Thanks again to all for the great discussion.

First off, no I do not think the purpose of CRM is to create a warm and fluffy environment. I do agree with you that it should be a RESOURCE MANAGEMENT tool. To clarify, when in the military I participated in critical incident response management; these teams were convened only when serious threat/actual harm events occurred. Usually there had already been one or more deaths, and occasionally we were in the position where no matter the course of action selected, more would be harmed or killed. (Talk about resource management!) As I indicated, in these situations what is not known is as or more important than what is known; decisions are very time pressured and information is limited at best. In these situations (as in many flight events), it was often the lowest-ranking personnel who held key knowledge and it was imperative for our success that these communications be made. So it was in this environment that I observed/acquired leadership behaviors; my commanders had an impressive combination of low power distance and decisiveness. So I look at CRM programs/aircrew behavior through this lens.

I’ve been quiet the last few days because TC (and later others) brought up two words that got me thinking: measurement and compliance. First measurement: as my peers and I explore the social, behavioral and cognitive aspects of aviation, we keep returning to the same question: what are the traits/behaviors/definition of a good pilot/crew and how can they be measured? I was sitting in the class partly to clarify this, but all I heard was very squishy. Any thoughts? Especially you, TC, how would you (if your customers would let you) define and measure CRM objectives and outcomes?

measuring crm
 

just picking up on something BOAC said, the UK RAeS HF CRM WG (phew) did discuss a few years back, at the behest of the CAA, the possibility of CRM for post-holders - so the recognition is there, just not the mechanism, yet.

As for measuring, from the little work that has been done, we know that current attempts to assess CRM are little better than random. However, the problem is that we lack a clearly-defined set of observable behaviours that are linked to safe performance; we lack a robust set of course design skills across the industry that allow facilitators to develop meaningful training events that change behaviour and, finally, we lack a culture within most organizations that recognise the frailty of human behaviour. To wax lyrical, while Rome burns, we are simply wondering if Nero's fiddle is in tune.

So, first, we need the right atmosphere in an organization. Metric 1 would be a measure of staff attitudes. Next, we need a more realistic and valid training curriculum. Metrics 2 and 3 would be training transfer and organizational benefit from CRM. Finally, we need better tools to evaluate performance in training situations in order to provide developmental feedback. Metric 4.

Problem. I did, once, run a fairly substantial attempt to look at training transfer in a big flag carrier. We surveyed nearly 5000 staff across the airline pre-training. We then did a repeat measure after passing an agreed milestone in terms of course roll-out and again a year later. It was logistic nightmare. What did we find? Well, quite a bit about attitudes to safety in different populations within the airline. We saw the impact of things like disputes between management and specific groups(pilots). Did I demonstrate that my course was the best thing ever? Well, the happy sheets were great. But was there a return on the investment? Hmmmmm...

We can all agree that it is the right thing to ask if CRM works and people have been doing that for years. The apparent failure of CRM spawned TEM. But getting data to demonstrate clear benefit is a problem. In my book on CRM Course Development (ouch! hideous commercial exploitation!) I cite some work done in Australia. The authors commented that "evaluation of training is not difficult; it's just not easy''

I was hoping for something a little more specific than “attitudes”….

I gave it some thought last night when I was out on trail… one thing that may have made it easy for the young airmen in my military example to break through was there was a certain amount of information management structure in place. In these situations we were concerned with change from OPSNormal (okay, if we were even talking we were well outside OPSNormal): values that were lower/higher than normal (threshold data), values that were fluctuating rapidly (abnormal rate data), and especially absent, intermittent or re-established communications. Personnel in the field were required to report when events exceeded certain pre-determined criteria; these reports were assessed and added to the ‘big picture’ with regular reports (or exception reports, as necessary) then forwarded to HHQ. During these events it was the DATA that was important, not the rank of the person delivering it, or the person receiving it. I think part of what made our ops so effective was the recognition that if certain conditions were in place, they needed to be reported and if these communications were present, they were important and evaluation/action needed to be taken. Also, certain events had to be reported to HHQ regardless of whether the CC thought they were of importance or not. It has me wondering how a similar system could be adapted to flight ops.

