Topic For Discussion - Keep It Nice, Folks ;)

[huski]

I don't want my dogs to have a high value for others. I prefer my dogs to see other dogs and think "meh". If I had my dogs out and they barked at other dogs we passed I would consider it a sign their arousal is too high considering I've asked them to chill out and walk nicely with me.

[LisaCC]

As a renter I don't want my dogs barking AT ALL in the house or backyard. And none of my dogs have ever barked while in the yard except if there is due reason. Due reason does not include communicating with the neighbouring dogs. If they did I would take steps to put a stop to it.

[Mrs Rusty Bucket]

MRB not sure why you've had more success with a flat collar than slip lead, which I'm presuming is a choker?

I wasn't either - worked just fine with other dogs I'd trained. But I had to learn new ways to train with this dog because nothing more aversive than a no-reward marker or no reward, works.

It's naive to believe there is no stress involved in "positive only" training, or that stress is a bad thing. Stress is present in all learning and dogs need to learn how to work through stress.

Huski, I agree with most of your post except the bit about which does more damage. Post Traumatic Stress Disorder doesn't come from too many rewards. I admit that Obesity related illness can. But I know which one I'd rather have. Bad unpleasant experiences even just once tend to be ETCHED forever in a way that good ones are not. But repeated good experiences create much value and joy in an activity.

And my dog does get stressed and frustrated when she doesn't get the treat. But she also gets very excited by the game of figuring what she has to do to get the treat. If I give the treats for crappy performance, or just get it wrong, I lose her engagement, it's too easy and she doesn't learn anything or improve. The Bob Bailey rules on this is if we can get it right 80% of the time or 4 out of 5 attempts - we will make progress. Part of success in our style of training involves teaching the dog to do something then testing the dog in their understanding... so the dog failing to get the reward is part of the training. Eg a stay with no distractions is pretty easy... but lets try it again with a small distraction... I can't understand the old school obedience insisting no distractions to the point of if a magpie lands near the dogs in a competition stay - they get a do-over.

So I agree that if one wants to have success at least two quadrants must be used. Eg treat or no treat (and or NRM) covers two quadrants. And I do like to stay away from the technical language because most people think +P is a good thing. It's positive right?

Bristol uni animal behaviour dept, specificaly in the 'dog training' area and senior lecturer Dr Rachel Casey was trashed in terms of its credibilty years ago.

Same with the alpha wolf study and the Ceasar Milan stuff.

I don't know Bristol Uni.

I do know Bob Bailey, Susan Garrett, Paul McGreevy, Ian Dunbar, Steve White...

Most of them are very good record keepers ie they keep records of their training plans, and their results. And they are all passing on their knowledge to the benefit of people who take them up. And people are taking up these methods because they work - it's most obvious on the agility field (the fastest dogs are trained this way) but becoming more obvious in other areas of training and competition.

I don't think the "purely positive" that the article writer is talking about is what these trainers and their students are using tho. I don't like seeing people who never ever train their dog (or children) to have good manners either. I don't want to be around them or their dogs or children. Reward based training is not about letting your dog be naughty and do whatever it wants. Control the environment, limit the choices, reward the good choices... If a bad choice is self rewarding - it has to be interuppted and stopped. But you don't need to yell at a dog or yank on the lead to do that.

•small numbers of respondents (plus 70% women - how can they assume that this is a representative survey?)...

70% women: that's probably a good representation of people obsessed with dog training. Have you been to any dog competition apart from IPO recently?

I find flat collars give them something to lean into and they can pull to their hearts content.

True but if you reward the dog (praise, attention, pats, treats) for keeping the lead loose and prevent the dog from pulling and self reinforcing... they stop pulling really quick.

Without that basic training, my dog would either lean into the chain, choke herself, cause herself pain and bruising - that she didn't feel until later... or she'd lie completely flat on the ground and not move in any direction. One old school instructor suggested dragging her - and then she'd get up he said. She did not. I can't believe I did some of that stuff to her because some old school trainer said it was a good idea. Sheesh. She is way smarter than me at dog training.

