6.6: Stimulus Control

Section Learning Objectives

• Define and exemplify stimulus control.
• Define stimulus discrimination.
• Define and describe discrimination training.
• Define and clarify why stimulus generalization is necessary.
• Describe generalization training and the strategies that can be used.
• Revisit the definition of discriminative stimuli.
• Clarify how stimuli or antecedents become cues.
• List and describe the 6 antecedent manipulations.
• Define prompts.
• List, describe, and exemplify the four types of prompts.
• Define fading.
• List and describe the two major types of fading and any subtypes.
• Define and exemplify shaping.
• Outline steps in shaping.

Stimulus Control

When an antecedent (i.e. stimulus) has been consistently linked to a behavior in the past, it gains stimulus control over the behavior. It is now more likely to occur in the presence of this specific stimulus or a stimulus class, defined as antecedents that share similar features and have the same effect on behavior. Consider the behavior of hugging someone. Who might you hug? A good answer is your mother. She expects and appreciates hugs. Your mother is an antecedent to which hugging typically occurs. Others might include your father, sibling(s), aunts, uncles, cousins, grandparents, spouse, and kids. These additional people fall under the stimulus class and share a similar feature of being loved ones. You could even include your bff. What you would not do is give the cashier at Walmart a hug. That would just be weird.

Do you stop when you get to a red octagonal sign? Probably, and the Stop sign has control over your behavior. In fact, you do not even have to think about stopping. You just do so. It has become automatic for you. The problem is that many of the unwanted behaviors we want to change are under stimulus control and happen without us even thinking about them. These will have to be modified for our desired behavior to emerge.

Stimulus Discrimination

We have established that we will cease all movement of our vehicle at a red octagonal stop sign and without thinking. A reasonable question is why don’t we do this at a blue octagonal sign, ignoring the fact that none exist? Stimulus discrimination is the process of reinforcing a behavior when a specific antecedent is present and only it is present. We experience negative reinforcement when we stop at the red octagonal sign and not a sign of another color, should a person be funny and put one up. The NR, in this case, is the avoidance of something aversive such as an accident or ticket, making it likely that we will obey this traffic sign in the future.

Discrimination training involves the reinforcement of a behavior when one stimulus is present but extinguishing the behavior when a different stimulus is present. From the example above, the red stop sign is reinforced but the blue one is not.

In discrimination training we have two stimuli:

• The SD or discriminative stimulus whose behavior is reinforced.
• and an SΔ (S-delta) whose behavior is not reinforced and so is extinguished.

When a behavior is more likely to occur in the presence of the SD and not the SΔ, we call this a discriminated behavior. And this is where stimulus control comes in. The discriminated behavior should be produced by the SD only. In terms of learning experiments, we train a pigeon to peck an oval key but if he pecks a rectangular one, no reinforcer is delivered.

Stimulus Generalization

As a stimulus can be discriminated, so too can it be generalized. Stimulus generalization is when a behavior occurs in the presence of similar, novel stimuli and these stimuli can fall on a generalization gradient. Think of this as an inverted u-shaped curve. The middle of the curve represents the stimulus that we are training the person or animal to respond to. As you move away from this stimulus, to the left or right, the other stimuli become less and less like the original one. So, near the top of the inverted U, a red oval or circle will be like a red octagon but not the same. Near the bottom of the curve, you have a toothbrush that has almost zero similarity to a stop sign.

