Pablo needed to learn to not run out the front door. We did not wait until he ran out and then scold or punish him when he did. Instead, he was taught that the door chime predicted a treat on the mat, and so when he heard the door chime, he went to the mat.
Let’s say you’ve just arrived for your first day at a new job. You walk into your new office, where you find a computer on a desk. “Oh, I know what to do with that,” you think, and push the space bar to wake up the screen.
Your new boss strides into the office, looking angry. “No!” he shouts.
What exactly didn’t he like? Should you never press the space bar again? Was this just the wrong time to press it? Does he not want you working on the computer at all?
How do you feel about your new boss?
And do you know what key you should hit instead, or whether you should touch the keyboard at all?
Dogs encounter this scenario pretty much every day. They do things that come naturally to dogs, like chewing on stuff, eating stuff, peeing on stuff, barking at stuff, or chasing stuff. And when we catch them in the act, as often as not, our response is to yell “no!”
While this sometimes interrupts the behavior, it probably won't prevent it from recurring again in the future unless the dog found it so aversive that he never finds it worthwhile to risk it again. Often that will not happen unless you follow "no" with something else unpleasant--for instance, snatching away items or yanking the leash. And if the dog finds the interruption that scary, but you haven't taught him what to do instead, he may begin to hide the behavior from you in dismaying ways–say, only peeing behind the couch, only chewing things if nobody is looking, or gobbling down forbidden items so fast that by the time you notice he has them, they're gone.
Behavior science--specifically what's known as "applied" research, from the helping arm of the field--tells us that the best way to reduce unwanted behavior is usually to replace it, not suppress it. In other words, once you’ve figured out what’s reinforcing your dog’s behavior, you’ll want to teach him a more acceptable way to get it—or if that’s inadvisable or impossible, something he likes equally well.
Note: I say reduce unwanted behavior, not eliminate, because once an animal forms a habit, it can’t be magically erased. It can, however, be overridden by a new habit. (For a fascinating long-form layman's read on the subject, pick up The Power of Habit by New York Times science writer Charles Duhigg. Here’s an excerpt describing how he tackled his own habit of buying a cookie every afternoon.)
The main components of a plan to change behavior are:
1) PREVENT THE BEHAVIOR. The old adage “practice makes perfect” also applies to unwanted behavior. Another word for “perfect” behavior is habit—a routine that the brain doesn’t even have to actively think about any more.
Prevent habits before they form. If it’s too late for that, start by removing the dog from the circumstances that trigger those routines. Gates, crates, exercise pens, leashes, and visual barriers are tools you can use to prevent unwanted behavior and set a dog up to do something you like better, which can then repeatedly reinforce. One of my favorite recommendations for jumping is to temporarily use a gate or ex-pen to make an "airlock" that guests can step into without the dog being able to immediately make contact with them.
2) TEACH YOUR DOG WHAT TO DO INSTEAD. The ethical use of operant conditioning starts with the premise that all behavior serves a purpose for the behaver, and strives to discover and accommodate that purpose.
So step 2 in problem solving with positive reinforcement is: using positive reinforcement, teach an alternative route to reinforcement that’s incompatible with the behavior you don’t like. The good news is that you may already be halfway there: the uses for simple behaviors like sitting, hand targeting, settling on a mat, or even just simply standing and looking at you are endless. The last is often my preference.
Once your dog is performing the new behavior quickly and enthusiastically, you can teach him that he can use it to get the thing that he is doing the problem behavior to get, and incrementally introduce it into the problem context. You can even do it in such a way that those same circumstances begin to trigger the new behavior instead of the old one: e.g., the approach of a person becomes another cue to sit, or the sight of a squirrel becomes a cue to check in with the handler. The tougher the conditions, the more important the incremental part is.
The video above is just one example of replacing an unwanted behavior with a desirable one. Pablo, a 3-month-old doodle, had started to run into the entryway when his owners went out to answer the door. The outer door opened to the street, so we wanted him to stay in the apartment. Rather than punish him for running out, we taught him to go to a mat, with the door "ding" as the cue.
3) DON’T REINFORCE THE UNWANTED BEHAVIOR. If, despite your best efforts at management, the behavior happens, and you can control whether it “works” or not, try not to have it work, or to have it work less than the unwanted behavior. (If you don't control the reinforcer, just get out of the situation and go back to step 1.) For instance, if your dog jumps on you and you have usually petted or pushed him around, don’t pet or push him around. You might not be able to remove all the reinforcement he gets--like smells, or physical contact.
I originally had this as step 2 when I wrote this blog in the mid-2010s. But there are perils to ignoring, so in 2022, I have moved it to step 3 to emphasize that you should teach the skills your needs to get attention before you stop reinforcing jumping. If the desired behavior isn't in the dog's repertoire, or has never been used to get this particular outcome, it isn't going magically appear when you ignore unwanted alternatives. Good trainers can help you navigate these types of nuances, and guide you away from the notion that "throwing cookies at the problem" has now failed and you must use punishment.