10.1 Differential Reinforcement Procedures (DRA, DRI, DRO, DRL, DRH)

Foundational Mechanics and Conceptual Basis of Differential Reinforcement

In applied behavior analysis, differential reinforcement is a foundational, evidence-based behavior-change strategy defined by the simultaneous application of two distinct operant principles: reinforcement for a designated, desirable response class and operant extinction (withholding the maintaining reinforcer) for a designated problem behavior or non-target response class (Catania, 1998; Cooper, Heron, & Heward, 2020; Skinner, 1938).

The dual operant mechanism functions as a dynamic system:

1. Reinforcement: Increases or maintains the future frequency, rate, duration, or fluency of the adaptive, functional, or acceptable response class.
2. Extinction: Systematically withholds the functional reinforcer maintaining the challenging behavior, thereby decreasing its future probability and weakening the motivating relation.

When differential reinforcement is implemented across varied environmental conditions correlated with specific antecedent stimuli, it establishes stimulus discrimination (differential responding in the presence versus absence of a discriminative stimulus, $S^D$). When differential reinforcement is applied to specific variations within or across response forms, it produces response differentiation (the emergence of a distinct, refined response class while unreinforced response variations extinguish). Together, these mechanisms allow behavior analysts to eliminate severe problem behaviors while concurrently building socially valid, habilitative repertoires without resorting to primary aversive contingencies.

---

Differential Reinforcement of Alternative Behavior (DRA)

Core Mechanics and Function Matching

Differential Reinforcement of Alternative Behavior (DRA) involves reinforcing a desirable, socially valid alternative behavior that serves the identical function as the problem behavior, while placing the problem behavior on extinction. DRA is widely recognized as the gold-standard intervention for socially maintained problem behaviors because it directly adheres to the Fair Pair Rule (White & Haring, 1980): whenever a challenging behavior is reduced or eliminated, an adaptive replacement behavior must be explicitly taught and strengthened to replace it.

A fundamental clinical mandate of DRA is functional equivalence. The alternative response must access the exact functional reinforcer maintaining the challenging behavior, as identified through a Functional Behavior Assessment (FBA) or experimental functional analysis (FA). For example:

• If a student engages in physical aggression maintained by escape from difficult academic tasks (social negative reinforcement), the alternative behavior must produce escape or task reduction (e.g., requesting a brief break, asking for assistance, or negotiating task modifications). Reinforcing quiet sitting with edible candies (social positive reinforcement) fails to address the underlying establishing operation (EO) of task aversiveness and will routinely fail.
• If a child engages in screaming maintained by caregiver attention (social positive reinforcement), the alternative behavior must produce adult attention (e.g., tapping the caregiver's shoulder, using a polite vocal greeting, or holding up an attention card), while screams are placed on attention extinction (planned ignoring).

Functional Communication Training (FCT)

Pioneered by Carr and Durand (1985), Functional Communication Training (FCT) is an application of DRA where the alternative response is an explicit communicative mand (e.g., vocal word, manual sign, picture exchange icon, or speech-generating device button). FCT restructures the client's communicative environment so that the functional reinforcer is delivered immediately and densely on a continuous schedule (Fixed Ratio 1, or FR1) contingent upon the communicative mand, while the problem behavior produces zero reinforcement.

The Matching Law in DRA Design

Herrnstein's (1961) Matching Law dictates that when two or more concurrent response alternatives are available, the relative rate of responding matches the relative rate, immediacy, quality, and effort of reinforcement associated with each alternative:

$\frac{B_1}{B_1 + B_2} = \frac{R_1}{R_1 + R_2}$

To ensure that the alternative replacement behavior outcompetes the problem behavior, the behavior analyst must deliberately bias reinforcement parameters in favor of the alternative response:

1. Schedule / Rate of Reinforcement: The alternative behavior must contact reinforcement on a denser schedule (initially FR1) than the problem behavior.
2. Immediacy of Reinforcement: The alternative behavior must produce instant reinforcement (zero latency), whereas reinforcement for problem behavior must be eliminated or substantially delayed.
3. Quality of Reinforcement: The alternative behavior must access higher-potency, preferred variations of the maintaining reinforcer.
4. Response Effort: The physical and cognitive effort required to emit the alternative behavior must be lower than or equal to the effort required to emit the problem behavior. If vocalizing "May I please have a three-minute break from this worksheet?" requires high cognitive effort while slamming a textbook onto the floor is effortless, the client will allocate responding to slamming under the matching law. The alternative response must be engineered to be effortless (e.g., touching a single break card).