I’ve gotten off on this tangent because of this: During my flight training I was taught that though I have responsibility for the flight, I am not the most important person on the aircraft. Rather it is the passengers (or payload) was the reason I and my flight team-mates were there, and that each team member (fueler, load, maintenance, ATC, etc.) was there to help me deliver them safely. But in the CRM class I attended recently was there seemed to be a caste structure in play; pilots were of supreme importance, with cabin crew, dispatch, baggage handlers, fuelers, etc viewed as second-class citizens to be tolerated as the flight progressed. The students came into the class with these attitudes, and I watched as they were re-enforced and became entreanched over the course of a few weeks. For some reason I thought the complexity of the system had forced us to acknowledge our interconnectedness so it disturbed me to see these (caste-based) perceptions transferred to another generation.

On to measurement: yes, we do need to measure the whole organization, but what exactly? “Attitudes” seems so general and non-descriptive, what traits exactly are we looking for? Looking back, the commanders I worked for were decisive, but they also demonstrated creativity, flexibility, humility, patience and a sense of justice and much more. I think what separated the best from the average was that the high-performers remained present, not superficially taking things in but truly paying attention to the details of what was going on and recognizing how they fit together. (One of the best leaders I ever worked for recognized me in a civilian setting when he had previously only heard my voice over a landline.) So which of these and other traits should we measure? Cultivate? Seek to extinguish? How do I quantify “awareness”? “Decision-making”? “common sense”? “elemental forces”? And what outcomes are we seeking with these measurements? Reduced number of accidents? Deviations? Or is there something we can measure on a day-to-day basis?

Thoughts please.

PS -
BOAC: The posts I referenced earlier are in the “Border between being assertive or arrogant/rude” thread under the CRM/Safety header, posts 30 and 44 if I recall correctly.

Thanks JG - I have looked at the thread and the posts. Granted that the 'Command management' was not apparently effective, they certainly did trigger some harmonics in following posts.

There is, as said on that thread, a danger that if CRM goes too 'warm and fluffy' (your point noted;)) it has the risk of engendering such attitudes in F/Os. Basically as with any 'rank' structure, it is encumbent on the subordinate to obey the orders of the superior, and for the superior to issue orders quote "Clear, concise and capable of being understood.' unquote (British military 'gospel'). The way you and I like things is for inputs as you describe.

I would suggest that wherever you experienced the 'CRM training' you described in para 3 of Post #23, that organisation should change their training system PDQ. I am not accustomed to the encouragement of the scenario of "second-class citizens to be tolerated as the flight progressed" you describe, although I have come across it in some individuals. However, the F/O and the rest of the crew MUST obey the Captain's commands with the proviso that he/she/they can challenge them if they are or appear dangerous, question them if they seem uneconomic or 'outwith' company procedures, or contradictory to good service delivery and of course ask questions at any time about them if they are not fully understood. That is what 'Command' means.

In the post #30 situation I would have simply taken control and discussed it on the ground. Post #44 MAY have fitted into the 'worth a learning curve' if it was not affecting safety. Always the option, as SNS3Guppy said, to over-ride. As always, almost impossible to draw real conclusions from one-side posted views.

I think comparing your military experience is not particularly relevant in our world. My experience has always been that we are swimming in a totally different fish tank, in which delegation and taking advice by a 'senior' can easily be interpreted by others as a weakness or 'inability to do the job'. Back to CompanyRM perhaps....? My common comment is that in my military (piloting) days, our 'managers' were actually trained as leaders, which faculty is so often sadly missing in this world and there is NO incentive to attempt to train our managers as leaders. I sometimes say to my F/O's how in the military you simply do not 'put on' respect in the morning when you don your 'respect' stripes, and also that in my miltary experience I followed my 'managers' out of respect and not as now sometimes to see when or how badly they were going to screw things up.:).

