Clicker Training: Thumbs Up or Down?

Editor’s Note:
In conversations, interviews, and observations of scores of horsemen, I’ve come to understand that ‘pressure and release’ or negative reinforcement, is an extremely effective method for working with horses. Since horses use negative reinforcement on each other, it’s easy to understand.

Understand more about negative reinforcement here.

But a Practical Horseman article states:
“In most cases the trainer prompts the desired behavior by guiding the horse with pressure (a negative) and rewards the behavior by releasing pressure. But if the behavior is evoked with something positive, the horse learns faster and remembers better, says Dr. Sue McDonnell. “There’s no question that positive training methods are more efficient.”
That’s the theory behind reward-based systems like clicker training. The horse learns to associate a click with a food reward, so the trainer can instantly reward behavior using a quick click as a stand-in for the food.

Is positive reinforcement (like clicker training) more effective than negative reinforcement (pressure and release)?

One study reports that positive reinforcement is, in fact, less effective than negative reinforcement.

We put it to frequent contributors and Best Horse Practices Summit presenters as well as board members: Katrin Silva, Amy Skinner, West Taylor, Cindy Morin, and Dr. Steve Peters.

Please Note: What follows here is not a study, but a conversation with folks who work with scores of horses other many years.

Learn more about learning theory and tribalism in the horse world.

Read this article on a horse trainer who uses Clicker Training

Dr. Steve Peters

Dr. Steve Peters:

Many of these training techniques are borne from Animal Behaviorism. Research involved simple stimulus response behaviors, many of which was reinforced with foods. Think pigeons tapping, rats in mazes, etc.

During its hey day, Behaviorism was never concerned with what was actually going on in the brain.
However, the world has changed and we are learning how animals’ central nervous system works. We are learning how to develop a language and conversation that works with what the horse can recognize and understand. It is reinforcing to the horse to simply realize that we are having a conversation and that certain behaviors result in certain changes from us.

I think a number of horseman feel that the simple clicker training type responses shortchange the horse and our ability to grow a working relationship that is dynamic and allows the horse to actively participate in a much more complex interaction.

Katrin Silva

Katrin Silva

I do agree with the principle that positive training methods work better than negative ones. If you’ve seen me work or heard me teach, you know that I’m always looking for a chance to reward a horse.
However, boundaries are also important, and I find horses actually like them. There are times when a firm “No!” is a perfectly appropriate response – with your voice, your leg, your spur, whatever is necessary. Then you move on, without holding a grudge, and look for a chance to reward again. This is not abuse. No horse will be psychologically scarred from an occasional correction in the form of a negative reinforcement.

A few months ago, I lost a client because I could not promise her to use exclusively positive training methods with her horses. She thought I was way too demanding. This woman is now working with a clicker trainer and has stopped riding her horses altogether.
Apart from the constant food rewards, my main problem with clicker training isn’t the positive approach – it’s the overly simplified communication style. Horses and riders can talk to each other in a much more subtle fashion, body to body, back and forth.
Good riding takes feel, sensitivity, timing, physical balance, etc., etc. It’s not about a tool that looks like a garage door opener, because horses are much more complex creatures than garage doors.

Amy Skinner

Amy Skinner:

It’s possible that learning is faster with positive reinforcement, but there are so many negative side effects. You end up working harder to fix behavioral problems that you could have otherwise avoided.

The pressure and release training we do on the ground works well because it ties into what we do in the saddle. Everything connects and the horse can understand all the ties from ground to saddle. Clicker training does not allow that. Horses like things to be smooth, and clicker training forces you to chop up your work, to stop in mid-movement to reward with a treat, instead of asking a horse to follow a feel of a rope, seat, or bridle. I’m currently working with a clicker-trained horse, and I find he’s much less frustrated, much happier and more peaceful without clicker training, now that everything ties together for him.

Katrin Silva:

Good points, Amy and Steve. Here’s another thought: I don’t think of “pressure” as something automatically “negative.” Framing it this way makes it sound like the horse is frantically trying to get away from the discomfort.
Instead, I think of training horses as communication. I think of the signals I give as conversation starters that prompt the horse to engage with me and to try different responses, some of which I will reward, many of which I will ignore, and a few of which I might reprimand. The horse does not try to “get away from pressure,” the horse tries to talk to me and vice versa.
George Lakoff talks about reframing as a tool to take back political power. We can do the same in the horse world!

Cindy Morin

Cindy Morin

Interesting that there are some people who swear by the clicker method. I “feel” it is a very small tool in the really big box of tools one needs for training horses. As with humans, there is not one training method that fits all. You need to understand your horse and how he thinks and the possible baggage he may have. Each horse is different.
I agree strongly with what’s been said here; I don’t think I will ever use it. It feels like something you have to do before you actually work on the issue at hand. And I don’t like treating my horse like a trick pony and feeding a million treats either.

