Clicker Expo 2011 (Chicago): Kathy Sdao- You're in Great Shape!

You guys, I loved, loved, loved Kathy Sdao. In fact, I loved her so much that I went to four of her sessions (the equivalent of a third of the weekend). I skipped sessions that I’d planned on seeing in order to attend her sessions. And I wanted more. Someone bring her to the Midwest, please; I’m dying for a working spot with her.

Why did I love Kathy Sdao so much? Well, she’s wicked smart and has had tons of practical experience, including training dolphins to do defense-related open-ocean work for the US Navy. How cool is that? Her lectures are just packed full of great information and fascinating stories. She’s energetic and engaging and enthusiastic and entertaining! Go see her… and take me with you.

Okay, enough gushing. Dog training is all about changing behaviors, right? The thing is, though, we can’t manipulate the behavior directly- the behavior belongs to the dog. So what we need to do is manipulate what happens both before and after the behavior.

There are many ways to get a behavior to happen from the front end, everything from physically prompting or luring the dog to capturing or shaping the behavior. The method you choose is up to you, and Kathy said that you should understand the science so that you can choose the method for yourself instead of letting someone else decide for you. However, no matter how you decide to train, remember that the fundamental law of behavior is that consequences drive behavior. No matter what you do on the front end, your power as a trainer comes from being a reinforcer, not a commander.

In fact, that’s so important that I’m going to repeat it: Your power as a trainer comes from being a reinforcer, not a commander. If you are labeling your dog in some way, be careful! Dogs that are “distractible” and “stubborn,” say more about you than about them. Specifically, those labels say that you don’t reinforce them enough! Awesome dogs come from awesome reinforcement histories.

Although there are lots of ways to get the behavior, as the title implies, this session was about shaping. Kathy defined shaping as “teaching new behaviors by use of differential reinforcement, systematically reinforcing successive approximations toward the goal behavior.” Or, to put it simply, teaching a dog to do something one small step at a time.

Kathy demonstrates shaping.

Shaping can take place through either reinforcement or punishment, but it should go without saying that Kathy focused on the reinforcement side. Click, then treat. You don’t need to punish wrong responses, you don’t even need to mark them with a No Reward Marker. The opposite of reinforcement is not punishment, after all- the dichotomy is reinforcement versus no reinforcement. You’re either clicking the dog for getting it right, or you’re not clicking.

So, why shape? Well, of course it’s based in the science of operant conditioning, which means that it takes advantage of using consequences in your favor. But it also makes the dog an active participant- there’s two of you present, so why not use both brains? What’s more, it gives the dog a sense of control, and can help him make sense of a seemingly random and punishing world. It allows him to find a way to make the world work in his favor, thus creating a sense of internal motivation and desire to interact with you. And of course, sometimes it’s the only way to get a complex behavior the dog wouldn’t offer otherwise.

Once you’ve decided to shape a behavior, there are a few pre-requisites. First and most importantly, you need to know what your goal behavior is. Be specific- does “come” simply mean “move towards me,” or does it imply that the dog will move towards as soon as you call, quickly, and directly? And what should he do when he arrives- stick around or dash off again? Know what you want.

Once you know what the completed behavior is, you need to know what the first criterion is. What sliver of behavior will you be clicking for in the first 60 seconds? That sixty seconds thing? Yeah, she means it. Your job as a trainer is to choose the criteria, do a brief bit of training, click when it happens (and withhold the click when it doesn’t), and then stop and ask how it went. In fact, that was sort of the refrain throughout the weekend. Many of the presenters I saw recommended working in really short bursts with lots of thinking time in between.

What are you thinking about? Well, you need to assess your dog’s progress and then plan the next 60 seconds. You’re trying to figure out what your criteria will be the next time around. Kathy recommended making the job easier when your dog is getting it wrong half the time, or if he’s getting less than 8 clicks and treats a minute. That’s one click every 8 to 9 seconds- not much time! But it helps underscore just how small your steps should be during shaping. If your dog is doing well, getting 8 to 11 clicks per minute, and getting things right somewhere between 60% to 80% of the time, you’re on the right track. Continue working on that criterion. And if your dog is getting it right more often than 80% of the time, or is getting clicked every 4 to 5 seconds, it’s time to move on to the next step!

Incidentally, if you’re working on a duration behavior and can’t get in enough repetitions in order to get a high rate of reinforcement, you should increase your density of reinforcement so that the dog is getting the same amount of treats- 12 to 15 per minute’s worth of behavior. Deliver them however you want- all at once or one at a time- but make it worth the dog’s while to do that behavior.

Kathy also covered Karen Pryor’s ten laws of shaping, paying special attention to number 5 (stay ahead of your subject by having a plan) and number 8 (don’t interrupt the session gratuitously, which includes talking. Don’t be more distracted than you expect your dog to be!). She also noted that Karen herself says that number 3 (put the current response on a variable ratio schedule of reinforcement before raising criteria) isn’t necessary with dogs, which I thought was fascinating, and relieving- I’ve never done it!

Remember that when you’re shaping, your main job is to sit back and watch your dog. You should not be doing the moving, your dog should. If your dog stalls out and gets stuck, your response should not be to help him out. Instead, have faith in your reinforcement history. He can figure it out. If need be, go ahead and reduce your criteria, but resist the impulse to help him out by pointing or luring him.

And that’s shaping in a nutshell. I feel like I glossed over so much of what she said- her sessions really are just jam packed with great information- but this should give you a good idea of what she said. Even though I shape a lot (I find it addicting, even if I’ll never use that behavior), I still took a lot away from this session.

I have to admit, I’m not very good at the planning/assessing piece. I just sort of… sit down and click. I usually don’t end up where I’m planning to go, either. That’s fine for casual sessions, but I’ve got to admit, it impedes my progress for tricks or competition behaviors. But between this session and Kathy’s presentation on cuing skills, I think I’ve figured out what I need to do! In fact Kathy’s cue seminar was amazing, and I cannot wait to share it with you.