Spin-off: what makes a positive trainer a positive trainer?

That is an excellent way to put it. I never thought about it that way. I think part of owning and properly training a dog is mutual respect and understanding. As fouriscompany said they need know what you want, what you will put up with, what you need and what you expect from them, an what you are willing to give in return.

I think of it like a bank account. Every time you are responsible in the eyes of your animal for something good, you make a deposit in the bank. The size of the deposit is directly proportional to the animal's delight in whatever you've done for them. When you do something to them that they don't like, you make a withdrawal. The size of the withdrawal is directly proportional to the level of dislike the animal feels for whatever you did to it. Thus, over time you develop a history which affects the degree of trust your animal has in you. As a positive trainer, my goal is to have my balance as high as I can while still having an animal that is safe, healthy, and that I can live with. Naturally, I want to keep my withdrawals at a minimum. I like my balance to be high because it makes less of a dent when I have to make a withdrawal. Ultimately, I invest in my animal's trust in me with positive experiences, while negative experiences set my investments back. I'm fine making withdrawals as long as my bank balance is high, and as long as I maintain my bank balance at a high level. Technically, I could make more withdrawals to get a dog that might respond quicker and I'd be safe as long as the balance stayed above the black, but as a positive trainer, I resist making withdrawals unless there's no other way to do whatever I have to do with my animal. That's just how I see it. Trust is a bank account I build up with positive actions and diminish with aversives, punishment, or corrections. That's not to say I never make withdrawals, just that my aim is to deposit as much as I can and withdraw as little as I can. Nor is it to say all animals respond the same way. What is a withdrawal to some may be a deposit to others. It all comes down to what they like and what stresses them. Does that make any sense?

Oh, and in the case of an animal you've just started with, of course your balance starts at zero, so you have to build it up a bit before you can afford to make a withdrawal.

What counts as an aversive depends on how the individual dog sees it, not the human. That can kind of cut both ways as a lot of people claim that what they are doing to their dogs is not aversive yet I can clearly see in the dog's body language that the dog disagrees. And in other cases, people tippy-toe around their dogs while their dogs are miles from feeling what is being done is anywhere near being aversive. Having two different dogs of two very different personalities, I definitely have to vary what I do in my attempt to keep from using aversives as much as possible. Conrad is softer than soft, submissiver than submissive, and is a naturally fearful and anxious personality and he'll stop drop and roll just at a raised voice or even a dirty look. I've literally made this dog cry (tears, sniffling, the works) by yelling at him in the past.  So I don't ever raise my voice with him in my effort to keep it minimally aversive. Marlowe on the other hand is unflappable and serious as a heart attack. I take a different tack with him and can push the line a bit further before we get into the realm of aversive. My approach with Conrad is all about building up his self-confidence and less about typical obedience stuff, but Marlowe has more than enough confidence already (it's what makes him so far such an outstanding agility dog), so we work on focus and drive. I know when I've gone over the line with Conrad when he starts to throw calming signals at me. With Marlowe, he just shuts down and goes away--he doesn't care about me being a threat or not a threat, he just checks out mentally and if possible calmly leaves the scene physically.

 What makes me LIMA (least invasive, minimally aversive) with Conrad is different for what makes me LIMA with Marlowe.

Incidentally, I've found this method of training to be highly successful with my very stubborn working coonhound rescue. It tricks him into thinking the things that I want him to do are actually his idea, and because he's a hound if it's his idea he thinks it's the best idea. Shaping is so fantastic with him because he actually thinks that touching my palm with his nose or laying down on his mat are purely his discoveries and he's got to be the luckiest, smartest, most intrepid coonhound on the face of the planet for discovering such treasure troves of goodness. There's a story in Leslie McDevitt's book about what happened when she was giving a demo with one of her dogs and mat training--she threw the mat up in the air and before the mat hit the ground her dog had jumped up and put all four paws on it while it was still flying through the air. Marlowe actually did that, though less dramatically, over the weekend. I was shaking the mat out in preparation for putting it down, but he already had two paws on it before I'd even had a chance to put it on the ground.
 

It's been thoroughly proven that "punishment", i.e. a correction applied after the "misbehavior" occurs, is the least effective method of modifiying behavior. Why haven't dog owners/trainers grasped this concept?  training with corrections, even if they are only 50% or 10%, is not the most effective thing you can do. Why use them at all? use something effective instead. Positive reinforcement (praise, treat, toy); negative reinforcement (self-correct on prong, ecollar, withholding attention until the dog complies) are the most effective methods of modifying behavior. Sorry, I just don't believe people who go around saying you can't get a reliable dog without delivering a few leash pops, or without being "balanced". You can. People who default to corrections are usually just being impatient or lacking in skill.

mudpuppy

It's been thoroughly proven that "punishment", i.e. a correction applied after the "misbehavior" occurs, is the least effective method of modifiying behavior.

NO it hasn't.

of course, YOU may have convinced yourself this by the mental agreements you make.