West Taylor

Horses are not capable of discerning between positive and negative. These are human words that we like to apply. Therefore, I don’t think of anything I do with a horse as “positive” or “negative.”
Pressure is the invitation to the horse to seek mental relief. Pressure is applied and the horse starts seeking mental safety. As the horse seeks relief I am guiding the horse to mental safety.
Let’s say I am asking a young horse to walk on a tarp:

West Taylor

Pressure is applied to the horse to walk forward to the tarp. As the horse seeks mental safety he will start to explore his options: He may want to avoid the tarp. As he does so there is a slight increase in pressure. As the horse seeks again, he may turn back toward the tarp. Then pressure is released or lightened.
I want the horse’s mind to be seeking mental relief. Each time the horse shows interest or moves toward the tarp, he will experience mental relief as I lighten or totally release pressure. Each time the horse seeks this mental relief he is also rewarded with a dopamine release which then reinforces the behavior of seeking mental relief.

Again, I want the horse to seek mental relief each time he senses pressure. I am simply there to direct his mind to the mental relief.
In the end, the horse is not focused on the tarp. Rather, he is seeking direction from me as to where he will find mental relief. As a result, his mind will find this mental relief at the tarp. The closer he gets to the tarp the more mental relief and dopamine rewards he will get. Ultimately he will seek relief so strongly that he will walk right onto the tarp.

That being said, I know mostly nothing about clicker training. From my minimal observations, it appears to me that the horses are simply performing maneuvers expecting a treat, not seeking relief. Then the question is:

What if I am in a real dynamic situation with my horse and I don’t have my clicker or any treats and therefore my horse is not able to figure it out on his own?
I feel the treats we can do best with neurochemical. There is an endless supply already built into the horse, and if your horse has been taught to seek mental relief to get his treat of dopamine, now you have a horse that can solve his own problems and pressures in life. This also works super well with humans! Teach our kids to find mental safety within their own mind and beliefs is way more impactful than teaching them to run from the pressures of life.

Amy Skinner:

Really well said! Clicker training involves performing movements instead of seeking and engaging. You hit the nail on the head, West.

This study reports that positive reinforcement is less effective than negative reinforcement.

Posted in BestHorsePractices Summit, Clinicians, Training.

42 Comments

  1. Excellent dialog here and points well taken by everyone. I especially loved the points about conversation between horses and humans. I agree that horses are complex beings and their natural curiosity caused them to explore and seek relief. What a joy it is when conversations lead to understanding which opens up opportunities for the horse to offer more with a willingness to do so. As Dr. Peters and Wes mentioned, the neurochemical treats are always readily available! What a super connection with science and behavior. This is one of my favorite article so far— more please. And thanks to all contributors on a very contentious topic.

  2. This seems like a very biased study and article.
    It doesn’t appear in the article that you talked to anyone that has had success with clicker training, or that really understands how to use it appropriately or has spent any significant time learning, studying and applying it with any expertise.
    I know many amazing trainers that use it successfully as a “tool” along with many other modalities. I would like to hear from someone that can share when this is an appropriate way of adding to the conversation of communication with the horse. Horses are aware of what you know and feel and what you don’t. If you don’t believe in what you are offering or have confusion about it, the horse will feel that from you. They like clarity. If you are consistent, clear and believe in what you are offering, have an awareness of the horse’s engagement of that. The horse will seek to understand and communicate. It is not about the food “treat”. It is about awareness for both the human and the horse. Who doesn’t like a cookie or payday? Would you pay a little closer attention if someone said their is possibly $100.00 in this for you? I bet most professionals are making a pay check. The food reward. Hmmm, how interesting….

    • Thanks for commenting, Jill. Please note that this is simply a conversation, NOT a study. As for bias, we know that virtually every written piece has bias. Sometimes it’s clear. Sometimes it’s more subtle. I do think that many, as Cindy Morin stated here, believe that positive reinforcement can be one tool in a greater toolbox. But stating, as McDonnell did, that it’s better, is worth questioning, especially as there is a study to the contrary.

  3. I’ll start first by saying that I’m not a proponent of clicker-training as a one and only method and that it can be misused in many ways. I use positive AND negative reinforcement with my horses. Aversives should be kept very low-level though (Mark Rashid’s work).
    I agree that initally behaviorism was sadly not concerned with relationship. Fortunately, a lot of modern trainers including Ken Ramirez, probably the best trainer in the positive reinforcement world, DO insist on establishing strong relationships with the animals they work with.

    That said, I think it’s important to point out that 2 things mentioned in the above interview are actually wrong because of lack of knowledge.
    1) In Katrin Silva’s first “speech” what she describes as “an occasional correction in the form of a negative reinforcememt” is NOT negative reinforcement, it’s POSITIVE PUNISHMENT, which is completely different.
    2) West Taylor is talking about “positive” and “negative” without knowing what these words actually mean in the behaviorist world. All living creatures know what “positive” and “negative” mean there: it’s nothing to do with moral judgement. “Positive” means that you ADD something (a reinforcer or an aversive), “negative” means you REMOVE something (again, a reinforcer or an aversive.
    Please look up BF Skinner’s work, the four quadrants of operant conditioning.

    If various methods of training must be discussed, I really wish people who do so really knew what they are talking/bitching about. This would make for a more balanced debate.

    • Interesting that you would call out individuals, but not post your own name in this discussion. Any form of training can be effective depending on a variety of factors, the least of which is a knowledge of the theories and definitions of operant conditioning. In the end, the effectiveness of any training methodology is reliant on three things from the trainer: Firmness, consistency, and patience.

