Click-L FAQ

What is Click-L?

Click-L, or the Clicker List, is an online newsgroup devoted to the exploration of behavior modification techniques as they relate to training. Subscribers to the list come from many different backgrounds and professions. However, all of us are pursuing, at some level, further understanding of the psychological, ethological, and biological principles underlying learning, motivation, cognition, and behavior. Training methods discussed (and advocated) on this list will favor the use of conditioned reinforcers with positive reinforcement. However, since the correct and appropriate use of punishment and negative reinforcement is a necessary part of understanding behavior modification, don’t be afraid to discuss those as well.

Subscriptions: email majordomo@click-l.com for full help.

 

I thought this list was for dog trainers; why are people discussing the California Highway Patrol?

Although many of the subscribers to Click-L are in some way connected with the field of dog training, there are numerous subscribers who are interested in other areas of behavior modification and training. On this list, you may find reference to anything from the training of the piglets in the movie, “Babe” to a discussion of alternatives for encouraging children to clean their rooms. One of the primary principles of behavior modification techniques is that like the laws of physics, they apply to all species (within their physical limits). So subscribers may speak in generalities of applications that are relevant to many different situations and concerns.

 

What is the difference between a Primary and a Conditioned (secondary) Reinforcer?

A primary reinforcer is anything the subject will work to get–that is, anything the subject finds desirable, whether it seems worthwhile to someone else, or not. Like: food; attention; patting; affection; chewing gum; loud noises (small boys are fond of loud noises, often) or whatever. A secondary reinforcer is an initially meaningless signal or stimulus that stands for one or more of the primary reinforcers: in other words, the term is synonymous with the term conditioned reinforcer. Sea World and some of the zoo people have taken to using the term, secondary reinforcer to mean things which do not seem as crucial to them as food and water–like toys, petting, etc. But this is misleading. For one thing, you are now calling into play the trainer’s opinion of what is truly important to the subject and what is not–when in fact only the subject can show you that. We all know dogs that are suspicious of free food, and other dogs that would kill for a favorite toy, dogs that don’t like toys, etc. etc. A primary reinforcer, to repeat, is ANYTHING the animal will work for. Secondary or conditioned reinforcers are learned symbols of primary reinforcers: as, for example, good grades, cash money, the bells on a slot machine, a whistle, or a clicker. Practically speaking, also, the term secondary reinforcer implies that the secondary or conditioned reinforcer would come AFTER the primary reinforcer, and that of  course does not work.

How is the Conditioned Reinforcer used in a training
session?
Once you have established a positive conditioned reinforcer by repeatedly pairing it with other reinforcers, you can use it to selectively increase the behavior you want the animal to perform.  (You have to maintain its conditioned reinforcing property by continuing to pair it with the unconditioned reinforcer at least some of the time.) Skinner’s research showed that the more frequently a behavior is reinforced, the more often that behavior will occur under those conditions. Any reinforcer, unconditioned or conditioned, selectively increases any behavior it follows, which is how you get targeted behaviors to increase while unreinforced behaviors do not. For example, if you are teaching your dog to sit and the reinforcer is delivered only when the dog is sitting, it will tend to sit until you deliver the reinforcer, whether it is a conditioned reinforcer such as a click or an unconditioned reinforcer such as food. If you progressively delay delivery of the reinforcer while the dog is sitting, the dog will learn to sit for longer periods of time because you are reinforcing longer, rather than shorter, sitting responses. If you then start using the command “sit” before the response-reinforcement contingency, the behavior will come under control of that stimulus.
What can you train an animal to do?

In the words of Dr. Marian Breland Bailey of Animal Behavior Enterprises, “Anything it is physically and mentally capable of doing.”

How are the methods advocated by this list different from other training methods?

