Shaping, Chaining, and Generalization

Slide 1: 

Section 3: Shaping, Chaining, and Generalization Created by Shane Alan Community Integrated Work Program

Slide 2: 

Section 3: Shaping, Chaining, and Generalization Created by Shane Alan Community Integrated Work Program

Overview : 

Overview Outline of skills training tools Shaping Topography Duration Frequency/rate Chaining Behavior chain Forward chaining Backward chaining Global chaining Step-by-step and whole-task presentation Generalization Review 1

Slide 4: 

The cue, the prompt, and the reaction focus on the events surrounding the behavior. Shaping, chaining, and generalization focus on the behavior. Shaping and chaining are types of skills training programs, so they each use cues, prompts, and reactions. Shaping and chaining are used when cues, prompts, and reactions by themselves do not reliably produce the desired behavior. Generalization should be factored into every skills training program. 2

Slide 5: 

In technical terms, shaping is the process of reinforcing successive approximations of a terminal behavior. To understand this, let’s look at each of the key terms. The terminal behavior is the behavior we want to see at the end of the intervention period (the intervention is terminated when the behavior is exhibited). An approximation of a terminal behavior is a behavior that resembles the terminal behavior, but is not equal to the terminal behavior. 3

Slide 6: 

Successive approximations of the terminal behavior are behavioral responses that gradually become closer to the terminal behavior. Reinforcing successive approximations of the terminal behavior means that we will provide reinforcement contingent on the learner exhibiting a behavior that is closer to the terminal behavior than the previous behavior was. By strengthening behaviors that the learner is already able to exhibit we can help them gradually come closer to and finally reach the desired skill. 4

Slide 7: 

Shaping generally focuses on one of three behavioral characteristics: Topography – this is what the behavior actually looks like. For example, the topography of wiping your mouth might be defined as “grabbing a napkin, bringing your hand to your mouth, and moving the napkin across your mouth”. Duration – this is how long the behavior occurs for. Frequency/rate – this is how often the behavior occurs. 5

Slide 8: 

To shape the topography of the behavior, you start by finding a behavior that the learner can already do that is physically similar to the behavior you want to see. For example, if you are teaching a learner to wipe off a table, you could first start with the behavior of putting their hand on the table. Once you know where you want to start, you outline benchmarks that gradually bring the learner closer to the desired behavior. Using the example above, we could set bench marks of moving hand across the table, then move hand around table in circle, then move hand around table until entire surface has been wiped. 6

Slide 9: 

When the learner has mastered the desired behavior (meaning that the behavior occurs at a pre-set level following the cue and using a pre-set maximum prompt), you will stop reinforcing the behavior (you put it in extinction) and begin providing reinforcement contingent on the next benchmark. This continues in a pre-planed and systematic way until the learner is consistently able to engage in the behavior following the cue. Once the learner has mastered the desired behavior you will begin to fade out the prompt and reinforcement. 7

Slide 10: 

We will shape the duration of a behavior when the learner is able to engage in the behavior, but not for as long as you would like. We start by reinforcing the duration that the learner can already meet. For example, if a learner is able to sort items as part of a vocational program, but only does so for 5 minutes at a time, we can can shape the behavior until it occurs for the desired 30 minutes. Once we know where we want to start, we outline benchmarks that gradually bring the learner closer to the desired behavior. Using the example above, we could set benchmarks of 10 minutes, 15 minutes, 20 minutes, 25 minutes, and 30 minutes. 8

Slide 11: 

When the learner has mastered the desired duration (meaning that the behavior occurs for a pre-set duration following the cue and using a pre-set maximum prompt), you will stop reinforcing the behavior at that duration (you put it in extinction) and begin providing reinforcement contingent on the next benchmark. This continues in a pre-planed and systematic way until the learner is consistently able to engage in the behavior at the desired duration following the cue. Once the learner has mastered the desired duration you will begin to fade out the prompt and reinforcement. 9

Slide 12: 

We shape frequency/rate when the learner is able to engage in the desired behavior, but does so too infrequently. We start by reinforcing the frequency/rate that the learner can already meet. For example, if a learner is able stuff envelopes, but only stuffs 5 every 10 minutes we can shape the behavior until the learner is able to stuff the desired 30 envelopes in 10 minutes. Once you know where you want to start, you outline benchmarks that gradually bring the learner closer to the desired behavior. Using the example above, we could set bench marks of 10 in 10 minutes, 15 in 10 minutes, 20 in 10 minute, 25 in 10 minutes, and 30 in 10 minutes. 10

Slide 13: 

When the learner has mastered the desired frequency/rate (meaning that the behavior occurs at a pre-set frequency/rate following the cue and using a pre-set maximum prompt), you will stop reinforcing the behavior at that frequency/rate (you put it in extinction) and begin providing reinforcement contingent on the next benchmark. This continues in a pre-planed and systematic way until the learner is consistently able to engage in the behavior at the desired frequency/rate following the cue. Once the learner has mastered the desired Frequency/rate you will begin to fade out the prompt and reinforcement. 11

Slide 15: 

Most skills are made up of several behaviors where the consequence of one behavior serves as the cue for another behavior. The relationship between one behavior’s consequence and another behavior’s cue is call sequencing. Sequencing comes easily for most neuro-typical adults. However, for individuals with pervasive and intellectual disabilities, this can be quite challenging. 12

