When your dog is being "bad", give more rewards, not corrections.
"Borrowed" from 4PawsUniversity in Sacramento's Facebook post:
TRAINING TIP TUESDAY: When your dog is being "bad," give more rewards, not more corrections. WAIT! Don't leave. Just hear me out. When our dogs begin to misbehave (a.k.a. act like dogs), they do so for one of two reasons: 1) stress, or 2) incomplete training. In both cases, a higher rate of reward for lower criteria will achieve better results than a louder, firmer voice. Here's one example: You've been working on having your dog stay on a mat while you open the front door. No problem. He can run to the mat on cue, lie down, and hold that position while you open your front door, walk in and out, and even ring your own doorbell. He's got it. Finally, a friend comes to visit. You give your dog his cue, he runs to his mat, you open the door and....WHAT??? He breaks position and runs to greet your friend. BAD DOG! He knows what he's supposed to do! Is he being stubborn? Are these stupid positive training methods not working? Of course not. You expected your dog to perform exactly the same way in a scenario you've never actually practiced. In other words, he wasn't ready to make that leap from empty room to birthday party. So, you're kicking yourself for your bad setup and thinking of all of the things you should have done, but what do you do RIGHT NOW? Ask your friend to become neutral - no looking, touching, or talking to your dog, just standing still and being as boring as possible. Now, patiently, work to regain your dog's attention, give him the cue for the mat, then turn yourself into a highly-caffeinated Pez dispenser. What does that mean? It means that, at first, you are either going to be clicking or feeding your dog, with no time in between for your dog to think about anything other than anticipating the delivery of the reward or chewing on the reward. You are now rewarding small moments of good behavior. It's not perfect behavior. It's not your final behavior, but it's the behavior you can get in that moment, which is holding position on the mat for one second (or one nanosecond) at a time while your dog's favorite Auntie Colleen is standing in your living room. Little by little, you can start lumping those seconds together - 2 seconds, click/treat....3 seconds, click/treat...etc. Until finally, your dog is back to their normal, sane self, fully focused on you and the exercise starts to feel as easy as it did when it was just the two of you. It won't take long if you've done your practice - your dog will start to remember pretty quickly. I think it was Ian Dunbar who said something like, "Your dog can do 100 things wrong, but only one thing right." This is why increasing rewards for smaller bits of good behavior yields faster results than corrections - rewards give clearer information. The same applies to a dog whose training falls apart because of fear or anxiety, but first we make sure we remove the dog to a distance where they are more comfortable, keeping them below threshold before resuming training (BTW thresholds also apply to overly-friendly dogs, too). When we learned to read, we didn't start with Hemingway. We started with the ABC's. By lowering your criteria and increasing your rate of reward, you are taking your dog back to the ABC's temporarily, so he can be successful. The more success he has for the behavior you want, the better that behavior will get. Don't forget - training isn't complete without practicing the 3D's: Distance, Duration, and Distraction. When one gets more difficult, making the others temporarily easier will speed the training process.