Doggie See, Doggie Do?

At my training center, AutumnGold, it is not unusual to enroll students who live with and train more than one dog. A common question that these clients have is how to arrange their training sessions to allow them to train one dog while the other dog “waits his or her turn”. In most of these cases, the student laments that the dog who is not chosen for training becomes upset, does not enjoy being isolated or confined, and may even show some separation stress or frustration with the overall unfairness (in their opinion) of the entire situation.

Or, if they are clever, as my four dogs appear to be, they attempt to all participate in the training session simultaneously.

Ally, Cooper, Chip, Vinny on Pause Table


Over the years, my personal solution to this problem has been to teach each of our dogs to stay on a pause table located off of the training floor while they await their turn to train. Although it can be challenging to teach (more about this in my next blog), I like this arrangement because it is convenient and saves me from having to return to the house multiple times to get a different dog. It is also fun for the dogs because they generally receive more training time and also get to play together after the session.

Chip Cooper Pause Tables


And it appears that there may be additional benefits for the dogs.  A recent research paper provides evidence for something that I anecdotally have observed with  my own dogs and suspect that other trainers have also experienced – that dogs seem to have some ability to learn new tasks by observing other dogs.  This type of learning is called social learning (or observational learning) and it has been studied in a variety of species and contexts. However, only recently has it been studied with dogs in a training environment.

Ally learning wave dogs pause table


Social Learning: It is generally accepted that social learning plays an important role in the lives of dogs. Observing the behavior of others helps dogs to learn about their environment, modifies their responses to new situations, and may even teach them new behaviors and solutions to problems.  Within the broad category of social learning, several sub-classifications exist and these presumably reflect different degrees of cognitive involvement. Although there is debate among social scientists about definitons, the types include: social facilitation, local/stimulus enhancement, response facilitation, social emulation and imitation (1). Much of the debate among those who study social learning revolves around what information the dog is actually using and how that information is processed cognitively or consciously to change behavior. For the purposes of this essay, we are most interested in how social learning (of any type) might occur between dogs in training situations.

Mimicry Cooper and Ally


Interestingly, much of the research with social learning in dogs has focused on their ability to learn by observing human demonstrators. Common study paradigms ask dogs to either solve a food acquisition puzzle or to maneuver around a fence after having watched a person perform the correct solution. Dogs have been shown to be quite successful at these tasks, although factors such as the identity of the demonstrator, the dog’s living situation, prior training experience and age can influence success rates.

Studies that have examined the dog’s ability to learn socially from other dogs are fewer in number. There is evidence that, just as with human demonstrators, dogs will show improved performance and problem-solving ability after watching a demonstrator dog successfully complete a detour or puzzle task. An earlier study reported that puppies who were allowed to observe their mother working at a scent detection task went on to become more successful as scent dogs in adult life.

Can Dogs Learn a Trained Task from Another Dog? Most recently, researchers asked whether on not learning would be enhanced in dogs who observed another dog who had been trained to perform a novel behavior (2). This study and its results have relevance to those of us who train multiple dogs and perhaps even for group training classes.

The Study: The objective of the study was to determine if learning a new task (jumping onto a trunk or small slide) was enhanced when dogs had the opportunity to observe another dog performing the exercise prior to being trained themselves.  A group of 33 adult Labrador Retrievers was tested. All of the dogs were enrolled in the training program at the Italian School of Water Rescue, lived in homes with their owners, and had been pre-screened to ensure that the exercises used in the study were novel to them. The dogs were divided into two groups; an observer group and a control group. Observer dogs sat next to their owners and watched another handler and his trained demonstrator dog approach the obstacle (trunk or slide). The demonstrator dog then jumped up and sat on the obstacle. The exercise was demonstrated twice and then the observer dog and handler approached the obstacle and the handler attempted to train the dog to perform the exercise. The control group of dogs were held on lead in the same position but did not have the opportunity to watch the demonstrator dog before being trained for the new task. Each dog’s success or failure to demonstrate the new behavior was recorded.

Results: The dogs who had the opportunity to observe a demonstrator dog perform the new exercises were significantly more likely to succeed at the same task when asked to perform it. Specifically, 62.5 % of the observer dogs were successful compared with 23.5 % of the control dogs. Neither a dog’s sex nor his/her level of prior training experience influenced the probability that they would perform the new task successfully. Age was somewhat important, as older dogs tended to be more successful than younger dogs.