Interesting stuff and good to see your interest.

“…the UK RAeS HF CRM WG did discuss a few years back, at the behest of the CAA, the possibility of CRM for post-holders - so the recognition is there, just not the mechanism, yet.” (#22)
A problem here; the responsibility for action in Europe is now with EASA, who may not have the same interest in RAeS. Furthermore, if EASA’s attitude towards HF is indicated by the vacancy listing EASA/AD/2008/046 for a ‘Human Factors Expert’ we might be very disappointed – it’s a part time job! (application now closed)

”The apparent failure of CRM spawned TEM. But getting data to demonstrate clear benefit is a problem.”
The apparent failure could be debatable, but this should not detract from the need to improve aspects of human performance.
It’s possible that the origins of CRM stemmed from an academic view of Airmanship, and the researches still had difficulty in defining the constituents and the training required. There were similar problems in defining the means of validating CRM, instead they look to the overall safety statistics. CRM was a convenient and refreshing acronym; we like to focus on new and shiny aspects, so it caught on.

The problem with the originating CRM concept, like airmanship, was that the means of teaching and assessing it were glossed over… similarly nowadays with TEM.
TEM is a rehash of an aspect of basic human behaviour, but has been launched on the industry as a refresh of CRM rather than a replacement – another ‘new and shiny’ aspect of safety training.

If we return to Airmanship to seek measures of behavioural traits, we might similarly be disappointed. At least one reference comments that we should not judge the outcome of an observable event, instead seek to understand the process being used; what, how, or why something is being “thought”.

“How do I quantify “awareness”? “Decision-making”? “common sense”? “elemental forces”? And what outcomes are we seeking with these measurements? Reduced number of accidents? Deviations? Or is there something we can measure on a day-to-day basis?”
How do we quantify a thinking processes; if the academics cannot provide answers or agree on a theme, then perhaps we should revert to intuition … isn’t that similar to the old-style judgement of crew performance associated with Airmanship?

Reason for the failure of CRM.
Airmanship training for modern aircrew.
Critical Thinking For The Military Professional.

attitudes
 

Attitudes have been described as 'proto-behaviours' in that holding an attitude towards a object in your world might indicate a disposition to act in a particular way. You only have to look at attempts to influence the attitudes of particular groups of voters in the recent US election. Why was Sarah Palin chosen as McCain's running mate? Because she was the best person for the job and would be able to step in should anything happen to the President? Possibly. But more likely, they gambled that in so doing they would change attitudes in certain voters and, as a result, change voting behaviour.

However, there is a huge gap between declared beliefs about something and then actually taking action. Look at smoking behaviour ( and set aside addictive properties). Lots of people declare their intention to quit smoking, knowing the harmful effects, but many of these still smoke.

In our work we came across some data for maintenance audit pick-ups out on the ramp. A quality manager had noticed a pattern. If a shift manager failed to attend the morning briefing, then his shift was usually significantly over-represented in quality failures. So, now I have a measure - quality failures - and an action by an individual. The reason given for late showing was usually traffic. However, I could investigate attitudes. So, I can ask about the importance of attending morning brief, the value of the brief, its usefulness etc. Although traffic is my reason for no-show, the actual explanation might lie in my attitude to the morning briefing ritual.

Our project was to improve the effectiveness of teamwork. We hypothesized that effective teams worked efficiently. This might be reflected in attitudes towards workload, cooperation etc. So, I could ask questions such as 'we always seem to be fighting one another'. 'I'm always having to pick up after other people', 'My day always goes according to plan', 'I am always thinking about the needs of others'. Our measure of efficiency might be minutes of delay. Through teamwork training, hopefully, attitudes would show a positive shift between pre- and post measures and this would be reflected in a reduction in lost time due to delay.

The problem with attitudes is finding the correct dependent variable and to accurately frame the attitude measure. But that doesn't mean that it cannot be done.

