Study the RBT exam by discriminating between adjacent concepts rather than memorizing terms separately. Work each content area through vignettes, compare the two or three procedures that fit one scene, and confirm your choices with a written rationale. Start each study session with one content area and end it by explaining, in writing, why the correct procedure fits the vignette and its closest neighbor does not. Then use readiness checks to find which content areas still blur together.
Continuous versus discontinuous measurement: knowing what each method actually records
Continuous measurement captures every instance of a target behavior; discontinuous measurement samples behavior during intervals or moments. The exam checks whether you can match a method to what the vignette says was tallied, timed, or observed at intervals.
Continuous procedures include frequency or rate, duration, latency, and interresponse time. The critical discrimination is between frequency (a raw tally of occurrences) and rate (a tally divided by observation time), and between duration (how long a behavior lasts) and latency (how long it takes the behavior to start after a demand). Read the vignette for the denominator: if observation time was measured, rate is in play; if a timer ran only during the behavior, duration is in play.
Discontinuous procedures trade completeness for feasibility. Partial interval scores an interval if the behavior occurred at any point in it, whole interval scores it only if the behavior covered the entire interval, and momentary time sampling scores only whether the behavior is happening at the instant each interval ends. Because each method defines an interval differently, the same observation can yield different numbers for the same behavior, which is exactly why the vignette's wording determines the answer.
The practical skill is translation: given a description of what an observer did, name the measurement system; given a measurement system, predict what it will overestimate or miss.
- Frequency: tally each hand-raise during a 20-minute lesson.
- Rate: those same tallies divided by 20 minutes, so two observers watching different lengths of time can be compared.
- Duration: seconds from the start to the end of a tantrum.
- Latency: seconds between the instruction 'Clean up' and the first response.
- Partial interval: mark a 10-second interval if a self-injurious response happened at any moment within it; this tends to overestimate occurrence.
- Whole interval: mark the interval only if the behavior occupied all 10 seconds; this tends to underestimate occurrence.
- Momentary time sampling: look only at the 10-second mark and record whether the behavior is happening right then.
Choosing among partial interval, whole interval, and momentary time sampling
Choose by purpose and observer capacity. Partial interval suits behaviors you want reduced, whole interval suits behaviors you want increased, and momentary time sampling fits observations too demanding for continuous interval-by-interval checking.
Each method carries a known direction of error. Partial interval recording flags any interval containing even a brief instance, so it generally overestimates how much the behavior occurred. Whole interval requires the behavior to persist across the whole interval, so it generally underestimates occurrence and is a natural fit for skills you are trying to grow, such as on-task behavior. Momentary time sampling records only the instant of each boundary, making it the least burdensome but also the least sensitive to brief responses.
Practice this as a decision, not a definition. When a vignette says an observer 'looked up only when the timer sounded,' momentary time sampling is the answer even if other details tempt you toward interval recording. When the description says 'any occurrence during the interval was marked,' partial interval is the answer. Build the habit of underlining the recording instruction in the stem before looking at the options, and when you write your own practice vignettes, deliberately pair real procedures that differ only in that instruction so the habit gets tested.
This section also holds your first practice exercise; run the observation drill below before moving on.
- Exercise: watch any 10-minute video of a classroom or home routine. Pick one behavior, then score it three times with the same 30-second intervals: once with partial, once with whole, once with momentary time sampling.
- Expected observations: your partial-interval percentage should come out highest, whole interval lowest, and momentary somewhere between or below partial. Write one sentence explaining why the numbers diverged.
