I started writing this article for two reasons. The first was that I found myself putting down some thoughts in response to some questions that were being asked in an on-line forum. There was a discussion about the difference between using clicker training to teach new behaviors and using clicker training at other times. The second was that when I went looking for resources to help explain how to transition from teaching new behaviors to integrating them into your horse’s repertoire, I couldn’t find much of anything.
There’s a lot of information available on how to use clicker training to train new behaviors or modify existing behaviors. There’s less information available on how to integrate these new behaviors into your daily routine or current training sessions. And even the useful tips that I have read over the years don’t outline the whole process. The information I’ve read ranges from “click what you like” to “don’t use a behavior until it is completely trained, proofed, and under stimulus control.” I actually like and have used both these pieces of advice, but I recognize that they are not really enough information to know how to get from point A (just starting clicker training and have a few behaviors) to point B (mixing in some clicker trained behaviors and working toward using it more).
So I decided to write down some thoughts that might help a new clicker trainer find their way to from point A to point B. What I am going to present here are some suggestions and guidelines based on my experience. The information is presented to help you make some choices and find the progression that works for you and your horse. I find that every horse is a little different in how they learn and what level of direction and support they need. Some horses benefit from a lot of structure and consistency. Other horses are better off when the training is discretely mixed in with other activities. And this can change over time so it’s important re-evaluate periodically to make sure the current system is working well.
STRUCTURE OF YOUR TRAINING SESSION: CLASSROOM VS. REAL LIFE
Basically, there are two ways I set up my clicker training sessions. One is to set aside a specific time slot to work on it and the other is to train it little by little as I ask for the behavior as part of my daily routine. Training sessions for these two strategies will look very different and I am going to give them the names “classroom” and “in real life.” I use these terms mostly to describe the “structure” of the training session, but there is a location component too, especially in the early stages of classroom sessions.
The title of this article is “Taking clicker training from the classroom to real life” so that should be a tip-off that both types of sessions are used with most horses and that I usually start with in the classroom and progress to training in real life. But this is not the only way to do it. In some cases, clicker training is introduced in real life training sessions and then individual behaviors are spruced up in the classroom as needed. I’ll write a lot more on this later, but I wanted to mention it here so you were not reading the next section thinking you had to choose one type and stick with it.
The word “classroom” comes from Alexandra Kurland who recommends that new trainers set up a “clicker classroom” which is a dedicated training space where you can safely work with your horse. You start out behind a barrier (protected contact) and later you can use the same space as a classroom you share with the horse to work on exercises like head lowering, backing in a square, targeting, etc…
She is using the word classroom to refer to a physical classroom. This is probably close to what many people would consider a formal training session. It’s how I teach many new behaviors. I have a plan that clearly defines the behavior I want for that session, how I am going to get the behavior, a way to keep track of correct responses (if needed), when to shift criteria, what to do about errors or resets, and as many other details as I can think of ahead of time. If I am working on improving an existing behavior, I still follow the same general outline (have a plan, set criteria, etc…). When I am using a classroom type structure, I am focusing on one behavior at a time, or simple loops.
In the beginning, it’s helpful to work in a physical classroom because It is easier to teach a new behavior in a controlled environment. Horse trainers can’t always control the environment as easily as other animal trainers and so we end up training in much less controlled environments than other animal trainers that I have seen. Dog trainers are often able to train inside in a distraction free room. Even if you train your horse in his barn, there are liable to be more distractions than if you were training in your living room (assuming your horse was used to being in there!).
I have noticed that some people are so used to this (working in less than ideal conditions) that they don’t realize how a good training space makes it easier for you and the horse to have more productive sessions. It’s really worth the time and effort to create a training space and then set aside some time to use it. A short dedicated session in a quiet environment can make the difference between getting a reliable behavior in a few sessions vs. working away at it with mixed results for weeks.
In summary, in a “classroom type” session:
- I am working in a controlled environment
- I am working on repetitions of one behavior, or a small set of behaviors
- I have planned ahead of time so that I know how to get the behavior, to collect data, handle errors, and when to take breaks
- I have set criteria based on previous sessions
- I have thought about my stimulus control, and in the early stages of a new behavior, I want the horse to repeat what was just clicked
- I am teaching a new behavior, modifying an existing behavior, or working on stimulus control
The other type of structure is what I am calling “in real life.” When I am using clicker training in real life, I am clicking based on what information my horse needs at that moment. There will be a mix of clicking for behaviors that I have already trained, or for existing behaviors that I want to strengthen, or I could even be capturing behaviors. This type of training tends to be a little more variable in that I might click for many behaviors over a short period of time and they might all be at different stages of training. I will also be working in lots of different environments. Using clicker training in real life looks more like “clicking for what I like,” but if done well, it’s not just random clicking. I am going to write a lot more about “real life” sessions later, but I wanted to introduce them here so you could compare them to “classroom” sessions.
