RBT (Registered Behavior Technician) — All Questions

Across a 40-minute session, an RBT counts 24 requests made by Jonah. The team graphs requests per minute. What should the RBT plot?

• a.1.67 per minute, because 40 minutes divided by 24 requests is 1.67
• b.0.6 per minute, because 24 requests divided by 40 minutes is 0.60✓
• c.24 per minute, because the raw count is entered on the graph directly
• d.36 per hour, because 24 requests in 40 minutes scales to 36 per hour

(b) is correct: rate is count divided by observation time, and 24 ÷ 40 = 0.60 requests per minute. (a) inverts the calculation and yields minutes per request rather than requests per minute. (c) plots the raw count, which ignores the length of the session and would not be comparable with sessions of other lengths. (d) is arithmetically right — 0.6 per minute is indeed 36 per hour — but the graph is scaled in requests per minute, so plotting the hourly figure puts the point far off the correct value.

An RBT records latencies of 8, 12, 4, and 16 seconds between the instruction and Hugo's first step. The BCBA asks for the mean latency. What should the RBT report?

• a.40 seconds, because that is the total latency across the four trials
• b.16 seconds, because the mean is taken from the longest latency timed
• c.4 seconds, because the shortest latency represents typical responding
• d.10 seconds, because the 40 total seconds are divided by 4 trials✓

(d) is correct: 8 + 12 + 4 + 16 = 40 seconds, and 40 ÷ 4 trials = 10 seconds per trial. (a) reports the sum, which is a true total but is not a mean and would grow simply by running more trials. (b) uses the slowest trial, which overstates typical responding. (c) uses the fastest trial, which understates it in the same way; the mean sits between these two extremes.

Using a 10-step task analysis for handwashing, Ivy completed 7 steps independently. The graph shows the percentage of steps completed independently. What should the RBT plot?

• a.70%, because 7 independent steps divided by 10 steps is 0.70✓
• b.7%, because Ivy completed 7 of the steps without any prompting
• c.30%, because 3 of the steps still required a prompt from the RBT
• d.1.4, because 10 steps divided by 7 independent steps gives 1.43

(a) is correct: 7 ÷ 10 = 0.70, so 70% of the steps were independent. (b) reports the raw count as if it were a percentage, which happens to look close but is off by a factor of ten. (c) computes the percentage of prompted steps — 3 ÷ 10 = 0.30 — which is the mirror image of what the graph tracks. (d) inverts the fraction and produces a ratio above 1, which cannot be a percentage of steps.

An RBT records 20 protests in a 50-minute session on Tuesday and 18 protests in a 30-minute session on Wednesday. The BCBA asks which day had the higher rate of protesting. What should the RBT report?

• a.Tuesday, at 0.4 per minute, because 20 protests is the larger count
• b.Tuesday, at 2.5 per minute, because 50 minutes divided by 20 is 2.5
• c.Wednesday, at 0.6 per minute, because 18 divided by 30 is 0.60✓
• d.Wednesday, at 1.67 per minute, because 30 divided by 18 is 1.67

(c) is correct: Tuesday's rate is 20 ÷ 50 = 0.4 per minute and Wednesday's is 18 ÷ 30 = 0.6 per minute, so Wednesday is higher despite the smaller raw count. (a) computes Tuesday's rate correctly at 0.4 but then draws the wrong conclusion by leaning on the raw count. (b) inverts the arithmetic, producing minutes per protest for Tuesday. (d) reaches the right day through inverted arithmetic, since 30 ÷ 18 gives minutes per protest rather than protests per minute.

Marcus engaged in pacing for a total of 12 minutes during a 60-minute session. The BCBA wants the result expressed as a percentage of session time. What should the RBT report?

• a.12%, because 12 minutes of pacing occurred during the whole session
• b.5%, because 60 minutes divided by 12 minutes of pacing gives 5
• c.80%, because 48 of the 60 minutes did not involve any pacing at all
• d.20%, because 12 minutes divided by 60 minutes equals 0.20✓

(d) is correct: percentage of session time is behavior time divided by session time, and 12 ÷ 60 = 0.20, or 20%. (a) reports the raw minutes as a percentage, which would only be correct in a 100-minute session. (b) inverts the fraction; 60 ÷ 12 = 5 is a ratio, not a percentage. (c) computes the share of the session without pacing — 48 ÷ 60 = 0.80 — which is the complement of what was asked.

Why does a team convert counts to rate before comparing sessions?

