The Theory of Positive Reinforcement in Dog Training
From the foundation of operant conditioning to the idea of mistake avoidance through positive reinforcement training
I have a problem and, as with many of the problems in my life, it has to do with dogs. Specifically with Spinoni. I love their drooly faces, friendly dispositions, and hunting ability delivered with style. That said, their body language can be subtle and they can be stubborn. This got me in a bit of a pickle while trying to force fetch Windy City’s Cattle Kate. Simply put, the traditional force fetch process was not working. So I called my mentor, Jonathan Paranjothy, who has 30-plus years of experience training spinoni. After much banter and degradation, it was agreed that Jonathan would take Kate for a month to figure where I was going wrong.
At the end of that month Jonathan and I talked and we were both disappointed. Kate’s retrieving wasn’t improving—in fact, it was regressing. It seemed nothing would work. At that time, Jonathan was a self described “Old School Trainer” which refers to the idea that the default training method is rooted in negative reinforcement. In a moment of reflection, Jonathan explained a lesson his father taught him about training protection dogs. “Scott, my dad always told me you have to be open to what works. I think we need to step back and rethink this.”
I did. What I eventually landed on was a positive reinforcement training approach. At the time I did not understand the model, which I will describe below. I did understand that the techniques worked and worked fast. Once I switched to positive reinforcement, Kate’s retrieving went from complete avoidance to acceptable in a matter of days. It then went from acceptable to field reliable much faster than I had experienced with traditional force fetching other dogs.
Based on this experience, and many more experiences to follow, I became more interested in positive reinforcement training. To understand what positive reinforcement training is we need to understand some theory first. This will be a bit academic, but understanding this will give you the ability to grasp any training method or technique and will provide a great starting point to comprehend any possible advantages or disadvantages contained within. Basically, understanding these concepts will make you a better trainer without consideration to technique or methodology.
Understanding the theory of operant conditioning
To gain a complete understanding it is necessary to zoom out and take a look at dog training in a broader sense. Modern dog training is derived from the theory of Operant Conditioning researched by B.F. Skinner and his contemporaries starting in the early 1900s. This model has been used widely but I will attempt to keep it germane to dog training.
Positive and negative punishment
Skinner and his contemporaries set out to define a model that describes learning through the use of rewards and punishments. Punishment is a training technique used to reduce the likelihood of a behavior. A dog receives a punishment directly after the undesired behavior. There are two types of punishment, positive punishment and negative punishment.
Positive punishment is an adverse event that occurs directly after the behavior. A good example of this is an electric bark collar. The bark collar is a training aid that provides a positive punishment to reduce or eliminate unwanted barking. This is done by automatically giving the dog an electric shock directly after the collar detects a bark.
Negative punishment can be thought of as removing something the dog wants after an undesired behavior occurs. I have seen this used when training pointing dogs for steadiness on game. If, during the steadiness sequence, the dog breaks point the dog is not allowed to retrieve the shot bird. The dog is not given the implied reward (the retrieve) because the dog did not satisfactorily complete the steadiness sequence.
Negative reinforcement
Reinforcement is a training technique used to increase the likelihood of a desired behavior. There are two types of reinforcement: negative reinforcement and positive reinforcement. Negative reinforcement has two variants, escape and active avoidance. An example of escape is using an ear pinch during force fetch training. The dog is given a negative reinforcement (ear pinch) when the dog takes the training bumper, the ear pinch stops. To escape the stimulus of the ear pinch a dog must perform the desired behavior.
The other type of negative reinforcement is active avoidance. I have seen this method used when training a dog to go away from the handler—for example, when teaching a dog to cross a body of water to search and/or retrieve. An often used technique is utilizing the e-collar to “drive” the dog away from the handler into the water. The stimulus of the e-collar is given until the dog moves in the desired direction.
When describing negative reinforcement, I provided examples of escape and active avoidance. In one example the dog is escaping the ear pinch and in the other it is avoiding the e-collar stimulation. Considering negative reinforcement, it got me wondering if a dog might make the secondary association that the negative stimulation is coming from the handler, activity, or training context. If so, would the theories of escape and active avoidance apply directly to the handler, activity, or training context? This is exactly what Robert Milner posits in his book “Absolutely Positively Gundog Training.” Milner roots this idea in his real world experiences training search and rescue dogs and their rookie handlers after 9/11.
Positive reinforcement
Finally, the last type of training is Positive Reinforcement. Positive reinforcement increases the likelihood that a behavior will occur by providing a positive stimulus directly after a desired behavior. An example of this is a method for teaching a puppy the sit command. You simply follow your puppy around until it naturally sits. When the dog sits you immediately provide it with a treat. Do this a few times and the dog will start sitting in your presence to get the treat. Next, you overlay the “sit” command as the dog sits, still providing the treat. Most dogs will associate the behavior and command with the treat remarkably quickly.
