I recently had someone ask a great question about one of my YouTube videos of my rat Georgie. In the video, Georgie gets clicked for going through a piece of PVC pipe. The person asked:
I don’t mean this as a criticism, but as an honest question– I’m looking at the timing of your click, and am not sure what you are clicking for– going to the edge of the table?
You can watch the video here:
Georgie goes through the PVC pipe tunnel
Trainers often think of a behavior as a single, discrete event. If the trainer is working on teaching the dog to sit, the trainer watches the dog and clicks whenever the dog’s bottom touches the ground.
However, it can be helpful sometimes to think of behavior as a continuous stream of of movement, rather than as a single discrete event. Behavior is always occurring. A dog’s sit can be thought of as a collection of many smaller behaviors:
The dog approaches in front of you, looks up at you, begins to lower his bottom, keeps lowering his bottom, and finally, touches settles the weight of his hind end onto the floor. You click your clicker. But, wait, there’s still more behavior that’s going to happen! The dog shifts his weight forward, stands up, walks forward, wags his tail, and takes a treat from you. Each of these could be broken into even smaller muscle movements.
Trainers sometimes don’t spend time assessing the behavior that comes before and after the click, because they are focussed on looking for the behavior they are going to click. However, depending on what behavior the trainer is trying to teach, these before and after behaviors can be very important. They can either help the training along or they can get in the way. Sometimes, a trainer might actually want to click for a behavior that is before or after the actual target behavior, in order to enhance or improve the target behavior or to eliminate other unwanted behaviors.
So, back to the video clip of Georgie the rat. If you watch the video clip, you’ll notice that my timing does appear to be pretty off! Rather than clicking when she was in the tube or when she was exiting the tube, I’m often clicking several seconds after she has exited the tube.
However, this was on purpose! What looks like “late” clicking was actually deliberate. Originally, I clicked approximations of going farther into the tube, until she was going all the way through. At the end of training, I was clicking right when she exited the tube. However, rats like dark, small spaces and she started slowing down in the tube and sometimes even stopping and waiting in it. Delaying the click until after she had exited meant that I was clicking for forward movement, rather than for stopping. This helped keep her moving through the tube and the stopping went away.
I wrote about some similar themes earlier in the summer when I was trying to figure out how to use clicker training to teach Beau to target, even though he was afraid of the sound of the clicker. This also relates to Alexandra Kurland’s concept of loopy training. If you found this post interesting, check out this post, which explains the concept of loopy training.