    • Dear F,
      Like you, and like Mark Rashid, whom I deeply respect, I do think positive reinforcement should be used most of the time when working with horses, though it can’t be used exclusively. I never pretended to be an expert on clicker training or behaviorism. What I know about positive reinforcement training I know from a couple of horse owners who practice it and have tried to explain it to me, not from years of study. Based on these explanations, I use the term “positive reinforcement” for any type of reward, and “negative reinforcement” for any type of reprimand. I have never heard the term “positive punishment.” Please accept my apologies for using behaviorist terms incorrectly, but I’m not the only one who does it.
      While we’re on the subject of semantics, don’t you think “bitching” is a little harsh as a descriptor of the discussion that took place? And why did you put quotation marks around “speech” – was it not something I actually said? English is not my first language, so I’m probably missing something, but I’m genuinely curious.
      Katrin

  4. Thank you to all the contributors on this “talk” about clicker training. I’ve never used it, so will not add to the commentary to it’s benefits or downfalls, other than to thank you for bringing up a subject to discuss. There is a time and place for many types of training. BUT the key is that you should be able to move beyond the training sessions to being able to work with your horse in a show arena or out in the woods. If your tools are your voice and body instead of “props” that you have to carry around, then you have your training toolbox ready at any moment.

  5. I agree with Jill and F, I like what was said, but it is undermined a bit by lack of balanced perspective of all in the conversation. A nice follow up would be cool. People are looking up to these mentors and not getting the whole picture. I could comment on a lot of things I know nothing about, how valuable would my information be to you? What does it say about my integrity if I’m commenting on something but can’t be bothered to do even basic research to understand the terminology?
    My horse has had a more than average share of extended medical downtime, clicker was a cool way to keep him from mentally shutting down without moving too much or running away during handwalking. In an emergency, clicker reinforcement can get a nervous horse tuned in quickly and even onto a trailer at times. I love soft feel and the dance of harmonious horsemanship, too. I realize you said this, yet I feel it needs a better explanation on all sides.
    I’m disabled now, and still have a language with my horse from both foundations. Not saying I’m good at either one, but good enough to have fun and the transition was a very smooth one, thanks to the multiple tools you mentioned. Glad to have this forum.

  6. Conversations are how we lead to better understanding. So thanks for introducing this topic.

    I have not studied clicker training though have watched demonstrations and others who use it. It would not be my approach of choice, I’d forever be losing the clicker device! All the comments here reflect a personal style, preference or a view to how the horse might interpret the training technique. I’m going to offer how the human as a learner might benefit from the style of training and why it has become popular.

    I’ve noticed of those who are drawn to it, again not a scientific study but rather an observation, happen to be female and appear to want their horse to “like” them. Yes they might be able to get their horse to do some impressive stuff on the ground, but they avoid riding them beyond a walk and it appears they like to focus on the relationship and not worry about getting a job done. I’ve just been writing on that very topic from a Leadership perspective – the limiting mindset of a Leader. Being liked as a boss versus being effective.

    The role of the horse has changed and for many owning a horse is a dream come true. The largest group of new entries into the horse ownership market are women in their 40s & 50s. Clicker training isn’t as threatening and reinforces that grade 5 science class where they learned and still remember, about behavioral science through the rat maze lab. Clicker training is similar to what they know that works with their Pug and it helps them manage their own fear of falling off.

    Clicker training does not take as big a leap of faith as being able to wait on the slightest of changes. The horse does start to seek out the treat. I avoid treating my horses because I don’t want them to turn into muggers with clients. The one positive reinforcement I use with food, I whistle at dinner time and they come running. They wouldn’t hear that clicker 1/2 mile away.

  7. I find the comments here by seasoned horsemen/women to be very disappointing for the following reason: the comments reflect the fact that they actually don’t know much about what clicker training can be. They seem to understand this type of training to be only superficial, rewarding a simple behavior in a mechanical way without the ability to work on anything more meaningful or complex. This can be true if a trainer doesn’t have the skills to use this type of training any further; similar to beginning dog trainers who learn how to lure and reward simple behaviors like a “Sit” by giving the dog a treat when they do it. However, “clicker training” (by the way one doesn’t have to use a clicker, one can use the voice or any other sound) is a means to shape behavior, and can be done on the back of a horse as well as on the ground. It can be used to teach any behavior: correct jumping, correct dressage movements, correct reining movements, fear issues, aggression issues, etc. However a trainer needs to be very skilled, both as a rider, horsemen and in the understanding of timing and horse behavior to reach the best effective levels. West Taylor talks about how pressure is the best way to get a horse to understand….this reminds me of myself when I was first learning how to train. The only thing I really understood and adamantly believed in was how to train through avoidance; in other words, by applying pressure and letting the animal figure out how to avoid it by choosing the desired behavior that I wanted. Really, West Taylor, you shouldn’t make statements about methods that you don’t actually understand and have never tried. Let me add here too that I strongly understand and believe that all horses are not alike and that a truly experienced trainer will have experimented with and understand many different tools, as every tool will have a place somewhere and sometime. Cindy Morin’s comments are probably the most disappointing as they reflect not only a complete lack of understanding of what shaping behavior consists of, they also demonstrate an arrogant disregard for the fact that something she doesn’t understand at all might have merit. Finally let me say that “clicker training” , or shaping behavior isn’t “necessary” in the basic training of horses as they will perform as desired through the application of pressure and negative reinforcement as they have ever since they were domesticated, and “cowboys” such as the above can continue forever training their way. However adding this tool can only heighten a trainer’s abilities when and if they might find it necessary. And , the results and deep relationship that one can find by working with the mind of an animal is incomparable. Finally: at some childish level I would love to see how the above would like it if the only way they could move through the world was by avoiding deliberately applied pressure, all too often applied by someone with poor timing and poor understanding of when to lift that pressure.

