Someone sent me this link, which I think says it about as well as it can be said:
The trainers that identify as “positive reinforcement dog trainers” tend to spend way more time explaining what methods, techniques and tools they don’t use, rather than explaining their own personal training ideology. Their website content, blog posts and Facebook pages are full of tired, now decades-old, rhetoric decrying any method that isn’t theirs as “un scientific” or “outdated” and “based on old ideas and mythology.”
In short, though they say they employ a positive methodology, their message is primarily negative. Before I go on, let me set the record straight. When a trainer identifies as a “positive reinforcement trainer,” what they are essentially implying, or often saying outright, is that they reject the use of one of the 4 learning quadrants of operant conditioning -the use of positive punishment – in their dog training protocol. By doing this, they also imply, or also come straight out and say, that if a trainer doesn’t follow their philosophy, they probably just rely on the use of just one of the quadrants, positive punishment, to achieve results. Furthermore, they will go on to cite examples of extreme punishment techniques that they never do. They talk about old, “military style” training techniques. They describe horrific things like hitting, kicking or hanging a dog by the leash, and then categorize these abusive techniques right along with the use of any type of training collar. On and on they go. “Never work with a dog trainer who does this”, “I never do that”. Negative phrase after negative phrase.
What’s missing from their websites though, is an explanation of exactly what it is they do to achieve results, and what those results are....
... Effective dog training is results driven, and so is having a successful dog training business, especially in these times of social media, reviews, and instant status updates. If a dog trainer is not getting results, they probably aren’t getting much business. If that’s the case, then just like a candidate with slipping poll numbers, they begin to switch their message away from what they are capable of achieving, and focus on what others, like their more successful competitors, might be doing wrong.
.
19 comments:
I'm sorry you find positive training websites to be "negative." I've had exactly the opposite experience. Since learning clicker training in 1995, I've had nothing but fabulous experiences with both clicker trainers and their websites. I love all the free clicker training information online, such as that on the "Kikopup channel" of YouTube. Non-positive trainers almost all charge for info, whereas clicker trainers share freely... so what's not to like? :-)
My last 5 dogs have been clicker trained their entire lives, and earned over 60 titles in a dozen sports. I invite you to visit my YouTube channel to see my training results with my dogs, cat and parrot. (www.youtube.com/user/k9shrink).
As you mentioned, positive trainers eschew the use of +P, but most have no problem with -P. I also use -R often, particularly when training stockdog work or working with our parrot. However, if there is a +P way to train something, I will normally chose that over any of the alternatives.
I started training with a club in 1977, as a kid, and saw some extremely nasty +P during that time. As a 12 year old, I was told to scream "sit" at my small puppy on the first day of class, then push on his butt. If he wasn't responding to the cue by the third try, he was "defying" me, and I was instructed to hang him by his choke collar until he almost went unconscious. By the time he was 6 months old, if he ignored a command, I had to pick him up and swing him by his choke collar, then fling him to the ground and do an "alpha wolf rollover." This was a highly respected city training club in Michigan with many talented competitors. BTW, the dog was a fearful English Springer Spaniel.
I continued training with +P clubs from 1977-1995. At one point, in an effort to get my ACD from pulling on walks, I was told to yank him with a choke collar until he fell over... every time he pulled. Starting from puppy-hood, I pulled him down over 200 times per walk (4 miles), and he continued pulling like crazy. Looking back, he was probably desperate to get away from me. Despite weekly obedience classes for 7 years, I couldn't get a single title on that dog--he wouldn't obey for any amount of punishment. At that point, I got my JRT puppy and decided there had to be a better way. I discovered clicker training and enrolled both of the dogs. After retraining the first dog for just two months, we went out and got 2 obedience titles, 2 agility titles, and I was able to call him off the cattle during herding lessons for the first time.
Having seen just how awful and ineffective +P is, I was happy to ditch it as a "training tool." No thanks!
"Shaping" works.
And I love to get a dog turned 'on' for learning the way shaping does... but I just don't have the patience.
So my two 'sport' dogs - agility and rally obedience (go ahead, laugh, we have fun) - are trained with a mixture of early luring into position, some shaping, a positive marker - 'yes!', a few negative markers - nope' and 'oops'. And they give me results I'm happy with.
And they both like to work for me. We have fun.
I still hold in mind the fellow you mentioned once (pretty sure it was here) whose dogs got along just fine with very little praise at all. I balance that image against what I am doing. What is motivating the dog to work? Am I getting in the way of the dog? How much do I need to give them to keep them engaged? Less can sometimes be a lot more. Dogs are much happier to read our bodies for cues than to wait for us to tell us what we want (maybe they have no patience for shaping, either... show me, don't bother telling me, they seem to say...)