• a.Rate removes the need to define the behavior in observable terms
• b.Rate accounts for differences in how long each observation lasted✓
• c.Rate converts a discontinuous measure into a continuous measure
• d.Rate is required whenever data are plotted on an equal-interval graph

(b) is correct: dividing the count by the observation time puts sessions of different lengths on the same footing, so a change on the graph reflects behavior rather than schedule. (a) is wrong because a clear definition is required no matter which summary measure is used. (c) is wrong because rate is computed from a continuous count; it cannot turn interval data into continuous data. (d) is plausible because rate is often graphed, but graphs also display counts, durations, and percentages.

Over ten sessions, Aria's independent responses were 2, 3, 3, 5, 6, 6, 8, 9, 9, and 11. The BCBA asks the RBT to describe the trend. What is the accurate description?

• a.An increasing trend, since the data path rises across the sessions✓
• b.A decreasing trend, since the values fall away from the starting point
• c.A flat trend, since several sessions repeat the same value in a row
• d.A variable pattern, since the values change from session to session

(a) is correct: the values climb from 2 to 11 with no reversal, so the overall direction of the data path is upward. (b) states the opposite direction. (c) is plausible because pairs of sessions repeat, but a few repeated values inside a rising series do not flatten the overall path. (d) is plausible because the numbers do change session to session, but variability describes bouncing without direction, and here the changes move consistently one way.

Baseline sessions for Rowan averaged about 12 tantrums, and the first five treatment sessions average about 4. The data path within each phase is fairly flat. Which term best describes the change?

• a.A change in trend, since the direction of the path shifted downward
• b.A change in variability, since the points cluster differently now
• c.A change in level, since the average value dropped between phases✓
• d.A change in latency, since the behavior takes longer to appear now

(c) is correct: level refers to where the data sit on the vertical axis, and a drop from about 12 to about 4 with flat paths in both phases is a level change. (a) is plausible because the numbers went down, but trend describes the slope within a phase, and both phases here are flat. (b) is plausible because phases can differ in spread, but the stem describes both paths as fairly stable. (d) is wrong because no time-to-onset measure was collected.

Dez's data for the last eight sessions read 2, 14, 3, 15, 4, 13, 2, and 16. The path shows no clear rise or fall. Which description fits the data best?

• a.A steep increasing trend with a high level across the sessions
• b.High variability, with the values bouncing between low and high✓
• c.A stable path at a low level with one unusual high data point
• d.A decreasing trend that is masked by the range of the values

(b) is correct: the values swing repeatedly between roughly 2 and 16 with no consistent direction, which is what variability looks like on a graph. (a) is wrong because the series does not climb; high values alternate with low ones. (c) is plausible because some points are low, but there are four high points, not one outlier. (d) is plausible because variability can hide a trend, but nothing here declines — the low points stay low and the high points stay high.

Kai's self-injury is targeted for reduction. Over the treatment phase the data read 9, 8, 6, 5, 3, and 2. What should the RBT report about the trend?

• a.The trend is increasing, which means the procedure is not yet working
• b.The trend is flat, which means the sessions are producing no change
• c.The trend is variable, which means the data cannot yet be interpreted
• d.The trend is decreasing, which is the direction the plan is aiming for✓

(d) is correct: the values fall steadily from 9 to 2, and for a behavior targeted for reduction a downward path is movement in the therapeutic direction. (a) reverses the direction of the data. (b) is wrong because the path clearly declines rather than holding steady. (c) is plausible because variability does block interpretation, but each value here is at or below the one before it, which is the opposite of a bouncing path.

Elin's graph shows baseline points at 10, 11, and 10, then a vertical line, then treatment points at 9, 7, 5, and 4. Which reading of the graph is correct?

• a.Baseline was decreasing, and treatment continued that same decrease
• b.Baseline and treatment show the same level with different variability
• c.Baseline was stable, and treatment shows a decreasing trend after it✓
• d.The vertical line marks a missing session rather than a new condition

(c) is correct: the three baseline points hover around 10 with no direction, and the four treatment points fall steadily, so a stable baseline is followed by a decreasing trend. (a) misreads baseline; 10, 11, 10 has no downward slope. (b) is wrong because the treatment points end well below the baseline values, which is a level change as well as a trend. (d) misidentifies the vertical line, which marks a phase change and is why the data paths are not connected across it.

After two treatment sessions, a caregiver looks at Rosa's graph and asks the RBT whether the program is working. Both points sit below the baseline values. What is the best response?

• a.Say that two points are too few to show a trend and that the BCBA reviews progress✓
• b.Say that the program is working because both points fell below the baseline level
• c.Say that the trend is decreasing and predict how many sessions remain in the plan
• d.Say that progress questions must be referred to the BCBA before anything is answered

(a) is correct: describing what the graph does and does not yet show is within the RBT's role, and two data points cannot establish a trend, so the honest answer names that limit and points to the person who evaluates progress. (b) overinterprets two points, which may reflect normal variability rather than treatment effect. (c) compounds the same error and adds a prediction about the course of treatment, which is a clinical judgment the RBT does not make. (d) is unnecessary deflection: the RBT can accurately describe the data and the review process without interpreting outcomes, and refusing to say anything damages the working relationship with the family.

