Should you only use positive reinforcement to train your service dog? Or will combining rewards and corrections provide greater reliability?

By Sharon Wachsler CPDT-KA KPA-CTP

If you are training your own service dog, or thinking of taking this journey, you should know how using exclusively positive reinforcement training can affect you and your dog.

As you’re starting the training journey with your precious service-dog-in-training (SDiT), upon whose furry shoulders rest so many of your hopes and dreams, you may wonder what type of trainer to work with. Given how important your dog’s training is for your future quality of life, it’s a big deal.

• Should you find a trainer who relies entirely on rewards, usually food, often referred to as a “force free” or “purely positive” trainer?
• Or should you work with one who also uses collar corrections or other types of punishment, usually referred to as a “balanced” trainer?

Just how important is your trainer’s approach? Will using punishments (“corrections”) during training harm your service dog’s chances of success? On the other hand, will training with lots of treats make your dog “food crazy,” and too excited and distractible to work in public? This post provides some answers.

My Background

I’ve been in the owner-training service dog community for over 25 years. I started as a disabled owner-trainer in 1998 and began working as a professional trainer with service dog owner-trainers in 2014. My decisions about methodology are based on these decades of combined personal and professional experience and the hundreds of SDiT teams I’ve worked with in those years.

Personal

When I started training Jersey, my first service dog, in 1998, I used a combination of praise, treats, and choke chain corrections. (Choke chains are also known variously as check chains, half-check collars, training collars, or slip leads.) This was effective for the great majority of what my dog needed to learn. There were a few things I didn’t know how to train, but I thought that might be due to my dog’s “soft” nature. My dog was extremely motivated by food, but she didn’t have a lot of “natural training drive.” I was already planning to look for more work drive in my successor service dog. Training was very slow, but I was happy with these methods because they were working! My dog became a wonderful service animal who changed my life.

At that time, living in a remote rural area, and long before the existence of Youtube, Facebook, or Zoom, there was no trainer to work with me. The idea of training your own service dog was almost unheard of. I got all my information from a couple of books and an email listserv with other trainers — some professional, and some owner-trainer pioneers. Some of these trainers urged me to try out this new fad I’d been hearing about called “clicker training.” With great skepticism, I eventually read up on it and tested it on my service dog. I was shocked to discover that I could now train things that had seemed impossible before. And it was FAST. What had previously taken months, my dog learned in days. What had taken days was now taking minutes.

While I was an immediate convert to the power of marker/reward training, learning to train without any punishment at all took time. It required more knowledge and skill. However, once I stopped falling back on corrections at all, I saw a huge surge in the speed of my dogs’ learning, and my teamwork and relationship with my dogs also improved. (More on that below.)

Professional

Many years later, I apprenticed with two pet trainers (one “purely positive” and one who was transitioning from positive reinforcement to “balanced” training) and also worked as a trainer for one of them. Then I opened my business coaching service dog owner-trainers. While I definitely considered myself a positive-reinforcement trainer, occasionally the pressure to try to get quick results for clients led me to try something mildly aversive (such as something seemingly so benign as a scolding noise or social pressure). This often seemed to work in the moment. But, over time, I always discovered some type of unwanted problem had come along for the ride that would also need to be addressed. This meant that the process was actually slower, in the end.

This phenomenon — unintended secondary consequences from aversives — is often referred to as “fallout.” With service dog training, the process is so long and fraught for the owner-trainer, anything that complicates it or has any possibility of causing behavioral harm to a SDiT must be avoided. Fortunately, the more expertise I gained, the less likely I was to try to cut corners, and I eventually swore off aversive approaches entirely.

Some of my service dog clients had done no training before we started. Some had worked with reward-based trainers before, and others had to switch from the aversive methods that they had learned from previous trainers or the culture at large. The impact of reward-based training that I observed with these teams continually struck me as very important, yet seldom (or never) recognized or discussed by the training community. I describe three of these “side effects” of positive reinforcement training for service dog teams below. I particularly noticed these trends in my disabled clients who switched from balanced training to positive reinforcement.

Three Surprising Reasons Why Positive Reinforcement Training Works Best for Service Dog Owner-Trainers

1. Positive reinforcement training enhances a service dog’s ability to make the right choices.

Service dogs often need to make decisions about what to do on behalf of their handlers. They may need to alert or respond to a situation that the handler isn’t aware of: alerting to a sound for a hard-of-hearing handler, stopping at a curb for a handler with sensory processing damage from a brain injury, waking up a handler with cognitive impairment or depression at the sound of an alarm, interrupting an anxiety-based behavior in a handler with PTSD, fetching a family member for a handler with epilepsy.

Because the handler is not aware of this cue, the dog needs to be comfortable taking the risk to respond without being commanded to perform the behavior. In fact, in these situations, the handler may even cue the dog to do something else because they are not aware of the “cue” and may think the dog is “misbehaving.” The service dog needs to correctly perceive the situation and feel confident enough to make the right choice. This type of training is often referred to as “intelligent disobedience,” but it’s actually just about teaching the dog which cue is more salient in a given situation.

