Reinforcing stimuli are environmental events that occur after a behavior that increase the likelihood of that behavior occurring in the future. Treats, praise, special privileges, music, toys, virtually anything can be used as reinforcement, if it serves to increase the occurrences of a particular behavior.

Primary and Secondary Reinforcing Stimuli

There are two varieties of reinforcing stimuli: primary reinforcing and secondary reinforcing stimuli.

Primary reinforcing stimuli (S^R) are unconditioned, that is they are events or rewards whose value we intrinsically realize: food, water, sex, warmth, etc. (I'll leave it up to you, reader, to determine which we're not likely to use in an educational setting).

The use of primary reinforcement has a few advantages. Primary reinforcement is intrinsic, it does not have to be taught. While some children do not yet find much value in hugs or praise, and consequently are not yet willing to work too strenuously to gain them, most children have a favorite snack or treat that they might be willing to do a little something for. Also, primary reinforcement will not extinguish, it will always remain reinforcing. One will always desire food or water. Primary reinforcement has some disadvantages however. Primary reinforcers are subject to satiation after a relatively short period of time. Believe it or not, there are only so many M&Ms that one can eat. And primary reinforcement is not especially representative of the natural environment. There are few situations in life where one is given a goldfish cracker every time he uses the right phrase to describe a picture.

Secondary reinforcing stimuli (S^r) are conditioned. These are intrinsically neutral stimuli that, through their association with primary, or other secondary reinforcing stimuli, have become reinforcing. There are two types of secondary reinforcement. The first type consists of praise, smiles, a sense of accomplishment, those reinforcers that are somewhat social in origin. Again, these are essentially neutral stimuli which have become reinforcing through their association with other more primary reinforcement. The second type of secondary reinforcement is a token economy. A token economy system is one in which a person earns tokens for desirable behaviors. Each token is a step towards acquisition of a primary reinforcer or more reinforcing secondary reinforcement. Money is the classic example of a token economy system. There is nothing particularly reinforcing about little, green bits of paper, but we are willing to work for those green bits because they allow us to obtain things that are quite reinforcing: food, a home, a DVD player, those lovely strappy-backed heels in the window at Nordstrom's. I've more on token economy systems in the section on visual supports.

Secondary reinforcement has a number of advantages. First, conditioned reinforcers are typically more convenient to use than primary. It is much easier to tell a child "Nice job!" or to give them a star on their math paper, than to carry around a pocketful of Cheez-its or candy at all times. In that vein, secondary reinforcement typically fits in much better in a natural environment. We all receive encouragement, praise, or tokens throughout our everyday lives. Secondary reinforcement lessens the need for proximity to a child. If one is constantly having to offer treats or toys while a child is working, one must remain at his elbow almost constantly. When one can offer smiles or a call of "Good job!" from across the room, it is typically far more convenient and much more reflective of natural environments. Secondary reinforcement can be used to broaden the interests of the child. By paring new or less reinforcing activities, foods, or objects with those that are more reinforcing, they can themselves become more desirable to the child. Finally, using a token economy system can increase the length of time between presentation of reinforcers, allowing for longer periods of instruction, with fewer interruptions, and a lessened likelihood of satiation. If secondary reinforcement has a disadvantage, it is that they must be taught, and occasionally maintained by pairing it once again with a primary reinforcer to reestablish interest.

When teaching kids with autism, any delivery of primary reinforcement should be paired with secondary reinforcement, especially social reinforcers: praise, eye contact, high fives, hugs, brief games. Some kids with autism do not yet find many of these things terribly reinforcing. It is our job to teach them to enjoy these things as much as possible.

Positive and Negative Reinforcement

In addition to primary and secondary, reinforcement can be divided into positive and negative.

Positive reinforcement (S^R and S^r) is the presentation of positive events after a particular behavior to increase the likelihood that the behavior will occur in the future. The examples that have been given throughout this section -- treats, hugs, toys, games, etc. -- are examples of positive reinforcement.

Negative reinforcement (S^-R and S^-r) is the removal of aversive events after a particular behavior to increase the likelihood that the behavior will occur in the future. Now, the word "aversive" often sends up red flags with many people, and rightfully so in many cases. Physical aversives used as punishments are certainly to be avoided. But less extreme aversives might have their place. As an example, after I pull into the school parking lot in the morning and turn off my car, more often then not it starts dinging at me, reminding me that once again I have forgotten to turn my lights off. Also, about 5:45 each morning, the most aversive event in the history of the world occurs when my alarm clock goes off. Each of these are somewhat unpleasant, but certainly not harmful, stimuli which are removed when I perform the intended response: to save the car battery by turning of the headlights, or to wake up and get ready for work, respectively. Likewise, assume that a child is somewhat hypersensitive to sound and finds loud speaking voices slightly discomforting. Currently, he yells at the person "Be quiet!" to attempt to get him to soften his voice. The child's teacher could set up situations wherein he (the teacher) talks too loudly (a mildly aversive stimulus). As soon as the child uses a more correct phrase "Could you talk more quietly, please?" the teacher adjusts the volume of his voice. Again, this is an example wherein the production of a desired behavior results in the removal of a less-than-pleasurable stimulus.

