What is a task analyses?
A task analysis is used by behavior analysts to break a complicated task down into a more easily accomplishable sequence of smaller, simpler steps.
Chaining:
This is typically referred to as a behavior chains: a specific sequence of responses that serve as individual members of a chain; each step serving as the reinforcer for the previous step and signaling the availability of reinforcement for completing the next response in the chain. The last step in the chain contacts the "terminal reinforcer" that maintains the whole chain. The terminal reinforcer for a behavior chain such as "making a sandwich" would be, of course, getting to eat the sandwich!
It can be helpful to think of task analyses / behavior chains as "recipes" - they're sort of like "behavioral recipes!" that produce consequences in the environment instead of food (although, of course, sometimes the consequence of our behavior is food).
General Teaching Strategies
There are three general methods for teaching a task analysis, each referring to when / how teaching will be organized:
-
Total or Whole Task Teaching refers to the teaching all of the steps in a task analysis and providing reinforcement differential to independence on any given step(s) in the task analysis.
-
Forward Chaining refers to starting at the initial step of a task analysis and providing prompting / reinforcement until that step is mastered. All subsequent steps are typically "teacher-completed," i.e., the teacher provides hand-over-hand assistance on all un-mastered steps. When the first step is mastered, the second step is introduced into teaching until all steps are mastered, from start --> finish!
-
Backwards Chaining refers to starting at the very last step of a task analysis and providing prompting / reinforcement until that step is mastered, and then introducing the next-to-last step into teaching after that, and so on, until all steps are mastered. The teacher will also complete any steps that are not mastered / currently in teaching. This type of chaining is used frequently when there may be some benefit to having the learner contact the "terminal step" from the get-go: e.g., if your learner likes to run around outside, having them complete only the last step to tying their shoes helps establish a learning history that compliance and completion of the shoe-tying task leads to an awesome time outside on the playground.
Limited Holds
Limited Holds are a special type of behavior chain. In behavior chains with limited holds, not only must the steps be performed correctly, but also within a specified period of time to produce the terminal reinforcer. For example, imagine a head chef working at a busy restaurant. In order to produce reinforcement for himself and the other cooks, he must not only complete all the steps necessary to produce a dish, but do so in a reasonable amount of time. If he fails to produce the dishes quickly, the clientele might become angry and leave the restaurant and / or never come back. The limited hold refers to the reinforcer (in this case, satisfaction of the customer) being only available for a limited period of time from the inception of the chain (creating the dish).
Another example could be a factory worker at a Widget factory. In order to satisfy his manager, the factory worker must create 30 widgets / hour. If the factory worker fails to produce the widgets at the required rate, he will not meet his production goal and may anger his boss and lose his job. If, however, he meets his production goal, he may receive a "bonus" to his weekly salary for his hard work. The reinforcement - the "bonus" to his weekly check - is on a limited hold contingent upon the rate of his widget production.
An Example Task Analysis:
Back to our original example of a task analysis - making a sandwich; here are some example steps:
- Open a bag of bread
- Take two slices out
- Lay them on a plate, side-by-side
- Get the jar of peanut butter
- Get a knife
- Spread the peanut butter on one side of the bread
- Wipe the knife off in the sink, and keep it with you
- Put the jar of peanut butter back
- Get a jar of jelly
- Spread some jelly on the other side of the bread
- Wipe the knife off in the sink and put it on the dry rack
- Put the two sandwich slices together
- Enjoy!
In the task analysis above, you might notice that some of the steps (such as steps 4 - 10) can be completed in a different order and still achieve the same effect. For instance, you could absolutely put the jelly on before the peanut butter. However, it would make little sense to do step 12 before doing steps 2 - 10. Sometimes, the logical order of the steps may be insensitive to change (that is, you can mix them up), or it may be sensitive to change (that is to say if you mix them up, the terminal reinforcer will change in quality or be unavailable).
How do you build a task analysis?
Generally, there are three ways one could create a task analysis.
- Observe a competent performer: observe and describe each discrete step a competent performer makes while engaging in a task analysis
- Ask an Expert: if you can't or don't have the opportunity to perform the task, ask someone experience with completing the task to write the steps down for you. In keeping with our "recipe" metaphor, a competent performer might be an experience chef. You might not have the resources (money, time) to practice baking a dish and, instead, might prefer to just ask an expert chef to do it for you.
- Do it yourself: Same as #1, except, you do it yourself, pausing to write each step down as you see fit. It can be trickier to do this in some cases because pausing to interrupt the chain may be impractical in some circumstances.
And that's it! Got any questions? If so, please feel free to leave a comment!