- Self-check rubric: you can name the method from the observer's described actions, you can state the typical direction of error for each method, and you can justify which method fits a target that is being increased versus decreased.
| Method | What you record | Typical fit | Known blind spot |
|---|---|---|---|
| Partial interval | Any occurrence within each interval | Reducing a problem behavior | Overestimates occurrence; brief blips score the whole interval |
| Whole interval | Behavior present for the entire interval | Increasing sustained behavior such as engagement | Underestimates occurrence; brief responses never score |
| Momentary time sampling | Behavior only at the instant each interval ends | Observations with limited observer attention | Misses brief behavior entirely; depends on boundary timing |
| Continuous (frequency, duration, latency) | Every instance and its timing | When an observer can attend fully | More demanding to collect accurately |
Preference assessment versus functional behavior assessment: two different questions
A preference assessment asks what might function as a reinforcer right now; a functional behavior assessment asks what conditions maintain a behavior. An RBT assists with both but designs neither independently.
Preference assessment procedures include free operant observation, single stimulus (no-choice) presentation, paired stimulus (two-choice), and multiple stimulus arrangements with or without replacement. The vignette tests whether you can identify the arrangement from its mechanics: two items presented side by side and the chosen one removed describes paired stimulus; an array of items with the chosen item staying in the array describes multiple stimulus with replacement. Repeating an arrangement over time matters because preferences shift, so what reinforces today may not reinforce next week.
A functional behavior assessment is a different project: it seeks the function of a target behavior, typically attention, escape, access to a tangible, or automatic reinforcement. The RBT's contribution is to assist the supervising analyst, for example by collecting ABC data (antecedent, behavior, consequence) during natural routines. Distinguishing the two assessments prevents a common exam trap: a stem describing what a client picks from a tray is a preference assessment, not a functional analysis, even though both eventually involve 'why' questions about behavior.
Keep the two roles straight in your notes: preference assessment identifies candidate reinforcers; functional assessment identifies maintaining variables; skill assessments such as criterion-referenced measures identify what a learner can already do.
DRA, DRO, DRI, and DRL: one vignette, four defensible-sounding answers
These differential reinforcement procedures differ in what gets reinforced: DRA reinforces a specific alternative behavior, DRI reinforces a physically incompatible response, DRO reinforces the absence of the behavior, and DRL reinforces lower rates. Identify the reinforcer's criterion to choose among them.
Read the stem backward from the reinforcer. If praise is delivered only when the learner raises a hand instead of calling out, that is DRA, because a specific alternative is specified. If praise is delivered for keeping feet on the floor, a response that cannot co-occur with kicking, that is DRI. If a timer ends and the learner earns reinforcement for having shown no aggression during the interval, that is DRO. If reinforcement is delivered when calling out occurs less often than before, that is DRL. The behavior described is often the same across options; the reinforcement criterion is what changes.
Worked scenario: a school team targets a student's out-of-seat behavior. A plausible first move is to reinforce the student for sitting for five-minute stretches with no out-of-seat episodes, which is a DRO-style criterion. The mistake is that no replacement skill is being built, so when the schedule thins the behavior has no competing response to sustain it. The better decision is a DRA: reinforce an appropriate alternative with the same function, such as asking for a movement break, while the supervisor plans how problem behavior contacts extinction. Why it matters: DRO reduces occurrence but leaves a functional vacuum; DRA strengthens a response that can meet the same need safely, which is the criterion your chosen procedure must satisfy.
Second worked scenario: a vignette asks what to do when the learner emits the DRA replacement and, seconds later, the problem behavior as well. Reinforcing either one after the chain risks reinforcing the problem behavior too; the supervisor's plan typically calls for withholding reinforcement when the problem behavior follows, and your job as the RBT is to follow that contingency exactly and report the pattern. The exam point is that the replacement response must contact reinforcement without the problem behavior bundled into the same episode.
Prompting and fading: response prompts, stimulus prompts, and how to remove each
Response prompts act on the learner's body or voice; stimulus prompts act on the teaching materials or the target stimulus. Fading means systematically removing the prompt, and least-to-most versus most-to-least describe opposite sequencing strategies.
