Partial Interval Recording in ABA: How to Score It Right
Picture a 10-minute group lesson where one learner keeps calling out, sometimes several times in a single minute. Logging each instance while you also teach is close to impossible, and timing each one is even harder.
Partial interval recording in Applied Behavior Analysis (ABA) handles this by asking one yes-or-no question per window of time: did the behavior show up at all?
The guidance below draws on peer-reviewed ABA measurement research. By the end, you'll know how to set it up, score it, read what the percentage does and doesn't tell you, and spot when a different method fits better.
What is partial interval recording in ABA?
Partial interval recording is a data collection method that marks whether a target behavior happened at any point during a fixed time interval, not how long it lasted or how many separate times it occurred.
To do this, you’ll split your observation period into equal chunks, like 10 or 15 seconds each. If the behavior shows up even once during that chunk, the whole interval gets scored "yes." If it never shows up, the interval gets scored "no."
This makes partial interval recording a form of discontinuous measurement, a category that also includes whole interval recording and momentary time sampling. All three record behavior in intervals rather than tracking it continuously, but picking the right interval type changes what the resulting data can tell you.
How does partial interval recording work?
Partial interval recording works by dividing a session into short, equal intervals and scoring each one based on whether the behavior occurred at all.
Here’s how to do it:
- Write an observable definition of the behavior: For example, a behavior like "calling out" should mean something specific, like speaking without raising a hand.
- Choose your interval length: Pick a length your observers can hold for the whole session.
- Start the timer and observe: Use a stopwatch, an interval recording app, or a printed data sheet.
- Mark each interval + or – : The moment the behavior appears, that interval counts as an occurrence. Mark it, then move on to the next one. Data sheets vary in how they record this: some use + and –, others use Y and N, or 1 and 0.
- Calculate the percentage of intervals: Divide the number of intervals with an occurrence by the total number of intervals, then multiply by 100.
Having multiple observers score the same session, then comparing the records, helps establish how reliable the approach is.
What is an example of partial interval recording?
The two examples below walk through how a Registered Behavior Technician (RBT) might score calling out during a group lesson and hand-flapping during a short session, one window at a time.
Example 1: Calling out during a classroom activity
An RBT is tracking a learner's calling-out behavior during a 10-minute group lesson, using 30-second intervals (20 intervals total).
That's 14 out of 20 intervals scored "yes," or 70% of intervals with calling-out behavior. It doesn't tell you the learner called out 14 separate times, only that the behavior showed up at least once in 14 of the 20 windows.
Example 2: Hand-flapping during a session
A Board Certified Behavior Analyst (BCBA) wants data on hand-flapping (a form of stereotypy) during a 5-minute session, using 15-second intervals (20 intervals total).
The RBT marks each interval + the instant hand-flapping appears, then waits for the next interval to start.
After the session, 9 of the 20 intervals are marked +, giving a score of 45%.
When should you use partial interval recording?
Partial interval recording works best for behaviors you want to decrease, especially ones that happen quickly or unpredictably.
It fits a session well when:
- The behavior is brief or hard to time exactly, like a shout, a hit, or repetitive movements or sounds (such as hand-flapping or humming) that don’t have a clean start and end point.
- You're targeting a challenging behavior for reduction. Because it tends to overestimate, the method gives a conservative (cautious) read, so a drop that shows up in your graph is unlikely to be exaggerated.
- Counting or timing every instance isn't realistic, like when one RBT is tracking a learner during a busy group activity.
Partial interval vs. whole interval vs. momentary time sampling
Each of these three methods measures a different side of behavior, so the right pick depends on your goal.
These three sit inside a broader set of ABA measurement methods, and behaviors that need an exact count or duration call for continuous measurement.
What are the limitations of partial interval recording?
The biggest limitation of partial interval recording is that it can make a behavior look more frequent, or longer-lasting, than it actually was.
A single brief occurrence fills the whole interval, so this method can overestimate occurrence compared with continuous duration or frequency recording.
The interval length you use plays a role in how large that error gets too. An analysis of more than 800 practice-based ABA sessions found that shorter intervals, 3 minutes or less, matched continuous recording far more closely than the longer intervals many clinicians use.