From the article.

Correction is about the dog, punishment is about you. Punishment is because YOU are pissed. YOU are frustrated.

This is a very classic misunderstanding of what Punishment means in the science jargon. All it means is to reduce the (undesirable) behaviour. If you can reduce the undesirable behaviour by not-rewarding it or training an incompatible behaviour, you've "Punished" the (undesirable) behaviour.

Collar pops, taps etc are all intended to reduce an undesirable behaviour - so if they actually work (and often they do not) - then they are +P punishment.

[Willem]

•small numbers of respondents (plus 70% women - how can they assume that this is a representative survey?)...

70% women: that's probably a good representation of people obsessed with dog training. Have you been to any dog competition apart from IPO recently?

no,...in our obedience classes the ratio is more 50%/50%, in the agility courses there might be slightly more female dog owners ....but yes, if they would really compare apples with apples in the study and would have an objective benchmark for the trainings success the gender of the trainer / dog owner shouldn't make a difference.

[huski]

It's not about what you think is capable of doing more damage, the damage is not measurable because the consequences for the dog can be severe either way. getting given lots of treats doesn't mean the dogs needs are actually being met. I see severe behaviour cases all the time where the dogs have never been given a physical correction, yet they still have severe behavioral problems. It's incorrect and naive to assume that the only real damage that can be done to a dog is caused by misuse of physical corrections. There are many factors that can contribute to a dog becoming stressed, anxious, fearful etc.

[Podengo]

http://denisefenzi.com/2015/12/28/have-to-vs-want-to/

[Simply Grand]

I'm wondering how much of this "argument" about positive only training vs other methods comes down to confusion about terminology...

Words like positive, punishment, correction clearly mean different things in different contexts and to different people.

And there is also apparently a lack of understanding about different methodologies. The article in Perse's OP talks about the use of luring and "cookie" and how using them means dogs won't work when they can't see the cookies, but the showing the food, luring and rewarding every time is only the beginning of the process of using positive reinforcement to train. If you don't get past the point of needing to have the food or toy reward visible to the dog, or even having it at all, in order for the dog to perform you haven't finished the training process, or haven't done it correctly.

I also wonder how many people actually do use 100% positive reinforcement only, even they believe or say they do (and I don't think there are that many around that even claim that). Never saying the word "no" doesn't mean that someone never uses negative reinforcement or punishment. For example the article refers to moving away when a dog jumps on you as part of positive only training, but that is actually (according operant condition) positive punishment - dog jumps up, something unpleasant happens ie human moves away - so I think there is some confusion going on.

[Willem]

I'm wondering how much of this "argument" about positive only training vs other methods comes down to confusion about terminology...

Words like positive, punishment, correction clearly mean different things in different contexts and to different people.

And there is also apparently a lack of understanding about different methodologies. The article in Perse's OP talks about the use of luring and "cookie" and how using them means dogs won't work when they can't see the cookies, but the showing the food, luring and rewarding every time is only the beginning of the process of using positive reinforcement to train. If you don't get past the point of needing to have the food or toy reward visible to the dog, or even having it at all, in order for the dog to perform you haven't finished the training process, or haven't done it correctly.

I also wonder how many people actually do use 100% positive reinforcement only, even they believe or say they do (and I don't think there are that many around that even claim that). Never saying the word "no" doesn't mean that someone never uses negative reinforcement or punishment. For example the article refers to moving away when a dog jumps on you as part of positive only training, but that is actually (according operant condition) positive punishment - dog jumps up, something unpleasant happens ie human moves away - so I think there is some confusion going on.

...giving no treats or not rewarding the dog is already a form of positive punishment ...'positive reinforcement' works only in combination with 'refusing positive reinforcement' as the negative consequence after an undesired behavior. Once the dog is used to positive reinforcement every refusal of positive reinforcement is obviously felt as punishment respectively unpleasant scenario.

