low experience, setting the scene for this classic latent error.
Overview - Method 1 - Design
A study was initiated to demonstrate the effect of memory bias, evidence of reinforced learning and evidence of how an over exaggerated learning technique can have a potentially bad effect on memory training. By monitoring Royal Navy helicopter crews flying simulators data was obtained from 30 military crews flying simulated malfunction sorties.
The methodology for this research included two elements which are summarised below.
Standard operating procedures
Standard operating procedures are used extensively in the field of aviation to trap errors or Slips of the mind (Reason and Mycielska, 1982). The same can be said for any industry or activity where an accident can be catastrophic. (Green, 1996) contends that aviations enviable safety record is based around development of rule-based behaviours or standard operating procedures and the methods by which these procedures are taught and tested. A pilot will go through a standard sequence of checks before and after any major evolution in the cockpit. These checks will be confirmed by use of a check list or challenge and response with another crew member or computerised display. Pilots flying on their own will commit the check sequence to memory which in turn will be regularly checked by an examiner.
Method 2 - Participants
Anecdotal evidence from aircrew
A questionnaire was designed based on the literature review and personal observations. A number of experienced aircrew were asked for their experiences of lapses and errors during aircraft emergencies. Specifically aircrew were questioned about occurrences where a procedure had been incorrectly selected in response to an emergency and asked to offer an explanation as to why this could have happened.
A questionnaire (Appendix A) was designed to ascertain if there was any evidence of latent errors being caused by over rehearsal of emergency procedures. Participants were given the option of anonymity, many elected to freely recount experiences. Ethical clearance was not considered necessary for the study, as participation was completely voluntary.
Method 3 - Materials/apparatus
Royal Navy Lynx Helicopter Full Mission Simulator (FMS)
The Royal Navy operates a Lynx helicopter FMS at Royal Naval Air Station Yeovilton, Somerset. The FMS is a CAE Maxview simulator with 6-axis motion, high fidelity day night visuals. The FMS is full mission capable and is used as an elementary aircrew trainer for crew learning to fly the Lynx Mk8 and continuation training for front line crews.
The Royal Navy Lynx in its operational role is manned by a pilot and observer. The pilot is responsible for the handling of the aircraft whilst the observer manages the tactical employment. The crew sit beside each other and both have visibility of each others controls and instrumentation. However the flying controls are not duplicated so the observer has no physical means of flying the aircraft. The crews are usually teamed up into a permanent combination and great emphasis is placed on crew resource management with a number of general and operational tasks being shared. When dealing with an aircraft emergency the crew work as a team utilising well rehearsed drills to identify, diagnose and respond to an emergency. Utilising an FMS malfunctions sortie profile crews were monitored during a malfunctions sortie and an assessment was made to see if the crew was subject to any bias introduced during the pre sortie briefing. The malfunctions sortie given to the operational aircrew is a free play sortie with the simulator instructor at liberty to give any aircraft system failure to the crew. Typically the sortie lasts 1 hour with a pre-sortie briefing conducted 30 minutes before the session and following the sortie a comprehensive debrief is conducted by the simulator instructor with the crew. The instructor has the capability to pause the sortie at any stage to discuss or emphasise instructional points. Each Lynx operational crew is mandated to complete at least 1 FMS malfunction sortie every 6 months. With the ability to conduct any type of malfunction during this evolution there was little difficulty in adapting the scenarios to fit the requirements of this research. Because of the time between malfunction sorties the effect of prior knowledge of the research scenario was also reduced.
The trial was designed to examine the effect of memory bias and to determine if memory bias produced an undesirable error in the crew's diagnosis of a malfunction. It was also designed to record if an error occurred it was possible to effect a regain of know procedures and recover from the situation. It was also considered desirable to record any temporary transgressions indicating the presence of any bias introduced at the pre sortie briefing. It was not possible to examine repeated bias effects during this research due to safety constraints. Therefore the research concentrated on establishing if there was any bias effect however slight.