 

mudpuppy

It's been thoroughly proven that "punishment", i.e. a correction applied after the "misbehavior" occurs, is the least effective method of modifiying behavior. Why haven't dog owners/trainers grasped this concept?  training with corrections, even if they are only 50% or 10%, is not the most effective thing you can do....Sorry, I just don't believe people who go around saying you can't get a reliable dog without delivering a few leash pops, or without being "balanced". You can.

This goes back to what I have been saying. For some dogs it IS the most effective way to teach them. For others, it is the least effective. I really depends on the dog. No one learns the same. My Beagles respond best to corrections. It has noting to do with my "lack of skill", as I consider myself a very competent trainer and a have had numerous people compliment me on my training abilities and my dogs. They wouldn't hire me to train their dogs if they didn't.
Your dog might do better with positive only, but mine take advantage of it. And by the way, my Brittany and my Shih Tzu learn best through positive reinforcement and not to leash corrections. They are clicker trainer and do very well with it. I wouldn't say that they learned any better or fast and they are not nearly as consistent as the beagles, but that was their learning style, so I worked with them. I have had several client dogs that want to learn through the clicker and do learn best that way. I work with them.

Years ago, on this board I believe, there was a women who was very insistent that clicker was the only way to go and everything could be trained with one. Anyone who used any other method was a horrible trainer, knew nothing about behavior or training, and obviously didn't get it. We got into several discussions about it. One day, she wrote in that her dog was chasing and taking down the cats. She had tried everything to stop him and was looking for suggestions. Out of fear for her cats and their safety, she tried a shock collar and it stopped that dog from killing the cats. Last I heard he was getting along very well with them. This is a prime example of a dog that could not be taught to avoid his instinct with a clicker. He took advantage of the fact that "mom" wouldn't punish him and the thrill of chasing the cats was more exciting and rewarding than making her happy. He ended up learning best with corrections.

"

NO it hasn't.

of course, YOU may have convinced yourself this by the mental agreements you make."

 

nothing to do with me. Scientists and psychologists and behaviorists have definitely proven that punishment is the least effective way to modify behavior in every species they've ever studied. If you don't believe science, think of it logically.

Punishment by definition STOPS a behavior. If you want your dog to perform a behavior, lets say HEEL, using punishment to create the behavior means you have to PUNISH (STOP) hundreds of alternative behaviors that are NOT heel before the dog figures out what you want and gives you a reliable heel. Much more effective to simply REINFORCE the behavior you want- use a ecollar to negatively reinforce the behavior, or use treats to positively reinforce the behavior.

A side effect of using PUNISHMENT to attempt to get the dog to do something is that it makes harder in future to train any new behaviors. Let's say you're training HEEL, and some of the non-heel behaviors you punished include lagging behind, lunging sideways, and going forward ahead of the handler. So you've effectively PUNISHED going forward ahead of the handler into extinction. Now you want to teach the dog to GO AHEAD of the handler and pick up the dumbbell. Guess what- you might have trouble doing that since you taught the dog to expect a hard collar pop for going ahead. You've shot yourself in the foot. If instead you used reinforcement to reinforce the heel position, the dog has no prior learning about going ahead. Much easier to teach this new behavior.

You might say, well, how about I first reinforce the dog for heel, then punish any deviations- surely that is more effective than just reinforcement. Perhaps, but see above about how that can and will reduce your efficacy in teaching new behaviors in future.

Now if you want to stop your dog from doing a self-rewarding behavior like chase the cat, well, go right ahead and reach for some punishment-- one of the few contexts where punishment makes any sense at all. It's very possible to PRVENT cat-chasing using reinforcement, but punishment is probably your best bet if you've foolishly let the dog learn the joys of chasing the cat.

A side effect of using PUNISHMENT to attempt to get the dog to do something is that it makes harder in future to train any new behaviors. Let's say you're training HEEL, and some of the non-heel behaviors you punished include lagging behind, lunging sideways, and going forward ahead of the handler. So you've effectively PUNISHED going forward ahead of the handler into extinction. Now you want to teach the dog to GO AHEAD of the handler and pick up the dumbbell. Guess what- you might have trouble doing that since you taught the dog to expect a hard collar pop for going ahead. You've shot yourself in the foot. If instead you used reinforcement to reinforce the heel position, the dog has no prior learning about going ahead. Much easier to teach this new behavior.

I highly, highly disagree with this. My obedience dogs know when heeling, to not forge or lag. Their command for heeling is heel. When they must run foward, I use "take it" to run out to the dumbell and retrieve it, and "GO" which means go out until I tell you to stop and sit. And when I say sit, they turn around, look at me, and sit. And these are dogs who have been given little leash pops for moving out of heel position. Since I use an out command (go and take it), they know that when I ask for that, they are allowed to run in front of me. I had no problems teaching either of these commands, so I do not think that I "shot myself in the foot". In fact, both of the commands were easy to teach and picked up very quickly.
In addition, my show dogs are taught to move out in front of me when running AROUND the ring. "Go" is the command and the choke is placed underneath the chin. When moving on the down and back, where the judge can see front and rear movement, the choke is placed on top of the head and the command is "move". They know to change pace depending on the command and where the collar is. Beagles have a tendency to have poor front movement, so in order to hide it, I teach them to move with me and slower (basically in a heel position), vs when looking at side gait, the judge can not see the front, so they are taught to give out as much reach and drive as possible
It is so very, very possible to train this way, you must have to know how to and it sounds as if maybe you don't, which might be why you are so insistent that it can not be done.

when one burns their finger on a hot stove, they learn REAL FAST not to go and touch it voluntarily again... much faster than if a parent were to say to a child "if you don't touch that stove today, i will give you a piece of candy after dinner" and then a few days later, the kid wants to satisfy his curiosity, touches the hot stove, cries, and never does it again.

