Clicker Training: The “How-To’s”

 

In clicker training philosophy, we use a “marker” (a word, noise, sound or gesture) that tells the dog when he’s done something to earn reinforcement.  Many clicker trainers will only use a clicker, and many will only use a verbal marker… I use either, dependent on how much precision I require in the behavior.  This really depends on the individual, and what you’re teaching, and things like whether your dog has a sensitivity that precludes using a clicker or other means.   Your dog won’t be confused if you switch from one to the other, after he’s learned the rules to the game.

 

If you’re simply teaching companion-style obedience, then using a verbal marker is fine.  Competitive sports like flyball, freestyle dance, and agility require more precision, though… the use of the clicker is recommended... but not necessary, if you know what you're doing.

 

There are 2 ways of teaching the dog the meaning of the marker: you can “charge” the marker (teach the dog that the sound means “you get a treat!”), or you can simply jump in, training… either way, he’ll get it.  If you’re brand-new to this, you might consider charging the marker first, so that you’ll understand the process involved.  Note: when I say “mark”, I mean either “click” or “verbal” marker.

 

The process:  Have your dog move around (it's not necessary to only have the dog assume a certain behavior wheile they learn this... in fact, it's important that he NOT think that the click/verbal is associated with a 'certain behavior'... so anything he does is fair game for teaching him that Mark = Treat.

 

Mark & treat

Mark & treat

Mark & treat

Mark & treat

 

Repeat the Mark & treat sequence every few seconds, for about 20-25 repetitions… then:

 

Mark & pause

 

If the dog looks to your treat hand when you pause, then the marker is charged – the dog understands that the sound means he’s getting a treat!  If he doesn’t look to your treat hand, then just continue on with the Mark & treat sequence for 10 or so more repetitions, and test it again with a pause.  Continue this way until the dog looks to your treat hand expectantly.

 

Caution: I have met 2 dogs who have sensitized to the sound of the clicker.  Sensitization is the opposite of desensitization… instead of becoming “used to” something (desensitizing), the animal develops a very strong FEAR of it -- and it can happen as quickly as 1-2 clicks .  If your dog shows ANY sign of sensitization to the clicker, stop using it immediately, and continue with the use of a verbal marker, or a click inkpen, or some other sound-maker that doesn't distress the dog.  Signs of sensitization include ducking away, jumping back, when they hear the “click”… at this point, if you see this type of behavior, STOP and use the verbal marker!  Granted, 2 dogs is NOT very many, when compared to all the dogs I’ve worked with… but sensitization is a very serious problem when it happens!

 

Now, the marker is “charged”… what next?  Well, you now have a way to mark those behaviors that you feel are appropriate (wanted) … and a behavior that’s rewarded is repeated… and repeated… and repeated. {grin}  You’re way over halfway there!

 

Now that you've started to teach your dog this way, it's time to figure out how to get those slight changes in body posture.  Clicker training allows me to manipulate rewards in a way that provides *feedback*, either that I like what the dog is doing, or that I would rather the dog find another way to earn reinforcement. If I like what the dog is doing, I say "GOOD!" or click to MARK the behavior, then offer reinforcement. If I don't like what the dog is doing, I say "huh-uh" to mark the behavior, and then the dog tries another behavior.

 

This brings up another point: what if you DON'T like the offered behavior?  Then you TELL him that you don't like it, in a way that won't scar him for life, or make him shut down emotionally... this is called the No Reward Marker.  “Huh-uh” is called a “No Reward Marker” (NRM)… all it means is “that behavior doesn’t get rewarded right now… try something else”. 

 

Example of using each marker:

 

I ask Zoe for a ‘sit’, she sits, I say “good!” and reinforce.  The “good” is my verbal marker, to let her know exactly what it was that I liked.

 

I ask Zoe for a ‘sit’, and she ‘downs’, I say “huh-uh”, and wait… she ‘sits’, I say “good!”, and reinforce.  The “huh-uh” (the NRM) helped her to understand that I’m not reinforcing ‘downs’ right now… and it allowed her to remain “in the game”, because it’s not overtly aversive in nature.  The opportunity for reinforcement is still there… just not for that particular behavior.

 

Another example:

 

Rody walks into the kitchen while I’m doing dishes, comes over to me and ‘sits’ for attention.  I say “Good!” and reinforce. 