    • Great comments Valerie,
      I use pressure as the invitation to solve the problem, much the same as life on this planet. I would love it if every time life put me under pressure all I had to do was run to the fridge to get a treat so I could feel better. That’s just not how life or horse training works for “me”. I have learned that the pressures of life have molded and formed me to who I am today. By me seeing the pressure of life as an invititation to solve my current life issue I have discovered great confidence for myself in life. By my seeking out these solutions in life I have shaped my behaviors to get me through anything this life can throw at me. I simply do the same thing in my horse training. I apply pressure to the horse, not for him to avoid, rather for the horse to “seek out” the answer. When the horse even thinks about looking for the answer I release pressure (my skill set in this is very good) whic then allows the horse to “seek mental relief” not avoid pressure, which in turn creates a Dopamine release within the horses brain rewarding his behavior for “seeking out” the solution. This Dopamine reinforcement will set the horse up to want to “seek out” more and more solutions to pressure. So let’s say my horse is tied up to my horse trailer at the horse show….somebody comes walking by dragging a metal garbage can in the ground making all sorts of racket (pressure), my horse will WANT to seek out the solution, rather than panicking and pulling back trying to escape. My horse will engage in problem solving because he has learned that when he feels pressure it is his invitation to seek a solution, and the solution that will work is to find calm in the chaos! Simply hold your feet still and relax, which the horse will then reward himself with yet another reinforcing dose of dopamine confirming to the horse that he found an approiate solution. I want my horse to be able to problem solve and reward themselves with the natural Dopamine release that comes from transitioning from the Sypathetic nervous system by down regulating to the Parasympathetic nervous system all on their own. No clicker, no trainer, no cookie!!
      Just my humble cowboy way of thinking :).

      West Taylor. 435.680.0946
      Science Based Horsemanship

      • West Taylor,

        Thank you for commenting on my response. Actually I think your method of applying pressure and having good timing in releasing it is an excellent way of training, especially with fear issues. What bothers me the most about this discussion is not the fact that no one interviewed uses “clicker training” but that none of the interviewees seemed to understand the process, only very superficially, and yet are asked to comment on it and discuss it. Right? And specifically to what you do, it’s different I’m sure you would agree when the environment or circumstances apply pressure in one’s life to when an individual decides to apply pressure including how much and when to release it. Someone like yourself with great timing can use this method to wonderful advantage – but all too often there are countless other “trainers” who don’t realize how much they don’t know, and that’s a shame for the horse. Anyway I appreciate your response and understand it. I have no argument with what you do.

  8. My contribution to the dialogue – I completely agree with everything that “F” has just said. Her take on learning theory is 100% correct. And both PR and NR are very useful tools in the tool box. Neither is better, it depends on the task.

  9. This has been a great discussion. This year I attended the Clicker Training Expo to dive a little deeper into the subject how the clicker is used in the positive perideim. I had the pleasure of three full days of learning from highly passionate and skilled trainers. I am always looking to find new tools to add to my base of knowledge. The number one thing I like about the use of a clicker is when used correctly it will mark the exact moment the horse did what you are asking for. I have used this method when my horse is struggling to find the answer to what I am trying to teach. I recently used this method to help my horses foot placement in a yield that I was teaching. I also will pair this method with something the horse finds negative. I had a super needle sensitive horse that was horrible for injections and this method corrected this issue. Clicker training taught me a lot about the shaping of behaviors. Feel, timing and my observation skills also improved. Positive reinforcement is also fun for both the trainer and the horse. I am not a clicker trainer but a person who will use this method with my horses when I feel it would benefit the situation. I love learning and trying new things. I think we all get exposed to various methods and techniques and we get to decide what we as trainers want to use in our programs. This is what is so beautiful. We are passionate and we create wonderful things. So every person is right in this debate. Only each one of use knows what is the right choice for ourself. We choose, we learn, we believe and we all have success.

  10. Hmmm. A lot to comment on here. What I observe is that most of this conversation seems focused on riding. As a functional movement specialist who has been learning to take my decades of training into the equine world, we use “clicker” training to get many behaviors on the ground. I mean how do you get a horse to do PT? I don’t use it much riding, but learning to effectively use +R/-R has very much changed the way in which I use +N/-N.

    The clicker is simply a “bridge” that tells the horse (or dog or cat or…) that yes, that was the behavior I was looking for. If the horse is mugging for treats, then the trainer is not rewarding correctly. One is using the clicker (or sound, hand signal, etc.) to MARK the desired behavior, often from a distance and increasingly delaying and/or even fading the “treat.” Many horses SEEK a requested new behavior like a game which becomes intrinsically rewarding. Call it “tricks” if you will. But who cares if it is achieving the end of creating a new desired behavior, setting up trust and creating a happy horse?