Response prompts include verbal, gestural, model, physical, and partial physical assistance; they add support to the learner's response. Stimulus prompts change the stimulus itself, for example by making the correct card larger, brighter, or positioned closer, or by adding a color highlight to the correct letter. The exam discrimination is where the support lives: pointing at the correct card is a gestural prompt, but enlarging the correct card is a stimulus prompt, even though both increase the chance of a correct choice.
Sequencing matters as much as labeling. Least-to-most prompting starts with the least intrusive level, such as a verbal instruction, and moves to more assistance only when the learner does not respond or errs; most-to-least prompting starts with full assistance and withdraws support as the learner improves. Prompt fading reduces the magnitude or intrusiveness of the prompt over time, and time delay inserts a pause between the instruction and the prompt so the learner has an independent opportunity. In your notes, trace one skill through both sequences: which prompt comes first, which is added or removed, and what data would tell you to shift levels.
Errorless teaching sits at the support-heavy end of this spectrum: prompts are delivered early so the learner responds correctly from the start, then faded carefully. The vignette checks whether you can name the procedure from its delivery pattern rather than from its label.
Documentation and reporting: session notes, data integrity, and when to communicate
Documentation must be objective, accurate, and timely: record what you observed and did, keep data faithful to the session, and report anything that affects the client to your supervisor promptly, following the behavior plan's instructions.
Session notes are tested as much for tone as for content. Objective language describes observable events: 'Client engaged in three instances of property destruction, each followed by the escape contingency described in the plan.' Subjective interpretation ('client was manipulative today') belongs in no note. The skill is separating observation from inference in practice: write what you saw, what you delivered, what the client did, and what variables changed, without diagnosing motives or editorializing about the client or family.
Reporting has defined triggers. A change in the client's behavior, health, environment, or the effectiveness of the plan is material information for the supervising analyst. Crisis and emergency procedures are followed exactly as specified in the behavior plan, and afterwards the event is documented and reported per the plan and organizational policy rather than improvised. Data integrity also lives here: record data during or immediately after sessions, avoid reconstructing numbers from memory, and know that altering or inventing data is a serious ethical violation, not a clerical shortcut.
In your study, convert each reporting trigger into a one-line rule you could apply to a vignette: behavior change of concern, new medical event, family question about the plan, schedule disruption, or equipment failure. Each has a common shape: follow the plan, document factually, and notify the supervisor.
Professional conduct and scope of practice: what an RBT does versus what a BCBA decides
RBTs implement behavior plans, collect data, and assist with assessments designed by their supervisor. They do not independently design interventions, conduct functional analyses, alter programs, or set their own clinical recommendations.
Worked scenario: during a session, a parent says, 'You know my child best, please just design a feeding program for us, the BCBA is hard to reach.' The plausible mistake is agreeing, which feels helpful and client-centered but exceeds the RBT's role: designing intervention is the supervising analyst's responsibility, and an unsupported program could also conflict with the existing plan. The better decision is to acknowledge the parent's concern, explain that program design must come from the supervisor, report the request and the access problem to the BCBA promptly, and document both the conversation and the report. Why it matters: the division of labor protects the client, keeps the RBT inside their defined role, and surfaces the supervision problem so the supervisor can act.
Other conduct boundaries follow the same logic. Confidentiality means client information is shared only with those authorized to receive it and only for authorized purposes. Dual relationships, such as providing services to a friend or accepting gifts that create obligation, are avoided and reported to the supervisor. Feedback from the supervisor is received and acted on, and the RBT seeks supervision rather than working through clinical uncertainty alone. The exam discrimination is between an RBT's implementation duties and any stem that quietly hands the RBT a decision reserved for the supervisor.
A second worked vignette to trace: a teammate suggests changing the reinforcement schedule mid-session because 'it isn't working today.' The correct action is to continue the plan as written, collect data, and raise the observation with the supervisor; mid-session protocol changes are not the RBT's call.
References and further reading
Use these references to explore the concepts and check the latest information from the relevant organizations.