[corvus]

I know the work of the Bristol group pretty well, and have met some of the scientists there. They are heavyweights in the canine science world for good reason. I am not aware of any of their work that has been discredited, except for perhaps the dominance study, which hasn't been discredited so much as generated some disagreement and robust discussion.

The general aim with positive training is to avoid needing to correct the dog. If you control the environment and step the dog up with their behaviour so they know what to do, there's not much to correct. Plenty of incorrect responses you can easily afford to ignore and they will just go away because they are not being reinforced. There are levels of precision and reliability that are challenging to create with rewards alone, unless you are Bob Bailey (who rewarded his open ocean working dolphins with sex toys on their return, as it happens - they were able to hunt and eat while working). And most people don't need that. They just need their dog to come when called and behave nicely in public. That is easy to get with rewards alone for the vast majority of dogs.

There's no way I would consider positive reinforcement remotely as risky as punishment or negative reinforcement. Aversive control is far more likely to be associated with aggression than rewards are, and there is science to support that. I am extremely cautious with punishment and pretty blasé with positive reinforcement. Because that's the nature of aversive and appetitive stimuli. They are not just as potentially bad as each other. Not by a long shot. If I could convince all of my clients of one thing only, it would be to stop correcting their dogs. Never needed to instruct anyone on how to punish, because it's not yet been necessary and often makes things worse.

[corvus]

...giving no treats or not rewarding the dog is already a form of positive punishment ...'positive reinforcement' works only in combination with 'refusing positive reinforcement' as the negative consequence after an undesired behavior. Once the dog is used to positive reinforcement every refusal of positive reinforcement is obviously felt as punishment respectively unpleasant scenario.

Not really. For that to be the case, you would see a reduction across the board of that behaviour that failed to be rewarded. This usually results in general behavioural suppression, which is a pain in the arse for training. Rather, the behaviours that are not rewarded are abandoned through extinction.

[The Spotted Devil]

Willem - not quite. Positive punishment ADDS something to DECREASE the likelihood of a behaviour occurring again. If anything it's -P BUT "punishment" doesn't describe the dog's emotion but the outcome.

As a scientist I think less about the quadrants and more about the dog's response. I want to foster joy and confidence. I want a dog that offers behaviours and sees failure as just another opportunity to earn a reward. Do I always get it right? Nope. Do my dogs work with enthusiasm and do stuff that I've been told is not possible? Absolutely :D

Edited December 31, 2015 by The Spotted Devil

[Denis Carthy]

Mr Rusty Bucket

Same with the alpha wolf study and the Ceasar Milan stuff.

I don't know Bristol Uni.

Ahh, it was to do with CMs' stuff starting to influence pet dog owners here in UK, they had enough of ineffective, expensive 'reward training', (which in fact is an attempt to use negative punishment as a base) and when CM, came on TV here he started to influence them as the way to go.....the same pet dog owners were potential clients or ex clients of reward trainers so Casey & the rest of them started to slag CM off to try to discredit him and put their own client base at risk........lol, instead they ended up with those vids filling about 2 3rds of page 1 of a google search for her name + the uni lawyer trying to stop them.......so thats how come those vids emerged......

.

Edited December 31, 2015 by Denis Carthy

[Simply Grand]

Mr Rusty Bucket

Same with the alpha wolf study and the Ceasar Milan stuff.

I don't know Bristol Uni.

Ahh, it was to do with CMs' stuff starting to influence pet dog owners here in UK, they had enough of ineffective, expensive 'reward training', (which in fact is an attempt to use negative punishment as a base) and when CM, came on TV here he started to influence them as the way to go.....the same pet dog owners were potential clients or ex clients of reward trainers so Casey & the rest of them started to slag CM off to try to discredit him and put their own client base at risk........lol, instead they ended up with those vids filling about 2 3rds of page 1 of a google search for her name + the uni lawyer trying to stop them.......so thats how come those vids emerged......

.

I think Cesar Millan's 'The Dog Whisperer' was one of the earliest dog training tv shows. He has been around a long time. And if you watch The Dog Whisperer throughout the seasons from 1 to 9 you can see how even Cesar has changed his methods to incorporate more reward based training.

