Behavior Reduction
Domain D is 14 of the 75 scored questions (19%) across seven tasks, D.1 through D.7, and it is where the exam does its hardest work: reducing behavior means responding correctly in the moment a client is escalating, when three of the four options in front of you all look like something a caring person would do. Two of its tasks are genuinely new in the 3rd-edition outline — D.5, implementing positive and negative punishment procedures, and D.6, describing the secondary effects of extinction and punishment — and neither appears in prep material built for the 2nd edition. The other thing this domain tests relentlessly is the boundary between implementing the written plan and escalating to your supervisor, which candidates routinely get backwards.
What this domain asks you to do (D.1–D.7)
Seven tasks, five of them "implement" tasks, and two that ask you to identify or describe. Read D.5 and D.6 twice: they are the additions that make this outline different from the one most study material was written for. D.5 puts punishment procedures explicitly in the behavior technician's implementation repertoire — under a written plan, with training, and with all the safeguards described later in this chapter — where the 2nd-edition task list did not mention punishment at all. D.6 asks you to describe what else happens when extinction and punishment are used: extinction bursts, response variation, resurgence, emotional responding, and the escape and avoidance that punishment can generate. Those effects are named in the outline, so they are fair game as vocabulary and as scenarios.
The four common functions of behavior (D.1)
Every behavior plan you will ever run is built on an answer to one question: what does this behavior get for the client? There are four common answers, and the exam tests them from scenarios rather than from definitions. Social attention: the behavior reliably produces someone's reaction — a reprimand and a hug are both attention, which is why "but I told him to stop" is not evidence that attention was withheld. Escape or avoidance: the behavior reliably ends, delays or reduces a demand or an unwanted situation; this is socially mediated negative reinforcement, and it is the function that punishes a well-meaning technician most often, because giving a break to calm someone down is exactly the consequence that maintains it. Access to tangibles or activities: the behavior reliably produces an item or an activity, including the return of one that was just taken away. Automatic or sensory: the behavior produces its own consequence without anyone else's involvement, which is why it continues when the client is alone and why it does not respond to changes in your attention. Read the scenario for the consequence that reliably follows, not for the emotion described. And note the scope line: knowing the four categories is your task, but determining the function of a particular client's behavior is your supervisor's determination from assessment data.
Antecedent interventions (D.2)
Antecedent interventions change the environment before the behavior happens, so the behavior becomes less likely in the first place — a far better outcome than managing it after the fact. Noncontingent reinforcement (NCR) delivers the maintaining reinforcer on a time-based schedule regardless of behavior: attention every two minutes on a timer, or a break every five minutes whether or not problem behavior occurs. The word to hold onto is noncontingent — nothing the client does earns or blocks the delivery, which reduces the motivation to behave problematically to get it. The high-probability request sequence (sometimes called behavioral momentum) presents two or three quick requests the client almost always follows — clap your hands, give me five, touch your nose — immediately before the difficult request, so the client is already responding and reinforcement is already flowing when the hard one arrives. Demand fading reduces the difficulty or quantity of demands and then builds them back up gradually, and is used when demands themselves are the trigger. Other antecedent strategies show up in plans too: offering choices, priming and previewing what is coming, visual schedules, and changing the environment to remove a trigger. All of these are strategies you implement as written; the exam's antecedent items usually turn on recognizing which one is being described, and on the fact that antecedent interventions are preventive rather than consequence-based.
Differential reinforcement: DRA, DRI, DRO, DRL and FCT (D.3)
Differential reinforcement means reinforcing one thing while withholding reinforcement for another, and the whole family is distinguished by what exactly gets reinforced. DRA — differential reinforcement of alternative behavior — reinforces a different, appropriate behavior that serves the same function, while the problem behavior no longer produces the reinforcer. A client who used to scream for a break now gets the break for handing over a break card. DRI — differential reinforcement of incompatible behavior — is a DRA in which the alternative is physically impossible to do at the same time as the problem behavior: hands in pockets is incompatible with hitting; sitting is incompatible with running. Ask yourself whether the two behaviors could physically co-occur; if they cannot, it is DRI, and if they could, it is DRA. DRO — differential reinforcement of other behavior — reinforces the absence of the problem behavior for a specified period. Nothing in particular has to be done; the client simply has to not do the target thing for the interval, and then reinforcement is delivered. It is sometimes called the omission procedure, and the giveaway phrase in a scenario is a time interval with no target behavior. DRL — differential reinforcement of low rates — reinforces a reduced rate of a behavior that is acceptable at lower frequency: asking to use the bathroom four times a session instead of twenty, raising a hand three times a lesson instead of thirty. Use it when the behavior is not dangerous and does not need to reach zero. FCT — functional communication training — is a specific and very common form of DRA in which the alternative taught is a communication response that matches the function: a sign, a card, a word or a device output that requests exactly what the behavior used to produce. The single sentence that separates the family: DRA and DRI and FCT reinforce a specified behavior, DRO reinforces the absence of one, and DRL reinforces less of one.
