The Dangers of "Purely Positive" Dog Training

One of Canada's top dog trainers is Roger Hild. He is a founding member of the CAPPDT (Canada's largest dog trainers' association) . Hild has trained thousands of dogs over the past 20+ years. The following post, which appeared on his blog a few years ago, is reposted here with his permission:   

 

Plan B - Kill the Dog

 The needle slid easily into the vein and its lethal contents were injected slowly. The final good-byes had been said and now, as tears streamed down his face, the owner reflected on what went wrong as he watched his dog die. He thought about how hard he’d tried - he had sought out the best trainer he could find and when the behaviour didn’t improve, he had sought out another. No one, it seemed, either could or would offer him a solution to solve the behaviour problem which ultimately had proven terminal.

If the above paragraph seems overly dramatic, I can assure you it is exactly how many dog and owner relationships end. What’s really sad, in my opinion, is that many such endings are needless and can be rightfully attributed to a philosophy that ironically calls itself "dog-friendly." The philosophy I’m speaking of is also known as "PP" (purely positive) and is based entirely on facilitating behavioral change by awarding or withholding positive reinforcement. Sometimes also referred to as PROC (Positive Reinforcement Operant Conditioning), in some areas of the country its trainers have become so numerous they are all the public can access.

The problem I have with the pure positive philosophy is that it lacks balance. Pure positive trainers don’t use corrections, claiming any aversive punishment is unnecessary. Many among their ranks refer to anyone who uses corrections or punishment as cruel and their equipment and methods as abusive. Having a real problem with the word "no," or any other word with a similar meaning, their pursuit of getting such words out of the trainers vocabulary has an almost fetish like quality. Both in their dealings with the public and their dealings with other trainers, who subscribe to more balance in their approach, I find many of its practitioners to be arrogant, zealous and dishonest.

I want to make it clear that my remarks are directed primarily at the evangelical promoters and the zealous fanatics within this movement; I do not, however, want to paint all trainers that follow such a philosophy with the same brush. Narrow-minded people occupy every ideology as do zealots. Within the dog training community there are some trainers who have developed very good skills with a positive approach and they freely admit they use this approach because it is their own preference. They concede there are other methods that at times might need to be used and if they aren’t in a position to use these alternatives, they will refer the client to someone who can. They do not use inflammatory and derogatory terms to label colleagues who have chosen different techniques. I have no problem with these trainers rather I hold them in high regard - they state their choice and are not attempting to limit others right to choose. While they may remain unconvinced another approach might work better, they will not try to eliminate it from the realm of possibilities and choices. The honest among them will acknowledge that just as they were free to operate from choice, so should others be free to choose.

There is nothing offered by PROC methods which is not also available in a more balanced approach. In addition to the tools and approaches used by the purely positive trainers, balanced trainers add the use of whatever corrections are necessary to make the point completely clear to the dog. Corrections are simply information to offer clarity about what is wrong, what the consequence is and what needs to be done to get it right. The PROC trainer, therefore, offers far less than the balanced trainer - less information to the dog and fewer options to the owner. Other than time-outs, "ignore him when he’s bad," controlling all ‘resources’ and withholding all rewards, there is a real poverty of effective consequences. If those don’t work, they will often label the dog as not trainable and, rather than try using effective corrections, the lethal injection is often all they are prepared to offer. There should be far more information (in terms of behaviour changing consequences) provided to the dog before that final and drastic step is taken.

Much about the PP movement is dishonest. Some of the more common claims, made by pure positive practitioners, are that their method is "the most effective, reliable and humane method for training the dog." Such claims are frequently attached to statements that, "punishment-based methods don’t work." Many among them consider a punishment-based trainer as anyone who uses any kind of physical correction and they classify any method, that includes corrections, as dangerous and "far too risky."

Contrary to their claims, a PROC training approach is not as effective and takes considerably longer to reach any level of reliability even close to what a balanced approach can produce. In some instances, reliability cannot be realized using a positive only approach and some dogs will not be trainable at all until appropriate corrections are included. When this lack of reliability, and the increased time PP training takes is raised with PROC trainers, a common response is that they are willing to take longer because their approach is humane. Several things about this attitude bother me: IT’S YOUR MONEY AND TIME THEY ARE WORKING WITH - THE LONGER THEY TAKE THE MORE IT COSTS YOU, their suggestion that PP is more humane is false and this assertion is paramount to emotional blackmail. Balanced training is about 15% more effective, is considerably more efficient with faster results and has a far better track record for producing the changes necessary to allow the dog to remain in his home.

The statistics I cited above are from a lecture given to a Toronto audience in the early 1990's by Dr. Ian Dunbar. (I refer to Dr. Dunbar because these days he is so highly regarded by many in the PP movement who view him as the vanguard of the "dog-friendly" training concept.) He stated that the most effective training occurs when positive rewards are combined with positive punishment (i.e., praise/reward combined with correction/punishment). He said that under controlled testing for reliability, when positive only methods were used the reliability factor could be brought no higher than the low to mid 80% range (ranged between 83 – 85%). He went on to say when appropriately timed punishments were added, the reliability increased by 12-14% to a maximum of 97%. He said that no one was able to observe a 100% reliable response, i.e., perfect on all aspects of the tests under a wide range of varying circumstances.

That 12-14% differential means a balanced approach can clearly help more dogs and is more effective. It means some dogs destined to die could be saved. However, in discussions looking at the relative effectiveness of a balanced approach vs. a PROC approach, a far more disturbing attitude sometimes emerges. When confronted with the suggestion that using a correction would successfully modify the potentially lethal behavior (which had not responded favorably to a PP approach), their rallying cry became, "The dog would rather be dead than corrected."

