The consequence obtained by an animal immediately after a behavior is performed, is called "reinforcement" if it causes the behavior to recur with greater frequency or intensity, and "punishment" if it causes it to decrease in frequency or intensity.
Also, as we have said on other occasions, that reinforcement will be "positive" if it involves the addition of a stimulus and "negative" if it involves its removal.
The most effective procedure to modify behavior is based on positive reinforcement, as generally the subject will be willing to work hard to get that reinforcement, whereas to avoid a negative reinforcement is usually applied the "law of least effort" to escape stimulus that produces aversion (plus it has been proved that the use of these measures increase aggression and apathy (Azrin & Holtz, 1966).
Moreover, reinforcements may be 'primary' or 'secondary'. Primary ones (or unconditioned) are those that have intrinsic value to the animal without being associated with something perceived as positive. These reinforcements don´t need to be learned by the animal as such, often associated with vital elements like food. Secondary ones (or conditioned) are those that the animal, a priori, doesn´t perceive as a reinforcement, being neutral. They must be associated or paired with a primary reinforcement for the animal perceives it as something positive. Clear examples are caresses, flattery, or the tool called “clicker”. So if we sound a clicker, which initially does not mean absolutely anything to the parrot, and simultaneously we offer him a sunflower seed, which is initially perceived as something good, after several repetitions, we condition the sound of the clicker (is said to be “charged”) so that the animal perceive it as something positive.
Several factors influence the effectiveness of reinforcement:
- CONSISTENCY: for something to be considered a reinforcement it should be provided consistently after a certain behavior is given (at least in the early stages of learning; later on I will explain variable schedules of reinforcement). If you teach a parrot to say his name after being asked, but only occasionally you reward him, he won´t understand the connection between the behavior and the treat, so it will lose some of its value as reinforcement.
- CONTIGUITY: reinforcement should be provided as soon as possible after performing the behavior. If after saying his name we don´t reward the bird until a few seconds later (during which several behaviors may have occurred, albeit being quiet), the strengthened behavior wont be the one we wanted to reward. This phenomenon has been studied by Lattal (Lattal, 1995), who documented that elapsing one second after catching a disc before rewarding a group of pigeons trained for it, the behavior was learned, while leaving 10 seconds elapse between response and reward, learning was impossible.
- FREQUENCY AND MAGNITUDE: Schneider (Schneider, 1973) discovered, through various experiments on rats, that frequent and smaller reinforcements are more effective than infrequent and larger ones.
Reinforcements, like toothbrushes, are personal and not transferable. What an individual perceives as a reinforcement could be a punishment for someone else, or at least something that does not motivate him enough to repeat the behavior that caused it. Suppose, for example, we offer a child a mint candy after having done something in particular. If that particular child loves mint, he will enjoy your candy and will see it as a prize, so it will probably tend to repeat the above behavior to obtain more candies. However, it is also possible that the child likes strawberry over mint candies, so that it is possible for him to repeat the behavior to gain another candy, but without much effort, or even he does not repeat it because he is not motivated enough to do so. Of course it could also be that the child hates mints, so he will not see it as a prize at all and will do nothing to encourage us to offer other candy.
The only real test we can do to find out if something motivates our parrot, or not, and to what extent, is whether, after offering it as a reward after a behavior, this is repeated or not.
In summary and following the experimental work conducted with children by Sulzer-Azaroff and Mayer (1991), but perfectly applicable to our parrots or any other animal, the "rules" to keep in mind to properly manage reinforcements are:
- Initially, we have to offer the reinforcement IMMEDIATELY after the desired behavior is given. Later, when the behavior is repeated often enough, we can be more flexible and slowly introduce some delay or latency.
- At first, we have to ALWAYS reinforce the desired behavior (ratio Behavior: Reinforcement 1:1). Later, when the behavior is already learned, we can be more flexible and reinforce only some of the responses, intermittently.
- We must be consistent and be clear about the criteria that make a performance worthy of an award (reinforcement). This criterion may be gradually amended, but in every training session we should be clear what criteria must be met to reward a behavior, and act accordingly.
- We must carefully select reinforcements. We should not assume that sunflower seeds motivate our CAG. Maybe half peanut stimulate him much more. Similarly, we must be careful with the size of the reinforcement, if it comes to be food, to prevent the animal is satiated too soon and lose its motivational value amid the session (without prejudice to the effect that such reinforcements can have in the animal's diet, something I will explain at another time).
- It is preferable to use a certain variety of reinforcements (seeds, petting, praise, games, peanuts, fruits, funny behaviors due to Premack´s Principle, etc).
Finally, it is essential to remember that the primary mission of the trainer is twofold: firstly, finding those reinforcements that truly reinforce the animal which we are working and, secondly, namely promoting ad maximum the number of opportunities to offer those reinforcements.
Herein lies the true art of training.