I'm not in a position to discuss the traits of leaders you have worked for JG; I don't know them. The old trait approach, though, has fallen out of favour as we have tried to move to trainability rather than emergence as ways of developing leaders. But one little story. When I took command of a field unit, my predecessor kept an album of all 250 staff with pics, personal details etc. He liked to be able to ask after wives and children by name as he moved around the unit. The guys were always impressed that he seemed to know children's birthdays. He left me a ton of operational problems to deal with but the guys loved his little party trick.

Alf - I'm happy to state categorically that LOSA/TEM is the product of an initiative that was prompted by a dissatisfaction with CRM training which had been deemed to have failed to deliver. Got it from the horses mouth. I personally didn't say that CRM had failed although I do question it's effectiveness - which is why I advocate attempting to measure, one such way being outlined above.

BOAC -
(Please note I am asking this out of curiousity and a desire to better understand your point of view)
What characteristics of the flight deck make it different than the war room? Isn't leadership leadership, no matter where it is applied?

TC-
(again, curiousity)
I haven't had a chance to fully go through your post yet. Also, I have one of your books on request at the library so i am behind that power curve as well.
But it stikes me as interesting you use a maintenance reference in response to a flight deck question. But this really gets at what I am interested in: How do we define desired performance? an optimum flight? What is the flight deck equivalent of a quality failure? What other behaviors should we be looking at? Curiosity? Persistance at task? Creativity? Compliance? (That word again, more on that later...)

I appreciate your patience with my young, impressionable mind.

-Jolly

JG

The maintenance example was useful to illustrate the concept. Alas, the pilot one I could have used, while relevant, did not show the pilot cadre in a good light. We asked groups to suggest ways that their non-specialist collaborative team members (cabin crew, maintenance, dispatch etc) could act in ways that made their jobs easier (i.e. team work across organizational boundaries). All groups except the pilots offered task-related activities, 'inform me of changes', 'communicate early', 'give me as much time as possible to meet your needs'. The pilots proposals were all ego-related 'treat me with respect', 'bring me things', 'do non-task related things for me to make my life easier'.

However, it does go to something you mentioned earlier. You said that your passengers (or payload) are the important thing and everyone else acts to support you to meet the needs of the customer. This still places you at the top of the hierarchy. However, the tasks of the other team members are actually equally directed at the goal of moving payload (pulsed and/or pulse-free). They are not there to help you; you are there to work with them to meet a collaborative goal. I'm sure in your case it was slip of the tongue, but in the case of the pilot group I studied it was a strongly-held attitude which was translated into observable behaviour. Importantly, I can measure that attitude, I can observe manifestations of it and I can evaluate it's impact on overall team effectiveness.

I didn't see yours as a 'flight deck' question (although that is what you clearly asked). I saw it as a 'behaviour in critical teams' question. I can point to exactly the same issues with consultants in hospitals, midwives, dispatcher/ops in some bizjets, specific nationalities in multi-cultural organizations. One of the reasons CRM fails to deliver is because we have framed the problem in a pilot-centric manner. Even though we have recognised that the same skills apply to other groups (maintenance, ATC, cabin) we have actually done little more that try to invent a different acronym (HF, TRM). What we haven't done is step back and look at the bigger picture. TEM is no different.

Primarily, CRM is about accommodating different personalities in a fast moving claustrophobic space for long hours with the objective of arriving alive.

You can be a philosopher, but it's pretty hard to get into people's minds, so I don't think that individual behavior is capable of being made into a science.

CRM is about ensuring the collective adherence to SOPs, common sense and basic survival instinct. . . . Because you may be having to share this claustrophobic space with people who chew with their mouths open, with people who you would never associate with on the street, with people who have nothing in common with you, except that of being a pilot. In this worst case scenario, it's only about tolerating each other for 14 hours from LAX to HKG, because we all want to be save and to live another day.

On the other hand, CRM is of less importance when one is flying with people whose flying skills are known and whose interests and personalities are compatible. But rostering does not include personality traits or individual preferences when pairing crews.

CRM is like a speed limit on a road; it's to protect the worst driver, driving the worst car under the worst conditions.

