Behavior Analysis LECTURE 10 MAINTENANCE
Review Last time we talked about generalization training The tendency for the effects of training to spread There are two types of generalization Stimulus generalization and response generalization These two principles refer to behavior generalizing from one situation to another, and from one behavior to another Today our topic is maintenance, which has to do with getting behavior to generalize from one time to another In other words, how can we get the changes in behavior to persist over time?
Maintenance This is sometimes called the maintenance problem. There are two basic solutions to the maintenance problem, these are: Maintenance schedules Maintenance training Today we will talk about how to maintain behavior via these procedures
Maintenance schedules Before we get into the use of maintenance schedules, you need to know about the basic types of reinforcement schedules A reinforcement schedule is a kind of rule about how reinforcers are to be provided Schedule of reinforcement: a rule governing the delivery of reinforcers There are many different kinds of reinforcement schedules, but we will discuss the three main types Ratio schedules Interval schedules Duration schedules
Ratio Schedules Ratio schedules are based on the number of times the behavior occurs They are called ratio schedules because the rule for the delivery of a reinforcer specifies ratio of performances of the target behavior to each reinforcement i. e. maybe the target behavior has to occur 3 times before the reinforcer is delivered There are two kinds of ratio schedules Fixed Variable
Fixed Ratio Schedule In a fixed ratio schedule the rule is that the target behavior must occur a certain number of times before it is reinforced Fixed Ratio Schedule: provide a reinforcer after the target behavior has occurred N number if times If the behavior is on an FR-1 schedule, then the ratio of target behavior to reinforcement is 1 to 1. i. e. each time the behavior occurs it is reinforced In fact, this type of schedule, in which the target behavior is always reinforced is called ‘continuous reinforcement’ Schedules that reinforce some occurrences of the target behavior are called ‘intermittent schedules’ For example, with an FR-10 schedule, the behavior has to happen 10 times before it is reinforced
Variable Ratio Schedule VR schedules are identical to FR schedules except that the number of instances of behavior required for reinforcement varies around an average Variable Ratio Schedule: provide a reinforcer after the target behavior has occurred a number of times, with the number varying around an average of n In a VR-5 schedules, for example, the behavior might be reinforced after occurring 3 times, then after 7 times, then after 2 times, then after 4 times etc The number of occurrences for reinforcement will change from 1 -10, but on average the ration will be 5 -1
Interval schedules In ratio schedules reinforcement is based on the number of times the behavior is performed In interval schedules, reinforcement is based on the length of the interval between reinforcements. These schedules are called interval schedules because the rule for the delivery of the reinforcer specifies an interval between reinforcements of the target behavior There again two types of interval schedules Fixed Variable
Fixed interval schedule In fixed interval (FI) schedule, a fixed amount of time must elapse since the last reinforced performance, only then does the behavior qualify for reinforcement again Fixed interval schedule: provide a reinforcer the first time the target behavior occurs after an interval of N length since the last reinforcement For example, if the behavior is on an FI-1 schedule, then after the behavior has been reinforced, a minute must elapse before the behavior will be reinforced again. Performing the target behavior is not reinforced until the minute has elapsed Importantly, after the time period has elapsed the behavior has to be performed again before reinforcement is delivered.
Variable interval schedule VI schedules are identical to fixed interval schedules except the length of the interval between reinforcement varies around an average Variable Interval Schedule: provide a reinforcer the first time a target behavior occurs after an interval, with the interval varying around an average of N length In a VI-5 schedule, the interval might be 2 seconds, 10 seconds, 7 seconds etc but with an average of 5
Duration schedules Ratio schedules are based on the number of times the target behavior occurs. Interval schedules are based on the time between reinforcements Duration schedules are based on how long the behavior must occur without interruption They are called duration schedules because they specify the duration of the behavior There are two kinds: Fixed Variable
Fixed duration schedule In an FD schedule, an activity must be engaged in for a specified period Fixed duration schedule: provide a reinforcer after the target behavior has been performed continuously for a period of N length Fixed duration schedules are very common, a typical example is a child who is required to practice piano for 30 minutes everyday, after which he receives ice cream!
