Sen Lin, PrepPass Founder · Verified against BACB — RBT certification (official) · How we review

Free · printable · no signup

RBT (Registered Behavior Technician) Exam Cram Sheet (2026)

A free, printable RBT (Registered Behavior Technician) exam cram sheet: the 544 highest-yield points to know, grouped into 6 sections that follow the exam's content areas, each section with its published weight.

  • Free to read, print and keep — no signup, no email, no paywall.
  • The real RBT (Registered Behavior Technician) exam: 85 questions · 90 minutes.
  • Prints from your browser (Ctrl-P, or Cmd-P on a Mac) — the navigation, buttons and links drop out of the printed copy.
  • 403 free RBT (Registered Behavior Technician) practice questions on the same material, every one explained.
  • Last updated: August 2026.
Free practice questions →Full study guide →
Studying in order?

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.

Get the book — $19.99

Data Collection and Graphing

17% of the exam
Study →
A.1 — Implement continuous measurement procedures
Frequency, duration, latency, interresponse time
A.2 — Implement discontinuous measurement procedures
Partial interval, whole interval, momentary time sampling
A.3 — Implement permanent product recording procedures
Measure the lasting outcome the behavior leaves behind
A.4 — Enter data and update graphs
Accurately, and in the session it was collected
A.5 — Describe behavior and environment in observable and measurable terms
Write what a stranger could count without asking you what you meant
A.6 — Calculate and summarize data in different ways
Rate, mean duration, percentage — new in the 3rd edition
A.7 — Identify trends in graphed data
Increasing, decreasing, flat, variable — new in the 3rd edition
A.8 — Describe the risks of unreliable data collection and poor procedural fidelity
Why bad data and drifting procedures lead to bad clinical decisions — new in the 3rd edition
Frequency (count)
How many times it happened. A tally, no timing.
Duration
How long it lasted, onset to offset. Timer starts when the behavior starts.
Latency
How long from the instruction or trigger until the behavior BEGINS.
Interresponse time (IRT)
Time from the END of one response to the START of the next.
All four are continuous
You capture every instance; nothing is sampled or estimated.
Partial interval
Behavior occurred at ANY point in the interval → score it.
Whole interval
Behavior occurred for the ENTIRE interval → score it.
Momentary time sampling
Behavior is occurring at the MOMENT the interval ends → score it.
All interval data are reported as a percentage of intervals
4 of 10 intervals scored = 40% of intervals
Interval methods estimate
They never produce a true count or a true duration
Partial interval OVER-estimates
One second counts as the whole interval. Good for behaviors targeted for reduction.
Whole interval UNDER-estimates
Anything short of the full interval scores zero. Good for behaviors targeted for increase.
MTS neither systematically over- nor under-estimates
It samples one moment; shorter intervals make it more accurate.
MTS is the practical choice when you are also teaching
You look up once per interval instead of watching continuously
Permanent product = measure the outcome after the fact
Worksheets completed, items put away, beds made
Strength
No real-time observation needed; easy for a second scorer to verify
Limitation
It cannot tell you WHO produced the product or HOW
Use direct observation instead
When the process, form, or prompt level is what matters
Observable
Someone else could see or hear it happen
Measurable
It can be counted, timed, or scored against intervals
Clear and complete
States what counts, what does not, and where the behavior begins and ends
Dead-man test
If a dead man can do it, it is not behavior — define the active response instead
Never write internal states
Frustrated, anxious, unmotivated, attention-seeking are inferences, not descriptions
Enter data during or immediately after the session
Delayed entry is reconstruction, not measurement
x-axis = time (sessions or dates); y-axis = the measure
One plotted point per session
Connect data points within a phase
Never connect across the phase-change line
Missing data stay missing
Report the gap to your supervisor; never estimate a value in
Rate = count ÷ time
15 aggressions in 90 minutes = 0.17 per minute = 10 per hour
Mean duration = total duration ÷ number of episodes
22 minutes across 4 tantrums = 5.5 minutes per tantrum
Percentage correct = correct ÷ opportunities × 100
12 of 20 trials = 60%
Percentage of intervals = intervals scored ÷ intervals observed × 100
4 of 10 = 40% of intervals
Use rate when session lengths differ
Raw counts are not comparable across different observation times
Level
Where the data sit on the y-axis, and whether that changed between phases
Trend
Direction over time: ascending, descending, or zero/flat
Variability
How much the points scatter; high variability = unstable data
A trend is only good or bad relative to the target
Descending is progress for problem behavior, regression for a skill
Phase-change line
The vertical line marking a change in condition; compare level and trend on either side of it
Read level first, then trend, then variability
In that order, and always on both sides of the phase line
One spike is not a trend
Direction is judged across a series of points, not across two
One dip is not a regression
Inside a clear ascending trend, a single lower point is variability
Check the y-axis label before judging
The same shape means opposite things for reduction and acquisition targets
Procedural fidelity
Running the intervention exactly as written, every session, by every technician
Bad data risk
The supervisor makes a clinical decision about a program that was never actually run
Drift between technicians
Inconsistent implementation slows learning and can worsen problem behavior
Never adjust the procedure to smooth the session
Report the difficulty; only your supervisor changes the plan
Report fidelity failures promptly
A disclosed lapse can be corrected; a hidden one corrupts the data set
Domain A weight
13 of the 75 scored questions, 17% of the exam, 8 tasks (A.1–A.8)
Four continuous measures
Frequency, duration, latency, interresponse time
Three discontinuous measures
Partial interval, whole interval, momentary time sampling
Partial over-estimates, whole under-estimates
MTS is a snapshot and gets more accurate as intervals shorten
Rate = count ÷ time
Always state the unit: per minute or per hour
Mean duration = total duration ÷ episodes
Mean latency = total latency ÷ opportunities
Interval data are always a percentage
Intervals scored ÷ intervals observed × 100
Three graph words
Level (where), trend (direction), variability (scatter)
Step 1 — find the phrase that fixes the procedure
"At any point," "for the entire interval," or "at the moment it ends"
Step 2 — check the reporting unit
Interval methods report percentage of intervals, never a count
Step 3 — do the arithmetic last
10 of 40 intervals = 25%
Best-answer habit
When two options name the same procedure, the difference is in the second half of the option
Looks like frequency, is actually rate
If session lengths differ, a raw count cannot be compared — the answer wants count per unit time
Looks like duration, is actually latency
Duration times the behavior; latency times the gap between the instruction and the start of the behavior
Looks like latency, is actually interresponse time
Latency starts at the instruction; IRT starts at the end of the previous response
Looks like whole interval, is actually partial interval
A brief one-second occurrence scoring the interval is partial interval, by definition
Looks like partial interval, is actually MTS
If the observer only looks at the end of each interval, it is momentary time sampling
Looks like a count, is actually a percentage
Interval data are reported as % of intervals — "4 occurrences" is never the right report for interval data
Looks like a good graph, depends on the y-axis
A descending trend is success for aggression and failure for a mastery target
Looks like a decision, is actually a report
Identifying a trend is your task; changing the program because of it is your supervisor's
Looks like helpful, is actually falsification
Filling in a missed data point with a reasonable estimate is fabricating data
Continuous captures everything; discontinuous samples
Only continuous measures give true counts and true durations
Partial over-estimates, whole under-estimates, MTS snapshots
Derive it from the definition rather than memorizing a table
Interval data are always percentages of intervals
Never occurrences
Rate beats count whenever observation times differ
And every rate needs a unit attached
Operational definitions are observable, measurable, and complete
No internal states, no opinions, no labels
Permanent product tells you the outcome, not the process
It cannot tell you who did it or how
Level, trend, variability read any graph
Compare on both sides of the phase-change line; never connect across it
Bad data and poor fidelity cause bad clinical decisions
Report lapses; never repair them quietly

