Number game
(In response to Dan Garry’s Sept. 14 column “Unethical detecting.”)
Dan Garry's thoughtful column offers a justifiably skeptical view of technology offering to detect terrorists by reading brainwaves. In fact, the situation is even more dubious than he suggests; and here is where a bit of math literacy can shed light on the problem.
We can assume that there is no more than one terrorist per 100,000 general population; there are probably actually far fewer, but that seems like a conservative working assumption. If the brainwave test is indeed 95 percent accurate as its developer Professor Rosenfeld asserts, that means it would detect 95 out of 100 terrorists; but it also means that it has a false-positive rate of 5 percent, so that if 100,000 persons were screened as potential terrorists, in the same way that employers screen applicants for drug use, we would catch 1 terrorist and falsely label 5,000 suspects as potential or probable terrorists.
Used as a screening tool then, false positives far outnumber true positives, simply because there are far more innocents than terrorists being tested. Retesting would likely not help, since whatever idiosyncrasies of brainwave pattern existed initially still remain and might even be exacerbated by the stress of suspicion or accusation.
Those who argue that this "tool" should be used as an adjunct to criminal investigations and not as a general screening test are left with the fact that the ratio of false positives to true positives will depend entirely upon the skills and prudence of investigators in deciding who to round up and that, over time, sufficiently large numbers of investigations and suspects may well result in an unacceptable number of false positives.
Nor is the argument valid that it's better to put a few innocents (or even a lot of them) under suspicion than to allow acts of terroristic mass murder to take place, since the diversion of limited police resources to concentrate on false suspects means that real terrorists have greater freedom of action in the meantime.
While Professor Rosenfeld may well be aware of the limitations of his test – though he has little incentive to emphasize them while seeking funding and support – can the same be said of the man in the street clamoring for authorities to prevent terrorist acts, or front-line police investigators eager to make arrests and be heros, or police administrators responding to political pressure from ignorant politicians catering to the wishes of an equally ignorant but angry "law and order" constituency?
Let's hope that bright young skeptical idealists like Mr. Garry will be around to remind others of the limitations of gadgetry as a method of police investigation. Terrorists will be caught by the diligent application of human intelligence – collecting information, infiltrating suspect groups or recruiting members as informants, observing individuals' actions, and, where warranted, intercepting their communications
Emil Pulsifer
Reader