Extinction (D.4)
Extinction means the behavior no longer produces the reinforcer that has been maintaining it, so over time it weakens. The critical word is "the" — extinction is function-specific, and applying the wrong version does nothing or makes matters worse. For attention-maintained behavior, extinction means the attention is withheld, which is what planned ignoring is. For escape-maintained behavior, extinction means the demand is not removed and the client is guided through it as the plan specifies, and here planned ignoring would actually deliver the escape and strengthen the behavior. For tangible-maintained behavior, extinction means the item is not delivered. For automatically maintained behavior, extinction is harder because the reinforcer is produced by the behavior itself, and the procedures used are specified by the supervisor. Extinction is almost never used alone: it is paired with reinforcement of an alternative, because a plan that only removes what worked leaves the client with no way to get their need met. That is why DRA-plus-extinction is the most common combination you will run. And extinction has to be consistent — if the behavior sometimes still works, you have effectively moved it onto an intermittent schedule, which makes it more persistent rather than less. That fact, more than any other in this domain, is why the exam keeps asking what you do when a parent, an aide or a colleague intervenes mid-procedure.
Secondary effects of extinction (D.6, new in the 3rd edition)
When reinforcement stops, behavior does not simply fade quietly, and the 3rd-edition outline now expects you to name what happens instead. The extinction burst is the immediate, temporary increase in the frequency, intensity or duration of the behavior right after extinction begins — the client screams louder and longer before screaming stops working. Response variation is the client trying different forms of the behavior to make the old consequence reappear: if screaming stopped working, they may throw, or grab, or find some other way. Some of these new forms are more severe than the original, which is why plans anticipate them and why an unexpected escalation is reported. Emotional responding — crying, aggression, distress — commonly accompanies extinction, particularly early on. Resurgence is the reappearance of a previously reinforced behavior that had stopped, when a currently reinforced behavior stops working; a client whose new break card stops producing breaks may return to the screaming that used to. Spontaneous recovery is the reappearance of the extinguished behavior after time has passed, even without any change in procedure. All of these are expected, documented and reported — not evidence that the procedure failed. The practical rule the exam is testing: a burst is the point at which giving in does the most damage, because the reinforcer then arrives after a bigger, louder version of the behavior. If the burst is beyond what the plan anticipated, or safety is at risk, that is a different situation and you follow the crisis procedure and contact your supervisor.
Punishment procedures (D.5, new in the 3rd edition)
Punishment, in behavior-analytic terms, is a consequence that makes the behavior less likely in the future — it is defined by effect, exactly like reinforcement, and not by how unpleasant it looks. Positive punishment adds a stimulus after the behavior: a verbal reprimand, an overcorrection procedure, effortful restitution. Negative punishment removes a stimulus: response cost takes away something already earned, such as tokens, and time-out from positive reinforcement removes access to reinforcement for a period. Time-out comes in nonexclusionary forms, where the client stays in the setting but loses access to reinforcement — the time-out ribbon, contingent observation, planned ignoring within an activity — and exclusionary forms, where the client is removed from the reinforcing setting. Two facts decide most time-out items. First, time-out only works if the "time-in" environment is reinforcing; taking a client away from a task they hated is not a time-out, it is escape, and it will strengthen the behavior. Second, time-out ends according to the plan's criteria, not when the client happens to calm down, and releasing a client from time-out during problem behavior teaches that behaving that way ends it. Now the part that matters more than the taxonomy: your position. Behavior-analytic practice puts reinforcement-based and antecedent procedures first, and reserves punishment procedures for situations where a supervisor has determined they are warranted, has written them into the plan, has obtained the necessary consents and approvals, and has trained you on them. You implement a punishment procedure only when all of that is true. You never add a consequence because a session is going badly, never improvise a time-out, and never use a procedure you have not been trained on. If you believe a written punishment procedure is inappropriate or is harming the client, that is a concern you raise with your supervisor promptly — and if you are asked to do something you believe is unsafe or unethical, you do not do it and you report it.
Secondary effects of punishment (D.6, new in the 3rd edition)
The outline pairs punishment's side effects with extinction's, and they are a big part of why the field prefers reinforcement. Emotional responses are common — crying, agitation, aggression directed at whoever delivered the consequence. Escape and avoidance develop not only from the situation but from the person: a client who has been punished by you may start avoiding you, leaving the room, or refusing to work with you, which damages the working relationship you need in order to teach anything. Punishment can also suppress behavior without teaching anything, so the moment the procedure stops the behavior returns, and it teaches nothing about what to do instead — which is why a punishment component always sits alongside reinforcement for an appropriate alternative. Other described effects include the client imitating the punishment they experience, and the technician's own behavior being reinforced by the immediate quiet that punishment produces, which is precisely how punishment procedures creep into sessions where they do not belong. Being able to name these effects is a scored task, and noticing them in a real client is a reporting obligation.