The crux of the problem is their presumption that they know something that escapes the rest of us.

Because they are adamant that aversive corrections are not necessary and since they are committed to that ideology, they would rather kill the dog than correct it. They would rather kill a few dogs than review and possibly change some of their beliefs. Since they hold the arrogant belief that they are kinder, more humane, enlightened and "dog friendly" than those who would correct, they are confident in believing they can speak for dogs. Whatever they want is what the dog wants! Keep in mind, THIS IS YOUR DOG THEY ARE TALKING ABOUT - their philosophy fails and your dog dies.

In my mind’s eye I can picture the dog (from the opening paragraph) on the table frantically struggling to live. As the death juice spreads through his body, as the dog’s vision goes to black, I can imagine his last thoughts as being, "Why this? What did I do wrong? Why didn’t somebody stop me?"

(If you're interested in learning more, check out this clip from a recent radio program which featured a debate between a balanced trainer and two purely positive trainers.)  

What is "Purely Positive" Dog Training?

So much has been written on this subject that I considered simply avoiding it all together. Frankly, I’m bored to death with this debate. It’s an emotionally charged issue and the two sides (balanced trainers vs. purely positive trainers) are about as likely to agree as any other polar opposites, like liberals vs. conservatives.   But I keep getting asked this question so I’ll answer it. 

First let me explain what I’m referring to in case you’ve never heard the terms “balanced” and “purely positive.” 

There are four quadrants in operant conditioning (i.e. the way that animals learn), first systematized by behaviorist B.F. Skinner in the early 20^th century. Animals learn from positive reinforcement (aka positive reward), negative reinforcement (aka negative reward), positive punishment (aka positive correction) and negative punishment (aka negative correction). So far, no problem - everyone agrees on what operant conditioning is. 

Where they differ is that balanced trainers use all four quadrants and purely positive trainers use mainly two: positive rewards (like a food treat), and negative corrections (like a time-out or the removal of attention). Sometimes they use the third, negative rewards, but not often. The most important difference is that they do not believe in positive correction.

My opinion – positive rewards are a great way to teach a dog what to do, but they are a bad way to teach a dog what NOT to do. When I see a bad behavior, I prefer to correct it. Correcting bad behavior reduces the likelihood of it recurring. 

When a purely positive trainer sees a bad behavior, they prefer to ignore it and reward (reinforce) the opposite behavior). Here’s a real-life example of the difference between these two dog training philosophies: If I am fostering a dog who decides to put his paws up on the top of my kitchen counter and steal food from it, I will wait for him to do it, catch him in the act, and correct him. If my correction is effective I will only have to do it once, and his counter-surfing days are over for good. But a purely positive trainer would ignore the bad behavior and reward the dog when all 4 paws are on the floor. The problem with that procedure is that counter-surfing, like many other undesirable behaviors, are self-reinforcing. The reward is built into the behavior (in this case, the fun of searching for food, and occasionally being rewarded by stealing some food).  But counter surfing can be a very dangerous – potentially life threatening – habit. If a dog snatches a bottle of medication, a large piece of chocolate, a sharp knife with meat residue on it, etc. he can be seriously injured, maybe even die.

I would rather protect the dog’s safety and hurt his feelings, than the other way around.  And I’m in good company:

Gary Wilkes, DVM, a nationally known veterinarian and advocate of balanced dog training has written extensively on this subject. I came across this recent post where he addressed this very issue:  “Positive reinforcement cannot inhibit behavior. If you choose “positive” training you cannot stop anything. The common suggestion of “teach an alternate behavior” is senseless. Learning to ride a horse doesn’t erase your knowledge of riding a bike – no matter who much you are rewarded for riding horses….The claim to be “positive” is a rhetorical tool that allows [purely positive trainers] to fool either themselves or those they wish to persuade. .. All manner of working dogs are trained to dependable performance using a balance of positive and negative…. if using “negative” training is bad, why don’t Seeing Eye dogs look terrified?”

The New York Times obituary of Arthur Haggerty (1932 - 2006), called him “The most famous dog trainer in the United States.” Haggerty had trained more than 100,000 dogs including dogs for more than 450 television commercials and more than 150 feature films. The Times article stated, “Haggerty… had little patience for the New Age dog-training methods of recent years.” In a 2003 Dog World article Haggerty said, “These methods, in which dog trainers...whisper in the dog’s ear kissy face nice and click a clicker, are almost always ineffectual.” In the article Haggerty gave an example using a dog who was nuisance barking. Instead of following complicated protocols he preferred to simply correct a dog for nuisance barking, especially in a case where a renter threatened with eviction might be forced to give up the dog. People needed fast, effective solutions. Correcting the behavior enabled the dog to remain in the home and live, rather than be surrendered to an overcrowded animal shelter and die.  

  

These are only two examples of hundreds of experienced, well known and highly respected professional dog trainers who believe in correcting bad behavior. (The most famous advocate of correction of course, is TV’s Dog Whisperer, Cesar Millan. The “purely positive” crowd doesn’t like Cesar, but millions of dog owners around the world visit his website and read his best-selling books.)

So this debate is sure to continue, but I’ll stick with balanced dog training methods using all four quadrants of operant conditioning. This approach is based on science, not emotion, and it’s what I’ve seen work for thousands of clients over more than 2 decades of professional dog training experience. In contrast, I’ve retrained thousands of dogs whose owners were close to giving up on them when purely positive methods failed miserably. 

One last point: if purely positive training is so great, why aren’t service dogs trained with purely positive methods? Search dogs, military dogs, police patrol dogs and seeing-eye dogs are all trained in balanced training programs, using both rewards and correction, not in purely positive programs.