In summary, in a “real life” session:
- I am clicking and reinforcing behaviors as part of my daily routine (feeding, leading, grooming, etc…)
- I am clicking and reinforcing many different behaviors over the course of a session
- I have planned ahead so I am taking advantage of environmental cues and training opportunities as they present themselves -or I have at least thought enough about them so that I recognize “training opportunities.”
- I have planned ahead of time so I am thoughtful about which behaviors to ask for where.
- the click means “yes, that was the right answer,” but does not necessarily mean “repeat what you just did.”
- I am usually modifying a new behavior, working on stimulus control, proofing or generalizing
In both types of sessions, I mention that you should have a plan. I don’t want to get too much into the nitty-gritty of plans because that is a topic on its own, but I would like to say that I think a little bit of advance planning goes a long way toward setting you and your horse up for a successful training session. The type of planning I do for a classroom session is going to be different than the type of planning I do for a real life session because they are structured differently and I often have different goals.
Plans are useful in different ways. One of the benefits of making a plan is that it requires you to sit down and think about what you really want and that makes you define the behavior more clearly. It also requires you to think about ways to train it. I find that as soon as I start working on a plan for training something, I start to observe my horse a little differently. I might notice other related behaviors or ways I could use the environment or existing cues as part of my new project. So a big part of planning is the thinking and observing that have to happen as part of the planning process.
The last thing I want to say about plans are that a good plan is a starting point, not a rigid set of instructions you have to follow. In any training session, you always have the option of stopping if you see you need to make adjustments or changing things as you go if you are comfortable with that. Ken Ramirez has a good saying and I can’t quote him directly, but it’s something about the importance of making a good plan, but being willing to throw it out the window if it’s clear that it is not working for you and your learner.
The last bullet point in each summary describes an important element of each training session. This is the focus of the session. When I am clicking and treating a behavior, I will have better and more consistent results if I identify what aspect of the behavior I am trying to improve. The focus can be about the behavior itself, how I want to use it, or where I want to use it. In training a new behavior, there are some steps that are more suited to a classroom type structure and others that are more suited to a real life type structure. Where you are in the training process will help you choose the type of session that will be the most successful and affect how you shift from classroom sessions to real life sessions. Before we get into that, I want to review the steps that I follow when I am training a behavior (from starting a behavior to having a finished behavior.)
FOCUS OF THE TRAINING SESSION: WHICH STEP ARE YOU WORKING ON?
The first step in clicker training is usually about getting the behavior. This could be shaping a new behavior and/or getting desirable behavior to happen more frequently. It’s the part of clicker training that gets the most attention. But that’s only the beginning of a much longer process.
If I write out the steps that describe the complete training process for a single behavior, it looks something like this:
- Get the behavior (shape it or capture it)
- Modify the behavior (through shaping)
- Add a cue
- Generalize (practice in new situations)
- Proof it (test the cue)
- Put it on a variable reinforcement schedule or start using different types of reinforcement/reinforcement substitutes
Not every behavior will follow this progression in a linear fashion and there is always some back and forth between steps as I redefine a behavior or realize I need to work more on some element that I thought I had already addressed. It’s not unusual for me to have a working cue from the very beginning so I can skip step 3. If I capture a behavior, I might skip step 2. Steps 4, 5, and 6 are about taking the behavior from the classroom to the real world so there might be some variation in the order or mixing of generalizing and proofing. I’m not so concerned with us all agreeing on the order in which things are done, as much as in pointing out that there are different steps that take a behavior from its initial stages to being a useful and fluent behavior.
I want to add something to this list, but it doesn’t fit neatly into any one step because it happens throughout this process. It is called building the “reinforcement history” for a behavior. The more times I reinforce a behavior (regardless of which step I am on), the stronger it gets. It happens as a natural by-product of the process, but I can also take time to work specifically on it if I think there would be a benefit. With these steps in mind, let’s see how I use combinations of changes in structure, focus, and location to move from working in the classroom to working in real life.
PROGRESSION FROM CLASSSROOM TO REAL LIFE: Change of location, change of focus, change of structure
I meet a lot of people who have trouble transitioning from the classroom to real life. Or they keep clicker training in the classroom and only use it for very specific aspects of horse training/handling and miss out on the many other ways that clicker training can improve your horse’s life (and yours!). Part of it is that the road from the classroom to real life can be kind of bumpy and if you don’t have someone to encourage you, or give you pointers, it can seem easier just to put clicker training back into that little box labeled “classroom.”