Clicker training relies greatly on the constructive use of conditioned reinforcers. The regularity with which we use a
conditioned reinforcer results in the development of a signal which functions, simultaneously, in the following ways:

as a conditioned reinforcer, i.e. a substitute for a primary reinforcer.
as a conditioned stimulus meaning “The PR is now available.”
as an “event marker” as Ogden Lindsley puts it, identifying for the learning organism what behavioral topography is being reinforced.
as a context marker, meaning, to put it informally, that “the game is On.”
as a generalized conditioned reinforcer–having the above qualities, the CR becomes more powerful as a motivator than the initial PR (we can also intensify this aspect of the CR by pairing it deliberately with many kinds of PR’s).
as information: for example the CR can convey to the subject that some new conditioned stimulus has been correctly
interpreted, even BEFORE the whole behavioral response occurs.
as reassurance (one aspect of the informational function) when for example the animal is working at a distance or under duress of some kind.
as an “ego boost,” again to put it informally: Animals are reinforced by their own success in manipulating their
environment; thus, in an experienced animal, the sound of the “click” can become an intrinsic or primary reinforcer.

The second biggest difference between traditional trainers and clicker trainers may lie in their understanding of successive
approximation. Clicker trainers are able to break behavior down into small, achievable steps, and thus approach any behavioral goal via conditioned reinforcement and shaping, rather than by more global methods such as modelling and compulsion. This is not to say that traditional trainers never do that, of course they do it; but they don’t do it as methodically and as generally; and often “trainer error” is incorporated rather than weeded out.

Finally, it is generally true that clicker trainers do not rely on punishers to teach behavior, but that is really a byproduct of
these applications; when you shape behavior with a CR you don’t need as much correction, and “correction” has toxic side effects, (Murray Sidman, _Coercion and Its Fallout_). But the “difference” really has nothing to do with “all positive,” in fact a conditioned negative reinforcer is very useful; the main difference is the knowledgeable and timely application of the CR.

What are the advantages of this training method?

Ultimately, the easiest way to find out the advantages and disadvantages is to try it on your own pets. However, if you read the list for any length of time, you are sure to come across posts from others who have tried it and found it helpful. Please pardon us if we sound like we’ve “gotten religion”…We’re having fun and can hardly wait to tell other people how to join us!

 

Why is it that the conditioned reinforcer elicits such a powerful response?

Conditioned reinforcers start out as neutral stimuli, having no effect on behavior. They become reinforcers by occurring at about the same time as established reinforcers, whether unconditioned reinforcers like food or play or other stimuli that have already become strong conditioned reinforcers. In other words, any stimuli that are consistently around when strong reinforcers occur are likely to become conditioned reinforcers, even if it is unintended. Conditioned reinforcers like a click sound are actually no more powerful than other reinforcers. What makes the “click” so effective in selectively increasing the behaviors they follow is

trainers have been taught to repeatedly pair the click with food, which keeps it a strong conditioned reinforcer,
the distinctive, momentary sound encourages trainers to be very careful about when they deliver the conditioned reinforcer.

This exact timing following a thoughtfully chosen behavior is the heart of a good reinforcement contingency. In fact, it is usually much more effective than what trainers were doing before they started using a clicker, which is why it seems extraordinarily powerful. In addition, because less precise reinforcement procedures often involve a slight delay between the exact behavior that should be targeted and the actual delivery of reinforcement, the behavior that is selected (increased) is sometimes not exactly the form of responding that is desired. The use of a conditioned reinforcer that can easily be delivered at the precise moment intended, particularly a sound, which does not require the animal to be watching, helps the trainer selectively increase the exact form of the animal’s behavior that is desired.

How can I use a clicker to teach my dog NOT to do something?

You can eliminate an undesired behavior is several ways. Most ways of using positive reinforcement to reduce the frequency of a problem behavior involve creating or increasing desired behaviors.

For example, one approach is to train an incompatible behavior. You can train the dog to sit when someone comes in the house. The dog can’t sit and jump up at the same time, so making sitting more likely under those conditions makes jumping less likely.  The dog can now earn reinforcement by sitting and being greeted by the guest. You can train any behavior to function in this way, even if it is not incompatible with jumping, as long as the behavior is intermittently reinforced.

