Friday, April 23, 2010

Suzanne Clothier Seminar: Where Clicker Trainers Go Wrong

It came as no surprise that Suzanne considers herself a humane trainer. I knew that by reading her book, by the fact that she calls her method “relationship centered,” and by the observation that the positive training community has rallied around her. So, I was completely unprepared for what she said next:

“All positive training is one of the stupidest things I’ve ever heard.”

Come again? Aren’t you for training our dogs in positive ways? What do you mean, it’s “stupid”? Needless to say, this statement really set me reeling. But you know what? She had some valid criticisms, and I think it’s important that as clicker trainers, we consider what she says, and find ways to improve our training as a result. Again, my understanding of her criticisms are my own, based on stray comments she made. This entry cannot be considered a complete commentary, and I only hope that it’s accurate. That said, in reviewing my notes, I found four ways we often go wrong:

Clicker Trainers are Too Cerebral
As I mentioned previously, Suzanne believes that consequences are useful in training dogs. Although I disagree with some of the consequences she believes are acceptable, I do concede that they can have a place in a thoughtful training program, and I’ve since found some ways to improve my training with Maisy as the result of using them.

The problem is, clicker trainers sometimes shy away from consequences entirely because we’re thinking too much. I know I fell into this category! When I was considering a consequence, I would think about which of the operant conditioning quadrants it fell in. If it fell in the “wrong” quadrant, I wouldn’t do it. Instead, Suzanne encouraged us to look at our dogs, and to decide if the consequence is humane or not. As long as we aren’t hurting or scaring our dogs, we’re probably okay.Instead of thinking about our actions based on some chart, we’re better off thinking about our dog.

Clicker Dogs Have No Impulse Control
Okay, this is a bit of an over-exaggeration of what she said, so don’t go around quoting this verbatim. What Suzanne actually said is that when she sees a dog throwing behaviors, she knows two things: it was clicker trained, and it has low impulse control. This does not mean that all clicker dogs struggle with impulse control, but it can be a by-product of the training method if we don’t get adequate stimulus control over the behaviors we teach our dog, and if we don’t take the time to teach our dogs how to relax.

I love having a dog who wants to work, but sometimes I think I’ve created a monster because I haven’t worked that much on teaching her how to just be. Sometimes, I don’t use the clicker at all because it is creates too much arousal. She loves it so much that she loses control over herself.

Clicker Trainers Micromanage Their Dogs With Cookies
There are two sub-points here. First, clicker trainers tend to rely too much on management, and don’t spend enough time teaching their dogs to be responsible for themselves. (I wrote about this previously in this entry.)

Second, clicker trainers tend to rely on cookies far too much. Suzanne is not against cookies, she just thinks we need branch out. She said that what we really need to do is harness intrinsic motivation. We need to teach our dog to work for other things. Life rewards such as getting to go for a walk or using play as a reinforcer fits here, but Suzanne is big on using praise and the social relationship as a reinforcer, too. Cookies are great, but they’re the icing on the cake of social relationships.

Clicker Trainers Rely on “Recipes” Too Much
Finally, Suzanne was critical of clicker training because it can result in the handler treating the dog like a computer: we train them based on stimulus and response, or input and output. We would find it insulting to be treated this way, so why wouldn’t our dogs? Instead, we need to see the whole animal, and tailor our plans to their needs and preferences. Even counter conditioning will fail, she said, if we don’t keep the dog safe. Training recipes might be a nice starting point, but we need to go beyond them.

As I mentioned, I know I fall in the “think too much” category. I also recognize that my use of the clicker could have contributed to Maisy’s impulse control problems, because I'm just not that good at getting things under stimulus control. So far as micromanaging with cookies… well, guilty again. I do rely a lot on management, and I’m just learning how to use things other than food as a reinforcer. I think I am safe on the last point; but I think my strength lies in being able to evaluate various “recipes” and tweak them so that they work to both Maisy’s strengths and my own.

So, fellow clicker trainers, do you recognize yourself in any of these descriptions? If so, what are you going to do to improve your training?


