Normalisation of deviance

[Genghis the Engineer]

This article popped up a couple of days ago...

Bedford and the Normalization of Deviance | (http://www.rapp.org/archives/2015/12/normalization-of-deviance/)

Analysing a G-IV accident I hadn't previously been aware of, but now am. There's a link in the article, but here's the NTSB report directly...

http://www.ntsb.gov/investigations/AccidentReports/Reports/AAR1503.pdf



It's a very interesting analysis, on a report where the causes of the accident are nonetheless distressingly obvious. It did get me thinking: in particular the value of this term "normalisation of deviance", which is related I think to James Reason's concept of the way in which human beings routinely using skill to circumvent failings in deficient systems: rather than to try and achieve corrections in a faulty system.

But how many times have any of us seen this? Aviation professionals (pilots being the obvious candidates, but I'd argue that it goes far wider than that) who routinely do their own thing because they have come to regard some aspect of published procedures as unnnecessary / overly onerous / pointless. Until, as in this case, it bit them.

G

[megan]

Professor Diane Vaughan developed the term "normalisation of deviance" in explaining the Challenger Shuttle accident at the inquiry. The term now much used in the safety industry to explain events.

[Armchairflyer]

The idea of normalization of deviance is also a central idea in Sidney Dekker's "Drift into Failure" (e.g., pp. 103ff.). Besides the Challenger accident, two other cited cases include Alaska Airlines 261 and a friendly fire incident in Iraq in 1994.

IMHO this is not a typical case, however. As I understand, one core element of normalization of deviance is (organizational) pressure towards increased productivity at the expense of time-consuming and/or costly safety procedures (what Reason terms "production vs. protection"). Granted, I am merely a hobby pilot, thus free from the pressures of commercial flying. Still, I fail to see any benefit (individual or organizational) gained from not moving the yoke during preflight, let alone continuing the takeoff even though something seems blatantly amiss with elements as crucial as engines and flight controls. And this with cabin crew and pax on board? Hard for me to disagree with the idea of this being involuntary manslaughter rather than a "normal" accident.

[Big Pistons Forever]

Normalization of deviance got them onto the runway with the controls still locked but continuing with the takeoff after it was obvious things were wrong is not normalization of deviance, it is expectation bias.

The crew knew the controls were locked and tried to fixed the situation with a made up procedure while the aircraft was accelerating down the runway.

I believe that was because both crew were highly experienced both in flying experience and time on type. I think their unconscious expectations was that no matter what happened they could fix it on the fly and continue.

I raise that point because I see this attitude occasionally expressed in posts on the this site. Pilots who think that they are going to save the aircraft from damage after the engine fails is one example of the attitude I am talking about.

One unfortunate part of this accident summary is that it is easy to make it a black and white discussions. This crews action were so wrong in so many obvious ways it is easy to say "well I will never be that dumb !" so nothing here for me to learn.

The reality though is many accidents start with a more subtle scenario and by the time the pilot realizes they are in trouble it is too late.......

One of things I encourage my students to do after they have gotten their licenses is to make a conscious effort to review each flight and identify things that they made a mistake on, or could have done better.

My 02 cents

Ultimately being a "safe pilot" is not one thing it is IMHO a state of mind. Safe flying comes from make an effort on every flight to be the best pilot you can and not accepting poor personal performance. More flying experience makes it easier but does not lessen the requirement.

[Chuck Ellsworth]

After reading the article I am left with a more simple reason for the accident.

It was the result of crew incompetence.

I can not recall ever starting a take off roll without checking for full control movements, and normal if the flight controls were visible.

Sooner or later you may find the controls will not move, it happened to me many years ago, we had gone to town to eat between fire bombing missions and just prior to entering the runway the ailerons would not move when I did the controls free check.

It was very embarrassing having to ask the tower for clearance back to the ramp to remove the aileron gust locks that the mechanics had put on while we were in town eating.

Failing to do a proper walk around was negligent but the ingrained controls free check caught the locked controls problem before we started the take off roll.

[Big Pistons Forever]

After reading the article I am left with a more simple reason for the accident.

It was the result of crew incompetence.

.

Absolutely and by any objective measure, they exhibited a near criminal level of lack of professionalism.

