Aidan

[itsadogslife]

I do a lot of long-line work with a tip of the hat to Koehler. But we know a lot more about learning now so to do it exactly the same way would not make sense to me.

Aidan, sorry to make the topic you, it isn't meant to be, I am just trying to catch your attention. I am just interested in your explanation or views on the above comment you made in the other thread. I am hoping to avoid the topic being on you know who, and simply wish to understand how knowing more about learning dissuades you from doing it exactly the same way.

[Aidan3]

Hi itsadogslife, it's a good question, so thanks for bringing it up

Back when Koehler came up with his long-line training it was cutting-edge. He had built on Thorndike (instrumental conditioning), Pavlov (classical conditioning), Lorenz (ethology) and Most (military trainer).

Of course, since then we've learned a lot more about learning, and animal training. The theory is better developed, and thoroughly tested using the scientific method so that we know that it is both accurate, replicable, and predictive. We also know the limitations of the theory, as a budding academic in this field I pay keen attention to the limits of behavioural science.

Of course, all this theory comes from collective experience. Outside of academia we have forums such as this one, dog clubs, behaviourists, and animal trainers from other fields all adding to the collective wisdom. There really isn't a big gap between theory and experience, theory is forged in the fire of experience.

The long-line training that I do is different to the long-line training that the Koehler method uses because I can accommodate a wider variety of dogs and handlers by applying what we know about learning today to the method. For me, the long-line is a very effective way to control the environment whilst giving the dog freedom to make choices, as it was for Koehler. For me, the focus is on letting the dog make good choices so that he can be rewarded, rather than setting him up to make bad choices early on that can be punished. I understand that Koehler trainers don't see his method as "setting the dog up to fail", but from a behaviour analysts perspective, that is what actually happens in practice. It becomes about avoiding the correction, rather than working for the reinforcer.

Another thing that you find is that a lot of dogs see the leash as a cue (or antecedent). I suspect Koehler and some of his proteges had a very deft touch and knew how to avoid this, but time and time again I've seen off-leash performance fall apart in dogs trained on a long-line. A lot of people get around that today by using an e-collar. I go to great lengths to explain to clients how to give guiding corrections appropriately without teaching the dog to wait for body language or leash signals.

But really, the main difference is close attention to "are we actually reinforcing this behaviour"? Praise will not do it, 99 times out of 100. Begin training without any leash at all, no corrections of any kind, and see how far praise gets you. Some dogs are particularly motivated by praise, others not so much. I'm lucky to have one who is, but it's still not a very good reinforcer compared to food, toys, retrieves, tugs, bites or pretty much anything else.

The other benefit of using positive reinforcement is that if the handler makes an error, it's almost never a problem. If the handler makes an error with aversives, it can be a big deal. A lot of people don't even notice how much trouble they have gotten themselves into. People can talk themselves into anything, particularly if someone seen as an authority tells them it's OK. Our reality is surprisingly malleable.

What else? I guess things like how I raise criteria such as distance, duration, and distraction. Rate of reinforcement. Schedules of reinforcement, making the behaviour stick in the absence of reinforcement. It's actually not very complicated, and easy to teach.

[itsadogslife]

What do you think is being taught by the long line? And how does learning theory improve upon it?

As a means of narrowing the answer: I think there is two things being taught, one is in the foreground and easily observed, the other in the backgound. The two are equally important in both the method and the philosophy that informs it.

Edited December 15, 2011 by itsadogslife

[pipsqueak]

Another thing that you find is that a lot of dogs see the leash as a cue (or antecedent). I suspect Koehler and some of his proteges had a very deft touch and knew how to avoid this, but time and time again I've seen off-leash performance fall apart in dogs trained on a long-line. A lot of people get around that today by using an e-collar. I go to great lengths to explain to clients how to give guiding corrections appropriately without teaching the dog to wait for body language or leash signals.

Could you explain more about "guiding corrections", please?

[itsadogslife]

Posted by Aidan2 Another thing that you find is that a lot of dogs see the leash as a cue (or antecedent). I suspect Koehler and some of his proteges had a very deft touch and knew how to avoid this, but time and time again I've seen off-leash performance fall apart in dogs trained on a long-line.

But the long line is 15' for a reason. You are instructed to turn whilst the line is slack, thereby avoiding giving the dog any cues. An experienced trainer could use a shorter line without giving cues, but the length of the line was specifically chosen so that even a novice could turn without giving a cue.