Lindsay said...

While I don't consider myself a clicker trainer, I think that I do fall into the micromanage category with cookies. More so with Heffner than with Bess. On the up side, I have been trying to use other things as a reward. Playing with toys in certain situations and lots of head hugs in others. When I'm waiting in line to go into the ring for agility or rally with Heffner, I've usually run out of treats because I can't take any into the ring with me. So we play little games while waiting to sort of keep him occupied and keep him from fixating on another dog. He likes me to hug his head, or scratch his face, or I have him give me "little" kisses on my cheek. Just little things that get his tail wagging. I know that they're things that he enjoys and it's rewarding for him on a one-on-one basis. And I certainly enjoy it!

Laura and The Corgi, Toller, & Duck said...

hmmm. Obviously these are really broad statements and you can make negative stereotypes about any group of people on any topic.

1. While I personally use a small amount of punishment as I mentioned previously I don't know that you would have to if you were a really great trainer. The best trainers break things down and shouldn't leave any holes in their training that would require the use of punishment. Would it speed things along and can it provide useful information to the dog? sure, but is it neccessary to have a well adjusted and well trained dog? I don't think so. Of course I don' tthink i'm that great of a trainer.

2. Lacking impulse control seems to me the fallout of many many dogs, regardless of method used to train. I don't think encouraging a dog to think and offer behaviors leads a greater risk of it then with other methods. I just honestly don't see how a dog throwing behaviors out is more of a symptom of impulse control. But yes, this is something I've worked very hard with on Vito and will continue to work very hard on since he needs it!

3. This point I agree on a bit more. but I also think that management can be very good training, especially if you are trying to train as positive as you can. If your dog isn't ready then if you're not managing you're going to have to resort to punishment. Of course ideally you will walk the fine line between challenging your dog and still setting him up for success. I definitely rely too much on cookies though :)

4. Actually I find that traditional training relies much more on recipes then clicker. Shaping is all about finding the steps that your dog is offering and building from there, there is no set plan. if the dog isn't responding the way you like then you need to think on your feet and the blame is all on you, not the dog and the pressure on you to think of a new way of communicating the exercise. Traditional trainers place more pressure on the dog so can follow the "recipes" more easily.

Not saying that you agree with all of her points, I just like to think about these things!

Anonymous said...

See, this is where I run into a problem. I find that so many trainers who use corrections seem to think that the dog should work for praise/relationship. However, if you're using corrections in your training program, your dog is not "working" for the praise: he's working to avoid a correction. Coercively trained dogs seem to really love praise. Of course they do! Praise is a "safe" signal to a coercively trained dog. If he's being praised, he knows he's not going to get corrected at that moment.

We have a local trainer who is VERY heavy into prongs (prong collars on every single dog who comes to their classes, even puppies). They advertise that the dog should not work for cookies, they should work to "please" you. Hogwash. The dog is desperate for signs of praise and approval from the trainer not because he has some "wants to please" drive, but because signs of praise and approval are mutually exclusive with prong collar corrections. *sigh* I'm not saying that this is what Clothier does (although I know she prefers prongs to head halters), but it's something I see quite a bit.

All this isn't to say I don't use praise in my training or that I don't think praise is useful in a training program. However, praise is a conditioned secondary reinforcer for my dogs. If I began using praise exclusively while continuing to avoid corrections, I would find it hard to train new behaviors. Praise is also a keep-going signal to tell my dogs they're on the right track.

I have one dog who will try her heart out for me with no rewards in sight, simply because of the relationship we've built up. Relationship = reinforcement history. My part of the bargain in our relationship is that I will be clear, fair, and generous. I agree that training with cookies alone is not as effective as other programs. I just prefer to use Premack rather than corrections.

Crystal (Thompson) Barrera said...

Yes, these are some rather broad statements, and these were things that she said in a mostly off-hand way. It certainly wouldn't be fair to say that Suzanne believes clicker training is misguided or wrong or... whatever. I didn't see her with a clicker during the weekend, but she does use them. They're an excellent tool, after all.

