WRITTEN BY: HARRIE PHILLIPS
PGCertClinEd, BAdVocEd (VocEd&Trng), DipVN (Surgical, ECC), DipBus, DipTAE (Development & Design), TAA
PUBLISHED: 18 May 2026
If equitation science tells us that positive reinforcement techniques like clicker training can be used to teach the highest movements of dressage, including piaffe and passage1, why is there such reluctance in the regular horse world to embrace positive reinforcement?
The Spanish Riding School in Vienna has built food rewards into its training method so deliberately that special pockets are sewn into the riders’ tailcoats so that sugar can be carried into the arena and given to the stallions after work2. And yet, in stables and arenas across the everyday horse world, owners are still being warned by their instructors that food rewards will ruin a horse. They will make them pushy. They will make them nippy. They will make them refuse to do anything without a treat in sight.
The horse world feels decades behind the rest of the animal training world. Marine mammal trainers, dog trainers, zoo trainers and laboratory animal handlers all incorporate systematic positive reinforcement into their core practice. Why are we so slow to catch up, and what is this reluctance costing the horses we work with?
So, why do food rewards still get such a bad reputation in our industry, what does the science actually say?
Why do horse owners think treats make horses pushy?
The myth goes like this. Horses are large prey animals with strong food drives, and using food rewards in their training inevitably produces horses who become pushy, demanding, and dangerous in their pursuit of treats. The horse mugs you for food. The horse refuses to do anything without a treat in sight. The horse becomes nippy and disrespectful. Far better, the argument goes, to keep food out of training entirely and rely on pressure-release and a good firm handler.
Most of the horse owners who hold this view have seen actual pushy horses around food. They are not making it up. These pushy horses are real. What they are wrong about, is the cause.
What does the research say about positive reinforcement in horses?
When food-based positive reinforcement is applied with appropriate protocols, the evidence base supporting it as successful is quite substantial. Hendriksen, Elmgreen and Ladewig compared positive and negative reinforcement for trailer loading in twelve horses with severe loading problems. Horses trained with positive reinforcement showed significantly less stress-related behaviour across all four observed categories (avoidance, nostril widening, eye widening, and tail movement) and required significantly shorter training sessions, with no significant difference in heart rate between the two groups3. Innes and McBride, comparing the two methods in rehabilitated ponies, trained to lead in hand over a seven-week period. They found similarly lower stress responses in the positive reinforcement group, while honestly noting that food anticipation can produce arousal which the trainer needs to manage4.
The benefits are not limited to the immediate training session. Hockenhull and Creighton found that horses whose handlers used food rewards in training had significantly fewer ridden behaviour problems, and that there was no evidence linking positive reinforcement to increased biting or pushy behaviour5.
It’s clear, the myth and the data go in opposite directions.
Sankey and colleagues demonstrated that horses trained with positive reinforcement retained positive associations with humans up to eight months after training had ended, with no further interaction in between, suggesting that the relationship effects of well-applied positive reinforcement are durable rather than fleeting6. Imagine that being applied to rescue horses, problem horses, brumbies being trained for the human world.
The pattern across the literature seems very clear. Positive reinforcement, applied with thoughtful and correct protocols, produces horses who are more willing, less stressed, more reliably trained, and better-disposed toward their handlers than equivalent work using pressure-release alone. So, why aren’t we using it more?
But, none of this means the method is foolproof. A poorly executed positive reinforcement protocol will produce a poorly trained horse, just as poorly executed pressure-release will. The variable that matters in both cases is us, the humans.
Pushy is a training error, not a method error
Mykle is one of the most food-driven horses I have ever owned. He starves on a daily basis, just ask him. When offered the prospect of a treat, he will work through every behaviour in his repertoire, good and bad, in rapid succession, on the off chance that one of them will produce the treat. He is, on paper, exactly the horse the “treats make horses pushy” myth predicts will become unmanageable on positive reinforcement.
But, he hasn’t, because the way I train him is built around the science, not around feeding the food drive.