Figure 4 Royal Navy Lynx Mk8 Full Mission Simulator
A search was conducted to establish is any research had been conducted into specific operator errors and gain a better understanding of how to record human factor events and grade their magnitude. Aviation specific trials turned up a number of relevant type experiments but yielded very little in the way of appropriate designs to use as a format for this experiment. Medical research generated better results in the form of (Galvan et al., 2005) who conducted a study to gain a better understanding of safety during paediatric cardiac surgery. The methodology they used was designed to study a human team and organisational factors which could influence the outcomes of paediatric surgery. By studying the way in which the research team made observations during surgery it was possible to categorize
Method 3 - Procedure
The possibility of problems occurring due to the over priming of an over rehearsed emergency response were understood by the Royal Naval simulator instructors. The effect had been witnessed on a number of occasions but had not been documented. To ensure that behavioural reactions encountered during this research would not create problems for the aircrew at a later date careful planning and considerations were untaken by the researcher and the simulator staff. A number of potential scenarios were considered and tested by the simulator instructors. It was decided not to attempt to replicate the Portland accident due to the uniqueness of the malfunction and its potential to remain as a latent failure of memory. The exercise to be tested needed to contain a unique emergency that was rarely encountered with the possibility of being easily diagnosed as a frequently exercised drill. In the last 30 years the Lynx helicopter has had two incidents involving a malfunction considered to be the most appropriate research tool. This emergency involved a free power turbine failure and uncontained explosion. The explosion ruptured the firewall situated between the two engines allowing hot gases and fire debris to enter the opposite engine bay and setting off the associated engine bay fire warning. The net result was; a failed engine with no fire warning and confusing indications of a failure, plus a functioning engine producing a spurious fire warning. When taken in isolation these malfunctions are relatively simple for an experienced and well trained crew to diagnose and deal with typically resulting in a satisfactory and safe outcome. When tied together the malfunction becomes complex and requires extensive knowledge of the aircraft systems as well as a good understanding of the emergency drills to effect a safe outcome.
It was decided to restrict the research to aircrew that had already graduated from the Lynx training squadron and were current and operational. Crews under training were considered too vulnerable to contamination of their strict training regime and output.
To test for any bias effect half the crews were given to pre sortie indication of the type of malfunction they could expect during the simulator sortie. This was the control group. The remaining half was give a briefing and rehearsal of actions in the event of an engine fire. To test the hypothesis that when a bias event is introduced the recipient will default to that event if there is confusion and pressure applied to the individual. This will cause the memory to recall the learnt event in preference to the correct procedure thereby causing an error. In this scenario the recently rehearsed engine fire drill would be triggered by the spurious fire indication causing the good engine to be shut down. With the other engine already suffering a drive shaft failure this would lead to a catastrophic event. The control group on the other hand would not be biased towards the engine fire drill and be forced to analyse the emergency using the information before them and their knowledge of the aircraft systems.
The outcome of each test situation was assessed by the simulator instructor conducting the sortie. Prior to the start of the research project three simulator instructors where briefed on the content of the research and a standard briefing for both groups was agreed. The output from the proposed test case was standardised to ensure continuity of assessment. The author was also present during the sortie to arbitrate and ensure consistency of interpretation of the crew reaction to the simulated event.
Once the malfunction had been given to a nominated crew the reaction to the emergency was measured by the instructor to determine if each individual had analysed the event or simply reacted instinctively to it. This was a simple method to establish if the subject had been influenced by the pre sortie briefing or was using long term memory to retrieve the relevant procedure to effectively deal with the problem. If the subject went through a logical process calling on system knowledge to firstly determine the nature of the malfunction then select an appropriate procedure it was categorised as an analytical response. If however the subject responded to the immediate indication presented to him then continued without further analysis of the malfunction this was deemed to be a reactive response. All the sorties observed were conducted by a crew of 2 a pilot and observer. Each individual was assessed to determine whether their reaction was analytical or reactionary. The progress of the crew's reaction to the emergency was closely followed and any errors recorded. The ability of the crew to recover the situation once an error was made was also recorded. The outcome to the emergency was also noted although this research was designed to determine if there was any trace of memory bias in the simulation rather than log a fully developed fatal error.
The tabulated results of the research are depicted at table 1 and show quite convincingly that the crews loaded with a pre sortie briefing of an engine fire emergency readily adopted this emergency procedure when faced with an emergency displaying similar symptoms. The occasions when one crew member in isolation adopted the pre briefed emergency drill was also significant with these events demonstrating that with well rehearsed crew resource management
Method 4 - Analytic method (and reflective discussion)
Risky shift or majority rule has potential to influence collective decision making choices. Using the majority rule presents two types of influence: The first influence decides whether a group will use the majority rule or not and the second what they will use in stead (Zhang et al., 2006).