What I find interesting based on these comments is that these mistakes aren't exclusive to "clicker trainers." Lindsay is identifying ways that she, who doesn't identify as a clicker trainer (what would you call yourself Lindsay?), falls into the same errors. And I have to admit, I know nothing about traditional training- I wouldn't even know the correct way to put on a choke chain- so Laura's assertion that traditional trainers also use recipes is interesting.

Sara, I think you'd find that Suzanne agrees with your comments regarding praise, relationship and corrections vs. reinforcements. Suzanne is not afraid to use treats, and would probably start out using more of them before moving to more praise. Certainly, at this point Maisy can work for longer periods of time with just praise... but I do need to back that up with treats or ball-throwing periodically or the praise becomes pleasant but less meaningful.

Kristen said...

Many of these comments leave me surprised...

I wonder if it's more of something that those teaching and writing about clicker training should be thinking of rather than something necessarily directed to those clicker training. While clicker training classes are much more accessible, there are still some less-recent information being passed on and many people still learning from books/online.

And if these things are what people come away from classes or reading with, then instructors are going wrong somewhere.

I'm happy to say that I really don't think I, or the majority of my students, fall into these categories. We talk about stimulus control from early on. I don't like it when dogs see treats and sit-down-sit. (...the sit is okay! For pet dogs, Sit is ALWAYS a right answer!). I don't like dogs with poor self control. I don't like it when people learn in classes to manage and not train their dogs, it gets stressful for the people! But I also dislike when people don't know how to manage before the training is in place. Recipes are great for beginners...whether they're training dogs basic behaviors, competition behaviors (...ever read a field training about recipes!), or cooking... learning how to create a training plan is a skill.

Crystal (Thompson) Barrera said...

I think the majority of the audience were people who are teaching classes... and it makes sense that she'd direct such criticisms to us as a result. These are things that need to be taught to those who are new to training. I certainly wish my early classes with Maisy had focused more on impulse control.

Lindsay said...

I'm sure what I would call myself actually!:) I use a clicker sometimes, but not frequently enough that I feel I deserve to fall under the category of clicker trainer. Although, because I do use it at some points, does that make me one? I'm not really sure. I guess that I just consider myself a trainer when it comes to my dogs and just a trainer in general.

Crystal (Thompson) Barrera said...

Lindsay, I suppose you don't have to call yourself anything other than a trainer.

I don't actually use a clicker that much- probably no more than you do- but philosophically I think of myself as a clicker trainer.

Ashley said...

I'm a little late to the party, but I found her last point about treating dogs as input/output machines to be interesting.

I don't know if you heard about this, but her and Jean Donaldson had a bit of a spat a while ago about a video that Jean posted of her dog, Buffy, humping her leg on cue. She was demonstrating, IIRC, that you can train (and put on cue) just about anything, even something the dog doesn't do "naturally"--it sounded like Buffy wasn't much of a humper to begin with. Clothier wrote a blog post that basically said that Donaldson was sexually abusing her dog.

I think that right there is the essence of both these trainers' styles. Donaldson is much more input/output, charts and graphs, and Clothier is more "touchy feely," relationship-based. Despite the fact that I would call them both "positive trainers," they're on completely opposite ends, I think. (Just for the record, I prefer Donaldson's philosophy, but that's just the kind of person I am)

Crystal (Thompson) Barrera said...

Good lord, Ashley. I didn't mean to ignore your comment! Just super behind in my email.

Anyway, I think your conclusion about the differing styles of Clothier and Donaldson is very interesting. I've never met Donaldson, nor seen/heard her lecture through video/audio, but I think I'm closer to her style than I am to Clothier's... Although, maybe not- I am very much into relationship. But I like charts and graphs, and while I don't think our dogs are input/output machines, I do think the science supports being able to train just about anything the dog is capable of doing. I just don't think we SHOULD ask our dogs to do stuff just because we can. So... whatever that makes me.