Why rewards work & I'm a force-free trainer

I often get asked about the methods that I use to train dogs and I have a very open and honest policy about which techniques I do and don't use (see my website). My decision to become a force free trainer and join the PPG is not only due to the personal choice that I do not want to inflict pain or cause an animal to feel discomfort or fear, but also because the scientific evidence shows that positive reinforcement is a successful, effective technique to alter the way an animal behaves. 

Before you can decide which side of the (electric!) fence you sit, I thought it would be useful to give a quick refresher on learning theory as it helps to put everything into perspective. There are two 'types' of learning which we can apply to training dogs. I use both of these with pretty much everytime I interact with a dog:

Classical conditioning - is concerned with the way a stimulus makes you feel. It's 'hard wired' so doesn't require you to react consciously, for example, the way the sight of a yummy meal causes you to salivate. We use classical conditioning to 'link' training tools such as clickers or whistles with food/toys so that the dog responds to the conditioned stimulus in the same way.

Operant conditioning - is concerned with the way a stimulus makes you act. It is a more conscious reaction that has been learned from previous associations and relates to particular behaviours leading to particular consequences.There are 4 quadrants to operant conditioning: Positive reinforcement,
positive punishment, negative reinforcement and negative punishment.

The positive/negative can lead to misunderstanding as in this case, it doesn't mean good/bad, it means something being added (+) or something being taken away (-). Reinforcement means that the behaviour will be encouraged and therefore increased (R), whereas punishment means that the behaviour will be discouraged and therefore reduced (P).

I find the easiest way to remember is to think of some examples:

Positive reinforcement is something being added, resulting in an increase in a behaviour e.g. say your dog's name, they look up, you give them a treat. You're adding the treat to increase the  behaviour of paying attention or recalling when they hear their name.

Negative punishment is something being taken away, resulting in a decrease in a behaviour e.g. if your dog is jumping up at a toy in your hand, they aren't given the toy. You're 'removing' the toy to reduce the jumping up behaviour when you have a toy in your hand.

Positive punishment is something being added, resulting in a decrease in a behaviour e.g. if your dog is barking, you spray them with water or a spray collar. You're adding the aversive spray to reduce the behaviour of barking.

Negative reinforcement is something being taken away, resulting in an increase in a behaviour e.g. a harness which gives relief when the dog is walking nicely at heel. You're removing the pain/discomfort to increase the behaviour of walking on a loose lead.

The terminology above can lead to confusion and misleading information within the dog training sector. For example, if trainer A has on their website that they are a positive trainer, using only positive methods that's good right? And if trainer B says they use punishment, that's bad yeah? Hmm, not necessarily. What if trainer A means they are a positive punishment trainer and only use positive punishment methods and trainer B only uses negative punishment alongside positive reinforcement?

As a force free trainer, I only use methods that fall within the positive reinforcement and negative punishment quadrants. I would be lying if I said I never used punishment because every time I withhold a toy/treat/access because the dog is not exhibiting behaviour I want to reinforce (e.g. barking/jumping up/pulling on the lead) I am technically using negative punishment. Withholding a treat/toy or access is as far as I take the punishment side training and I most certainly do not use positive punishment or negative reinforcement of any kind.

One of the most common complaints or concerns I see about positive reinforcement or force free methods is that training with food leads to dogs getting fat and only responding when you have food in your hand. This is absolutely not true. My clients are asked to use their dogs daily allowance of food in their training (but substitute it for higher value treats in certain environments) to ensure that the food they receive in training is not 'extra' food. I also use toys a lot in training to reward dogs and toys are not only calorie free but provide mental and physical exercise. I also teach my clients to use food as a reward, not a bribe. We use luring for very young puppies or for new tricks, but quickly phase out the food lure. The reward needs to occur because of the behaviour, not the behaviour occurs because of the reward.

It would be misleading if I denied that positive punishment 'works' to reduce an unwanted behaviour when executed efficiently. The same scientific evidence that shows the merits and effectiveness of positive reinforcement, also shows the effectiveness of positive punishment techniques such as shock collars. However, there are knock-on consequences to positive punishment techniques and there is often behavioural fall-out when these methods are used because they are based on fear, intimidation and pain. The question of why people are seduced by shock and painful, forceful methods is something I may have to save for another blog (or PhD!) but I think that Ian Dunbar's quote sums it up very eloquently; 

To use shock as an effective dog training method you will need:

A thorough understanding of canine behavior.

A thorough understanding of learning theory.

Impeccable timing.

And if you have those three things, you don’t need a shock collar. 

— Dr. Ian Dunbar

Also, the DEFRA code of practice for the welfare of dogs states that  "All dogs should be trained to behave well, ideally from a very young age. Only use positive reward-based training. Avoid harsh, potentially painful or

frightening training methods". 

In summary, learning theory is very complex and any trainer or behaviourist worth employing should have a thorough understanding of classical and operant conditioning, and which quadrant their training methods are categorised under. In my opinion, there is simply no need to use aversive methods which involve frightening or forcing your dog into doing/not doing a particular behaviour. The vast majority of behaviour cases that I see have been caused or exacerbated by the use of aversive methods. I just cannot fathom why some trainers still follow the more 'old school' fear-based techniques that inflict pain and cause the dog to become more fearful. If a dog is exhibiting fear-based aggression due to a previous aversive experience, why on earth would inflicting more fear make things 'better' and improve the behavioural response? Of course the underlying cause for behaviour is what a reputable trainer or behaviourist would be looking for before even thinking about how to resolve the issues using counter conditioning, desensitisation and environmental management.

I have barely scratched the surface here today and there are lots of links below which you can work through if you want to find out more. I hope that you feel a bit better informed about the learning processes that take place during training and why I choose to use force-free, reward-based methods and all of my dogs and doggy clients are very thankful for that!

Now stop reading this, close your laptop and go outside to play with your dog :)

Further reading

An animal trainer's guide to learning theory: http://www.wagntrain.com/OC/
What are the implications of using training methods that cause pain or suffering in animals? http://www.dogwelfarecampaign.org/implications-of-punishment.php
APBC - Shock Collars, the shocking truth: http://www.apbc.org.uk/articles/shockcollars
PPG - training tools to avoid: http://www.petprofessionalguild.com/petprofessionalresources
Why we don't use aversives: https://www.facebook.com/notes/dog-training-advice-and-support/why-we-dont-use-aversives-by-emma-judson/649293208419573
APDT coe of practice - http://www.apdt.co.uk/members-only/code-of-practice
DEFRA Code of Practice for the Welfare of Dogs - https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/69390/pb13333-cop-dogs-091204.pdf