The question in my mind, however is, how did they get to that stage ? The PIC had been flying for 30 years and had 15,000 + hrs and the FO over 9000 hrs, both up to then accident and incident free. These guys did not all of sudden wake up and decide "let's break every rule of good airmanship and common sense today"

[Chuck Ellsworth]

These habits become ingrained through lack of oversight by the people who are supposed to ensure the company flight safety rules are being followed.

In other words the whole company culture was to blame.

[tmmorris]

After reading the article I am left with a more simple reason for the accident.

It was the result of crew incompetence.

I can not recall ever starting a take off roll without checking for full control movements, and normal if the flight controls were visible.

Sooner or later you may find the controls will not move, it happened to me many years ago, we had gone to town to eat between fire bombing missions and just prior to entering the runway the ailerons would not move when I did the controls free check.

It was very embarrassing having to ask the tower for clearance back to the ramp to remove the aileron gust locks that the mechanics had put on while we were in town eating.

Failing to do a proper walk around was negligent but the ingrained controls free check caught the locked controls problem before we started the take off roll.

That's why I'm grateful to a number of instructors who have ingrained in me various checks like that. I remember my IMCR initial test where the comment from the examiner (whom I'd never met before) was 'You made a few mistakes, but you always trapped and rectified them in good time. So you passed!'

I hope I can keep it up...

[118.9]

'Deviance normalisation' are fine, sophisticated-sounding words, but wouldn't 'bad habits' do just as well?

Perhaps there is a simple word that describes arrogance + incompetence + bring it on (throw any emergency at me, I'll handle it)?

Anyway, a simple TAFFIO check would have prevented this calamity!

[Pilot DAR]

This aspect of normalization of deviance has as an element the silly long checklists which are being evolved. Aircraft manufacturers are creating longer ones in more recent years, but then "operators" tend to add more and more. The result is near boredom completing one, and the temptation to skip it in part or in the whole. For this, operators who create longer checklists are a factor in the normalization of deviance, and I appoint some flying schools I have seen as major offenders.

There are some basics pilots must do, and I lump it all in to what I call and train as "configuration assurance". Is the aircraft properly configured for the next thing you're going to do with it? If you're not sure, take a few seconds now to review (like use the checklist) and assure it.

At the most primal: You're going to fly it; do the flight controls move freely, and will fuel get to the engine? If you as the pilot can't take responsibility for those two basic element without paper to remind you, you really need to re-evaluate your place in the sky!

In the mean time, as I review the Cessna 182 checklists for the "A" model to the "S" model, the Preflight goes from 26 items to 53, Prestart 9 to 11, Start (is that really a checklist, or just checklist format instructions?) 9 to 15, and Before Takeoff 12 to 22 items. Admittedly, the "S" model has a fuel pump, which the "A" model does not. But mostly I see this as Cessna learning expensive lessons about corporate liability over the decades, and attempting to solve the problem with longer checklists, rather than we all insisting on training and maintaining better basic discipline in pilots at all levels of skill.

We are normalizing of deviance from pilots maintaining basic cockpit discipline, and configuration assurance as an element of piloting skill. We allow those actions to be moved into aircraft specific checklists, which then become needlessly long, and return to promoting skipping.

I opine that "today's" pilots are trained with silly long checklists, and then use them as a crutch to proper self discipline, until they think they just don't need them any more. Did an instructor ever tell them to simply take mental responsibility for the plane?

I remember my Caravan type training, the Chief Pilot said: "Please always use the Cessna checklist for prestart, start, and pre takeoff." After that, she trained me specific flows in the cockpit, and refer to the checklist if uncertain - but always stop and think about what your were going to do with the plane next, and was it configured correctly?

But it is inexcusable for any pilot to fail to assure that the flight controls move before flight, the fuel will reach the engine(s) and the landing gear is in the correct position for landing. It's like failing to lower the AoA at the indication of an impending stall! Oh... Wait....

[RatherBeFlying]

Some pilots and operators have a Killer Items checklist for just before advancing throttles.

It wouldn't hurt to have it on a placard in an obvious location.

[Chuck Ellsworth]

It wouldn't hurt to have it on a placard in an obvious location.

And the obvious location is your brain.

[Capot]

I hope that it's not too much of a drift to remark that "normalisation of deviance" is a concept that has been endorsed by EASA who, totally unnecessarily, decided that Europe's airports should be subjected to the dead hand of EASA regulation.

