You've mastered the fundamentals of reinforcement. Now, let's move into more nuanced strategies that involve reinforcing specific behaviors while addressing others. This is where differential reinforcement comes in. It's not about just increasing a behavior; it's about systematically shaping behavior by reinforcing one response class and withholding reinforcement for another.

Let's break down DRA with an example. A child shouts out answers in class to get the teacher's attention. A simple reinforcement plan might not work, but a DRA procedure would. The function of the behavior is attention.

An appropriate alternative is raising a hand. Using DRA, the teacher would ignore the shouting (extinction) but provide immediate, enthusiastic praise when the child raises their hand. The new behavior serves the same function (getting attention) but is a much more adaptive alternative.

Stimulus control is about getting the right behavior to happen at the right time. You already know that a discriminative stimulus ($S^D$) signals that reinforcement is available. But we can build on this by teaching finer discriminations.

Imagine teaching a child to identify a cat. At first, they might call any four-legged furry animal a "cat." This is stimulus generalization. To refine this, you'd use discrimination training. You would reinforce them for saying "cat" when they see a cat ($S^D$) but not when they see a dog (a stimulus delta, or $S^Δ$). This involves presenting both examples and only reinforcing the correct response.

Teaching a concept like "red" requires showing many different red objects (generalization within the class) and many non-red objects (discrimination between classes). This ensures the learner doesn't just associate the word "red" with one specific red ball but understands the property of redness itself.

Achieving behavior change is one thing; making it last and apply to new situations is the real goal. This is the domain of generalization and maintenance.

Response maintenance is the extent to which a learner continues to perform the target behavior after the intervention has been partially or completely terminated. If you have to reinforce a behavior forever, it isn't truly learned.

Setting/situation generalization is the extent to which the learner performs the behavior in environments different from the instructional setting. For example, a child learns to ask for a break at a therapy clinic and then begins asking for breaks appropriately at school.

Let's explore these strategies:

• Teach Sufficient Examples: Don't just teach one way. If you're teaching a student to greet people, have them practice with different people and use different greetings ("hi," "hello," "what's up"). This is called using multiple exemplars.
• Program Common Stimuli: Make the instructional setting as similar to the generalization setting as possible. If you are teaching vocational skills, try to simulate the actual work environment, using the same tools and materials.
• Teach Loosely: Vary non-critical aspects of the instructional setting, like the tone of voice you use, the time of day, or your location. This reduces the chance that the behavior becomes too narrowly controlled by irrelevant stimuli.

Finally, all of these advanced procedures must be tailored to the individual. A functional behavior assessment (FBA) is non-negotiable. It helps you identify the function of a behavior, which is critical for choosing an effective intervention, especially a DRA. What works for one person may not work for another, even if their behaviors look identical on the surface. Effective practice is always individualized.

Now, let's test your understanding of these advanced concepts.

Applying these strategies correctly and ethically is central to the work of a BCBA.