On The Effective Use Of Punishment

[corvus]

Steve White lists 8 rules for using punishment in dog training: http://rewardingbehaviors.com/forum/viewto...+8+rules#p14815

This has stimulated an interesting but at times annoyingly childish discussion on another forum. Thought it would be interesting to see which rules DOLers agreed with and why, but hopefully we can skip the childish mud-slinging.

Here are the 8 rules:

Eight Rules for Using Punishment:

1. The punishment must be something the animal dislikes and something the animal does not expect.

2. The punishment must suppress behavior. (This is, in fact, the very definition of something that is a punisher.) If something is being used for punishment, but it does not suppress behavior, it’s ineffective and often just plain abuse.

3. The punishment must be of the perfect intensity. Too much and there will be negative fallout. You’ll end up hurting your relationship with the animal and loosing more than just that behavior. Too little and the punishment will only serve to desensitize the animal and build resistance.

4. The punishment must happen immediately after the behavior it is to be associated with. Otherwise, a clear enough association between the wrong behavior and the punishment will not be made.

5. The punishment must be associated with the behavior, but not with the trainer. Otherwise, the trainer becomes part of the punishment and the animal starts fearing and disliking the trainer.

6. The punishment must happen every time the behavior occurs. If punishment does not happen every time the behavior occurs, the behavior gets put on a variable schedule of reinforcement. Depending on the behavior and how often the punishment actually occurs, the animal could decide that performing the behavior was worth the risk of getting punished.

7. There must be an alternative for the animal.

8. Punishment must never be used to the extent that punishment outweighs positive reinforcement (from the animal’s perspective, not yours!)

[Tilly]

The rules make perfect sense to me ...

[persephone]

Sounds a very logical and sensible set of rules to me .

[Cosmolo]

Don't disagree with any although you could have an interesting discussion to clarify exactly what point 5 means and how it relates to different dogs within training.

[Bullbreedlover]

Sorry but those rules dont make sense to me at all.

But I guess I am old school.

I dont think of it as punishment nor do I call it punishment.

I think of it as enforcing an action on an animal in a humane way.

I hope the author doesnt use this punishment on their children.

[bedazzledx2]

What makes the most sense to me is the omitted paragraph at the bottom of the quote

"If you can’t follow all 8 of these rules, you’re probably better off avoiding the use of punishment. Heck, even if you can follow all 8 rules, you’re probably better off avoiding the use of punishment, as punishment can result in so many unintended and undesirable side effects."

Sorry but those rules dont make sense to me at all.

But I guess I am old school.

I dont think of it as punishment nor do I call it punishment.

I think of it as enforcing an action on an animal in a humane way.

I hope the author doesnt use this punishment on their children.

[Okami]

Those rules make sense to me

I re-read through them twice... and can't disagree on any point

[angelsun]

I'm old school obedience as wel and don't have any issues with this list overall. I will state however that there is a time to not use punishment when correcting and that depending on the dog, distraction techniques work much better.

I'm not a fan of no punishment for anything however as some training regimes follow. And yes I do shape/clicker but not with all dogs and only for certain aspects after I lay down the foundations.

[WildatHeart]

I agree with that set of rules. But i also agree with angelsun that 'there is a time to not use punishment when correcting and that depending on the dog, distraction techniques work much better".

[corvus]

I dont think of it as punishment nor do I call it punishment.

I think of it as enforcing an action on an animal in a humane way.

Maybe you don't think of it as a punishment, but does your dog? If your actions result in a decrease in the behaviour following them over time, then it is a punishment whether you call it one or not. If your dog works to avoid these "enforcing actions", then they are punishments. If not, then they are not punishments. You can reason with a kid and explain to them why they mustn't do something, but you can't reason with a dog.

For the record, I agree with the rules as well. Rule 3 is basically the reason why I have moved away from using punishments in training.

[gundoglover]

Yes, I agree with all of these rules. I have just started back in obedience after a break of 20+ years and I am astounded at what is being taught at the obedience school. I train mostly at home, but need the distraction of other dogs to test the training. My dog is an entire male in the throws of puberty and if I used only reward-based training as advocated at the school we would be proceeding much more slowly, if at all. The only negative reinforcement that I use is a quick tighten-release of the check chain and occasionally placing the dog in the position that I want. I use liberal positive reinforcement of food, voice and pats.