Mykle’s primary reinforcer is not food. It is a well placed scratch — he loves human interaction so it works for him. The best reward is the thing that the horse finds rewarding and it may not be what we think it should be. For the vast majority of his training, including all maintenance work on behaviours he already knows, the reward is a scratch and a quiet word. Food comes out for new behaviours, for very high-value skills I am still building, and for moments where I genuinely need to override an established response with a stronger reinforcer. Once a new behaviour is established, food is phased out and the scratch takes over again. The food reward is a tool I deploy deliberately, not a default.
Recent neuroscience research on dopamine and learning supports this kind of phasing approach for reasons beyond simply managing a food-driven horse. The brain’s dopamine response, which drives the strength of learning, is shaped not only by the reward itself but by whether the reward is predicted or surprising7. Rewards delivered in highly predictable patterns generate weaker dopamine responses than rewards that vary in timing and value. Reducing food frequency as a behaviour becomes established is therefore not just about preventing pushiness; it is also how you keep the reinforcement working at full power.
This is the part of positive reinforcement that the myth fundamentally misunderstands.
Pushy behaviour is not caused by the use of food. It is caused by the use of food in response to demanding behaviour.
If you reward the horse who is mugging your pocket with a treat, even an apparently unrelated treat given a moment later, you are reinforcing the mugging. If you reward the horse for offering calm, polite behaviour and reliably do not reward demanding behaviour, you build a horse who offers calm, polite behaviour. The horse is doing what works for them. You get to decide what that should look like.
The same principle applies to inconsistent timing. A trained cue is a learned response that has to be maintained through consistent reinforcement, whether the reinforcement is pressure-release or a food reward. Inconsistent delivery of a reward, whatever that is, produces inconsistent behaviour.
A food-sensitive horse like Mykle is not a contraindication for positive reinforcement. He is a horse who requires me to be careful, thoughtful, and consistent in my use of the method. The food sensitivity is a feature to manage, not a defect that rules out the technique.
What is clicker training and how does it work?
Food based rewards are most commonly associated with clicker training and the like. The clicker is a marker. Specifically, it is a bridge signal: a clear, distinct, consistent sound that tells the horse, at the precise moment of correct behaviour, that what they just did was right and that a reinforcer is coming. The mechanism is operant conditioning paired with classical conditioning, with the click acting as a conditioned reinforcer in its own right after the horse has made the association between the click and the food.
To understand why a marker matters, it helps to understand the contiguity window. For an animal to associate a behaviour with a consequence, the connection between the two has to occur very quickly8. Even a brief delay weakens the link, and the horse starts learning something other than what the trainer intended. In practice, you cannot deliver a piece of food from a pocket to a horse’s mouth quickly enough to keep the operant link clean, and that is the problem the marker solves. The click happens at the exact moment of correct behaviour, and the click itself buys you the time you need to deliver the food. The horse hears the marker, knows the food is coming, and the operant link between behaviour and consequence stays intact.
The clicker itself is not magic. Williams and colleagues tested 60 horses on an operant task and found no significant difference in either acquisition or resistance to extinction between horses receiving a clicker plus food and horses receiving food alone9. A clear verbal marker, a short word used consistently and only in this context, does the same job. Some trainers prefer the click because of its distinctiveness and consistency; others prefer a verbal marker for its simplicity. The choice of marker is a matter of practical preference, not of mechanism.
What the clicker does well is buy you precision. It can mark the exact moment of correct behaviour with millisecond accuracy, well ahead of your ability to deliver the food reward. That precision is what makes the technique powerful for shaping fine behaviours, and it is why marker-based training has been so effective across species where precise behaviours need to be marked at exactly the right moment.
How do you start clicker training a horse?
Clicker training is real training. It depends on the same three things that any good training depends on: clarity of the cue, good timing, and consistency. The method works when those three things are in place and falls apart when they are not, like all methods.
A few practical points about how the method actually starts.
