RBT common mistakes
RBT · Understand the traps and how to avoid them.
Computes interobserver agreement with the wrong formula for the method named in the question - e.g., dividing agreements by the number of intervals when the scenario describes total-count IOA, or reporting occurrence-only agreement as if it were trial-by-trial IOA.
Candidates memorize the phrase 'agreements divided by total times 100' without attaching it to a specific IOA method, and under time pressure they grab the first formula that comes to mind instead of the one the item specifies.
What to do instead
Before calculating, write down what is being compared: counts (smaller total divided by larger total), intervals (agreements divided by total intervals), or trials (agreements divided by total trials). Then multiply by 100 and check that the answer is between 0% and 100%.
Selects a measurement procedure that does not match what the question says the data are needed for - for example, choosing partial-interval recording when the item asks for the exact number of times a behavior occurred, or choosing frequency for a behavior with no clear onset and offset such as pacing.
Definitions of continuous and discontinuous measurement are memorized as vocabulary rather than as answers to the question 'what information does this method give you?' So any method whose name sounds like 'recording' gets chosen.
What to do instead
Read the stem and state the required output in one word first - count, duration, latency, or estimate. Then pick the method that produces that output: event recording for count, duration recording for time, latency for time to onset, whole/partial interval or momentary time sampling only when an estimate of occurrence across time is acceptable.
Mistakes one preference assessment for another - labeling a free-operant observation as a paired-stimulus or forced-choice assessment, or calling a multiple-stimulus-without-replacement probe a 'free operant' because the client 'freely chose' an item.
The names are learned as a list rather than as descriptions of the arrangement, and the phrase 'the client chose' appears in several procedures, which triggers the wrong label.
What to do instead
Identify two features in the stem before answering: how many items are available at a time (one pair, an array, or the whole environment) and whether access is contingent on a selection or just observed. Two items at a time means forced choice; an array presented repeatedly with selected items removed means MSWO; observation with no structured trials means free operant.
Confuses indirect and descriptive assessment with experimental functional analysis - for example, choosing 'functional analysis' for an item that describes caregiver interviews, record review, and scatterplot data, or choosing 'indirect assessment' for an item where a therapist systematically manipulates antecedents and consequences.
All three are filed under 'figuring out the function,' and candidates answer based on the words 'function' or 'assessment' rather than on whether the variables were observed, reported, or manipulated.
What to do instead
Scan the stem for the verb describing what was done. If someone was asked or records were read, it is indirect. If behavior was watched in the natural setting with no manipulation, it is descriptive. If conditions were arranged and compared, it is a functional analysis.
Reverses chaining procedures - describing forward chaining when the learner is taught the final step first while the trainer completes all earlier steps, or describing backward chaining as starting from the beginning of the task analysis.
Both procedures are taught as 'chaining' and the direction word is treated as a label rather than as a description of which step the learner performs independently.
What to do instead
Ask one question: which step does the learner do on his own in the first teaching session? First step independently with the rest prompted means forward chaining; last step independently with earlier steps done for him means backward chaining.
Mixes up prompt-fading strategies or prompt types - calling a least-to-most sequence 'most-to-least' because the prompt got stronger, confusing a gestural prompt with a model, or describing time delay as a fading procedure when the stem describes adding a prompt after the instruction with a gradually increasing wait.
Prompt terms are learned as a vocabulary list, and the directionality of 'least' and 'most' is easy to invert under pressure, especially when the stem describes the prompt changing across trials.
What to do instead
Draw the sequence on scratch paper as a ladder: write the prompt level used on trial one and on the last trial, then check whether it moved up or down. For time delay, confirm the stem mentions a wait interval before the prompt rather than a change in prompt strength.
Chooses the wrong differential reinforcement procedure - answering DRO when the stem reinforces a specified alternative behavior, DRI when the stem reinforces a lower rate of the same behavior, or DRA when the stem reinforces behavior that cannot occur at the same time as the target.
The four acronyms differ by one letter and are recalled as a set, so candidates match on the word 'differential' or on the general idea of 'reinforcing something else.'
What to do instead
Name exactly what is being reinforced in the stem. Reinforcing a specific alternative behavior is DRA; reinforcing a behavior physically incompatible with the target is DRI; reinforcing the absence of the target for an interval is DRO; reinforcing a reduced rate of the same target is DRL.
Recommends abandoning or altering an extinction procedure when the client shows a temporary increase in response intensity, or selects punishment-based procedures as the default first step because the behavior is dangerous or disruptive.
The extinction burst looks like treatment failure in a written scenario, and candidates are drawn to the answer that makes the discomfort stop rather than the answer that is technically correct.
What to do instead
Treat an increase in rate, intensity, or variety at the start of extinction as an expected and temporary effect: the correct action is to continue implementing the plan as written and report the change to the supervising behavior analyst. For any question asking for a first intervention, prefer reinforcement-based strategies and reserve punishment for items that state less intrusive options have failed.
Records data with subjective or interpretive language ('was frustrated,' 'had a bad day,' 'acted out') and reconstructs session data at the end of the shift from memory rather than recording at the time of occurrence.
Narrative charting habits from other fields carry over, and end-of-shift data entry feels efficient when sessions are busy, so the numbers get estimated after the fact.
What to do instead
Write only what a camera could capture: observable actions, quoted statements, and measurable values with the time and setting. Record each response as it occurs or immediately after the trial, and if data are missed, note the gap honestly rather than filling it in later.
Accepts a request outside the RBT role - revising a behavior intervention plan, changing reinforcement schedules, or answering a parent's clinical question with advice - or discusses a client by name in a public place or on social media.
The desire to be helpful and the closeness of in-home and school settings blur role boundaries, and confidentiality lapses usually happen in casual moments rather than deliberate ones.
What to do instead
When a decision requires clinical judgment, implement the plan as written and route the request to the supervising behavior analyst with a factual report of what was observed. Share client information only with those who need it for service delivery, and use initials or a client code in any message that leaves the file.