Clean Run Magazine - April 2007 - 38

In some instances I find it more palatable to use a conditioned reinforcer on those “correct but no reward” attempts. Let’s compare two examples, in which, over a series of 10 cues, the dog: Example 1 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. Jumps correctly Jumps correctly Runs by the jump Jumps correctly Jumps the wrong jump Jumps correctly Jumps correctly Jumps correctly Jumps correctly Jumps correctly is rewarded with click/toy is rewarded with a click is ignored is rewarded with a click is ignored is rewarded with click/treat is rewarded with click/toy is rewarded with a click is rewarded with click/treat is rewarded with a click Example 2 is rewarded with click/toy is ignored receives a NRM such as Wrong is ignored receives a NRM such as Wrong is rewarded with click/treat is rewarded with click/toy is ignored is rewarded with click/treat is ignored In the first example, the dog is on a schedule where 50% of the correct responses earn primary reinforcement. The dog is on a continuous schedule of conditioned reinforcement (100% of correct responses earn clicks). Only incorrect responses result in no consequence. This provides a set of clear outcomes for the dog. In the second example, the dog is on a 50% schedule of both conditioned and primary reinforcement. He is still able to discern correct from incorrect responses because of the NRM (no reward mark). There is also the option of merging both these approaches, in which you incorporate clicks on correct but nonrewarded jumps and NRMs on incorrect attempts. How do you decide which approach is best to use? I suppose it comes down to what you are comfortable with as a trainer. I tend to avoid using NRMs so I prefer the first example. But is the dog really gaining the benefit of intermittent reinforcement? Well, yes and no. Some studies demonstrate that using a conditioned reinforcer on intermittent schedules increases the rate of responding over the same intermittent schedule with no conditioned reinforcement. Yet, other studies find that an animal maintained on a combined 100% (continuous) conditioned reinforcement and 50% (intermittent) primary reinforcement is going to show patterns of responding that are between the two schedules. Either way, the animal is learning persistence. The animal is learning to keep trying in the face of nonreward. I don’t view it as much of a handicap for agility training because most handlers provide their dogs with positive feedback while they are on course anyway, whether through cheering, moving on, and/or cueing the opportunity to perform the next obstacle in the sequence. Psychology researchers sometimes use conditioned reward as a way to maintain animals on very sparse reinforcement schedules. Typically, animals will be reluctant to work or even quit on very lean schedules of reinforcement—this is termed “ratio strain.” But, if conditioned reinforcement is doled out occasionally, animals will do amazing amounts of work for an occasional piece of food. In one study, chimpanzees had to press a button 4000 times for a food reward. It’s hard to believe that the chimps would even do this. They did, but they worked very slowly, often taking breaks of minutes to hours. However, if the chimps were rewarded “It’s official: my entire crew loves this toy!” © DOGHOUSE ARTS The Crunch ‘N’ Tug “Most dogs love to crunch on an empty water bottle, others love the feel of a fleecy exterior. This toy has both. I love it because once you remove the water bottle, it folds flat for easy packing. Plus the bungee makes tugging easy on my back (and I’m sure on my dog’s body as well!).” Susan Garrett Twister, Buzz, Decaff & Encore www.dogdreamstoys.com 38 Clean Run | April 07

http://www.dogdreamstoys.com

Clean Run Magazine - April 2007

Table of Contents for the Digital Edition of Clean Run Magazine - April 2007