I do rely on treats and toys - unlike dirt work, the action is not it's own reward. Sad, I know. My poor dogs.
Anyway, bit of a stream of consciousness ramble here. I do thank you for giving me something to think about.
Eek, I made a typo:
>if there is a +P way to train something, I will normally chose that over any of the alternatives.
I meant to say "+R" there!
The first 'positive only' trainer I met was the Battersea Dogs' Homes when we worked on my learning how to use clicker training with the dogs I handled as a volunteer. It was a very useful and interesting session from which I did learn much. However, her sense of anger at other workers was palpable... and audible. Didn't think much of it, but didn't make anything of it either.
I have to say that reading blogs by 'positive only' trainers, it looks like that attitude isn't rare. There seems to be a deep-seated defensiveness and an 'us vs. them' attitude which for me is a huge turn-off. It doesn't seem useful, necessary or productive to me. I don't know if commercial jealousy plays any role in it but that article is an interesting read.
Life isn't all clickers and sunshine. I do primarily positive training with dogs, although I don't use a clicker.
They're all just tools. And while choke collars, prong collars and e-collars can all be misused, they can also be highly effective. For example, I used to have a Border Collie that hated the sound of a clicker (God knows why) and was a horrible leash puller. I went through various no-pull harnesses, which he would ignore and/or break in short order. I got him a prong collar. He immediately stopped pulling. AND, before anybody says he was traumatized, this dog would run to the door super excited, tail wagging and then sit down and wait for me to put the prong collar on him, just by watching me pick it up. Because he knew that collar meant he was about to go somewhere fun. Meanwhile, my mom had a GSD that was terrified of her prong collar, but walked fine on a choke collar. Different dogs. Different techniques. As long as it works.
I used to think that I was deprived because where I grew up, dog trainers were rarer than hen's teeth (well, there was *one*, but that's another story), equipment consisted of whatever I could persuade non-dog-owning relatives living abroad to buy and send and it was the same with books. If my mom let me have kitchen scraps, I had treats, otherwise nothing.
Nowadays, I think I wasn't so badly off! I was flexible, tried whatever I read, kept what worked (and kept in mind what didn't. Often it works elsewhere). If you're clear in the head about what you want, a dog will learn whether you've got food or not, whether you've a cheap length of chain or a nice leather leash or a flexi, a slip or a buckle collar.
I think where people who claim 'positive only' part with other trainers is over the entire training process. To my not-terribly tutored mind, a dog can be said to be trained when it understands three things:
1. 'X is easy to do and fun!'
2. 'I can do X anywhere!'
3. 'Even when I have the opportunity to do Y, I must choose to do X.'
1 and 2 are places that few people have issues with. That's where shaping and marker training work. However, 3 is essential and the idea of letting a dog make mistakes and correcting it is where it gets controversial.
Dear Patrick,
Skinner/clicker/positive training does work for most people because it is breed non specific, most people don't ask much of their dogs and the tricks which endear dogs to pet owners are best taught with simple rewards.
And, of course, very good trainers (and some S/C/P trainers are very good and could train irrespective of method.
These methods are contraindicated for breed specific and some high standard all breed work (I am told nobody has ever earned an OTCH with purely positive methods and ATF and war dogs are trained traditionally and/or with ecollars.)
Your cited poster was correct. One doesn't need to spend many minutes amongst S/C/P trainers to hear horror stories - with no evidence offered (save Bill Koehler's books) that anyone uses the terrible abusive techniques that distress these gentle souls. I have been training sheepdogs for 25 years and have visited many of the top traditional pet dog trainers yards. Yes, there are brutes, but they are rare - I'd estimate fewer than one percent.
Mental abuse: confusing the dog, ignoring and mistaking dog behaviors, asking for fantasies,
and assuming that dogs that get treats and are never corrected are happier than dogs trained by traditional methods is common as dirt.
Nonsense, human vanity, and feelgood beliefs can, and often do far more lasting harm to dogs than corrections.
Donald McCaig
Without using any positive punishment, the Brelands and Baileys trained 12 dogs to find land mines. Because these were military dogs, they were required to proof the behaviors at the end using positive punishment. To do this, they withheld food for 36 hours, then baited the "land mines" (in this case, a loud, painful explosion) with meat. 10 of the 12 dogs, who had been trained primarily with R+, and without any P+, still behaved as trained and found the land mine without setting it off.
The Brelands and Baileys ran a company called Animal Behavior Enterprises. They were scientists (the Brelands were formerly Skinner's graduate students) and they were interested in producing their product (trained animals) as quickly and efficiently as possible. While they wanted to be humane, they were otherwise not looking to be "nice and fluffy". They wanted results. They studied all methods of training. In the end, they use R+ as their primary training method, and never chose to use P+ (it was used only in the example above, and only because of the client's request). Their choices were supported by about 50 years of data on 15000+ animals.
