The Best Way to Train Your Dog According to Science

Woman w dog

Could Confucius’ quote shed light on dog training?

Clicker training is widely popular method by professional dog trainers, but is it the most successful and efficient way to train your canine best friend?

Researchers out of Eötvös Loránd University in Hungary, Claudia Fugazza and Dognition’s Dr. Ádám Miklósi, looked into the most successful and efficient way to train a dog. Fugazza and Miklósi focused their attention on two methods of training: shaping/clicker (SHA) and Do as I do (DAID). Shaping/clicker is an operant conditioning approach that uses a clicker to reinforce a dog’s behavior. Do as I do is a social learning approach that has the dog imitate the human’s actions.

They assessed dog-owner pairs to see the success and efficiency with both methods in 3 categories: simple actions, complex actions, and sequences of two actions (Table 1). All actions chosen for this study are object-related actions, which are particularly useful for assistance dogs to learn.

Table 1

The Subjects

The researchers recruited a total of 30 dog-owner pairs for the study. Fifteen of these dog-owner pairs had previously earned a certificate for shaping/clicker, while the other half had one for Do as I do. Each group (SHA and DAID) had a diverse mix of dog breeds, from Border Collies to mixed breeds (full list of dog breeds for each group in the study (paywall)). The dogs’ ages ranged from 2 to 11 years old in both groups with the mean age being 6 years old. All dogs involved practiced some sports and training activities with their human partner.

The Test

Each trainer had 15 minutes to teach their dog a specific novel object-related action. The timer was stopped the first time the dog demonstrated the task. If the dog did not demonstrate the action within 15 minutes, that trial was counted as a fail. Each human trainer taught their dog 3 separate tasks: one simple, one complex, and one sequence.

Simple tasks involved the dogs touching a specific object. Complex tasks needed more elaborate manipulation of an object. Sequence tasks consisted of two actions. The researchers doled out the same set of training tasks to pairs of subjects (i.e. they matched up a clicker trainer with a “Do as I do” trainer for comparison sake). The researchers made sure that the training tasks were novel to each dog, which is why the same set of training tasks were not given to all subject pairs.

The Take Home

The researchers found no significant difference between success or efficiency in training dogs in simple tasks. Whether using DAID or clicker training method, all 30 dogs succeeded in demonstrating the touch-related tasks within 15 minutes with similar efficiency.

The real difference occurred in training for complex and sequence tasks.

With complex tasks, all DAID group dogs succeeded in demonstrating the complex task within 15 minutes, while 11 out of 14 dogs were able to demonstrate the task within 15 minutes with clicker training. The significant result is that the DAID group dogs were much quicker to demonstrate the complex task than those dogs with clicker training, 55.71 seconds and 356.18 seconds, respectively. In other words, it took dogs under a minute to demonstrate the predetermined complex task by social learning and close to 6 minutes for dogs learning by clicker training.

With action sequences, significantly more dogs were successful at demonstrating the predetermined action in the DAID group. In particular, 13 dogs out of 15 in the DAID group were able to demonstrate the task within 15 minutes; this is compared to 7 dogs out 15 in the clicker training group.

Table 2 sums up the results above:
Table 2

Things to Note

This is the first study to show the efficiency of Do as I do method. Though, the researchers did not factor in the time it took for the dogs to learn the initial imitation rule. All dogs in the DAID group were familiar with the imitation rule prior to the study. As well, clicker training relies on dogs learning by trial and error, which is a slower process. The researchers suggest that further studies should investigate the consistency of the dog’s behavior after demonstrating the task for the first time and on its resistance to extinction.

The researchers also noted that the tasks were object-related actions, and the results should not be extended to training of body movements.

At Dognition, we think this research should be very intriguing to those who train assistance dogs. With time being a valuable resourceful, Do as I do could be a more viable method in teaching service dogs in a time effective manner. Yes, there’s some initial set-up work in that the dogs need to learn the imitation rule, but once the dogs have learned this rule, think of how many more dogs could be trained in the same time period as dogs who were being trained with clicker! That’s a savings in time and in money for assistance dog organizations. We’ll definitely be staying tuned to further research in this subject matter!

We want to hear from you! What are your thoughts or questions about this study?