The good news is that once people see how effective clicker training can be, and how much their horse likes it, most people do want to try using clicker training in real life. My new students are happy to stay in the classroom for a while, but at some point they usually ask “when can I start asking for the behaviors I am training in other situations?” Please note that this question is about ASKING for behaviors, not about clicking and treating behaviors. There is a difference. Right now I am going to talk about taking behaviors from the classroom to real life. In the other case, the training is more focused on capturing or strengthening existing behaviors, which is also a valid way to incorporate clicker training into real life and I’ll get to that in a little bit.
There are lots of ways to move from the classroom to real life. If you go back and look at the summaries for classroom vs. real life, you can see that there are differences in location, structure, and focus. Changing one of these will often lead you naturally to changes in the others and if I do this in little steps, I can make the transition smoothly. The idea is to change things a little bit at a time so that I am asking new questions, but also keeping enough things that are familiar so that both the horse and I remain in our comfort zones.
Before I start the transition from the classroom to real life, I make sure that the horse has enough understanding of the behavior and has appropriate manners around food so that it’s reasonable to think of taking the next step (there’s more on this later). This does not mean that the behavior is “finished.” It just means that the behavior is solid enough that I can ask for it and get something that meets my basic criteria. I am going to list some specific examples below, but first I want to provide a very general progression:
- teach the behavior in the classroom: classroom structure in classroom setting
- practice the behavior in a new location: classroom structure in real life locations
- practice the behavior mixed in with other behaviors: real life structure in classroom setting
- practice the behavior mixed in with other behaviors in new locations: real life structure in real life locations
- ask for the behavior in real life
When I train a behavior, there are always times when I go back to a classroom environment or structure to clean up some details or add the next piece. So the transition from classroom to real life happens many times. In the progression above, I might find myself going back to the classroom setting between any of the steps if I decide there is some detail that needs more attention and it is easier to work on it in a controlled environment. It’s normal to start working on a behavior in the classroom, take it out and use it a while in real life, decide you want to make some changes, go back to the classroom structure or environment to do some fine tuning, add cues, work on fluency etc…. Sometimes you don’t really know what you need until you test it out.
Here are some specific examples of changing one element at a time to go from the classroom to real life.
CHANGE OF LOCATION -> practicing in other environments -> change of focus
Here’s an example of a way to use location to expand the classroom, which eventually leads to practicing new behaviors in situations where you might want them in real life.
Most of the behaviors that I train are started in the classroom. In the beginning, the classroom might be my horse’s stall and I will be working behind a barrier. Over time my classroom can expand to new locations that include places where I work my horse on a regular basis. A fairly normal progression for me would be:
- horse in stall, me outside with a stall guard or door in between.
- both of us in the stall
- both of us in the aisle, or grooming area
- both of us in the ring
- both of us in the pasture, yard, or areas near the barn
If you look at this list, you can see that this is one way to transition from working in the classroom to real life. The last locations on the list are places where I might also be passing through as part of my daily routine. If I have practiced head lowering near the barn as a classroom session, then it is reasonable to ask the horse to stop and lower his head as he passes by the barn on the way from his stall to turnout, or turnout to the grooming stall, etc.
I also want to point out that by keeping the classroom structure (one behavior, repeated a few times before changing anything) as I move to new locations, I am building my horse’s confidence because practicing a few times before changing anything helps keep him on a higher rate of reinforcement and gives him time to be comfortable with the behavior in the new place. Over time these new locations will become familiar enough that I don’t have to return to the original physical classroom if I need to do a little refresher or fine tune a behavior. This is one way that I can expand the size of our physical classroom. It is also one way in which the lines between classroom and real life become a little blurred, which in this case is a good thing, as I want my horse to respond in a real life location just as well as he does in the classroom.
For example, if I wanted to start training head lowering, I might start by teaching head lowering in the stall (classroom structure/classroom location). When the horse is doing well in the classroom, I can move to a new location, but do the same type of session (classroom structure/new location). As I do this in more and more locations, my focus will change because I am now working on generalizing and proofing the behavior as well as just getting the behavior. Depending upon what the horse needs, I can start to work on the other steps in the process so that the behavior becomes stronger and more reliable.
By the time I am done with this, I have changed my location to include new places. I have changed my focus because I have now worked on asking for the same behavior in many places (generalizing and stimulus control) and I have changed the structure because I am probably moving around more to ask in different spots so there are other behaviors mixed in with head lowering.
CHANGE OF FOCUS -> change of structure, change of location
Here’s an example of how changing my focus can help me transition a behavior from the classroom to real life. This would be a fairly normal progression for adding a cue to head lowering.
