Reward based training

The Australian Veterinary Association (AVA) recommends the use of positive reinforcement training methods for dogs as the preferred method of training.

INTRODUCTION

The use of positive punishment (applying something the dog doesn’t like to make behaviour less likely to occur in the future) is often a first approach used in dog training and for modifying problems in behaviour. However this method is not preferred by most behavioural specialists as dog training techniques that a dog doesn’t like can be dangerous for both owners and dogs.

HOW DOGS LEARN

It’s important to recognise that learning is an adaptive process used by a dog to cope with a changing environment. When there is a change in the environment the dog will respond physiologically or with a change in behaviour.

Some responses are instinctive (such as a newborn pup seeking a nipple), some responses are due to the environment (such as drinking when thirsty), and some responses are due to learning. The consequence of this response will influence the likelihood of the behaviour occurring again.

In many ways dogs learn the same way we do, by making connections between neurons. Pathways are strengthened through repetition and are weakened through lack of use.

Classical conditioning is an involuntary process that happens when a dog makes an association between a previously “neutral” object (such as a set of car keys) and a reflexive action or emotional response (such as the excitement of the possibility of going to the beach). The dog has no control over its behavioural response.
Instrumental conditioning

takes place when an association is made between a neutral stimulus (such as the word “sit”) and a voluntary response (placing bottom on ground). This type of behaviour is modified by the consequence of the dog’s response. For example, if the consequence is rewarding, the behaviour is more likely to increase in frequency. If the consequence of the chosen behaviour is unrewarding to the dog, the behaviour is likely to decrease in frequency. Dogs, like all animals, learn to respond to stimuli that enhance the possibility of a good outcome.

Rewarding means that the dog receives something it wants, or values, as a consequence of its behaviour. These are called “reinforcers”.

Primary reinforcers are essential for survival, such as food, water and reproduction. Secondary reinforcers are neutral things such as words which have been associated with primary reinforcers to become desirable to the dog. For example, the words “good dog” or the sound of a clicker means nothing to a dog until the dog has associated it with a treat for example.

PUNISHMENT AS A TRAINING METHOD

Punishment uses the principle that if the consequence is painful or unpleasant, the dog is less likely to behave that way again. This can involve two strategies:

Positive punishment is applying something the dog doesn’t like to make the behaviour less likely to occur again. For example, smacking a dog on the nose (an aversive stimulus) when it approaches to take a biscuit off the table reduces his scavenging behaviour around the table.
Negative punishment means taking away something the dog wants to make the behaviour less likely to occur again. For example, removing all biscuits from the table as the dog approaches the table (removal of something the dog wants) reduces the likelihood of the dog scavenging around the table.

Pitfalls of positive punishment

  • Positive punishment has to be delivered immediately and consistently and be of the appropriate intensity to be effective in reducing the likelihood of a behaviour occurring (Overall 2006). This can be very difficult to achieve.
  • It stops a behaviour but fails to replace it with a desired behaviour, thus creating a “behavioural vacuum”.
  • The dog doesn’t learn what is “right” or what is wanted from him.
  • It’s likely to increase anxiety in dogs (Schilder 2004, Overall 2007).
  • It can lead to defensive aggressive behaviour in many species of animals, putting the person administering it or any person or animal near the dog at risk of being bitten or attacked (Hetts 1999).
  • Animals can habituate to punishment so the owner may feel the need to escalate the intensity of the punishment or to cause physical injury in order to be effective (Landsberg 2005).
  • Punishment may be used as a substitute for losing one’s temper and lashing out.
  • Punishment may elicit a strong fear response and this response can generalise to things that sound or look similar to the punishment.
  • Learned helplessness can develop (i.e. depression, learning problems, decreased motivation) (Lindsay 2000).
  • It doesn’t support a healthy humananimal bond.

REINFORCEMENT AS A TRAINING METHOD

There are two types of reinforcement techniques:

Positive reinforcement involves giving the dog something it wants to make the behaviour more likely to occur again. This occurs if asking the dog to sit and stay and giving it a reward (something the dog wants) for staying away from the biscuit on the table increases his likelihood of doing as asked again.
Negative reinforcement involves removing something aversive to make the behaviour more likely to occur again. This occurs if holding the dog back from the table by a tight choker chain (an aversive stimulus), and releasing the pressure when the dog sits down and stops leaning in towards the biscuits. This results in the dog being less likely to try and approach food on the table in the future. Note that an aversive stimulus had to be present (the choker) for the dog to have to work to escape, so negative reinforcement can also be called escape or avoidance learning.

The use of positive reinforcement is the most humane and effective training method as it avoids undesirable behavioural side effects. It also makes training more enjoyable and helps to improve the bond with the pet.

PROBLEMS WITH USING PUNISHMENT IN TRAINING

Training methods that use positive punishment and negative reinforcement have been linked with undesirable side effects for dogs and behavioural problems such as escape and avoidance behaviour (to avoid the punishment), aggressive behaviour (in self-defence), response suppression (habituation or learned helplessness) and fear of people or things in the environment where the aversive stimulus was present (fear conditioning and generalisation). (Blackwell 2008)

HELPFUL ADVICE FOR DOG OWNERS

Other strategies to modify behaviour include giving advice and help to owners such as:

  • Seeking the help of a qualified trainer who uses reward-based training techniques.
  • Having the dog medically checked by a veterinarian to rule out any underlying medical problem that might be contributing to the undesirable behaviour. •
  • Seeking further advice from a veterinary behavioural specialist or veterinary behaviourist if required so that a diagnosis can be established, and a behavioural modification plan can be tailored to the dog’s needs.