I'm pretty sure even Cesar would not claim his techniques were a response to people being unhappy with "expensive, ineffective reward training" ?!?

[Willem]

Willem - not quite. Positive punishment ADDS something to DECREASE the likelihood of a behaviour occurring again. If anything it's -P BUT "punishment" doesn't describe the dog's emotion but the outcome.

hm, IMO it depends on what becomes the 'normal' scenario. A dog used to treats (or a pat) if he makes the right choice get used to it...I recall my dog and she comes as she knows she gets the treat, and sometimes she even comes without having here called, sits in perfect heel position and her eyes asking me, 'hey where is my treat, that was a perfect recall...come on..'...so this behaviour becomes the 'normal scenario' after intensive training. If she does a lousy job I can 'add' the scenario where I refuse to give here the reward...IMO a positive punishment once the good behaviour pattern with the reward is established.

As a scientist I think less about the quadrants and more about the dog's response. I want to foster joy and confidence. I want a dog that offers behaviours and sees failure as just another opportunity to earn a reward. Do I always get it right? Nope. Do my dogs work with enthusiasm and do stuff that I've been told is not possible? Absolutely :D

I think one of the problems is that we humans tend to believe that a new approach has to be 'better' read more successful, more effective....more convenient etc. While this might be justified wrt technical gadgets, IMO it is not always appropriate regarding dog training. Most of us don't own a dog because they have to, but because of the fun factor, hence the chosen trainings approach should reflect the fun factor for dog and owner too. Seems to be easy when getting the dog as a puppy without having to deal with the load of a 'bad history', however, for adopted problem dogs the question about the best trainings approach can be pretty tricky I guess, and I believe it is in the best interest of the dog and the owner of such a problem dog to consider all available options including the not popular once.

EDTA the following question for the experts: you are speeding and the police officer gives you a ticket...positive or negative punishment? :D

Edited December 31, 2015 by Willem

[Simply Grand]

Giving the ticket is a consequence. The fine involved is intended to be a positive punishment, as in it is added to the situation in an attempt to reduce the behaviour. Whether it is an effective punishment depends on whether the behaviour reduces or not.

Were you at the point where your licence was taken away as a result of the ticket that would be intended to be a negative punishment, in that a privilege was taken away from the situation in an attempt to reduce the behaviour.

I'm not sure what your point is.

[Weasels]

.

Edited December 31, 2015 by Weasels

[Willem]

Giving the ticket is a consequence. The fine involved is intended to be a positive punishment, as in it is added to the situation in an attempt to reduce the behaviour. Whether it is an effective punishment depends on whether the behaviour reduces or not.

Were you at the point where your licence was taken away as a result of the ticket that would be intended to be a negative punishment, in that a privilege was taken away from the situation in an attempt to reduce the behaviour.

I'm not sure what your point is.

at the end the money (my reward for work) I won't have for me due to the fee is what hurts me...sometimes more than 50% of my dog's food ration are treats as reward for good behaviour, so giving her a 'ticket' for bad behaviour means less treats (no, I don't let her starve if she doesn't earn enough treats :D )...hence IMO a positive punishment...

[JulesP]

I think Cesar Millan's 'The Dog Whisperer' was one of the earliest dog training tv shows. He has been around a long time. And if you watch The Dog Whisperer throughout the seasons from 1 to 9 you can see how even Cesar has changed his methods to incorporate more reward based training.

I'm pretty sure even Cesar would not claim his techniques were a response to people being unhappy with "expensive, ineffective reward training" ?!?

Barbara Woodhouse had a dog training show in the 1980s :) Walkies.... :laugh:

Victoria Stillwell had a show around the same time as CM too.