[Aidan3]

Another thing that you find is that a lot of dogs see the leash as a cue (or antecedent). I suspect Koehler and some of his proteges had a very deft touch and knew how to avoid this, but time and time again I've seen off-leash performance fall apart in dogs trained on a long-line. A lot of people get around that today by using an e-collar. I go to great lengths to explain to clients how to give guiding corrections appropriately without teaching the dog to wait for body language or leash signals.

Could you explain more about "guiding corrections", please?

Sure. Let's say we're using the long-line to teach a recall (this is not what itsa is referring to, but it serves a better example). It's difficult to achieve errorless learning in practice, so we're going to encounter some errors where the dog does not respond to the trained cue.

We give a moment's grace to respond. Then, with a harness or flat collar, we give a little pop. When the dog starts coming in, we take up the slack in the line. If he starts to wander off or stop, we keep bringing in the line gently.

The trick is to avoid teaching the dog to wait for the pop. It needs to be used very sparingly.

[Aidan3]

Posted by Aidan2 Another thing that you find is that a lot of dogs see the leash as a cue (or antecedent). I suspect Koehler and some of his proteges had a very deft touch and knew how to avoid this, but time and time again I've seen off-leash performance fall apart in dogs trained on a long-line.

But the long line is 15' for a reason. You are instructed to turn whilst the line is slack, thereby avoiding giving the dog any cues. An experienced trainer could use a shorter line without giving cues, but the length of the line was specifically chosen so that even a novice could turn without giving a cue.

They know the line is there. Unless the dog is highly motivated by praise, you can't teach anything without giving a correction.

[itsadogslife]

Posted by Aidan2 I understand that Koehler trainers don't see his method as "setting the dog up to fail", but from a behaviour analysts perspective, that is what actually happens in practice. It becomes about avoiding the correction, rather than working for the reinforcer.

Sorry to keep asking questions, I keep seeing different things in your response. But my question is, on what basis should I accept the "behaviour analysts perspective"?

[itsadogslife]

Posted by Aidan2 They know the line is there.

But you said that Koehler and his proteges must have had a deft touch and so knew how to avoid this. How did they avoid this? Obviously they were also working with the long line.

[Aidan3]

What do you think is being taught by the long line? And how does learning theory improve upon it?

As a means of narrowing the answer: I think there is two things being taught, one is in the foreground and easily observed, the other in the backgound. The two are equally important in both the method and the philosophy that informs it.

The basic idea is for the dog to pay attention to what you are doing, that the right choice is to follow you.

Learning theory makes this accessible to a wider range of dogs and handlers. By using positive reinforcement we can ensure that the dog is willing. We can step outside the rigid formulae and work on principle, rather than method.

[Aidan3]

Posted by Aidan2 They know the line is there.

But you said that Koehler and his proteges must have had a deft touch and so knew how to avoid this. How did they avoid this? Obviously they were also working with the long line.

Everyone following the same recipe will get different results. Otherwise every dog at the local obedience club would progress at the same rate and get the same score in trial.

Some are better at it than others. Good timing and a good sense of when to progress and how far can condition the dog to stop paying attention to the line.

[OSoSwift]

A very interesing and insiteful topic, thank you for starting it, I read with interest.

[Aidan3]

Posted by Aidan2 I understand that Koehler trainers don't see his method as "setting the dog up to fail", but from a behaviour analysts perspective, that is what actually happens in practice. It becomes about avoiding the correction, rather than working for the reinforcer.

Sorry to keep asking questions, I keep seeing different things in your response. But my question is, on what basis should I accept the "behaviour analysts perspective"?

You don't have to accept anything, but behavioural science is a rigorous process. We have literally thousands of experiments that try to dissect every little facet informing us.

If you really want to satisfy yourself, find a fenced area. Use no collar or line. Use no verbal corrections or threatening postures. See how far you get with the Koehler method sans collar.

Then, for good measure, repeat the experiment with a check chain, long-line, but in the absence of praise. I'm willing to bet this experiment gets much better results.

[itsadogslife]

Posted by Aidan2 Everyone following the same recipe will get different results.

How do you know this? This goes to the question of why should I accept the behaviorist's perspective? This is also goes to your suggestion that there is not much of a gap between theory and practice?

Can you point me to where a behaviorist has studied and observed a qualified Koehler class over a period of time to reach such a conclusion?

Otherwise every dog at the local obedience club would progress at the same rate and get the same score in trial.