I believe the evidence is mounting that positive reinforcement training methods are more efficient in the training of the dog and handler. The reason for this efficiency is that when positive reinforcement training goes wrong, escape and/or active avoidance is not triggered. In other words there is a dramatically reduced chance of missteps between the handler-dog team. Jason Carter describes the dangers of a misstep using a negative reinforcement method in his article “Understanding Timing In Bird Dog Training.” Jason describes what happens when e-collar training goes wrong.
“In that moment, the handler unraveled nearly two years worth of work with the press of a button. It took five years before that dog eventually learned to trust decoys again. It had taken hundreds of hours of training to get to that point and within less than a second, all was dismantled. Negative experiences—especially with young dogs—can cut like a knife.”
Most hunting dog owners are not professional trainers and do not have the experience needed to properly utilize negative reinforcement training. They are going to have to gain that experience with their dog and all too often it goes dramatically wrong as Jason describes above. One could argue that the skill set needed to effectively train a dog with negative reinforcement is the exact same skill set needed to use positive reinforcement methods. I wholeheartedly agree with this. But, when positive reinforcement training goes wrong, training progression may slow but is not reversed. And, most importantly, the mental well being of the dog and handler is not negatively affected.
Based on my experiences, the experiences of professional trainers using positive reinforcement, logic and reason, I believe positive reinforcement training not only to be more efficient but also reduces the likelihood of mental harm to the dog. Should positive reinforcement methods be the default training methods for our communities’ clubs, professional trainers, and individual handlers?
I want to hear about your experiences. Have you incorporated positive reinforcement training? Do you think that I am completely misguided? Do you think that I couldn’t be more correct? I am especially interested in scientific evidence and real world experiences. Please comment below.
SUBSCRIBE to the AUDIO VERSION brought to us by: ESP – Digital Hearing Protection for FREE : Google | Apple | Spotify
I am curious as to the methods for the Force Fetch training with positive reinforcement.
A couple months ago I learned that duck hunting my LM was detrimental to her training for duck search in the NAVHDA tests. She was used to a visual of seeing ducks fall or bumpers thrown, so she wasn’t wanting to send let alone search the water for birds. There were a ton of suggestions on how to “fix” the dog. Many of the suggestions used an e-collar. A friend who knows more than me decided that the best way was a slower reward/success based system. We place ducks, both dead and alive, out in the water for her to get her success. It started out in the open at 10-20 yards, and then worked to 40 yards out and 40 yards wide in cover and so on. Each time she gets sent there is a duck out there to get that success and reward. At the moment her bubble is probably 80 yards out and the same distance side to side, but she will search that area until she is dead tired. It won’t take long now to have her out there at 120yds, 200yds, 300yds, to satisfy the UT test. I don’t need a gunshot. I don’t need an e-collar. I can send and re-send. The dog knows that if she comes in that’s ok, but she needs to get back out there in the search without me telling her, and she will do just that. This system has taken longer to accomplish the goal but the dog is excited and confident to do the work. She loves it. Seeing her so happy to do this work let me know that we did it the right way for her.
The method you described is used commonly within NAVHDA circles and works well for many. I am glad you were able to find a method that fit well for you and your dog. Good luck running Utility!
It is my opinion that you can complete a trained retrieve using only positive reinforcement. The topic is the subject of multiple books so a comment isn’t the best format to explain. The book I mention in this article has a good overview and I suspect I will be writing future articles.
Hi Dale,
My first NAVHDA dog was a PP who received a 204 Pz1 without a “Force Fetch.” We spent HOURS training correct hold, retrieve, and delivery using positive reinforcement methods. You asked what are the steps? It was a very slow and repetitive process much like the “retraining” scenario you describe above working to get your dog to expand and learn the duck search.
George was my first hunting dog, and this would be my first NAVHDA UT test, so I was constantly seeking guidance and advise from more experienced bird dog trainers. Eventually I was convinced that I should “overlay” the e-collar to my dog’s fetch so I would have more control (ie, be able to use either negative reinforcement or positive punishment) at a distance. When I tried to add e-collar conditioning to the process, my dog’s fetch immediately fell apart, so I went back to the positive methods I started with. This involved MANY, MANY MORE hours of training and inventing creative “set ups” so George would have an opportunity to make a mistake that could be either positively reinforced… or ignored. It helped that he was a natural retriever, loved “games” and that my background was in training Schutzhund dogs (to annoying precision). At the time, I was also a Certified Pet Dog Trainer through an organization (APDT) that promotes positive reinforcement first and foremost. In other words, this was not my first dog, or first experience with using primarily positive reinforcement techniques.
My second dog was positively reinforced in play for fetching, taught hold and delivery position all using positive reinforcement but when birds were introduced, her motivations changed. With her, I had to also use negative reinforcement to get consistent, reliable retrieving behavior.
