The Efficacy and Flaws of Generalizations and Profiling in Decision-Making and Law Enforcement

Key Insight

Generalizations, while often inevitable and desirable in decision-making, become perilous when they morph into stereotypes. For instance, following 2005 London bombings, the New York City Police Department (NYPD) opted for random subway bag searches over profiling, despite arguments that terrorists often appear as young Arab or Pakistani men. However, defining and identifying such categories is challenging; for example, 'Mafiosi' characters in 'The Godfather' were played by actors of diverse European descents, not exclusively Italian-American. Similarly, identifying an 'Islamic terrorist' is complex, as Islam is a global religion with adherents of all appearances, unlike groups with identifiable costumes. NYPD Commissioner Raymond Kelly asserted that racial profiling is 'wrong' and 'ineffective,' citing the diverse backgrounds of actual London bombers (British citizens of Pakistani descent, a Jamaican, East Africans) and a Chechen woman in Moscow. Given that 40% of New Yorkers are foreign-born, Kelly questioned the practicality of identifying whom to profile.

Profiling faces a fundamental 'category problem,' as successful generalization requires clearly defining and identifying the target category. Terrorists are acutely aware of ethnic characterization and adapt to avoid detection; the 9/11 hijackers, for example, deliberately altered their appearances and behaviors to blend in, demonstrating that they are 'not dumb people.' Kelly highlighted that a terrorist could easily disguise themselves, making profiling 'just nuts.' Furthermore, the effectiveness of generalizations is undermined by instability. For instance, New York City's crime rate, once notoriously high, significantly decreased between 2004 and 2006 (murder down 10%, rape 12%, burglary 18%, auto theft 11.8% in 2005), now ranking 222nd out of 240 major US cities. This demonstrates that historical generalizations can become rapidly inaccurate, unlike stable phenomena such as Kenyan dominance in distance running.

Instability also plagues law enforcement profiling. The Drug Enforcement Administration (DEA) agents historically employed an absurd and self-contradictory list of suspicious traits for identifying drug smugglers, including behaviors like 'arrived late' or 'arrived early,' 'carried no luggage' or 'carried four pieces.' These were unstable traits that smugglers could easily alter (e.g., switching from one-way cash tickets and bulky bags to round-trip credit card tickets and single carry-ons) to evade detection, especially as law enforcement became aware of previous patterns. This dynamic changes the very basis of the generalization, rendering it ineffective. Raymond Kelly, during his tenure as head of the US Customs Service, recognized this flaw and reformed the system by replacing 43 unstable criteria with 6 broad, stable indicators: suspicious physical appearance, nervousness, specific intelligence, drug-sniffing dog alerts, discrepancies in paperwork or explanations, and implicating contraband. By focusing on these unchanging aspects, such as nervousness or detection by dogs, Kelly's reforms led to a 75% reduction in searches while simultaneously improving successful seizures by 25%, making customs officers significantly more efficient and effective.