Behavior Assessment

11% of the exam
Study →
B.1 — Conduct preference assessments
Multiple stimulus, paired stimulus, free operant
B.2 — Participate in assessments of relevant skill strengths and deficits
Curriculum-based, developmental, social skills
B.3 — Participate in components of functional assessment procedures
Descriptive assessment, functional analysis
The verb is the boundary
You conduct preference assessments; you only participate in skill and functional assessments
Reinforcer = defined by its effect on future behavior
Not by what the client is assumed to enjoy
Preference assessment identifies likely reinforcers
It does not by itself prove reinforcement
Preference shifts
Satiation, deprivation, illness and time of day all change what works today
Brief pre-session assessments are common
Run them as often as the plan directs, not once and forever
Free operant
Items freely available, no demands, nothing removed; records engagement
Single stimulus
One item at a time; for clients who cannot yet make a choice between two
Paired stimulus (forced choice)
Two at a time, all pairs; most precise ranking, most time-consuming
MSW — with replacement
Chosen item goes back into the array before the next presentation
MSWO — without replacement
Chosen item is removed; yields a full hierarchy quickly
Free operant is the safe default
When removing items evokes problem behavior, or the client is new to you
Curriculum-based assessment
Structured skill lists across domains; results become program targets
Developmental assessment
Compares skills against typical developmental sequences
Social skills assessment
Initiating, responding, turn-taking, and other interaction skills
Do not prompt during assessment unless the protocol says to
Prompting inflates the baseline and misdirects target selection
Score what happened today
Not what the client has managed before
Interpretation is the supervisor's
Route "what does this mean" questions to them, not to the caregiver
Indirect assessment
Interviews, rating scales, questionnaires — no direct observation, hypotheses only
Descriptive assessment
Direct observation, usually ABC data; shows correlation, not causation
Functional analysis
Experimental manipulation of conditions; the only level that demonstrates a functional relation
Typical FA conditions
Attention, demand/escape, alone, play/control, sometimes tangible
You participate in an FA only when trained and directed
It deliberately evokes problem behavior and runs under the supervisor's protocol and safety parameters
A = what happened immediately before
Observable events, not guesses about mood
B = the behavior, in the plan's operational definition
Not a summary label like "tantrum"
C = what happened immediately after
What the environment did, including what you did
Never write the function into an ABC record
"Attention-seeking" and "escape-maintained" are conclusions, not observations
Record promptly
ABC data reconstructed hours later lose the immediate antecedent and consequence
Domain B weight
8 of the 75 scored questions, 11% of the exam, 3 tasks (B.1–B.3)
Preference assessment formats
Free operant, single stimulus, paired stimulus, MSW, MSWO
MSWO gives a ranking fastest; paired stimulus gives the most precise one
Free operant places no demands and removes nothing
Preferred ≠ reinforcing
Only an effect on behavior confirms a reinforcer
Three levels of functional assessment
Indirect (report), descriptive (observation/ABC), functional analysis (experiment)
Only the FA demonstrates a functional relation
Descriptive data are correlational
Skill assessment families
Curriculum-based, developmental, social skills
Step 1 — is this within the plan and my training?
If yes, implement rather than escalate
Step 2 — match the format to the client's behavior
Removal evokes problem behavior → free operant
Step 3 — weigh session cost
The most precise procedure is not best if it consumes the session it was meant to serve
Step 4 — prefer current, direct data over second-hand report
Caregiver information is background, not measurement
Looks like MSW, is actually MSWO
If the chosen item is removed and the client picks again from what is left, it is without replacement
Looks like paired stimulus, is actually single stimulus
Paired stimulus always presents two items and forces a choice
Looks like a preference assessment, is actually a reinforcer assessment
Preference predicts; only an effect on responding confirms reinforcement
Looks like descriptive assessment, is actually indirect
A rating scale or interview involves no observation, however detailed the informant
Looks like a functional analysis, is actually a descriptive assessment
No conditions were manipulated — you just watched what happened
Looks like a conclusion you may draw, is actually the supervisor's
Naming a client's behavioral function or skill level is interpretation
Looks like helpfulness, is actually contaminating the data
Prompting a client through an assessment item inflates the score
Looks like a good reason to escalate, is actually covered by the plan
Conducting a preference assessment you were trained to run needs no permission
You conduct preference assessments; you participate in the rest
The outline's verbs are the scope boundary
Free operant when removal evokes problem behavior
MSWO when you need a fast ranking; paired stimulus when precision matters most
Preference is a prediction, not proof of reinforcement
Behavior data settle it
Re-check preference often
It changes with satiation, illness, and the day
Indirect → descriptive → functional analysis
Increasing evidence; only the FA demonstrates a functional relation
ABC data are observations, never conclusions
Write what happened before, what the client did, and what followed
Do not prompt or over-score during a skill assessment
The baseline is only useful if it is honest
Route interpretation questions to your supervisor
Including caregiver questions about level, progress, or diagnosis