Limitations of this study
Methods of investigation including incident reporting and latent observation studies are hindered by hindsight bias and labour intensive information overload, respectively. Research in human factors is strengthened and more effective when using several methods of investigation. Limitations when interpreting these events include the possibility of missing events that occurred inside the cockpit not in the view of the observer.
Results 1 - Descriptive statistics (analysis of transcript, supporting quotations)
Results 2 - Inferential analysis (justification of analytic themes, discussion etc.,
Discussion 1 - Discuss results; place in context of research in introduction; critique the study; conclude (general discussion section)
Human information processing, error identification, triggering of actions and the selection of inappropriate output functions due to memory bias and habit capture.
School of Engineering
This thesis is submitted in partial fulfilment of the requirements for the degree of Master of Science
Cranfield University, 2008. All rights reserved. No part of this publication may be reproduced without the written permission of the copyright holder.
Experience, expertise, training and memory are seen as fundamental reasons why aviation has become as safe as it is. If a procedure is repeatedly practiced in preference to other drills it is possible that this rule-based behaviour will achieve dominance leading to a mental lapse. Research has been undertaken in an attempt to explain an aircraft accident, which shows no evidence of mechanical failure and a clear example of habit capture or source confusion. The effect of continuous rehearsal might lead to latent conditions prompting human error. This study reflects current research into the memory retrieval process the errors attributable to it and methods by which an accident investigator may use such research to explain causes and causal human factor occurrences.
Click here to enter a minimum of 5 keywords (not contained within the thesis title)
AAIB (2007a) AAIB Bulletin: 3/2007 TF-CSB.
AAIB (2007b) AAIB Bulletin: 12/2007 G-TBRD.
AAIB (2008) AAIB Bulletin 8/2008. AAIB.
ATKINSON, R. C. & SHIFFRIN, R. M. (1968) Human memory: A proposed system and its control processes. Spence, Kenneth W, ; Spence, Janet T.
BADDELEY, A. D. & PATTERSON, K. (1971) THE RELATION BETWEEN LONG-TERM AND SHORT-TERM MEMORY. Br. Med. Bull., 27, 237-242.
BEATY, D. (1995) The naked pilot : the human factor in aircraft accidents, Shrewsbury, Airlife.
BENNETT, S. (2001) Human error - by design?, Leicester, Perpetuity.
BESNARD, D. & CACITTI, L. (2005) Interface changes causing accidents. An empirical study of negative transfer. International Journal of Human-Computer Studies, 62, 105-125.
BOGACZ, R., MCCLURE, S. M., LI, J., COHEN, J. D. & MONTAGUE, P. R. (2007) Short-term memory traces for action bias in human reinforcement learning. Brain Research, 1153, 111-121.
BRAINSKILLS (2007) How Memory Works. IN MARSHALL, J. (Ed.) Amazing Brains. Cheshire, BrainSkills.co.uk.
BRAITHWAITE, G. R. (2001) Attitude or latitude? : Australian aviation safety, Aldershot, Ashgate.
CAA (1998) Global Fatal Accident Review 1980-1996 - CAP 681. IN GROUP, S. R. (Ed.). Civil Aviation Authority.
CAA (2003) Helicopter Tail Rotor Failures. IN GROUP, S. R. (Ed.). Civil Aviation Authority.
CROSS, P. (2008) Switch Hitter Hits Hard. DDAAFS and Aviation Safety Spotlight. Canberra ACT 2600, Directorate of Defence Aviation & Air Force Safety.
DE VOOGT, A. & VAN DOORN, R. (2007) The Paradox of Helicopter Emergency Training. International Journal of Aviation Psychology, 17, 265-274.
DEKKER, S. (2003) Illusions of Explanation:A Critical Essay on Error Classification. International Journal of Aviation Psychology, 13, 95-106.
EYSENCK, M. W. & KEANE, M. T. (2005) Cognitive psychology : a student's handbook, Hove, Psychology.
FANG, X., SINGH, S. & AHLUWALIA, R. (2007) An Examination of Different Explanations for the Mere Exposure Effect. Journal of Consumer Research, 34, 97-103.
FAULKNER, J. P. E. (1996) Evaluating crew resource management in ab initio pilot training - The first seventy hours. Applied aviation psychology - Achievement, change and challenge; Proceedings of the 3rd Australian Aviation Psychology Symposium, Manly, Australia; UNITED KINGDOM; Nov, 143-147.
