Choose the measure, then keep its units
A count answers how many events occurred. Rate divides that count by observation time. Duration measures an episode from its defined onset to offset. Latency begins at a specified instruction or stimulus and ends at response onset. IRT needs explicit response-to-response anchors; do not assume the same anchors in every program.
Do not turn intervals into event counts
Partial interval records occurrence at any point in the interval. Whole interval requires occurrence throughout it. Momentary time sampling observes a scheduled instant. A positive interval can contain several responses, so eight positive intervals does not mean eight responses. Interval percentages also do not reconstruct exact time spent behaving.
Keep graphs and records interpretable
Use the units named on the data sheet and graph. A cancelled session is not an observed zero. A permanent product must be attributable to the response being measured. A trend describes a pattern; it does not prove treatment caused it. Report missing observations and fidelity concerns instead of filling gaps with invented values.
A mistake worth catching
Dividing minutes by responses produces time per response, not responses per minute.
The matching quiz uses 13 questions from our shared mixed bank. Each answer includes the reasoning for the alternatives.
Practice this topicScope and references
Measurement concepts and the explicitly stated definitions in each question. These examples are educational, not client-specific clinical instructions. An RBT implements assigned procedures with appropriate supervision.
BACB RBT Test Content Outline (3rd ed.)
RBT Ethics Code (2.0)
Editorial revision: October 20, 2026.