Clicker training that uses shaping and encourages the dog to think for itself is ideal for giving dogs this level of confidence, creativity, and discernment. Clicker trained dogs often learn to “offer” behavior, which is just what you need if you want a dog to do any sort of alerting or responding task.

But won’t corrections for the wrong behavior make the dog more reliable in performing the right behavior?

It would seem like that should be true, but that’s generally not how behavior actually works because of the interplay of learning with emotions and the generalization of emotional states, including fear, along with learning. The purpose of punishment is to suppress behavior. If your dog jumps up on you when you walk in the door, and you say “NO!” in an angry voice, your hope is that in the future, she will not just up on you. When punishment works correctly, the unwanted behavior occurs much less frequently in the future until it never occurs again.

Suppressing behavior is certainly beneficial if you are only concerned with stopping one particular undesirable behavior. However, training with punishment typically has an overall suppressive effect on not just the target behavior, but all behavior. The dog tends to learn, “I better not put a paw out of line or it might be bad news for me. Better to wait and just do only what my human tells me I must do.”

Punishment is very appealing and seems a good idea when you’re trying to get a dog to STOP doing X, but it’s not that great for teaching a dog to START doing Y. Even though it is crystal clear to us, the humans, that these two situations have nothing to do with each other, the dog may conflate the situations. This is why punishment often backfires.

This can happen with very related behaviors that easily make sense to most people, such as…

• I trained a very sweet sporting breed dog as a pet and ESA for a family that included a developmentally disabled young woman. The mom really did not appreciate the pup’s enthusiastic jumping up for greeting and scolded or turned her back on the dog when she jumped. Later, the family asked me to train the dog to lie across the daughter’s lap for deep pressure. It took a lot of time and careful shaping to overcome the dog’s reluctance to put a paw on a human.
• I worked with a seizure response service dog that had been originally been trained by a “balanced” trainer. As part of the dog’s obedience training, the trainer had “proofed” the down-stay, which included correcting the dog for breaking the down-stay. After that, when the handler had a seizure, the dog stayed in his down, rather than responding to the seizure. It was safer to stay in the “down.” He didn’t want to risk being punished. 

Because fear generalizes, less obvious associations can also occur. A dog that has been punished in the kitchen for counter surfing may be hesitant to perform any offered behavior in the kitchen. In fact, dogs who are punished enough may simply stop offering any behavior and become unnaturally quiet and still. I have seen this occur most often in dogs who have been trained with shock collars (also known as electronic collars, e-collars, stimulation collars, static collars).

Service dogs need to be able to make decisions without fear of punishment for “disobedience.” They need to be willing to take risks and think for themselves. Skilled, positive reinforcement training is great for empowering dogs to make good choices. (By the way, rewards-based training is also the best for training the rock-solid manners that service dogs need for public access. But that’s a post for another time!)

2. Positive reinforcement training helps dogs cope better with the stress of working in public, reducing the chances of burnout and early retirement.

Working as a service dog means going into public frequently, which is inherently stressful. It requires a lot more of a dog than being a pet does. Michele Pouliot, former director of research and development for Guide Dogs for the Blind (GDB), cites these stressors as a factor for why GDB switched to clicker training in 2006:

“The working environment for guide dogs had become progressively more difficult… Examples of environmental challenges included an increased number of encounters with ‘less than friendly’ humans on the street, more free-running dogs (often aggressive), higher volume vehicle traffic, increased pedestrian traffic, and louder noise volume. It became essential to address the question of how to make the job of a guide dog in the modern world more positive and less stressful.”

Very few dogs are cut out for the life of a service dog. Those special dogs that are born with the confidence and resilience to thrive and succeed as service dogs need ongoing support to maintain maximal mental, physical, and behavioral health. Stress accumulates and is one reason many service dogs develop physical or behavioral issues that lead to early retirement. Some stress in life is unavoidable. But when you add fear, uncertainty, intimidation, pain, or even discomfort to a dog’s life, you add stress. Training entirely with rewards cuts out the stress associated with punishment.

Of course, positive training is not enough. To maintain that very even keel (often referred to as a ‘bombproof” temperament) and that ability to think well in challenging environments, service dogs also require additional help to cope with stress. Positive enrichment, frequent breaks, the optimal amount and type of exercise, and effectively reading your dog’s subtle body language cues are also essential. Skilled reward-based trainers are usually adept at helping owner-trainers incorporate these other forms of behavioral support, too.

3. The third – and perhaps most important – reason is that rewards-based training is much better for the handler. 

I’ve never heard anyone else address this, but I feel really strongly about this one. Punishment-based service dog training is oftentimes inhumane to the disabled handler. This is particularly true for someone who is training their own service dog, and it goes double for a trainer/handler with a psychiatric disability. 

