The Effect of Training Methods on the Efficiency of Learning
Punishment based dog training results in reduced learning.
Posted April 29, 2014
While I was attending the Western Psychological Association meeting in Portland, Oregon, I was approached by a psychologist who does some consulting work for the US Army concerning the training of military service dogs. He told me that he was a regular reader of my work, and he wanted to talk about my recent article describing how so-called "discipline-based training" methods, which use punishment and physical force, seem to result in increased stress in dogs (click here to see that article).
He told me, "I recognize that you are concerned about stressing the dogs, but the use of an occasional punishment focuses the dog on the trainer and the task that they have been assigned and that, ultimately, produces better learning and performance in the dogs. When the dog is no longer in training those stress affects will dissipate, but the better learned behavior will be left behind."
Unfortunately both of his major contentions, namely that dogs learn better when physical and punishing techniques are used, and that when such methods are used it results in better attention on the part of the dogs, appear to be wrong. I directed him to some research on the training of military service dogs that was conducted a few years ago. The team of researchers was headed by Dr. Anouck Haverbeke of the University of Namur in Belgium, and was published in the journal Applied Animal Behavior Science*.
That study involved 33 military dog and handler teams from the Belgian Army. The dogs were selected to be representative of the military canine population regarding sex and breed (Belgian shepherds and German shepherds). The dog handlers worked with their dogs and used whatever rewards or punishments they thought to be effective in their training. The positive training procedures involved rewarding the dog by stroking or patting him, verbal praise, toys, and food. The more negative techniques involved punishing the dog for poor or wrong behavior by a variety of methods ranging from speaking harshly or yelling, up through tugging sharply on the leash using a choke or prong collar, hanging the dog by their collar, hitting the dog with the hand or the leash, or using an electric shock collar. All of the handlers used some mixture of these techniques, however for certain trainers the proportion of punishments was significantly greater, while for others the predominant interactions involved more positive rewards.
Each of the dog and handler teams was tested twice in an open field that was specifically set up for assessing how well the dogs performed. The evaluation included eight obedience exercises (Walk-at-heel, Sit, Down, Stand, Positions-at-distance, Recall-to-heel, Down-out-of-view-of-the handler, Jump) followed by five protection work exercises (Handler’s-defence, Attack, Attack-with-gunshots, Attack-with-threatening-behaviour, Stand-off). These test exercises are versions of those tasks that the dogs are expected to know after training. All the tests were videotaped so that the team's performance, the handler's behavior, and the dog's behavior could be measured. To analyze the dog's accuracy of performance the number correctly performed exercises were tallied and the numerical scoring method used by the Belgian army was applied.
In addition to performance, the amount of attention that the dog paid to his task and to the handler was measured. Dogs that failed to look at their handler or to look in the direction that their task required were considered to be distracted. Finally, some measures of the dogs stress were included, the most sensitive being whether the dog was carrying his body and a low cringing posture.
Consistent with previous research, the dogs that received punishment showed signs of stress in the form of significantly lower body posture — but that was to be expected. Let us now turn to the actual performance of the dogs.
The dog and handler teams were divided into two groups, a high performance group with good accuracy and control, and a low performance group. When the researchers looked at the amount of punishment administered to the dogs, they found that the dogs who received the most punishment were significantly more likely to be in the low performance group. In other words they did not learn as well.
Another important finding is that the dogs who received more discipline and punishment were found to be more distracted, paying less attention to their handler and the tasks that they were supposed to complete. This makes complete sense, since the handler and the task have been associated with numerous punishments. Thus simple conditioning of emotional responses would say that the handler and the task are much more likely to evoke negative emotions, so in order to protect themselves from feelings that are unpleasant, the dogs avoid looking at the handler and items associated with the tasks that have caused them discomfort in the past. This may, in fact, be the primary reason why the punished dogs do not perform as well, since attention also reflects the dog's motivation to learn. When the experimenters looked at the high performance versus low performance dogs they found that the most distracted dogs were significantly more likely to be in the low performance group.
So contrary to the beliefs of my psychological colleague, and apparently contrary to the beliefs of some military service dog trainers, the use of discipline-based training, including physical and punishing techniques, does not produce a better trained dog who is more focused on his handler and the tasks he has been set. Rather based upon these experimental results we can draw a simple relationship which says punishment produces distraction, and that poor attention results in worse learning by the dogs.
Stanley Coren is the author of many books including: The Wisdom of Dogs; Do Dogs Dream? Born to Bark; The Modern Dog; Why Do Dogs Have Wet Noses? The Pawprints of History; How Dogs Think; How To Speak Dog; Why We Love the Dogs We Do; What Do Dogs Know? The Intelligence of Dogs; Why Does My Dog Act That Way? Understanding Dogs for Dummies; Sleep Thieves; The Left-hander Syndrome
Copyright SC Psychological Enterprises Ltd. May not be reprinted or reposted without permission
* A. Haverbeke, B. Laporte, E. Depiereux, J.-M. Giffroy and C. Diederich, Training methods of military dog handlers and their effects on the team’s performances. Applied Animal Behaviour Science, 113 (2008) 110–122