With so many people claiming to have the best training methods, which should you choose – punishment, rewards or a “whatever works” approach, combining both?

There’s an old saying among trainers:  “The only thing two dog trainers can agree on is what the third trainer is doing wrong.”  That certainly used to be true in the early days of training when there were few standards for what qualified as good training.  But now, the science of learning and animal behavior has taken away the mystery of dog training, explaining how and why certain methods work.

In reality, there are really only two types of training methods:  Reward-Based and Aversive-Based.  While the techniques and tools used may vary widely within those methods, the underlying philosophy is always broken into one of those two categories.

AVERSIVE TRAINING

Aversive-based training methods, also referred to as compulsive methods and traditional training methods, are centered on applying and removing something the dog finds unpleasant, relying on two principles of operant conditioning:

R- Negative Reinforcement

Negative reinforcement is the process of removing or withholding an aversive when the dog performs correctly.  It is considered reinforcement because it increases the likelihood that the behavior will be repeated.  It is negative because something is removed when the dog performs correctly.

One common example is putting pressure on a dog’s collar, while asking the dog to “Sit.”  When the dog sits, the pressure on collar is released. Another example is the seat belt buzzer in your car. When you put your seat belt on, the buzzer stops. You have just been negatively reinforced for fastening your seat belt.

P+ Positive Punishment

Positive punishment applies something aversive to decrease the chance that a behavior will be repeated. With positive punishment, an “incorrect” response produces an unpleasant or aversive consequence.

A shock from an electronic collar when a dog barks is positive punishment. It is positive because something is applied.  It is punishment because the behavior stops.

Aversive-based methods often rely on an array of training equipment, including chain (“choke”) collars, metal prong (“pinch”) collars, and electronic (“shock”) collars.  Even when these types of collars are not used, punishment-based training is reliant on a collar and a leash in order to administer “corrections” with a sharp jerk or “pop” of the leash

Some trainers have realized that more and more dog owners are seeking positive reinfocement methods and have started labeling their methods as such.  They claim that because they praise the dog after removing an aversive, they are using positive reinforcement.  However, the removal of an aversive is negative reinforcement. If the trainer were only using praise without aversive collars or methods, then they would be using positive reinforcement.

Note:  Like the seat belt buzzer, there are negative reinforcement methods such as “Behavior Adjustment Training” (BAT), that don’t cause the dog physical discomfort.  They do, however, still require exposing the dog to something aversive, but at a level that causes the least amount of stress possible during the behavior modification process.  I employ this approach when working with reactive dogs and like that it gives the dog a choice to avoid the aversive by giving a natural signal that the owner rewards by removing the dog from the situation.  In other words, the dog doesn’t have to learn a new command in order to avoid the aversive.

REWARD-BASED TRAINING

Reward-based training methods, usually referred to as positive reinforcement training, focuses on giving and taking away things the dog wants.  Reward-based training methods rely on two principles of operant conditioning:

R+ Positive Reinforcement

Positive reinforcement rewards a behavior by giving the dog something he wants or likes. By applying or adding something pleasant, you increase the likelihood a desired behavior will occur again.

For example, if your dog comes when called and you give her a treat, she has been positively reinforced.  She will be more likely to come to you next time you call her.

P- Negative Punishment

Negative punishment is the act of taking away something the dog wants as a way to decrease unwanted behavior.

One example is a dog that jumps and barks when the owner prepares to throw the ball.  If the owner withholds the throw of the ball or, even better, walks away with the ball when the dog jumps and barks, the dog will begin to learn that jumping and barking makes the ball go away.  Another example of negative punishment comes from the cable company.  If you don’t pay your bill, the cable company takes away your service until you pay your bill.

Within reward-based training, there are many different techniques, including lure/reward and clicker/marker techniques.  Some reward-based trainers recommend special equipment, such as head halters or front-clip harnesses, to prevent pulling on walks while the dog is undergoing the training process.  However, this equipment is generally a management tool and not a training tool, in that it is not used to teach new behaviors.

The Chair Test

If you’re still confused, take a look at what happens if you come into my office and sit in a chair:

The consequence determines whether or not you will sit in that chair again in the future. If the consequence is rewarding, you will be more likely to sit in that chair again in the future, especially if the reward is consistent. Now, what happens if I vary the consequence ieach time you sit in that chair?  Sometimes you get $20, sometimes you get shocked.  Are you going to run to the chair in hopes you get $20, or would you hesitate?

I frequently work with dog owners that have tried various forms of “punishment” for bad behaviors.  Everything from noisy cans filled with coins, to squirt bottles filled with water, to yelling, shoving or even hitting.  However, in order for these things to truly be called punishment,the unwanted behavior has to decrease.

“If a behavior increases in frequency or intensity, it is being reinforced,
not punished.