Variable duration schedule In a VD schedule, the requirement for reinforcement is the same as in an FD schedule except that the period during which the behavior must occur varies around an average Variable duration schedule: provide a reinforcer after the target behavior has been performed continuously for a period, with the period varying around an average of N length For example, the mother may require the her child to practice 10 minutes one day and 50 minutes another day, but the average will be 30 minutes
Maintenance schedules Ok so we have discussed three kinds of reinforcment schedules: ratio, interval and duration There are others, but these are the main ones One reason that reinforcement schedules are of interest to us is because of their value in maintaining behavior for long periods One solution to the maintenance problem is to put the behavior on a maintenance schedule, which is simply: Maintenance schedule: a reinforcement schedule that maintains a target behavior at a desired rate
Maintenance schedules Much research has been conducted investigating which schedule is ideal at maintaining behavior at high rates with infrequent reinforcement The ideal maintenance schedule seems to be the variable ratio schedule A reinforcer is provided after the target behavior has occurred around an average of N However, when shaping a behavior with continuous reinforcement, it is likely that when you switch to a VR schedule you may get some sort of extinction
Maintenance schedules Imagine for example, that you gave jonnie a sweetie every time that he finished his math's work (CRF). When you switch to a VR-5 schedule you may see a drop off in his behavior i. e. he will finish his math's work less. You may also see increases in emotional behavior This has been called ratio strain Ratio strain: a reduction in the rate of target behavior and an increase in emotional behavior resulting from increases in the ratio of behavior to reinforcement
Maintenance schedules To prevent Ratio Strain you need to avoid abrupt (sudden/sharp) increases in the ratio Instead, you gradually increase the requirement for reinforcement, this is known as stretching the ratio Stretching the ratio: gradually increasing the number of times a behavior must be performed to qualify for reinforcement For example, with jonnie we wouldn’t move straight from CRF to VR-10. Instead we would start at VR-1, then maybe VR-2, VR-4 etc. There is a limit to how much you can stretch the ratio, but you would be surprised at how few reinforcers it takes to maintain target behavior at high rates
Contingency Contracts So in order to maintain behavior at a high rate, all we have to do is put the behavior on a maintenance schedule. The trouble is that often, people don’t do that. Once they have produced the desired changes, they stop reinforcing the behavior or they reinforce it so little that the schedule is not rich enough to maintain the behavior For example, imagine you get jonnie to do his work by giving him sweets. As soon as jonnie starts doing his work, the teacher may stops providing the reinforcers, therefore putting the behavior into extinction
Contingency Contracts Why? The reason it drops off is because once the changes in the target behavior take place, maintaining that behavior becomes less reinforcing i. e. when the teacher sees jonnie doing his work the use of the reinforcement schedule is reinforced, but once this increase in target behavior slows down, the use of reinforcement schedule is less reinforced! In other words, we need make sure that we control the behavior of the person providing the reinforcement To ensure that they are correctly putting their target onto a maintenance schedule. You could even say that we need to reinforce the teachers behavior of providing a maintenance schedule, and put that behavior on maintenance schedule itself
Contingency Contracts We need to find some way of reinforcing the act of maintaining the maintenance schedule Remember that if a maintenance schedule is to stay in force, the effort required in managing such a schedule must itself be reinforced So were talking about reinforcing the behavior of the person who’s reinforcing someone’s behavior! That can sometimes be achieved with a contingency contract Contingency Contract: And agreement between two or more parties about what each is to do for the other
Contingency Contracts For example, a contingency contract between a teacher and study would specify what the teacher would do that the students find reinforcing, and what the students would do that the teacher finds reinforcing For example, the teacher could want a quiet classroom, without fooling around and passing of notes. This would give him time to mark papers instead of telling people off The students could want the freedom to approach the teacher when they find an assignment very difficult and once they had finished the assignment they could have time in the playground
Contingency Contracts So the teacher and the students must agree on the contingency contract together. If the teacher just makes it up without negotiation then it is not an agreement. Then someone writes everything down, and all parties sign the contract From the teachers point of view it specifies the target behavior and the consequences he has to provide to maintain them From the students view it specifies the teachers target behavior and what they have to do to maintain them It is therefore a reciprocal relationship