- teach head lowering in a classroom setting (stall, aisle, ring, …) until the behavior is reliable
- decide to put the behavior on cue (classroom setting/classroom structure)
- test the cue by mixing in other behaviors (classroom setting/real life structure)
- test the cue in new places (real life structure in either less familiar classroom settings or real life settings)
- ask for the behavior in real life
When I work on adding a cue, I start in a familiar place so I know I will get the behavior, but then I have change things around a bit by asking in different places or mixing in cues for other behaviors. This is the best way to make sure the horse is not just repeating the behavior that I last clicked. By slowly making these changes, I can shift from a classroom setting/structure to working in different environments. I also start preparing my horse for the structure of a real life session which is less likely to require the horse to repeat the same behavior multiple times.
CHANGE OF STRUCTURE – > change of location -> change of focus
In a classroom type session, I usually work on one behavior at a time. In “real life” I need the horse to be able to shift gears from one behavior to the next. It’s not uncommon for horses to develop favorite behaviors that they offer all the time and sometimes I do this on purpose when I create default behaviors. But aside from that, I do want my horses to respond to my cues correctly and to be able to flow from one behavior to another. When I start mixing up requests for different behaviors, this will be important. In real life, there are a lot of context cues that will help this process, but I can get a head start by changing the structure of my sessions to include more behaviors and the types of questions I might ask in real life.
Here’s an example of how I can change the structure of my training session to prepare for real life. I start by creating small chains of behaviors that go together nicely. If my horse is used to doing only one behavior in a session, I will choose behaviors that have very different cues and find ones that flow nicely together.
- start in grown-ups
- ask for targeting
- walk off casually
- grown-ups – > leads back to targeting
I can create lots of little loops, starting with easy patterns and I can also practice patterns that I might need as part of my daily routine. Teaching a horse to walk, stop, and back up is similar to what I need to do if I am walking to a field and opening a gate. I can add in grown-ups to build some patience or I can add in some yielding of the hips or shoulders if I have a gate that requires the horse to move in any direction other than straight.
In addition to changing the structure, working in chains like this is going to change the focus because I will be working more on cues. It will also eventually change the location as I take these little chains of behavior and start using them in real life. Once I am working in real life, I can start to add in other types of reinforcement because there will be opportunities to reinforce in ways other than just using the food I am carrying.
If you look at the examples above, you should see a common pattern which is:
- teach the behavior in a controlled environment
- change one aspect of the training (location, focus, structure)
- change another aspect of the training (location, focus, structure)
- change another aspect of the training (location, focus, structure)
- repeat until you have the behaviors you need where you need them, inserting classroom sessions as needed.
I hope you can see how structure, focus and location are interconnected and lead to behaviors that are reliable in many situations. I hope you also see that there is a lot of flexibility in how you go about doing this. I do want to stress that it is important to try and change as few things at a time as you can. This goes back to one of Karen Pryor’s principles of training which is that you relax one criteria when you add a new one. It’s completely normal to relax the standard of behavior a little when you work in a new location or are working on cues.
This may seem like a lot of work, but once you understand the process with a few behaviors, it becomes easier with each new one. You understand more about how to prepare your horse and your horse has a better understanding of cues and reinforcement. A good quote to keep in mind comes from Alexandra Kurland who writes “You can’t expect to get a behavior from your horse on a consistent basis if you have not gone through a teaching process to teach that behavior to your horse.”
Having a well thought out training plan should not prevent you from being flexible. I may decide ahead of time that my criteria is nose to the floor regardless of location. But if I go out and it’s clear that I am not going to get nose to the floor in any new location, then I either need to put my horse away and do some thinking, or I need to be flexible enough to adjust my criteria or choose different locations or make some other change so that the horse can still be successful.
The examples I listed above show possible ways to take a behavior from the classroom to real life. In some cases, you will find that once you have worked through those steps, it is easy to ask for the behavior in new places so that you slowly expand how you use it. If you have built a solid foundation, you should find that the behaviors become stronger and more useful as you ask for them under many different conditions. In real life, you will find that steps 4,5, and 6 of the development of a behavior are easier to do because there are many opportunities to ask for a behavior in new places (generalization), clear up confusion over cues (proofing) and introduce new reinforcers.
WHAT ABOUT ADDING NEW “CLICKER TRAINED BEHAVIORS” TO THE TRADITIONALLY TRAINED HORSE?
So far, I have been writing about the situation where I am teaching and introducing a new clicker trained behavior into my horse’s repertoire. I chose that scenario because it is easy to break that process down into small steps without adding in too many other variables. If I work through the steps listed above, I will have a horse that has been taught some new behaviors in such a way that he can do them in various places and in combination with other behaviors. These are also behaviors that he is going to want to do because he associates them with positive reinforcement.
