Wow, plowing thru those scientific writings is a challenge!

Since I use the clicker to teach NEW behaviors mostly. And once my dogs 'get it' start fading the click/rewards for that behavior and start teaching new things, I think that is the difference with the studies/vs. what I use the clicker for.

In the present example, the sound of the food magazine acquires reinforcing value (becomes a secondary reinforcer) through its correlation with ultimate (primary) reinforcement. It can function as a reinforcing agent even when this ultimate reinforcement is lacking. Its reinforcing power will be weakened through the resulting extinction, but considerable conditioning can be effected before a state of more or less complete extinction is reached.” (Skinner, 1938)
I added the bold to the above. I do NOT want the power of the click to be weakened, even if it works better FOR A WHILE when the dog is trying to figure out what's going on, and then shuts down. I want no extinction.

although Hull knew in 1943 that a habit can be made more resistant to extinction if established by means of intermittent reinforcement, he thought that the most effective secondary reinforcement was established by associating the secondary reinforcement stimulus “repeatedly and consistently” with primary reinforcement . Once again I added the bold part. The 'he thought the MOST effective was the part I bolded' and I agree!

So in simple words. You can actually make the click intermittent and the treat intermittent AND INCREASE the number of repetitions (or time) it takes for the behavior to become extinct from lack of primary reinforcement. I believe all that you cited Robert. I swear I do. But their examples were different from what we are doing when TEACHING something new to our dogs.

Their tests were on teaching ONE thing to animals (I think it's rat hit bar get click and food in chute). And WHEN THEY STARTED, every time the rat hit the bar, the click would happen, and the food would come. Only after they knew the rats 'got it' did they start NOT clicking and/or clicking and NOT giving the food. And the resultant 'thinking' that the rats went thru to now figure out this new puzzle was what resulted in the subsequent INCREASE in their hitting the bar for a while (which is your point I think about why it is good to do) BEFORE the rats slowed down because and even then, the SR had not lost all of its potency.
I want the potency of the clicker to always be high to my dog. I don't want to push them to find out the level of random reinforcement start their extinguishing a behavior, even if initially they may perform better. For the reward after the click, I use what MY DOG thinks is a reinforcer. So if my dog works best for liver, tug, or praise, or whatever. And once my dog has learned a behavior well, frankly, I rarely use the clicker at all. And if I do, it's to 'randomly' click what I've decided is brilliant so that keeps them on their toes and working for the 'surprise' of the click/treat.

Jenn, you are selectively taking parts of comments from the quotes I posted.

You're not seeing the big picture. Zimmerman got 2000 responses!!! So, you think that a few here or there is going to make a difference in motivation? It isn't, sorry. (Well, unless you consider that the variable ratios will increase overall persistance of the learning.)


The learning will be longer lasting, as well as the conditioned reinforcer. So the benefits outweigh the possible downside of the extinction. . .and you'll only really see the effects of extinction if you WAY over do it!

That is the reason why the clicker folks say you MUST click/treat. Very inexperienced trainers or pet dog owners who aren't even inexperienced trainers (LOL) may go wild. . .not get good results. . .and say clickers are a sucky gimmick that don't work. Happens all the time.


Now, I'm not going to pick apart the whole post, but this re-quote right here is just glaring.

Your quote on Hull.

You highlighted Hull's comment that he thought continuous pairing would be stronger (to support your argument), while managing to pretend that the very next sentence says that a few years later, he realized that he had NOT been correct, and that intermittent pairing was actually stronger. You're suffering from selective perception - we all like to only read the parts that support our argument.


"Plowing thru those scientific writings" is the only way to educate yourself so that you can sift the bullshatt from the ACCURATE science on those fluffy pet-owner websites that you've quoted so well from.

I wonder if you are still missing the point. Most of us DO fade out the use of an actual clicker cuz of it's clumsiness but, we continue, for the life of the dog, to use a conditioned reinforcer, albeit a verbal one. I use "GOOD".