Differential Reinforcement With and Without Extinction

Task G.18 asks you to implement differential reinforcement with and without extinction. With extinction, problem behavior no longer produces its reinforcer, which is the most reliable arrangement. Sometimes extinction is impossible (peers keep laughing at call-outs) or unsafe (severe self-injury cannot simply be ignored). Differential reinforcement can still work if the reinforcement parameters strongly favor the alternative behavior: the alternative produces a longer, higher-quality, more immediate, or more frequent reinforcer than the problem behavior, which still contacts some reinforcement. Athens and Vollmer (2010) reduced problem behavior this way without extinction by manipulating the duration, quality, and delay of reinforcement. The trade-off is that the effect depends on keeping that advantage; if reinforcement for the alternative weakens, problem behavior can return quickly.

---

Differential Reinforcement of Incompatible Behavior (DRI)

Anatomical and Topographical Incompatibility

Differential Reinforcement of Incompatible Behavior (DRI) is a specialized subtype of DRA in which the alternative response class is topographically and anatomically incompatible with the target problem behavior. The learner physically cannot emit both behaviors at the exact same instant in time (Cooper et al., 2020).

The defining feature of DRI is physical impossibility. Incompatibility must be rooted in biomechanics rather than subjective interpretation:

• Problem Behavior: Hand-to-head self-injurious hitting $\longleftrightarrow$ Incompatible Behavior: Keeping both hands clasped firmly in lap, sitting on hands, or holding a weighted ball with both hands.
• Problem Behavior: Out-of-seat running in a classroom $\longleftrightarrow$ Incompatible Behavior: Sitting with buttocks flat on seat cushion and feet touching the floor.
• Problem Behavior: Mouthing non-food objects / pica $\longleftrightarrow$ Incompatible Behavior: Chewing sugar-free gum or biting on an oral motor chewy tube.
• Problem Behavior: Scratching, pinching, or gouging staff members $\longleftrightarrow$ Incompatible Behavior: Keeping hands inside zippered jacket pockets or interlocking fingers.

Clinical Strengths and Limitations of DRI

• Clinical Strengths: DRI provides an immediate safety buffer. Because the incompatible response physically prevents the emission of the problem behavior, it reduces the need for intrusive physical blocking or mechanical restraints. It provides clear, unambiguous observation criteria for direct-care staff and Registered Behavior Technicians (RBTs).
• Limitations: In daily living, identifying an incompatible behavior that is both anatomically exclusive and functionally equivalent to the problem behavior is often challenging. For example, sitting with hands in pockets does not provide functional escape from a difficult reading task; thus, DRI must often be combined with functional communication to address the maintaining establishing operation.

---

Differential Reinforcement of Other Behavior (DRO - Zero-Rate Responding)

Core Contingency and Definitions

Differential Reinforcement of Other Behavior (DRO) delivers a specified reinforcer contingent upon the complete absence (zero occurrences) of the target problem behavior throughout a specified temporal interval or at a specific momentary point in time. DRO is technically a misnomer; it does not reinforce a specific "other" behavior, but rather reinforces the non-occurrence (omission) of the target response class. For this reason, it is frequently termed differential reinforcement of zero responding or omission training (Reynolds, 1961).

Primary DRO Subtypes

Behavior analysts employ four primary variations of DRO, structured along two methodological axes: interval versus momentary and fixed versus variable.

