How Dogs Learn: by Vanessa Lee
Learning Theory, Teaching New Behaviors, Putting Behaviors on Cue, & Proofing
Positive Reinforcement - add something the dog values to increase a behavior
Example: Dog sits, you reward. Dog is more likely to sit in the future.
Negative Reinforcement - take away something the dog finds aversive to increase a behavior
Example: You tighten your dog's collar only releasing the pressure once he sits. Dog is more likely to sit in the future.
Positive Punishment - add something the dog finds aversive to decrease a behavior
Example: Dog jumps up on you, you squirt him with a blast of water. Dog is less likely to jump up on you in the future.
Negative Punishment - take away something the dog values to decrease a behavior
Example: Puppy nips during play, you end the play session. Puppy is less likely to nip during play sessions with you in the future.
Extinction - extinguish a behavior by removing the reward that was maintaining it
Example: Dog barks at the dinner table and has previously been rewarded with tidbits from the table. Dog barks at the dinner table, owner ignores and no longer offers tidbits. Dog eventually stops barking at dinner table.
Reward Markers (RM)
A reward marker is a unique word or sound that is paired with rewards. It is used to mark a correct behavior the instant it occurs. It works as a bridge between the behavior and the reward.
Examples: "Yes!", "Right", the sound of a Clicker
To create a reward marker simply pair your chosen marker with a reward such treats. If you're using a Clicker, you'll click/treat for no behavior in particular about 20 times in a row. If your using a word such as "Yes!" you'll say "Yes!"/treat about 20 times in a row. The dog should begin to associate the reward marker with the reward.
No Reward Markers (NRM)
A no reward marker is a unique word used to communicate to the dog that the behavior he just offered was not what you wanted and to try again.
Examples: "Ah Ah", "Too Bad", "Try Again"
To create a no reward marker you can take a treat enclosed in your hand and offer your closed fist to the dog. When the dog paws, licks, or tries to chew on your hand say "Too bad". Wait for the dog to stop trying to get the treat and then reward him. Repeat several times.
Release Words (RW)
A release word is a unique word used to mark the end of a behavior.
Examples: "Release", "Free", "All Done"
To create a release word give your dog a know cue. Once he complies reward him, wait one second and reward him again while he is still in position, wait one second and release him with "All done" and then walk away from him.
If the dog breaks position before being released, give him the NRM and get him back into position, wait one second and then release him.
Teaching New Behaviors & Putting them on Cue
When teaching a new behavior its a good idea to get the behavior reliably on visual cue (hand signal) before adding a verbal cue.
Food Lure - Food Reward (Day 1)
Start out by luring the dog into position using a treat wrapped tightly in your fist. The treat should be potent enough to entice your dog to follow your fist with his nose as if it were a magnet. The second your dog moves into position give your RM, open your fist and allow him to take the treat, give your RW. Repeat several times in three 5 minute training sessions throughout the day.
Hand Lure - Food Reward (Day 2)
Begin with Food lure - Food Reward as described above. Do this 2-3 times during your first 5 minute training session but then remove the food from your luring hand and place it in your opposite hand behind your back. Lure the dog into position with your closed fist. The second your dog moves into position give your RM, bring out the hand that was behind your back and offer him the treat from it, give your RW. Repeat several times in 5 minute training sessions throughout the day.
Hand Lure becomes Hand signal - Food Reward (Day 3)
At this point your closed fist will become your hand signal and you can modify it a bit if you wish. Continue practicing as described above using your modified hand signal. Repeat several times in 5 minute training sessions throughout the day.
Verbal Cue - Hand signal - Food Reward (Day 4)
By now your dog should be reliably responding to your hand signal and you can now introduce the verbal cue of your choice. To begin, state your verbal cue immediately followed by your hand signal. Repeat several times in 5 minute training session throughout the day.
Verbal Cue - Food Reward (Day 5)
Continue as described above but begin to wait a second or two after saying the verbal cue before giving the hand signal so that your dog has the opportunity to respond to your verbal cue. Repeat in several 5 minute training sessions throughout the day.
Verbal Cue - Variable Rewards (Day 6)
By now your dog should be reliably responding to the verbal cue. At this time you should begin to offer a variety of rewards. Rewards can be anything the dog likes or needs. Practice throughout the day using every opportunity available.
Verbal Cue - Intermittent Variable Rewards (Day 7)
Begin to offer rewards intermittently. The dog will now only be rewarded sometimes. Be sure you are random so the dog is not able to figure out exactly when he will be rewarded.
Can you repeat that? sit, Sit, SIT, SIT!!
Dogs have better hearing than we do so there is no need to yell or repeat the cue over and over again. There are many reasons your dog may not respond to a cue but deafness is probably not one of them.
Here are a few reasons a dog might not respond:
He doesn't understand the cue yet
There are more distractions than he is ready to handle yet
Your training session is too long
Repeating the cue can cause it to lose it's meaning. The dog might learn that he should only respond after hearing the cue repeated 3 times. "Sit" now becomes "Sit, Sit, Sit". If you are sure the dog understands the cue, stop and take a moment to consider why he might not be responding.
Reward for position!
Be sure your dog is still in position when you offer the reward. If he moves from position get him back before you reward him.
Example: You say "sit", dog sits, as you are about to reward him he stands up. Lure him back into sitting position and reward him there. Otherwise you will be rewarding him for standing and he will be more likely to break the sit in the future.
Timing is everything!
In order for your dog to understand what you want, you must have good timing. If your timing is off you might end up rewarding a behavior you did not intend to, which will slow down your training progress.
Proofing - teaching your dog to generalize
Work on one of the following at a time.
Rather than always working with your dog directly in front of you, change your body position from time to time.
Examples: stand with your back turned to the dog, sit and stand to the right and left of your dog, go out of sight
Duration is the length of time the dog remains in position before being released. Randomly vary this time.
Examples: 2 secs, 3 secs, 2 secs, 5 secs, 3 secs, 6 secs, 4 secs, 5 secs, 7 secs
Practice in a wide variety of locations.
Examples: kitchen, dining room, living room, bathroom, backyard, porch, driveway, parking lot
Vary your distance starting only one tiny step away, gradually and randomly moving further away and closer again.
Examples: 1ft, 2 ft, 1ft, 3 ft, 4ft, 2 ft, 3 ft, 5ft
Practice with minimum distractions and then move on to medium distractions and then high distractions
Examples: Min.- Backyard, Med.- low traffic park, High - busy parking lot
Set your dog up to succeed!
If you find that your dog is failing, you've probably advanced too quickly. Take a few steps back and lower your criteria by making things easier so he CAN succeed. Advance more gradually this time.
Consistency & Persistence
To make your dog reliable you must be consistent AND persistent. Once you give the cue be sure you get the behavior. Don't allow the dog to constantly break position without being released. Don't change cues and markers, choose one and stick with it. To make things clear for your dog, a single cue should be designated to a single behavior. Your dog may get confused if you sometimes use "sit" and other times use "sit down" to cue him to sit.
Practice Makes Perfect!
Training is life long. Be sure to practice daily otherwise your dog might forget the things you've worked so hard to teach him.