If every behavior I was going to ask for was one I had previously clicker trained, then this would be all I needed to do. But most of us start with horses that have some training. And whether we intend to re-train all their behaviors with clicker training or not, we can’t do it all at once. So we have a combination of new clicker trained behaviors and existing behaviors while we are in a transition period. This can create some confusion for both the horse and the trainer. It is one of the areas where I see people get frustrated, so I’d like to look at it in a bit more detail.
Let’s assume you have worked on a behavior in the classroom and the horse is doing it pretty reliably. You can think of a few places in your daily routine where this behavior might be helpful and you would like to start using it in real life. There are a few things you need to consider that can make this more successful.
1. Decide what to do about stimulus control. You need to have an appropriate level of stimulus control to keep you both safe, and to keep either of you from getting frustrated. One of the things about stimulus control is that you never know how much you need until you don’t have it! But.. there are things you can do to stack the odds in your favor. At the very least, you need to have some kind of working cue to ask for the behavior. The working cue could be one that you used to shape the behavior, or it could be some kind of context or environmental cue. If you want to use your working cue, it’s a good idea to check to see how reliably the horse does the behavior on cue and honestly evaluate if that’s good enough for how you want to use it.
If you want to use a context or environmental cue, then you have to think about what cues will work for you. For example, I find it is easy to start clicking horses for backing up when I enter their stalls. I teach them about backing in a classroom setting. Then I ask for it when I open the stall door. Within a few sessions, my presence at the stall door often becomes the cue to back up. This is a nice environmental cue because it contains information about when to start, when to stop, and whether or not to offer it again. Most horses quickly learn that my presence means to start backing and that they can stop once I am in the stall. It helps that there is already a set routine for this because they are used to the pattern of me coming in, doing something, and going back out. There are always a few overachievers who keep offering backing, but I can either redirect them or ask for something else and they usually figure it out.
2. Decide what to do about offered behavior. I could have lumped this in with number 1, but it’s important enough that I think it deserves to be discussed separately. If you are new to clicker training, you might not have considered what to do about a horse that wants to offer behavior because you are used to a horse that is more passive. I think it’s important to decide ahead of time if and when you will accept a behavior that is offered without you intentionally cueing it (note I said “you intentionally”). This can save you both a lot of frustration.
Most people end up with different behaviors on different levels of stimulus control because we all have different preferences and priorities about how we want our horses to behave. The right level of stimulus control is the one that works for you and your horse, but it really helps if you think about this ahead of time and try to be consistent. You can always change your mind if the option you pick is not working, but don’t be so flexible that the horse is unclear about what it can and cannot offer.
I have to add a note here which is that if you decide not to reinforce offered behavior, but your horse keeps offering behavior and is getting frustrated, then you need to take a closer look at what is going on. Cues can be tricky and what we intend to be the cue and what the horse uses as the cue are not always the same. You can create a huge amount of frustration if you ignore offered behavior and don’t look more closely at why the horse is offering it. Clicker training allows us to communicate with our horses and it also allows them to communicate with us. We need to be aware of this and use that feedback wisely.
3. Look at what other behaviors you are already clicking and treating to see if you have created a complementary set of behaviors, or that each behavior is at least balanced by an opposite behavior. I have people stay in the classroom until the horse has a big enough repertoire and is fluent enough in a few behaviors that the handler is confident that the horse will respond correctly most of the time.
Alexandra Kurland often talks about how important it to keep behaviors in balance. For every behavior you teach, you also need to teach the opposite behavior. This is a very simple way of managing behaviors before you have finished working on stimulus control. I take that idea a little further by teaching behaviors that are not necessarily opposite, but that complement each other. My goal is to have a selection of behaviors that I can use to redirect a horse that gets stuck on one behavior in particular, or that is unable to do the behavior I want.
This means that in addition to behaviors that balance each other out, It’s also helpful to have some overlapping behaviors so that if one is not working, you have another option. One of the mistakes I sometimes see is that people start working on “real life” training before they have created a nice group of complementary behaviors. This is especially important if you are working directly with the horse. If the horse is behind a barrier, then there is more flexibility, but I still like to have a few behaviors so the horse doesn’t get fixated on doing one thing. The more options I have for things I can ask my horse to do, the easier it is for me to work through any unexpected training snags when I start to explore working in real life.
I’ve been reading the yahoo lists for a lot of years and it’s not uncommon to have someone who has trained a few behaviors and now writes and says their horse is out of control. The horse only wants to do those behaviors and is cranky if asked to do anything else. The simple answer is that they don’t have enough stimulus control, but I think it’s also worth thinking about what motivates horses to choose one behavior over another.
