| name | differential-reinforcement |
| description | Use when selecting and implementing differential reinforcement procedures (DRA, DRI, DRO, DRL, DRH) to increase desirable behavior or decrease problem behavior, setting reinforcement parameters, or planning schedule thinning for DR-based interventions. |
Differential Reinforcement
Differential reinforcement (DR) procedures are a family of interventions that systematically reinforce one class of behavior while withholding reinforcement for another. DR procedures are among the most versatile tools in applied behavior analysis — they can be used to increase desirable behavior, decrease problem behavior, or both simultaneously. They are generally categorized as reinforcement-based approaches and are preferred over punishment procedures when clinically effective.
DRA: Differential Reinforcement of Alternative Behavior
Definition
Reinforce a specific desirable alternative behavior while withholding reinforcement (extinction) for the target problem behavior. The alternative behavior should ideally serve the same function as the problem behavior.
Procedure
- Identify the function of the problem behavior through FBA/FA
- Select an alternative behavior that is:
- Already in the learner's repertoire (or teach it)
- Functionally equivalent (accesses the same reinforcer)
- Socially appropriate
- Requires equal or less effort than the problem behavior
- Reinforce the alternative behavior on a dense schedule (initially CRF)
- Withhold reinforcement for the problem behavior (extinction)
- Thin the reinforcement schedule for the alternative behavior over time
Clinical Example
A student raises their hand instead of calling out. Calling out previously produced teacher attention (attention function). Now, hand-raising produces teacher attention; calling out is placed on extinction for attention.
Functional Communication Training as DRA
FCT is the most common form of DRA. The alternative behavior is a communication response that accesses the same reinforcer as the problem behavior. FCT is DRA with a communicative alternative.
Critical Considerations
- If the alternative behavior requires substantially more effort than the problem behavior, the problem behavior will persist (matching law)
- Initial CRF for the alternative is critical — the learner must contact dense reinforcement for the new behavior before the schedule can be thinned
- Extinction for the problem behavior produces an initial extinction burst; prepare for temporary increases
DRI: Differential Reinforcement of Incompatible Behavior
Definition
A specific form of DRA where the reinforced behavior is physically incompatible with the target problem behavior — the two behaviors cannot occur simultaneously.
Procedure
Same as DRA, with the additional constraint that the selected behavior makes the problem behavior physically impossible.
Examples
- Reinforce keeping hands in lap (incompatible with hitting)
- Reinforce sitting in seat (incompatible with wandering)
- Reinforce chewing food with mouth closed (incompatible with talking with mouth full)
- Reinforce flat hands on the table (incompatible with finger flicking)
When to Choose DRI over DRA
DRI is preferred when a physically incompatible alternative exists and is contextually appropriate. DRI provides an additional layer of behavior reduction because the target behavior literally cannot co-occur with the reinforced behavior. However, DRI limits the pool of potential alternative behaviors to only those that are physically incompatible, which may not always include the most functional or desirable replacement.
DRO: Differential Reinforcement of Other Behavior
Definition
Reinforce the absence of the target problem behavior for a specified time interval. Any behavior other than the target behavior is reinforced — the learner receives reinforcement if the problem behavior does not occur during the interval.
Types of DRO
Momentary DRO (mDRO): At the end of each interval, the clinician checks whether the target behavior is occurring at that precise moment. If not occurring at the moment of observation, reinforcement is delivered. Easier to implement but less sensitive to behavior that occurs during the interval.
Whole-interval DRO (wDRO): The target behavior must be absent for the entire interval. If the behavior occurs at any point during the interval, reinforcement is withheld and the interval resets. More rigorous; produces greater behavior reduction.
Fixed vs. Variable Intervals
Fixed DRO: Intervals are the same length throughout (e.g., DRO 3-min). Predictable; the learner may learn to "time" the intervals.
Variable DRO: Intervals vary around an average (e.g., DRO 3-min average, ranging from 1 to 5 min). Less predictable; often produces more stable behavior reduction.
Setting the Initial DRO Interval
- Collect baseline data on the rate or interresponse time (IRT) of the problem behavior
- Set the initial DRO interval at approximately 50–75% of the mean IRT (ensures the learner can succeed early)
- If the learner earns reinforcement on 80%+ of intervals across 2–3 sessions, increase the interval
- Typical thinning: double the interval length (DRO 1-min → 2-min → 4-min → 8-min → 15-min)
Resetting vs. Non-Resetting DRO
Resetting DRO: If the target behavior occurs, the interval timer resets to zero. Reinforcement cannot be earned until a full interval passes without the behavior. More effective for behavior reduction.