I have noticed that not only is my dog proceeding more rapidly than the rest of the class, he is also one of the few happy workers with the fastest recall. Now, I have no doubt that it is possible to train a dog with positive reinforcement only, but surely this can be confusing for the dog while he tries to read the handler's mind and offer the required behaviour. I do understand that novice handlers can do a lot of damage to a dog by incorrectly using check chains, etc., and that positive reinforcement should vastly outweigh the negative, but it seems that we have gone to an extreme where some novice handlers with large badly controlled dogs are not being shown how to correctly use negative reinforcement when appropriate.

[poodlefan]

They sound reasonable to me but I'd add one other.

Punishment has a very limited place in training. I see too many handlers use 'punishment' for non compliance when they haven't clearly indicated what they want or the dog doesn't know what is expected.

[Cosmolo]

But what kind of punishment are we talking about PF. I disagree, i don't think punishment has a limited place in training more than any other technique- suitable for some, not for others- i just think it needs to be taught properly!

Edited January 3, 2010 by Cosmolo

[poodlefan]

But what kind of punishment are we talking about PF. I disagree, i don't think punishment has a limited place in training more than any other technique- suitable for some, not for others- i just think it needs to be taught properly!

Aversives for failing to produce the result the trainer wants. Eg. hitting a pup for not sitting or jerking and dragging a dog into a down.

[Tilly]

Punishment can range from a verbal correction, ignoring the dog for bad behaviour or a physical correction on the collar.

I am all for positive rewards for training young pups but at some point you are going to need to correct the dog ... and this correction is a form of punishment.

[bedazzledx2]

Or putting a choke chain on a dog in order to teach it to heel

But what kind of punishment are we talking about PF. I disagree, i don't think punishment has a limited place in training more than any other technique- suitable for some, not for others- i just think it needs to be taught properly!

Aversives for failing to produce the result the trainer wants. Eg. hitting a pup for not sitting or jerking and dragging a dog into a down.

[huski]

I think it's also important to note that what constitutes a correction for one dog may not constitute a correction for another.

I know a dog who cannot handle verbal correction, just a slight 'ah' will make him slink back with his ears down. I can use the same 'ah', with the same tone, with Daisy as a NRM as a way of telling her nope, you got it wrong, try again, obviously with totally different results.

I think it's impossible to completely train a dog with only positive reinforcement, as at some point, you will do something to the dog that they find unpleasant - for example, Daisy finds the removal of a reward as unpleasant, and it's obvioulsy not PR.

Edited January 3, 2010 by huski

[PAX]

6. The punishment must happen every time the behavior occurs. If punishment does not happen every time the behavior occurs, the behavior gets put on a variable schedule of reinforcement. Depending on the behavior and how often the punishment actually occurs, the animal could decide that performing the behavior was worth the risk of getting punished.

Point six is weird, if punishment has to happen every time, then clearly the punishment hasn't worked. I would be looking for a better training system, dogs never chose to be in trouble.

[Cosmolo]

Thats true Huski- and thats why its so important for techniques to be tailored to a dog. I saw two dogs recently and one of them was so sensitive i had to be very careful using a lure because it was almost too much for her. We took tiny little steps and seeing her progress has made me so proud! I would have hated to have seen this particular dog in the hands of someone who did not realise just how sensitive she really was!

ETA- Pax, dogs won't learn to the point of reliability with one repetition though- regardless of the training technique so consistency with any training is key.

Edited January 3, 2010 by Cosmolo

[PAX]

I can use the same 'ah', with the same tone, with Daisy as a NRM as a way of telling her nope, you got it wrong, try again, obviously with totally different results.

I think it's impossible to completely train a dog with only positive reinforcement, as at some point, you will do something to the dog that they find unpleasant - for example, Daisy finds the removal of a reward as unpleasant, and it's obvioulsy not PR.

I agree, but if Daisy finds the removal of the reward unpleasant, why would you bother with a harsh verbal reprimand as a NRM, why not just let her figure it out if she is truly is working for the reward?