Positive
Reinforcement: Using Food to Train Your Dog
By Marty Guerra
What Positive Reinforcement Is:
Positive reinforcement is anything your dog values and will work to earn.
Dogs value food, as it is necessary for their survival. Dogs also value
other things such as praise, toys, and attention. To positively reinforce
a behavior you like, you simply give your dog something he/she values
immediately after he/she has performed the behavior. By positively reinforcing
the behavior you are increasing the likelihood of the behavior occurring
again. For example, I know my dog values hotdogs and will work to earn
them. When I ask her to sit and she then sits, I immediately give her
a piece of hotdog. Therefore, she is more likely to sit again the next
time I ask. Why? Because her "sitting" behavior could earn more
reinforcement (hotdog) which is what she wants. This creates reliable
behavior. Additionally, by using positive reinforcement my dog learns
to think and offer behaviors of her own accord, without me forcing her
to do it. She must figure out what behavior will earn hotdogs and must
offer the behavior whenever asked in order to earn the hotdog. If she
doesn't do what I ask, she doesn't get what she wants. My goal from the
outset is to get my dog (or whatever animal I'm working with) to perform
as much behavior for the least amount of reinforcement (treat) as soon
as possible. So, as she learns each behavior with more reliability, we
wean the hotdog away. We put the behavior under a variable schedule of
reinforcement. This serves to increase the reliability of her responses
because she never knows when she'll get the treat; it could happen at
any response, so she keeps working in hopes that "the next one"
will pay off. I'll fill in the gaps with petting and/or verbal praise.
If you've ever spent any time in front of a slot machine or have purchased
lottery tickets before, you've experienced variable reinforcement; you
don't get a pay off every time. It's little pay offs every now and then
that keep you going and thinking that "the next one" will be
the big pay off.
What Positive Reinforcement Is Not:
Positive reinforcement is not the same as physically forcing or compelling your dog to do something and then giving your dog a treat or praise. For example, the trainer who pulls up on the dog's collar while pushing on the dog's hips until the dog is physically forced into position, and then says "good!" or gives the dog a treat, is not using positive reinforcement. When compelling or forcing or "correcting" a dog to get it to perform a behavior, the dog is working to avoid punishment (as opposed to working to earn reinforcement). When using punishment to get a dog to perform a behavior, it is punishment that produces the behavior. The dog has not learned to perform the behavior on his own. In the absence of punishment the dog will often not perform the behavior. For example, a dog trained using punishment often will not obey a command when out of reach or when the "training collar" (choke chain, pinch collar, shock collar) is removed. Alsol, a dog trained using punishment often learns to avoid the trainer because they are the source of unpleasantness.
Dogs Work Either to Earn Reinforcement or to Avoid Punishment:
Dogs work either to earn reinforcement or to avoid punishment; dogs do what works for THEM. Some people prefer to train their dog using punishment rather than reinforcement. A common explanation is they do not want the dog "working for food," they want the dog "working for them" or to "please" them. Paradoxically, these people use punishment to train their dog. They fail to realize that simply not using food in training does not mean the dog is "working for them," it means the dog is working to avoid punishment from the trainer. Think about it, trainers that do not use food for training, do use choke chains, pinch collars, shock collars, throw chains, bopper-sticks, yelling at the dog, hitting the dog, "corrections," squirting Bitter Apple or other bad tasting things in the dog's mouth, setting up booby traps, and/or a myriad of other unpleasant things to train the dog. Not a one says, "Take off the collar and leash and just tell your dog what you want–he wants to please you and will do it just because you ask." Just because the source of punishment comes from the trainer does not mean that the dog is working "for" the trainer; the dog works to avoid the punishment that comes from the trainer. Punishment comes from you and reinforcement comes from you; in neither case is the dog working "just for you." You might as well use the tool that works, is easy to use, and does the least harm.
How to Train Using Positive Reinforcement:
Positive reinforcement trainers use food that the dog really likes to establish behaviors. Training with food does not result in harm to the dog or the dog's avoidance of the trainer. Food is positively reinforcing and therefore, is a valuable training tool. Like any tool, there is a right and wrong way to use it. To properly use food, the trainer must establish the behavior to the point that the dog understands what behavior it must perform on its own in order to earn reinforcement, and then, the trainer must quickly begin weaning away the food as discussed above. Because food is so successful in getting behavior, many trainers become too dependent on it and don't reduce and eliminate it soon enough. Like the dog that only responds to commands when the choke chain is on, dogs can learn to only respond when food is present. This may be referred to as "the dog only listens when I have food" syndrome. The trainer should also avoid using the food as a "bribe." Like the dog that only does what you want when you show him the rolled up newspaper, the trainer should avoid first showing the dog the treat and then asking for the behavior ("I'll give you this treat if you sit.") Instead, the trainer should avoid showing the dog that they have a treat and only after the dog performs the requested behavior should they give the dog the treat. The dog's behavior should earn the treat; the treat should not earn the dog's behavior. The trainer should also strive for good timing when delivering treats to the dog. For the reinforcement to be associated with the intended behavior, the food must be delivered within 1 second of the occurrence of the behavior. If you wait too long to deliver reinforcement, the reinforcement may be associated with a behavior other than the one you intended. (To ensure accurate timing in delivering positive reinforcement, we use a "clicker.") Most positive reinforcement trainers eventually use petting and praise as reinforcement, however, initially when teaching a new behavior food is used because it is quick and effective. Dogs receive non-contingent petting and praise from us all the time. It may be hard for the dog to discriminate when the petting or praise is intended as reinforcement for a behavior and when it is simply because the owner feels like it. The well-timed delivery of a piece of hotdog will be more effective and clear as a reinforcer in your initial training efforts. Some people claim their dog is not "food motivated." If their dog eats it's meal, the dog is food motivated. They may simply have not found something that the dog really values or the dog may be averse to taking food from their hands. Raid the cupboard until you find what the dog really likes and toss a piece across the room to the dog. Other people claim their dog is "food crazy." I suggest that these dogs are the best to train using positive reinforcement. They will work their tails off for food. To reduce some of their zeal the dog should be fed part of its meal before training to reduce their hunger, and training treats should be something the dog only moderately likes. Bits of apple, dog food, bread or other unexciting treats will help the dog focus. When using food we recommend using food low in fat and is healthy because during the initial stages of training your dog will getting a lot of treats. Commercial brands of treats are filled with preservatives and may not be the healthiest thing for your pet. Use dog food, meat (chicken, liver, etc.) cheese, tortellini, whatever the dog will like. I always caution to be aware of how much treats you give for training and cut back on the dog's regular feedings accordingly. Use bite-size pieces. Use moist treats that can be quickly consumed. Dry treats such as biscuits will crumble. You'll lose time while the dog vacuums the floor of crumbs. Also, some dogs may experience temporary stomach and/or intestinal upset after receiving food they are not accustomed to. In addition, if your dog has a sensitive system or has food allergies, consult with your veterinarian. Food is a valuable training tool. Its use does not result in harm to the dog and helps build the dogs bond with you. You're going to feed your dog anyway. Why not make your dog "work" for its food?
Back
to Articles and Tips
e-mail: mguerra@dogmanners.com
Phone: 342-2494
© Copyright Good Dog 1999