Also, one should be aware that positive and negative reinforcement are at work when children "misbehave," too. For instance, if a child is screaming in the toy store for the newest Pokemon card set, and his parent finally breaks down to buy the child the set to stop his screaming, than the parent has just positively reinforced that behavior, and the child is more likely behave similarly next time to obtain a desired item.

But look how negative reinforcement might play into the above scenario, too. The child is presenting a stimuli which the parent finds rather aversive (the screaming). When the parent performs the response required of the child, the child turns off the screaming, negatively reinforcing the behavior of the parent, and possibly making it more likely that the parent will behavior similarly in the future.

Differential Reinforcement

Reinforcement should be differential. Differential reinforcement involves reinforcing most any positive response (successively closer approximations of the correct behavior, for example) to some degree, but providing very strong reinforcement when the child completely exhibits the target behavior or skill. Take, for example, the earlier scenario of teaching Joey to receptively identify the color blue. If Joey were to require hand-over-hand assistance and was shaped to point to the blue cube, he might be reinforced with, "Yes. That is blue." If Joey needed only a touch on the shoulder to remind him to touch the blue cube, the teacher might give him a high five and say, "You got it! Nice job." But, if Joey independently points to the blue cube, the teacher might pull out all the stops. "Woo-hoo! You're right! Great job." She may give him big hugs and a few goldfish crackers. Of course, what one child may find reinforcing another may find very uncomfortable. Some kids may not find the drama and volume of that last example reinforcing at all. For them, a simple, "Great job, Joey," and an offer of the cracker (or a chance to play with a toy train, or whatever) might have been much more comfortable and reinforcing.

Error Correction

Error correction procedures should be well defined. If an incorrect response is given, typically it is correct with a simple, neutral "No," or "Nope" and the trial is often repeated to give the child another chance. If the child makes a number of incorrect responses, the level of prompting may need to be increased. If the child fails to make any appropriate response whatsoever, the prompting level is again increased, and the efficacy of the reinforcer is re-examined.

Guidelines for Reinforcement

There are a few additional guidelines that may prove useful when dealing with reinforcements:

If the reinforcement is to be consistent and effective the criteria for the response need to be planned out in detail, understood, and used consistently by everyone involved in the child's program.

Consequences for correct and incorrect responses should be easily distinguishable. This may seem obvious, but is often harder than it sounds. For instance, I happen to have a great affinity for naughty kids. So, for example, when James takes his shoes off in class, which he does regularly, and I know that he is just doing it to be naughty, I often crack a smile, and tell him jovially, "C'mon James. Put your shoes on." Is this a bad thing? Well, if I really wanted James to keep his shoes on, it probably was not the best thing I could do. Since I had determined that James takes his shoes off to be naughty -- that is, to gain the attention of an adult -- I gave him just what he wanted by having fun with it, smiling at him, etc. I did ask him to put his shoes back on, but the manner in which I did it really reinforced the undesirable behavior (the removal of his shoes) rather than the desirable. If I was truly concerned about James's shoes staying on, a better approach might have been to say, with a straight face, in a flat tone, "No James. Put your shoes on." This takes some of the fun out of this particular naughty behavior, lessening the chances that it may happen again. Of course, if that behavior has been reinforced for a long time, knowingly or not, it may take an equally long time to establish that it is no longer going to be reinforced. Now, am I likely to be so concerned about James's taking his shoes off that I am going to eliminate a chance for him to have fun with me? Not on your life.

If reinforcement is being presented after each trial (see schedules of reinforcement) short-lived reinforcers should be used. A single M&M (baking M&Ms are great as well), one blow on the bubble wand, a quick high five and a "Great job!" are all good choices. Longer lasting rewards -- time on the computer, a chance to play with playdough, gummy bears -- may detract from the inter-relatedness of consecutive trials. These rewards are great, however, as a break after a series of trials, or as a big reward "paid for" with a token economy system.

The next section is quick look at schedules of reinforcement.

[ Previous | Contents | Next ]