Partial interval recording works well for day-to-day progress tracking, but a reauthorization report that hinges on exact minutes needs a duration measure.
Tips for accurate partial interval recording
The most reliable partial interval recording comes from short intervals, clear definitions, and consistent observers.
- Keep intervals short: Shorter intervals, often 5 to 15 seconds for frequent behaviors, tend to produce more accurate data than longer ones.
- Train every observer on the same definition: Run a quick practice round together before scoring begins.
- Use a timer or app instead of a wall clock: Losing track of interval boundaries is a common source of scoring errors.
- Pair it with periodic continuous checks: Running a few sessions of full data collection alongside your interval data helps confirm your interval length gives you an accurate picture.
Score partial interval recording data with Passage Health
Scoring 20 intervals by hand means splitting your attention between a stopwatch, a data sheet, and the learner. Passage Health takes over the timing and the math, so RBTs can stay with the session. Here’s how the platform handles all this, plus documentation and practice management:
During the session
- Automatic timers: Frequency, duration, and rate targets can start automatically in the phases you choose, and interval targets run on the interval length and count you set.
- Scoring on a phone or tablet: RBTs mark each interval in the mobile app, and the data fills the session note in real time. Custom permissions control who can view, edit, or manage session notes for each role.
After the session
- Graphs that build themselves: Interval scores turn into graphs automatically, with phase lines added when a target changes phases and trend lines the BCBA can add for any date range.
- Customizable treatment reports: Session data flows straight into the reports you send for reauthorization. Saved signatures go onto any session note with a single click.
- Claims tied to session data: Automation rules turn documented sessions into claims, and you can adjust service lines before you submit.
Across the practice
- One system for the rest of the week: Color-coded schedules, calendar filters that stay applied as you switch views, authorization utilization reports across every client, and Client Intake to manage leads and convert them into clients.
We ship new features every quarter, and our Frontera AI integration adds clinical AI features. Every new practice gets one-on-one onboarding, backed by a responsive support team.
Book a demo to see how Passage Health can turn partial interval recording into cleaner treatment data.
Frequently asked questions
What is partial interval recording used for in ABA?
Partial interval recording is used in ABA to measure whether a behavior happened at any point during a set interval, most often for brief or high-frequency behaviors that are hard to count one at a time. It works well when the goal is reducing a challenging behavior rather than building a new skill.
How do you calculate partial interval recording?
Partial interval recording is calculated by dividing the number of "yes" intervals by the total number of intervals, then multiplying by 100. For example, 16 "yes" intervals out of 24 total intervals works out to about 67%.
What is the difference between partial interval and whole interval recording?
Partial interval recording marks an interval as an occurrence the moment a behavior appears, even briefly, while whole interval recording only counts an interval if the behavior lasts the whole time. That difference is why partial interval scores tend to come out higher than whole interval scores from the same session.
Does partial interval recording overestimate behavior?
Yes, partial interval recording generally overestimates behavior, since scoring only requires one brief instance to mark the entire interval as an occurrence. That tendency works in a clinician's favor when the target is a challenging behavior you want to see drop, since the estimate errs on the side of caution.
How long should intervals be for partial interval recording?
Shorter is better: In a large study of real-world ABA sessions, accuracy fell off once intervals passed 3 minutes, and research on frequent behaviors typically uses windows of 5 to 15 seconds.
References
Fiske, K., & Delmolino, L. (2012). Use of discontinuous methods of data collection in behavioral intervention: Guidelines for practitioners. Behavior Analysis in Practice, 5(2), 77–81. https://doi.org/10.1007/BF03391826
LeBlanc, L. A., Lund, C., Kooken, C., et al. (2020). Procedures and accuracy of discontinuous measurement of problem behavior in common practice of applied behavior analysis. Behavior Analysis in Practice, 13(2), 411–420. https://doi.org/10.1007/s40617-019-00361-6
Meany-Daboul, M. G., Roscoe, E. M., Bourret, J. C., et al. (2007). A comparison of momentary time sampling and partial-interval recording for evaluating functional relations. Journal of Applied Behavior Analysis, 40(3), 501–514. https://doi.org/10.1901/jaba.2007.40-501