So a huge document was eventually produced, after many man-years of bureaucratic endeavour, which restated the ICAO Annexe 14 requirements with some minor changes to prove how valuable EASA is.

But since no airport was actually going to change anything as a result, to avoid any unpleasantness about just ignoring EASA every airport was allowed to have "allowable deviations" and carry on just as before.

That's a slight simplification, but the general principle is clear.

[eckhard]

Pilot DAR

Hear, hear!!

It's been my experience that as the aircraft that I flew grew larger and more complicated, the checklists became shorter and more relevant. I think that reflects the better training courses and emphasis on briefing and crew co-operation that pertained on bigger machines and which perhaps was missing on the smaller types.

I agree that flight-school checklists have evolved into quasi-Operation Manuals, that try to instruct the pilot in the method to operate the aircraft, rather than providing an easy reference that the pilot may consult before committing him/herself to the next phase of flight operations. As you so aptly say, 'configuration assurance' should be the key message.

Example: I remember seeing a B.767 checklist. The 'Before Take-off' checklist had one item on it:

Stab Trim ----- ______units

I think that everything else was left to a 'scan flow' and reliance on the TOCW system.

The trouble with checklists is that manufacturers use them to cover their corporate a$$es, students use them to help them figure out how to fly the plane and instructors who fly multiple types in one day find them unnecessarily prescriptive and difficult to manage. What is the answer?

For simple, small aircraft, I feel that in the 'ideal world' rather than trying to tailor the checklist to the aircraft, one should try to tailor the checklist to the user. In other words, each pilot should be presented with a checklist that best serves his/her needs. For large aircraft, or multi-pilot operations, especially those that involve pilots flying together who have never met (freelance/contract operations), the checklist must reflect 'best practice' on that type.

As we don't live in this 'ideal world', my suggestion may seem an impossibility; however, some critical items (FASTER, TAFFIOHH, BUMPFGH?) lend themselves to almost any type and I would certainly recommend, 'controls full and free', 'flaps set' and 'fuel supply' to the before take-off list. Similarly, I would include, 'gear down, brakes off, flaps set' to any before landing checklist. Everything else is gravy.

Maybe individual pilots should develop their own checklists for the types that they operate? I see the big problem with that: two pilots flying together by chance who both have their own and different checklist. As I said previously, this solution really only lends itself to the single-pilot, simple aircraft.

But if the training organisations could manage to instill a basic 'survival kit' into their students, on top of all the other requirements that they are obliged to fulfil, we might find fewer of these appalling and unnecessary accidents in the future.

[Chuck Ellsworth]

What about flying fixed wing and rotary wing aircraft randomly day after day?

Is that acceptable?

[eckhard]

Yes, probably! As you are doing it 'day after day' you will be familiar with the different skills required in each machine. Every pilot drives/bikes/roller-blades/sails/walks to the airport and has apparently little difficulty transitioning to a new form of transport when they get airborne and back again when they transport themselves home afterwards.

[9 lives]

Occasionally I do fly helicopter and fixed wing in the same day. I do use the prestart and pretakeoff checklist in the helicopter, as I don't fly it often enough to be in "flow" with it, and I don't need the distraction of catching something up later.

When I do transition training to amphibians, I do not make a big deal about checklists, instead, I make a very big deal about the configuration assurance mentioned by Pilot DAR. If you're changing the phase of flight, your brain, without the aid of paper better have the configuration correct. It's all memory items! If you want to add time in there to put the next phase of flight on hold for a few seconds to use the checklist, I can't knock that, but it will be just that, a checklist - everything better have been correctly set and confirmed already - because you know what it should be from your training and skill, and your brain reminds you to check.

Our culture does not allow for abandoning checklists entirely, but I will openly relegate them to their intended role - to aid the pilot in confirming that all required actions have been done. It is not a "to do" list, and I won't solo a pilot who relies on it that way...

[Big Pistons Forever]

Pilot DAR


But if the training organisations could manage to instill a basic 'survival kit' into their students, on top of all the other requirements that they are obliged to fulfil, we might find fewer of these appalling and unnecessary accidents in the future.

I think when teaching people to be "safe" pilots there is a tendency to swing for the fence "do This and you will be safe", or binary arguments "written checklist are Bad, no no no written checklists are Essential"

Personally from a flight s training perspective the creation of safe pilots is all about having the student sweat the boring, unsexy little details all the time, every time, so that it becomes habit.