It's the pairing of the primary reinforcement with the conditioned reinforcement (regardless of whether that's an actual clicker or a verbal CR) that was the subject of discussion - not whether you should use a real clicker forever. After your dog has learned the new behavior, you will continue to ask for it thru the dog's life, and you should intermittently reward that behavior (I think you would agree with that!). I'm saying that when you mark the response, you don't need to always follow up with a primary reinforcer.


That's the point I made in my first big post.

You said you MUST click and treat, and that is not accurate. You can expand on that to an extent and still see the same level of success, as long as you are careful about extinction.


. . .and you might also help yourself a little in the persistance of your CR.

NO!


AAAAAAAAAAAAAAaaaaaaaaaaaaaaaaaaaaaaaaaaaaaggggggggggggggghhhhhhhhhhhhhhhhhhhhhh...............


lol


I freaking give up.


You don't HAVE TO treat and click while teaching, you don't HAVE TO wait till the dog has learned the behavior to work with a variable ratio of reward.


All you have to do is condition a level of emotional value for the click, making it a conditioned reinforcer/secondary reinforcer and then you can use that click to reward/teach the behaviors you want. . .even without having to give the primary reward after every single click.

AND you'll see more persistant results as well as having a more persistant conditioned emotional value for your click (secondary reinforcer).


<img src="/ubbthreads/images/graemlins/mad.gif" alt="" />


<img src="/ubbthreads/images/graemlins/laugh.gif" alt="" /> <img src="/ubbthreads/images/graemlins/laugh.gif" alt="" /> <img src="/ubbthreads/images/graemlins/laugh.gif" alt="" />


I can't present the information any better than I already have. . .my blood pressure is soaring.


<img src="/ubbthreads/images/graemlins/laugh.gif" alt="" />

Well, and speaking with low blood pressure, I'll just say we are both reading the same material and getting different information from it. All the general studies you cite are about the entire subject of the Principles of Operant Conditioning which consists of Reinforcement, Punishment, Shaping, Extinction, Generalization and Discrimination. So just the 'training new behaviors' are hard to pull out.


And were you able to find any dog (or specific animal) training sites, that agree with your slant on the scientific information? Cause I can probably find over 20 who's understanding of operant conditioning is the same as mine, and can't find any that have your read on the scientific papers. For TEACHING a new behavior and using the click.


Once again, my point is to educate everyone that is willing to take up the clicker to give it a try and make it the easiest for them and their dogs. So all the websites you can provide would help. I'm still open minded enough to read any DOG related (or other specific animal related) training sites cause I learn new things about training all the time. And other professionals that have trained hundreds of dogs with the method you explain would be interesting to read about.

First of all, I learn a ton from this discussion board and appreciate the information. There have been some good links and quotes posted on this thread.

Anyway, Van Camp said, "I can't present the information any better than I already have. . .my blood pressure is soaring."

I heard a great example for variable ratio of reward works in a psych class I took a few years ago. Basically slot machines work on a variable ratio reward system on humans and people continue to feed the money in without any guarantee of payout.Heres a site that briefly explains that
http://www25.brinkster.com/ranmath/skinner.htm

I was wondering if the clicker training works on the same principals?

Michael, good point about the slot machines and shows what Robert is saying about the variable ratios. And I agree with this with an ALREADY LEARNED BEHAVIOR. Once I have won at a slot machine and had the joy of the money coming out once, it is worth trying again (and again and again and again). But will say, if I have a bad hour or so at the machine, I'll be quite discouraged and not be inclined to gamble for awhile.

Where I disagree is for learning a completely new behavior. Especially if you are trying to shape something new (like the pigeon on the site you posted that they were teaching to spin) If you don't pair the 'click' with a reward each time to confirm the animal is on the right track and keep them willing to participate and think thru the game, I think you are lessening the 'power' of the click.

The site you posted says:

 Quote:

Animal trainers know it is not necessary to reward a trainee every time it does the right thing. Every few times will do. This is called fixed ratio reinforcement. Reinforcement on an unpredictable schedule is known as variable ratio reinforcement.

It has been shown experimentally that behavior induced by variable ratio reinforcement has the greatest resistance to extinction.