    I have to agree with some of the commentators regarding the bandying about of terminology. Pressure is an added negative (+N) and removal of that pressure is not positive, it’s removal of the negative (-N). Positive reinforcement is giving a reward (+R) and (-R) is removing the object of desire. Positive and negative in this context are not simply opposites of each other. I know. It can be confusing and I may not be explaining it well.

    I just got in a very fearful, highly reactive mustang. Her biggest and only fear so far seems to be being chased. By humans. The last thing I’m going to do to win her trust, is chase her around my round pen. Initially I used the slightest advance and retreat so that she was able to tolerate me in the pen. That’s +N/-N. I ignored her and cleaned her pen. Her bubble got smaller. Last couple of days I was able to extend the muck rake and she took a few steps toward me and sniffed the rake, I took it away. Did this dozens of times. Her bubble got smaller. She stayed calmer. She got used to me moving around. Still all +N/-N, though I would argue that her getting calmer each time is an intrinsic +R. Tonight, on her own, she took hay from my hand maybe twenty times. So we’ve switched to +R. If I’d wanted to punish her for coming to me, I would have taken the hay away. -R, but why would I do that? We’re making progress. She is learning, AND she is stressed. Dr. Peters has talked about the threshold of stress or “discomfort” zone. Somehow there was an implication in the comments that +R/-R means the horse never feels discomfort and is always completely relaxed. Not true. It’s an art to keep any being in the sweet spot. I can’t say I’m real good at it, but I’m working hard on my observational skills. (I read the referenced study about stress and learning in horses and it’s full of holes IMHO.)

    I have the luxury of time. I’m not getting paid to get this horse ready in 90 days. In fact the horse is a TIP fail because she didn’t make the 90 day cut. I can afford to, as West says, “Go slower to get faster.” (Also, the spin on this horse might send me off into orbit and kill me.) Some people might run this horse, rope her, lay her down. Then we could have a discussion about learned helplessness.

    I guess my point is that using a marker and positive reinforcement to get certain behaviors from certain horses at certain times in their development is just another tool in the toolbox. It depends on what you’re trying to achieve and your understanding and skill at using the tool. Like a lot of training tools, it’s often used without skill or understanding.

    I recall a woman at a dog park explaining to me that the electric collar didn’t work on her dog. She kept pressing and pressing the button on the controller and the dog wouldn’t come. I’ve been read the riot act by people about the inhumanity of a small E-collar on my JRT who had the most insane prey drive on earth. I’d only used it a couple of times. But he could enjoy the freedom of being off leash because it helped him remember some semblance of self control. Just a tool in the toolbox.

    PS: there are other studies that show a combination of both positive reinforcement and pressure, ARTFULLY combined are most effective, but I can’t cough them up right now. Of course that is anathema to the 100% +R folks, but hey… just another tool in the toolbox.

  11. Reward based training has its merits and place in the trainer’s toolbox or at least this is so recognized by the Spanische Hofreitschule in Vienna where Lippizaner horse are trained in traditions going back to the 1700’s. Read the last paragraph of Alois Podahskjy’s (Former director of the Spanish Riding School) book “the complete training of horse and rider in the principles of classical horsemanship”. What he describes under “rewards” is paramount to what we call clicker training. Someone who has had the chance to witness the riders of the Spanish Riding School train their horses in groundwork may have noticed the little pouches with treats they carry on their belt and which they distribute sparingly.

    As the effectiveness of “clicker” training depends on the immediacy of the reward after the desired behaviour, food based reward training in the saddle seems at best impractical. But haven’t we all not caressed our horse on the withers after a real good effort?

  12. just one comment…love the conversation. FOOD is the strongest reinforcer of any kind. For all animals, us and them! So using with this in mind is important.
    Also if you ever walk with your horse and stop to “let him or her graze” along with other positive or calming experiences, you are consistently using food as a reinforcer. It is the same concept as clicker training…you are just (or maybe you are) marking the behavior as immediately as you mark it with a sound…that’s all..

    thanks everyone! Stimulating conversation and very useful.

  13. I think we are finding some common ground here! I’ve been thinking a lot about clicker training lately because of all the thoughtful comments, and I’ve asked myself where my less than enthusiastic attitude toward it comes from. The answer has two parts:
    1. I’ve learned about this method from observation, but also from a few clicker trained horses I’ve worked with – mouthy , pushy, confused horses wo made me feel like a human Pez dispenser. I realize that the clicker training was not done correctly in these cases. Judging the entire method based on a few bad examples is like judginig the discipline of dressage based on watching a few riders who go around the arena wth their horses’ noses cranked to their chests – it’s not fair. I have one friend who is a professional dog trainer and also an accomplished horsewoman. She has used clicker training for some behavioral issues with her OTTB, which worked great. She has used the clicker with dogs for many years. She is an expert. Her timing is impeccable, and because of this, clicker training is an effective tool in her toolbox.
    2. What I do with horses actually has a lot incommon with clicker training. I keep negative reinforcement/positive punishment to a minimum. I use food rewards sparingly, but I do use them, mostly when a horse has fear-based behavioral issues, like bolting as soon as a rider gets on. I use other rewards every day, with every horse. My frustrated tone in the original post (I still would not call it “bitching”) comes from discussions with a coulpe of recent clicker training converts who don’t see that I’m not a mean, abusive trainer because I choose not to use the clicker tool.
    So, yes, clicker training can work, but it’s not a panacea. If you choose to use it, you still have to learn how to use it correctly. I also believe it’s mostly applicable to ground work. Once I’m riding, the conversation between the horse’s back and mine becomes more and more involved, to the point where the clicker would just interrupt us.
    Thank you for all the constructive discussion!