The way I look at it is that most horses want to get along with us and they learn to cooperate and do what we ask, but they have not invested any particular amount of time or energy in learning any particular behavior. They are used to the idea that people ask for things and they just do it. This is especially true for horses that have already had a lot of traditional training. Then we come along and say “hey, I’m going to pay you for doing this behavior.” Now suddenly, they have one behavior that pays and all these other behaviors that don’t. So what do they do? They become very determined to do that one behavior. And if you don’t let them do it, they get really cranky.
I have had a number of emails from people who taught targeting. They taught the horse to touch a target stick and then started having it target other objects. At first this is great. They have a horse that loves the game and is now exploring its environment and having a great time. But pretty soon the horse is dragging them around in order to target things, and if they don’t let him, he gets frustrated and acts out. So now they aren’t sure they like clicker training at all. The problem is not with the targeting, although it would help if they had a cue. The problem is that the horse only has one behavior that he can do to earn reinforcement. Even if you work to put it on cue, there’s always going to be that little desire to do the behavior that gets better reinforcement.
That’s why I said it might not be enough to just put targeting on cue. I find it works better if the trainer adds some more behaviors so that the horse can get paid for doing a greater variety of behaviors. This is what I mean by having a complementary set of behaviors. If the horse is intent on targeting, but it also knows backing or head lowering, then you can ask for one of them when it wants to target. Instead of being in a position of having to prevent the horse from doing the thing it wants to do, you are asking for an alternative behavior which is also clickable. Having a nice set of clickable behaviors can make a huge difference when you try to work outside the classroom.
WHAT ABOUT STARTING BY USING CLICKER TRAINING IN REAL LIFE?
The last topic I want to address is an alternative way to introduce clicker training into real life. In general, I prefer to train behaviors and then gradually put them to use or incorporate them into what I am already doing. But this process takes time and there’s always the question of what to do in the meantime. And it doesn’t take advantage of capturing or strengthening existing behaviors.
The focus in clicker training tends to be on short sessions where we shape behavior because that is one of the most powerful ways to use clicker training and we can shape behavior so quickly. This is why classroom sessions are so useful. But that’s not the only way we learn. We also learn by asking questions over time and putting together those answers to get a better understanding of a subject or by making small changes and seeing their consequences. Some children learn their colors by sitting down with colored blocks and learning the names. Other children learn by seeing something, asking the color, and being told the answer. There’s no right or wrong way to do it and some children will learn better one way than the other.
The same is true with horses. Some horses and/or lessons are more easily taught in classroom sessions, whereas others can be learned gradually as each new piece of information is needed. In the latter case, Iskip the classroom sessions and just look for opportunities to shape or reinforce behavior as I spend time with the horse. This is another way to use clicker training, but it starts start by using the real life structure. I want to be clear that one is not better than the other, and they are not mutually exclusive. Some horses do better with one approach than the other and some horses do well with a mix. I think that most people eventually end up doing both, because there are times when it really helps to be able to do a classroom session and there are other times when it is more productive to shape the behavior as you go about your daily routine.
There are a few rare horses that don’t seem to do well if they are put in “training sessions” and they need to be taught in very tiny pieces and in such a way that they do not feel pressured. Any kind of training can put stress on a horse because being asked to change your own behavior is not always easy. This can show up in different ways. Some of them shut down, some act out and get frustrated, others just don’t seem to retain the information well. Luckily I have met very few horses that didn’t improve over time (with clicker training) until they were able to go back and forth between different types of training sessions. But if you feel your horse gets more stressed when you are doing classroom sessions, you might want to experiment with changing around the structure of the sessions, as well as looking at all the other elements that might be causing frustration or confusion (food delivery, rope handling, clarity of cues, etc….)
If you do decide to try using clicker training in real life without doing a lot of classroom sessions, or in addition to classroom sessions, then I think it’s important to be a little thoughtful about it. Before I start any “real life” work, I want to make sure that the horse understands what the click means so at the very minimum I do some basic targeting. Once the horse has the “light bulb” moment, then I might start by just clicking for a few simple things, so I can see how he responds to being clicked in a different context. This is a nice way to evaluate what he does after the click and treat. Does he suddenly change his focus to trying to get me to click more, or does he go back to what he is doing? Does he repeat what was just clicked? Does he start to mug me? In the early stages of real life clicking, I don’t necessarily want a click to mean that we are starting any kind of focused training session. It just means “I like what you are doing.”
