premack principle

Saya

I was wondering is anyone familiar with this and use it?

What are your thoughts on it a few dogster members talked about it, but I never really looked much into it till now.

http://blogs.dogster.com/dog-training/the-premack-principle/2010/09/

http://www.teachingdogobedience.com/2009/01/premacks-principle.html

I don't know much on it I just started to read on it out of curiosity.

BradA1878

The use of nontraditional reinforces to chain behavior is something Jen and I have used for awhile now, however I had never heard of the term "Premack Principle". Interesting, thanks for sharing.

----

sandrat888

I use Premack Principle to transfer value from something the dog wants to an object or activity that the dog is not too keen on a lot in training.

I will present whatever I want to build value on to the dog and if he chooses to engage it (initially the criteria is set to be really easy, so the dog gets the reward pretty fast, but the criteria will be raised, so we can move along toward the end behavior I want), then he will get something he wants in return.

Once you have enough value transfer to the new object or behavior, it can sustain itself and be just as rewarding and valuable to the dog as the original reward he gets.

JaanaS

I have never heard of that term.

But it sounds like how some people prepare for obedience events. It works in lower levels where different movements are usually in same order. Lets say that there is movements: heel, down from heeling, recall and jump.

When they have learned all the movements they start with the last one: Jump -> big reward. This has to be done several times. Dog will learn that jump ends up with something really good.
Then they do recall -> jump ->reward. This some time then they add next one and so on. In the end dogs will learn that when they do these it's rewarding to get to next one because they have linked them all to nice reward.

It takes time to get whole obedience class together like this but usually dogs end up doing even better after few movements at competitions. (opposite to that "I didn't get any reward from first one. Why I should do the next one?")

sandrat888

Premack Principle can be used in many scenarios. You pair a low probability behavior with a high probability behavior to increase the probability of the first behavior. If you constantly do this, you will increase the probability of the low probability behavior.

Scientific research has shown that you can even pair an aversive (I believe in the experiment, mild electric shock was used) with a reinforcer (such as food). If presented consistently, in the end, the aversive itself will become an reinforcer, which is very fascinating.

kbbd83

My dog trainer taught this method, but I also have not heard it called Premack Principle. Very interesting. It worked very well with my Shiba.

InuRyuu

In short, the principle is making a high probability behavior contingent on a low probability behavior. Such as, when you tell a child, no ice cream if you don't eat your vegetables. Basically you are conditioning a behavioral chain where the low probability behavior becomes the discriminative stimulus for the high probability behavior.

@Sandrat888- as an undergrad I performed a similar experiment in avoidance conditioning where the rat was able to avoid the shock and get water (+ reward) if he would press a bar on one side of the cage and then pressed another on the other side in under 2 seconds. The experiment was nick named The Rat Race.

Losech

Yes, that is one of the methods I used with Conker's food aggression. I use it now to keep him calm when I make his food, and before he gets his food. I also use it for recall and off-leash privileges.

Yandharr

I use this too. Never really put a name to it though. I always thought of it as more of a bridge and transfer process- as in you build a bridge to a wanted behavior and then transfer the behavior completely over time to a different object/behavior. Wait, isn't this just operant conditioning??

sandrat888

Operant conditioning is used by B.F. Skinner to describe the effects of the consequences of a particular behavior on the future occurrence of that behavior. This is sometimes referred to as trial and error, where a living organism changes its behavior based on the consequences of said behavior.

There are four quadrants in operant conditioning - positive reinforcement, negative reinforcement, positive punishment and negative punishment.

Premack's principle states that more probable behaviors will reinforce less probable behaviors and for this to work, the less probable behavior has to happen first, before presenting the more probable behavior.

A simple example would be I ask my dog for a sit first, then offer him a yummy treat as a reward, rather than using the food as a bribe/lure to ask him to sit, and then give him the treat when he sits. In the first example (asking for a sit, then reward), you will increase the value of the sit (transferring from the reward given after the sit) and as you keep doing this sequence, "sit" itself will have more value to the dog and can be reinforcing to build value for other behavior once the "sit" itself is strong enough to be an reinforcer itself.

InuRyuu

It's interesting to note this works both ways. My dogs perform certain behaviors to get me to do certain things. Hmm, whose manipulating whom?