Positive Reinforcement 101
February 28, 2010 at 10:53 am
If you have been following this blog from last year then you would know that I have a pretty awesome little dog. You would also know that this Boogie has bitten people and became dog-aggressive after getting attacked three times last year. His people-biting history actually goes way back before we adopted him.
So last year we went through a hardcore Obedience training program which taught us crate-training, basic commands and corrections… but the program didn’t address our specific behavior problems. What we really needed was a “Behavior Modification Program” that focuses on uniquely Boogie issues.
I first encounted Sarah of Bridges Dog Training on the Functional Rewards Yahoo Group (see my earlier blog post)and was really impressed by her posts and Youtube videos. And what’s more, she lives near me!
So far we have had one consultation session and today — our first clicker-training session!
click to view larger
There is just way too much info to share on this blog without it turning into a novella so I’ll summarize with a few training tips and pieces of Positive Reinforcement wisdom that I have learned from Sarah. (I am paraphrasing here, and adding my own bits and pieces)
1. Aggression is a behavior.
Boogie isn’t an aggressive dog. He has aggressive behaviors. This is a really useful distinction. Sarah explained that aggression is a behavioral response to stress, it does not define the dog’s personality. We want to teach Boogie to replace aggressive behaviors (lunging, growling, biting) with friendly behaviors (turn his head away, sniff the ground, do a shake etc.) when he is stressed by something. These friendly behaviors are also known as “Calming Signals”.
Corrections are counter-productive because when we correct or punish a dog, we are adding stress to his already stressed-state and this raises cortisol levels and inhibits learning. If we are to encourage Boogie to do socially-acceptable behaviors, we have to minimize stress as much as possible, and learn to read and pay attention to dog body language, so that we can mark and reward him at the precise moments that he offers any friendly behaviors. With time and practice, he will respond differently to stress.
I am already seeing signs of improvement! :)
So a part of our homework is to study dog body language. This is not as easy as it sounds because Boogie is a very “stoic” dog (quote Sarah). He is not a wiggly dog, he rarely lets his tongue hang out (except when he is hot). He is a calm and serious dog….a Mr. Poker Face of boston terriers!
[dvds and books- thank you, Sarah!]
2. Positive Reinforcement isn’t about “bribing with treats”. We have to be careful about using a LURE — we could be reinforcing the wrong behavior.
A lure is like a treat used to illicit a behavior… like a bribe to get a dog to do something. Sarah gave us some excellent examples on how this can be a problem:
Say we call Boogie and he doesn’t come. So we get a treat and call him again, and he comes. What we have just done is reward him for not coming when called. We have just taught Boogie that if he doesn’t come when called, he gets a treat. The more often we do this, the less often Boogie will come when called… unless if there is a treat in our hand.
Similarly, if Boogie jumps up and we hold out a treat and say “Sit”, we are reinforcing the JUMP. Boogie’s learns that if he jumps up, he gets to sit for a treat.
The correct way is to mark (“click!”) at the precise moment that Boogie does the correct behavior and offer the treat-reward after the click, so that he knows the behavior that he is being rewarded for (and not the behavior preceding it).
To read: Fifteen tips for getting started with Clicker Training (the site requires free registration)
Today we did the “101 Things to Do With A Box” clicker-game, which is designed to encourage movement, thinking and trying new things. It is said that dogs that have been punished or abused in the past are slower to respond because they are sort of repressed. They don’t know how to “think”. This is a game to loosen Boogie up.
The idea is to click & treat for ANY behavior that Boogie does with the box, beginning with a simple head-turn towards the box.
Warning: The videos are actually very boring to watch … you had to be HERE:) I also apologize for the terrible lighting.
We did two sessions of this game. Boogie got as far as LOOKING at the box. Most of the time he stared at my hand — waiting… waiting… waiting….
It’s only our first day so hey, give us some time! :)
I look forward to the day that Boogie jumps inside the box or picks up the box and moves it around. Ha – maybe 20 sessions later!
Other homework: “Boogie!”(Whiplash head turn) cue and “Here” cue in different locations around the apartment, with distractions added. We need to strengthen his recall before we take the lessons outside.
This is all so interesting, and it is great to see Boogie perkier and more responsive.
One question I had to ask : When do we stop using the clicker (and treats)? Do we always have to use it?
Answer: When Boogie has learned a cue and he responds to it quickly 100% of the time, then we no longer have to use the clicker. The clicker is a tool that is primarily for training sessions when we need to make it clear to Boogie which behaviors we are reinforcing. The more often we reinforce, the stronger the behavior.
However, Sarah says that the recall cue (“Boogie!” or “Here!”) should ALWAYS be reinforced with something good, like a treat, and never anything bad so that he will reliably come to us when called in a case of emergency.
Entry filed under: Articles, links, Play, Reads, Training. Tags: .