Behavior Acquisition

25% of the exam
Study →
C.1 — Implement positive and negative reinforcement procedures
Immediately, contingently, on schedule, and along dimensions such as magnitude, intensity and variety
C.2 — Implement procedures to establish and use conditioned reinforcers
Pairing, and using the conditioned reinforcer once established
C.3 — Implement discrete-trial teaching procedures
Structured, adult-led trials with a clear beginning and end
C.4 — Implement naturalistic teaching procedures
Incidental teaching, natural environment training
C.5 — Implement task-analyzed chaining procedures
Forward, backward, total task
C.6 — Implement discrimination training
Teaching responding in the presence of one stimulus and not another
C.7 — Implement stimulus and response prompts with appropriate fading
Errorless, least-to-most, stimulus fading, time delay
C.8 — Implement generalization procedures
Across settings, people and stimuli
C.9 — Distinguish between maintenance and acquisition procedures
Keeping a mastered skill versus building a new one
C.10 — Implement shaping procedures
Reinforcing successive approximations toward one terminal behavior
C.11 — Implement token economies
Tokens, backup reinforcers, and the exchange rules
Positive reinforcement = add a stimulus → behavior increases
Praise, token, preferred item delivered after the response
Negative reinforcement = remove or postpone a stimulus → behavior increases
Demand withdrawn, aversive noise stopped
Both forms INCREASE behavior
Negative reinforcement is not punishment
Immediacy
Deliver within a second or two; delays strengthen whatever happened in between
Contingency
Only after the target response — never as consolation
Magnitude, quality, variety
Rotate reinforcers to prevent satiation
Motivating operations change reinforcer value
Deprivation raises it; satiation lowers it
Continuous (CRF)
Every correct response reinforced — use during acquisition
Intermittent
Some responses reinforced — use for maintenance; produces resistance to extinction
FR / VR
Ratio schedules count responses; VR produces high, steady rates
FI / VI
Interval schedules depend on time elapsed before the next response
Thinning is gradual and planned
Follow the plan's schedule change; never cut reinforcement off abruptly
Unconditioned reinforcer
Effective without learning history — food, water, warmth
Conditioned reinforcer
Acquires its effect through pairing — praise, tokens, stickers
Pairing procedure
Deliver the neutral stimulus with or just before an established reinforcer, repeatedly
Generalized conditioned reinforcer
Paired with many backups; works across motivating conditions
Pair yourself with a new client first
Build value before placing demands
Unpaired conditioned reinforcers weaken
Keep pairing praise with things that already work
The DTT trial
SD → (prompt) → response → consequence → inter-trial interval
Adult-led and structured
The technician decides the target, the materials and the moment
Strength
Many learning opportunities per minute; highly consistent
Limitation
Contrived context and reinforcers; generalization must be programmed
Keep the inter-trial interval brief and clean
It marks the end of one trial and the start of the next
Naturalistic teaching follows the client's motivation
The current interest creates the teaching opportunity
Natural reinforcers
The reinforcer is the direct product of the response — the requested item itself
Two-question test
Who chose the activity? What was the reinforcer? That separates DTT from NET
Trade-off
DTT gives more trials; NET gives better generalization
Contriving motivation is legitimate
Placing a preferred item in sight but out of reach creates a request opportunity
Task analysis
The written list of steps making up the chain, sized to the client
Forward chaining
Teach step 1 first, prompt the rest; add steps in order
Backward chaining
Teach the last step first; the client independently completes the step that earns the natural reinforcer every trial
Total-task chaining
Every step, every trial, prompting only where needed
Each completed step cues the next
The chain is held together by the product of each step
Match the procedure to the repertoire
Total task when most steps are already there; backward when early success matters
Shaping
Differential reinforcement of successive approximations to ONE terminal behavior
Raise the criterion only when the current approximation is reliable
And stop reinforcing the older, weaker approximation
Stalled progress usually means the step was too big
Go back to a smaller approximation, per the plan
Shaping vs chaining
Shaping refines one behavior; chaining links several behaviors into a sequence
SD (discriminative stimulus)
Responding is reinforced in its presence
S-delta
Responding is not reinforced in its presence
Stimulus control
The response reliably occurs when the SD is present and not otherwise
Rotate positions and orders
Prevents the client learning location or sequence instead of the target
Watch for unintended stimulus control
Your gaze, position or card handling can become the real cue — report it
Response prompts
Full physical, partial physical, model, gestural, verbal — help directed at the client
Stimulus prompts
Position, salience, or within-stimulus changes — help built into the materials
Most-to-least
Start with enough help to guarantee success; for new, complex, or safety-critical skills
Least-to-most
Start with an independent opportunity, add help as needed; for partly-mastered skills
Errorless teaching
Prompt immediately so errors do not occur, then fade systematically