  14. A friend who knows I’m pretty passionate about the benefits of clicker training sent me a link to this article. It is obvious that no one who was involved in the interview really understands clicker training, so I won’t get into a discussion about the finer points and the multiple uses of it, under saddle as well as on the ground. I’m just going to make a couple of general statements that I hope will give people pause for thought before making negative statements about a training approach they are unfamiliar with.

    1) There is good and bad training, and it depends much more on the skill of the trainer than on the specific method used. Good training is methodical, consistent, and respects the nature of the horse. Good training *always* results in the communication between rider and horse becoming ever more subtle and complex, regardless of the method used to achieve those ends. Sure there are “clicker trained” horses who mug for food and have learned no complex behaviours. There are also “western trained” horses who go when you kick them with a pair of spurs, and stop (most of the time at least) when you yank back on a harsh bit in their mouth. Neither are representative of a method, they are just bad, incompetent training.

    2) Absolutely clicker training can be used under saddle. It’s not a magic bullet, but once a horse is “clicker wise” he actually learns faster, because guess what – he has learned that he can find the right answer to your question every time, and it’s not going to be super hard, so he puts his mind to learning rather that trying out avoidance behaviours. If you think that sounds a lot like a horse that has learned that there will always be a release from pressure when he makes an effort, you’re right. Good training produces similar results regardless of method.

    3) It’s one thing if you are training horses, and quite another if you are teaching people to train their own horse. Maybe one of the greatest things about clicker training done right is that it teaches the human to be aware of timing and the subtleties of horse body language in a very effective, non-stressful way.

    I dislike having clicker trainers dismiss all traditional horse training methods out of hand without ever having studied any of them enough to understand how they work. I feel the same way about horsemen from other traditions when they dismiss clicker training out of hand. A closed mind and unwillingness to learn is never in the best interest of the human, or the horse.

  15. As I was reading this excellent conversation, I started thinking about the labels of “positive” and “negative” reinforcement. Perhaps it is the connotation of these labels that has become problematic for owners (rightly) seeking a positive interaction with their horses. The positive-versus-negative, black-versus-white, good-versus-bad, connotation of terms used for behavioral reinforcement may be the mental sticking point for owners who equate negative reinforcement with punishment. Now that we have evidence-based neuroscience on our side, maybe we can reframe this concept of ‘positive’ and ‘negative’ reinforcement into an immediate (positive) or a provisional reward (positive) training. Since the relief of mental pressure results in the release of the feel-good substance, dopamine, which is an exceptionally powerful positive reinforcer, I would argue that a provisional reward training system based on pressure and release results in a very powerful positive reinforcement – but only if the horse seeks the appropriate behavior. How is this much different from a food reward (positive reinforcement) for seeking an appropriate behavior and withholding it if that behavior is not forthcoming? I would argue that a horse seeking a food reward and being frustrated when repeated attempts to attain it are not forthcoming becomes mentally stressed therefore, you have set up a negative stimulus for the horse under this situation as well. One can never avoid negative stimuli in an enriched environment. I do agree that there is more fluid conversation between horse and human as well as a more nuanced interaction with the system of provisional reward rather than the ‘positive’ reward system of clicker training. What I am saying is that maybe reframing or rewording the concept of reinforcement is what we are after. To me, conditioned response (clicker) is different than behavioral shaping (“natural horsemanship”).

  16. I’d like to jump in to the conversation as someone who studies and practices clicker training with donkeys! I have been fortunate to attend clinics, webinars and conferences with world-renowned trainers (Alexandra Kurland, Ken Ramirez, Dr. Susan Friedman, Dr. Jesus Rosales-Ruiz, Peggy Hogan, Monty Gwynne) to name a few.

    I agree with comments that effective training lies in the hands of the trainer – any technique can be done poorly. Thus the need for study, consistency and an understanding of the very nuanced approach to the science of operant conditioning that is called clicker or positive reinforcement training. It is anything but simplistic!

    For the donkeys, this approach to training makes them sparkle! They are engaged and stimulated by the opportunity to think for themselves, to problem-solve. They get to make choices and we get to have a conversation! This training method has built a deep and trusting relationship between me and each of them. I have used it to ease us out of some potentially “scary” situations on the trail, to load them into trailers, to calm and refocus during vet visits and husbandry procedures and to relieve boredom during inclement weather. We do obstacle course work, hiking and interact with visitors of all ages.

    Of course we also use pressure and release, as does anyone who uses a halter and lead rope – one method does not negate the other. The common element is timing – precise release of pressure, precise marking of behavior (BTW, I use a tongue-click, so no need to carry a mechanical clicker!)

    One final comment – my job as a trainer is to avoid my animals feeling frustrated! I do this by breaking my lesson plan down into tiny steps – building blocks that the individual animal can understand. If he or she does not understand or shows any frustration, the responsibility is mine, not theirs and I must back up and find an even smaller step to reinforce. My animals never mug me for treats – that’s the very first thing I teach and that behavior never gets reinforced so they drop it easily. I am continually amazed at how much we have learned but also how much there is still to learn – it’s been a fascinating journey and way more effective than any other method I have tried.