Take Away for Dog Folks

On one level, these results are not surprising. Anecdotes abound among dog folks regarding our dogs’ ability to learn from one another through observation. Ask anyone who lives with more than one dog and they will relate numerous examples of their dogs sharing information (and not always in a good way). For as long as I can remember our dogs have learned to “wait” at the door and in the car by watching each other. While I do train this command, our young dogs learn to wait very rapidly when they notice that the entire family is frozen in its tracks. Similarly, because we hike a lot with our dogs, one dog finding something yummy or smelly on the trail is quickly observed and acted upon by the others. Still, these examples may arguably fall relatively low on the social learning scale as they probably reflect social facilitation or simple stimulus enhancement.

What is exciting about the recent results is that they demonstrate, albeit with a small number of dogs, that a dog who has the opportunity to observe another dog who is performing a trained exercise (and by extension, perhaps the training of that exercise itself?) can benefit from that observation. The researchers provide several possible explanations for their positive results, one of which is that dogs may show enhanced learning when they are highly motivated to engage in the task. For example, teaching something that is target-oriented, such as jumping up onto an obstacle or retrieving a toy, may be more successful than training a static exercise such as a down stay. This difference is reflected in the results of a previous study reporting that untrained dogs did not perform well in learning a positional behavior (lying down on command) after watching a trained dog perform it (3). While dogs are often naturally interested in examining and engaging with new objects, most are decidedly less motivated to spontaneously offer a static behavior such as a sit/stay or down/stay.

These results make me consider the relative ease with which both Ally and Cooper have learned platform positions, retrieving tricks such as “put your toys in a basket” and go-outs to a target. Both regularly watch and get excited as the other is being trained in these behaviors. (And, not to put too fine a point on it, I rarely see that level of interest or excitement when I am training sit/stays and down/stays). While I have no control group for my own anecdotal experiences, these results suggest to me that having all of my dogs present and attending during a training session may have benefits that go beyond convenience. Watching the other dogs learn new things may help my observing dogs to learn more rapidly, at least in those exercises that interest and engage them.

And, for my clients who are attempting to train two dogs at the same time, I will now recommend that, if possible, they find a way that the dogs are able to observe each training session with the other dog. They may be very pleasantly surprised at the results.  Happy training everyone!


Cooper and Alice Standing Platforms


Cited References:

  1. Kubinyi E, Pongracz P, Miklosi A. Dog as a model for studying conspecific and heterospecific social learning. Journal of Veterinary Behaivor 2009; 4:31-41.
  2. Scandurra A, Mongillo P, Marinelli L, Aria M, D’Aniello B. Conspecific observational learning by adult dogs in a training context. Applied Animal Behaviour Science 2016; 174:116-120.
  3. Tennie C, Glabsch E, Tempelmann S, Brauer J, Kaminski J, Call J. Dogs, Canis familiaris, fail to copy intransitive actions in third-party contextual imitation tasks. Animal Behavior 2009; 77:1491-1499.

Excerpted from: “Only Have Eyes for Your: Exploring Canine Research with the Science Dog” (2016).


20 thoughts on “Doggie See, Doggie Do?

  1. Pingback: Wait. Should I like that? – The Science Dog

  2. Pingback: Wait. You Can Eat THAT? – The Science Dog

  3. Pingback: Doggie See, Doggie Do? – K9 Jocks

  4. Pingback: Just Show Me A Sign – The Science Dog

  5. Pingback: Manners Minder and Me – The Science Dog

  6. Pingback: Weekly Woof from the Web | AniEd Ireland

  7. I think they do more than just copying, too. Year ago I was trying to teach my ‘trialling’ dog how to heel while she was carrying the dumbbell. I felt that I was just banging my head against a brick wall! Talk about not chewing gum ad walking up some steps!
    But the little old Kelpie was sitting and watching and quivering with the “Give me a go!” behaviour. so I swapped dogs, and Little Old Sammy did it perfectly. After watching him for a while, Kelly was able to heel and carry the dumbbell too!

    Liked by 1 person

    • I’ve always felt, (even before this incident) that in dog clubs, the beginners classes would do much better if those with old retired dogs joined the class too. I know how much my old dogs have missed “training” and I try to always give them a session at home — just the ‘same old’ but they find it important, I think.

