In dog training with positive reinforcement, especially when you are teaching a new behavior, food is great to use in adding value to that behavior as it is quick to deliver which means you can get many more practice repetitions in, in a short period of time. If you are teaching your dog using shaping, you can move through more approximations toward the final behavior goal more quickly too.
However, people sometimes tell me dog training with food does not work. Often when I watch clients train, I see things that could help them to be more effective teachers.
Here are a few tips for making your food reinforcers more effective in your lesson.
Use high value treats. Only the learner gets to decide what is of high value to him. Just because you think your dog really likes that store-bought dog treat or you think your dog will work for just kibble, is not necessarily true depending on other factors such as the environment where you are training (are there other competing reinforcers?), the hunger level of your dog, the difficulty of the lesson. Remember, when you are teaching with positive reinforcement, you are teaching by building value for a given behavior by associating it with a high value consequence.
Use a high rate of reinforcement. In teaching a new behavior (or changing criteria of a behavior), your pet will learn faster with more frequent feedback from you that he is on the right path toward the goal. This means, splitting the goal into small steps or approximations that can be marked (with a click or verbal marker) and then reinforced (with the food treat) quickly. Finding behaviors to click 10 to 20 times in one to two minutes is a lot more feedback than simply waiting out the learner for a long time to try and figure out what you are looking for. This is why, when I am teaching something like ‘sit’, once the dog has had some history of reinforcement for sitting, I often will get into a game of clicking and immediately tossing a treat to get the learner out of position so that he will need to go back into a sit to earn another click/treat. Then I may do this over and over and over to get a deeper reinforcement history in a short window of time.
In shaping a behavior (teaching it by teaching small approximations toward the final behavior), I try to look for small steps to reinforce. As an example, in teaching a dog to go to a mat, I may click for just looking at the mat, then any small movement in the direction of the mat, then a step toward the mat, then two steps toward, the mat, then looking at the mat at a closer distance, then touching a paw to the mat, then either touching more of that paw to the mat or two paws to the mat, etc. There are times where I may simply toss a treat away too, just to reset the learner if needed, to keep him in the game.
If your rate of reinforcement is too slow, you risk having your learner check out of the lesson to find greater reinforcement elsewhere. Or your learner could show stress and avoidance behaviors. Think about yourself trying to learn something complicated. It is pretty tough to do without much feedback as to whether or not you are moving in the right direction.
Practice your timing for delivering reinforcement. It takes a lot of practice to be able to accurately mark your pet’s behavior at the very specific moment that you intend. Ideally, for your marker to be effective, it will come just as the learner is doing the behavior, and your treat delivery hand DOES NOT move toward the food until AFTER you mark the behavior. Additionally, it is more effective if your treats are delivered quickly after your marker. If you are fiddling to get your food out of your bag or pocket, you may want to come up with other plan for holding the food.
Learn to understand your dog’s communication and listen. If your dog is backing up from you when you lean over to deliver a treat, your dog is telling you it is uncomfortable. If you are outside and your dog just sniffs your otherwise super high value food and then looks away, that is feedback too, just as if your dog begins doing other stress related behaviors such as a lot of sniffing, avoiding eye contact, becoming hyper vigilant about the environment, yawning, laying down or rolling over as examples. On the other hand, if your dog has loose body muscles and is actively looking at you to keep the lesson going, that is a sign that you are doing a great job with all of the above. That is our goal!
Hold lessons when your dog is hungry. While I am by no means advocating for holding off on food, a dog that is not satiated from a full meal will tend to value eating more in that moment. This is why whenever possible, I time my walks with Dawson to occur around his meal time because there is usually some training involved with our outings. You can make your high value food (and thus the behavior you want to teach, since it will be associated with that reinforcer) even more valued by using it only for certain training.
Keep food available throughout the day for teachable moments. If your dog only pays attention to you when you walk to the cookie jar or take out your treat pouch, your dog has learned those are times when doing what you ask will pay off but other times. Once your dog has learned behaviors, you can keep jars of treats in different rooms without having the food on you. Then when you ask for a behavior (or capture your dog doing something you like), you can mark the behavior and go to the jar for a cookie. (You can also use environmental reinforcers but that is not the focus of this post.) If you wear your treat pouch throughout the day, your dog will get used to seeing it without it signaling the start of official training.
Be very careful with how you pair food when it comes to teaching your dog to do or feel something to which it already has established negative association. Through classical conditioning, remember it is what happens AFTER something that affects how the learner feels about what happens before. If you bring out that piece of steak to lure your dog into a crate that has a negative history, your dog may come to learn that the steak coming out is a predictor of the evil crate. If, on the other hand, your dog goes into the crate and THEN gets a piece of steak, the crate can become a predictor for a nice, tasty snack.
