4.5 Discrimination Training and Stimulus Control

Teaching the Relevant Cue

Discrimination training is used when a client needs to respond one way in the presence of one stimulus and a different way, or not at all, in the presence of another. In RBT work, this may include selecting the correct picture when asked, following sit down but not stand up, identifying red versus blue, sorting clean and dirty clothes, answering yes and no questions, or using a restroom sign to choose the correct door. The procedure is important because many skills are not just responses; they are responses under the right conditions.

A discriminative stimulus, often abbreviated SD, signals that a particular response will contact reinforcement. An S-delta signals that the same response will not contact reinforcement in that context. RBTs do not need to use these terms with clients, but they need to understand the implementation. If the instruction is Touch spoon and the array includes spoon, cup, and ball, touching spoon should be reinforced. Touching cup should not produce the same reinforcement for that trial. If both responses produce the same outcome, the client has little reason to attend to the difference.

Stimulus control is demonstrated when the response reliably occurs in the presence of the relevant cue and not in other conditions. A learner who says dog when shown any animal picture does not yet have precise stimulus control for dog. A learner who touches the left card regardless of instruction may be under position control. A learner who responds only when one RBT teaches may be responding to person-specific cues. These are not failures by the client. They are data patterns that guide programming, and the RBT reports them objectively.

Scenario: A client is learning to identify community signs. The RBT presents stop, restroom, and exit cards. The plan says to rotate positions randomly and reinforce independent correct responses. During the first five trials, the stop card happens to appear on the left three times. The client begins choosing left before the instruction is finished. The RBT should rotate positions as planned, avoid predictable placement, record the errors or prompted responses, and report the possible position bias.

The RBT should not conclude that the client is being careless or decide to remove all left-side choices without direction.

Scenario: A learner is working on yes and no discrimination. The RBT asks, Is this a shoe? while holding up a shoe, and the client says yes. Then the RBT asks, Is this a shoe? while holding up a spoon, and the client also says yes. The RBT follows the correction procedure and records the response pattern. The teaching issue may involve overselectivity, a response bias toward yes, unclear prompts, or insufficient practice with no trials. The RBT does not redesign the program but can provide the supervisor with specific data: The client answered yes on 9 of 10 trials, including all 5 no trials.

Discrimination training depends heavily on data quality. Overall percentage may hide important patterns. A client might score 80 percent because they identify cup and ball perfectly but miss spoon every time. A client might perform well with photos but not real objects. A client might respond accurately when comparison items are very different but not when they are similar. RBT notes should be objective enough to help the supervisor see those patterns.

A stimulus control troubleshooting map for RBT reporting:

• Position pattern: Client selects left, middle, right, top, or nearest item regardless of instruction.
• Response bias: Client says yes, no, same word, or same sign across many trials.
• Prompt dependence: Client waits for a gesture, model, or facial expression before responding.
• Stimulus overselectivity: Client responds to one feature, such as color, while missing shape or function.
• Person-specific control: Client responds with one staff member but not another trained person.
• Setting-specific control: Client responds in the therapy room but not classroom, home, or community routine.
• Material-specific control: Client responds to one card set but not real objects or different examples.

RBTs also support discrimination training by using clear pacing. If trials are too fast, the client may respond to routine instead of the SD. If trials are too slow or wordy, motivation may drop. If instructions are repeated several times, the repeated instruction may become part of the cue. Many plans specify one presentation, a response interval, a prompt if needed, and an error correction sequence. The RBT follows that sequence even when the answer seems obvious.

Discrimination training should remain respectful. The RBT should avoid saying, No, you know this, or making the client repeat errors for embarrassment. Correction procedures are teaching procedures, not scolding. A neutral tone, brief representation, and immediate opportunity for reinforcement after a correct or prompted response can keep learning clear. If the client shows frustration or refusal, the RBT follows the behavior support plan and reports objective information.

For the RBT exam and daily practice, the key distinction is between knowing a response and knowing when to use it. A client may be able to say bathroom but needs to discriminate when to request a bathroom break, which door sign to follow, and which adult can help in a school hallway. The RBT implements the programmed discrimination steps, minimizes accidental cues, reinforces the response under the correct stimulus conditions, and gives the supervisor data that are detailed enough to guide next steps.