How accurate are Drug Detector Dog?

 The data gathered is from various states and departments. These data will change as more teams are reviewed. It is taken from vehicle searches were the dog's response was used as probable cause.

What is the probability of a vehicle that is stopped using interdiction tactics will in fact possessed drugs?


Of the vehicles that are searched by a drug detector dog to provide probable cause what is the probability of the dog responding?


What is the probability of drugs actually being discovered after a dog's response?


Accuracy claims including drugs actually found after the dog's response, with included claims of residue and currency related to drug traffic. 67.94%

* These numbers are from records provided from 14 different detector dog teams.*

There are many factors that influence these numbers. After hearing many handlers testify concerning what they were taught concerning record keeping and analysis I would have to say the main reason is failure by trainers to properly instruct handlers in the need for detailed records, and the importance of monthly analysis of those records. This opinion is not only supported by the testimony of handlers, I saw it with interaction with other trainers. For example, I was invited to instruct at a seminar in Jackson Ms. some years ago. Of the instructors present I was the only one who actually ran a training session without an aid present. Not one handler was able to complete the search area, that would taken a well trained team about 10 minutes to clear, without a response. One dog responded 22 times in less than 3 minutes. The handler told me his dog had never responded I asked him if he ever did a training exercise without aids being present and he said no.

If handlers would plot their records they would see trends they cannot see without such information. This graph was taken from a month to month analysis of a handler's field reports during searches to provide PC during traffic stops.

What is not shown is the team during training claims to be at 98% efficient with a 97% accuracy rate. The handler claims the dog made 100% during certification. What these numbers tell us is

  1. training is not realistic
  2. certification is not valid.

If training was realistic, and the certification both would have detected the problem. Since neither did, and the true results of training is what actually occurs in field conditions. it is clear this handler is poorly trained or just doesn't care.

revised 04/26/08