Quality Failure
 

JG, I noticed that I didn't cover some of your questions. By 'quality failure' I am referring to any performance that does not meet a defined specification. In the case of maintenance we used things like sticking to cooling times before replenishment, using the correct grades of lubricant, making sure lifed items were checked.

LOSA is only a flight deck quality audit so that would show up some things that might be used as measures.

Curiosity, persistence, creativity and compliance are all characteristics that have a place in critical-task management. First, though, I'd like to see the terms elaborated as a set of observable, trainable behaviours (accepting that some component of each might well be genetic and then distributed variably across the population). I'd also like to see the risks attached to each cluster identified. Compliance, especially, is not universally 'good'.

- I have no knowledge of what a 'war room' is, so I cannot say. If you refer to an 'Ops Room' then I see quite different requirements to the flight deck in terms of time to take inputs and input options. Most 'Ops Rooms' have a significant amount of info coming in and resources to tap. 9/10 in the flight deck one is left with only the resources on board and quite limited info from outside. It would indeed be good to be able to 'stop our involvement in the war', knowing that others will keep on running it for us and have a quick conference about things, take inputs from lots of people etc, but we have a different time frame most of the time, and a lot of the time our 'war room' is humming along at 500mph+ and running out of options as it does. As I said in my penultimate paragraph, it is a different world, I think and I would be wary of the drawing of too close comparisons and procedures. Perhaps you could explain the 'war room'?

Sorry for the silence. I had a project due today.

BOAC –
I am not familiar with the term “Ops Room,” other than as the flight-following desk at a squadron. The best I can explain it (other than the description in my original post) is this: While you are triaging/managing your unprogrammed compound emergency in the air, we are the team on the ground providing you support. It is up to us to get you access to data you need, not only providing real-time access to in-house maintenance (sub-system experts) or even the component/platform designer/team at the manufacturer if necessary, but also managing the ground response – what equipment/capabilities are available at what ground facilities, weather/conditions at these locations, hazmat requirements, crowd/media control, PR response, logistics (transfer of pax to primary destination/hospital) – so you can focus on flying the aircraft. For us the critical situation proceeds at the same speed as you – or the bullet, the toxic spill or (in a mass casualty event) the speed of blood – in short, at the speed of life. Thus it surprises me to read your perception that you have limited resources on the flight deck in the event of an emergency – it has been my experience that even a pan-pan will have the world rush to your aid. Perhaps CRM training needs to expand to include descriptions of available resources and exercises that include accessing these off-aircraft supports.

TC – Your latest (2005?) has arrived. More soon.

All – I had an interesting conversation with a friend yesterday. By luck, it was a ride on the way-back machine, to when we were both in a location and he was a pilot occasional assisting with the critical incident response team I was a part of. He noted that he had always been impressed that while we had incredible amounts of information coming at us we were able to differentiate between important cues and noise. Over the course of the discussion we arrived at the idea that my communication style is a result of being part of high-impact time-critical decisions several times a week over the course of years. It got me thinking, maybe it’s not the elements of communication/behaviors during normal everyday ops that are important, but rather that we need to be focusing on critical decision-making skills instead.

Thoughts?

- hmm! I do not know of any such 'war room' in my experience of civil aviation so I do not see the 'read across'. I think perhaps we should focus on Crew CRM on this thread?

I'm not sure of your flying experience, JG, but I do think that is a 'rose-tinted spectacles' view. In my experience, in the few minutes I have to action an emergency and land safely, I am very much 'alone' with my crew. I appreciate that the well-known Nimrod with 22 hrs endurance and a jammed u/c who requested the board of enquiry be convened to reach an intial finding before any drills were actioned by the crew might well fit your scenario, however.:) CRM in the limited world of the cockpit means 'crew'. Of course there are 'resources' outside, but timescales often override their usefulness. The only experience I have of any of that 'support' was a possible bomb threat during which we gave up waiting for 'war room' to get started.

