Chaining is a method for teaching sequential skills - skills that require several steps, accomplished in a set order. For example, all kinds of household jobs (doing laundry, cooking) and hygiene tasks (tooth brushing, shampooing hair) have multiple steps, with a particular order to them. So do many academic skills, such as multiplying two double-digit numbers in math class.
Chaining involves reinforcing individual responses occurring in a sequence to form a complex behavior. The chain is broken down into steps, or links, using task analysis. The chain initiated by a discriminative stimulus, whether it be a verbal prompt or a naturally occurring stimulus. When planning to teach a challenging sequential lesson, it's imperative to start with a task analysis - a well-reasoned set of written, ordered steps. For example, let's consider teaching a person, who seems to enjoy coffee, how to get a cup of coffee independently.
The two major types of chaining are backward chaining and forward chaining ...
When conducting backward chaining, the teacher provides substantial assistance, even hand-over-hand guidance, through the initial steps in the task analysis ...until she gets to the last step that the student can't do independently. It's on that step that prompting and then prompt fading methods are used, until the student completes the step independently (without prompts). Once that occurs regularly, the teacher moves backward and applies prompting and prompt fading to the next previous step that the student has not yet mastered. One big advantage to backward chaining is that the student's hardest work at learning a new step occurs near the end of the chain, and leads immediately to the reinforcers associated with completing the whole chain. When teaching a child to make French toast, for example, it's the final flips with a spatula that lead to a meal that smells great and is ready to eat. Teaching can then proceed backward to the messy steps of coating the bread with eggs!
Backward chaining does not involve teaching a behavior chain in reverse order, as the name suggests. Rather it involves reinforcing links in the chain, beginning at the back end of the chain and working toward the beginning. This is done such that in the initial stages of training, the final step is performed and a conditioned reinforcer is delivered. Eventually the trainer requires that the two final links of the chain must be performed in order to occasion the reinforcer, then three, four and so on until the entire chain has been taught. Through operant conditioning, the stimuli (physical changes to the environment) associated with completing each link in the chain becomes the discriminative stimulus for the link that comes after it. Similarly, as an earlier link becomes associated with the link that comes after it, that latter link takes on reinforcement value. When backward chaining is used, the effect of reinforcement associated with the last step is transferred "back up the chain," that is, in a backwards direction. Thus, the power of the positive reinforcer at the end of the chain is transferred down the line to the beginning of the chain. In this way, backward chaining has the advantage of starting with relatively access to a terminal reinforcer that works to condition reinforcement for each response added to the chain.
Forward chaining is what you might expect. It involves teaching the initial step first, with conditioned reinforcement following that first link. Instead of directing your teaching efforts on the last step that isn't done independently, you find the first step the child needs to learn and work forward through the task analysis. For instance, you may first teach a child to measure the right amount of laundry detergent on his own before proceeding to teach the next steps in order. In forward chaining, the teacher then usually guides the child through the remaining not-yet-learned steps in the task analysis.
Total task presentation is a variation of forward chaining, but with the addition of supplemental reinforcers after each step. This method has the advantage of stronger reinforcement for each step. However, the practitioner should be cautious to ensure the terminal reinforcer is the strongest. Moreover, this method requires that the supplemental reinforcers be faded. For example, social reinforcement (praise) might be applied after each link in a chain is performed. Perhaps the sequence is learned more rapidly. However, over time, the practitioner must apply less praise, and allow the terminal reinforcer to take over.
Unfortunately, very little research exists on which of the chaining approaches works best for whom, or for which different types of tasks. You will want to base your decision on your observations and analysis of the child's behavior, and by trying it out!