  17. Clicker training isn’t really about the clicker, it is the best form of communication we have to date with animals. We ask a question, they try to answer, and we say ‘yes’ for any approximation with a food reward. Powerful stuff for an animal that eats 70% of the time. With only 1 cup of food, doled out in pieces the size of a nickel, horses will do astonishing things that can’t be trained any other way. We work closely with dog trainers, and zoo keepers because science tells us all organisms learn the same way. That being said, it always comes down to the individual animal in the end. The horses love it, because every try is rewarded (you like getting a pay check right?) and they never get punished. We work to substitute unwanted behavior with something that we do want. They will go above and beyond, and practically beg to work. It is the only form of training where they will offer the behavior without being asked. We literally change the chemistry in their brains. Horses trained with pressure and release will always give you minimum effort (just enough to make the pressure go away) because there is nothing in it for them. Most clicker trainers know how to train with traditional methods, we all started at a different place.

    • Exactly! It makes sense that clicker training works for all animals, including exotics like tigers. Yet try to use negative reinforcement on a tiger and you’d probably get bitten.

  18. My disappointment with this discussion is that it didn’t include a single person who actually uses or more importantly understands clicker training. Some of the participants don’t understand the definition of negative and positive reinforcement. That’s certainly problematic. There are plenty of people to ask about clicker training. It’s hardly unheard of at this point. Please ask those that have experience with it and an opinion with relevant knowledge to to back it up.

  19. How unfortunate that you didn’t inform yourselves about equine learning theory and positive reward based training (clicker training) before writing this article. All you have demonstrated here is your ignorance about this topic, and your ignorance of your own training style. Of course Your horse is seeking “mental relief” , from the aversive you are applying, pure and simple. That is not a good thing. Research shows that positive punishment and negative reinforcement (what you folks are using) are detrimental to the psychological wellbeing of horses . Don’t rely on the few clicker trainers you know, or 1 study, look at the multitude of research that is out there.

    • Hi Catherine and thank you for your comment. Please share the research you mentioned. Negative reinforcement is more effective to learning as we cited in research for this article. Thanks again.

      • No it isn’t – except in the specific stress circumstances provided during that study. The study concluded that IN THOSE CIRCUMSTANCES negative reinforcement is more effective. Of course it is. Any animal under stress doesn’t want to think they just want to react. The first lesson an animal has in positive reinforcement training is that stuff will only happen when he’s ready for it to happen and not when he’s forced into it

  20. Wow I have to say I agree with all the commenters that are disappointed that no one with clicker experience was interviewed. And it does seem that many of the people interviewed really don’t understand positive reinforcement and learning theory in general. Although I do appreciate Katrin’s comment that she shouldn’t judge it on her personal experience of a few clients with poorly behaved horses- that’s exactly it. You guys should all check out Georgia Bruce and Rumby the wonder horse to see a brilliant example of how amazing clicker training can be when it’s done correctly! Rumby can do some very advanced dressage moves and his attitude to training sessions is one of eagerness and willingness.

    Also to the admin- your linked article only says that positive reinforcement didn’t work as well as negative reinforcement when the horses were stressed. It doesn’t look at any other scenario, just when they are under stress. And guess what! Most animal behaviourists know this because when animals are checking out potential threats or feel threatened they stop eating, they go into the sympathetic nervous system and ready themselves to fight flight or freeze.

    I have to add that I agree there are many militant +R trainers who tell people that they are being cruel to use -R or pressure/release methods and I disagree with that. Good training is good training and you can tell who is a good trainer because they will have a willing and calm horse who trusts them and who the my can communicate with effectively. That being said I think +R really adds to the positive feelings a horse has for you and it can greatly increase motivation and increase learning speed. As an ex zookeeper it is clearly demonstrated as an incredibly powerful teaching tool because there are so many animals that you can’t use much -R or punishment with because they are too dangerous or completely behind bars. It’s amazing to see tigers having a voluntary blood draw from their tail or to see rhinos calmly walking onto the scales for weighing or giraffes having ultrasounds to see it they’re pregnant! Very powerful training technique and I’m pleased that more people are interested in it in the dog and horse world.

  21. Can you explain your comment “Negative reinforcement is more effective to learning as we cited in research for this article”? You mention this in several ways and at various points.

    The study quoted concludes:-
    “. Our study shows that exposure to stressors before testing has a negative effect on learning performance, mainly under conditions of positive reinforcement. In addition, learning performance appears to be differentially related to personality according to the type of reinforcement and the presence of extrinsic stress. In the absence of stressors unrelated to the task, the most fearful horses were the best performers when they learned with negative reinforcement but the worst when they learned with positive reinforcement. When stressors unrelated to the task were applied, the most fearful horses were consistently the worst performers, particularly with negative reinforcement learning.”

    This is a study concerned with the effects of stress on learning. It says nothing about which form of learning is “more effective to learning”. All of the quadrants work, it’s why Skinner describes them.

    As others have noted, the reason for precise descriptions of behaviour is that it facilitates discussion. You need to learn more about the behavioural terms you are utilising and understand what they are describing. Whatever form of training you utilise, the ability to analyse the process clearly will be invaluable. It prevents superstitious behaviours on the part of the trainer and stops emotional interpretations of behaviour.

