Dear Arsalan,

as you are considering the possible consequences used in dog training (positive / negative reinforcement / punishment), I would like to suggest Michael Domjan (The principles of leaning and behaviour) and/or Stepehn R. Lindsay (Handbook of applied dog training and behaviour and training) as very useful ressources.

Both authors go into learning theories and the possible consequences administered to dogs in instrumental conditioning in detail, often quoting from original sources in the literature.

I hope this helps!

Stefanie

Dear Arsalan,

I suggest you read also Karen Pryor's books on training methods. There is a short book called Zoomility: Keeper Tales of Training With Positive Reinforcement (I forgot the authors' names...) which is quite interesting.

Sincerely, Fabienne

Positive

Hi Arsalan,

I don't train animals but I spent a lot of hours studying behavior in other model that trainers uses positive reinforcement… I think that negative reinforcement and punishment will just increase frustration in the animals. 

I recommend you to read Karen Pryor's books as Fabienne Delfour suggested… She is an expert in the field… she trained a lot of models, including dolphins, cats and dogs. She only uses positive reinforcement and a clicker.  

HI, I am a dog trainer and studying dog cognition at Porstmouth Uni and will say, for me, the best way to train dogs is primarily positive reinforcement and occasional negative punishment. These techniques reduce stress in learning which helps them to learn more quickly. Positive reinforcement means something good happens to encourage the behaviour to be repeated. Negative punishment means something good does NOT happen, and the prospect of this is removed to deter a behaviour.

An example of  negative punishment is in leash pulling. I would stand firm if a dog is pulling on the leash so that the dog cannot be reinforced for it. This means that the dog does not gain the anticipated reward and is negatively punished. However, if I was using a clicker and treats, the dog would be positively reinforced for walking nicely. But if the dog pulled he would not get the click and treat, nor should he get the other reward he is after (sniff) and is technically punished (neg punishment).

For negative reinforcement to be applied so must positive punishment as it works by stopping the unpleasant stimulus thereby operating on the emotional state of relief. This leads to stress and reduces learning in my experience. See Burch and Bailey How Dogs Learn for a nice scientific introduction.

I hope this helps?

Kind regards

Denise

In my experience positive reinforcements are the best. Negative reinforcements always elicit stress and fears which might be hided but appears unexpectedly harming the wanted behavior in dogs

Hi.

I'm a PhD student studying the effects of delayed positive reinforcement in dogs ... If you would like a very quick and simple explanation of some learning principles as they are applied to dog training, Sophia Yin has a poster explaining operant conditioning: http://drsophiayin.com/blog/entry/which-category-of-operant-conditioning-is-it

Otherwise some good books (as recommended already) will provide a more comprehensive understanding.

The risk of using aversives incorrectly (inappropriately, at incorrect strengths, poor timing, etc.) is always stress, fear, and learned helplessness (worst case scenario). These things have welfare implications and thus should inform training decisions. Those books will explain things in more detail if you're interested.

Regards,

Clare.

Operant learning occurs when there is an outcome the animal would prefer, and an outcome the do not prefer.  So some form of punishment is always present in that scenario.  However the punishment need not be intensely aversive and this is obviously preferable. And a great deal of learning can be through shaping and minimize the role of punishment.

What is your goal is asking this question?  If you provide more context the answers could be more specific.

The most effective punishment is removal of what is desired.  For example, most dogs jump on people because they want attention.  Train the dog not to jump by removing what is desired, which is attention.  When the dog jumps on a person, that person must move away, turn their back to the dog, and ignore the dog.  This type of training needs consistency and many repetitions.  Every person the dog interacts with must respond to being jumped on the same way, every time.   

The most effective reward is giving the dog what that particular dog wants.  In the example of the jumping dog, the dogs wants attention.  To reward the dog, give attention when the dog is not jumping.  Give the dog attention at different times, when the dog is laying down, sitting, standing, and walking.

During this training, it does not help to speak to the dog at all.  The dog learns the association best with simple actions, not verbal rebuke.

This is the same thing as Denise Nuttall said, only with different words.