Contingency Contracts When contract problems arise you deal with it by re- negotiating the terms of the contract The contract creates a situation in which it is likely that the consequences necessary to maintain target behavior will continue to be provided This is because if either party breaks the contract then both parties will lose their reinforcers It’s a bit like a job, your employer agrees to pay you, and you work. If your employer stops paying you, then you stop working Contingency contracts have been used in many studies and have consistently got positive results
Token Economy Contingency contracts are great ways to maintain target behavior. A variation of a contingency contract is token economy A token is something that you can exchange for a reinforcer The reinforcer that you get for a token is called a back-up reinforcer Its called that because it is what makes the token valuable; it backs it up Token Economy: a form of contingency contract, usually involving a group of people, in which the reinforcers are tokens
Token Economy There are two defining characteristics of token economies There is a contractual arrangement involving a group of people Tokens are used as reinforcers, that can at a later stage be exchanged for back up reinforcers Most token economies involve students or people in institutions, but the study I am about to describe was set up in the home
Case Study Wolfe and colleagues devised a simple way to help parents reduce the amount of time their children spent watching TV They found 3 two parent families whose children, between the ages of 8 and 12, we addicted to TV The parents had tried the usual things of telling their kids off and reminding them that they should be doing other things, but this did not work The behaviorists first asked the parents to monitor TV watching behavior over a 7 week period to get a baseline. This measure showed that kids watched 21 hours per week
Case Study After this the researchers asked the parents to give their child 20 tokens. For every 30 minutes of TV watched, the kid had to give up one token They could not watch TV unless they paid for it in tokens This meant that the kid could watch no more than 10 hours per week If the kids abided by the rules for 4 consecutive weeks then they received a reward The results of the study shows that all of the kids reduced their TV watching behavior to less than 10 hours per week Interestingly, these kids actually spent a little more time doing homework, and much more time reading. They also reported liking the arrangement and even saved their tokens for special occasions
Social contracts So you have seen that we can create conditions to maintain behavior. But society also creates conditions to maintain behavior that’s beneficial to the society Society tries to maintain high rates of certain behavior and low rates of other behaviors by means of contingency contracts For example, criminal behavior occurs at a relatively low rate compared to lawful behavior. Of course one reason for this is due to our learning histories. However, another reason is due to contingency contracts; we have laws that specify that certain behaviors are not allowed and will results in punishment
Social contracts generally go against what we know about human behavior E. g. there is an over emphasis on punishment, and consequences tend to be delayed and uncertain Nevertheless social contracts can be used to improve behavior for the better For example, Yokley and colleagues noticed that less and less parents were getting their children vaccinated against various diseases. They therefore set up a contingency contract with those parents so that if they had their child vaccinated then they would receive money. As predicted, those parents were much more likely to have their children vaccinated What do you think of this?
Social contracts One lesson that we, as a society, haven't quite learned, is how to make reinforcement contingent on socially desirable behavior Behavior is not random The kinds of behavior we get, depend on the kinds of consequences that follow it If we want people to behave differently then we have to provide different consequences for their behavior So we have discussed how to maintain behavior via maintenance schedules, now we must talk about maintenance training and how that can also help with the maintenance problem!
Maintenance training The problem with maintenance schedules is that invariably come to a point where they no longer can be maintained For example, the students might move on to a new teacher It is quite often the case that behavior cannot be put on long term maintenance schedules Imagine the mentally retarded man who improved his behavior in an institution due to token economy, what happens when he moves to a different care home As helpful as maintenance schedules are, they seem to be a temporary arrangement So what can we do to prevent the effects of our training to continue once the intervention is over?
