30/11/2022
Really great explanation on why we don’t agree with ‘ balanced’ training
THE PROBLEM WITH “BALANCED” TRAINING
A friend and colleague sent me a video this morning of a group of “balanced” trainers working with a dog in a shelter, supposedly demonstrating how “corrections” are not punishment, but guidance or non-verbal communication. The first couple of minutes of the video show a very confused Rottweiler being taken out of a cage in the shelter on a slip lead and walked around amongst the group of trainers. The handler feeds the dog treats and pats the dog “making friends” with her and assessing how she will likely respond to him during training (this is what the commentary tells us). Then a second dog is brought into the area and as the Rottweiler starts looking at this dog, the trainer yanks on the slip lead so hard that the dog comes off her feet. He begins walking her around and repeatedly yanking her off her feet. I stopped watching after a few seconds of this, as it was utterly sickening and gave me flashbacks to a previous trainer I worked with many years ago, who I left because he started doing things like this. I refuse to watch abuse.
So, what is “balanced” training? The term BALANCED is likely a label used in reaction to the POSITIVE label used by many trainers, indicating that they continually aim to use positive reinforcement (adding a pleasant consequence e.g. food, toys, play or anything the dog enjoys to increase behaviour) as a training method, rather than positive punishment (adding something nasty or scary e.g. yanking a dog off their feet to stop or reduce behaviour) or negative reinforcement (taking away something nasty or scary when the dog complies with what you want, to increase behaviour e.g. ceasing yanking the dog when one perceives the dog to be doing what one wants). The label “balanced” indicates that the aim is to use a mix of positive reinforcement and positive punishment (corrections). So balanced trainers may use food and shock collars on the same dog. Put simply it is a “carrot and stick” approach, rather than a carrot only approach or stick only approach.
A lot of people think this sounds like a good idea. It is all very well to reward a dog when they do something you like, but surely you have to also be able to “correct” them when they do something wrong? Well, the science does not support this. Research shows several things about the use of positive punishment in training:
1. It increases stress levels in dogs: https://www.psychologytoday.com/za/blog/canine-corner/201404/the-effect-training-method-stress-levels-in-dogs
2. It is NOT more effective than positive reinforcement: https://www.researchgate.net/publication/261106650_Dog_training_methods_Their_use_effectiveness_and_interaction_with_behaviour_and_welfare
3. It has a negative effect on the relationship between dog and owner: https://www.psychologytoday.com/za/blog/canine-corner/201910/how-training-methods-affect-dogs-attachment-its-owner
4. It increases aggressive responses in dogs: https://www.psychologytoday.com/za/blog/canine-corner/201205/is-punishment-effective-way-change-the-behavior-dogs
In addition to this evidence, I also want to explain why trying to combine positive punishment (and negative reinforcement) with positive reinforcement is fundamentally a flawed notion. To do that I need to talk a little about what is happening in the brain during these types of learning. Stay with me though, as I am going to explain it in very simple terms:
Positive reinforcement adds something pleasant as a consequence of behaviour in order to increase the likelihood of that behaviour occurring again in the same situation. Working (performing a behaviour) in anticipation of gaining something pleasurable (for dogs: food, toys etc) occurs in the Behavioural Approach System or SEEKING system of the brain. In other words, a certain type of brain activity is switched on when an animal is engaged in a goal-oriented task that is likely to end in gaining something good. Activation of this system is associated with anticipatory pleasure – much like we feel as we sit down to our favourite meal. When we train dogs with positive reinforcement, we are activating their Behavioural Approach or SEEKING system (they are working for food or the chance to chase, grab and bite a toy) and so are highly motivated to engage and “problem solve” to gain rewards.
On the other hand, when we use positive punishment (and verbal or physical corrections are punishment – don’t be fooled by euphemisms), we are adding something unpleasant (painful or scary) in order to stop a behaviour or reduce the likelihood of it occurring again in the same situation. Positive punishment causes a very different type of brain activation: The survival System or 5F System (Fight, Flight, Fiddle About, Freeze or Faint) kicks in. This type of brain activation is pretty much designed to save us from death. Its first line of defence is to trigger innate species-specific behaviour patterns designed to get us out of trouble quickly (the 5F’s), which is why we see an increase in aggression (and also avoidance or escape behaviours) with the use of punishment. We then learn which behaviours are more successful at escaping the nasty thing e.g. if a dog lunges on lead and the other scary dog walks away, then lunging worked and will be repeated (negative reinforcement). There is great urgency associated with the Survival System - if you are not quick you could be dead. Not a lot of rational thinking or healthy problem solving takes place here, so this type of brain activation does not create a healthy learning “environment” for dogs. They may learn how to escape danger, but they are not going to learn much more than that and the physiological and emotional toll from the activation of stress hormones that go along with this type of brain activation, makes this an extremely “expensive” training method in terms of welfare.
The balanced trainer’s philosophy implies that one can somehow mix or combine these two types of brain activation for a “healthy” balance, but this is not the case. Interestingly, the two systems cannot operate at the same time. You cannot be simultaneously engaged in a pleasurable activity and be escaping danger. Activation of the Survival System does not allow for simultaneous activation of the Behavioural Approach System. Imagine a buck fleeing from a lion, only to stop and start grazing on the way, because she happened to run over a particularly lush bit of grass. Ridiculous, right? In order for the brain to be able to engage in pleasurable goal-oriented behaviour, it first has to “relax” after the threatening event. If the frightening situation is not resolved and the dog still feels under threat, the SEEKING system will be inhibited, and you will not be able to use positive reinforcement effectively. It seems rather obvious that this is an incredibly inefficient and cumbersome way of training. In contrast, positive reinforcement training attempts to work continuously in the Behavioural Approach System, where the dog feels good, the learning environment is optimal, the dog is able to think more clearly and there is no physiological fallout from stress hormones.
So why do balanced trainers sometimes seem to get “good” results? Well, for much the same reason that people stay in abusive relationships. Abusive relationships are usually characterised not only by abuse, but also by opposite extremes of “love”, attention and gifts. Abusers are effectively using “balanced” training on their victims. This creates a rollercoaster of emotions due to the unpredictability of the situation, a dependence on the abuser and a fear of doing anything that could trigger the abuse. Many dogs that appear “well-behaved” are simply too terrified to do anything at all (shut down emotionally and behaviourally) and simply “obey” with robotic obedience or inhibit most behaviours to avoid punishment.
None of us are perfect and I am not saying that there won’t be times where we have a seriously bad day, get frustrated and raise our voices or that we won’t get a fright and grab hold of our dogs to keep them safe. However, the difference is that positive reinforcement trainers will see this as a MISTAKE, ERROR, OVERSIGHT or HUMAN WEAKNESS and something we should do everything possible to avoid in the future. We will consider what we should have done differently to prevent that situation from having arisen in the first place and make plans to avoid it happening again. We will never see using force, intimidation or pain as an acceptable training method, because “balanced” training is not healthy – it is abuse.
Photo credit: Peter Merle
