In the world of dog training, we often refer to balanced and positive reinforcement training as two different things: both being polar opposites that are an incompatible. So lets dig in and discuss what reinforcement in training really is. Spoiler: it’s not always food.
Training labels
Like most of the world, people like to label themselves, and be part of similarly labeled groups. Dog training is no different. There are many labels in the industry – some often meaning the same as others but most completely hinging on the use of quadrants. Anyone that trains with me knows that there is far more to training than quadrants; yet, for some reason many are fixated on it and create a level of toxicity based on these often bizarre and completely misleading labels.
To name a few, we have: +R trainers, reinforcement based, force free, fear free, balanced, punitive, punishment based. However, to burst the bubble most trainers, apart from the extremists fall into the balanced category. They are aware of the quadrants; and educated enough to know that we can’t pick where something will fall into the quadrants but smart enough to understand that we can plan well based on previous experiences. We don’t bully dogs with threats and punishment and accept that food can be coercive and a slip can be just that: a lead.
Reinforcement in training
When it comes to using reinforcement in training, it is not always food. It is something that causes a behaviour to be repeated with improved latency and desire. Hence, the name +R – we are adding a reinforcer that is causing an increase in the behaviour.
Depending on the dog, what is deemed reinforcing will be different! For example, when teaching a recall.
- With Oscar, if I recall him, then throw a ball behind me as he draws level he will chase it. And the next time I recall him, he will run as fast as he can at me, as the reinforcement last time was SO good, he wants to do it again!
- With Bear, if I done that he would be less inclined to run to me, as he doesn’t really care for a ball. However, if I throw a tasty treat to the side of me, it will definitely increase the behaviour next time.
- For Holly, either of these would equally motivate her; both contain movement and that is what her core reinforcement hinges on.
Have a think about your dog, which would they prefer?
Marker training
When looking at reinforcement, it’s also worth considering the association to clicker training. Often, as a mentor I see new trainers click and feed at the same time. I see it in classes all the time too. Or a hand subconsciously tracking towards that treat bag or pocket before they have even marked the event.
This not only causes conflict, but it’s not actually effective clicker training because the marker is not indicating the dog got it right, the dog is just being rewarded and often the clicker / marker becomes a source of frustration of where the f++k is the treat! Marker training should be super clean.
Marker –> get reinforcer –> give reinforcer
If you are saying yes whilst feeding you dog, then it’s not quite right. Both are rewarding – but not necessarily reinforcing. When building behaviour we really want to be specific and seek to reinforce for clarity and speed, not just reward.
Quadrant considerations
In order to build a behaviour, make it reliable and solid we need a solid history of reinforcement. When we use a behaviour to replace an unwanted behaviour, we are technically punishing that unwanted behaviour regardless of how reinforcing the replacement behaviour is to the dog. Whether you like to admit it or not – it’s not possible to purely train a dog and behaviour problems with purely positive reinforcement. It’s also not possible to teach a dog new skills by using purely punishment – there needs to be some level of reinforcement. Whether we like it or not, us good trainers that build behaviours, condition emotions and change behaviours use all of the quadrants. Whether you do so intentionally or not, recording data and videos of your sessions will allow you to plan better in the future to create a slicker system and more responsive behaviour.