  22. There is ONE study that shows negative reinforcement more effective and it gets trotted out every time (see the equine reference there? Lol)
    This one study found that horses stressed or in pain or fear didn’t learn from positive reinforcement training but they could still learn with negative reinforcement training. This is why so many trainers diss positive reinforcement training – because their normal methods stress horses. I use both positive and negative reinforcement but I understand how each works and I prefer positive.
    Positive reinforcement training (note I don’t say clicker ;because that’s just shorthand and many positive reinforcement trainers don’t use one) works on people, dogs,cats, mice,rats, wild animals in zoos, honey bees and some research shows it may work on plants and individual neurons in a petri dish. Why would it not work on horses?

  23. In general, I’m not in favor of rewarding with food. It may be a shortcut to get the behavior you want; but it is not leadership. It’s a bribe. Ultimately, we want to assume the position of leader for our horses; we want them to see us as a safe place, as someone they can trust, as someone they respect. I want my horse to do what I ask because I ask it, not because she wants the food reward. And, if one day I need her to, for example, load in a trailer and I don’t happen to have a treat handy…then she learns that I am inconsistent, sometimes there will be a treat, sometimes not. Maybe she’ll still load even if she only gets the treat occasionally; but I don’t ever want my horse thinking “sometimes she does, sometimes she doesn’t” about me. That I am inconsistent, in anything. That inconsistency is okay sometimes. That is only one step away from “sometimes I have to do what she wants, sometimes I don’t”. I believe if I want 100% consistency from my horse, I have to be 100% consistent with her. So, I don’t use food as a training tool.

  24. Shocking how misinformed this article is! Shocking that you feel comfortable publishing this!
    My relationship with my horse has improved exponentially since I have started CT. Finally I found a way of communicating to him that serves us both well. He is learning that he can have an opinion & offer suggestions on what & how he wants to interact at any given day, within the boundaries that we have both set for each other.
    As writers & publishers of information, you have a responsibility to educate yourselves. I challenge you to go find a trained CT trainer, take some lessons, and then write a new article. Oh, & learn some behavioural & learning theory. Not only is it facinating, but you will be able to correctly speak about it.

  25. “No horse will be psychologically scarred from an occasional correction in the form of a negative reinforcement.”

    This is not negative reinforcement. Corrections would be termed positive punishment within the use of operant conditioning.

    In scientific terms related to operant conditioning (BF Skinner);

    Positive is the addition of stimuli.
    Negative is the removal of stimuli.

    Reinforcement of a behavior promotes / increases the repetition of the behavior.
    Punishment of a behavior extinguishes / decreases the repetition of the behavior.

    *Positive reinforcement = the addition of a stimuli to help promote the increase of the behavior. The most thought of is probably clicker training (using the clicker as a bridge for a food reward after a behavior is performed).

    It is interesting to look at feeding time in a typical barn with stall kept horses. The feeding routine starts, horses start pawing at doors, or twirling around in circles in anticipation of the food (a desired or pleasant stimuli). The horse gets it’s grain dumped through the bars into a bucket for them to eat. So whatever behavior the horse has started to perform at feeding time gets reinforced (Positive reinforcement) when the grain hits the bucket and the horse is allowed to eat.

    A foal will root it’s nose up into it’s mothers udder (the behavior) and the milk will get released (the stimuli that will cause the behavior to be reproduced).

    *Negative reinforcement = the removal of a stimuli to help promote the increase of the behavior. Example, using your leg aid to get a behavior like walking from a halt, then removing the leg aid when the horse does walk.

    Positive punishment = the addition of stimuli to decrease the repetition of the behavior. Example, the electric fence.

    Negative punishment = the removal of stimuli to decrease the repetition of the behavior. Example, if your dog is pawing you for attention and you walk away. This is also the flip side of the positive reinforcement coin. Withholding something desirable (food item given for positive reinforcement) can become a punishment if your timing / your ability to be congruent suck.

    This science has to take into consideration how the operator (the horse in this case) truly feels. So within the operant conditioning quadrant there is quadrant within each quadrant. Is the stimuli perceived by the horse as it is intended? Is it effective? Does it increase the behavior or decrease the behavior as wanted?

    We may think a scratch is ‘rewarding’ and appreciated by the horse, so we believe we are using the positive reinforcement quadrant . But many horses do not like our delivery of scratches or pats, so to the horse it is aversive. Is the stimuli perceived by the horse as WE intended? NO. Will it help increase the behavior? Probably not.

    Going back to the electric fence, and corrections for behaviors, or positive punishment. Do most horses perceive the fence as aversive, as intended? Yes. Do most horses stop the behavior of touching the fence? Yes. So that would be an example of that quadrant working as it should.

    But there are empowered and disempowered states of the quadrants, because there is the relationship of the one giving the stimuli to the once receiving the stimuli. There are many horses that will stand near their electric pasture fence very relaxed, because the consequence of touching the fence is immediate and very clear, and the horse is in control of touching the fence. If a human stood at the fence with a hand held cattle prod, the end results probably wouldn’t be desired for human or horse.

    This is definitely a deep and fascinating rabbit hole. 😉

Leave a Reply to Mary Ellen Spaite Cancel reply

Your email address will not be published. Required fields are marked *