is the new 鈥渄riving while black鈥�. So say travellers who believe they are being targeted at airport security on the basis of their race or religion. Yet new analysis confirms warnings that such profiling is no better at weeding out risky individuals than random screening.
Bill Press of the University of Texas at Austin modelled situations in which members of a profiled group are, for example, 100 times as likely to be terrorists as typical travellers. You might think the best approach would to run extra security checks on them 100 times as often. In fact, this is no better than random screening, because you would waste resources screening innocent people in 鈥渉igh risk鈥� groups and so might miss terrorists who fit 鈥渓ow-risk鈥� profiles (Proceedings of the National Academy of Sciences, ).
A better approach might be to take the square root of the extra risk posed by the profiled group, and use that number to decide how often someone from that group will be pulled out of line. Then members of a profiled group who are 100 times as likely to be terrorists would be 10 times as likely to undergo secondary checks. Even then terrorists are so rare this may still be little better than random screening. 鈥淲e probably shouldn鈥檛 be doing profiling at all,鈥� Press suggests.
Advertisement
The US denies screening .