Time delay
Insert a gap before the prompt — constant (same every trial) or progressive (lengthening)
Prompt dependence
The client waits for help instead of responding; caused by failing to fade
Fade on the plan's schedule
Never invent your own fading steps because the session runs more smoothly with the prompt
Purpose
The trial ends with the correct response performed, without reinforcing the error
Run the program's exact procedure
Identically every time — inconsistency stalls acquisition
Neutral acknowledgment for the prompted correct response
Full reinforcement is saved for independent responding
Never punish an attempt
Reacting with frustration suppresses responding altogether
Stimulus generalization
The skill occurs with new people, settings and materials
Response generalization
Untaught variations of the response appear
Program generalization deliberately
Multiple exemplars, multiple people and settings, teach loosely, use natural stimuli and reinforcers
Maintenance
Keeping a mastered skill: thinned schedules, periodic probes, interspersed mastered targets
Acquisition vs maintenance
Building a new skill vs preserving an existing one
Loss of a mastered skill is reported, not re-designed
Record the data and raise it with your supervisor
Three components
Tokens, backup reinforcers, and the earning/exchange rules
Tokens are generalized conditioned reinforcers
Their value comes from the backups they buy
Deliver immediately and contingently, with brief specific praise
The same rules as any other reinforcer
Honor the exchange every time
An unreliable exchange collapses the system
Keep backups genuinely preferred
Re-check preference; a stale menu kills the economy
Response cost is a punishment procedure
Only if written in the plan and you are trained on it
Domain C weight
19 of the 75 scored questions, 25% of the exam, 11 tasks (C.1–C.11) — the largest domain
Positive and negative reinforcement both increase behavior
Plus = added, minus = removed; neither means good or bad
CRF for acquisition, intermittent for maintenance
Intermittent schedules produce resistance to extinction
Conditioned reinforcers are made by pairing
Generalized ones are paired with many backups
DTT trial anatomy
SD → prompt → response → consequence → inter-trial interval
NET follows the client's motivation and uses natural reinforcers
DTT is adult-led with contrived reinforcers
Forward / backward / total task
First step first / last step first / every step every trial
Most-to-least prevents errors; least-to-most promotes independence
Errorless = prompt immediately; time delay = a planned gap before the prompt
Shaping refines one behavior; chaining links several
Successive approximations vs a step sequence
Generalization is programmed, not assumed
Multiple exemplars, people, settings, natural reinforcers
Step 1 — name the teaching arrangement
Every step every session with help as needed = total task
Step 2 — name the prompting hierarchy
Least help first, increasing as needed = least-to-most
Step 3 — identify the failure
A prompt retained for convenience produces prompt dependence
Step 4 — pick the response that keeps fidelity
Follow the written fading steps and report the stall; do not improvise a fix
Looks like punishment, is actually negative reinforcement
If removing something made the behavior MORE likely, it is reinforcement
Looks like DTT, is actually naturalistic teaching
The client's interest set the occasion and the reinforcer was the natural product of the response
Looks like naturalistic teaching, is actually DTT
Adult chose the target and the moment, reinforcer unrelated to the task
Looks like shaping, is actually chaining
Distinct steps in sequence = chaining; successive versions of one response = shaping
Looks like forward chaining, is actually total task
If every step runs every session, it is total task regardless of where the prompting is heaviest
Looks like backward chaining is old-fashioned, is actually indicated
It gives the client independent contact with the natural reinforcer on every trial
Looks like least-to-most is always better
Most-to-least is indicated for new, complex, or safety-critical skills and error-prone learners
Looks like a helpful prompt, is actually prompt dependence
A prompt kept because the session runs smoother is a fidelity failure
Looks like the client mastered it, is actually unintended stimulus control
Perfect scores while watching your face, or a card that is always in the same position
Looks like a mastered skill lost, is actually a maintenance issue
Report it; it is not evidence the acquisition procedure failed
Looks like a fair consequence, is actually an unauthorized punishment
Taking earned tokens away is response cost and requires the plan and training
Reinforcement is defined by its effect
Positive adds, negative removes; both increase behavior
Immediate, contingent, and worth having
Those three properties do most of the work in any teaching procedure
Continuous while acquiring, thin for maintenance
And thin gradually, as the plan directs
DTT for trial density, NET for generalization
Neither is universally better; the plan decides
Chaining links steps; shaping refines one response
Forward, backward and total task are chaining variants
Most-to-least minimizes errors; least-to-most builds independence
Errorless prompts immediately; time delay inserts a planned gap
Always fade, and always fade on the written schedule
The unfaded prompt is the commonest fidelity failure in this domain
Generalization and maintenance are programmed
Multiple exemplars and settings; thinned schedules and periodic probes