 

Rody walks into the kitchen, comes over and jumps up on me for attention.  I say “huh-uh”, he ‘sits’, and I reinforce.

 

These examples illustrate the use of the 2 markers in teaching the dog which behaviors are unacceptable, without removing the opportunity for reinforcement.  This is very important, because removal of the opportunity for reinforcement can cause a sensitive dog to shut down, emotionally.

 

Note #1: before I teach the dog that “huh-uh” means that they won’t earn reinforcement for a behavior, I first make sure that there is another, more appropriate behavior to reinforce.  I actively teach that behavior.  For example, if I don’t want the dog to jump on me for attention, I first teach that sitting gains attention, consistently.  This gives the dog an alternative way to earn reinforcement. 

An extinction burst: when a behavior has earned reinforcement in the past (like jumping up for attention), and then you STOP the reinforcement, the behavior will extinguish -- it will stop.  Before it stops, though, it will go through extinguishment... the animal will try that same behavior over and over, harder and harder, in an attempt to earn the behavior.  You have to stand strong, and just let this happen... if you mess up and GIVE UP on the process (if you inadvertantly reinforce) then you'll make the behavior even stronger.  In fact, to make a behavior stronger, you use the principle outlined here (the extinction burst) to either make the behavior stronger or get a new behavior tossed into the mix for possible reinforcement.  Be very careful here, if you don't want to keep a behavior... jumping, for example.  A 150-pound dog going through an extinction of jumping behavior isn't a place where you want to accidently reinforce, making the behavior stronger and even more resistant to change.

Note #2: If your dog has learned, for example, that jumping on you will always (or, almost always) be reinforced (through yelling, eye-contact, pushing him off, kneeing him in the chest, etc.), then withholding the reinforcement may cause an extinction burst.  Read the last paragraph: an extinction burst is, simply, trying the behavior over and over, in an effort to gain the expected reinforcement -- because This Behavior Has Always Worked Before.  You probably see this all the time, with humans... what happens if your pen starts to skip?  Do you throw it away and get out a new one?  Chances are, you go through a whole repertoire of things to TRY before giving up... this is an extinction burst.  Watch people at an elevator... do they press the button once, and wait??  Usually, you'll see several people who repeatedly press the button until the doors open... this is an extinction burst.  Dogs who spend a lot of time on a chain are very prone to extinction bursts... barking, jumping, mouthing, pawing are all behaviors that dogs use to get attention, especially when that's all they have.

 

 Continuing to ignore the inappropriate behavior while reinforcing the appropriate will allow the behavior to extinguish.  Trust me on this -- DO NOT revert to your old ways of handling the situation, or you’ll reinforce the extinction burst, and cause it to become more resistant to extinction.

 

Note #3: Training is a bargain, and involves reciprocal trust.  It’s IMPORTANT to always follow the marker (either verbal or clicker) with reinforcement.  Even if you “mess up” and mark a “wrong” behavior as deserving of reinforcement, don’t forget to reinforce.  The Training Game is a BARGAIN between you and your dog… if you don’t carry through on your end of the bargain, then how can he trust the “marker” the next time you give it?

 

Note #4: The marker ENDS the behavior.  This means that it’s OKAY for your dog to “pop” out of the ‘sit’ position to come get his treat, the instant he hears the marker – he understands the game!  When you’re ready to increase the duration of the ‘sit’, you simply gradually increase the length of time between giving the command, and marking it.

 

What Words to Use as the Markers?

 

That’s you’re preference… lots of people use “good”, a tongue-click, or “yes” to mark appropriate behavior as worthy of reinforcement – you can use “potato” if you’d like.  It really doesn’t matter, so long as you’re consistent in using the same word. 

 

For the NRM, lots of people use “huh-uh”, “wrong”, “nope” – you can use “frog”, so long as you’re consistent in its use.  What IS important is that you NOT inject negativity into whichever word you choose for this… you simply want it to communicate to the dog that “this particular behavior isn’t being reinforced, at this time” – injecting negativity into the marker may cause the dog to feel like he’s been harshly corrected, which may mean that you won’t see that behavior again.

