Sometimes you cannot record every instance of a behavior: it happens too often to count, has no clear start or end, or you are busy teaching. Discontinuous measurement solves this by recording a sample of the behavior instead. This lesson covers the three methods you will see most: partial-interval recording, whole-interval recording, and momentary time sampling.
Why it matters
Discontinuous methods are practical, but they only estimate behavior. Each one has a predictable bias, and your supervisor chooses the method knowing that bias. If you score intervals the wrong way, the graph tells a different story than what really happened, and treatment decisions follow the wrong story.
In practice questions, you will often name the method from how the RBT scores intervals, or say which method tends to over- or underestimate.
Key ideas
- The session is split into equal intervals. For example, a 10-minute observation might be divided into 60 ten-second intervals. A timer, app, or vibrating prompt signals each interval so you do not have to watch a clock.
- Each interval gets one mark. You score whether the behavior occurred (often a + or a check) or did not occur (a − or blank). You do not count how many times it happened.
- Results are a percentage of intervals. Divide the number of scored intervals by the total number of intervals and multiply by 100. The result is not a count, and it is not the exact time the behavior lasted.
- The scoring rule is what separates the three methods:
- Partial-interval: score the interval if the behavior happens at any point during it.
- Whole-interval: score the interval only if the behavior lasts for the entire interval.
- Momentary time sampling: score the interval only if the behavior is happening at the moment the interval ends.
- Shorter intervals give better estimates. In general, the shorter the interval, the closer the data come to what continuous measurement would show. Interval length is set in the plan, so follow it.
Definitions
- Discontinuous measurement: recording a sample of behavior during an observation rather than every instance. It produces an estimate.
- Partial-interval recording: the interval is scored if the behavior occurs at any time during it, even once and briefly. It tends to overestimate how much of the session the behavior took up, and it can underestimate how often a high-rate behavior occurred. It is often used for behaviors the team wants to decrease, because an overestimate keeps progress from looking better than it is.
- Whole-interval recording: the interval is scored only if the behavior continues throughout the whole interval. It tends to underestimate the behavior, so it is often used for behaviors the team wants to increase, such as staying engaged in an activity.
- Momentary time sampling (MTS): the interval is scored only if the behavior is occurring at the instant the interval ends. It can overestimate or underestimate, and it easily misses brief behaviors that happen between checks. It suits ongoing behaviors and lets you observe several learners or keep teaching between checks.
Examples from a session
Same learner, three scoring rules
Diego, a 10-year-old, is working at his desk in a resource room. His RBT, Hana, uses 15-second intervals. In one interval, Diego works for 12 seconds, then looks out the window for the last 3 seconds, and he is still looking out the window when the interval ends.
- Partial-interval (target: off-task looking): score it. The looking happened at some point during the interval.
- Whole-interval (target: on-task work): do not score it. Working did not last the entire 15 seconds.
- Momentary time sampling (target: on-task work): do not score it. At the moment the interval ended, Diego was looking out the window.
In a different session, Andre, an RBT at a clinic, runs a circle-time group with four preschoolers. He cannot watch every child while leading a song, so the plan calls for momentary time sampling of engagement: each time his watch vibrates, he glances at the next child on the sheet and records yes or no. In a home session, a teen's brief vocal outbursts last one or two seconds each, so the supervisor chose partial-interval recording to make sure even short outbursts are captured.
If you notice the method does not seem to fit what you are seeing, for example a target behavior that keeps happening between MTS checks and never shows up in the data, tell your supervisor. Do not switch methods on your own.
Common mistakes / exam traps
- Reporting a count. Interval data tell you how many intervals contained the behavior, not how many times it happened. Ten scored intervals could hold ten hits or forty.
- Turning intervals into time. Twelve scored 20-second intervals do not mean the behavior lasted 4 minutes. With partial-interval, a one-second behavior scores the whole interval.
- Watching the wrong moment. For MTS, you look at the instant the interval ends. Behavior earlier in the interval does not count, even if you happened to see it.
- Scoring "mostly" as "whole." If the learner was engaged for 9 of 10 seconds, the whole-interval rule says do not score it.
Exam trap
If the stem says the RBT recorded whether the behavior happened "at any time" during each interval, it is partial-interval. "Throughout the entire interval" is whole-interval. "At the end of each interval" or "at the moment the timer signals" is momentary time sampling.
Exam trap
Partial-interval vs. whole-interval bias: partial-interval scores any brief occurrence, so it tends to overestimate; whole-interval needs the full interval, so it tends to underestimate. MTS has no fixed direction. Scored on the same intervals, the true share of time falls between the whole-interval and partial-interval percentages.
Quick check
An RBT uses momentary time sampling with 30-second intervals to record humming. In one interval, the learner hums for about 20 seconds and stops 5 seconds before the timer signals. How should the RBT score that interval?