The Profiling Game at Border Crossings

Developing effective screening processes in border crossings, in order to identify violators within large groups of mostly innocent people, is an important and difficult task, as it is not possible nor effective to screen every passenger with the same intensity required to detect a violator. Profiling has been applied for several decades as a tool to deal with this task, but there is still no proof of its usefulness. Our main motivation is to study whether profiling is indeed helpful, and if so, how it should be used so as to maximize its effectiveness. We consider an interaction that takes place in some crowded border crossing, where passengers can be affiliated into different groups. We offer a sequential game-model with three players: a defender, who acts first and decides on a screening process, an attacker, who acts second and may recruit a passenger as a violator, and the recruited passenger, who acts last and may choose not to violate, as it has its own private violating motivation. We study variants of the base game, which differ by the choice of screening policy (an announced profiling, an unannounced profiling, no-profiling), the attacker's knowledge of the screening policy, and the costs of recruiting passengers of different groups. These variants help us to answer the extremely challenging social and strategic questions regarding the controversy over the need of profiling.