An arrest in Utah and the subsequent dismisal of the charges has again highlighted the potential issues with drug-sniffing dogs in law enforcement, and how the application of modern principles of scientific behaviorism could make trained dogs into better evidentiary and investigatory tools.

The dog’s handler told fellow officers that, although the dog had not alerted to the presense of drugs like he had been trained to do, he could tell from the dog’s behavior that drugs were present. The Utah court found the arrest to be unlawful after reviewing video of the search, and questioned the efficacy of the police department’s training of its animals, ruling that the dog was not even certified in a single-blind test.

Most police departments nationwide use single-blind tests to certify their drug-sniffing dogs. This means that, during the test, the handler is not aware of the quantity or location of the drugs the dog is searching for. Failing to conduct a certification test to this standard means that the handler — who has an interest in the outcome of the dog’s test — could intentionally or unintentionally influence its outcome.

A better test than the common standard would be a double-blind, in which the examiner is also unaware of the location of the drugs. A dog is not an objective instrument in the way that a metal detector is; he will use any available cue to reach the result he thinks he’s supposed to. Even an examiner disinterested in the outcome may still influence it.

Even better would be to include false-positive testing. This means that in some of the tests, the handler is be mislead about what the dog should find. This helps to select for dogs who can ignore the unconscious cues of their handlers.1

A more rigorous national standard consistent with this Utah ruling is the first step. The second would be an evidentiary ruling that the entitles the defense to review full footage of the dog’s search and alert. In the absense of a practical way to cross-examine a canine witness, an arrest based solely on a dog’s alert should stand only if the alert can be seen and reviewed on video.

Unscientific methods in the legal system have consequences both for accuracy and for civil rights. A well-trained, tested, and certified detection-dog can be an asset for law enforcement, but too often under-trained dogs are used as rubber stamps for probable case.

  1. Disappointingly, false-positive testing seems not to be a part of any department’s training. A study using this type of test was done at UC Davis in 2011 with 18 teams of law-enforcement officers and their dogs, who succeeded in 21 out of 144 searches. This suggests an 85% false alert rate.