Home / RBT Study Guide / Behavior Acquisition / Discrimination training: SD, S-delta and stimulus control
Domain C · Behavior Acquisition · Task C.6
Discrimination training: SD, S-delta and stimulus control
Discrimination training teaches a learner to respond in the presence of one stimulus and not in the presence of others. You reinforce the response when the SD is present and withhold reinforcement when the S-delta is present.
Last reviewed November 10, 2026 against the BACB RBT Test Content Outline (3rd ed.)
What discrimination means in ABA
In everyday talk, discrimination means treating people unfairly. In behavior analysis it means something neutral: telling things apart. A learner discriminates when they respond one way in the presence of one stimulus and a different way, or not at all, in the presence of another.
Almost every skill depends on it. Saying "dog" when you see a dog and not when you see a cat. Stopping at a red light and going at a green one. Putting the fork with the forks. Task C.6 on the RBT outline is to implement discrimination training, the procedure that builds this.
SD, S-delta and stimulus control
Two terms carry the whole topic.
- A discriminative stimulus (SD) is a stimulus in whose presence a particular response has been reinforced. It signals that reinforcement is available for that response.
- An S-delta is a stimulus in whose presence that response has not been reinforced. It signals that reinforcement is not available for that response.
A ringing phone is an SD for answering: pick up and someone is there. A silent phone is an S-delta: pick up and you get nothing.
When a behavior happens reliably in the presence of the SD and rarely in the presence of the S-delta, the behavior is under stimulus control. That is the goal of discrimination training.
Common mix-up
An SD is not the same as a motivating operation. The SD signals that a reinforcer is available. The MO changes how much the learner wants it. See motivating operations for a side-by-side table.
Simple and conditional discriminations
In a simple discrimination, one stimulus sets the occasion for the response. Leo presses a button when its light is on and does not press when the light is off. Light on is the SD. Light off is the S-delta.
In a conditional discrimination, which stimulus is correct depends on another stimulus. A dog picture and a cat picture are on the table. If you say "touch dog", the dog picture is the SD and the cat is the S-delta. If you say "touch cat", the roles switch. The instruction is the condition that decides which choice gets reinforced.
Most receptive identification and matching programs are conditional discriminations. At the RBT level you need to recognize the two types. Choosing targets and designing the sequence is the supervisor's work.
How a discrimination trial is run
Discrimination training is often run in a discrete-trial format, though it also happens in play and routines. The core rule never changes: reinforce the response in the presence of the SD, and do not reinforce it in the presence of the S-delta.
- Set out the materials as the plan describes: the target and one or more other items.
- Get the learner's attention and give the instruction once, in the wording the plan specifies.
- Prompt if the plan calls for it, then fade the prompt as written. See prompting and prompt fading.
- Correct response: reinforce immediately.
- Incorrect response: no reinforcement. Run the error correction in the plan.
- Record the data and rearrange the materials before the next trial.
In session
Diego is learning to pick a shoe from a shoe and a ball. You place both, say "find shoe" and wait. He touches the shoe: you praise and hand him a token at once. On the next trial you swap the positions and say "find shoe" again. He touches the ball: you give no praise or token and follow the plan's error correction (here, repeating the instruction while pointing to the shoe, then trying again without the point).
Small habits in how you present materials can teach the wrong thing. Watch for these:
- Position. If the correct item is always on the same side, the learner may learn "pick the left one". Rotate positions.
- Accidental cues. Looking at the correct item, holding it a little forward or changing your tone can become the real signal.
- Reinforcing errors. A warm "good try" with a smile can work as reinforcement for the wrong choice.
| In the presence of the SD | In the presence of the S-delta | |
|---|---|---|
| What the stimulus signals | Reinforcement is available for the response | Reinforcement is not available for the response |
| What you do when the response occurs | Reinforce it | Do not reinforce it. Follow the plan's error correction |
| What happens over time | The response becomes more likely | The response becomes less likely |
Stimulus generalization: the other direction
Stimulus generalization happens when a response taught with one stimulus also occurs with similar stimuli. Ava learns "dog" with one photo and then says "dog" for the neighbor's dog and a cartoon dog. That spread is usually wanted, and plans build it on purpose, as described in generalization and maintenance.
Discrimination pulls the other way: it narrows responding to the right stimuli. If Ava also calls a cat "dog", generalization has gone too far and she needs more discrimination training between dogs and cats. A well-taught skill has both: wide enough to cover every dog, narrow enough to leave out cats.
How C.6 questions tend to look
Common formats ask you to label the SD or S-delta in a scenario, to choose what the RBT does after a correct or incorrect response, or to tell discrimination from generalization.
On the exam
To find the SD, ask "in the presence of what does this response get reinforced?" To find the S-delta, ask "in the presence of what does the same response get nothing?" If a response is spreading to new, similar stimuli, the answer is generalization, not discrimination.
Practice questions
Original questions in the style of the exam. Pick an answer before you open the explanation.
1. Noor is learning to select a spoon. The RBT places a spoon and a cup on the table and says 'find spoon'. Noor touches the cup. In discrimination training, what happens with this response?
- It is reinforced, but with a smaller amount of praise
- It is punished with a firm verbal reprimand right away
- It is ignored, and the target is dropped for the day
- It is not reinforced, and the plan's correction is run
Show answer
D. It is not reinforced, and the plan's correction is run
A response to the S-delta is not reinforced, and the RBT follows the error correction written in the plan. Giving a smaller reinforcer would still reinforce the wrong selection and weaken the discrimination.
2. An RBT always places the correct picture on the left. Leo scores 100 percent with her, but with another RBT who moves the pictures around he is correct about half the time. What most likely happened?
- The reinforcer lost value through satiation effects
- The skill moved to maintenance before it was ready
- Position, not the picture, controlled his responding
- The second RBT used a schedule that was too thin
Show answer
C. Position, not the picture, controlled his responding
Because the correct item was always on the left, Leo's selecting came under the control of position, not the picture. That is faulty stimulus control, and it shows up as soon as positions are rotated.
3. Priya learned to label her red plastic cup as 'cup'. Now she also says 'cup' for a blue mug and a paper cup that were never taught. This is an example of what?
- Stimulus discrimination
- Stimulus generalization
- Response generalization
- Conditional discrimination
Show answer
B. Stimulus generalization
The same response is occurring in the presence of new, similar stimuli, which is stimulus generalization. Response generalization would be Priya using new, untaught responses, such as saying 'mug' or 'glass'.
Common questions
What is the difference between an SD and an S-delta?
An SD (discriminative stimulus) is a stimulus in whose presence a response has been reinforced, so it signals that reinforcement is available. An S-delta is a stimulus in whose presence that response has not been reinforced. For the instruction touch the dog, the dog picture is the SD and the cat picture is the S-delta.
What is stimulus control in ABA?
A behavior is under stimulus control when it happens more often in the presence of a particular stimulus than in its absence. Answering the phone when it rings, and not when it is silent, is an everyday example.
Is discrimination the opposite of generalization?
They are two ends of the same scale. Stimulus generalization means a response spreads to similar stimuli, and discrimination means the response is limited to specific stimuli. Good teaching needs both: saying dog for every kind of dog, but not for cats.