So I am probably still not explaining this right, it's the using the clicker/ without any reinforcer each time I have issues with. To me that is NOT a variable ratio of reinforcement. Using a variable ratio of reinforcement (to me) is (for instance) once your dog knows a sit, I am sure not going to continue to click/treat every time the behind hits the floor when I ask it to!

Interesting that the Cambridge Center for Behavioral Studies suggests the same sites I have previously referred to.

MICHAEL, I'm still looking up sites and found a dog training/clicker site that mentions your slot machine comparison! Slot machine!

 Quote:

If reinforcement fails to occur after a behavior that has been reinforced in the past, the behavior might extinguish. This process is called extinction. A variable ratio schedule of reinforcement makes the behavior less vulnerable to extinction. If you're not expecting to gain a reward every time you accomplish a behavior, you are not likely to stop the first few times your action fails to generate the desired consequence. This is the principle that slot machines are based on. "OK, I didn't win this time, but next time I'm almost sure to win!"

When a behavior that has been strongly reinforced in the past no longer gains a reinforcement, you might experience what's call an extinction burst. This is when the animal performs the behavior over and over again, in a burst of activity. Extinction bursts are something for trainers to watch out for!

Once again, I want to stress that this variable schedule of reinforcement is (the way I read it) for an already learned behavior NOT when in the learning stages.

The reason I like to look at the dog training sites with their specifics of training is they seem clearer to me with exactly what I am supposed to do. Variable Reward Giving says:

 Quote:

6. Become a Variable Reward Giver
Now try to get your dog to do 2 or 3 repetitions of the trick before you click and treat it

Hint: it’s still a good idea to say "Good dog!" each time, just to let the dog know she's getting it right.

This is called "putting the behavior on a variable reinforcement schedule". The dog doesn’t know when she’ll get the big reward, so she keeps trying — just like people playing at slot machines or the lottery.

Important: Since you’re rewarding less often, you can also get picky — only reward the straightest sits or the highest paw-lifts. This is where behaviors get perfected. Note: Some people prefer to perfect the behavior before they add the cue.

All the sites I have found, specifically for training dogs/animals seem to pair the click/reward each and every time. And they all are proponent of variable reinforcing!

Jenn wrote: "I'll just say we are both reading the same material and getting different information from it."


That isn't accurate.

Many of the works/information put out by top clicker gurus are either omitting information to keep it simple for the dummies, subverting the science to keep nay sayers and other dummies from screwing up and raining down more training hell on their own poor little dogs, or are just plain wrong because they got the information from another clicker site/book and are reposting the same information over and over and over without ever putting any study into the originating science behind their methodologies. <img src="/ubbthreads/images/graemlins/wink.gif" alt="" />

I've posted scientific studies with proof, so post as many websites with clicker trainer's opinions as you want. It doesn't change the results of the experiments and studies that I've referenced.


I'm not saying this is the ONLY way to use the clicker. What I am saying is that it is an option, it is correct use of a conditioned reinforcer, and a good one for someone savy enough to do it correctly.


The click is ONLY a conditioned reinforcer. . .nothing else, so it can be used in any way a conditioned reinforcer can be used.

It's that simple.

I can't control myself. . .I must post more. LOL


Jenn you posted this quote below. Thank you, it's spot on. (Remember that verbal praise is itself a secondary reinforcer, or a conditioned reinforcer. . .just like a clicker.)


"quote:6. Become a Variable Reward Giver
Now try to get your dog to do 2 or 3 repetitions of the trick before you click and treat it

Hint: it’s still a good idea to say "Good dog!" each time, just to let the dog know she's getting it right."


What exactly is "Good dog"? A Conditioned Reinforcer!!!!!!!!!

Which is NOT followed by a primary reinforcer.

In fact, every single time a trainer says "good boy" without adding a primary reinforcer behind it, he is doing exactly the same thing as clicking without treating. Yet no one ever seems to have a problem with doing THAT. They all believe their 'good boy' has power, yet they think the clicker will become powerless. . .or lose it's potence. Go figure?!

The clicker and "good boy" are exactly the same thing, Jenn.