Pinpointing terror passengers poses problems with profiling

As airline passenger profiling is introduced to the UK under a mist of speculation, panic and alleged discrimination, forensic psychologist Dr Jeremy Quayle claims “there is little an individual can do to avoid a potential threat using appearance and character descriptions.”

Dr Quayle is a leading researcher in the study of cognition and behaviour at Sheffield Hallam University and has experience in the prison service. He explains: “A key problem in the area of passenger profiling is that genuine threats to airline security are extremely rare, meaning that data concerning the characteristics of airline terrorists is scarce. Most terrorists would be classed as 'first-time offenders'. Consequently accurate risk levels derived from this data is likely to be poor in comparison, for example, to predicting the likelihood of a sex offender re-offending. Therefore predictions have to be based on indirect factors and informed guesses rather than reliable statistical evidence.”

“Characteristics of offenders such as being socially isolated, associating with criminal peers or having a cold, callous attitude towards victims of offences can be used as risk predictors, but many of these ideas cannot be generalised to predicting who will commit a terrorist attack. It is however widely assumed a terrorist is unlikely to participate in a frequent flyer programme or pay for a return ticket with a credit card, since these would reveal information about their identity.”

In spite of recent high profile interest in passenger profiling, it has been used as a security measure by US airlines for more than twenty years. There are two approaches. Firstly, passenger lists are cross checked against government lists of suspected terrorists. Secondly data concerning the characteristics of non-offending passengers are compared and analysed with those who have gone onto commit terrorist acts to establish 'risk factors'. Computer passenger profiling systems then assign a risk assessment 'score' colour for each passenger in green for low risk, yellow for moderate risk and red for high risk.

“Actuarial risk prediction is based on the notion that the greatest predictor of future conduct is past behaviour. This type of passenger profiling will work provided large enough bodies of accurate, representative data concerning passengers and terrorists are collected and analysed. If someone has offended in a specific way in the past, they are more likely to offend again in a similar way.

“The actual risk factors airlines use are secret however they insist they do not include factors such as race, religion or national origin. Indeed this makes a great deal of sense. Whilst individuals involved in recent threats to airline security may have been Muslim and Arabic or Asian in appearance, many millions of airline passengers who would never consider threatening airline security also fit this description.”