 

How to get those behaviors, in the first place… just in case you’re working with a dog who has NO manners. (You can't MARK a behavior that isn't OFFERED in the first place, right?)  There are several ways to Get the Behavior, although one is difficult (my opinion), and one has undesirable consequences (again, my opinion).  Click the link to read the article on this.

 

About Training Sessions:

 

Training sessions should be SHORT, and FUN!  In fact, I keep treats all over my house, and just reinforce appropriate behavior as it happens… I firmly believe that Every Opportunity Is A Learning Opportunity, and the environment works 24/7 to Untrain our dogs -- so it’s incredibly important that we teach those behaviors that we want, throughout the day!  As an example, it's 1:30 in the afternoon, and I have done 7 sessions (3-5 minutes each) with the dog I currently have as a foster.  She's special needs because of her lack of training and socialization, so it's important to make Everything into a session.

 

When an average (not special needs) dog comes into my home, I’ll set up little “formal” sessions (maybe 5 minutes each, 3-4 times per day) to help him become acclimated to how we do things… otherwise, I just reinforce things as they occur.  These ultra-short, frequent sessions allow dogs to learn A LOT of things, in a very short period of time – and, it allows the entire training game to remain just that… a GAME.

 

How Much Reinforcement?

 

Many people are concerned with how many treats their dogs get in a day, for many reasons.  So:

 

1) If it’s not a health concern, don’t concern yourself over it.  If he gains a pound or 2, he’ll take it off later.  He’s not going to be training forever.  Really. 

 

2) If it IS a health concern, use his meals as treats.  Yes, that’s right… just don’t give him his food as MEALS.  Instead, use his meals to provide reinforcement for his training.  This way, you won’t be adding anything to his diet that may be potentially hazardous to his heath.

 

The rate you use to reinforce will affect the rate with which your dog learns.  In the early learning phases, the rate of reinforcement should be relatively high – about once per second.  As he becomes consistent in his responses, you can start to reinforce less and less often, and then only use treats when “brushing up” a behavior, or teaching a new one.

 

What Kind of Reinforcement?

 

This depends on what your dog loves.  The easiest way to start, of course, is in the use of food treats – I don’t use store-bought doggie treats, unless they’re the very good ones (few preservatives, high real meat/cheese content).

 

 I prefer to use “human” foods… cheese, hotdogs (all-beef only), leftover chicken or steak, popcorn, Squeeze Cheese, etc.  It’s best to find several food rewards that your dog loves – and it’s important to understand that rewards/treats will have different “values” under different circumstances. In a low-distraction setting, you can get away with Cheerios and popcorn -- less messy stuff. In high distraction, though, you'll want to bring out the Big Guns: Messy Foods Taste Better. {grin} And, I still want Kraft to do peanut butter, limburger cheese, and liverwurst in those squeeze cans. How cool would THAT be?? lol

 

For example, in my house, with no distractions, popcorn is a big treat for my dogs… but, in the yard, with a dog across the street, popcorn has no value – I’d better come up with something better, if I want their attention! 

 

As you progress in your training, you can use other forms of reinforcement: toys, petting, praise, chasing squirrels, other dogs, etc… but, at the outset, food rewards are more convenient, and easier to control.

 

Where to Put the Reinforcement?

 

This isn’t as simple as it sounds… because if you keep the treats on your person all the time, then the dog will never learn to Trust In The System Of Reward.  You’ll end up with a dog that only works when you have the treat-bag.  So, to teach this trust, you must become creative… sometimes the treats are in the fridge, sometimes they’re on the table, sometimes in the treat-bag, sometimes on the mantle… you get the picture. It's important to change the location where you keep the treats, so that the dog doesn't think that training depends on that treat in your hand. 

 

Who Can Train Your Dog?

 

Everyone in the family!  I’ve had kids as young as 6 or 7 years learn to use the clicker effectively.  The only person who shouldn’t be training a dog is someone who becomes angered or nasty.  In fact, your dog will learn even faster, with more people working with him!  It’s best, of course, that only one person be in charge of the clicker at a time (don’t have 3-4 people standing around, clicking him at the same time). 

 

I highly suggest that you work with very tiny pieces of behavior, moving ever closer to the goal behavior.  This is called “successive approximation” – moving successively closer to the goal, in small increments.

 Next: Behavior Modification 3

Help!