Clicker training is a great way to train, and also fun for the person training and the trainee. While many trainers and behavior consultants use or recommend the use of clicker training and have great success, it’s always helpful to know some of the science behind the technique.
Theory behind clicker training
Clicker training is really interesting in that it’s used to describe both a technique and a philosophy of dog (and cat or other species) training (Feng et al., 2017).
First, let’s chat about the theory behind clicker training technique. In case you aren’t familiar with clickers, they are small devices that typically have a button that is pressed and makes a distinct mechanical click noise. There are now many varieties available, even apps that can make the noise, but I personally like a good old fashioned clicker that’s attached to my wrist with a stretchy band. While I’ll be talking about clickers throughout this, and much of the research has been done with clickers specifically, other noises can be used such as verbal cues or clicks, different tones, and others.
Clicker training uses a couple different aspects of learning theory. The first is positive reinforcement. In positive reinforcement we are giving the animal something in order to increase a behavior we’d like to happen more often. Most of the time, this is through the use of treats.
So, if we’re using treats with clicker training the treat is technically a primary reinforcer – the animal finds the treat itself rewarding and will likely work to get more treats. When we add the clicker to the equation, pairing it with the treats, it may become a secondary reinforcer. The animal knows that when they hear the click, the food is coming. There is debate about this, as technically a secondary reinforcer would become rewarding itself after a number of pairings, and in general it’s not believed that dogs find the click itself rewarding. Despite the debate, dogs do absolutely learn to associate the sound of the clicker and treat delivery.
When pairing, we want to make quite a few repetitions of click and treat to make sure the animal understands really well that the click means the food reward is coming. There is no consistent information in the literature about the number of pairings that should happen. Many of the studies mentioned below used up to 20 repetitions or pairings and felt that the clicker had been appropriately ‘charged’.
There was actually a study done that looked at two groups of dogs, one where the trainer used a click followed by a treat every time, and one where a treat was used after 60% of clicks (Cimarelli et al., 2021). The researchers looked at the learning speed when shaping the behavior of putting two paws on a target, but also the affective state of the dog. The learning speed wasn’t any different between the two groups of dogs, but dogs that were only given a treat some of the time showed a pessimistic bias in a cognitive bias test.
One of the benefits of clicker training is that you can usually click much more quickly than get a treat out, and into, an animal’s mouth. With animals that move quickly this can be helpful as they can already be onto the next behavior by the time you get them a treat. Also, when learning, even a brief delay of one second between delivery of a reward has been shown to impair the speed of learning. Clicker training can therefore allow us to teach more quickly. In this manner, clickers have also been described as being a bridge or a marker for treat delivery.
In a qualitative study looking at perception of clicker training, some of the benefits listed from proponents included having dogs that were more eager to learn, better at problem solving, and dogs that learned tasks more quickly (Feng et al., 2017). In dogs, clicker training was popularized with Karen Pryor’s book, ‘Don’t shoot the dog’ which is a great book covering the basics of learning and clicker training.
As a philosophy, clicker training is typically referred to in relation to positive reinforcement styles of training and by those who advocate against the use of aversive techniques. It’s also used as a philosophy description in relation to creating a good relationship with a dog and when using scientific approaches to training (Feng et al., 2017).
Research in dogs
On the whole, in research so far in dogs, we don’t see the big differences in learning speed when using a clicker compared to simply using food rewards. This could be due to a few different factors. The first is that laboratory studies are really well controlled, and things can be precise. With applied studies in pet dogs and with different trainers, there can be so many issues with timing, distraction, previous learning, and many more that can affect results. In order to get a really accurate understanding of what is happening, we need huge studies looking at many different trainers and many dogs, as well as very controlled and precise ones. At the moment, we don’t have this. Another factor could be that dogs actually don’t learn any faster with clickers than with simply food rewards.