Non-resetting (fixed-interval) DRO: The timer continues regardless of behavior occurrence. If the behavior is absent at the end of the interval, reinforcement is delivered. Less commonly used; weaker contingency.
Limitations of DRO
- Does not teach a specific replacement behavior (reinforces "not doing X," not "doing Y")
- The learner may engage in other undesirable behaviors during the interval and still earn reinforcement
- Best combined with DRA to teach what the learner should do, not just what they should not do
- Requires precise timing and monitoring
DRL: Differential Reinforcement of Low Rates
Definition
Reinforce the target behavior when it occurs at or below a specified criterion rate. Used when the behavior is acceptable at low rates but problematic at high rates.
Types of DRL
Full-session DRL: Reinforcement is delivered at the end of the session if the total count of the behavior is at or below the criterion. Example: If the student asks fewer than 5 questions during the class period, they earn the reinforcer.
Interval DRL: The session is divided into intervals. Reinforcement is delivered if no more than a specified number of responses occur per interval.
Spaced-responding DRL (IRT-DRL): Reinforcement is delivered for each response that occurs only after a minimum time has elapsed since the last response. This directly targets the interresponse time. Example: Raising hand is reinforced only if at least 5 minutes have passed since the last hand-raise.
Setting DRL Criteria
- Collect baseline data on the current rate
- Set the initial criterion slightly below the baseline rate (the learner should be able to succeed)
- Gradually lower the criterion across sessions as the learner demonstrates control
- Continue until the behavior reaches the desired terminal rate
Clinical Examples
- Reducing excessive hand-raising (appropriate at a low rate, disruptive at a high rate)
- Reducing repetitive questioning (functional at low rates, stigmatizing at high rates)
- Reducing bathroom requests that are partially escape-maintained
DRH: Differential Reinforcement of High Rates
Definition
Reinforce the target behavior when it occurs at or above a specified criterion rate. Used to increase a behavior that is already in the repertoire but occurs too infrequently.
Procedure
Same logic as DRL but the criterion is above baseline. Reinforce when the behavior meets or exceeds the criterion, then raise the criterion.
Clinical Examples
- Increasing rate of social initiations
- Increasing number of math problems completed per session
- Increasing frequency of independent self-care steps
Selecting the Right DR Procedure
| Clinical Goal | Recommended Procedure |
|---|
| Replace problem behavior with a specific alternative | DRA (or DRI if incompatible alternative exists) |
| Replace problem behavior with functional communication | FCT (a form of DRA) |
| Eliminate a behavior entirely | DRO + DRA (DRO reduces; DRA teaches replacement) |
| Reduce a behavior that is acceptable at lower rates | DRL |
| Increase a behavior that occurs too infrequently | DRH |
| Behavior reduction without extinction (when extinction is not possible) | DRA with enriched reinforcement for alternative (NCR may supplement) |
Combining DR with Extinction
DR procedures are most effective when the problem behavior is simultaneously placed on extinction. Without extinction, the problem behavior continues to contact reinforcement and competes with the target behavior. When extinction is not possible (safety concerns, environmental constraints), DR procedures can still be effective but typically produce slower and less complete behavior change. In these cases, use the richest possible reinforcement schedule for the desired behavior to create a strong differential.
Procedural Fidelity
DR procedures are only as effective as their implementation. Common fidelity failures:
- Inadvertently reinforcing the problem behavior (extinction break)
- Failing to reinforce the alternative behavior promptly
- Using the wrong DRO interval (too long → never earns reinforcement; too short → no behavior change)
- Thinning schedules too rapidly
- Not specifying the alternative behavior clearly enough for all implementers
Train all implementers to criterion and conduct regular fidelity checks with direct observation.
Key References
- Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). Applied Behavior Analysis (3rd ed.). Pearson.
- Vollmer, T. R., & Iwata, B. A. (1992). Differential reinforcement as treatment for behavior disorders: Procedural and functional variations. Research in Developmental Disabilities, 13(4), 393–417.
- Petscher, E. S., Rey, C., & Bailey, J. S. (2009). A review of empirical support for differential reinforcement of alternative behavior. Research in Developmental Disabilities, 30(3), 409–425.
- Repp, A. C., Barton, L. E., & Brulle, A. R. (1983). A comparison of two procedures for programming the differential reinforcement of other behaviors. Journal of Applied Behavior Analysis, 16(4), 435–445.
- Athens, E. S., & Vollmer, T. R. (2010). An investigation of differential reinforcement of alternative behavior without extinction. Journal of Applied Behavior Analysis, 43(4), 569–589.