Slide 16: 

Cue Behavior Consequence

Slide 17: 

Chaining is the process of treating individual behaviors in a behavior chain as separate skills to be mastered individually. Once an individual behavior in a chain has been mastered it can more effectively be linked with another behavior in the chain. Each step of the chain is taught using a cue-prompt-reaction procedure described in the previous training. 13

Slide 18: 

There are three chaining programs: Forward chaining teaches the behavior chain from the first step. This is best when the learner works well in a sequential order, and when the easiest steps are at the beginning of the chain. Backward chaining teaches the behavior chain from the last step. This is best when the learner works well with immediate reinforcement and when the easiest steps are at the end of the chain. Global chaining teaches the behavior chain from the easiest or most important step. This works best when there is a large difference in the difficulty of various steps in the chain. 14

Slide 19: 

Any of the three chaining programs can use either step-by-step presentation or whole-task presentation. With step-by-step presentation we teach the targeted step alone or only with another step that has already been mastered. The learner is not requested to go through the remaining steps of the chain until the current one is mastered. With whole-task presentation we request that the learner engage in the entire behavior chain but we only target one step at a time for training. 15

Slide 20: 

When starting a chaining program we first conduct a task analysis of the behavior. This provides us with the specific steps in the chain. To do this we observe the learner as they currently go through the behavior chain and outline the various steps. Any behavior chain can have several different task analyses depending on the needs of the learner. The behavior chain should be broken down into steps that the individual learner is likely to be able to complete with a cue-prompt-reaction style training. 16

Slide 21: 

Once you have the steps outlined in the task analysis, choose which chaining program and presentation to use. Programs: Forward chaining Backward chaining Global chaining Presentation: Step-by-step Whole-task 17

Slide 22: 

Once you know your steps/benchmarks and what program/procedure you will use, outline the cue-prompt-reaction program for each individual step and outline the criteria to move from one step to the next. For example, using chaining to teach someone to wash their hands could include a plan to fade from full-physical prompts to partial-physical, then to gesture prompts, and finally independence for each of the steps. If the learner is able to successfully complete the step at the current prompt level for 3 consecutive trails they will move on the the next prompt level (or next step when they are independent at the step). 18

Slide 23: 

The techniques outlined so far in these trainings have been proven effective to teach specific, targeted, behaviors and skills. However, in the real world we have to be able to respond to different situations by adjusting our behavior. If we cannot do this, the skill might not actually be mastered. For example, if a learner becomes independent in washing their hands using only single-handled faucets and drying their hands on cloth towels, they might still be completely prompt dependent when presented with a two-handle faucet and paper towels. 19

Slide 24: 

To help the learner develop the ability to adjust their behavior to different stimuli, we need to train across stimuli at all levels. This is called generalization. There are three categories of stimuli that we target for generalization: People Settings Tools 20

Slide 25: 

It can be very easy to become dependent on people when learning a new skill. Remember that when you start a new skills training program you are introducing reinforcement contingent on a behavior that has not been reinforced in the past. By doing this, you are also adding yourself as a contingency for reinforcement (remember, reinforcement only occurred after you started intervention). Most children (even many adult children) respond differently to situations depending on if they are around their mother or their father, because they have learned that various behaviors have different consequences around different people. 21

Slide 26: 

To generalize across people, we need to reduce “people” as a contingency of reinforcement. Ensure that EVERY person who will be responsible for training the skill does so according the the agreed-upon plan. If one trainer is reinforcing one step in a behavior chain while another trainer is reinforcing a different step, the learner is likely to confuse when they should engage in the behavior and may only respond correctly when cued by a certain trainer. To the extent possible, have several people provide training on the skill during intervention. The more people that deliver reinforcement contingent on the behavior, the more generalized the behavior will be. 22

Slide 27: 

The setting is both the environment in which the behavior occurs and the time at which it occurs. If a skill is only taught in a certain location or at a certain time of day / day of the week, the learner is likely to view the setting as a contingency for reinforcement and might only exhibit the skill in that setting. To generalize settings, we should teach the skill in several different settings, including both environments and times. 23

Slide 28: 

The tools are the actual mechanics involved in a skill. Very often a skill must be slightly modified to fit different tools. For example, there are countless examples of different ways of turning on a water fountain or using a sink to wash your hands. Tools are probably the most challenging thing to generalize because we cannot possibly factor in every type of tool. 24

Slide 29: 

Like other types of generalization, we start by exposing the learner to as many different types of tools as possible. For skills that are likely have many different tools, the learner should be taught the skill using as indirect of methods as possible, even if the skill is acquired slightly more slowly this way. For example, when teaching someone to wash their hands using a direct verbal prompt, we should state things such as “turn the water on” instead of “pull the faucet up”, and allow the learner to discover how to turn the water on if possible. 25

Slide 30: 

Shaping and chaining are used when the cue-prompt-reaction style of training by itself does not reliably produce the desired behavior. Shaping is used to gradually turn a behavior that the learner has already mastered into a more functional behavior. Chaining is used to train individual steps in a behavior chain and then link the steps together. Generalization is the ability for a skill to be used across people, settings, and tools. Generalization should be factored into every skills training program before skills training starts. 26

Slide 31: 

The End