RBT Study Guide 2026: What to Know for All 6 Domains

What the RBT exam covers in 2026

The RBT exam is pass or fail on the whole test, and every scored question counts the same. The questions follow the43 tasks of the 3rd-edition outline, split into six domains with fixed weights. Study time should follow the weights:

Data Collection and Graphing (Domain A)

Domain A is how you measure behavior, record it and read the graph. Most questions describe a behavior and ask which measurement fits, or give you numbers to summarize.

Measurement methods and when each fits

Observable and measurable definitions

A good definition names what anyone watching would see or hear, so two observers count the same thing. "Aggression" is a label; "hitting another person with an open or closed hand" is a definition. Feelings and intentions (upset, bored, trying to get attention) never go in a definition.

Calculating and summarizing data

Graphs, trends and unreliable data

On a line graph, read three things: the level (where the data sit), the trend (rising, falling or flat) and the variability (how much points bounce around). Your job is to enter each session's data promptly and accurately and to update the graph; deciding what the trend means for the plan is the supervisor's job.

Unreliable data lead to wrong decisions, such as calling a skill mastered too early. Poor procedural fidelity, meaning running a procedure differently from how it is written, has the same effect: the data then describe your version, not the plan. Interobserver agreement, where two people record the same session, is how teams check reliability.

Behavior Assessment (Domain B)

Domain B is the smallest domain, 8 questions, and it is about carrying out assessments the supervisor designs and interprets.

Preference assessments

Skill and functional assessments

Skill assessments measure strengths and gaps: curriculum-based tools such as the VB-MAPP or ABLLS-R, developmental assessments and social skills assessments. You run the items as instructed and record the responses exactly; the supervisor scores and interprets them.

Functional assessment finds why a behavior happens. You are most likely to collect descriptive data, usually ABC data (antecedent, behavior, consequence) during normal routines. In a functional analysis, the supervisor may have you run test conditions (attention, escape, tangible, alone) against a play or control condition while you follow the protocol precisely.

Behavior Acquisition (Domain C)

Domain C is the biggest, 19 questions, and covers how new skills are taught. Expect scenarios where you pick the next step of a procedure.

Reinforcement and schedules

Reinforcement increases a behavior. Positive means something is added (praise, a toy); negative means something is removed (a demand, a noise). Neither word means good or bad. Reinforcement works best delivered immediately, only after the target behavior, at a magnitude that fits the effort, and with variety so the client does not tire of one item.

A conditioned reinforcer starts neutral and gains value by being paired with an existing reinforcer: praise delivered with a snack, or tokens exchanged for a favorite activity.

Discrete-trial teaching and naturalistic teaching

Naturalistic teaching uses the client's own motivation in everyday settings. In incidental teaching, the client reaches for a toy car on a shelf, the RBT waits for or prompts "car", and the car is the reinforcer. Natural environment training teaches the same targets during play and routines rather than at a table.

Chaining, discrimination and prompting

A task analysis breaks a skill into steps. Forward chaining teaches the first step first; backward chaining teaches the last step first, so the client always finishes the task and contacts reinforcement; total task chaining runs every step each time with help where needed.

Discrimination training reinforces a response in the presence of one stimulus (the SD) and not in the presence of others (S-deltas): saying "dog" for the dog picture is reinforced, saying "dog" for the cat picture is not.

Response prompts are verbal, gestural, modeling and physical (partial or full). Stimulus prompts change the materials, such as placing the correct card closer. Every prompt needs a fading plan: most-to-least starts with full help and reduces it, least-to-most adds help only when needed, time delay waits a set number of seconds before prompting, and stimulus fading shrinks the added cue. A client who waits for help every trial has become prompt dependent, usually because fading stalled.

Generalization, maintenance, shaping and token economies

Generalization is a skill showing up across settings, people and materials. Maintenance is a skill lasting after teaching ends. Acquisition procedures teach new skills with dense reinforcement and prompts; maintenance procedures keep mastered skills going with thinner reinforcement and occasional practice mixed into sessions.

Shaping reinforces closer and closer approximations of a target and stops reinforcing the earlier ones: "buh", then "ba", then "ball". A token economy gives tokens for target behavior that are exchanged for backup reinforcers, with the number needed and the exchange times set in the plan.

Behavior Reduction (Domain D)

Domain D is 14 questions on reducing problem behavior, and nearly every procedure depends on the behavior's function.

Antecedent interventions and differential reinforcement

Antecedent interventions change what happens before the behavior. Noncontingent reinforcement delivers the maintaining reinforcer on a time schedule, whatever the client is doing. A high-probability request sequence gives two or three easy requests before the hard one. Demand fading drops demands, then brings them back gradually.

Extinction, punishment and their side effects

Extinction withholds the reinforcer that maintained the behavior, so it must match the function. Planned ignoring is extinction only for attention-maintained behavior; for escape-maintained behavior the demand stays in place.

Punishment decreases behavior. Positive punishment adds something; negative punishment removes something, as in response cost or time-out from reinforcement. RBTs use punishment-based procedures only when they are in the client's written plan and only after training on them.

Expect side effects. An extinction burst is a temporary rise in the behavior when extinction starts; response variation brings new forms of it; spontaneous recovery is the behavior briefly returning after it had stopped; resurgence is an old behavior returning when a newer alternative stops working. Punishment can bring emotional responses and escape or avoidance of the person or setting.

In a crisis, follow the written crisis plan and your organization's procedures, keep everyone safe, use only the techniques you are trained in, then report and document what happened straight away.

Documentation and Reporting (Domain E)

Domain E is 10 questions on what you write down and who you tell. Session notes describe what happened, in terms another person could verify.

Report anything that could affect progress promptly: illness, a medication change, lost sleep, a schedule change or a family event. Pass caregiver and teacher concerns to your supervisor rather than answering clinical questions yourself. Ask for direction when data look wrong or a procedure is unfamiliar, and follow the chain of command. Notes also have to meet legal and workplace rules, including privacy law such as HIPAA.

Ethics (Domain F)

Domain F is 11 questions built on the RBT Ethics Code (2.0). Most are scenarios with a clear rule behind them:

Know the supervision rule too: at least 5% of your service hours each calendar month, with two face-to-face contacts, at least one of them individual. Good supervision uses instruction, modeling, rehearsal and feedback, plus direct observation of your sessions.

A three-week RBT study plan

RBT exam day

RBT study guide FAQ