 

(the curious behavior to touch the stove has been MODIFIED)

likewise... a dog can learn to fear heights REAL fast by slipping off a 10 foot cliff and then going into a small panic while trying to get back to his pack. next time, he'll change his behavior and be weary and perhaps, not even want to jump out of the bed of a truck..... try teaching that sort of behavior REAL FAST LIKE using +R.

 

(the dogs behavior has been MODIFIED from wanting to peer over the edge to being fearful of falloffs)

FourIsCompany

What I do with dogs is more about relationship. They're like family. I don't "train" my husband. I let him know what I want, what I will put up with, what I need and what I expect from him, AND what I'm willing to give, and he does the same with me, and then we each choose what we give to the relationship. Depending on that, we either have a great relationship or not.

You NEED to read Bones Would Rain.... YOU JUST NEED TO.  I think you will LOVE LOVE LOVE IT. 

FourIsCompany

Depending on that, we either have a great relationship or not. We happen to have a great one. So, someone could say I have "trained" him, using "positive training" or they could say we have a great relationship.

You could say I have "positive trained" my husband.  I try to ignore behaviour I don't like.  I don't draw attention to it by nagging or becoming sulky or irate.  But I make a big effort to smile/kiss/hug/laugh/thank/whatever whenever he does something I DO like.  I also try to communicate clearly and make sure he has heard me and understood rather than getting pi**y if he appears to be ignorant, rude or stubborn.  He does know where I "draw my line" and what I am/am not willing to do.  I think we have a great relationship too!

I totally agree that "I train using mostly positive methods" is a world away from "I am a positive/positive only trainer".  The latter is, for some reason waaaay more loaded.

Chuffy

I totally agree that "I train using mostly positive methods" is a world away from "I am a positive/positive only trainer".  The latter is, for some reason waaaay more loaded.

the latter is very misleading. as has been stated, there is no such thing as 100% positive training. to call oneself such is inaccurate. to say that you use "mostly positive methods", though, just doesn't have that same "feel good" sound to it ;)

mudpuppy - I think it would be beneficial if you could pointg us in the direction of some of the studies/reports you have cited.  I would be interested to find out whether the studies that show rewards are more effective than punishment included a group that were "trained" using mostly reward but SOME correction and if so how this was controlled.

I notice that you say that rewards are MORE EFFECTIVE.  Depends what "effect" you want.... do you want the dog to be very reliable?  Do you want him to lean fast?  Do you want him to "learn how to learn", think, problem solve and get creative?  Do you want him to accomplish the task with minimum input from the handler?  Do you want him to accomplish the task with minimum effort/management from the handler? 

I also notice that it has been mentioned that not using punishment is not effective for SOME DOGS.... I approach from the other angle.  Such dogs require greater effort from their teacher to motivate them and communicate what is required.  Occasionally, I have corrected my own dogs.  But I suspect that, in the hands of a more skilled positive trainer, the correction would not have been necessary because the person would have been able to communicate with the dog more effectively.  This does NOT mean that my dogs need corrections.  It means that *I* have felt the need to use them sometimes with these dogs.  BIG diff. IMO.  To me, the fact that I have corrected a dog signals that I failed to communicate what I *did* want in the first place.  With hindsight, I wonder whether the correction really DID assist with learning, or whether it would have been just as quick, if not quicker, to simply stop the exercise, back up to a point where they were successful and try again.  Impossible to know as I can't rewind, erase, try again and compare with that dog.

There are some areas of training where I REFUSE to use punishment and I refuse to accept that any kind of punishment is necessary or even OK.  House training for example.  You know what, our dogs are so danged CLEAN.  I sometimes wish that they were as "good" in other areas as they are about toileting in the right place.  Can you guess how often they have been punished for house training mistakes? 

corgipower

Chuffy

I totally agree that "I train using mostly positive methods" is a world away from "I am a positive/positive only trainer".  The latter is, for some reason waaaay more loaded.

the latter is very misleading. as has been stated, there is no such thing as 100% positive training. to call oneself such is inaccurate. to say that you use "mostly positive methods", though, just doesn't have that same "feel good" sound to it ;)

Oh I know that positive only is a total misnomer and doesn't actually exist.  You could say its the "ideal" we strive for, despite knowing it is unattainable and therefore that makes us a positive based as possible?  Rather like breed standards describe the IDEAL pedigree, which is unattainable as no one is perfect.  But it is still what breeders aim for and the standard against which they are judged.  I sometimes think of the "positive only" myth as like a unicorn.... a lovely idea in a way, but does not exist.... but we nevertheless chase after it!