A double-blind experiment involving 18 drug and/or explosive detection dogs was reported in January 2011 in the journal Animal Cognition. The experiment didn’t just discover that all 18 scent dogs were incapable of accurately detecting the presence of drugs and/or explosives, but overall the dogs were wrong more than 100% of the time.

Belgian Shepherd (dogchannel.com)

All 18 dog/handler teams were trained and certified by a law enforcement agency for either drug detection, explosives detection, or both drug and explosives detection. The 18 dogs included a mixture of male and females, there were 4 different breeds plus mixes, and the dogs ranged from 2 to 10 years old.

Each of the 18 teams conducted two searches of four rooms that the dog’s handlers were led to believe had hidden drugs and/or explosives depending on the dog’s training. So a total of 144 searches were conducted.

However, unbeknownst to the dog handlers there were no drugs or explosives involved in the experiment. Half of the rooms did not have anything hidden in them, while the other half had “Two Slim-Jim sausages (removed from their wrappers and stored with their wrappers in an unsealed plastic bag) and a new tennis ball hidden in the bottom of a pot and placed in a metal cabinet with the doors closed.” (3) Consequently, every “alert” by a dog to finding drugs or explosives would be a “false” alert. All the dogs were trained to signal an “alert” by barking and/or sitting where they had detected drugs or explosives.

One of the things the experiment was intended to measure was the effect of a handler’s non-verbal cues on their dog’s “alert” to finding drugs or explosives. That is known as the “Clever Hans” effect. In the early 20th century a horse named Clever Hans was believed to have amazing knowledge and mathematical abilities because he so often tapped with a hove the correct answer to a question. However, it was eventually discovered that Hans the horse was indeed clever, because he detected subtle cues from both bystanders and his owner about the correct answer.

To measure the “Clever Hans” effect, half of the rooms had a piece of 8-1/2” x 11” red construction paper taped to the location where the drugs and/or explosives were allegedly located, while the other half of the rooms had no visible indication of their location.

The experiment was double-blind because the persons stationed in each room to observe the dog/handler teams did not know any more than the dog’s handlers did about the details, so they couldn’t unintentionally transmit any information to a handler or their dog.

The dogs “alerted” that they found drugs or explosives a total of 225 times during the 144 searches — even though the handlers had been told that in each search there was only one location that contained drugs or explosives. So the false “alert” rate was in excess of 100% (225/144 = 156%). The dogs identified non-existent drugs or explosives in 123 searches (85%). Although the number of false “alerts” varied by only two between the four rooms (ranging from 55 to 57), what was significant is that in the two rooms that had a cabinet marked with the red paper (visible to the dog’s handler) that supposedly identified where the drugs or explosives were, over 50% of the false alerts were at that cabinet. So indeed, each dog’s handler either consciously or unconsciously telegraphed those locations to their dog.

Overall the dogs identified non-existent drugs or explosives at 39 different locations in the four rooms.

All the dogs were certified, so the experiment provides evidence there are significant deficiencies with the process used to certify dog/handler teams for detection of drugs and/or explosives.

The experiments findings are significant because they provide evidence undermining the reliability of drugs or explosive evidence in a criminal case allegedly found by a dog. The experiment suggests that in a significant number of those cases the drugs or explosives may have been planted so they could be found by a dog given cues during the search. Consequently, the experiment could prove valuable for a defense attorney in a case involving key prosecution evidence of drugs or explosives allegedly “found” by a scent dog.

Click here to read “Handler beliefs affect scent detection dog outcomes,” by the researchers who conducted the experiment.

By Hans Sherrer
Justice Denied

Police scent dogs were wrong 100% of the time in alerting their handler to the presence of drugs during
sweeps of schools in Janesville, Wisconsin. During sweeps at six middle and high schools the scent dogs gave 80 alerts for the presence of drugs in a student’s locker — and no drugs were found in any of the lockers. The dogs also falsely detected drugs in 13 cars during a sweep of the parking lot at two high schools. The sweep of a high school in nearby Edgerton by police scent dogs resulted in 9 alerts for drugs when none were present. So in those sweeps the dogs falsely alerted for drugs a total of 102 times — without correctly detecting drugs a single time.

Belgian shepherd (dogchannel)

The inability of the scent dogs to reliably detect drugs during the school sweeps is consistent with the finding of a recent double blind experiment involving 18 certified police scent dog/handler teams that found they were wrong 85% of the time in detecting drugs and/or explosives. An article about that experiment, “Handler beliefs affect scent detection dog outcomes,” was reported in the January 2011 issue of the journal Animal Cognition.

The scent dogs used to sweep the schools were certified to detect marijuana, cocaine, heroin and methamphetamines. The scent dogs inability to reliably detect drugs suggests the certification process is flawed since it does not involve a double-blind test to determine a dog’s competence. A double-blind test is one in which the dog’s handler doesn’t know if there are any drugs in an area the dog is assigned to search or where those drugs might be located. The experiment of the 18 police scent dog/handler teams also supports that a handler sends conscious or unconscious cues to his dog where drugs or explosives are believed to be located. Consequently the only way to determine a scent dog’s competence is to administer a double-blind test that prevents the handler from tipping off the dog where to search.

Although it is now known that a scent dog’s alert for drugs and explosives is overwhelmingly likely to be false, the widely believed myth that dogs can accurately detect the presence of contraband is relied on by judges to almost automatically issue a search warrant when presented with scent dog evidence that is actually in the realm of “junk science.”

If you go to the Janesville Gazette’s website at http://gazettextra.com and enter “drugs dogs schools” into the search box a series of articles will be listed about the scent dog searches in area schools.

By Hans Sherrer
Justice Denied