A BCBA tells the RBT that baseline will continue until the data path is stable. After five sessions the values read 4, 9, 3, 10, and 5. What does the graph show?

• a.A stable baseline, because the values stay within a narrow band
• b.An increasing trend, because the last value is above the first one
• c.A decreasing trend, because the highest value came in the middle
• d.An unstable baseline, because the values vary widely across sessions✓

(d) is correct: the values swing between 3 and 10 with no direction, so the path is variable and baseline has not yet met the stability criterion. (a) is wrong because a range of 3 to 10 is not a narrow band. (b) is plausible because 5 is above 4, but a trend is read from the whole path, not from the first and last points. (c) is wrong because the position of a single high value does not establish a downward slope.

Which three features are used to describe a graphed data path?

• a.Reliability, fidelity, and the agreement between two independent observers
• b.Level, trend, and variability of the data points across the sessions✓
• c.Frequency, duration, and latency of the target behavior being recorded
• d.Baseline, treatment, and maintenance phases shown along the x-axis

(b) is correct: level describes where the data sit, trend describes the direction of the path, and variability describes how much the points bounce — together they summarize what a graph shows. (a) names data-quality concepts, which matter but describe how data were collected rather than what the path looks like. (c) names measurement procedures, which determine what is plotted rather than how the path is described. (d) names conditions, which are labeled on the graph but are not features of the data path itself.

An RBT gets busy during a session and reconstructs the frequency data from memory during the drive home. The counts look about right. What is the main risk?

• a.The counts will be recorded as a rate rather than as a raw frequency
• b.The counts will belong to the caregiver rather than to the RBT's record
• c.The counts may be inaccurate, and a working plan may be changed on them✓
• d.The counts will need a second observer before they can be graphed at all

(c) is correct: unreliable data lead to unreliable decisions — a clinician looking at inflated or deflated numbers may drop an effective procedure or continue an ineffective one. (a) is wrong because how the number is summarized is unrelated to when it was written down. (b) is wrong because the RBT still authors the record; the problem is its accuracy, not its ownership. (d) is plausible because agreement checks improve confidence in data, but a second observer is not a precondition for graphing, and it would not repair data written from memory.

Two RBTs collect data on the same learner's 'disruption.' One counts each shout; the other counts each minute containing shouting. Their graphs disagree sharply. What is the underlying problem?

• a.The two RBTs are not using the same definition and measurement rule✓
• b.The two RBTs are working on different days, so the data cannot match
• c.The two RBTs need to graph the results on two separate line graphs
• d.The two RBTs should average their two totals before entering the data

(a) is correct: one is counting responses and the other is scoring intervals, so the two data sets are not measuring the same thing and cannot be compared or combined. (b) is plausible because behavior does vary day to day, but that would produce differences in size, not two incompatible measures. (c) is wrong because splitting the graphs preserves the inconsistency rather than resolving it. (d) is the worst option: averaging a count with an interval score produces a number that means nothing.

An RBT has been delivering a token after every third correct response, although the written plan specifies a token after every correct response. Progress has stalled and the BCBA is considering dropping the procedure. What is the risk created by the RBT's practice?

• a.The learner will earn tokens too quickly and lose interest in them
• b.The data will need to be collected by a second observer next session
• c.The graph will show variability that must be relabeled as a new phase
• d.A procedure that was never run as written may be judged ineffective✓

(d) is correct: when the procedure delivered differs from the procedure described, the flat data describe the RBT's version, not the plan, so an effective procedure can be abandoned for the wrong reason. (a) reverses the effect of the error — the learner is earning tokens less often than the plan specifies, not more. (b) is plausible because agreement checks are part of data quality, but the problem here is implementation, not observation. (c) is wrong because a phase change marks a planned condition change, not an unreported drift in how a procedure was run.

Why does poor procedural fidelity threaten the value of the data?

• a.Data collected with poor fidelity cannot be plotted on a line graph
• b.Decisions get made about a procedure that was not actually delivered✓
• c.Data collected with poor fidelity must be reported to the certifying body
• d.Decisions get delayed because a second observer must repeat the session

(b) is correct: the graph is only meaningful if it reflects the procedure as written; when implementation drifts, the team draws conclusions about something that never happened as described. (a) is wrong because such data can still be plotted, which is precisely the danger — they look identical to good data. (c) is wrong because fidelity problems are addressed through supervision and training within the workplace. (d) is plausible because fidelity issues can prompt retraining and re-measurement, but the core threat is invalid decisions rather than scheduling delay.