I dislike that the traditional training community is vilified by positive trainers. I dislike that positive trainers are vilified by traditional trainers. There ARE bad people in each community, but neither should be judged as a whole by the bad apples. Having said that, the science supports the use of R+ and the avoidance of P+ for effective and efficient training. If one chooses otherwise, one should recognize that it is for other reasons (personal preference, tradition, the recognition that one can personally get faster results with traditional methods due to one's experience).
Well, i'm a force free trainer and must be getting results because my business is 6-figures, doubling yeae-on-year, and I am hiring two new trainers this year. And for every P+ trainer on the internet that comes across as defensive or negative, trust me, there are the same or more compulsion-based trainers casting aaspersions, name-calling, ("cookie-tosser" is my fave; it makes me smile) and generally not encouraging intelligent and collegial debate. Ok, have to go run the business. Happy training!
I am told nobody has ever earned an OTCH with purely positive methods
Not true! Denise Fenzi has done it twice! I know there are others out there as well but she's the only one whose name I can remember.
So, to recap, ONE trainer has gotten an OTCH. That's a pretty small number.
In fact, far more people have walked on the moon!
Bottom line, I think the case was made, not the exception proven.
And YES, clicker training works for MOST everything. But not all. And it is NOT the only way to train or the right way in all circumstances, as has been noted. Surely people know this? Surely people will admit this?
Bob Bailey was mentioned.
You DO know that Bob Bailey is quite up front that he is NOT a pure clicker trainer, right? He will tell you that himself. Yes, he thinks there IS a place to "shoot the dog." If a lion is attacking you, a clicker is not much use!
Bob Bailey and I have corresponded a little bit, and he is pretty upfront that he has used SERIOUS aversives on rare occassion when that was needed to prevent a dog or other animal from doing something that would seriously injure it or others.
To make it simple, Bob Bailey worked with animals that were up close and personal with explosives, such as mines.
You know why they call it "bomb-proofing" an animal?" Because you do not want to lose the dog, or the porpoise, or the sea lion when it goes in to mark on a bomb or attach something to a sunken item.
Ditto for a dog that barks in war. The tolerance for noise when a dog is on patrol is zero. No exceptions, ever, because human lives, as well as canine lives, are in the swing.
Bob Bailey, it should be said, was mostly training animals back when e-collars only give very powerful aversives, and the idea back then was that you turned it up high and nailed the dog very hard just once or twice or maybe three times, and if your timing was good that was all that was ever needed.
That still works, and it's still true if you are trying to send a NEVER signal.
But technology changes, and so too does what we can do with it.
Today's new e-collars come with simple tones, with simple vibrations, and with very low nicks (so low a dog or human can barely feel them) that enable a trainer to use an e-collar like a tap on the shoulder. The goal here is NOT to train a command (e-collars are almost never used to train basic obedience) OR to create a powerful aversive (i.e. to stop a bad self-rewarding behavior), but to reach out and REMIND a dog that it knows what to do in this situation, and it is being watched and will be rewarded when it does the right thing. A tap on the shoulder on these occassions can break a dog out of an obsession or a mental wandering, which is very simple thing to do, but very necessary with many dogs who (like humans) can have "selective hearing" when it comes to birds and other wildlife in the field, other dogs in the field, etc.
I know of NO e-collar trainer that does not understand how clickers work, and most e-collar trainers (I am not one) use clicker or verbal and food rewards for 90 percent of what they train. But e-collar trainers also know what an e-collar can do and what it can do well for the benefit of the dog. Have clicker trainers learned how to use an e-collar? Not generally. They do not own one, have never been trained to use one, have not tried it, and simply demonize it based on what they imagine it is about.
As Bob Bailey has noted, clicker trainers have gone a few stages too far -- at their most extreme they have taken on a lot of the trappings of a religion or a cult, with absolute dicta (such as nonsensical pledges, etc.) that ignores all discussion about time-place-and manner. But religion is not science, is it? It is the opposite of that. So, unlike the clicker trainers, Bob Bailey has kept all the tools, which is to say he is still suited up for science rather than religion.
I am in total agreement with what you said : )
I am well aware that Bob Bailey does not consider himself a clicker trainer. I don't consider myself a clicker trainer either (and I think the clicker trainers are just fine with that). ;) I am also well aware that just because he himself has used serious aversives, does not mean that he easily supports the use of them. If you have listened to his lectures or attended his classes (I have done both), you would understand this. He even mentioned in a discussion that he felt the 2 bomb dogs that failed could have been brought up to the necessary level of response WITHOUT the use of aversive proofing. But it was required of him to do the aversive proofing, even with the 10 dogs that did not fail the final test.