[The Spotted Devil]

Willem - not quite. Positive punishment ADDS something to DECREASE the likelihood of a behaviour occurring again. If anything it's -P BUT "punishment" doesn't describe the dog's emotion but the outcome.

hm, IMO it depends on what becomes the 'normal' scenario. A dog used to treats (or a pat) if he makes the right choice get used to it...I recall my dog and she comes as she knows she gets the treat, and sometimes she even comes without having here called, sits in perfect heel position and her eyes asking me, 'hey where is my treat, that was a perfect recall...come on..'...so this behaviour becomes the 'normal scenario' after intensive training. If she does a lousy job I can 'add' the scenario where I refuse to give here the reward...IMO a positive punishment once the good behaviour pattern with the reward is established.

As a scientist I think less about the quadrants and more about the dog's response. I want to foster joy and confidence. I want a dog that offers behaviours and sees failure as just another opportunity to earn a reward. Do I always get it right? Nope. Do my dogs work with enthusiasm and do stuff that I've been told is not possible? Absolutely :D

I think one of the problems is that we humans tend to believe that a new approach has to be 'better' read more successful, more effective....more convenient etc. While this might be justified wrt technical gadgets, IMO it is not always appropriate regarding dog training. Most of us don't own a dog because they have to, but because of the fun factor, hence the chosen trainings approach should reflect the fun factor for dog and owner too. Seems to be easy when getting the dog as a puppy without having to deal with the load of a 'bad history', however, for adopted problem dogs the question about the best trainings approach can be pretty tricky I guess, and I believe it is in the best interest of the dog and the owner of such a problem dog to consider all available options including the not popular once.

EDTA the following question for the experts: you are speeding and the police officer gives you a ticket...positive or negative punishment? :D

I can't be bothered quoting in parts but:

1. Learning theory is not so simple as just quadrants. It's just one part of the puzzle. Timing, criteria, choice of rewards, distractions, environment, failure, focus...If you don't reinforce a behaviour it will fade. Although if you reward intermittently it's less likely to fade - I prefer to do the latter through criteria, only rewarding 50% or better but ensuring success 7-8 times out of 10 when training.

2. I have trained a lot of dogs. I use the same approach with puppies as I do with adolescents, adults, rescues, puppy farm, behavioural problems...of course I tweak what is emphasised depending on the individual. In actual fact I can take most dogs and have them working with enthusiasm pretty quickly. The skill comes in teaching the owners how to do this. My clients have FUN training their dogs. They get results. The vast majority continue on with the training because they enjoy it and the dogs enjoy it. I teach them about timing and criteria - so even the kids can do it. I actually teach them to be "hard arses" as I like to put it - minimal luring (I don't use it at all for my own dogs), realistic expectations, split behaviours, consistency and timing. Don't reward rubbish but pay big time for trying and improvement.

3. +P (and you can have your opinion about that particular scenario but you are still wrong - you keep talking about what the dog expects or feels. It's not about that. It's about the outcome plain and simple. I know someone who uses e-collars and has a lot of experience and success with them. Twice recently 2 of her dogs have ignored it on its highest setting and gone hunting. It didn't change the outcome (as far as we know) so it doesn't fall into a quadrant).

The vast majority of people cannot recognise emotions in dogs by the way. We still have discussions about dogs experiencing guilt when it's fear or appeasement. It's something I teach my clients so they can see the relationship between their dogs having an awesome time AND the owner getting the response they want. I recall a poster of CM advertising his stage show. The body language of the 8 dogs pictured with him made my skin crawl. People think it looks amazing though.

Funny, this whole "real world" stuff. Firsf time I took Em on a game shoot she blew them away. Perfect combination of ridiculously high drive, focus and impulse control.

Edited December 31, 2015 by The Spotted Devil

[Willem]

...giving no treats or not rewarding the dog is already a form of positive punishment ...'positive reinforcement' works only in combination with 'refusing positive reinforcement' as the negative consequence after an undesired behavior. Once the dog is used to positive reinforcement every refusal of positive reinforcement is obviously felt as punishment respectively unpleasant scenario.

Not really. For that to be the case, you would see a reduction across the board of that behaviour that failed to be rewarded. This usually results in general behavioural suppression, which is a pain in the arse for training. Rather, the behaviours that are not rewarded are abandoned through extinction.

...I made a second attempt to digest this, but na, I don't get it...???...