When deciding whether a baseball player is likely to get a hit, we look at his history of success at the plate. When deciding whether to listen to the advice of a stock analyst, we look at whether the prices of her past recommendations rose or fell. But when police officers claim that they have probable cause to believe a certain location contains evidence of a crime, we do not look at whether they have been right or wrong when they have made the same claim in the past. This is a mistake.
Law-enforcement-search success rates vary widely, even when the same legal standard applies. Searches pursuant to warrants issued on a probable-cause standard recover evidence at very high rates, usually exceeding 80%. By contrast, warrantless searches, even when officers allege they have probable cause, succeed at far lower rates, recovering evidence as infrequently as 12% of the time. Similarly, some officers are far more successful than others when they conduct probable-cause searches. Some almost never recover evidence; some almost always do.
What role should these different success rates play in probable-cause analysis? The current answer is none. Judges are not presented with the success rates of the law enforcement officers who appear before them. I argue that law enforcement should be forced to present success-rate data to judges when making probable-cause claims and judges should be allowed to consider the data when deciding whether to issue a warrant or approve a previously conducted search. These success rates capture information not currently analyzed in the search process and their addition would improve the accuracy of probable-cause decisions. Most significantly, we would learn private information in the possession of law enforcement that is not currently presented to judges.
You are a magistrate judge. Officer A is seeking a search warrant for a house.1 He knows the house belongs to an individual with a history of narcotics trafficking. When the officer went to the home and interviewed the owner, the owner seemed nervous and appeared to be concealing something behind his back. Officer A tells you that, in his view, this information constitutes probable cause to believe that the house contains narcotics.
Officer B performed a search of an automobile as part of a routine traffic stop. When he pulled the car over, the officer ran the license plates and learned that the car's owner was a known drug dealer. When he approached the car, the driver seemed nervous and appeared to be concealing something beneath his seat. Officer B searched the car and found narcotics.2 He claims the evidence should be admissible because he had probable cause to believe there were narcotics in the vehicle.
These cases may appear similar and may seem close to the probablecause threshold but are, in fact, very different. There is far more reason to rely on Officer A's judgment. When he believes he has probable cause, he is usually right - he recovers evidence almost 90% of the time.3 On the other hand, Officer B is almost always wrong - he recovers evidence just under 12% of the time.4 Searches pursuant to a warrant are far more likely to recover evidence than warrantless searches.
Such variation does not simply exist between different types of searches. Different officers conducting the same type of search succeed at very different rates. Here we can take as an example two real officers of the Florida State Police.5 Between January 2000 and September 2001, each performed a similar number of probable-cause searches of automobiles: Officer A conducted eighteen searches, while Officer B conducted fifteen searches. When Officer A thought he was likely to recover evidence, he was almost always wrong. Only one of his eighteen searches led to a seizure - a success rate of 5.6%. Officer B, by contrast, was almost always right. He recovered evidence in thirteen of the fifteen searches, for an 86. …