You can also put the undesired behavior on cue. For example, you teach the dog to rear up on its hind legs with a command like “paws up” and offer your forearm as a place for the dog to rest its front legs. In training this performance, it will be important to never reinforce the behavior if it occurs in the absence of the command. After this training is well-established, the dog will be unlikely to jump up when the command is not offered.

Of course, there are many other ways of getting rid of unwanted behavior. However, they do not involve the use of a clicker.

If I just click to get the behavior, when do I introduce
the command word?
 

Sit and be patient until the subject does whatever you want to reinforce (if there is a specific behavior you’re wanting to train, it’s legal to find ways to set the subject up to offer a behavior in the right direction)
Click a behavior that meets your criteria (or maybe a small piece of a behavior that is a step in the right direction on the subject’s part). If you’re not getting any behaviors you probably need to readjust your criteria (expectations).
Give the primary reinforcement.
Wait for the behavior to repeat itself.
c/t and repeat until you are consistently getting the behavior you want. (Until, as Gary Wilkes suggests, “You would be willing to bet $5 that the subject will repeat that behavior within the next 30 or 40 seconds.”)
Give the verbal command or signal (or both), *as* the dog is performing the behavior, pairing it directly with the C/T. So, using “sit” as an example, *as* the dog’s rump is about to hit the ground you would say “sit”/C/T. This associates the word “sit” with the actual physical act of sitting
When the subject is performing on cue, put him/her on a variable schedule of reinforcement.

 

What is a clicker, and where can I get one?

A clicker is a small metal/plastic object that makes a clicking noise when pressed. You can make a quite good clicker out of a lid from a jar (if it has a “pop-up freshness seal”: baby jar lids can work well).  Experiment until you find one that makes an easily perceivable sound. Ogden Lindsley uses the acronym SURA to list the   qualities of a conditioned reinforcer:

S – Short
U – Unique
R – Ready
A – Ambient

Obviously, the subject must be able to perceive it in order to become attuned to it. Deaf dogs, for example, cannot hear the sound of a clicker, but they can *see* a flashing light, or a “thumbs up” signal with the hand, or *feel* the tingle of an
e-collar set on its lowest setting (which, in a good collar, is not painful at all).

If you want a “real” clicker, you can obtain them from Sunshine Books (1-800-47-CLICK). These are small (approx. 1″ x 2″ x 3/4″)  plastic boxes with a metal interior which clicks when it is flexed with  your thumb or finger. These make a quite loud click, so if your pet is quite sound sensitive, you might want to muffle it when you first introduce it.

Where can I get more information?

Some excellent resources for those who are interested in a basic understanding of the principles we discuss on this newsgroup are available in any bookstore. You do not have to be a rocket scientist to understand how operant and respondent conditioning work, but you will better understand the posts on this list if you do some preliminary reading. Some suggestions follow; the first six are available through Direct Book Services (1-800-47-CLICK), from your local library, or as listed. The psychology  textbooks are probably available through many college bookstores.

 

B.F. Skinner (1951). “How to teach animals.” Scientific American 185 (12): 26-29.

 

Karen Pryor, _Don’t Shoot the Dog! The New Art of Teaching and
Training_
 
Karen Pryor, _A Dog and a Dolphin_

 

Karen Pryor, http://www.karenpryor.com, has clicker related information.

 

Gary Wilkes, _CLICK! & TREAT (TM) TRAINING KIT_ VHS – 56 minutes
Includes two clickers and an instruction book
Available from Click & Treat (TM) Products, Mesa, AZ
Order line (800) 456-9526
 
Karen Pryor, _Lads Before the Wind_

 

Steve Martin, _The Positive Approach to Parrots as Pets_
Available from Direct Book Services. There are two tapes in
the series, but the one specifically related to training is
the second.
 
The list administrator (Kathleen Weaver) has written a couple
of informational articles about clicker training which you can
read on her web page: http:\\www.metronet.com\~kathleen
All previous digests are also available at this site and can be
downloaded from here (see below for further information on
how to download the digests).
 
