A Consultant's Casebook

              about me:   Andres Inn, Ph.D.


[Under Construction]



Kiss.com Picture Personals and Dating LowerMyBills.com

Police Field Interrogation

Work with the police was never dull.  While we studied police shooting data, I was given the responsibility for developing evidence in support of a high-speed pursuit policy.  A patrol car had given chase to a teenager, and during the 15-minute pursuit through the central business district, the suspect’s car and patrol car damaged thirteen other cars.  The teenager could only be cited for a traffic offense, but the Police Department was legally liable for much of the damage.  The personal injury potential from such incidents demanded immediate attention from the Chief Frank Dyson. 

I reviewed two years of data and found that in 96% of the cases involving high-speed pursuit, the maximum charge against the suspect was “failure to yield to police.”  If police had not given chase, there would have not been a crime!  And, our finding was corroborated by anecdotal evidence.  Many senior officers recalled how they rescued an evening from boredom.  If they spotted a car with a jacked-up painted rear-end with dual, chrome exhaust pipes they would close in and follow it, bumper-to-bumper.  After some time, the patrol car would fall one stoplight behind, and the officers were almost assured a chase.

Our data and conclusions were straightforward, and the Dallas Police Department issued a policy statement limiting high-speed pursuit.  I had further interest, however.  During our interviews, police officers stressed the need to stop “suspicious” people and their anecdotal evidence suggested it was a useful tool in preventing and solving crime.  

Such anecdotal evidence appeared absurd.  There is a wealth of literature in psychology that suggests that in controlled laboratory circumstances, people can’t even identify emotions from facial features.  Yet, police officers claimed they could identify “suspicious” criminal-looking people from a moving patrol car.  “Stop-and-frisk,” they claimed was a legitimate tool for policing.  I expanded the mandate to investigate high-speed pursuit to include the policy of field interrogation or “stop-and-frisk.”   

There was a grizzled old traffic sergeant that assisted me on this project.  He always remarked that his line of work represented the front line in “the war between the humans and the poh-liis” and that “the humans were winning.”  The traffic sergeant and I divided the city into grids, randomly chose a sample of locations, and then arranged for traffic blockades to be conducted at two-hour intervals throughout the city.  Two patrol cars and four motorcycle officers blocked the streets.  The motorcycle officers chased down any cars attempting to turn off and evade the blockade.  From each motorist stopped, the officers collected the driver’s license number.

The data from this experiment included the date and time of the traffic stop, and from the drivers’ license numbers we obtained race, sex, and age information as well as the arrest record of each person stopped.  These data were compared with (1) the race, sex, age, and arrest records of people actually arrested, and (2) the race, sex, age, and arrest records of persons stopped for field interrogation (“stop-and-frisk”)[1],

The results were predictable. Police are inaccurate in their perception of “suspicious” looking people.  They disproportionately “stop-and-frisk” males, blacks, and the young, That is, the sex, race, and age distributions of actual arrested suspects differed significantly from the sex, race, and age distributions of people subjected to field interrogation.  Inevitably, if the distribution of personal characteristics of arrested people does not match the distribution for field-interrogated persons, police are performing their “stop-and-frisk” ineffectively.

It might be argued, that really “suspicious” people tend to avoid the police and are therefore unavailable for field interrogation.  That might explain why the personal characteristics of interrogated people differs from the distribution of arrested people.  But, in fact, when we compared the distribution of interrogated people against our distribution of randomly stopped people – those availabe for field interrogation, again we found that police dis-proportionally “stop-and-frisk” males, blacks, and the young.  Worse yet, the proportion of the people who were stopped by our experimental traffic blockades and had outstanding arrest warrants was the same as the proportion of outstanding arrest warrants discovered during “stop-and-frisk” procedures.  In addition to being biased, field interrogation proved to be no better than random chance at identifying the “suspicious” persons that had outstanding arrest warrants.

[1] This comparison was done matching for day of week and time of day.


(New!)  Investigation of Police Response Time
See Response Time for more details.
All material is copyrighted!  Not to be used without the written permission of Andres Inn.  ©2000.
Send mail to Andres.Inn@mail.ee with questions or comments about this web site.
Last modified: October 19, 2000