In behavior modification (the applied side of learning), we want to promote generalization meaning that if we teach someone how to make a desirable response in a training situation, we want them to do that in all relevant environments where that behavior can occur, whether that be a child in a classroom, at home at the dinner table or in his/her bedroom, on the playground at recess, at the park, with the grandparents, etc. This is called generalization training and is when we reinforce behavior across situations until generalization occurs for the stimulus class. The desirable behavior should generalize from the time with a therapist or applied behavior analyst and to all other situations that matter. To make this happen you could/should:

• Always reinforce when the desirable behavior is made outside of training. By doing this, the desirable behavior is more likely.
• Teach other people to reinforce the desirable behavior such as teachers and caregivers. The therapist cannot always be with the client and so others have to take control and manage the treatment plan. Be sure they are trained, understand what to reinforce, and know what the behavioral definition is.
• Use natural contingencies when possible. Let’s say you are trying to teach social skills to a severely introverted client. In training, she does well and you reinforce the desirable behavior. Armed with new tactics for breaking the ice with a fellow student in class, she goes to class the next day and strikes up a conversation about the weather or the upcoming test. The fellow student’s response to her, and the continuation of the conversation, serve as reinforcers and occur naturally as a byproduct of her initiating a conversation. Another great example comes from a student of mine who was trying to increase her behavior of eating breakfast before class. She discovered that she felt more alert and energetic when she ate breakfast then when she did not, which are positive reinforcers, and naturally occurring. In fact, she was so happy about this, she jumped four goals and went from her initial goal of eating before class two times a week, to eating breakfast 6-7 times a week. Her behavior generalized beyond simply eating before class to eating breakfast every day when she woke up. It should be noted that her distal goal was 5 days, so in her first week of treatment she had already exceeded this goal. Way to go.
• Practice making the desirable response in other environments during training. You can achieve this by imagining these environments, role-playing, or setting up the environments to some extent.
• Related to the previous strategy, use common stimuli that are present in other environments as much as possible. An example is a stuffed animal that a child has at home. Or have the special education teacher bring the child’s desk to the training environment and have them sit in it.
• Encourage the client to use cues to make the desirable response outside of the training environment. These are reminders to engage in the correct behavior and can be any of the antecedent manipulations already discussed.

Stimulus Control Procedures: Antecedent Manipulations

One critical step is to exert control over the cues for the behavior and when these cues bring about a specific behavior, which, if you recall, are termed discriminative stimuli (also called an SD). So, what makes an antecedent a cue for a behavior? Simply, the behavior is reinforced in the presence of the specific stimulus and not reinforced when the stimulus or antecedent is not present.

The strategies we will discuss center on two ideas: we can modify an existing antecedent or create a new one. With some abusive behaviors centered on alcohol, drugs, nicotine, or food, the best policy is to never even be tempted by the substance. If you do not smoke the first cigarette, eat the first donut, take the first drink, etc. you do not have to worry about making additional problem behaviors. It appears that abstinence is truly the best policy.

But what if this is not possible or necessary? The following strategies could be attempted:

