« What's wrong with "near-match" DNA database searches? | Main | Helping Legal Actors with Bayes’ Theorem »

August 25, 2006

The Perils of Profiling

Professor Bernard Harcourt has an interesting op-ed in today's New York Times regarding the Transportation Security Administration's intention to expand behavioral profiling at the Nation's airports.  [See here]  He questions the validity of such techniques and recommends a straightforward thorough search of each passenger as a more effective solution.  I agree.  But I would add that there is another danger to adoption of such profiles.  They, in fact, tend not to narrow the population worth targeting for greater intrusion, they expand it.  Experience with behavioral profiles of the past, especially drug courier profiles and hijacker profiles, indicates that just about any behavior eventually finds its way into the profile's definition.  This means that law enforcement personnel are granted a large measure of discretion under the cover of a purportedly objective test.  For instance, courts have upheld searches of suspected drug couriers on the basis of a DEA profile that listed suspicious behavior as including (1) getting off an airplane first (United States v. Millan (8th Circuit)), (2) getting off an airplane in the middle (United States v. Buenaventura (2d Circuit)), and (3) getting off an airplane near the end (United States v. Mendenhall (S.Ct.)).  Other suspicious activity identified in court opinions includes (1) carrying a large heavy bag, (2) carrying little or no luggage, (3) paying for the ticket in cash with large bills, (4) paying for the ticket in cash with small bills, (5) paying for the ticket in cash with a combination of large and small bills, (6) leaving the airport by taxi, (7) leaving the airport by hotel courtesy van, (8) leaving the airport by private transportation, and so on.  The reader, I am sure, gets the point.  The problem with behavioral profiles is that they actually permit screening personnel to pursue hunches and intuition, possibly based on illegitimate if not unconstitutional bases, but to hide behind scientific claims of accuracy.  In the past, courts have not demonstrated either the inclination or the capacity to demand valid data to support reliance on behavioral profiles.  It is not that such profiles could not prove to be an effective law enforcement tool, but that they never are proved to be effective.  If TSA goes forward with its plans to step-up use of behavioral profiles, courts must demand the data that would support their continued use.
-- DLF

August 25, 2006 | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference The Perils of Profiling:


Post a comment