Differential Reinforcement Schedules in Dog Training
What exactly are Differential Reinforcement Schedules? As mentioned in my last article on the four quadrants of dog training, positive reinforcement is what in my opinion works best when it comes to dog training. In that article, we saw a video of how positive reinforcement instills enthusiasm in the dog when compared to negative reinforcement. Today, I would like to tackle the variety of types of reinforcement at your disposal to train your dog.
The beauty of these training methods is that reinforcement drives behavior, and that you as a trainer for your dog, make all these great things happen. It's inevitable for the bond between owner and dog to strengthen as they work together as a team. When I start classes, many times dog and owner are just two separate entities who lead their own lives, not really much happening there. At the end of class, the biggest satisfaction is seeing the start of a bond. It's this bond that makes training through positive reinforcement so worthy, and what really makes me feel like I have done a good job in helping dog and owners. The following are different types of differential reinforcement schedules you can use when training your dog.
In differential reinforcement schedules you're often bringing training to the next level. You may have taught your dog to sit using positive reinforcement, but the training doesn't stop there. Initially, to help your dog, you may have given him a treat every time he sat. Now that your dog fully starts understanding what you want, you can shift into giving the treats every now and then, and this is when you start getting picky. You need to proof the behavior and make it more reliable. When you are using these methods, you are basically choosing which behaviors to reinforce and which not to. This means you will be providing positive reinforcement for the behaviors you like, (you give a treat so the behavior repeats) and you will be providing negative punishment for the behaviors you don't like (you don't give the treat so the behavior extinguishes).
This abides to Skinner's new term into Thorndike's Law of Effect which states that: "Behavior which is reinforced tends to be repeated (i.e. strengthened); behavior which is not reinforced tends to die out-or be extinguished (i.e. weakened). Let's take a looks at these differential reinforcement schedules.
Differential Reinforcement of Incompatible Behaviors (DRI)
This training method comes handy in force-free dog behavior modification. I use it a lot when dealing with cases of aggression. Basically, in this schedule, you will reinforce only behaviors that are incompatible with the unwanted behavior. What this means is that a dog who jumps, cannot jump up if he is sitting nicely with all four paws on the floor. In the case of a dog who bites, the dog cannot bite if he is asked to carry a toy in his mouth. In aggressive behavior, a dog cannot lunge if he is asked to heel nicely and is lavishly rewarded for that. In this schedule, we will therefore, find an incompatible behavior, train it systematically in a controlled setting and reward it lavishly, so the dog feels more compelled to perform the incompatible behavior and makes good choices.
Differential Reinforcement of Alternate Behaviors (DRA)
Sometimes, it's not that easy to find an incompatible behavior, in this case an alternate behavior will do. In this case, you will reinforce the wanted alternate behavior and ignore the unwanted one. This should result in Skinner's theory where "Behavior which is reinforced tends to be repeated (i.e. strengthened); behavior which is not reinforced tends to die out-or be extinguished." Of course, it's best to minimize the chances for the dog to rehearse the unwanted behavior. It's best therefore, to work with the dog under threshold. The alternate behavior in this case gives the dog something to focus on so to to hopefully not engage in the unwanted behavior.
Differential Reinforcement of Other Behaviors (DRO)
In this case, any other behaviors other than the unwanted behavior are reinforced. In other words, you have no specific target behavior in mind just as when you are free-shaping in clicker training. So say your dog jumps, you would reward any other behaviors other than the jumping behavior. This leaves you with quite an assortment! I find this technique helpful for cases where the dog needs that extra encouragement to get things started. It's useful for foster dogs or dogs who don't have a history of training and training is very new to them.
Differential Reinforcement of Excellent Behavior (DRE)
This is where you start getting picky when you are training. Basically, once your dog learns a new command well, you start reinforcing only those excellent behaviors that impress you the most. For instance, let's say your dog knows well how to sit, but does so quite slowly, in this case, you will start reinforcing only those sits that are a little faster than others. This brings your training to a whole new level and is a great method if you're into the precision training required to compete in obedience.
McLeod, S. A. (2007). B.F. Skinner | Operant Conditioning. Retrieved fromhttp://www.simplypsychology.org/operant-conditioning.html
The Canine Aggression Workbook; James O'Heare Dip.C.B