• Create a Cue for the Desirable Behavior — If we want to wake up in the morning to go to the gym, leave your gym clothes out and by the bed. You will see them when you wake up and be more likely to go to the gym. If you are trying to drink more water, take a refillable water bottle with you to classes. Hiking around campus all day can be tough and so having your water bottle will help you to stay hydrated.
• Remove a Cue for the Undesirable or Problem Behavior — In this case, we are modifying an existing antecedent/cue. Let’s say you wake up in the morning, like I do, and get on your phone to check your favorite game. You initially only intend to spend a few minutes doing so but an hour later you have done all the leveling up, resource collecting, candy swiping, structure building, etc. that you can and now you do not have the time to do a workout. In this case, phone use is a problem behavior because it interferes or competes with the execution of the desirable behavior of going to the gym. What do you do? There is a simple solution – do not leave your phone by your bed. If it is not in the room, it cannot be a reminder for you to engage in the problem behavior. The phone usage in the morning already exists as a behavior and the phone serves as a cue for playing games. You enjoy playing the games and so it is reinforcing. If the phone is not present, then the behavior of playing the game cannot be reinforced and the cue loses its effectiveness. In the case of water, if we do not carry tea with us, we cannot drink it, but can only drink our water bottle, thereby meeting our goal.
• Increasing the Energy Needed to Make a Problem Behavior — Since the problem behavior already exists and has been reinforced in the past, making its future occurrence likely in the presence of the stimulus, the best bet is to make it really hard to make this unwanted behavior. Back to the gym example. We already know that our phone is what distracts us and so we remove the stimuli. One thing we could do is place the phone in the nightstand. Out of sight. Out of mind, right? Maybe. Maybe not. Since we know the phone is in the nightstand, we could still pull it out in the morning. If that occurs, our strategy to remove the cue for phone usage fails. We can still remove it, but instead of placing it in the nightstand, place it in the living room and inside our school bag. So now it is out of sight, out of mind, but also far away which will require much more physical energy to go get than if it was in the nightstand beside us. Think about this for a minute. The strategy literally means that we expend more energy to do the bad behavior, than…..
• Decreasing the Energy Needed to Engage in the Desirable Behavior — …we would for the good behavior. Having our clothes by our bed is both a cue to go to the gym, but also, by having them all arranged in one place, we do not have to spend the extra time and energy running around our bedroom looking for clothes. We might also place our gym bag and keys by the door which saves us energy early in the morning when we are rushing out to the gym. What about for drinking water? Instead of carrying a water bottle with us we could just drink water from the water fountains at school. Okay. But let’s say that you are standing in the hallway and the nearest water fountain is all the way up the hallway and near the door to exit the building. You have to walk up the hall, bend over, push the button, drink the water, remove your hand from the fountain, walk back down the hall, re-enter the classroom, and then take your seat. Not too bad, right? WRONG. If you had your water bottle in your backpack, you would only need to reach down, pick it up, open the bottle, take a drink, cap the bottle, and set it back down on the floor or on the desk. You never have to leave your seat which means you are making far fewer behaviors in the overall behavior of drinking water, and so expending much less energy. Now you can use this energy for other purposes such as taking notes in class and raising your hand to ask a question.

Another way you can look at antecedents is to focus on the consequences. Wait. What? Why would that be an antecedent manipulation? Consider that we might focus on the motivating properties of the consequence so that in the future, we want to make the behavior when the same antecedent is present. Notice the emphasis on want. Remember, you are enhancing the motivating properties. How do we do this? From our earlier discussion, we know that we can use the motivating operations of establishing and abolishing operations. See Section 6.3.5 for the discussion. But they make up the last two antecedent manipulations that can be employed to bring about the desired behavior.

Transfer of Stimulus Control: Prompting and Fading

One way to help a response occur is to use what are called prompts, or a stimulus that is added to the situation and increases the likelihood that the desirable response will be made when it is needed. The response is then reinforced. There are four main types of prompts:

• Verbal — Telling the person what to do
• Gestural — Making gestures with your body to indicate the correct action the person should engage in
• Modeling — Demonstrating for the person what to do
• Physical — Guiding the person through physical contact to make the correct response

These are all useful and it is a safe bet to say that you have experienced all of them at some point. How so? Let’s say you just started a job at McDonald’s. You were hired to work the cash register and take orders. On your first day, you are assigned a trainer and she walks you through what you need to do. She might give you verbal instructions as to what needs to be done and when, and how to work the cash register. As you are taking your first order on your own, you cannot remember which menu the Big Mac meal fell under. She might point in the right area which would be making a gesture. Your trainer might even demonstrate the first few orders before you take over so that you can model or imitate her later. And finally, if you are having problems, she could take your hand and touch the Big Mac meal key, though this may be a bit aversive for most and likely improper. The point is that the trainer could use all of these prompts to help you learn how to take orders from customers. Consider that the prompts are in a sort of order from the easiest or least aversive (verbal) to the hardest or most aversive (physical). This will be important in a bit.

It is also prudent to reinforce the person when they engage in the correct behavior. If you told the person what to do, and they do it correctly, offer praise right away. The same goes for them complying with your gesture, imitating you correctly, or subjecting themselves to a physical and quite intrusive or aversive prompt.