There is a journal for trainers who want to use Operant Conditioning.
It is called _The Clicker Journal_, and is published bi-monthly. Send
$15.00 (U.S.) to:
The Clicker Journal
Corally Burmaster, Editor
20146 Gleedsville Rd.
Leesburg, VA 22075

 

Many Intro to Psychology textbooks have one or more chapters
on behavior modification. There are also some excellent
textbooks on Behavior Modification: the following are highly
recommended as informative and readable:
Garry Martin & Joseph Pear. Behavior Modification: What it is
and how to do it (5th ed). (1996). Prentice Hall. Upper Saddle
River, NJ 07458.
ISBN # 0-13-310947-X

Alan Kazdin. Behavior modification in applied settings (5th
ed.). (1994). Brook/Cole (division of Wadsworth). Pacific
Grove, CA (1-800-423-0563).
ISBN # 0-534-21786-9

Richard Malott, Donald Whaley, & Maria Malott. Elementary
principles of behavior (2nd ed.). (1993). Prentice Hall.
Englewood Cliffs, NJ 07632.
ISBN # 0-13-260241-5

David Watson & Roland Tharp. Self-directed behavior:
Self-modification for personal adjustment. (1993). Brooks/Cole.
Pacific Grove, CA
ISBN # 0-534-18978-4

Holland, JG; Skinner, BF The Analysis of Behavior. (1961) New York:
McGraw,-Hill.

A list of trainers who use operant conditioning can be found at
http://www.wazoo.com/~marge/Clicker_Trainers.html

 

What are some common mistakes likely to be made by new
clicker trainers?
a) Setting goals too high: Remember that you should train in
small increments towards an ultimate goal. Success breeds
success, and if your training steps are smaller, you should
easily be able to complete them and move on to the next
thing.

b) Not giving enough reinforcement: Especially when you are
training a new behavior, make sure you reinforce every time
the behavior occurs. Take your time about moving into the
variable reinforcement schedule.

c) Using the CR as a “recall” device: Often, when people see
how their pet responds to the CR, they make the mistake of
using it as a way to get the animal to come to them. Since
the animal hears the sound, and knows it will get a treat,
it approaches the trainer. This can lead to some unfortunate
consequences, because the CR signals reinforcement at the
time it sounds. If your dog is digging in your neighbor’s
flower bed, and you use your clicker get him to come, you
have just rewarded him for digging in the flower bed. You
must always think about what you are reinforcing, because
behavior that is reinforced gets repeated more frequently.

 

A number of abbreviations are frequently used on this list.
Below is a list of the most common ones, with definitions
where relevant.
Common Abbreviations

RC: Respondent Conditioning
OC: Operant Conditioning
CR: Conditioned Reinforcer (also sometimes called a secondary reinforcer)
R+/R-: Positive/Negative Reinforcer
PR: Primary Reinforcer
P+/P-:Positive/Negative Punisher
C/T: Click and Treat
VR/VSR: Variable Reinforcement/Variable Schedule of Reinforcemt
FR/FSR: Fixed Reinforcement/Fixed Schedule of Reinforcemt
DSTD: Don’t Shoot the Dog (book by Karen Pryor)
LBTW: Lads Before the Wind (another book by K. Pryor)
OTOH – On the other hand
IMHO/IMO/IME – In my (humble) opinion/experience
BTW: By the way
ROTFL: Rolling on the Floor Laughing
FWIW — For What It’s Worth
AFGE — Another F*** Growth Experience (we get a lot of those)
J&P — “Jerk and Praise” (or “Jerk and Puke/Pull”) — refers to traditional “pop and release” training methods

 

Can I get copies of previous posts?

Yes, these are available in digest format at web site:  http://www.ListService.net/clickerlist

When you get it back to your machine, you’ll need to “unzip” it as it is a compressed file. You’ll need PKUNZIP (pc) or UNZIP (os/2) or WINZIP (win95) or a compatible uncompression program. The digests will uncompress into a bunch of seperate files (2 a day). They are in ASCII format so most viewers, word processors, or the PC command type filename.ext | more will work.