Crisis and emergency procedures (D.7)
A crisis is a situation involving risk of harm to the client or someone else, and it is one of the few moments in this domain where escalation is unambiguously correct. Your obligations are simple to state and must be automatic. Follow the client's written crisis or emergency plan and your employer's emergency procedures, exactly as trained. Prioritize safety: block or move to protect the client and others, remove hazards, clear the area of other people if that is what your training and setting call for. Use only procedures you have been trained and authorized to use — physical restraint and any similar restrictive procedure is used only where it is permitted by the plan, your employer's policy and applicable law, only by staff trained in the specific protocol, and only for as long as the emergency requires. Call for the help your setting requires, whether that is a colleague, your supervisor, or emergency services. Notify your supervisor as soon as it is safe to do so, and document what happened objectively and promptly — what preceded it, what was done, how long it lasted, and the client's condition afterwards. Two habits protect you and the client: never improvise a physical procedure you have not been trained on, and never treat a crisis as something to sort out quietly and mention later.
When you implement the plan and when you escalate
Candidates walk into this exam believing that "contact the BCBA" is always the safest option, and that belief costs points, because the whole point of the behavior-technician role is implementing written procedures faithfully in real time. If the situation is covered by the written plan and you have been trained on the procedure, you implement it — with fidelity, consistently, even when someone in the room is urging you to do otherwise, and even when it is uncomfortable. Consistency is not a preference, it is the mechanism: a procedure that works only sometimes is worse than one that never ran. Concerns can and should be raised afterwards, in your session note and at your next supervision contact. So when do you escalate? When there is a risk to safety, immediately and during the session. When the situation is not covered by the plan at all. When the plan cannot be implemented as written in the current circumstances. When you are asked to do something outside your scope, your training, or the plan. When the data show the client is getting worse, or a mastered skill has been lost. When something has changed for the client — illness, medication, a major event at home — that could affect the program. When something makes you think the plan may be causing harm. And when you have made a mistake or the procedure was not run correctly, which is reported rather than hidden. Notice that most of that list is about the plan's coverage, not about your comfort. "I felt bad running the procedure" is a reason to talk to your supervisor after the session; it is not a reason to stop mid-session.
Key facts and definitions for Domain D
This is the memorize-and-apply list for behavior reduction. The differential-reinforcement contrasts and the extinction side effects earn the most points, and the punishment material is where 2nd-edition study guides will leave you exposed.
Worked scenario: three plausible options and one best answer
During a table-work session, Leo drops to the floor whenever a matching task is presented. His written behavior intervention plan states that floor-dropping is maintained by escape, and directs planned ignoring of the floor-drop with continued presentation of the demand and guided completion. The RBT has been trained on this procedure. Today, on the third demand, Leo drops and begins to cry loudly. The classroom aide, who does not work on Leo's program, urges the RBT to give Leo a break so he can calm down. Four options: give Leo a short break and re-present the task once he is calm; continue presenting the demand and follow the planned procedure in the plan; pause the session and contact the supervising BCBA before responding; ask the aide to manage Leo while the RBT records ABC data on the episode. Every one of these is something a real person has done. Work through them. Giving a break delivers escape — the exact consequence the plan says maintains the behavior — so it strengthens floor-dropping and undermines the intervention, however humane it feels; the crying is emotional responding, which is a documented secondary effect, not a signal to abandon the procedure. Contacting the BCBA is the option candidates over-choose, and it is wrong here for two reasons: the plan explicitly covers this behavior and this response, and pausing the session to make the call would itself deliver escape. Handing Leo to the aide breaks treatment integrity by putting implementation in the hands of someone untrained on the plan, and collecting ABC data is not a substitute for running the procedure. The best answer is to continue presenting the demand and follow the plan. Afterwards, the RBT documents the episode objectively — including the emotional responding and the aide's request — and raises it at the next supervision contact, because a recurring disagreement with another adult in the room is exactly the kind of thing the supervisor needs to know about. And if Leo's behavior had escalated to a safety risk, the answer would change: that is the crisis plan, and that is when you escalate immediately.
Exam traps in Domain D
Almost every trap in this domain is either a differential-reinforcement confusion or a misjudged escalation. Learn each pair as a pair, and read every option for whether it hands the client the very consequence the plan is trying to withhold.
Key takeaways
Eight sentences for the domain that decides most borderline results, because it is where scenario items are hardest and where the two brand-new 3rd-edition tasks live.
Practice stays free. The full RBT (Registered Behavior Technician) study guide is the material itself, taught start to finish — a downloadable PDF + EPUB you keep.