Maintenance training: intervention procedures that increase the likelihood that changes in the target behavior will persist when the intervention has ended. In other words, what can we do during the training to make sure that the behavior endures after the training has ended? Imagine you had shaped Billie to be very creative in your class The following year he may have a teacher that puts creativity into extinction, and reinforce very conventional behavior
Maintenance training Your intervention needs to make the desired changes resistant to the influences of a less than ideal future. You may need to immunize Billie against future experiences that may undermine the work you have done in making him more creative There are 5 maintenance training procedures that help us to do this Continue the intervention well after the target behavior has been changed Expose the target behavior to its natural reinforcers Teach the person to obtain reinforcement for the target behavior Shape tolerance for delayed and uncertain reinforcement Fade the intervention program
Continue the intervention well after the target behavior has been changed ABA seems to get results very quickly It has been recorded that behavior can be made to occur at a high rate with reinforcement, that shaping has happened in a couple of minutes and that years of a problematic behavior have been reduced by differential reinforcement in dramatic fashion The problem is that once the behavior has changed for the better, people often assume that’s the end of it. And there is no need to continue the intervention They are mistaken, the longer the intervention endures, the longer its affects will last Therefore the training program should continue ling after the target behavior has occurred, even when the program seems to produce no additional gains
Expose the target behavior to its natural reinforcers Ultimately, if changes in behavior produced by an intervention are to endure, then the behavior must be maintained by reinforcers it produces in the natural environment It is therefore important to expose the target behavior to those consequences For example, if a child initially put on her coat because of reinforcement from her parents, eventually you want to the behavior to come under the control of the natural consequences; the coat keeps her dry and warm.
Teach the person to obtain reinforcement for the target behavior Sometimes it is possible for the person to increase the availability of reinforcers One way to do this is to reinforce people who reinforce the behavior For example if Billie’s creative behavior is reinforced by a sticker, then Billie might reinforce the teachers behavior of getting a sticker, by saying ‘I really like stickers Dr Nic!’ Even more obvious techniques have been used ‘Hey Dr Nic, I've been paying attention, do I get a cookie? ’
Shape tolerance for delayed and uncertain reinforcement Important words for reinforcement are immediate and certain, however in the natural environment, things don’t often work as smoothly Therefore, if after the intervention ends there is an abrupt shift from immediate and certain to delayed and uncertain, there is likely to be an affect on the behavior To prevent this the intervention should include a gradual transition from nearly ideal reinforcement to those that resemble contingencies in the natural environment For example, you could gradually increase the interval between the target behavior and the delivery of reinforcement
Fade the intervention program At some point the intervention will have to be terminated In order to maintain the behavior, this should be done gradually For example, imagine someone was trained to live by themselves in a rehabilitation center. Once the training was done they might first live in a half way house, then an apartment where someone checked on their everyday, then a house where someone checked on them every week. Fading the intervention program narrows the gap between the intervention environment and the natural environment
Flashcard schedule of reinforcement a rule governing the delivery of reinforcers fixed ratio schedule provide a reinforcer after the target has occurred a number of times variable ratio schedule provide a reinforcer after the target has occurred a number of times, with the number varying around an average of N fixed interval schedule provide a reinforcer the first time the target behavior occurs after an interval of N length since the last reinforcement
Flashcard variable interval schedule provide a reinforcer the first time a target behavior occurs after an interval, with the interval varying around an average of N length fixed duration schedule provide a reinforcer after the target behavior has been performed continuously for a period of N length variable duration schedule provide a reinforcer after the target behavior has been performed continuously for a period, with the period varying around an of N length
Flashcard maintenance schedule a reinforcement schedule that maintains a target behavior as a desired rate ratio strain a reduction in the rate of target behavior and an increase In emotional behavior resulting from increases in the ratio of behavior to reinforcement stretching the ration gradually increasing the number of times a behavior must be performed to qualify for reinforcement
Flashcard contingency contract an agreement between 2 or more parties about what each is to do for the other token economy a form of contingency contract, usually involving a group of people, in which the reinforcers are tokens maintenance training intervention procedures that increase the likelihood that changes in a target behavior will persist when the intervention is ended
Скачать презентацию Behavior Analysis LECTURE 10 MAINTENANCE Review Last
85f849a66eaad614dc09e6226458541c.ppt