Occasionally I hear service dog trainers mention that punishment-based methods may be physically difficult for some handlers to carry out, depending on their physical abilities, size, or dexterity. That can certainly be true and is one of the reasons that the first manual for physically disabled people on how to train their own service dogs relied heavily on treat-based training.

However, there’s another, much bigger factor. Disabled people who are training their own service dog are investing not just huge amounts of money and time into training their dog. They are also investing all their hope and their hearts. I have seen over and over again how relieved and happy owner-trainers and handlers are when they’re given the tools to train effectively entirely with rewards. 

Training with punishment requires the handler to be on the watch for every time their dog makes a mistake – so they can correct it. This puts the dog and handler in an adversarial relationship. It means the handler’s brain has to keeps firing those “oh no” neurons. This naturally breeds anger and frustration. It can even lead to despair. It keeps the handler focused on everything that is “going wrong,” and because SDiT trainers are training a lot, it may feel like training is constantly going wrong. Some start to feel like the dog is intentionally misbehaving or disrespecting them. It causes great harm to the trust and relationship between the dog and handler. 

Most of the people who contact us for help training their own service dogs have psychiatric disabilities. If you have a psychiatric disability, by definition you are already coping with chronic negative emotional states. People with physical disabilities are usually also coping with all the stress that living with a disability in today’s society involves. People want to train their service dog to reduce stress and difficulty. They want to have a loving bond and a relationship built on joy and trust. This is much easier to achieve when you are not being instructed to constantly notice everything your dog is doing “wrong.”

I went through this with my first service dog, Jersey. Most of our training was done with rewards, but some skills I only knew how to train with corrections — skills like “wait” or “leave it.” I remember how angry, frustrated, and ruminative I would feel, sometimes for hours afterward, when I trained Jersey with the choke chain on “leave it.” I didn’t know that the problem was our training method. I just knew I was frustrated with her being so “bad” and with myself for not being able to train her as easily as I thought I should.

On the other hand, if I support a handler in noticing and rewarding good behavior, it has multiple positive “side effects.” First, it leads to better training. That person is more tuned in to their dog. They are able to read the dog’s body language better and make better training choices. They become more skilled at delivering reinforcement, which drives better behavior.

Secondly, it makes training much more fun and rewarding for the handler and massively improves the relationship between the dog and handler. That handler is more likely to do more training because it’s enjoyable, and not so much of a chore. 

Some of the cases that really stand out to me in this regard are the veterans and first-responders I’ve worked with who previously had trained with corrections. These handlers have typically been hard-working, great about following instructions, and dedicated to carrying out the homework. I think because of their history with following a chain of command, they were less likely to question methods or slack off. They did what they were told, always, and to the best of their ability.

When they learned they could train without corrections, their relief was incredibly touching to me. All service dog owner-trainers love their dogs. When we train a service dog, we are looking to our dog to help and support us, to build a mutually beneficial and trusting relationship. If you are also following a trainer’s instructions that you know will cause your dog discomfort, stress, or even pain, it puts you into a state of conflict that can be extremely uncomfortable. For people with psychiatric disabilities or a trauma history, it can aggravate symptoms.

“[Our first service dog trainer] wanted us to use a prong collar. She was very different from [Sharon]. I don’t like the choke collar. I don’t like harshness. Sharon’s the type of trainer we want.” Frankie Mazzei, Veteran, US Air Force, US Army

This relates to one last “side effect” of positive training that I’ve noticed for handlers. I’m not a mental health professional, so this is just based on my layperson’s observations.

I think there is often a subtle therapeutic effect for the handler to train with positive reinforcement. Many of my clients’ disabilities were caused by abuse, assault, or other trauma. And even in those cases that weren’t, living with a disability in this culture is often emotionally brutalizing. People with disabilities are typically disenfranchised. 

A big part of the way I train is to promote empathy and connection with the dog’s emotions — simply because it makes you a better dog handler and trainer. It is often a lot easier to be kind and focus on a relationship with a dog than it is with other people or with ourselves. I have seen how this encouragement to be kind and considerate to the dog, and to think about ways to empower the dog, has helped clients over time to see that the same is true for them. Their dog learns better if she is happy and secure. The client realizes he doesn’t have to be punitive toward himself, either.

Compassion, empowerment and freedom from suffering can be a two-way street for disabled owner-trainers of service dogs. Training service dogs is about adding freedom, independence, and greater well-being to the lives of disabled individuals. To me, it’s only natural that if we want to bring about these positive changes, we do it in a positive way.

To get started with training your service dog, watch the free informational video on our home page or register for our level one class, PEARL DISC. Happy training!

P.S. This post does not go into the general topic of how aversives affect dogs and their training, in general. To learn more about this topic, please visit Eileen Anderson’s blog. It provides accurate information on the fallout from aversives and dispels confusion about objections to rewards-based training.

https://atyourservicedogtraining.com/benefits-of-positive-reinforcement-training/