If the horse understands about the click (I can usually tell by their reaction to hearing it) and accepts the food politely, then I can start to click for desired behaviors. When I am clicking in this real life structure, I might be capturing behavior (if it’s behavior I want more of), strengthening behavior (building reinforcement history), or putting behavior on a variable reinforcement schedule (if it’s behavior I want to maintain).
While this all falls into the category of “clicking for things I like,” I do try to be aware of why I am clicking and what I hope to accomplish by doing it. This means that I have to think about what kinds of changes I want to see and take the time to observe and evaluate what effect clicking and treating is having on these behaviors. It can be difficult to keep track of a lot of different behaviors at once, although I find it gets easier with time and practice. One solution is to only actively work on shaping a few behaviors at a time. I can click and treat other behaviors, but more with the idea of building reinforcement history or maintaining them. So even though it sounds easy to just go out and click behaviors you want to reinforce, it does require some thought.
If you are considering doing this, then here are some points to consider. Actually they are useful points to consider if you are transitioning from the classroom to real life too. Some of these points may seem to overlap or even contradict each other (yikes!), but that’s because each situation is different and these are general comments. Like most training there are a lot of “it depends” that I could make about each point. So rather than try to cover every possible situation, my intention here is just to provide some information and tips so that you can be more thoughtful about how you choose to do your training.
- What’s your motivation? Are you clicking/treating for you or the horse? I think people often start clicking for existing behaviors because they are anxious to start using c/t outside of “formal” training sessions. It makes them feel better because they are rewarding the horse for things they like. I use the word “rewarding” on purpose as I think that just because you start clicking and treating, it doesn’t necessarily mean you are reinforcing specific behaviors, in the sense of making them more likely to happen in the future. There’s not necessarily anything wrong with doing this, For some people, this is an important step as it gets them more comfortable with the idea of using food as reinforcement. You just might want to ask yourself if you are rewarding behavior or reinforcing behavior. If nothing else, asking yourself this question will start to make you more aware of what you are doing.
- If you are clicking for the horse, how do you want this to affect him? I started my clicker training with an aggressive horse. I clicked her for EVERYTHING for a long time. This was important because she needed the clarity of the click and to change how she felt about interactions with people. But over time, I did slowly start to look for ways to change to other reinforcers, chain behaviors together, and teach her that we could communicate in other ways besides clicking and treating. In the same way I said it was important in point #1 to ask yourself if you were rewarding or reinforcing, it’s important to recognize that in some cases clicking and reinforcing for everything is the answer, at least for a while.
- When I start clicking and treating for lots of things, the effect is more like using classical conditioning because my use of positive reinforcement is creating an association between my presence/requests and good things, rather than modifying specific behaviors. This is a good thing and for some horses it is HUGE. A horse that has had really negative experiences with people will often benefit from just being clicked and treated for lots of things so that he starts to associate people with positive experiences. It’s always nice if your clicks are also being used to shape behavior at the same time, but I have had cases where it really was just simpler and more effective to focus on keeping a high reinforcement rate and change the horse’s emotional response and then deal with the rest later.
- It’s worth noting that as soon as I start clicking and treating something, I create an expectation that this behavior could be clicked. Therefore, before I start clicking an existing behavior, I try to think a little bit about if this is something I want to continue to click and treat in the future. If the horse is doing it well and I don’t want to be clicking and treating it in the future, then I’m probably not going to choose that behavior to reinforce that way. It’s important to re-evaluate this on a regular basis as it’s easy for behaviors to get out of balance so different behaviors may need reinforcement at different times. And don’t forget that you can reinforce in other ways to help keep behaviors in balance.
- Previously in this article, I used the example of a horse that got frustrated because his only reinforceable behavior was targeting. I talked about how important stimulus control can be for avoiding frustration and I want to emphasize that again. Horses love being clicked and treated for doing things, but they can get incredibly frustrated if they think you are withholding clicks or they don’t know how which behavior will earn reinforcement. If I have a horse that is at the stage where it will throw behaviors at me to get me to click and it doesn’t really understand about cues, then I am going to be very careful about what I click in real life. I am going to choose behaviors that already have cues attached to them. These can be context or environmental cues as well as cues I have taught. The point is that I want the horse to know what was clicked and what to do next.
- Sometimes it’s easy to click a behavior that is in the middle of an established sequence so that the horse gets clicked in a situation where it knows what happens next. This can make it easier to click/treat and move on to the next thing. If I have a horse that is a bit sticky about haltering, I can often improve its behavior by clicking for one aspect of haltering (the crown piece going over the ears) and feeding when the halter is on. The horse is not going to expect me to take the halter off and start again because the normal routine is that I just put the halter on once. Over a period of a few days or weeks, that one click will often make a significant improvement in the horse’s attitude about haltering.