In any event, can you get a UT prize in NAVHDA without a FF? Yes. Is training a UT dog the fetch using positive reinforcement any easy process? No. To teach using positive reinforcement, I basically follow the same process of introducing the hold, teaching the carry, the pick up, and the finish. I introduce different items, different scenarios, etc., and reward the dog for doing it right. I’ve used clickers and I’ve used voice to mark correct (and incorrect behaviors). Dogs that are motivated by praise, problem solving, and food respond the best. For me, it’s difficult to use positive reinforcement to teach retrieving to dogs with no food drive and/or no relationship to the trainer.
Another question I hear a lot is: Does positive reinforcement produce a completely reliable retrieve in a test situation? I say, “it can.” You will need some luck. Your dog will also have to have the right temperament to think clearly in a test situation, and hopefully there will not be any competing motivations for the dog (like marking a bush, chomping on a duck because you are 100 yards away, etc.).
As a rule, I feel using negative reinforcement to train fetch is easier, faster, and possibly produces a more reliable fetch. There are many reasons profession trainers start with negative reinforcement to train the FF, but ultimately, it depends on the dog and the trainer. I like this quote, “Scott, my dad always told me you have to be open to what works. I think we need to step back and rethink this.” That’s great advise for any dog trainer!!
Thanks Scott for explaining operant conditioning in more detail for the readers/listeners of Project Upland.
Thank you for sharing your experiences. I think experiences like yours are valuable both as lessons learned and good data.
After training my Spinone I totally agree with positive reinforcement training, especially when dealing with a Spinone! Being a laid back (bullheaded) breed I found my spin would do just about anything for a reward/treat or even just verbal praise. E collars have there use but I think very few people know how to properly use them and end up doing more damage as you stated. Praise, praise and more praise has worked for me, when I’m happy my Spinone is happy! Good article I hope people take your advise.
Thank you, always nice to hear from another happy spinone owner!
Thank you, always nice to hear from another happy spinone owner!
I’ve been moving away from ‘traditional’ training methods over the past few years. With the pup I retained last year, I finished my transition to 100% Force-Free training. The results have been amazing in all aspects of her ‘training’. Makes me wonder why I was so slow to crossover, and has left me wondering why the ‘positive methodology’ has been so slow to gain traction in the birddog community.
Glad to hear you are having success!
Just like Susan, we have had great success with praise when training our Spinoni. Glad to hear from other Spinone owners. Great article, Scott.
Dave
Thank you
I’ve always used positive reinforcement when introducing new commands to a young dog beginning with the dogs name. When I bring a pup home I begin by just saying the dog’s name and immediately treating when the dog makes eye contact. This provides a life long benefit. When I begin teaching recall I continue to reward/praise proper response but do so with a leash graduating to checkcord to insure a proper and prompt response. This is where I begin to question the solely reward based methods. It seems that at some point the praise or food reward I provide will be over shadowed by the dog’s desire to do something else (give chase to a deer rather than heed my recall for example) and that at this point a (properly introduced negative) collar correction is appropriate.
I use the e-collar as safety in the woods as well it just makes good sense. And I think you touched on a misconception of positive reinforcement training. The idea that you have to continue to reward the dog to get the desired behavior. The idea with any style training is to be able to command your dog against their natural drives. The same is true with positive reinforcement training.
Only a very inexperienced trainer or a sadist relies on negative reinforcement only. Dog trainers for ever have known that all dog’s need a balanced mixture of positive and negative reinforcement. By combining the two, not relying on either, a dog is trained in the most solid, efficient, and humane manor. I was a pro for quite a few years and this is the way I learned back in the 70’s.
I have always been told that there is nothing new under the sun in dog training! But you are correct trainers almost always operate in both negative and positive methods. I didn’t mean to imply that this was not the case. I am saying that I think a greater focus needs to be brought to the positive side. Especially if you are a new trainer.
I think you hit on some great key points and hope the conversation about positive gun dog training being more accepted in the wider gun dog community can continue. As a positive trainer myself, I am often left to train on my own but would rather have that than be heckled as to why I am not doing it the “old way”. I train and hunt with a Brittany and an English Setter.
Thank you! Not that you should ever put up with heckling! But also don’t pass up an opportunity to learn… This tradition has been around for a long time and you never know when someone elses methods might strike a chord. Although I know people can get set in a methodology and become rigid and that does not make for a fun training day.
I am happy to see this article! I have a 1 year old Spinone, I joined NAVHDA (Yankee Chapter where Jason Carter is at) last year. I am a positive trainer, and learned much this past year. I was surrounded by knowlegable trainers, and a near equal amount of dogs with e-collars, and never once was I heckled. Animals have been trained positively for years with great success. I do not see why there is a need to use the e-collar(though I hear and understand the reasons ). I also realize I have a long way to go to prove my point, and I hope others will work towards that goal.