When you use prompts, you also need to use what is called fading, which is the gradual removal of the prompt(s) once the behavior continues in the presence of the SD. Fading establishes a discrimination in the absence of the prompt. Eventually, you transfer stimulus control from the prompt to the SD.

Prompts are not a part of everyday life. Yes, you use them when you are in training, but after a few weeks, your boss expects you to take orders without even a verbal prompt. To get rid of prompts, you can either fade or delay the prompts. Prompt fading is when the prompt is gradually removed as it is no longer needed. Fading within a prompt means that you use just one prompt and once the person has the procedure down, you stop giving them a reminder or nudge. Maybe you are a quick study, and the trainer only needs to demonstrate the correct procedure once (modeling). The trainer would simply discontinue use of the prompt.

You can also use what is called fading across prompts. This is used when two or more prompts are needed. Maybe you are trying to explain an algebraic procedure to your child who is gifted in math. You could start with a verbal prompt and then move to gestural or modeling if he/she has a bit of an issue. Once the procedure is learned, you would not use any additional prompts. You are fading from least to most intrusive. But your other child is definitely not math-oriented. In this case, modeling would likely be needed first and then you could drop down to gestural and verbal. This type of fading across prompts moves from most to least intrusive. Finally, prompt delay can be used and is when you present the SD and then wait for the correct response to be made. You delay delivering any prompts to see if the person engages in the desirable behavior on their own. If he or she does, then no prompt is needed, but if not, then you use whichever prompt is appropriate at the time. For instance, you might tell your child to do the next problem and then wait to see if he/she can figure it out on their own. If not, you use the appropriate prompt.

Shaping

Sometimes there is a new(ish) behavior we want a person or animal to make but they will not necessarily know to make it, or how to make it. As such, we need to find a way to mold this behavior into what we want it to be. The following example might sound familiar to you. Let’s say you want a friend to turn on the lights in the kitchen. You decide not to tell them this by voice but play a game with them. As they get closer to the light switch you say “Hot.” If they turn away or do not proceed any further, you say “Cold.” Eventually, your statements of “Hot” will lead them to the switch and they will turn it on which will lead to the delivery of a great big statement of congratulations. “Hot” and “Thank you” are reinforcers and you used them to make approximations of the final, desired behavior of turning on the light. We called this ‘hot potato-cold-potato’ when we were a kid but in applied behavior analysis, this procedure is called shaping by successive approximations or shaping for short.

To use shaping, do the following:

• Identify what behavior you want the person or animal to make. Be sure you create a precise and unambiguous behavioral definition.
• Determine where you want them to start. This can be difficult but consider what others have done for the same problem behavior. When all else fails, start very low and make your steps small. More frequent reinforcement will help you too.
• Determine clear shaping steps; the successive approximations of the final behavior.
• Identify a reinforcer to use and reinforce after reaching the end of each step. This steady delivery of reinforcers, due to successfully moving to the next step, is what strengthens the organism’s progression to the final, target behavior.
• Continue at a logical pace. Don’t force the new behavior on the person or animal.

For shaping to work, the successive approximations must mimic the target behavior so that they can serve as steps toward this behavior. Skinner used this procedure to teach rats in a Skinner box (operant chamber) to push a lever and receive reinforcement. This was the final behavior he desired them to make and to get there, he had them placed in the box and reinforced as they moved closer and closer to the lever. Once at the lever the rat was only reinforced when the lever was pushed. Along the way, if the rat went back into parts of the chamber already explored it received no reinforcement. The rat had to move to the next step of the shaping procedure. We use the shaping procedure with humans in cases such as learning how to do math problems or learning a foreign language.

We are almost finished with our coverage of operant conditioning. Before moving onto the final two sections, take a break. Section 6.6 was pretty long and full of a ton of information. Go do something fun, get some food, take a nap, etc. but return once you head is clear, and you are ready to finish up.