I would be far more impressed with this comment, if you could show me a study of Koehler class over an extended period of time. Has any such study been conducted? Again, on what basis should I accept the behaviorists perspective? You have said the science is rigorous - okay, where can I find a rigorous empirical study done on the long line use as instructed by Koehler?

[itsadogslife]

Posted by Aidan2 You don't have to accept anything, but behavioural science is a rigorous process

You have misunderstood my question. You said: "I understand that Koehler trainers don't see his method as "setting the dog up to fail", but from a behaviour analysts perspective, that is what actually happens in practice."

On what basis should I accept the behaviorist's perspective as opposed to the Koehler perspective that the "method sets up the dog to fail".

[Aidan3]

On what basis should I accept the behaviorist's perspective as opposed to the Koehler perspective that the "method sets up the dog to fail".

Try the experiment I suggested earlier. If you can get equally good results without corrections as you can with, then you know that I'm wrong.

[Aidan3]

You have said the science is rigorous - okay, where can I find a rigorous empirical study done on the long line use as instructed by Koehler?

The science of learning theory is rigorous. We know how operant conditioning works, we know how to identify antecedents and consequences of behaviour. There have been studies showing that praise is (with rare exception) an ineffective reinforcer (http://www.amsciepub.com/doi/abs/10.2466/pr0.1967.20.3.843) unless coupled with a primary reinforcer, but you can satisfy yourself of this in 5 minutes. Try teaching a new trick using praise and praise alone. Then try it using praise followed by food. Which one works best?

We also know about the function of discriminative stimuli (for e.g you could look up what an S-delta is). We know about conditioned reinforcers (both positive and negative). We know about Sidman avoidance procedures (signaled reduction in the frequency of aversives functions as a reinforcer). We also know a tonne of stuff about the problems associated with aversive conditioning.

But common sense would have to tell you - if you want to train a dog efficiently, to a level that is reliable, you do what people who have to train dogs efficiently to a high level do. That hasn't been Koehler for a very long time.

Posted by Aidan2 Everyone following the same recipe will get different results.

How do you know this?

I've never seen a dog club using Koehler long line in a group class, the comment was referring to any group of people following the same recipe, perhaps Koehler long-line is the sole exception? I doubt it though. Go to any dog club and see if every dog is doing exactly the same thing. Better still, follow up in 12 months.

[itsadogslife]

Posted by Aidan2 But common sense would have to tell you - if you want to train a dog efficiently, to a level that is reliable, you do what people who have to train dogs efficiently to a high level do. That hasn't been Koehler for a very long time.

Ok, I will ask the same question that I asked in the previous thread. Do you know of a method that can take a class of handlers and their dogs and get a 80% success rate in passing a AKC Novice test in ten weeks?

Go to any dog club and see if every dog is doing exactly the same thing. Better still, follow up in 12 months.

I took my first dog to an obedience club and can assure you that nobody was doing the same thing. For one everybody was using different tools, some on flat collar, others on gentle leaders, others on harnesses other still on martingale. To say that such a club followed a 'method' in teaching would be a gross overestimation. Having people walk about doing the same things, but with different tools during class is not a method. I expect, that most obedience clubs are similar.

I've never seen a dog club using Koehler long line in a group class, the comment was referring to any group of people following the same recipe

To say it again, obedience clubs are the very last place I would go to see people following the "same recipe".

perhaps Koehler long-line is the sole exception?

But why don't you know whether it is or isn't the exception? Surely a simple empirical study would very easily establish it one way or another. Why has such studies not been conducted by behaviorists? You have said that there is a close relationship between theory and experience - but here, I only see a gaping void. Surely you can see that there is a vast gap between experiments conducted in controlled enviroments as opposed to real people with real dogs in real life.

If you want to know whether the long line brings uniform results one need only seek out a qualified Koehler instructor, spend six months observing and keeping records and you will have your answer. No matter how many experiments behaviorist do in controlled enviroments it doesn't add up to anything (at least in recommending a method of training) until they take those experiments out into the real world of people and dogs and have them empirically tested in real world conditions.

[huski]

Ok, I will ask the same question that I asked in the previous thread. Do you know of a method that can take a class of handlers and their dogs and get a 80% success rate in passing a AKC Novice test in ten weeks?

itsadogslife... just out of curiosity how many obedience trials have you been to? Have you seen many novice run outs?

[itsadogslife]

itsadogslife... just out of curiosity how many obedience trials have you been to?

one.

Have you seen many novice run outs?

see above. lol