A while ago I was flying with one of the students who I had trained for their PPL. One of the things I require of my students is that before they turn the key they verbalize to me "tow bar out, area clear". The idea is that this will become a habit and they will mentally do that check before starting the engine after they have finished their PPL.

So we are strapped in, seat adjusted, doors closed and he goes to turn the key to start the engine; and then stops. He turns to me and says "I am 99 % sure the tow bar is out but I want to check". So he pushes the seat back, un does his seat belt, opens the door and leans out to see that in fact the tow has been removed.

That young man aspires to be a commercial pilot. I would like to think that I have inculcated the kinds of good habits that means he will never be hurtling down a runway with locked controls while his FO tries some home made procedure to release the control lock........

[118.9]

What about flying fixed wing and rotary wing aircraft randomly day after day? Is that acceptable?

Why not? I have often flown fling-wing and fixed-wing on the same day. The pre-flight (and all) checks are vital for both.

A killer item lurking in helis is lifting off with the collective and cyclic frictions tight. In light helis, you check for free and full movement before starting and then tighten the frictions again. Obviously you can't do free and full movement when the system is up to speed, so checking the frictions before lift-off is critical - as critical as checking that control locks are removed in a fixed wing.

[Chuck Ellsworth]

My question was rhetorical. ;)

I used to be chief pilot for a company that operated both fixed wing and rotary wing aircraft and flew two traffic patrol flights a day five days a week in our helicopters.

When getting into the helicopter I never had any problem figuring out they were different machines than the fixed wing ones.

[Genghis the Engineer]

I don't fly rotary, but do fly both 3-axis and flexwing microlights, and light aeroplanes.

Going from 3-axis anything to a flexwing I find as difficult as getting out of my car and onto a bicycle. That is to say, not at-all.

I have concerned myself occasionally, although not made any serious mistakes thankfully, going between 3-axis microlights and larger light aeroplanes as whilst the controls work basically the same, the speeds, heights and attitudes are all different.

G

[Chuck Ellsworth]

My reason for bringing this up is that here in Canada the system is breeding generations of pilots who are brain washed into thinking that unless you have a PPC check ride and the paper work to show you have a given type of airplane on your existing license you can not fly that type commercially, hence we have one or two trick ponies who believe flying a new type of airplane is some kind of black magic that requires you have paper work to fy it.

Now don't get me wrong here, I have no problem with pilots who believe that it is best for them to jump through more and more hoops to be blessed by the bureaucracy that controls their ability to fly for a living. Knowing your limitations is perfectly fine.

I came from a different mindset where if we could demonstrate to a company chief pilot that we knew the limitations and pilot operating instructions for a given airplane and then demonstrated we actually could fly it we could then go to work for that company.

Of course we had to show that our license covered that type and class of airplane, for instance we held a ATPL if we were applying for transport category airplanes.

[Ridger]

'Deviance normalisation' are fine, sophisticated-sounding words, but wouldn't 'bad habits' do just as well?

I totally agree with you; bad habits describes it well but deviance normalisation (while just another label) is perhaps a little more informative a term,

Even better is just to describe the issue at hand rather than using such a label (the basis of Sidney Dekker's rationale in many ways) in the first place. Long checklists aren't ergonomic to use and probably contain many items which rarely (if ever) change state. Over a long time (years perhaps) the crew probably gradually omitted more and more of the 'static' items until they ended up with three actions: fire it up, taxi out and take-off.

Unless I missed it, one thing the report doesn't really mention is time - I would expect rushed or omitted checks to be partly driven by time pressure. If this was the case it may help explain the lateness of the t/o abort although many factors could have equally contributed to it.

Whatever the case, it's a very alarming occurrence which was completely avoidable and particularly harrowing, given that all souls survived the initial impact.

[mary meagher]

CB SIFT CB E. only a glider. but this list is graven (or beaten) into the brain of every student pilot.

Controls - Full and Free.
Ballast - weight and balance correct
Straps - securely fastened
Instruments - working and set
Flaps (if fitted) - set for takeoff
Trim - set for airspeed in case of launch failure
Canopy - closed and locked
Brakes - airbrakes in this case, not wheelbrakes
Eventualities - your plan of what to do in case of launch failure

When you have completed this memorised check list, then you may accept the launch cable, whether airtow or winch. The club member on the wingtip will check that above and behind is clear to go. The signal, by radio or by arm for takeoff is given to the tug pilot or the winch driver. Then up you go!