Behavior Reduction

19% of the exam
Study →
D.1 — Identify common functions of behavior
The categories a plan is built on
D.2 — Implement antecedent interventions
NCR, high-probability request sequences, demand fading
D.3 — Implement differential reinforcement procedures
DRO, DRA, DRI, DRL, FCT
D.4 — Implement extinction procedures
Withholding the reinforcer that maintains the behavior
D.5 — Implement positive and negative punishment procedures
For example time-out — NEW in the 3rd edition
D.6 — Describe secondary effects of extinction and punishment
Extinction burst, response variation, resurgence, emotional responding; emotional responses, escape and avoidance — NEW in the 3rd edition
D.7 — Implement crisis/emergency procedures
As written, as trained, and reported
Social attention
The behavior produces someone's reaction — reprimands count as attention
Escape / avoidance
The behavior ends, delays or reduces a demand or situation
Access to tangibles or activities
The behavior produces an item or activity, including its return
Automatic / sensory
The behavior produces its own consequence; it continues when the client is alone
Read the consequence, not the emotion
What reliably followed the behavior identifies the likely function
Naming a specific client's function is the supervisor's job
Knowing the categories is yours
Noncontingent reinforcement (NCR)
Time-based delivery of the reinforcer, independent of behavior
High-probability request sequence
Two or three easy, reliably-followed requests just before a difficult one
Demand fading
Reduce demand difficulty or quantity, then build back gradually
Other antecedent strategies
Choice, priming, visual schedules, environmental changes
Antecedent = before, and not contingent on behavior
This is what separates NCR from every differential-reinforcement procedure
DRA
Reinforce an appropriate alternative that serves the same function; problem behavior no longer works
DRI
A DRA where the alternative physically cannot occur at the same time as the problem behavior
DRO
Reinforce the ABSENCE of the behavior for a set interval — the omission procedure
DRL
Reinforce a lower rate; for behaviors acceptable in moderation
FCT
A DRA teaching a communication response matched to the behavior's function
NCR is not differential reinforcement
NCR is time-based and antecedent; DR procedures are contingent on behavior or its absence
The separating question
Is a specific behavior reinforced (DRA/DRI/FCT), an absence (DRO), or a lower rate (DRL)?
Extinction = the maintaining reinforcer is withheld
Not "ignoring," unless attention is the maintaining reinforcer
Attention function
Withhold attention — planned ignoring
Escape function
Do not remove the demand; follow the plan's escape-extinction procedure
Tangible function
The item is not delivered
Extinction is paired with reinforcement of an alternative
Removing what worked without teaching a replacement is not a complete plan
Inconsistency makes behavior stronger
Occasional success converts the behavior to an intermittent schedule
Extinction burst
Temporary increase in frequency, intensity or duration when extinction begins
Response variation
New forms of the behavior appear as the client tries to restore the old consequence
Emotional responding
Crying, distress and aggression often accompany early extinction
Resurgence
A previously reinforced behavior returns when the current one stops working
Spontaneous recovery
The extinguished behavior reappears later without a procedural change
Giving in during a burst is the worst possible moment
It reinforces a larger, more intense version of the behavior
Bursts are documented and reported
Expected effects still belong in the session note and the next supervision contact
Punishment is defined by its effect
It decreases future behavior; positive adds a stimulus, negative removes one
Positive punishment examples
Reprimand, overcorrection — added after the behavior
Negative punishment examples
Response cost (loses something earned), time-out from positive reinforcement
Nonexclusionary vs exclusionary time-out
Client stays in the setting without reinforcement vs is removed from it
Time-out requires a reinforcing time-in
Removing a client from an aversive task is escape, not time-out
Release on the plan's criteria
Ending time-out during problem behavior reinforces it
Punishment is never improvised
Written plan, supervisor's decision, required consents, and your training — all four
Reinforcement-based procedures come first
Raise concerns about a punishment procedure with your supervisor promptly
Emotional responses
Crying, agitation, aggression following the procedure
Escape and avoidance
Of the setting, the task, and of the person who delivered the consequence
Suppression without learning
The behavior returns when the procedure stops; nothing was taught instead
Modeling
Clients may imitate the punishing behavior they experience
The technician trap
Immediate quiet reinforces the person delivering punishment, so its use creeps
Always paired with reinforcement of an alternative
A reduction plan without a replacement behavior is incomplete
Safety first
Protect the client and others; remove hazards
Follow the written crisis plan and employer procedures
As trained — a crisis is not the moment to improvise
Restrictive procedures only if trained, authorized and permitted
By the plan, employer policy, and applicable law
Get help
Colleague, supervisor, or emergency services per your setting
Notify your supervisor as soon as it is safe
A crisis is always reported, never handled quietly
Document objectively and promptly
Antecedents, what was done, duration, and the client's condition afterwards
Covered by the plan + you are trained → implement it
Consistently, even under pressure from others in the room
Safety risk → act on the crisis plan and escalate now
This is the clearest escalation trigger
Not covered by the plan → escalate
Do not invent a procedure to fill the gap
Outside your training or scope → escalate
Including any request to modify the plan
Data show deterioration or a lost skill → report it
You bring the data; your supervisor decides what changes
Client circumstances changed → report it
Illness, medication change, major home event
Fidelity slipped or you made a mistake → report it
Disclosure is part of the job, not an admission of failure
Discomfort alone is not a mid-session escalation
Raise the concern in your note and at supervision
Domain D weight
14 of the 75 scored questions, 19% of the exam, 7 tasks (D.1–D.7)
Four common functions
Attention, escape/avoidance, tangible/activity, automatic/sensory
NCR is antecedent and time-based
Delivered regardless of behavior
High-probability request sequence
Easy requests first, then the difficult one
DRA reinforces an alternative; DRI's alternative is physically incompatible
FCT is a DRA using a communication response
DRO reinforces absence; DRL reinforces a lower rate
The clearest split in the family
Extinction must match the function
Planned ignoring is extinction only for attention-maintained behavior
Extinction side effects
Burst, response variation, resurgence, emotional responding, spontaneous recovery
Punishment side effects
Emotional responses, escape and avoidance of the person and setting, suppression without learning
Positive punishment adds, negative punishment removes
Response cost and time-out are negative punishment
Time-out needs a reinforcing time-in
And ends on the plan's criteria, not on the client's mood
Step 1 — does the written plan cover this exact situation?
If yes, escalation is not the best answer
Step 2 — would this option deliver the maintaining reinforcer?
For escape-maintained behavior, any pause or break is escape
Step 3 — does this option keep treatment integrity?
Handing implementation to an untrained adult never does
Step 4 — what changes the answer?
A safety risk moves you to the crisis procedure and immediate escalation
Step 5 — what happens afterwards?
Objective documentation and a supervision conversation, not a mid-session plan change
Looks like DRO, is actually DRA
If a specific appropriate behavior is being reinforced, it is DRA — DRO reinforces an interval with no target behavior
Looks like DRA, is actually DRI
Check whether the alternative physically prevents the problem behavior
Looks like DRO, is actually DRL
DRL still allows the behavior at a lower rate; DRO requires none of it in the interval
Looks like a DR procedure, is actually NCR
If the reinforcer arrives on a timer regardless of behavior, it is noncontingent
Looks like extinction, is actually reinforcement
Planned ignoring of escape-maintained behavior delivers escape
Looks like the procedure failing, is actually an extinction burst
A temporary increase right after extinction begins is expected — document it
Looks like a new problem behavior, is actually response variation or resurgence
Both are named secondary effects of extinction
Looks like time-out, is actually escape
Removing a client from a task they dislike reinforces the behavior that got them out
Looks like negative reinforcement, is actually punishment
Ask whether the behavior went up or down; that decides it, not how the consequence felt
Looks like the safe answer, is actually unnecessary escalation
When the plan covers it and you are trained, implementing IS the correct action
Looks like compassion, is actually reinforcing the behavior
Giving in during a burst reinforces a bigger version of the behavior
Looks like a plan you can adjust, is actually a plan only the BCBA changes
Report what you observe; never modify the procedure yourself
Function drives everything
Attention, escape, tangible, automatic — read the consequence that reliably follows
Antecedent strategies prevent; DR procedures respond
NCR is time-based and noncontingent
DRA/DRI/FCT reinforce a behavior, DRO an absence, DRL a lower rate
One question separates the whole family
Extinction must match the function and stay consistent
Occasional success makes behavior more persistent
Bursts, variation, resurgence and emotional responding are expected
Document and report them; do not read them as failure
Punishment is implemented only under a written plan and training
Reinforcement-based procedures come first; time-out needs a reinforcing time-in
Crisis means safety first, trained procedures only, and immediate reporting
Never improvise a physical procedure
Implement what the plan covers; escalate what it does not
Safety, gaps, scope, deterioration and changed circumstances are the escalation triggers

