Pat Miller© 1997
Approximately 2850 words
CLICK OR JERK?
Let The Dogs Decide
It’s a sunny summer day, and you’re taking a stroll around the neighborhood. You pass a park and notice a dog training class in progress. The instructor is standing in a meadow with a dozen dogs and owners circling her. Each of the dogs is wearing a shiny metal chain collar, and from time-to-time when a dog forges ahead or lags behind the owner jerks on the leash to bring the dog back to heel position, then pets and praises it. You hear an occasional “No!” issued in a commanding tone. The dogs appear well-behaved, and all of them are doing the same thing at precisely the same time.
You continue on, and reach another park where you see another training class. This is a more ragged-looking bunch, although also well-behaved. A half-dozen dogs are walking in different directions with their owners, turning, stopping and starting up again apparently at random. The dogs are wearing regular flat collars. Some of them also wear something around their noses that looks like a muzzle, but on closer inspection you realize it is more like a horse halter. There is no jerking, but there is much treat-tossing, talking and laughing, and you hear a lot of “Yes!” and an occasional, odd, clicking noise. Since you have been thinking about signing your dog up for training, you pause to ponder the obvious differences between the two groups.
These are both beginning dog training classes. They both can produce dogs that are well-trained. The main differences between the two are the methods used in training, and the philosophies and behavior theories behind those methods.
The Training Continuum
All dog training techniques fit somewhere on a long continuum, from seriously harsh and abusive punishment-based methods at one extreme, to pure positive reinforcement at the other. As is often the case with extremes, neither of these is likely to be very practical or effective, nor will you find many trainers who recommend them. Most trainers use a combination of techniques that place them somewhere between the two ends of the continuum. Which side of center they are on defines them as primarily compulsion-based trainers or primarily positive ones.
Within the dog training community the debate about methods is generally good-natured, albeit spirited. Hackles get raised when trainers, who tend to be an opinionated lot, disagree on the very best method to resolve a particular canine behavior challenge. But when the dust settles, good humor returns, and on at least one e-mail list trainers tease each other and mock themselves with self-deprecating labels like “Treat Slinging Weenies” (TSW’s) and “foodies.”
Why the huge diversity in training philosophies? Because there are, in fact, several different training approaches that can successfully teach a dog to do what we ask. We can teach our dog to sit by saying the word “sit,” jerking up on the collar and pushing down on her rump to force her to sit, then patting her on the head, verbally praising or giving her a cookie. We can lure her into the sit position by moving a treat over her head, then saying “Yes!-Sit” when she does. And we can choose to wait until she decides to sit on her own and then give her a verbal marker and treat reward for sitting.
In behavioral terms, training is known as “conditioning behavior.” We really aren’t teaching our dog any new behaviors when we train. She already knows how to sit, lie down, stay in one place, walk by our side, or come running to us from far away -- when she wants to. She just may not know how to do it (or may not choose to do it) when we ask her to. Training is conditioning (or teaching the dog) to reliably give us the behaviors we ask for, when we ask for them.
In classical conditioning, as first described by Pavlov, there is an association between a stimulus and a response, or behavior. (A stimulus is something that elicits a response.) This is the famous “ring a bell, the dog salivates,” experiment that most of us are familiar with. Classical conditioning can generally be used to teach only very simple behaviors.
Operant conditioning is most commonly used for training, because it can be used to teach complex behaviors and behavior “chains.” (A behavior chain is a series of behaviors strung together.) With operant conditioning there is an association between a behavior and its consequence. The dog does something, then something happens as a result of the dog’s behavior. There are four ways that this works.
1. Positive reinforcement: The dog’s behavior makes something good happen, so the desired behavior increases. For example -- when the dog walks next to you without pulling on the leash, she gets a treat (treat = good thing).
2. Positive punishment: The dog’s behavior makes something bad happen, so the undesirable behavior decreases. For example -- if the dog pulls on the leash, her neck gets jerked to bring her back to heel position (jerk on neck = bad thing).
3. Negative punishment: The dog’s behavior makes something good go away, so the undesirable behavior decreases. For example -- when the treat is used as a lure to keep the dog walking in heel position, she may jump up to get it. The treat is hidden until she stops jumping. Every time she jumps up the treat is hidden, until she stays on the ground as the treat is offered (treat = good thing; hidden = “goes away”).
4. Negative reinforcement: Dog’s behavior makes something bad go away, so the desired behavior increases. For example -- a no-pull harness puts pressure on the dog’s chest as long as the dog puts pressure on the leash. When the dog stops pulling, the pressure stops. (pressure = bad thing; no pulling = bad thing “goes away”).