FREED, M. & REMINGTON, A. (1998) A conceptual framework for predicting error in complex human-machine environments. IN GERNSBACKER, M. A. & DERRY, S. J. (Eds.) Proceedings of the 20th Annual Conference of the Cognitive Science Society.
FREUD, S. (1922) Introductory lectures on psycho-analysis, [S.l.], Allen & Unwin.
GALVAN, C., BACHA, E. A., MOHR, J. & BARACH, P. (2005) A human factors approach to understanding patient safety during pediatric cardiac surgery. Progress in Pediatric Cardiology, 20, 13-20.
GLANZER, M. & CUNITZ, A. R. (1966) Two storage mechanisms in free recall. Journal of Verbal Learning & Verbal Behavior, 5, 351.
GREEN, R. G. (1996) Human factors for pilots, Aldershot, Avebury Aviation.
HAWKINS, F. H. & ORLADY, H. W. (1993) Human factors in flight, Aldershot, Ashgate.
HURST, R. (1976) Pilot error : a professional study of contributory factors, London, Crosby Lockwood Staples.
HYDEN, H. (1961) Satellite cells in the nervous system. Scientific American, 205, 62.
KARPICKE, J. D. & ROEDIGER III, H. L. (2007) Expanding Retrieval Practice Promotes Short-Term Retention, but Equally Spaced Retrieval Enhances Long-Term Retention. Journal of Experimental Psychology: Learning, Memory, and Cognition, 33, 704-719.
KLIEGEL, M. & JÄGER, T. (2006) Can the Prospective and Retrospective Memory Questionnaire (PRMQ) Predict Actual Prospective Memory Performance? Current Psychology, 25, 182-191.
KORIAT, A. & GOLDSMITH, M. (2000) Toward a psychology of memory accuracy. Annual Review of Psychology, 51, 481.
MAURO, R., BARSHI, I., AMERICAN INSTITUTE OF AERONAUTICS AND, A., MEETING, A. A. S. & EXHIBIT (2003) Training smart using principles of cognitive science in aeronautical education and training, Reston, AIAA.
MILLER, G. A. (1956) The magic number seven, plus or minus two: Some hints on our capacity for processing information. Psychological Review, 63, 81-93.
MOSIER, K., SKITKA, L., DUNBAR, M. & MCDONNELL, L. (2001) Aircrews and Automation Bias: The Advantages of Teamwork? International Journal of Aviation Psychology, 11, 1-14.
NIEHOFF, D. (2005) The language of life : how cells communicate in health and disease, Washington, D.C. ; [Great Britain ], Joseph Henry Press.
NIELSON, K. A. & POWLESS, M. (2007) Positive and negative sources of emotional arousal enhance long-term word-list retention when induced as long as 30 min after learning. Neurobiology of Learning and Memory, 88, 40-47.
NIELSON, K. A., RADTKE, R. C. & JENSEN, R. A. (1996) Arousal-Induced Modulation of Memory Storage Processes in Humans. Neurobiology of Learning and Memory, 66, 133-142.
PIETERS, R., BAUMGARTNER, H. & BAGOZZI, R. (2006) Biased memory for prior decision making: Evidence from a longitudinal field study. Organizational Behavior and Human Decision Processes, 99, 34-48.
POPE, A., AUDRA, E., WILLIAMS, A. L. & BUTT, J. (1961) Pastoral poetry and An essay on criticism, London, Routledge, 1993.
PRIBRAM, K. H. (1969) Brain and behaviour ... Selected readings. Edited by K. H. Pribram, Harmondsworth: Penguin Books.
REASON, J. T. (1990) Human error, Cambridge, Cambridge University Press.
REASON, J. T. & MYCIELSKA, K. (1982) Absent-minded? : the psychology of mental lapses and everyday errors, Englewood Cliffs, N.J., Prentice-Hall.
SKITKA, L., MOSIER, K., BURDICK, M. & ROSENBLATT, B. (2000) Automation Bias and Errors: Are Crews Better Than Individuals? International Journal of Aviation Psychology, 10, 85-97.
STRAUCH, B. (2002) Investigating human error : incidents, accidents, and complex systems, Aldershot, Ashgate.
TSANG, P. S. & VIDULICH, M. A. (2002) Principles and practices of aviation psychology, Mahwah, NJ, Lawrence Erlbaum.