- One of the benefits of clicking a variety of behaviors without focusing on one in particular is that it is easier to keep behaviors in balance. Because I am clicking for many different things, the horse will not get focused on any particular behavior. I can keep the rate of reinforcement high. If I find behaviors are a bit out of balance, I can just adjust as necessary so it’s a very “go with the flow” type of training.
- If your goal is to start with a high rate of reinforcement and then slowly shift over to using other reinforcers, working in chains, fading the click/treat, etc… then you need to have a plan for how to do this. I’ve done it a few different ways. The first step is usually to take a few days and just click any response that is vaguely correct and not encouraging unsafe behavior. I know this sounds like a very sloppy way to set criteria, but I can’t think how else to say it. The point of the first few days is not to get or train behaviors. It is to start to connect with the horse and collect data. You can also do this step without clicking and just collect data by observing the variation in how your horse behaves at different times. One advantage to clicking (and treating) during this data collection period is that there will be some things that improve right away and then you know you don’t need to spend a lot of time on them. The point of collecting data is so that I can find a reasonable starting place and set criteria that the horse can meet. It’s really hard to set criteria for a behavior if I have no idea what the horse can and cannot do. The first day I might ask my horse to stop and he stops but takes 4 steps and swings his hips a bit. If stopping has been an issue in the past, I may click it because at least he did stop. Over the course of the next few days I may ask for 20 stops and discover that actually 4 steps with a swing is pretty good, compared to not stopping at all, 10 steps and stop, 2 steps/lean into me/step on my toe, etc… Or I may discover that 4 steps with a swing is not good at all and that sometimes the horse stops in 2 steps nice and straight, or 2 steps with a swing, or 3 steps with a slight lean but no swing. After a few days I will have a better idea of what kind of variation there is in his existing responses or behaviors and I may see that some behaviors are already improving just by being clicked. Once I and the horse start to find some connection and the horse knows reinforcement is possible, a certain amount of stuff will sort itself out. This is a great example of where I could collect some data by clicking in real life and then go back to the classroom and train the behavior I want from the ground up. But if I can’t do that, then I can still make progress in my real life training by setting criteria based on what I have learned in my data collection phase.
The normal way would be to pick a starting point (what can the horse do?) and click for all efforts that meet your criteria. Another way (from Kay Laurence) is to look at the other end of the spectrum and decide what makes the effort so “bad” that you can’t possibly click it. They are just two different ways of thinking about criteria. You can set your clickable criteria or you can set your “too bad to click” criteria. I have found that for a real life situation, and especially if you are working on multiple behaviors at a time, it is often simpler to just decide what is not deserving of a click.
This still allows you to shape behavior. You are just shaping it by weeding out the worst instead of picking out the best. I recommend choosing the method that will help you be more consistent about what you click and makes it so that most efforts by the horse are going to be reinforced. You want to help the horse be successful so don’t set your criteria so he’s only going to meet it one out of ten times. Both methods introduce the horse to the idea that not every behavior will get clicked, which is helpful if you are interested in going toward variable reinforcement or using alternative reinforcers. Sometimes I start this process and then decide that there is no way to improve the behavior in a real life setting and I will come up with a management solution while I set aside specific “classroom” time to shape the behavior from the start.
- Clicking for lots of things in real life can be helpful for horses that are tense/nervous/anxious about food in general. In some cases, I think horses get overly muggy or anxious because they think food is only available for certain things. Or maybe food is such a novelty that they don’t know how to stay calm around it. These horses benefit from getting regular reinforcement for simple things over a longer period of time so that they learn that there are plenty of opportunities to get food. There seems to be a point at which they just breathe a sigh of relief and become calmer about food in general.
- I think there is something about this approach that can affect your relationship with your horse in ways that go beyond reinforcing for specific behaviors. It’s hard for me to put my finger on it, but it allows for a little more flexibility on your part as to what you reinforce and it feels less like “training” where you are teaching or telling the horse what to do.
One of the best things I heard in recent years was that it is the nature of behavior to change (Kay Laurence.) Even if you work hard and train a behavior to perfection, it’s going to change. That’s normal and it’s one reason we need to be able to adjust our training based on what the horse needs today. The classroom setting provides an environment where we can control some of the variables and it is invaluable for fine tuning and modifying behavior. But it’s out in the real world where we really get to see how things work. Being able to flow back and forth between different locations, types of sessions (structure), and changing the focus as needed, are important skills. As you and your horse get more comfortable with clicker training, it will become easy to find the right mix of classroom and real life to keep your training moving forward.
Katie Bartlett 2014