1. Fixed-Interval DRO (FI-DRO): A fixed temporal duration (e.g., 5 minutes) is established. If the learner does not engage in the target problem behavior at any point during the entire 5-minute interval, reinforcement is delivered immediately at the conclusion of the interval. If the problem behavior occurs at any point during the interval, reinforcement is withheld.
2. Variable-Interval DRO (VI-DRO): Reinforcement is delivered following the absence of the target behavior across intervals of varying durations that revolve around a predetermined mathematical mean (e.g., VI 4 minutes, consisting of randomized intervals such as 2, 4, 6, 3, and 5 minutes). Varying the interval makes the end of each interval less predictable for the learner.
3. Fixed-Momentary DRO (FM-DRO): The observer schedules a fixed temporal interval (e.g., every 5 minutes). Reinforcement is delivered contingent on the learner not engaging in the target behavior at the exact moment the interval expires (e.g., at minute 5:00). If the behavior occurs at minute 2:30 or 4:15, but is completely absent at minute 5:00, the learner earns reinforcement.
4. Variable-Momentary DRO (VM-DRO): The observer checks for the presence or absence of the target behavior at variable, unpredictable moments throughout the day. If the learner is not emitting the behavior at that specific momentary observation point, reinforcement is delivered.

Interval Resetting vs. Non-Resetting Mechanics

When implementing interval DRO (FI-DRO or VI-DRO), clinicians must specify how the interval operates when a problem behavior occurs:

• Resetting Interval: The instant the target problem behavior occurs, the timer is immediately stopped and reset to zero, beginning the interval anew. This establishes a strict contingency where every occurrence of problem behavior directly postpones reinforcement by the full interval duration.
• Non-Resetting Interval: If the target behavior occurs, the timer continues running until the interval ends. Reinforcement is withheld at the interval's conclusion, and the subsequent interval begins automatically at the scheduled time. While resetting intervals generally produce faster behavioral deceleration, non-resetting intervals are significantly easier for classroom teachers and vocational supervisors managing group schedules.

Setting and Thinning Initial DRO Intervals

A critical, frequently tested rule of thumb governs initial DRO interval calibration: The initial DRO interval must be equal to or slightly less than the learner's baseline mean interresponse time (IRT):

$\text{Initial DRO Interval} \le \text{Baseline Mean IRT}$

$\text{Baseline Mean IRT} = \frac{\text{Total Duration of Observation Period}}{\text{Total Number of Target Responses Emitted}}$

Clinicians typically set the initial interval between $50\%$ and $80\%$ of the baseline mean IRT. For example, if a client emits 20 instances of severe self-injurious skin-picking during a 60-minute baseline observation, the mean IRT is 3 minutes (180 seconds). The initial DRO interval should be set between 90 seconds and 140 seconds.

The Trap of Setting Intervals Too Long: If the clinician sets the initial DRO interval at 10 minutes when the baseline mean IRT is 3 minutes, the client will almost certainly emit the behavior before the interval expires. The client will experience prolonged extinction without ever contacting reinforcement, precipitating severe extinction bursts, emotional aggression, and treatment failure.

Systematic Schedule Thinning: Once the learner consistently earns reinforcement across consecutive intervals (e.g., achieving $\ge 80\%$ of available intervals over 3 consecutive sessions), the interval is systematically thinned. Thinning can occur by increasing the interval duration by fixed increments (e.g., adding 30 seconds or lengthening by $10\%$ to $20\%$) or by transitioning from an interval DRO to a momentary DRO.

The "Behavioral Vacuum" and Critical Limitations of DRO

While DRO rapidly suppresses problem behavior, it possesses two major, inherent clinical vulnerabilities:

1. Zero Alternative Skill Acquisition: DRO reinforces the absence of a behavior, meaning it fails to teach, strengthen, or promote an adaptive replacement repertoire. It violates the Fair Pair Rule if implemented in isolation.
2. The Behavioral Vacuum and Accidental Reinforcement: When a high-rate problem behavior is suppressed, a "behavioral vacuum" is created. Because the DRO contingency reinforces whatever behavior is occurring when the timer expires, it can accidentally reinforce novel, secondary, or dangerous problem behaviors. If a student refrains from head-banging for 4 minutes and 59 seconds, but kicks a teacher or screams profanities at minute 5:00, delivering reinforcement at the 5-minute mark directly reinforces kicking or screaming. Therefore, DRO should rarely be used as a standalone intervention; it must be coupled with DRA, DRI, or skill-acquisition programming.