I'm not sure why e-collars are suddenly brought up. I'm simply talking about the science, not about specific techniques.
I know that Bob Bailey agrees that some clicker trainers have gone too far. Once again, this is a specific technique, not the science, so I won't talk about clicker trainers specifically either.
The science says that the best, most efficient and effective way of getting the behavior is to use primarily R+. P- and R- can also be minimally used. P+ rarely needs to be used. In the case of the scientists at Animal Behavior Enterprises, they used P+ 12 times out of over 15,000. That's less than 0.08% of the time.
Again, I don't care what any trainer uses, as long as it is humane. But I do care when certain methods are promoted as being better or necessary when the science isn't there to back it up. That's what's been happening in the comments, and I simply want to correct the misinformation. If you use P+ regularly, then it is a personal choice (which may be made for very legitimate reasons) but not a scientifically-supported choice *with regards to efficiency.* That is very different from saying "P+ doesn't work." It does work, it can work quite nicely, and I dislike when "positive trainers" try to imply or flat out say that it doesn't work. I don't like misinformation on either end of the spectrum.
You have have just restated what I said ... as if I did not write the word "rare" and as if you had actually mentioned the points I made about Bailey.
I am not sure what point you are trying to make. My only point (and the one made by others, and by the author of piece quoted here) is that clicker trainers spend a lot of time demonizing every other form of training, creating straw men, and acting as if their method is "science" while all other methods fail. n fact, if you actually had read Skinner, you would realize that the "science" of training (to the extent there is such a thing) talks about aversives quite a lot. In fact, the famous "Skinner box" had an electric floor! But are aversives needed very often? No. But does the U.S. military and the Secret Service, etc. still use them because they WORK? Yes. Aversives also work to train U.S. troops where rubber bullets, shock vests and shock knives teach real field combat. A drill instructor is not much for click and treat!
My question is, are there ever situations where pure positive does not work? I frequently get told by pure positive trainers that the reason they don't use P+ is because they can achieve all the same results without it. The argument isn't that P+ doesn't give results, it's "why would I use P+ if I don't have to?"
I find it hard to defend the use of P+ in the face of statements like that because the answer is usually interpreted as the trainer being lazy or wanting a quick fix, which is not how I see it at all. In my view, pure positive training does not have the ability to teach a dog to NEVER do something the way that P+ does, because there is never any serious consequence for breaking the rules. Unless a dog is starving to death or lives in total isolation, witholding food or praise isn't going to be much of consequence if the dog really wants to break the rules (think dog chasing cars).
Dear Patrick,
Stouthearted hounds wondered, "... are there ever situations where pure positive doesn't work?"
Many top pet dog trainers have told me that purely positive methods don't produce reliable results. I defer to their judgement.
I have never known a Novice trainer who used Pure positive methods in sheepdog training. Ethology trumps behaviorism every time. For sheepdogs, as Jack Knox says, "Allow the right, correct the wrong."
Donald McCaig
Donald McCaig
Puppy trainers or trainers that have not worked much with hard-wired behaviors simply do not understand the limits of cheese. A small blind spot. Most, of course, have never actually read Skinner either. See my post this morning on "Shocking News About B.F. Skinner" and the one on Kingergarten training versus Marine Corps Recruit training and the difference between cake and cookies.
P
Donald, that is exactly what I was thinking; that while pure positive does produce results the results are not always going to be reliable, particularly when it comes to hard-wired, self-reinforcing, genetically-coded behaviors.
I'm sure there are pure positive trainers that will cite anecdotes to the contrary, but having worked closely with dogs that have extreme prey drive I would never put faith in a pure positive training system to curb that behavior even if I wanted to. The same is true for dogs that are genetically wired to be aggressive towards humans (think guardian breeds like the Fila and the Presa).
Pure positive never begats reliability in any training. Because life is not always positive. Training with positive and negative reinforcement mimics life. Dogs that faithfully and reliably do their work have to be exposed to life. Not a set up platform of controlled situations for them to repetitively do. They are taught how to learn and how to be bomb proof by understanding negative and positive reinforcement as well as being able to receive it. Unwarranted dog bites will never be negated by a clicker or treat. Much of what has been said already. Treats and clickers have their place
But all subjects of training have to experience the full range of life to have the ability to learn rather than being trained by someone all the time. Dogs like people can be taught how to learn as well as being trained. And, the best dogs are those that have been trained how to learn. Thanks for your blog. Apologize for any repetitive words or commemts. I have learned from your blog.
Post a Comment