Thursday 26 March 2015

#Germanwings: the speed of judgment

On Tuesday morning this week a plane crashed, and this Thursday lunchtime an official investigator has stated the cause: deliberate action on the part of the co-pilot. Before we look at the psychological aspects, consider the cultural ones: very few crash investigations have ever proceeded this fast. Every textbook would recommend a much more lengthy, considered and cautious approach, spread over several months, and sometime a year or so. A textbook, I should explain, is a printed text, bound and published, so that it can be taken down from a bookshelf and read.

Textbooks are historical documents. The contemporary equivalents can be updated every month or so, or over several days when important changes are required. Essential safety directives to airlines are not transmitted by post, but come directly by email. Data is gathered so fast that the path of the a plane can be tracked in real time, and in this case transmitted widely within moments of the crash. I use the app to track family flights, and it can be a disturbing experience, particularly when a plane is stacked out at sea for no stated reason. The crash site in the Germanwings case was shown within 24 hours, the damaged black box located and analysed very rapidly, the names of the pilots partly revealed 2 hours ago, and now given fully. Facts gleaned from Facebook pages and interviews with colleagues will be common currency by this evening.

For grieving families the speed of enquiry may be a blessing. The news is already so terrible that the terrible fact of deliberate pilot action is less bad than weeks and weeks of uncertainty. According to long observation, anger will now be directed at the airline company and regulators. This may seem unfair, but many people will assume that there has been a preventable error in the selection and monitoring of these pilots.

Can psychology offer hard data on the success of psychological screening and monitoring? Lufthansa CEO Carsten Spohr, ultimate owner of Germanwings has has apparently confirmed that as a general practice, pilots in the Lufthansa group do not undergo psychological testing. Is that something which should change?

Pilots are a well studied group, particularly fighter pilots. Test of intelligence are extensive, and in the US will normally involve a full Wechsler Adult intelligence assessment, and the Reitan Trail Making Test, on which there are extensive pilot norms. There will also be standardised personality inventories. Unless you have access to the materials, intelligence test answers cannot be faked. You either get the solution in the allotted time or you don’t. Personality tests are easier to fake. It takes little intelligence to work out that airlines want dependable, sober, resilient people able to deal with stress and uncertainty in a calm and effective manner. Nowadays they also want team players who can take criticism and listen to advice, and keep their egos well under control. It requires no tuition to realise that if you are asked “Would you describe yourself as a worrier” or “Do you do most things on the spur of the moment” it might be silly to admit it. Self evaluations, in my opinion, have at best a tenuous relationship to usual behaviour. The observer evaluations of people who know you well and have worked with you are more informative. As far as I know, these are not collected.

Once hired, commercial pilots have their physical health monitored regularly, and with even greater frequency once over 40 years of age, but psychological evaluations are less frequent, and most depend on self-referral, or on alerts raised by colleagues. The usual triggers of drink, divorce, debt and depression apply. Many pilots with a living to earn seek private treatment and keep flying, unless things get very bad. Being a pilot is a cherished job, and too much honesty about personal problems could lead to too much unemployment.

Surely psychologists have some tricky tests that can detect a death wish in a superficially happy and adjusted persons? These have been proposed for many decades. Most are attempts to detect highly defensive emotional reactions, leading people to effectively block out perceived threats, at the cost of eventually being overcome by a powerful emotional event or sharply depressed mood.

The base rate problem affects the detection of all rare conditions. Pilot selection is good enough to weed out marginal candidates. There are so many applicants that employers can be very choosy. In this occupation no busybody can tell you that you have to employ someone so as to make the profession representative of the general public. Not yet, anyway.

Can psychologist rise to the challenge of detecting unsafe pilots with an acceptable rate of false positives?

Should we tolerate stripping some capable pilots of their occupation (with compensation) so as to protect ourselves against a potentially unsafe one? What if that mean unfairly removing 5o pilots to nail the 1 that harbours bad intentions?

Finally, some words on suicide. Usually, that means going somewhere quiet, usually in the throes of profound despair, and killing one’s self. Killing 150 people, even out of an egotistical wish to end one’s life in a spectacular manner, is usually described as murder.


  1. "Can psychologist rise to the challenge of detecting unsafe pilots with an acceptable rate of false positives?

    Should we tolerate stripping some capable pilots of their occupation (with compensation) so as to protect ourselves against a potentially unsafe one? What if that mean unfairly removing 5o pilots to nail the 1 that harbours bad intentions?"

    No. How many commercial pilots are there? How many of them commit suicide by crash? We can probably count that on one hand.

    People may not want to hear this, but the risks are as low as they possibly can be, given our current technology. Sometimes there is no solution, certainly not a fool proof one.

  2. @James - I have argued that the Weschler is not an appropriate IQ test when used in this kind of situation, for this kind of purpose:

  3. I'm a professional pilot. I have not applied for jobs with US or European airlines, but I did apply to Emirates and went through the testing. I was not offered further interviews but those tested were invited to call for the results of their psychological screening. I was described as introverted, works well with others but is possibly not assertive sometimes with others. I think that's pretty fair.

    FedEx had an attempted pilot suicide incident involving a pilot riding in the cockpit to headquarters, where he knew he was going to be fired. FedEx now only accepts applicants with multiple internal referrals, I suppose they figured that was the best way to avoid unstable candidates.

    It's often a very stressful and difficult life. You have to go to work right after a fight with your wife where she threatens to leave. It may be an easy day, it may be a very hard day, but you have to work all the same.

    I doubt that any psych test is better than extensively investigating the background and character of the candidate, which the airlines already do, at least the big ones.

  4. Just a note: I watched the prosecutor's press conference (visible at and it was obvious that he meant to say "given the evidence that we have, the crash seems like a voluntary act on the part of the copilot".

    The English part (to which I'm referring) begins at roughly 14:00.

    I don't think that it was a rush to judgement at all. He even said (in French unfortunately) that he would wait to change the charges from involuntary homicide to voluntary until more is known.

  5. I read that 6 years ago the pilot had interupted his traing because he was suffering from depression. I have no knowledge of psychology at all, but I assume that any prior record of mental problems increase the risk of such problems later in life. If that assumption is true, Lufthansa has not been all that picky in ist hireing.