I take my 4-year-old boxer, Bella, on a walk every day. I keep her on a short leash, and she is usually good about walking at my side. However, sometimes if she is nervous or overly alert, she will begin to pull. Usually, I will react with a quick tug on the leash to remind her that she's not supposed to be pulling. If she continues to pull, I say "No. Sit." Then I wait for her sit before we continue.
My method obviously doesn't work, because she's four years old and I still have not been able to break her of it... I also worry about making her sit too often, because she has joint problems. But she KNOWS where she is supposed to walk, she is just very strong-willed.
So, when a dog KNOWS where she is supposed to be and is happy being there most of the time, but willfully chooses to disobey commands and disregard rewards, what should I do?
At Wit's (and Leash's) End
Dear At Wit's (and Leash's) End,
First off, I'd change your focus from teaching her not to pull, which isn't working, to teaching her where you want her by building and maintaining a strong connection with her. That shift in mindset alone will help a great deal.
Second, when I see noncompliance like you describe, I try not to frame it in "she knows but she chooses not to" terms. That puts the onus on the dog rather than on the human, and I'm not sure it's all that accurate. I'd say "she knows it when she's not distracted, but she either lacks the motivation, the self control, or both to do the behavior when she's distracted." That's more accurate and puts the onus squarely on the human to solve the issue, rather than punishing the dog for it. In Bella's case, it sounds like her leash behavior isn't really strong enough to handle the context of a high level of distractions.
What you're doing now is punishing the pulling with the intent of making it go away. A leash correction constitutes positive punishment in behavioral terminology, because you are adding (positive) something the dog doesn't like with the intent of reducing the undesired behavior (punishment). Second, stopping your motion is what's called a negative punishment, because you are taking away (negative) something the dog likes with the intent of reducing the undesired behavior (punishment).
While punishment can reduce undesired behavior, it's not going to do so as quickly and reliably as rewarding an alternative behavior, and punishment carries the potential for side effects. Plus, in a scenario like this, the dog might respond to the punishment in many situations but not end up with a behavior that's reliable enough to hold up to serious distractions, which is what you're describing. Even the gentler negative punishment of simply stopping the forward motion, while it does work with some dogs, can sometimes end up teaching the dog simply to run out to the end of the leash and choke herself for a while.
So instead of punishing the pulling, I'd change the dynamic entirely and teach her to connect with you, even when she's overstimulated. If you're walking along and you lose your connection with her, try walking backwards excitedly in a different direction so you can make eye contact, clap your hands, shuffle your feet, and talk to her. That makes you much more worthwhile to stay connected with with. And while I would not lure her with food in order to get the attention, I certainly would use food to reward her once you do have her attention (i.e., don't put it in her face, but rather keep it in your pocket or a closed hand until you get some attention, and then reward her with the food).
When you get the hang of it, the dog isn't pulling at all, but you're taking an active role in maintaining that connection with her, and then you can pivot and walk forwards together again with her by your side, delivering rewards to her right at your side where you want her (again, with the food not in her face until it's time to actually feed it to her). There's a lot more on this loose leash walking method here.
With practice around gradually increasing distractions, you should be changing direction and backing up less and less. However, if you do end up in a bad situation, you always have the option to try to back up and reconnect with your dog by using eye contact, your voice, and your motion, just so you can get away from whatever is causing the problem.
Full disclosure: reader questions are edited for clarity, concision, and anonymity. They are always based heavily on the real text of a question somebody has asked me and the real text of my response, but I mess around quite a bit with the text in order to make a good post out of it. Also, pretty much nobody actually signs their questions with a goofy Dear Abby style name, so those are made up by me.