WIEGMANN, D. A. & SHAPPELL, S., A (2001) A Human Error Analysis of Commercial Aviation Accidents Using the Human Factors Analysis and Classification System (HFACS). Washington, D.C. 20591, U.S. Department of Transportation FAA, Office of Aviation Medicine.
ZHANG, J., HSEE, C. K. & XIAO, Z. (2006) The majority rule in individual decision making. Organizational Behavior and Human Decision Processes, 99, 102-111.
From this research it follows that there is little evidence of memory causing latent errors to emerge in aircraft accident data. This is unsurprising given that the move away from quoting one main cause to all accident reports is a recent phenomenon. An attempt to give historical data more depth by discovering all accident causal factors would be extremely difficult and not welcome as the trends identified in current statistics would become distorted and less relevant. Existing studies into the effects of rehearsal for successful memory clearly show that there is a substantial bias effect, whether that bias is good or bad is dependent on the situation or scenario the human subject finds it's self in. Clearly the mainstay of established aviation instruction is based around repeated rehearsal; this has proved a successful strategy from the Wright Brothers to the current day. More sophisticated techniques have been developed over the years and the plethora of psychological, educational and needs analysis studies have helped hone aviation instruction to a highly efficient degree.
Research by (Dekker, 2003) questions the accepted practise of forcing error-counting into specific categories. His detailed discussion shows how organisations are content to gauge their own safety by "counting errors". He express an opinion that this simplistic system appeals to managerial practices and intervention which happily produces bar charts showing how an organisation is recording the accepted criteria without accepting advances in the understanding of human factors and being able to adjust how their institution to influence or controls the human element.
Operators or in this case pilots are susceptible to two types of unsafe acts: errors and violations (Reason, 1990). Errors are defined as an action that fails to achieve its desired outcome, whilst a violation is an act of wilful disobedience or negligence. Assuming that violations are rare and the solution, once the transgression is discovered, relatively simple. The errors of interest in this research involve highly practised and automatic behaviours which are particularly susceptible to memory failures. (Wiegmann and Shappell, 2001) purports that most accident and incident reporting systems are designed and used by aircraft engineers. The data provided is highly analytical, clearly defined, easy to categorise and subsequent interventions highly effective. Traditional systems have little or no human factor analysis framework consequently the research is 'fad-driven' and the intervention is ineffective.
The CAA Safety Regulation Group commissioned research into the study of tail rotor failures in helicopters and the associated consequences. The research was conducted due to a previous study by the UK Tail Rotor Action Committee which concluded that the incidence of tail rotor failures was far greater than the required airworthiness design standard. The CAA report concluded that "regular focussed training is essential for the efficient application of emergency procedures." Being particularly pertinent to helicopter tail rotor failures where the correct and immediate diagnosis of the emergency is vital to a successful outcome. The report goes on to state that "inappropriate training could exacerbate the problems encountered during emergencies, particularly tail rotor failures" (CAA, 2003).
The scope of this study has been designed with an accident investigator in mind. Examining the human factors involved in learnt behaviour and errors in memory recall, then analysing the findings as possible accident causal factors.
The thesis will examine existing human science studies on how the human memory stores and then retrieves information. Given the vast amount of academic research available the intention is to correlate relevant information to provide accident investigators with a simplified source of key enablers to provide a source of potential causal factors in memory bias.
Research conducted to examine the phenomenon of habit capture and the effect it has on aircrew examined using military aircrew in a flight simulator. A control group assessed in an unusual emergency procedure will be compared to a group primed with a routine emergency drill before the sortie. The data will show that even small amounts of bias can have a disproportionate effect on the emergency outcome. Anecdotal evidence collected from experienced aircrew will be assessed to test the hypothesis outlined below.
Air Accident Investigation: Repeated rehearsal, the effect on human memory and the potential to trigger latent conditions for human error. An over emphasised emergency procedure may cause an operator to recall and select this action in preference to the correct one producing an error.
Experience, expertise, training and memory are seen as fundamental reasons why aviation has become as safe as it is today. If a procedure is repeatedly practiced in preference to other drills it is possible that this rule-based behaviour will achieve dominance leading to a mental lapse. This research is an attempt to explain causal factors into an aircraft accident, which failed to provide evidence of mechanical failure and potential of habit capture or source confusion or memory bias.
The long established aviation training precedence prevalent world-wide suggests that the results of this initial research would show little effect on aircrew performance due to a memory bias. Any effects demonstrated would be easily corrected by any capable instructor. Would there be a result