Plan how you will deliver the food before you start. Where the food comes from, where you deliver it relative to your body and the horse’s body, and how the horse is expected to receive it. Most pushy problems are made or avoided in this step. Delivering food away from your pocket, slightly away from the horse’s body, with the horse expected to remain in position rather than move toward you, prevents mugging from being reinforced before it ever becomes a habit.
The clicker has to be charged before it means anything. Out of the box, the click is just a sound. The horse has to learn through repeated pairing that the click predicts food, so that the click itself becomes the marker. This is straightforward classical conditioning. You stand with your horse, click, and immediately deliver a small piece of food, repeated many times across short sessions. Within a session or two most horses will visibly anticipate the food on hearing the click, and the marker is now ready to do its job. It is worth noting that the horse is being reinforced for whatever they are doing in those early sessions, which is why you charge the clicker while they are standing calmly and quietly. Calm presence becomes the foundation behaviour from the very first repetition, and the pushy problem is pre-empted before training proper has even begun.
Plan the session, do not wing it. Short sessions, one behaviour at a time, clear criteria for what you will mark, and a stop point you decide in advance. Five minutes of focused, well-planned work will produce more reliable learning than thirty minutes of drift. The horse cannot read your mind, and unclear criteria during a long session is how cues become diluted and behaviours become inconsistent.
Start with a behaviour the horse already offers easily. Targeting, the horse touching their nose to a held target, is the most common starting point because it is something most horses will offer naturally within a few seconds. From there you can shape almost any other behaviour. Trying to capture a complex behaviour from scratch is one of the most common reasons clicker training fails, because the criterion for the click becomes unclear and the horse stops being able to work out what is being marked.
Build complex behaviours through small steps. Successive approximations, not leaps. The horse offers a small piece of the behaviour, you mark and reward, the next click waits for slightly more, and so on. The art of this is in the splitting, knowing how small a step to ask for next, so that the horse keeps succeeding and stays engaged. This is shaping, and it is the same shaping principle that underlies any well-built behaviour, including the ones traditionally trained with pressure-release.
None of this is unique to positive reinforcement. Clarity of the cue, timing, and consistency are what make any training method work. Clicker training is simply one of the more honest tests of those three things, because the precision of the marker reveals very quickly whether you have them or not.
Where positive reinforcement shines
The horse who has let me refine my rusty clicker training the most in recent years is Tedi Ber, the first Fjord I bred. Tedi Ber was intended to be my future riding horse. He is not going to have a ridden career, because his ability to “horse” and injure himself is very strong, but he is here, he has a working brain, and he really needs a job.
The job we have built together is clicker training tricks. Tedi Ber and I have nailed target training, “smile”, and spin around the handler, and we are close to a reliable bow. We will keep adding skills over time. The work is not preparation for anything ridden, because there will not be a ridden career to prepare for. The work is the work. I can train anything I can imagine and not inadvertently damage something he’d need for a ridden career.
What I have noticed is that the work is doing several things that the literature predicts. His ability to focus on me has deepened. His willingness to try things, to offer behaviours, to engage with the puzzle of what I am asking, has increased. Our relationship was already strong; it is stronger now. And, almost as a side effect, his responsiveness to my cues in everyday handling, things that have nothing to do with his trick repertoire, has noticeably improved. He leads better, stands quieter, and pays closer attention to what I am asking of him. Others have commented that he appears more relaxed and content in his overall demeanour, compared to before we started clicker training.
For a horse like Tedi Ber, this is not a consolation prize. The clicker work gives him a job, an engaged brain, and a real partnership with me, none of which he would have as just a paddock ornament. Engagement with us and a brain being used are valuable for welfare, regardless of whether a horse will ever have a ridden career.
Can positive reinforcement be used for advanced dressage?
It is sometimes assumed that positive reinforcement is fine for tricks and basic husbandry but does not scale to serious technical work. The assumption is wrong on both the science and the practice.
McGreevy and McLean discuss the application of clicker training to the higher dressage movements including piaffe and passage, taught first in-hand and then transferred under saddle1. The principle is the same as in trick training: a precise marker captures the moment of correct effort, the reinforcer follows, and successive approximations build the movement up over time. Belynda Moore, an FEI dressage rider and USDF bronze, silver and gold medalist, has publicly described using the clicker for the piaffe specifically, on the grounds that the movement can be very stressful to the horse, who can feel trapped in it if the rider is not careful10.
