Data Collection and Graphing

17% of the RBT exam · blueprint 3rd edition (2026)

Short answer

Domain A is 13 of the 75 scored RBT questions (17%). It tests whether you can implement continuous measurement (frequency, duration, latency, IRT), discontinuous measurement (partial interval, whole interval, momentary time sampling), permanent product recording, enter data and update graphs, describe behavior in observable terms, summarize data as rate or percentage, read trends, and explain why unreliable data is dangerous.

Why it matters on this exam

Every other domain depends on data you collected correctly. The exam does not ask you to design a measurement system; it asks whether you can implement the one in the plan and recognise when it is being done wrong.

Key concepts (tasks A.1–A.8)

Continuous measurement records every instance: frequency/rate (count per unit time), duration (how long), latency (time from SD to response), interresponse time (time between responses).
Discontinuous measurement samples time: partial interval (did it occur at any point in the interval; overestimates), whole interval (occurred for the entire interval; underestimates), momentary time sampling (occurring at the moment the interval ends).
Permanent product: measure the outcome after the fact (worksheets completed, items put away).
Observable and measurable terms: "hits peer with open hand" not "gets aggressive".
Summarising: rate = count ÷ time; percentage of opportunities = correct ÷ opportunities × 100; mean duration = total duration ÷ occurrences.
Trends: increasing, decreasing, or zero-celerating; variability; level.
Procedural fidelity: if the procedure is not run as written, the data cannot tell you whether the program works.

Common mistakes

Choosing partial interval when the plan calls for whole interval because it is "easier".
Reporting frequency without the observation time (12 in 10 minutes is not the same as 12 in an hour).
Graphing on the wrong phase or forgetting the phase-change line after a program change.

Worked example

A session lasts 30 minutes. Elopement is recorded 6 times. Rate = 6 ÷ 30 = 0.2 per minute, or 12 per hour. If the same 6 occurrences were recorded with 10-second partial-interval recording over the 30 minutes, the data would read as a percentage of intervals with elopement, not a count; comparing it with last week's frequency data would be invalid.

Original practice questions

Data Collection and GraphingDifficulty 2/5

During a 20-minute session an RBT records that a learner asks for help 8 times. Which measure has the RBT collected, and what is the rate?

AFrequency; 0.4 responses per minute
BLatency; 2.5 minutes
CPartial-interval; 40% of intervals
DDuration; 8 minutes

Show answer and rationale

Correct: A. Counting every instance is frequency (a continuous measure). Rate converts the count to a standard time unit: 8 ÷ 20 minutes = 0.4 per minute, which lets sessions of different lengths be compared.

D — Duration measures how long each response lasts, not how many occurred.
B — Latency is the time from the instruction to the start of the response.
C — No intervals were used; the RBT counted each instance.

Data Collection and GraphingDifficulty 3/5

A plan specifies 10-second whole-interval recording for on-task behavior. In one interval the learner is on task for 7 seconds, looks away for 2 seconds, then returns. How should the interval be scored?

AOccurred, because on-task behavior was present for most of the interval
BDid not occur, because the behavior was not present for the entire interval
COccurred, because the learner was on task at the end of the interval
DHalf occurred, recorded as 0.5

Show answer and rationale

Correct: B. Whole-interval recording scores an interval only if the behavior occurred throughout the entire interval. A 2-second break means it did not, so the interval is scored as not occurring. This is why whole-interval tends to underestimate behavior.

A — Majority of the interval is not the whole-interval rule.
C — Scoring at the end of the interval is momentary time sampling.
D — Interval recording is binary; there is no partial credit.

Data Collection and GraphingDifficulty 2/5

Which of the following is written in observable and measurable terms?

AStudent becomes frustrated during math
BStudent is defiant toward staff
CStudent pushes the worksheet off the desk with one or both hands
DStudent does not want to participate

Show answer and rationale

Correct: C. An operational definition describes what can be seen and counted by any observer. Pushing the worksheet off the desk is observable; frustration, defiance and wanting are inferences about private states.

A — 'Frustrated' is an inferred emotional state, not an observable action.
B — 'Defiant' is a label that different observers would score differently.
D — 'Does not want to' describes a motive, not a behavior.

All 4 Data Collection and Graphing questions →

4 flashcards for Data Collection and Graphing

Practice this topic

Blueprint-mapped questions with full rationales for every choice.

Start practice

Primary references

RBT Test Content Outline (3rd ed.), effective January 1, 2026 — Behavior Analyst Certification Board (BACB) · primary · retrieved 2026-09-21, verified 2026-09-21

Last reviewed: 2026-09-21Next review by: 2026-11-20Draft — pending editorial reviewHow we verify facts