JollyGirl - you seem to have a very cynical view towards civil aviation operations. Your supposition that CRM is a failed model, has no statistical evidence. CRM is an ongoing programme of loose definition, designed to enhance safety. It was introduced to reduce the accidents caused by the steep authority gradient in the cockpit (many pilots having been military trained and taken their rank into civil operations).
To state that all F/Os are "captains-in-training" and all captains are training captains is incorrect and somewhat naive.
I started my career at a time when Captains assumed they had god-like status and only to be spoken to through the No. 2 in command. When I retired even the most junior cabin crew member or ramp staff felt at ease in pointing out something that they felt could be important to the flight at that time.
Rather than disparage an evolving system, why not work with it to improve it's concept? Or would you rather rename it and push paper?

As I read this thread I wonder if I'm working in the same industry as some of you.

As far as I'm concerned, the skills, knowledge and attitudes required to operate an aircraft safely are all trainable. To suggest that CRM is a failed model is to throw the baby out with the bathwater. Failures of communication between crew members are major risk factors, and CRM training has the potential to improve that.

The fact that such programs may not be conducted effectively is a training standardisation issue. There is no doubt in my mind that not all operators adequately train to a standard that meets regulatory requirements or recommendations. I've certainly had a "you want a half day in the class room? Preposterous!" conversation with management.

As an industry, we have the knowledge to provide 'best practice' training in a number of areas, yet on a worldwide basis it has not been deseminated. Ueberlingen being the classic case in point.

JG, I'm with BOAC on the emergency handling side of things. When it's time critical we triage our own priority list and the majority of our informational needs will already be catered for via ATC, datalinking or good old fashioned forwards planning. Most of the things in the later half of your list wouldn't even figure on our radar. e.g. We do not fly along thinking 'oh, what about the PR implications of this - never mind, ops will take care of that'.

pb

point8six – Yes, I will concede that I am cynical about aviation operations. But I am cynical about everything. Please try not to take it personally. Also, you would be surprised how many hours of my day are spent trying to make the system better. Pushing paper is the unfortunate price for my vice.

BOAC – I have to be honest, my experiences with aircraft assistance (which has been from both sides of the scope, in the US and East Asia) has been vastly different than yours. I’m sorry to hear the ground handlers you encountered in the past have not been up to the challenge of your situation(s) and hope that action was taken following your event(s) to improve their performance.

GlueBall, BOAC and others - Some of your comments have given me the feeling you are trying to tell me that being a pilot makes you so special that nothing anyone else does has any application in the cockpit… is this an accurate perception?

So back to CRM. Turbocharged, as I read your book (still reading) I see several definitions for the term CRM that have been proposed over the years: to reduce accidents, to reduce human error, to make aviation safer. We will skip my first thought for now (and yes, I will make it back to compliance one day), my second being this: Regulatory agencies set out guidelines for CRM/HF training. Operators must then take this guidance, interpret it and develop curricula. This curricula is then used by trainers, who may modify it as well. Doesn’t this result in an unlimited df regarding the trajectory and objectives of training?

-Jolly

- there is actually very little that 'ground handlers' can offer in these situations. Needless to say, however, ATC, in most places - especially the UK - are always superb. We do not have 'war rooms' and by the time any commercial company has assembled your team for "real-time access to in-house maintenance (sub-system experts) or even the component/platform designer/team at the manufacturer if necessary, but also managing the ground response – what equipment/capabilities are available at what ground facilities, weather/conditions at these locations, hazmat requirements, crowd/media control, PR response, logistics (transfer of pax to primary destination/hospital)" I and my crew would (hopefully) be checking in to our 'ad hoc' hotel. On the ground, my mobile phone is invaluable and I have time normally to seek those resources. In the air, it is the crew and ATC that are invaluable.

No, I (and, I trust, the other 2) do not think "that being a pilot makes you so special that nothing anyone else does has any application in the cockpit" - it is just that being the Captain makes me the team leader on the spot.