If positive reinforcement can refine a piaffe, it is not a parlour trick limited to teaching ponies to smile for treats. It is a valid training option that scales from basic behaviours up to the most demanding technical work in the equine athletic disciplines. The argument that the method is somehow beneath serious training, or only suitable for hobbyists and pet owners, is fundamentally wrong.
Where to from here
Used well, positive reinforcement is one of the most powerful tools available for the work you and your horse are doing together. It is also one of the kindest. The next time someone tells you that treat training will ruin your horse, you might gently ask them what evidence they have, and where it came from.
Read next: Your Horse is Always Learning. The Question is What.
References
- McGreevy, P., & McLean, A. (2010). Equitation Science. Wiley-Blackwell. ISBN 978-1-4051-8905-7. View on Google Books
- Vienna Tourist Board. The Spanish Riding School in Vienna: Lippizzaner horses at the Hofburg. Available at: www.wien.info/en/art-culture/imperial-sights/spanish-riding-school-359386
- Hendriksen, P., Elmgreen, K., & Ladewig, J. (2011). Trailer-loading of horses: Is there a difference between positive and negative reinforcement concerning effectiveness and stress-related signs? Journal of Veterinary Behavior, 6(5), 261–266. https://doi.org/10.1016/j.jveb.2011.02.007
- Innes, L., & McBride, S. (2008). Negative versus positive reinforcement: An evaluation of training strategies for rehabilitated horses. Applied Animal Behaviour Science, 112(3–4), 357–368. https://doi.org/10.1016/j.applanim.2007.08.011
- Hockenhull, J., & Creighton, E. (2013). Training horses: Positive reinforcement, positive punishment, and ridden behavior problems. Journal of Veterinary Behavior, 8(4), 245–252. https://doi.org/10.1016/j.jveb.2012.06.002
- Sankey, C., Richard-Yris, M.A., Leroy, H., Henry, S., & Hausberger, M. (2010). Positive interactions lead to lasting positive memories in horses, Equus caballus. Animal Behaviour, 79(4), 869–875. https://doi.org/10.1016/j.anbehav.2009.12.037
- Schultz, W. (2016). Dopamine reward prediction-error signalling: a two-component response. Nature Reviews Neuroscience, 17(3), 183–195. https://doi.org/10.1038/nrn.2015.26
- McGreevy, P. (2012). Equine Behavior: A Guide for Veterinarians and Equine Scientists (2nd ed.). Saunders Elsevier. ISBN 978-0-7020-4337-6. View at Elsevier
- Williams, J.L., Friend, T.H., Nevill, C.H., & Archer, G. (2004). The efficacy of a secondary reinforcer (clicker) during acquisition and extinction of an operant task in horses. Applied Animal Behaviour Science, 88(3–4), 331–341. https://doi.org/10.1016/j.applanim.2004.03.008
- Lotz, K. (2024). Clicker Training Your Horse. Young Rider Magazine. Available at: www.youngrider.com/clicker-training-your-horse
- Jones, J.L. (2020). Horse Brain, Human Brain: The Neuroscience of Horsemanship. Trafalgar Square Books. ISBN 978-1-57076-948-1. View at Trafalgar Square Books
Every due care has been taken to ensure the information herein is based on sources Veterinary Nurse Solutions believes to be reliable, but is not guaranteed by us and does not purport to be complete or error-free. As such, we do not warrant, endorse or guarantee the completeness, accuracy, and integrity of the information. You must evaluate, and bear all risks associated with, the use of any information provided hereunder, including any reliance on the accuracy, completeness, safety or usefulness of such information. As part of our quality control of information contained within this document, it has been peer-reviewed by qualified animal care professionals.
Veterinary Nurse Solutions acknowledges that there is more than one way to carry out many of the tasks described within this website, and techniques omitted are not necessarily incorrect.