To learn more:
Azrin, N. H. & Holz, W. C. (1966). Punishment. In W. K. Honig (Ed.), Operant behavior: Areas of research and application. New York: Appleton-Century-Crofts.
Friedman, S. G., Martin, S. & Brinker, B., (in press, 2005). Behavior analysis and learning parrot. In A. Luescher (Eds.), Manual of Parrot Behavior, (pp. xx). Ames, NY: Blackwell Publishing.
- CONSISTENCY: for something to be considered a reinforcement it should be provided consistently after a certain behavior is given (at least in the early stages of learning; later on I will explain variable schedules of reinforcement). If you teach a parrot to say his name after being asked, but only occasionally you reward him, he won´t understand the connection between the behavior and the treat, so it will lose some of its value as reinforcement.
- CONTIGUITY: reinforcement should be provided as soon as possible after performing the behavior. If after saying his name we don´t reward the bird until a few seconds later (during which several behaviors may have occurred, albeit being quiet), the strengthened behavior wont be the one we wanted to reward. This phenomenon has been studied by Lattal (Lattal, 1995), who documented that elapsing one second after catching a disc before rewarding a group of pigeons trained for it, the behavior was learned, while leaving 10 seconds elapse between response and reward, learning was impossible.
- FREQUENCY AND MAGNITUDE: Schneider (Schneider, 1973) discovered, through various experiments on rats, that frequent and smaller reinforcements are more effective than infrequent and larger ones.
Reinforcements, like toothbrushes, are personal and not transferable. What an individual perceives as a reinforcement could be a punishment for someone else, or at least something that does not motivate him enough to repeat the behavior that caused it. Suppose, for example, we offer a child a mint candy after having done something in particular. If that particular child loves mint, he will enjoy your candy and will see it as a prize, so it will probably tend to repeat the above behavior to obtain more candies. However, it is also possible that the child likes strawberry over mint candies, so that it is possible for him to repeat the behavior to gain another candy, but without much effort, or even he does not repeat it because he is not motivated enough to do so. Of course it could also be that the child hates mints, so he will not see it as a prize at all and will do nothing to encourage us to offer other candy.
The only real test we can do to find out if something motivates our parrot, or not, and to what extent, is whether, after offering it as a reward after a behavior, this is repeated or not.
In summary and following the experimental work conducted with children by Sulzer-Azaroff and Mayer (1991), but perfectly applicable to our parrots or any other animal, the "rules" to keep in mind to properly manage reinforcements are:
- Initially, we have to offer the reinforcement IMMEDIATELY after the desired behavior is given. Later, when the behavior is repeated often enough, we can be more flexible and slowly introduce some delay or latency.
- At first, we have to ALWAYS reinforce the desired behavior (ratio Behavior: Reinforcement 1:1). Later, when the behavior is already learned, we can be more flexible and reinforce only some of the responses, intermittently.
- We must be consistent and be clear about the criteria that make a performance worthy of an award (reinforcement). This criterion may be gradually amended, but in every training session we should be clear what criteria must be met to reward a behavior, and act accordingly.
- We must carefully select reinforcements. We should not assume that sunflower seeds motivate our CAG. Maybe half peanut stimulate him much more. Similarly, we must be careful with the size of the reinforcement, if it comes to be food, to prevent the animal is satiated too soon and lose its motivational value amid the session (without prejudice to the effect that such reinforcements can have in the animal's diet, something I will explain at another time).
- It is preferable to use a certain variety of reinforcements (seeds, petting, praise, games, peanuts, fruits, funny behaviors due to Premack´s Principle, etc).
Finally, it is essential to remember that the primary mission of the trainer is twofold: firstly, finding those reinforcements that truly reinforce the animal which we are working and, secondly, namely promoting ad maximum the number of opportunities to offer those reinforcements.
Herein lies the true art of training.
To learn more:
Azrin, N. H. & Holz, W. C. (1966). Punishment. In W. K. Honig (Ed.), Operant behavior: Areas of research and application. New York: Appleton-Century-Crofts.
Friedman, S. G., Martin, S. & Brinker, B., (in press, 2005). Behavior analysis and learning parrot. In A. Luescher (Eds.), Manual of Parrot Behavior, (pp. xx). Ames, NY: Blackwell Publishing.
Lattal, K. A. (1995). Contingency and behavior analysis. The Behavior Analyst, 24, 147-161.
Schneider, J. W. (1973). Reinforcer effectiveness as a function or reinforcer rate and magnitude: A comparison of concurrent performance. Journal of Experimental Analysis of Behavior, 20, 461-471.
Sulzer-Azroff, B., & Mayer, G. R. (1991). Behavior Analysis for Lasting Change. Orlando, FL: Harcourt, Brace, Jovanovich
Sulzer-Azroff, B., & Mayer, G. R. (1991). Behavior Analysis for Lasting Change. Orlando, FL: Harcourt, Brace, Jovanovich
No comments:
Post a Comment