RBT cheat sheet

Continuous measurement methods: frequency (count), rate (count ÷ time), duration (total duration or duration per occurrence), latency (stimulus onset to response onset), and interresponse time (time between consecutive responses).

Rate formula: count ÷ observation time; use rate instead of frequency whenever session or observation lengths differ across sessions.

Percentage of opportunities formula: (occurrences ÷ total opportunities) × 100; trials-to-criterion counts the number of trials needed to reach a preset performance level.

Total count IOA formula: (smaller count ÷ larger count) × 100; trial-by-trial IOA formula: (agreements ÷ total trials) × 100; interval-by-interval IOA: (agreements ÷ intervals) × 100.

Acceptable IOA is generally 80% or higher, with 90% or higher preferred for discrete, easily counted behaviors; IOA is calculated by a second observer recording independently.

Partial interval recording overestimates behavior, whole interval recording underestimates behavior, and momentary time sampling records only whether behavior is occurring at the end of each interval.

Permanent product measurement records the tangible environmental effect of a behavior after it occurs, such as completed worksheets, folded laundry, or a clean table.

Graph conventions: x-axis shows time or sessions, y-axis shows the behavior measure, data points are connected within a condition but not across phase change lines.

Preference assessment methods in order of increasing structure: free operant, single stimulus, paired stimulus (forced choice), multiple stimulus with replacement (MSW), multiple stimulus without replacement (MSWO).

Paired stimulus and MSWO assessments produce a ranked hierarchy of preferred items; MSW does not produce a rank order because each chosen item is returned to the array.

Indirect assessments (interviews, FAST, MAS, QABF) rely on caregiver or client report and do not involve direct observation of the behavior.

Descriptive assessment (ABC recording, scatterplot) documents behavior as it naturally occurs, suggests possible functions, but does not demonstrate a causal or functional relation.

Functional analysis is an experimental assessment that manipulates antecedents and consequences to identify the function of behavior; it is the only standard method that demonstrates function.

Standard functional analysis conditions: alone (automatic), attention (social positive), demand/escape (social negative), tangible, and play/control for comparison.

Behavioral definitions must be observable, measurable, objective, clear, and complete, including examples and non-examples of the target response.

Baseline data are collected before intervention begins so that later performance can be compared and treatment effects evaluated.

Discrete trial teaching sequence: antecedent (SD) → learner response → consequence → brief intertrial interval, with massed trials in a structured setting.

Prompt types: verbal, gestural, model, visual, positional, and physical (full or partial); physical prompts are the most intrusive.

Prompt hierarchies: most-to-least prompting supports errorless learning of new skills; least-to-most prompting allows independent responding for skills already in the repertoire.

Prompt fading and time delay: insert a progressively longer delay between the SD and the prompt so stimulus control transfers from the prompt to the SD.

Chaining methods: forward chaining teaches step 1 first, backward chaining teaches the last step first, and total task chaining teaches the whole chain each session.

Schedule effects: fixed ratio produces a high rate with a post-reinforcement pause, variable ratio produces high steady rates resistant to extinction, fixed interval produces a scalloped pattern, variable interval produces a steady moderate rate.

Discrimination training: reinforce the response in the presence of the SD and withhold reinforcement in the presence of the S-delta; stimulus control is shown when the response occurs mainly to the SD.

Generalization tactics: train across settings, people, and stimuli, use natural reinforcers, program common stimuli, and teach loosely rather than under one narrow set of conditions.

Four common functions: attention (social positive), escape or avoidance (social negative), access to tangibles (social positive), and automatic or sensory (nonsocial).

Function-based treatment matches the intervention to the identified function; a replacement behavior must produce the same reinforcer as the problem behavior.

Differential reinforcement variants: DRA reinforces an alternative behavior, DRI reinforces an incompatible behavior, DRO reinforces the absence of the problem behavior, DRL reinforces lower rates, DRH reinforces higher rates.

DRO intervals deliver reinforcement when the problem behavior does not occur during the whole interval; interval length is set from baseline data on the behavior.

Extinction variants: attention extinction withholds attention, escape extinction continues the demand, tangible extinction withholds the item; an extinction burst is expected at the start.

Antecedent interventions include noncontingent reinforcement, functional communication training, high-probability request sequences, priming, offering choices, and environmental modification.

Response cost removes a specified amount of reinforcement contingent on behavior, and time-out removes access to reinforcement; both are punishment-based and require plan authorization.

Reinforcement increases behavior and punishment decreases behavior; the label depends on the observed effect on the response, not on whether the consequence is pleasant or aversive.

SOAP note format: Subjective, Objective, Assessment, Plan; the objective section contains only directly observed, measurable facts.

Session notes include date, time, duration, target behaviors with data, interventions implemented, learner response, and the recorder's signature and credentials.

Data are recorded at the time the behavior occurs or as soon as possible afterward; reconstructed or estimated data are not acceptable.

Correct documentation errors with a single line through the error, the correction, initials, and date; never erase, scribble out, or use correction fluid.

Line graph: x-axis is time or sessions, y-axis is the behavior measure, and vertical phase change lines separate conditions.

Report to the supervisor immediately: injuries, suspected abuse or neglect, medication changes, illness, and sudden or significant changes in behavior.

Confidentiality: store data in secured locations, use client identifiers rather than full names where possible, and do not discuss clients in public spaces or on social media.

RBTs practice under the supervision of a qualified BCBA, implement programs designed by the supervisor, and do not independently design or modify assessments or interventions.

Supervision standard: at least 5% of hours worked per month, at least two face-to-face supervision meetings per month, at least one of which is individual.

Mandated reporting: suspected abuse, neglect, or harm is reported to the supervisor and to the appropriate authorities as required by law.

Informed consent is obtained before assessment, intervention, or changes in services, from the client or legally authorized representative.

Avoid multiple or dual relationships with clients and families; do not accept gifts of more than trivial value or enter into business or social relationships.

RBTs may not falsify data, claim competencies they do not hold, or perform procedures outside their training and scope of practice.

Ethical decision steps: identify the issue, consult the Ethics Code and supervisor, consider all stakeholders, act, and document the resolution.

Confidentiality limits: information is shared only with the treatment team or when legally required, disclosing the minimum necessary information.