Gladwell's Culture & Air Crashes Analysis Badly Flawed

Koreantoast writes "As a recent Slashdot article showed, interest in Malcolm Gladwell's theory on the impact of culture on airline crashes has come up again following the tragic accident of Asiana Flight 214. Yet how good was Gladwell's analysis of the Korean Air Flight 801 accident which is the basis of his theory? A recent analysis by the popular Ask a Korean! blog shows serious flaws in Gladwell's presentation: ignorance of the power dynamics amongst the flight crew, mischaracterizations of Korean Air's flight accident record (three of the seven deadly incidents characterized as 'accidents' were actually military attacks or terrorism) and manipulative omissions in the pilot transcripts to falsely portray the situation. 'Even under the most kindly light, Gladwell is guilty of reckless and gross negligence. Under a harsher light, Gladwell's work on the connection between culture and plane crashes is a shoddy fraud.' Perhaps Gladwell should have asked a Korean before writing the chapter."

Comments

Nevil Shute worried about this problem in 1940's

[jbrohan]

Nevil Shute worked on the problem of making sure that aircraft were properly repaired. When the engine cowling is closed who knows if the work was done properly? His solution is a new religion of aircraft mechanics. Ordinary people pray 5 times a day, but we are special people responsible for keeping aircraft safe, we need to pray 50 times a day, each time we start a task, and each time we finish a task. The book he wrote "Round the Bend" by Nevil Shute is widely available in the bookshelves of elderly engineers. The problem is still alive and dangerous today and we approach this with code walk-throughs and such like.

How can this be?

[Ambitwistor]

I was assured on Slashdot that Gladwell was supported by evidence and logic and science, and anyone who disagrees is just being politically correct.

Or simply

[Virtucon]

The pilot was a trainee learning the capabilities and handling of the 777 and his co-pilot, the instructor was merely incompetent? I'll believe that before I believe cultural hierarchies resulted in the crash.

This incident reminded me of another aircraft mishap involving SFO, a Compressor Stall with a somewhat rusty first officer at the stick on a 747

[quote]
On June 28, 1998, a UAL 747-400 that had just taken off from San Francisco International (SFO) experienced a number-three engine compressor stall. The plane shook violently, and the crew shut down the number-three engine. Then, instead of applying rudder, the first officer, who was piloting the plane, used ailerons and spoilers, further slowing the heavily-loaded plane. The stick-shaker stall warning activated, and the F/O pushed the nose over, getting so low that the ground proximity warning activated. The 747 cleared San Bruno Mountain, which is dotted with 600-foot TV towers, by less than 100 feet. At that point, the captain took control, dumped fuel and returned to SFO. In the aftermath of the incident, it was discovered many of the airlines' F/Os were flying for years without making any real-world takeoffs and landings.
[/quote]

What also came out of that incident was the fact that the first officer was getting instructions yelled at him from others in the cockpit while a more experienced captain sat there with his hands off the controls in the left seat. Eventually the more experienced captain finally took control of the plane and landed it back at SFO. Never mind the fact that there's passengers in the back and that you nearly hit a mountain letting the first officer get some experience. It could have been a very bad catastrophe but instead it was a near miss.

The FAA after that mandated that pilots had to do more "real" takeoffs and landings instead of mostly simulator runs.

Re:Or simply

[Virtucon]

Read the one post above about how little actual flying time pilots get these days. Takeoff, get to cruise altitude, switch on auto pilot.

How much actual stick and rudder time, I mean actual handling the aircraft do pilots get these days? Not in simulators, but in the cockpit, actually handling the throttles, the flaps and all the other controls. For all we know those 10,000 hours were really more like 1000 in terms of actually taking control. You had other pilots in that cockpit and nobody saw the problem, typical. But at least they're still alive and now they can tell their side of the story and maybe something good will come out of this in terms of training or better automation, cockpit warnings etc. to help in these kinds of situations.

Not to be macabre, but this happens all the time in air disasters.
Look at the Airbus 330 crash from a few years ago.

http://www.guardian.co.uk/world/2011/jul/29/air-france-crash-pilot-error

[quote]
Captain, Marc Dubois, 58, was resting when the Airbus began encountering turbulance, leaving co-pilots David Robert, 37, and Pierre-Cedric Bonin, 32, in the cockpit.

Bonin was at the controls when the speed sensors failed. When the autopilot reacted to the confused readings by disconnecting itself and handing control of the plane to the pilot, he reportedly hauled the aircraft up to 37,500ft in an apparent attempt to slow it down. As a consequence the A330's stall warning sounded, meaning that the plane's aerodynamics were not generating enough lift even though its twin engines were working normally.

Robert, Bonin's co-pilot at the time, supposedly check-listing the emergency procedures, lost precious seconds calling the captain and failed to correct his colleague's error as the plane plunged towards the sea, said the report. Dubois had returned to the cockpit seconds before the crash but was unable to save the situation as it hit the Atlantic belly first.

A French pilot told Le Figaro newspaper: "This manoeuvre (the pulling up of the plane) is totally incomprehensible. My colleague must have panicked."
[/quote]

Inexperience cost all those people their lives. Yes there was a mechanical failure in sensing true airspeed but the guys in the cockpit didn't have enough experience actually flying the plane, ignoring stall warnings and were relying on the autopilot.

Somebody has to fly that plane and personally I'd prefer it to be somebody who's got experience at actual control vs. simulated runs or hours logged on auto pilot.

Not too surprising

[nine-times]

I have to say, I enjoy Gladwell's books. They're interesting and thought provoking. However, I've noticed a sort of pattern. He gives lots of examples of his theories, and the examples always sound compelling, but whenever I know about the example he's using in detail, his analysis is generally wrong. They're not patently provably wrong, but just wrong enough to make me uneasy and think, "This is a really weak argument here. If I knew about his other examples in detail, would they be equally weak?"