Old-fashined, compulsion-based training works on the philosophy that we have to show the dog who is boss. She must do what we say, and quickly. If she doesn’t, we immediately correct her or she will learn that she can ignore our commands. The primary tool for compulsion trainers is positive punishment, often followed by a treat, a pat, and or verbal praise to keep up the dog’s enthusiasm for the training process. 20 years ago, traditional trainers abhorred the use of food treats as praise. This thinking has changed in the last decade, as more and more “foodies” have demonstrated the effectiveness of food as a training motivator.
Compulsion training works, as demonstrated by decades of well-behaved dogs. A skilled trainer uses the minimum amount of force necessary to get the job done. Proponents argue that the small amount of discomfort this may cause is worth the end result of a reliable, promptly responsive dog. It can be problematic, however, with very assertive or independent dogs who don’t take kindly to being pushed and pulled around and may decide to argue back. You must be prepared to use enough force to get your message across quickly, and be willing to escalate the level of force if necessary. Techniques like scruff shakes and alpha rolls only work if the trainer is strong enough to persevere if the dog fights back. Many owners and trainers are either unwilling or unable to use this kind of force with their dogs – thank goodness.
Timid, submissive or sensitive dogs may also not do well with positive punishment. Forceful corrections can cause them to melt into a puddle on the floor, and a slight miscalculation can cause irreparable damage to the owner’s or trainer’s relationship with the dog.
Yet another concern about compulsion training is the possible damage to a dog’s throat from a standard choke chain collar, which can exert tremendous pressure on a dog’s trachea. They are not recommended for puppies under the age of six months, yet it is more and more widely accepted that starting puppies in training classes at the age of 10 weeks is ideal, in order to take advantage of a pup’s important socialization and learning period. Prong collars reputedly distribute the pressure more evenly around the neck and are less likely to do damage, but many owners understandably shy away from using the medieval looking spikes on their tender baby puppies.
“Clicker trainers” is a slang term for trainers who use positive reinforcement as their first method of choice, combined with an audible reward signal to mark right behavior. These trainers operate on a different training philosophy from the compulsion trainers, preferring to get the dog to offer the desired behavior voluntarily, then mark and reward it when it does. (The marker signal, or “bridge,” can be the Click! of the clicker, a whistle, some other mechanical sound, or a word. “Yes!” is frequently used to mark a right behavior.) Since all living creatures tend to repeat behaviors that are rewarding, behaviors that are repeatedly marked and rewarded by a dog’s owner get offered more and more frequently. Behaviors that are ignored (not rewarded) tend to go away, or “extinguish.”
Take, for example, the puppy who wants to jump up on everyone. Dogs greet each other face-to-face, so it is natural for our dogs to want to greet our faces. Plus, when they are cute little puppies we pick them up and cuddle them in our arms, thereby rewarding them for being “up.” Small wonder that so many dogs jump on people!!
Many of the suggested compulsion approaches to correcting jumping behavior actually reward the very behavior we are trying to extinguish. When the dog jumps up, she touches us. That’s a reward. We look at her. Eye contact is a reward. We speak to her to tell her to get off. We are paying attention to her -- that’s a reward! We reach down to push her away. We touched her -- another reward!! For some rowdy dogs, even the time-honored “knee her in the chest” is an invitation to start a rousing game of body-slam.
The positive reinforcement approach relies on the principal that behaviors that are ignored will extinguish. But how do you ignore an enthusiastic, obnoxious canine who is leaping up to greet you nose-to-nose, inflicting multiple bruises and lacerations in the process? Just standing still doesn’t work; she gets all kinds of self-rewards by jumping all over you. Instead, we turn our back on the dog and step away. As the dog tries to come around to face us, we do it again. Turn away and step away. Turn away and step away. Over and over. Sooner or later (and with most dogs this happens much sooner than you would imagine) the dog gets frustrated and confused, and sits down to puzzle out your bizarre behavior. Bingo! Now you turn toward her, tell her “Yes!” and feed her the treat from the stash you keep in your pockets in anticipation of opportunities just like this. You can also pet her and praise her. If she jumps up again, repeat the process. Before you know it, she will have figured out that in order to get the attention she craves as quickly as possible, she needs to sit when she approaches you, not jump.
Clicker trainers use primarily positive reinforcement, but will also use varying degrees of negative punishment, negative reinforcement and positive punishment, depending on the dog and the individual trainer’s own comfort level and skill with the various methods. The jumping up example above actually uses negative punishment -- the dog’s behavior (jumping up) causes something good (you) to go away. Then, when she sits and you give her a treat and attention, it is positive reinforcement -- the dog’s behavior (sitting) causes something good (treat and attention) to happen.
Proponents of positive reinforcement training claim that a training approach based on rewards rather than punishment builds trust in the human-canine relationship and encourages the dog to think for herself and freely make deliberate choices of rewardable behaviors rather than living in fear of being punished for making a wrong choice. Positive-trained dogs often tend to be more willing to think for themselves, choose “right” behaviors, take risks, and offer new behaviors than do dogs who have been physically corrected for making mistakes.
