Differential reinforcement
Reinforcing one behavior or response class while withholding reinforcement for others, usually problem behavior, so the reinforced behavior increases and the others decrease.
Differential reinforcement has two parts: reinforcement for a specified behavior (or for the absence of problem behavior) and extinction for problem behavior, meaning problem behavior no longer produces the reinforcer that used to follow it. Because the reinforcer is available for the right behavior and not for the wrong one, behavior shifts toward what is reinforced. It is one of the most common reduction procedures because it decreases problem behavior while building or strengthening appropriate behavior.
The main forms an RBT will see are DRA (reinforcing an alternative behavior), DRI (reinforcing a behavior that is physically incompatible with problem behavior), DRO (reinforcing the absence of problem behavior during an interval), and DRL (reinforcing a behavior when it occurs at or below a set low rate). Functional communication training is a widely used type of DRA. Shaping also relies on differential reinforcement, applied to successive approximations.
The procedure is most effective when the reinforcer used is the one that maintains the problem behavior, when the reinforcer is delivered immediately and consistently, and when problem behavior reliably does not produce it. If problem behavior is still reinforced some of the time, it can persist even while the new behavior is also reinforced.
The supervisor chooses the type of differential reinforcement, the reinforcer, the schedule and how it will be thinned. The RBT implements those details exactly and records both the target behavior and the problem behavior.
Example and non-example
| Example | A child's grabbing of toys is maintained by access to toys. The plan says to give the toy when the child says "my turn, please" and not when the child grabs. The RBT follows this every time. |
|---|---|
| Non-example | An RBT praises a client for sharing but also returns toys immediately when the client screams. Problem behavior is still reinforced, so the procedure is not differential reinforcement as written. |
Don't confuse it with
| Extinction | Extinction only withholds reinforcement for problem behavior, while differential reinforcement also provides reinforcement for another behavior or for its absence. |
|---|---|
| Noncontingent reinforcement (NCR) | NCR delivers reinforcement on a time schedule regardless of behavior, while differential reinforcement depends on which behavior occurs. |
Test yourself
-
Question 1A plan says independent correct responses earn a preferred toy and praise, and prompted correct responses earn praise only. The client answers correctly after a gestural prompt. What should the RBT do?
Show answer
D: Give praise only and record the trial as prompted. This plan uses differential reinforcement: stronger reinforcement for independent responses than for prompted ones, which helps move the client toward independence. The RBT follows it exactly and records the prompt.
- A: The response needed a prompt, so it is not independent and does not earn the toy under this plan.
- B: The response was correct with a prompt; the plan calls for praise and a prompted score.
- C: The data are right, but the plan reserves the toy for independent responses.
Task C.3 · 40-hour curriculum section E
-
Question 2Many plans give a stronger reinforcer for independent correct responses than for prompted correct responses. Why?
Show answer
C: It makes responding without help more worthwhile, which supports prompt fading. When independent responses earn more reinforcement than prompted ones, the learner has a reason to respond before the prompt. This differential reinforcement helps shift control from the prompt to the instruction.
- A: Differential reinforcement of independent responses does not prevent errors; prompts do that.
- B: The number of trials is set by the plan, not by how reinforcement is divided.
- D: Prompts still need to be faded; differential reinforcement supports the fading.
Task C.7 · 40-hour curriculum section E
-
Question 3Which two processes make up the differential reinforcement used in shaping?
Show answer
A: Reinforcing the current approximation; extinction of earlier ones. In shaping, responses that meet the current criterion are reinforced, while earlier approximations that used to be reinforced are now placed on extinction. This shifts responding toward the terminal behavior.
- B: Shaping does not use punishment for earlier approximations; they simply stop being reinforced.
- C: Prompting and prompt fading are a different procedure.
- D: Noncontingent reinforcement is not contingent on any response, so it cannot select closer approximations.
Task C.10 · 40-hour curriculum section E
Answer every question to see your score.
On the RBT exam
This term belongs to task D.3 Differential reinforcement in content area D, Behavior Reduction, which is 14 of the 75 scored questions.