The most frequent example that I see are dog owners who attempt to use aversives to stop rough puppy biting.  The puppy bites at their hands or clothes and the owner immediately stops, turns their attention toward the puppy and scolds, shoves or otherwise “punishes” the puppy.  Instead of the behavior decreasing, it often increases.

Why? If a behavior increases in frequency or intensity, it is being reinforced, not punished.  Rough puppy play is a behavior that thrives on attention.  The goal is to initiate play, a type of attention.  So every time the owner gives the puppy attention (looking, touching or talking), theypositively reinforce the behavior.

Rewards and Aggression 

A common misconception is that reward-based methods won’t work with aggression or that reward-based trainers recommend ignoring all bad behavior, including aggression. Sound too silly to be true?   It is.

Reward-based trainers only advise ignoring behavior that does not threaten the safety of property, people or animals.  Instead, they outsmart the dogs by planning ahead and preventing these problem behaviors from recurring during the training process.  They do not put the dog in a situation where it reacts aggressively and then punish the dog, thinking the dog will learn aggression is “wrong.”  In such cases, these so-called trainers are reacting, not training.

Reward-based trainers first teach the aggressive dog basic skills that it may be lacking (such as walking nicely on leash, holding a down-stay at the front door, etc.), then practice those new skills at greater intensities, gradually building the dog’s tolerance to the situations that normally trigger aggressive behavior.  It might not look as flashy on television, but produces longer-lasting results.  Crash diets might work in time for the wedding, but permanent results require long-term changes in lifestyle.

“BALANCED” TRAINING:  BALANCE IS GOOD, RIGHT?

“Reward the good, punish the bad.” When faced with two extremes, we often tend to drift toward the middle, trying to find a balance between the two.  To many dog owners this seems to be the most reasonable approach; No extremes, just a middle ground between the two methodologies.  However, there can be many reasons for “bad” behavior that have nothing to do with disobedience.  If you believe that punishment is necessary in SOME instances, is needed for SOME behaviors, or that SOME dogs/breeds need punishment, ask yourself these questions before punishing the dog:

1.  Just how mad ARE your training skillz?  Have you ever videotaped yourself during a training session?  If you are a dog trainer, have you ever submitted video of a training session to your peers (as I was required to do for one part of my certification).  If not, you can’t really know how good your skills are…or aren’t.  Granted, it’s not easy to watch, but once you curse yourself for getting off your diet or vow never ever wear those jeans again, you can start to focus on your training skills and look for these common mistakes:

  • Unclear and/or inconsistent cues.  Is the cue the same each time?  Do you only say your cue once, or do you chant “Down, down, lie down, ah-ah! DOWN.”  Do your visual cues (hand signals) look the same each time?  Are you mixing verbal cues (commands) with visual cues (hand signals, pointing, etc.)?  Are you giving cues before the dog has learned the complete behavior?
  • Poor timing.  Are your rewards delivered at the exact moment the dog performs correctly?  If not, how many seconds pass between the behavior and the reward?  Is the dog doing something else by the time he gets rewarded?  Consider learning how to use a marker to improve your timing.
  • Inconsistent Rewards.  Not only is there more than one type of reward, there are many different ways to deliver rewards. Can’t think of any?  Then you still have more to learn.

2.  Is the dog exhibiting signs of stress that could be causing bad performance?  This is a big one.  If you don’t know how to recognize common stress sign, you could be missing a lot. Fear, anxiety, and over-arousal/over-excitement are all forms of stress.  Stress affects behavior in dogs, just as it does in humans.

3.  Do you know multiple ways to get the same behavior?  Just because 9 out of 10 dogs respond to your method of teaching “heel,” Dog #10 may not.  Are you familiar with – and proficient in the use of – the following methods to teach different behaviors?

  • Lure
  • Shape
  • Target

If not, it’s time to brush up on your dog training knowledge and skills before administering punishment.

“I don’t need to learn that stuff.  I’ve been training dogs a really, really long time.”

No matter how long you’ve owned dogs, no matter how long you’ve considered yourself a trainer, you can always improve your knowledge and skills.  NFL quarterbacks with multi-million dollar contracts still make bad throws and NBA stars still miss the basket.  To assume that one has achieved total perfection in communicating with a completely different species is the height of arrogance.  In other words, get over yourself.  

Somewhere, somehow, you make mistakes in your training. The mark of a good trainer is to spot the holes in your skills or knowledge and fix it, either by reading, attending seminars, or reaching out to other professionals who can offer advice.  The mark of a bad trainer is to punish the dog to cover bad knowledge/skill.

WHY METHODS MATTER

I frequently see dogs with behavior problems that are the result of aversive methods, whether at the hand of the unskilled owner or at the hands of a “professional.”  A recent study found that aversive methods, specifically the use of physical punishment, did not produce better-behaved dogs.  Another study, found that aversive methods can increase aggressive behavior in dogs.