Documentation and Reporting

13% of the exam
Study →
E.1 — Communicate concerns and suggestions from the intervention team to a supervisor in a timely manner
Caregivers, teachers, other service providers
E.2 — Seek and prioritize clinical direction from a supervisor in a timely manner
Training needs, data irregularities, following the chain of command
E.3 — Report/document variables that might affect client progress in a timely manner
Illness, medication, schedule changes
E.4 — Communicate objectively what occurred during the session
In accordance with applicable legal, regulatory and workplace requirements
Pass concerns on promptly and unfiltered
In the words the person used, whether or not you agree
Do not agree to a plan change on the spot
Route the request to your supervisor
Clinical questions from caregivers go to the supervisor
Acknowledge warmly, then route — and tell the supervisor it was asked
Suggestions count too
The task names "concerns and suggestions" — good ideas from the team are data for the supervisor
Immediately
Safety, harm, suspected abuse or neglect, crisis events
Before your next session with that client
Untrained procedures, unclear programs, missing materials, anything that blocks correct implementation
At the next supervision contact
Data trends, minor questions, team suggestions
Follow the chain of command
Assigned supervisor first; use the organization's path if unavailable or if the concern is about them
Document what you raised and when
The record of a report is part of the report
Never fill a gap by improvising
If you have not been trained on it, you do not run it
Report anything that could plausibly explain a change in the data
Illness, sleep, medication, schedule change, staffing change, major home events
Write observations and attributed reports
"Caregiver reported…" rather than a conclusion about the client's condition
Timeliness matters for interpretation
A late report cannot be aligned with the point where the data shifted
Never advise on medication
Documenting a reported change is in scope; commenting on it clinically is not
Objective = what a camera would capture
Observable, measurable, dated, and free of inference
Ban the mind-reading words
Refused, unmotivated, defiant, manipulative, attention-seeking, happy, had a good day
Write during or immediately after the session
Memory-based notes are inaccurate and hard to defend
Record session times accurately
Times usually drive billing; never document a session or service that did not occur
Correct errors, never erase them
Use your employer's correction procedure; never backdate
Follow your employer's format and requirements
The task explicitly includes legal, regulatory and workplace requirements
Store and transport records as your employer directs
Never in personal accounts, personal devices, or unsecured places
Share only what is necessary, through approved channels
Minimum necessary information, and only with those who need it
No client discussion in public or semi-public spaces
Hallways, lifts, waiting rooms, car parks, staff kitchens
Never post client photos, video or written information online
This is an absolute prohibition, not a consent question
Identifying information is broader than a name
Schools, addresses, diagnoses and recognisable images all identify
Report a breach promptly
A disclosed error can be contained; a hidden one cannot
Know your jurisdiction's and employer's rules before you need them
Who is mandated, where reports go, and by when varies by location
Report suspicion, not proof
Certainty is not required and investigating is not your role
Do not investigate or interview
No questioning the client for detail, no confronting a family
Telling your supervisor may not discharge a personal duty
Follow your employer's procedure and make any report you are required to make
Document observations objectively and separately from conclusions
What you saw and heard, and when
Keep it confidential
Not a topic for colleagues, other families, or anyone outside the process
Domain E weight
10 of the 75 scored questions, 13% of the exam, 4 tasks (E.1–E.4)
Three of the four tasks say "in a timely manner"
Timeliness is scored, not implied
Objective = observable, measurable, camera-verifiable
No inference, no diagnosis, no emotional characterization
Urgency triage
Safety immediately; implementation blockers before the next session; the rest at supervision
Route, do not answer
Clinical questions from caregivers and teachers go to the supervisor
Session times drive billing
Accurate times, no note for a session that did not happen, no backdating
Correct errors by your employer's procedure
Never erase, overwrite or rewrite the record
Mandated reporting is state- and employer-specific
The constant is: report suspicion, do not investigate, document objectively
Step 1 — lead with the data
Numbers, and the comparison that makes them meaningful
Step 2 — describe behavior, not intent
"Pushed materials off the table," not "refused" or "non-compliant"
Step 3 — attribute reported information
Who said it, what they said, and when the event started
Step 4 — do not connect cause to effect
Record both facts; the interpretation belongs to your supervisor
Step 5 — record the notification
That you reported it, and when
Looks objective, is actually an inference
"Refused," "non-compliant," "unmotivated" and "tried hard" all describe intent, not behavior
Looks like a helpful observation, is actually a diagnosis
"Seemed anxious," "is probably sick," "the medication is affecting her"
Looks like initiative, is actually a plan change
Recommending reduced demands in a note is a clinical recommendation
Looks like a small detail, is actually a reportable variable
Illness, sleep, medication and schedule changes are named in E.3
Looks like being responsive, is actually answering out of scope
Telling a caregiver what the data mean or what to expect
Looks efficient, is actually a documentation failure
Writing several sessions' notes at the end of the week
Looks like a favor, is actually falsification
Adjusting a session time, or documenting a session that did not happen
Looks like discretion, is actually a failure to report
Deciding a caregiver's concern is not worth passing on
Looks like a private channel, is actually a confidentiality breach
Client details in a staff group chat, or on a personal device
Looks like due diligence, is actually investigating
Questioning a client to find out more about a suspected abuse situation
Write what a camera would have captured
Observable, measurable, dated; no inference and no labels
Timely means graded by urgency
Safety now, implementation blockers before the next session, the rest at supervision
Carry the team's concerns up, unfiltered
And route clinical questions to your supervisor rather than answering them
Report the context that explains the data
Illness, sleep, medication, schedule and staffing changes
Record facts, not causal connections
Two facts side by side is documentation; linking them is interpretation
Accuracy in times and services is non-negotiable
Notes are billing and legal records
Protect client information everywhere
Storage, transport, messaging, conversation, and never online
On suspected abuse: report, do not investigate
Follow your jurisdiction's and employer's rules, which you should know in advance