      Liked by 1 person

  8. HI Linda

    I enjoyed your piece “Doggie See, Doggie Do?”. As you described the study done with the group of Labradors, I wondered what impact the individual handlers had on their dog’s learning ability. It seemed that in the study you presented, the owners of each dog were the instructors of their own dog. In my experience there is a wide variation in communication skill among dog owners. I think it would be difficult to determine how much of the difficulty of learning the behavior was due to not seeing another dog, or in the handler’s ability to communicate with their dog. Also, the trust between handler and dog would vary. Anyway, some thoughts that I wanted to share.

    Thanks for your interesting posts!

    Susan Signor, Certified Dog Behavior Consultant CBCC-KA, ACDBC, CPDT-KSA, KPA-CTP 509-594-6337



    • Hi Susan – The handlers were all inexperienced trainers, though I am not certain what that actually meant, since the dogs were being trained in Water Rescue. Regarding variation among trainers in ability; such differences, just as differences among dogs in terms of learning aptitude, temperament, etc, would be accounted for by the design of the experiment. The use of groups of dogs plus control groups, allows reserachers to statistically remove variations that not related to the treatment being studied (these are called error terms). Remember that the differences in training ability among owners would vary around a partilar mean and that this variability would be represented in both the control and the treatment group. Conclusions are then drawn by comparing the two groups, thereby “cancelling out” the effects of trainer and dog variation. While researchers cannot account for every single variable that might affect response, using a well-designed experiment and balancing the groups effectively accounts for factors such as differences in training skill, ect. That said, I think it is important to point out (as I did in the essay) that the groups were pretty small in this study, which can lead to a statistical problem called “the problem of small numbers”. So, personally, I would look at this particular paper as a promising pilot study and would love to see more studies of this type! (If you are interested in learning more about how these studies are conducted and controlled to account for issues such as the one you mention, this is exactly what I will be discussing in my PPG webinar on Tuesday – would love to see you there!). Thanks for posting – Linda


  9. Pingback: Can Dogs Learn a Trained Task from Another Dog?...

  10. My example is interspecies – when I brought my second pup home I did lots of work on polite turn taking to avoid problems down the line. I asked both dogs to sit, and gave each a treat, naming the dog it was for. Tille-cat watched this game for a while, worked out that treats were apparently being handed out simply for parking bums on floor, and next time I looked was sitting sedately next to the dogs, awaiting her turn…

    Liked by 3 people

    • I used to work with a woman who was rearing an orphaned flying fox “fruit bat”. She told me that at meal times and when her dogs were otherwise interacting with her, the little flying fox would try to do the same as the dogs 🙂

      Liked by 1 person

  11. I work with breed rescue and have often seen a foster dog look up sharply at a new sound and watch the reaction of my own dogs, then follow it. Even a feral pup, the first day here, hid and watched, and watched as I sat outdoors with her and my own dogs, playing with them, petting them, and talking quietly over the fence with a neighbor. Finally, she got up, came over to the stool I was sitting on, crawled underneath, and heaved a great sigh. And that was the end of her distrust. (She was only about 8 or 9 weeks old, though.)

    Liked by 2 people

  12. You should also look at the research by Dr. Claudia Fugazza, who did her PhD work on this using her “method” called Do As I Do, which takes the idea of social learning to the interspecies level. It is of course based upon the original theories and work by Arthur Banduraher work starting in the late 90’s, plus the further work at Adam Mikolsi’s Hungarian Ethology institute. I was in contact with her yesterday and she told me, that the research concerning social learning is not finished at the insititute, now that she has her PhD. that she has a project underway – sorry, no leaks from me concerning her work. Personally, the implementation of other techniques from others beside Skinner & Co. is long, LONG overdue. I’ll be looking at what “we” now say we use for fear reduction (and hopefully showing that it’s probably NOT what’s going in under the hood according to more current research) as well as making some suggestions of what “we” might borrow from other techniques.


    • Yes, I am aware of Dr. Fugazza’s work. Her research examines interspecies (dog and human) observational learning, while the topic of the current essay examines intraspecies social learning in a training situation. I will be writing about Dr. Fugazza’s “Do As I Do” research in a future essay – stay tuned! Linda Case


Have a comment? Feel welcome to participate!

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s