I find myself wondering, as we exchange here, exactly what your flying experience is, both mil and civil, and, for example, how long your "head has been in the clouds"? I'm sure we would all benefit from a brief resume of your aviation experience so we can focus our responses. As a gesture, mine, for you, is 17 years mil fast jet and 26 years civil flying.

So continuing with the df thought, doesn’t this variability make measurement of efficacy almost impossible? Am I correct to understand that the regulations set out what is essentially a wish list of objectives with no quantitative measurement guidelines; doesn’t this set up an issue with validation/repeatability?

When I add a little Reason to my thinking (along with some Weigman and Shappell) I get this: deviations on the flight deck fall into four categories: skill-based errors, decision-based errors, perceptual errors and violations. Now skill-based errors arise from simple lack of practice – the remedy here is additional training, whether it be self-study or in a more formal setting. Decision-based errors are the aviation equivalent of turning off the highway to go to work rather than the mall as you intended; these are attention based and can be reduced by reducing the number/frequency/intensity of distractions in the cockpit and increasing focus on task. Perception errors are the result of incorrect detection and/or interpretation of cues; here the fix is to increase the quality and integrity of cues/data/information displayed on the flight deck. Violations are, well, violations; routine violations can be checked with the enforcement of standards, with not much to be done about the exceptional ones.

Now, looking at this with my view of CRM (which has apparently been distorted with training that emphasized it is a communication model designed to reduce errors/accidents): There is not much communication models can do to reduce the frequency of skill-based errors. Control over this variable is a result of management providing sufficient training to allow pilots to demonstrate inert knowledge as opposed to rote recitation of system architecture, EPs and system functionality (or an individual making the effort to accomplish these on their own). Now there are occasions where captain/copilot intervention can forestall a decision error; when communication flows captain to copilot these actions are classified as training. But really, to reduce this number we need work flows that support a variable range of workload (with a minimum of distractions and that allow for breaks) that does not allow (or at least reduces the potential for) channeling of attention and/or periods of excessive workload, especially during critical stages of flight. Perception errors are based on the quality of information presented to crews; here again a word here and there can assist in comprehension but the underlying issue is the layout of both display screens/pages/cues/architecture/feedback and that of the physical cockpit. As for violations, again, an in-flight admonition may have some effect, but in the long-term they are reduced by management practice.

So here we go – of the four conditions CRM seeks to mitigate/eliminate, three have underlying causes which can be only slightly affected by flight crew interactions, with the fourth entirely out of their hands (at least at the time of occurrence). While I am all for pilots receiving all the HF training they can get, and I agree that a reduced cockpit authority gradient is a good thing, I wonder whether CRM is the correct remedy for our ills? Are we spending more time/money/effort on a band-aid than we would on a cure? Turbocharged – I have to admit that my first thought as I read through your book was that a perhaps the explanation for the reduction in fuel exhaustion events (such as Portland) is not that we have this improved cockpit communication/HF model but rather that the newer advanced automation cockpits (which normally include a LOW FUEL caution) are a more salient authority than other (liveware) crewmembers.

As always, let ‘er rip.

BOAC:
If there is one thing you take away from this thread I hope it is this:
(In my neck of the woods and I’m assuming yours is the same) Aircraft operators have a dedicated factory service (maintenance) representative assigned to them by both the airframer and the engine manufacturer. These representatives provide on-site and on-call assistance with the troubleshooting of maintenance issues. In addition, both the airframer and engine manufacturer have a 24-hour help desk, accessed by a 1-800 number or international equivalent. I would hope that, in the event of an issue outside the LASA Possible/Practicable /other programmed criteria, your dispatch function woud be able to (once you ask them) get in-house maintenance, a factory service representative or the factory help desk on the line to begin triaging your (in-flight) issue within a reasonable amount of time (3-5 minutes?) of your request. If your experience(s) during your event(s) was different than this, I sincerely hope the operator/union involved initiated an investigation to determine what could have been done differently to increase the likelihood the next crew who needs this level of assistance has ready access to these resources.
With sincere respect,
Jolly

Some hard views earlier in the thread. CRM historically is a diverse subject, often misrepresented or couched in cultural terms best fitting ‘the way we do things here’.