For all that, I had to stop a launch when I was only a beginner holding the wingtip; the highly experienced chairman of the gliding club still had the gust locks on his wings...........the rule is that anyone may STOP a launch if you see danger.

[akaSylvia]

I've written up the accident here:

Fear of Landing ? How to Drop a Gulfstream IV into a Ravine: Habitual Noncompliance (http://fearoflanding.com/accidents/accident-reports/how-to-drop-a-gulfstream-iv-into-a-ravine-habitual-noncompliance/)

I think the most telling aspect was not just that they routinely skipped the pre-flight checks but that having discovered that the throttle was being limited, they tried to solve the issue while continuing the take-off run.

[PLovett]

There was lurking on this forum about a month ago a link to a very interesting article on normalisation of deviance written by an RAF Tornado pilot. It was so interesting that I lifted it and included it in a safety newsletter for the pilots in the company that I work for.

The article referenced the Challenger and Columbia crashes as well as the G4 crash referred to in the OP but the primary interest was in his own tale of very, very nearly smashing a Tornado into the ground while trying to raise the undercarriage.

He was on a NATO deployment at the time engaged in an exercise in Belgium. All was fine until one day the undercarriage refused to lock up. The inspection showed a severely worn part that wouldn't let the undercarriage lock up. It couldn't be fixed. However, and this is where the deviation started, the pilot thought that if he slightly bunted the jet after takeoff thereby creating some negative g the undercarriage might lock up and the exercises could be continued. It worked.

However, during the week the weather worsened requiring more extreme handling to get the gear up (it took about 10 sec to cycle and there were speed limitations on the u/c). It all came to a head on the day they were to return to the UK when it went pear shaped as the cloud base was the lowest it had been. The aircraft wound up heading for the ground outside of the ejection parameters and was only saved by some extremely quick thinking. A very chastened crew then flew the aircraft back to the UK with the u/c down.

What he wanted to emphasise was that the deviation initially was not great but that it gradually grew as the week progressed and the weather worsened. The deviation had worked initially without problem so perhaps some further deviation will work without problem. And this is the rub, its more likely to be found the more experienced the crew.

They are the ones more likely to cut a corner because they have done it in the past without a problem and thus the deviation gets larger. That G4 crew were experienced and the NTSB report notes that on the recurrency training they did everything by the book but the flight data recorder noted that it was common practice for them not to do the pre-takoff checks at all. It had worked for all those times before why shouldn't it work this time?

I discussed the article with the CFI of the company I work for and he admitted to being guilty of it in the past and spoke about that occasion. He subsequently pulled himself up for his lapse and is well aware of the danger. We also had an example of a check-list not been done in circumstances where there was enormous pressure on the pilot to get going (weather, last light etc.). He neglected something that very nearly was fatal.

Want another example, the Costa Concordia. The company navigation plan called for (I think) a 5 mile distance from the island. That had been progressively lessened by the ships crew, all without problem until, of course, that night. Normalisation of deviance at its best.

I get the impression from reading some of the posts here that people think it can't/wouldn't happen to them. I beg to disagree; it can happen to the best of pilots when complacency becomes a factor.

[Piltdown Man]

Bad habits would be a term given by punitive managers trying to make a point rather than trying to fix the problem. Because too often the problem is the system itself. And the things that make it work are the people who run it. And we are not talking about un-motivated local authority workers skiving off work because they can be bothered. I'm talking about highly motivated people who want to do nothing less than a good job. Unfortunately, they are also the people who unknowingly cross the grey margin from relatively safe to highly dangerous.

It has another name: Practical drift. It is defined as the unintentional adaptation of routine behaviors from written procedures. The occurrence of practical drift can result in a catastrophic disaster in normally highly reliable systems.

These are also the little tweaks we all put in to make a system work. We do them just to get the job done more efficiently. Things like not wearing your HiViz vest, releasing the brakes early so your friend the Disatcher can get an on-tine departure, taxiing at 35 Kts to ensure your passengers make their connections, and so on. And then sometimes we have to do them because the job as described would be impossible. I remember once that we had to release the brakes to perform a power-back and when we only when we were reversing could we start the engines. Place this particular crew in the system they operated and then think again before you start throwing rocks at them.

PM