Ethics

15% of the exam
Study →
F.1 — Identify and apply the core principles underlying the BACB's ethics code for RBT certificants
Benefit others; compassion, dignity and respect; integrity; competence
F.2 — Provide behavioral technician services only after demonstrating competence
Trained first, then service
F.3 — Provide services only under ongoing supervision from supervisors who meet BACB requirements
The supervision requirement is a condition of practicing
F.4 — Identify effective supervision practices
Training with instructions, modeling, rehearsal and feedback; observation of your service delivery
F.5 — Identify and comply with requirements for confidential information
Collecting, using, storing, protecting and disclosing
F.6 — Identify and comply with requirements for public statements
Social media activity; misrepresenting credentials, behavior analysis or outcomes
F.7 — Identify types of and risks associated with multiple relationships
And how to mitigate those risks when unavoidable
F.8 — Adhere to the gift giving and receiving guidelines in the ethics code
A specific monetary rule you must know
F.9 — Identify and apply interpersonal and professional skills
Accepting feedback, listening actively, seeking input, collaborating
F.10 — Engage in ongoing cultural humility and responsiveness
Including identifying personal biases — NEW in the 3rd edition
RBT Ethics Code (2.0)
Replaced the 2018 code; effective January 1, 2022
29 standards in 3 sections
General Responsibilities; Responsibilities in Providing Behavior-Technician Services; Responsibilities to the BACB and BACB-Required Supervisor
Core principle 1 — benefit others
The client's interest sets the direction
Core principle 2 — treat others with compassion, dignity and respect
In person, in front of others, and in the record
Core principle 3 — behave with integrity
Records match reality; behavior does not change when unobserved
Core principle 4 — ensure their own competence
Know your limits and say so
Trained and demonstrated, not just informed
Reading a plan is not training on it
Competence is procedure- and often client-specific
Experience with a similar program elsewhere does not transfer automatically
A revised plan is a new procedure
Training on the old version does not carry over
Untrained request → say so and seek direction
Never attempt an unfamiliar procedure to keep the session running
Ask for a refresher when your implementation has drifted
Competence is maintained, not banked
5% minimum
Of the hours you spend providing behavior-analytic services each calendar month
The denominator is service hours, the period is the calendar month
Not all hours worked, and not a rolling 30 days
At each organization
Meet the requirement separately wherever you provide services
At least 2 face-to-face, real-time contacts per month
Phone and email do not count
At least 1 contact includes observation of you providing services
In person is preferred; web camera or video-conference is permitted
Video monitoring alone may not count
Without real-time interaction or feedback
At least 1 of the 2 must be individual
No other RBTs or trainees present
Small group = 2–10 RBTs, hard cap 10
Regardless of the number of supervisors present
Supervision and professional development are separate
One event cannot satisfy both; supervision must be client focused
Retain supervision documentation at least 7 years
Both you and your supervisor, even after the relationship ends
Effective training = instructions, modeling, rehearsal, feedback
The four components named in the outline
Plus observation of your service delivery
Supervision without observation evaluates your description, not your work
Bring data, questions and honest accounts
Describe what happened, not what should have happened
Try the feedback
Implement it as given before concluding it does not work
Missing supervision is a compliance problem
Raise it with your supervisor and then through your organization's path
Follow BACB, employer and legal requirements together
All three apply; the strictest governs
Discuss confidential client information under your supervisor's direction
Unless disclosure is allowed by law for a valid reason, such as protecting someone from harm
Share only necessary information in job-related communications
Emails and documentation included
Anecdotes identify people
Stories told to friends, family or in trainings are disclosures
Other professionals are not automatically entitled to information
A teacher's question about diagnosis or history routes to your supervisor
Ask before disclosing, not after
An unauthorized disclosure cannot be undone
Never share identifying client information on social media or websites
Photos, videos or written information — an absolute prohibition
Consent is not the issue
The standard contains no "unless the family agreed" exception
Social media is defined broadly
Any digital platform for consuming, creating, sharing or commenting — private groups and disappearing stories included
De-identify properly or not at all
A blurred face with a named school still identifies
Do not misrepresent your credentials
You are a behavior technician, not a therapist, analyst or specialist
Do not promise or imply outcomes
No guarantees, no predicted results, no client testimonials
Multiple relationship
Mixing professional and personal roles with a client, stakeholder, supervisor, coworker, or someone close to the client
If one develops
Immediately inform your supervisor, work to resolve it, and document the actions
If it involves your supervisor
Report to their manager or another appropriate entity (HR, BACB) and document it
Never the employer of your supervisor
Fieldwork trainees may separately contract for supervision toward a future BCBA/BCaBA application
Current clients, stakeholders, supervisors
No romantic or sexual relationships, full stop
Former clients and stakeholders — minimum 2 years
From the date the professional relationship ended
Former supervisors — no fixed period
Gated on documenting that all professional duties are complete
Receiving supervision from a past partner — at least 6 months
After that relationship ended
The threshold
More than $10 US dollars (or equivalent purchasing power) is prohibited — $10 itself is not over the limit
Both directions
You may neither give nor accept above the limit
Covers clients, stakeholders AND supervisors
Not just client families
Occasional gratitude only
A gift must not result in financial benefit to the recipient
Cumulative gifts can violate the standard
Even small ones, if they become a regularly expected source of income or value
A stricter employer policy wins
Including a total prohibition on gift exchange
Decline warmly and tell your supervisor
The offer itself is worth reporting
Self-report to the BACB within 30 days
Of the event, or of becoming aware of it
Reportable categories