In saying that, let’s look into some of the research that has been published. The first study that looked at clicker training in dogs included a number of pet Basenjis taught to nose touch a cone (Smith & Davis, 2008). This study compared those trained with a clicker and food vs simply food, and they used luring to get the behavior. There was no difference in the amount of time or number of trials it took the dogs in the different groups to learn the new behavior. The researchers also looked at extinction of the behavior, where they stopped rewarding the behavior. They continued to click without giving food, and simply did not give food to the dogs in the food only group. Those still hearing the click took longer for extinction of the behavior.
Another study compared training methods, looking at a group of dogs trained with a clicker and food, one trained just with food, and a control group that didn’t receive any additional training (Feng et al., 2018). During the study the dogs were taught a variety of tricks first by a trainer and then by the owner who was given instructions. These included: nose target a hand, nose target an object, spin in a circle, chin rest on the ground, play dead, and station on a mat. While the authors wrote that positive reinforcement was used with no aversives, the exact methodology of training was not detailed. The dogs in the two training groups for the most part had similar levels of success with training, except for a nose target that was reported to be less challenging by those in the clicker training group. The relationship between the owner and dog, as well as dog impulsivity, also didn’t differ between the two groups. Those that were using the clickers did not report more difficulty with training than those not using clickers.
Another study looked at training new behaviors to pet dogs and compared clickers, a verbal marker (the word ‘bravo’), and just food (Chiandetti et al., 2016). The dogs were first taught to touch their nose to a handle, and then this was generalized to a new device. There was no difference in the length of time it took to learn new behaviors between the three groups.
And finally, one last study looked at the use of clickers in shelter dogs, but also using the theory that clickers are beneficial when using shaping (Dorey et al., 2020). The authors trained both a stay and a wave, using a specific set of successive approximations to reach the behavior. Dogs were assigned to three different groups – training with a clicker and food, a verbal marker (the word ‘next’) and food, or food alone. The researchers did not find that either the clicker or the verbal marker outperformed the food alone. There were differences with the two behaviors though, and they suggest that different reinforcement strategies might be more or less effective depending on the type of behavior you’re training.
Future Research
There is definitely a need for further research in this area, looking at dogs with varying levels of experience with clickers, training different behaviors, looking at the methods and timing of trainers, and more.
Despite the inconclusive research results, I still highly encourage clicker training in clients. In my experience, it’s a noise that can stand out from background noise, it helps to get people into a training mentality, and both people and their dogs seem to enjoy it.
If you’d like a less technical version of this to share with clients, one can be found at the Landmark Behaviour blog that is written for pet owners.
References
Chiandetti, C., Avella, S., Fongaro, E., Cerri, F. 2018. Can clicker training facilitate conditioning in dogs? Applied Animal Behaviour Science. 184:109-116.
Cimarelli, G., Schoesswender, J., Vitiello, R., Huber, L., Viranyi, Z. 2021. Partial rewarding during clicker training does not improve naïve dogs’ learning speed and induces a pessimistic‑like affective state. Animal Cognition 24:107-119.
Dorey, N.R., Blandina, A., Udell, M.A.R. 2020. Clicker training does not enhance learning in mixed-breed shelter puppies (Canis familiaris). Journal of Veterinary Behavior. 39:57-63.
Feng, L.C., Howell, T.J., Bennett, P.C. 2017. Comparing trainers’ reports of clicker use to the use of clickers in applied research studies: methodological differences may explain conflicting results. Pet Behaviour Science 3:1-18.
Feng, L.C., Hodgens, N.H., Woodhead, J.K., Howell, T.J., Bennett, P.C. 2018. Is clicker training (Clicker + food) better than food-only training for novice companion dogs and their owners? Applied Animal Behaviour Science 204:81-93.
Smith, S.M., & Davis, E.S. 2008. Clicker increases resistance to extinction but does not decrease training time of a simple operant task in domestic dogs (Canis familiaris). Applied Animal Behaviour Science. 110:318-329.