Of course, it is not always possible to ignore a dog’s inappropriate behavior. Some unwanted behaviors are self-rewarding, destructive or unsafe, like barking at the mail carrier, chewing electrical cords or chasing cars. All trainers use a variety of approaches to correct unwanted behavior, but clicker-trainers generally apply methods that stop short of harsh physical corrections.
One such method is management. It is easier to prevent unwanted behaviors than it is to correct them. It is far easier to keep your dog properly confined in a fenced yard or on a leash than it is to stop a dog with a strong prey drive from chasing cars, cats, joggers or skateboarders. While you are managing the behavior, you also work to train a better level of control so that she becomes more reliable around highly enticing stimuli.
Another approach is the use of a “No Reward Marker” or NRM. The NRM is a signal to let the dog know she made a mistake. It is not applied angrily, just used in a neutral tone to let the dog know that the behavior didn’t earn a reward. Commonly used NRM’s include “Oops,” “Try Again,” or the sound “Uh!” or “At!” A properly-used NRM tells the dog that the behavior offered was not the one requested, and encourages the dog to try again.
Yet another positive behavior-correction method is to ask for (and reward) an incompatible behavior. A dog can’t lie on her rug in the living and bark at the visitor on the front porch at the same time. If we teach her that the doorbell is the cue to go lie down on her rug and stay there, she will no longer greet your guests with her sometimes unwelcome exuberance.
The Ongoing Debate
There is no lack of debate between trainers about the effectiveness of their various training approaches. Compulsion trainers believe that an aggressive dog must be physically corrected for the least sign of aggression: hackles raised, intense stare, growling. This teaches the dog that the behavior is not acceptable. Clicker trainers believe that this supresses the dog’s warning signals, but that the aggression is still there, waiting for the opportunity to be unleashed without warning on some hapless victim.
Positive reinforcement trainers suggest that a better approach is to change the way the dog thinks about the aggression-causing stimulus by associating it with positive things. Take a dog who wants to bite children. If every time he sees a child he gets a treat before he has a chance to act aggressive, he will begin to associate the presence of children with “Good things happen.” Eventually he will be eager to see children, and the aggression will fade. Aggressive behavior is not lurking beneath the surface, because the dog no longer thinks of children as a threat; they are now a source of good things.
Clicker trainers tend to believe that force-based training dampens a dog’s enthusiasm for learning, and “stifles their creativity.” Compulsion trainers may mistakenly assert that reward-trained dogs won’t perform reliably under stress. Clicker trainers say that violence elicits violence, and that many dogs who are euthanized for biting were made worse by physical corrections. Compulsion trainers argue that their methods are faster, and that sometimes the use of force can cause quicker behavior changes that save a dog’s life whose owner is at the breaking point and on the verge of sending the dog to the shelter.
Most trainers agree that owners apply whatever training methods they are using with varying degrees of skill and success. Trainers from both sides of the continuum talk about owners who “just don’t get it.” Other arguments aside, it would seem logical to conclude that much more harm can be done by an owner improperly jerking on a collar than by one who tosses a few extra treats.
Deciding on what training methods to use is a personal choice. Pet owners left to their own devices are more likely to follow their hearts and choose a gentle, non-violent training methods, while those owners who have been conditioned by past trainers and the pressure of competition to believe that a little “pop on the collar” won’t hurt the dog, will more quickly accept force-based training.
In the end, our dogs tell us the truth. We can find pet dogs and obedience show ring competitors from both training styles that are happy, reliable, willing workers. We can find dogs from both training styles that are poorly trained and out of control. But in general, a larger percentage of dogs in a compulsion-based class will grudgingly comply with commands or look bored or disgruntled than will dogs in a positive reinforcement class, where enthusiasm usually abounds among all students in the class, two-legged and four-legged alike.
Now let’s go back to our imaginary stroll around the neighborhood. You’re ready to sign up for a class, and just have to decide which one. Just put yourself in your dogs place for a moment and ask yourself which kind of class she’d prefer to go to. She’ll give you the answer.
Peaceable Paws LLC
Pat Miller, CPDT, CDBC
Pat Miller is a Certified Dog and Horse Behavior Consultant and Certified Professional Dog Trainer. She offers classes, behavior modification services, training clinics and academies for trainers at her 80-acre Peaceable Paws training facility in Fairplay, Maryland (US), and presents seminars worldwide. She has authored “The Power of Positive Dog Training,” “Positive Perspectives,” “Positive Perspectives 2,” and “Play With Your Dog.” Miller is training editor for The Whole Dog Journal, writes for Tuft’s University’s Your Dog, and several other publications. She shares her home with husband Paul, five dogs, three cats, five horses, a donkey and a potbellied pig. www.peaceablepaws.com