The risk of causing or exacerbating unwanted behavior is high with punishment-based methods for the following reasons:

Negative Association – All punishment runs the risk of creating a negative association to something in the environment – or the environment, itself – or the person that is handling the dog when the punishment is delivered. A recent study in Germany measured the Cortisol (stress hormone) levels of dogs trained with shock collars. In the dogs that received random, poorly timed shocks (much as would be delivered by the average dog owner), the Cortisol levels increased over 300% just by entering the room where the dogs had been shocked one month after the training had taken place. As stress is the main cause of most behavior problems, increasing a dog’s stress levels runs the risk of causing behavior problems that were not present or exacerbating existing problems.

One example of how these methods increase stress involved a dog that suffered severe diarrhea at each training session with her former trainer, a man who used collar corrections on a prong collar, causing the dog to cry in pain. Shortly after her first training session, the dog severely bit a man who was visiting the owner, although she had shown no fear of or aggression toward men prior to her training.

Sensitization– Because traditional behavior modification methods involve setting the dog up to exhibit the undesired behavior and then punish it, the dog must be repeatedly exposed to the trigger that incites the behavior. The more a dog is exposed to a trigger and has a negative experience, the more likely the chance of sensitizing the dog, making the dog more sensitive to that trigger.

Positive training methods, on the other hand, work towards desensitizing the dog through carefully controlled, limited exposure. Each exposure is paired with something pleasant, often food, which changes the dog’s association to the trigger. At the same time, the dog is taught an alternate, more acceptable behavior to perform when the trigger is present, such as looking at the owner instead of lunging or barking at another dog.

Suppression of Warnings – Barking, growling and snapping without making contact with the skin, even an inhibited bite that does not break the skin, are all normal communication from dogs.  These warnings are designed to increase distance, keeping the dog safe. A dog that can warn and have the warning respected doesn’t intend to, or need to bite.

This is not to say that the only alternative is to ignore or reward the barking or growling.  Instead, these warnings should be respected, and then the owner should work with a qualified professional to change the dog’s behavior in those situations.

Aversive methods can and do work on some dogs in some instances.  Primarily dogs of sound temperament, bred for work, who are learning obedience and when applied by trainers with excellent timing and skill, who know how to use such methods sparingly and when not to use them.  Unfortunately, with training being an unregulated industry, trainers who have such advanced skills and knowledge are few and far between.

KNOWLEDGE, SKILL & EXPERIENCE

Methods do matter, but the use of certain methods is not always an indication of the quality of the trainer.  I have seen some highly-skilled and judicious use of aversive methods by trainers who had in-depth knowledge of when, why and how to use such methods and, more importantly, when NOT to use them.  I have also seen some very badly executed attempts at reward-based methods by inexperienced trainers.  The difference is that badly-applied reward-based methods will have little to no affect on the dog’s behavior, while aversive methods applied without skill or sufficient understand can have serious repercussions

Dog training is a science, an art, and a mechanical skill.  For some, learning the science behind the methods is easy, but learning the art of how and when to apply various techniques, how to adjust to each individual dog, and the mechanical skills needed to be effective is much more difficult.  And there are those who can make working with a dog look like a dance, but struggle when instructing a dog owner on how to get the same results.

Experience is important, but if the experience is only with aversive-based methods, the trainer’s knowledge is lacking.  If the trainer is highly skilled in a certain method, but doesn’t have the knowledge of why the method works or how stress affects learning, they may not be able to adapt to the needs of the individual dog.

This is why it is so important for dog owners to observe a dog trainer in action before beginning a training program.  A professional trainer will never refuse to let a dog owner observe a class and ask questions before/after. I recommend observing several different trainers for comparison.  Unlike nail technicians, massage therapists, and contractors, dog trainers who don’t seek certification through reputable and independent organizations are only regulated by themselves.

CONCLUSION

“If we can teach a Killer Whale to pee in a cup, you can train your dog without the use of punishment.”

~Ted Turner, SeaWorld

The vast majority of reward-based trainers, like myself, who have been training for
10, 20 even 40 years, started using some form of aversive methods.  Whether those methods were traditional, dominance-based, or modern aversive methods, including the use of electronic shock collars, most of us have been there, done that.

And, yes, we work with the same behavior cases as other trainers, everything from severe separation anxiety to serious aggression and with the same breeds.  In fact, I work with dogs who were rejected by local aversive-based trainers because their dogs were “too aggressive.”

The difference between the two types of trainers is that, although many of us were skilled in the use of aversive methods, reward-based trainers continued to learn from other professionals, stayed open-minded to the possibility that there were other ways and even better ways, and, eventually, we abandoned aversives as we learned how to better teach dogs to do what we want.

Over the last fifteen years, I’ve learned that there are many ways to train a dog.  And the more I have learned, the less I have needed to use punishment.  I don’t need brute force when I can outsmart the dog.  Knowledge is the ultimate form of dominance.