Solving Behavior Problems Like a Dog Trainer (Part II)
Last week, we discussed the hypothetical case of Sally and Harland (her jumping dog). This story is actually only hypothetical in that I've never had a client named Sally who owned a dog named Harland, but their situation is identical to one I've seen dozens of times. If you haven't yet, take a moment to read the last post so that you're familiar with their story (we'll wait).
We finished by identifying opportunities which elicited Harland's jumping behaviors. These are, essentially cues (or "antecedents"), much like the word "sit" is often a cue for a dog to place his rear on the floor. Trainers know that behavior is controlled by its consequences. If a behavior is followed by a consequence which the dog deems favorable, he is more likely to do it in the future (reinforcement). If a behavior is followed by consequences which the dog finds unpleasant or undesirable, he is less likely to engage in that behavior in the future (punishment). We call this the "ABC" contingency: Antecedent (cue - what elicits the behavior) --> Behavior (what the dog does) --> Consequence (what happens to the dog?) Control the consequence, and you control the behavior.
If you read the other entry, you'll find that the consequences for Harland are ambiguous at best, and downright confusing at worst! Dogs thrive on structure and consistency, both are critical in establishing reliable control of behavior. Sometimes Harland gets things he likes (attention - not only when the owners encourage him, but also when they think they are "correcting" him!). If Harland is jumping for attention (and most dogs do), even pushing him away or verbally reprimanding him qualify - the opposite of attention is not "other" (scolding) attention, the opposite of attention is ignoring or reward removal. Sometimes, Harland gets pulled back on the leash, and he probably doesn't like this, but it's not functioning as a true punisher because it doesn't stop the behavior (probably because it is not a strong enough aversive and because the behavior has been placed on a variable schedule of reinforcement).
This ambiguous blend of consequences (sometimes good stuff - attention happens, sometimes bad stuff - leash jerks happen) has created one confused dog. Harland has learned that most of the time jumping earns him attention, that the owners can only control his behavior when his leash is on, and is likely in the process of building a punishment callus to leash corrections (because the "aversive" was not strong enough to stop the behavior when applied, harder and harder "corrections" are required to get his attention).
What Harland needs is clear-cut consequences. We need to split this into rewarding him for the appropriate behavior (which the owners identified as standing ("four on the floor"), sitting, or lying down and also, removing reinforcement (attention) for the unwanted behavior. Let's ditch the collar corrections altogether, for poor Harland's sake!
Rewarding the Correct Behavior
- The behaviors must be taught and put on cue before we can begin practice. For the purposes of this exercise, default behaviors lacking stimulus control are fine - while a person hoping to trial may not want a dog to offer an uncued sit, Sally would be thrilled if Harland automatically greeted a new person or favorite friend with a default sit behavior.
- The behaviors should be proofed for a variety of distractions systematically before introducing big-time distractions like new friends entering the house or kids playing in the yard.
- The reward level should correlate with the importance of the behavior and the difficulty level of the environment - if this behavior is important to Sally, she should "pay" her dog like it is through pairing these behaviors with high level reinforcement.
- Once the behavior is on cue, Sally and Harland can begin practicing greeting rituals. Sally will need to line up some volunteers, the more the better. Sally will watch Harland's feet as the visitors approach. She will continue clicking and treating him as long as all four paws are on the floor. But what if, during this process, Harland jumps at the approaching visitor?
Find out next time on the Dogster Guide to Behavior & Training!
P.S. Last time on the blog, there were a couple of comments on the recommendation of stepping on the leash to manage behavior. Thanks, dogsters, for reminding me to introduce the caveat - this technique is not recommended for use with fearful or reactive dogs but can be useful with friendly dogs who are over-exuberant greeters. If you choose to use this technique, it is recommended that you practice a lot in gradually increasing distracting environments, and that you pair the tethering experience with copious and appropriate reinforcement to teach your dog it is a safe, comfortable, and enjoyable experience well in advance of using it in your greeting rituals.