Legal charges and related actions; investigations naming you; disciplinary actions including terminations for cause; conditions that may impair safe service delivery
Certification status change — within 24 hours
Report to your supervisor, then comply with BACB requirements on practice, billing and title use
Check your BACB account regularly
The code gives monthly as an example, not as a fixed requirement
Accepting feedback
Thank, clarify, implement — do not defend first
Listening actively
Let people finish; reflect back and check your understanding
Seeking input
Ask caregivers and teachers what they are observing
Collaborating
Partner with other professionals; never argue in front of the client
Disagreement has a script
Run the plan as written today, raise the concern with your supervisor after
Cultural humility is ongoing
A practice of assuming incomplete understanding, not a course you complete
The code makes it a supervision matter
Work directly with your supervisor to ensure your work is culturally responsive
Actively evaluate your own biases
Notice which families and clients you react to differently, and raise it in supervision
Family practice conflicts are clinical information
A plan a family cannot live with will not be implemented — report it
Ask rather than assume
How a family does things is not a deviation from a neutral default
Use proper interpretation services
Not a family member, and never a child
New in the 3rd edition
F.10 has no 2nd-edition counterpart — prep material built for the old task list omits it entirely
Never modify a plan or program
Not even slightly, and not for a good reason
Never design or select assessments
You administer what you are trained and directed to administer
Never interpret results or declare a client's behavioral function
You collect the data the determination is made from
Never diagnose or give a prognosis
Including informal predictions about a client's future
Never advise on goals, services or placement
Route the question to your supervisor
Never practice without required supervision or beyond your training
Both are conditions of providing services at all
In scope means do it, not ask about it
Implementing the plan you are trained on needs no permission
Domain F weight
11 of the 75 scored questions, 15% of the exam, 10 tasks (F.1–F.10) — there is no F.11
RBT Ethics Code (2.0)
Effective January 1, 2022; 29 standards across 3 sections
Four core principles
Benefit others; compassion, dignity and respect; integrity; competence
Supervision: 5% minimum
Of hours providing behavior-analytic services, each calendar month, at each organization
Supervision: 2 face-to-face real-time contacts per month
Not phone, not email
Supervision: at least 1 observation and at least 1 individual contact
Observation may be by web camera or video-conference
Small group: 2–10 RBTs
Cap of 10 regardless of the number of supervisors
Supervision records: keep at least 7 years
Both RBT and supervisor, even after the relationship ends
Gifts: more than $10 US is prohibited
Both directions; clients, stakeholders and supervisors; stricter employer policy wins
Relationships: 2 years / documented completion / 6 months
Former clients and stakeholders / former supervisors / receiving supervision from a past partner
Self-report to the BACB: within 30 days
Of the event or of becoming aware of it
Certification status change: within 24 hours
Report it to your supervisor
Step 1 — name the ethical issue
Personal connection with a caregiver = multiple relationship
Step 2 — rapport is not a justification
Professional rapport lives inside the service relationship
Step 3 — the relationship is the risk, not just the posts
A private commitment not to post does not remove it
Step 4 — boundary plus disclosure beats silent avoidance
Decline courteously, then tell your supervisor
Step 5 — this is the general shape for gifts too
Decline warmly, explain the guideline, and report that it was offered
Looks like 5% of hours worked, is actually 5% of service hours
And per calendar month, not a rolling 30 days
Looks like a valid supervision contact, is actually not
Phone and email do not count; video monitoring without real-time interaction may not count
Looks like one group meeting is enough
At least one of the two monthly contacts must be individual
Looks like the 2-year rule, is actually the 6-month rule
2 years applies to former clients and stakeholders; 6 months applies to receiving supervision from a past partner
Looks like a $10 gift is banned
The prohibition is on gifts of MORE than $10; a $10 gift is not over the limit
Looks like small gifts are always fine
Cumulative or expected gifts can violate the standard even under the threshold
Looks acceptable with parental consent, is actually prohibited
No identifying client information on social media or websites — the standard has no consent exception
Looks like 30 days, is actually 24 hours
30 days is self-reporting to the BACB; 24 hours is telling your supervisor about a certification status change
Looks like helpfulness, is actually practicing out of scope
Answering a caregiver's clinical question, predicting outcomes, or tweaking a plan
Looks like professionalism, is actually defensiveness
Explaining why the feedback does not apply instead of implementing it
Looks like respecting a family, is actually avoiding a report
A cultural conflict with a plan is information your supervisor needs
Looks like the safe answer, is actually unnecessary
Ethics questions have real absolutes — "never" and "always" options are sometimes correct here
The RBT Ethics Code (2.0) governs
29 standards, 3 sections, effective January 1, 2022
Four core principles decide the unwritten cases
Benefit others; compassion, dignity and respect; integrity; competence
Train first, then serve
Competence is procedure-specific and must be maintained
Know the supervision numbers cold
5% of service hours per calendar month, 2 contacts, 1 observation, 1 individual, groups of 2–10, records 7 years
Gifts over $10 are out, both directions
And a stricter employer policy always wins
No identifying client information online, ever
Consent does not create an exception
Multiple relationships: avoid, disclose, document
2 years for former clients and stakeholders; 6 months before receiving supervision from a past partner
Cultural humility is ongoing and belongs in supervision
F.10 is new in the 3rd edition and is missing from older prep material
Implement, collect, report — never design, interpret, diagnose or modify
That single line settles most scope questions on the exam

Now prove you know them

Reading an outline is not the same as recalling it under exam pressure. Drill the free RBT (Registered Behavior Technician) questions to find the areas you keep missing, then sit a full timed mock.

Free practice questions →Timed mock exam →

Study aid, not a substitute for the official material — always confirm the current rule with BACB — RBT certification (official).

Report