---

Differential Reinforcement of Low Rates of Responding (DRL)

Clinical Rationale and Scope

Differential Reinforcement of Low Rates of Responding (DRL) delivers reinforcement contingent upon the target behavior occurring at a rate lower than a predetermined, specified criterion. DRL is uniquely indicated when the target behavior is acceptable, appropriate, or socially valid in moderation, but problematic, disruptive, or counterproductive when emitted with excessive frequency.

Common clinical applications include:

• Asking a teacher for help or clarification.
• Raising one's hand during classroom instruction.
• Requesting permission to use the restroom or drink water.
• Eating rapidly (shoveling food during meals).
• Washing hands or checking door locks.

Three Subtypes of DRL

1. Full-Session DRL: Reinforcement is delivered at the end of an entire instructional session or observation period if the total number of responses emitted during that entire session is equal to or less than a specified criterion (e.g., student earns a preferred activity at the end of a 60-minute period if they asked to leave their seat 3 or fewer times).
2. Interval DRL: The total session is divided into equal, smaller temporal intervals (e.g., four 15-minute segments in an hour). Reinforcement is delivered at the conclusion of each interval if responses during that specific interval do not exceed the criterion (e.g., no more than 1 question per 15 minutes).
3. Spaced-Responding DRL: Reinforcement is delivered immediately following an occurrence of the target behavior, provided that a specified minimum interresponse time (IRT) has elapsed since the preceding response ($IRT \ge \text{Criterion}$). If the response occurs prior to the criterion IRT, it is placed on extinction, and the IRT timer resets to zero. As with DRO, set the initial criterion at or slightly below the baseline mean IRT and lengthen it gradually.

Absolute Contraindication of DRL

DRL must NEVER be used for severe problem behaviors, including self-injurious behavior (SIB), physical aggression, property destruction, or elopement. In DRL, responses that meet the schedule criterion are explicitly reinforced; reinforcing even a "low rate" of head banging, biting, or violent assault is a dangerous and unethical clinical practice. Zero-tolerance behaviors mandate DRA, DRI, or DRO.

---

Differential Reinforcement of High Rates of Responding (DRH)

Differential Reinforcement of High Rates of Responding (DRH) delivers reinforcement contingent upon the target behavior occurring at or above a predetermined rate or frequency criterion, or with interresponse times shorter than a specified duration ($IRT \le \text{Criterion}$). DRH is designed to build behavioral fluency, increase the speed of academic or vocational task execution, and promote rapid, active student responding (ASR).

Variations parallel DRL:

• Full-Session DRH: Reinforcement delivered if total responses across the session meet or exceed criterion (e.g., completing at least 40 math facts during a 2-minute timing).
• Interval DRH: Reinforcement delivered if responses in a discrete block meet or exceed criterion (e.g., reading at least 25 sight words in a 30-second interval).
• Spaced-Responding DRH: Reinforcement delivered immediately following a response if the IRT since the preceding response is shorter than a specified maximum duration (e.g., assembling two mechanical components within 4 seconds of each other).

---

Differential Reinforcement Procedures Matrix

The following matrix contrasts the operational contingencies, target behavioral effects, skill acquisition properties, clinical indications, and primary risks across the differential reinforcement family:

---

Clinical Troubleshooting and Common Exam Pitfalls

• Pitfall 1: Confusing DRO with Extinction: Operant extinction involves withholding the maintaining reinforcer whenever the target behavior occurs. DRO involves delivering a reinforcer contingent on the non-occurrence of the target behavior. Extinction is a single operant process; DRO is a schedule of reinforcement.
• Pitfall 2: Setting DRO Intervals Based on Clinical Hope Rather Than Baseline IRT: Setting a 10-minute initial DRO interval for a client who engages in aggression every 2 minutes guarantees treatment failure. The initial interval must be calculated using baseline mean IRT (Initial Interval $\le$ Baseline Mean IRT).
• Pitfall 3: Failing to Combine Extinction with DRA: If a teacher provides a 3-minute break when a student says "break, please", but also allows the student to escape the task when they throw a textbook, extinction is absent. Under the Matching Law, both behaviors remain in concurrent competition, and the problem behavior will persist if it requires less effort.
• Pitfall 4: Utilizing DRL for Self-Injurious Behavior or Aggression: Board exams routinely test this trap. DRL explicitly reinforces low rates of behavior. Never select DRL for behaviors that pose physical danger.