Tomorrow the Supreme Court is scheduled to hear two Florida cases that cast doubt on the routine use of drug-sniffing dogs to generate probable cause for searches. The cases give the Court an opportunity to reconsider what has until now been an excessively deferential approach to a law enforcement tool that is far less dependable than commonly believed.
Florida v. Harris raises the question of how much evidence is needed to establish that a dog is reliable enough for its "alert" to justify a vehicle search. The case involves Clayton Harris, whose pickup truck was pulled over twice in 2006 by Officer William Wheetley of the Liberty County Sheriff's Office, once for an expired tag and once for a malfunctioning brake light. On both occasions Wheetley walked a German shepherd named Aldo around Harris' truck, and on both occasions he reported that Aldo indicated the presence of illegal drugs by "becoming excited and then sitting" near the door handle on the driver's side. Wheetley searched the car both times without finding any substances Aldo was trained to detect. But during the first stop, he found 200 pseudoephedrine pills in a plastic bag, eight boxes containing about 8,000 matches, a bottle of muriatic acid, two bottles of antifreeze, and coffee filters holding iodine crystals. Charged with possessing pseudoephedrine with the intent of using it to make methamphetamine, Harris unsuccessfully sought to have the evidence suppressed, then entered a no-contest plea while reserving the right to appeal the legality of the search. Last year the Florida Supreme Court ruled that the search was invalid because the state had not shown Aldo's alert was enough to establish probable cause.
One difficulty in assessing this issue is that "probable cause" has never been precisely defined. In the context of a drug search, the Supreme Court has said it amounts to a "fair probability" or a "substantial chance" that contraband will be discovered. That does not mean a reliable dog must be right 100 percent of the time. But how high an error rate is acceptable? In a 2005 dissent, Justice David Souter cited examples from court cases of dogs with error rates of up to 38 percent, adding that "dogs in artificial testing situations return false positives anywhere from 12.5 to 60% of the time." Last year a Chicago Tribune study found that vehicle searches justified by a dog's alert failed to find drugs or drug paraphernalia 56 percent of the time. While that error rate might seem surprisingly high, the Court might nevertheless deem a 44 percent chance of finding drugs "fair" or "substantial."
In Aldo's case, there is no field performance record to check, because police did not keep track of his errors. (After all, why would anyone be interested in those?) The state argues that Aldo's two unverified alerts to Harris' truck can be explained by traces of meth that Harris left when opening the door. In other words, the alerts were not, strictly speaking, false positives, because the dog really did smell meth, just not enough to be visible. Along with the possibility that drugs were hidden so cleverly that the cops could not find them, this "residual odor" explanation is a common excuse for apparent errors by drug-sniffing dogs. But it tends to undermine the argument that an alert provides probable cause. While it makes sense that a meth cook (and meth user) would have traces of the drug on his hands, such odors might also be left behind by passengers, passers-by, or even previous owners of the vehicle. If a car once carried a pot smoker or its trunk once contained several pounds of cannabis, for how long could marijuana's ghost be detected by a dog? Depending on how persistent and common residual odors are, they could play havoc with the argument that a dog sniff reveals nothing but the presence or absence of contraband.
In addition to residual odors, a fruitless search based on a dog's supposed alert might be due to other distracting smells, poor training, poor handling, a police officer's subconscious cues, misinterpretation of the dog's behavior, or even outright lying about what the dog did. Given all the potential sources of error, the Florida Supreme Court said, "the fact that a drug-detection dog has been trained and certified to detect narcotics, standing alone, is not sufficient to demonstrate the reliability of the dog," especially since there are no uniform standards for certification. In addition, "the State must present evidence of the dog's training and certification records, an explanation of the meaning of the particular training and certification, field performance records (including any unverified alerts), and evidence concerning the experience and training of the officer handling the dog, as well as any other objective evidence known to the officer about the dog's reliability." Then a judge can decide, based on "the totality of the circumstances," whether the dog's alert amounts to probable cause.
The state of Florida, of course, disagrees, arguing that certification and a handler's confidence in a dog are enough to establish its reliability. The state's Supreme Court brief claims that "no one is in a better position to evaluate the reliability of a well-trained dog's alert than the trained K-9 officer who has spent countless hours training and working with that dog." That's one way of looking at it. But a skeptic might point out that, aside from whatever emotional attachment the officer may have to this dog with which he works so closely, he has a strong incentive to believe (or at least claim) the dog is reliable, since that belief allows him to search people he suspects are carrying drugs and makes the evidence collected during such searches admissible—assuming courts agree with the officer's assessment, which they generally do. It is long past time for judges to examine such assertions more critically.
The other drug dog case, Florida v. Jardines, poses the question of whether a warrant is required for a canine inspection of a home's exterior (which can then be used to obtain a search warrant for the home). In cases involving traffic stops and luggage at airports, the Supreme Court has said the sniff of "a well-trained dog" does not count as a search within the meaning of the Fourth Amendment, based on the theory that it does not reveal anything people have a right to conceal. Does a different analysis apply to a home, given the traditionally stronger privacy protections in that setting? The question of a dog's reliability should matter here too, since 1) the idea that a sniff is not a search rests on an unrealistically sanguine view of the average police dog's abilities, and 2) a mistaken alert to a house results in a much more serious invasion of privacy than a car or suitcase search. If courts are putting their faith in "certified" dogs that are wrong most of the time, the upshot could be a lot of legally unjustified home invasions.