”Doesn’t this result in an unlimited df regarding the trajectory and objectives of training?” Unfortunately yes; these aspects are often seen as weakness or failure. It is equally difficult to find or prove the successes of CRM – thus the quest for assessment, but is it necessary to assess this form of training?
If CRM is viewed as an insurance (difficulty in selling it to management), something which has to be embedded in the operation as opposed to biannual training, then whilst it is important to renewed an insurance policy and not ‘call it in’, likewise, CRM’s effectiveness might only be seen during an incident or event.
Thus evaluate your incidents and events for aspects of CRM – human factors – determine what aspects are done well and contribute to safety; there should be more of these than there are failures.

Returning to the issue of communication (#1 and within the thread), the following from New Scientist (5 Oct) may be relevant.
“Our interpretations determine what we do. Some of them we reveal to other people, but we do not pass on information, ideas or memes in a kind of "pass the parcel". Someone speaks to us and we create a personal version of what was said.
These interpretations can come from only one place: past experience. Since everyone's past experience is different, no two people can ever interpret anything in exactly the same way. There will be as many interpretations of this Comment and Analysis as there are readers of it.
There will be as many interpretations of this article as there are readers of it
The wealth and complexity of the scientific study of meaning, via the discipline of psychology, is immense. It should be relished, not run from. As yet we do not know how brain activity becomes meaning, how the firings of neurons translate into subjective experience. So the only fantasies we need right now are the kind that turn into testable hypotheses.
"Ask better questions," as my mentor would still say.”

Concerns over CRM failure also highlight a problem associated with accident investigation - what do the statements of ‘CRM failure’ in accident reports actually mean? What aspect of human performance, mental activity, behaviour, etc, failed? This question is rarely answered, even if it were possible to do so.
In the CFIT example (#9), there appears to be a failure in the communication process – but not necessarily that of the Captain disregarding the message, as presumable this could not be determined.
Thus the alternatives might be:-

• That the message was poorly constructed – “I don’t think this is a good idea” - is an ambiguous statement, it fails to state what ‘it’ is, or the degree of concern, i.e. the message fails to convey ‘information’.
• It’s difficult to infer the tone of the message from text, but there could have been lack urgency and the need for action.
• The message may not have been ‘received’ due to cognitive overload – the focus of attention was elsewhere
• The Captain really did disregard the message - he thought he knew better (attitude), or had a ‘strong but wrong’ mental model.

Most of the above should be included in ‘classic’ CRM communication training. However, it may be debatable if this can be taught in a classroom which is often the expectation of CRM training.
Pilots require both the theoretical knowledge of communication, and the practical ability (experience) of being able to relate their understanding of the situation to a style of communication required – it all depends on context.

Inter crew communication could be an inquiry about knowledge (what) or about the likely outcome (how/why). In the first instance (knowledge) the least experience crewmember could be making a knowledgeable statement – “have you seen …” or be asking a question to enhance an apparent lack of knowledge (mental model mismatch). The second instance (know how) could state an appreciation of outcome – ‘high ground 5nm ahead’, or question the apparent outcome – ‘why are we flying this route’, again because there is a mismatch between the crew’s mental models.
The corresponding alternatives, where the more experienced pilot initiates the communication, often occurs with an educational or commanding manner, either of which may result in the failure to appreciate the other crew’s mental model (perception) by lack of enquiry.
In parallel with CRM training we should require continued learning/enquiry routines amongst crews. All Captains should continue to teach – pass on experience; First Officers must continue to enquire, be this to learn or correct an erroneous mental model (either Capt’s or F/O’s) – or as part of the process of monitoring.

So why do we use accident reports for our examples in CRM training? Either because the regulatory authorities (mis)use the reports – ‘CRM’ failure’, and thus demand more training, or because the (better) CRM instructors can build up a scenario from the report to make a training point; what ever the reason we are still in reactive mode.