As a dog trainer, the questions I am asked the most start with “How do I stop my dog from ________?” Fill in the blank. How do I stop my dog from chasing the cat? Pulling on the leash? Jumping to greet people? Charging at the door when the doorbell rings? The list goes on.
My answer is the same for all of these: “It’s hard to teach a dog to NOT do something. So, what do you want your dog to do instead?” (Yes, I know that I split an infinitive, but I think it conveys the message better.)
So, why do I say this?
What Will You Reinforce?
When training dogs, I rely heavily on positive reinforcement methods. Recall that to reinforce a behavior is to increase the likelihood that a dog repeats that behavior. Positive reinforcement means that we are adding something that will increase the likelihood of a behavior. I ask my dog to sit. She sits. I reward her with a treat. She likes the treat so she is likely to sit again when I ask her.
Now we are in a somewhat different situation. The client has asked me how to stop the dog from doing something. Under operant conditioning, to decrease the likelihood that a dog repeats a behavior is to punish. Also, we can use either positive (add something) punishment or negative (take something away) punishment. These are easier to understand with examples.
There’s More Than One Way to Cook an Egg
Suppose you want your dog to stop jumping to greet you when you get home from work. Let’s consider some different approaches.
Positive punishment: When you arrive home and your dog jumps to greet, you knee him in the chest. Presumably the knee to the chest is uncomfortable (or even painful) enough to decrease the likelihood that your dog jumps to greet you next time.
Negative punishment: When you arrive home and your dog jumps to greet, you turn away from your dog so that he doesn’t get your attention. Your dog is probably jumping on you to get attention. Removal of your attention is unpleasant enough so that he’s less likely to jump to greet next time.
Positive reinforcement: When you arrive home, you ask your dog to sit before he jumps. When he sits, you give him attention and possibly a treat. Because your dog got the attention he wanted by sitting, he is more likely to sit to greet the next time.
Will all three of these approaches work? Probably. If the first two, which focus on stopping a behavior, will work, why do I say that it’s hard to teach a dog to NOT do something? Because I always strive to use the least aversive method available, and that means using a positive reinforcement approach.
But Some Ways Are Better
Why strive to use positive reinforcement methods? Well, first, for exactly the reason that I’m writing this blog post. With positive reinforcement, you can clearly communicate with your dog what you want him to do, rather than what not to do. It’s easy for both you and your dog to get frustrated if you’re focusing on what not to do. He jumps to greet and gets a knee in the chest. So, he starts tugging at your jacket, and you swat him away. He barks and runs circles around you, and you scream out of frustration. Sadly, all of these things result in attention for your dog, even if not the attention he wants, so these behaviors may continue. Asking for that sit, however, is a clear message of what to do to get attention.
Second, positive reinforcement is associated with the lowest incidence of aggression, attention-seeking, avoidance, and fear in dogs [*], or learners more generally. It should go without saying that we all want our dogs to be content. Positive reinforcement methods simply have the last chance of backfiring on us. (More on this in an upcoming blog post.) Keep in mind that the dog determines what is reinforcing. In my example of positive reinforcement, giving your dog attention when he sits as a reward is only reinforcing if you give him attention that he enjoys.
Third, but certainly not last, with positive reinforcement my dog can choose to do the replacement behavior or not without the fear of adverse consequences. (See my post on consent.) Frankly, I like my dog to have the option to say “no.” Do I like it if my dog jumps on me if I ask for a sit? Of course not. Will this happen occasionally? Yup. And the consequence is that he doesn’t get the attention he wants. What happens if I’ve been using positive punishment and my dog jumps to greet? Knee to the chest. What if he does it again? A knee to the chest didn’t work the first time so now what? The consequences may need to be harsher to get your dog to stop. Then what? Frankly, I don’t like where this leads.
Let’s go back to that litany of behaviors that we want to stop. In each case, what replacement behavior could we choose to reinforce?
|Behavior we want to stop||Possible replacement behaviors|
|Chase the cat||Initial tug with owner|
|Pull on leash while walking||Walk nicely at owner’s hip|
|Jump to greet people||Sit|
|Charge at door when doorbell rings||Go to special place/bed/mat/kennel|
Whatever you choose, the replacement behavior should be one that you have worked on extensively already.
Also, you need to consider what reward you will use to reinforce the replacement behavior. It some cases, a tasty treat will do the trick, especially if your dog is food-driven. However, food may not be the most reinforcing reward in all cases. For example, when your dog sits instead of jumping to greet, you can and should reward your dog with the attention he seeks. When your dogs is walking on a slack leash instead of pulling, your dog will find the forward motion itself to be rewarding.
Need some help finding a replacement behavior for that annoying behavior that you want to stop, contact me!
[*] “[The] use of positive reinforcement alone was associated with the lowest mean scores (attention-seeking score 0.33; fear (avoidance) score 0.18; aggression score 0.1). The highest mean attention-seeking score (0.49) was found in dogs whose owners used a combination of positive reinforcement and negative reinforcement. The highest mean avoidance score (0.31) was found in dogs whose owners used a combination of all categories of training method. Owners using a combination of positive reinforcement and positive punishment had dogs with the highest mean aggression score (0.27).” Emily J. Blackwell, Caroline Twells, Anne Seawright, Rachel A. Casey, The relationship between training methods and the occurrence of behavior problems, as reported by owners, in a population of domestic dogs, Journal of Veterinary Behavior: Clinical Applications and Research, Volume 3, Issue 5, September–October 2008, Pages 207-217, ISSN 1558-7878, http://dx.doi.org/10.1016/j.jveb.2007.10.008.