Momentary Time Sampling in ABA: A Practical Guide

Key Takeaways

Momentary Time Sampling (MTS) records whether a behavior is occurring at a single moment in time, at the end of a fixed interval. It’s the right method for high-rate, continuous, or distributed behaviors that frequency counting can’t catch, especially when one observer is tracking multiple people at once.

• MTS is best for engagement during shared instruction, parallel play, on-task behavior, and other sustained behaviors observed across multiple students.
• MTS is the wrong method for low-rate or discrete behaviors - use frequency for those.
• Interval lengths typically range from 15 seconds to 5 minutes, depending on behavior rate and observation length.
• TallyFlex shipped MTS in February 2026 with a full-screen overlay and a 15-second entry window for marking the sample moment.

If you’re trying to track on-task behavior across 6 students at once, frequency counting will collapse.

A clean way to do MTS without juggling a paper grid, a kitchen timer, and six students at the same time was the most common request school-based teams sent us in late 2025. TallyFlex shipped Momentary Time Sampling in February 2026 with a full-screen overlay and a 15-second entry window for marking the sample moment. This guide is the full picture: what MTS is, when to pick it, when not to, and how to set it up well.

What is Momentary Time Sampling?

Momentary Time Sampling is a discontinuous observation method. You divide the observation period into equal intervals, and at the end of each interval (the “moment”), you record whether the target behavior is occurring or not. You don’t track the behavior continuously. You sample it.

The output is a percentage: number of moments the behavior was observed, divided by the total number of moments observed. If you observed 20 moments and the behavior was present at 12 of them, the data point is 60%.

This contrasts with the other interval methods:

• Partial interval recording: records whether the behavior occurred at any point during the interval. Tends to overestimate behavior duration.
• Whole interval recording: records whether the behavior occurred for the entire interval. Tends to underestimate.
• Momentary time sampling: records whether the behavior is occurring at the moment the interval ends. It can provide a practical estimate of the proportion of time the behavior occurs, especially for continuous or high-rate behaviors.

A brief history note: MTS was developed in the 1970s as a way to make observational research feasible in classrooms and group settings, where continuous frequency or duration recording was impossible for a single observer. Cooper, Heron and Heward describe it as “particularly well suited to behaviors that occur at relatively high frequencies, behaviors of long duration, or when multiple individuals must be observed simultaneously.”

The trade-off is that MTS is a sample. Anything happening between the sampling moments is, by design, not recorded. That’s the point - and it’s the limitation.

When MTS is the right choice

A few scenarios where MTS is the cleanest method:

1. Classroom-wide on-task behavior. A SPED teacher needs to know whether their three students are on-task during morning circle. Frequency would require tracking dozens of behaviors per minute. Duration would require three stopwatches. MTS with a 30-second interval gives a clean percentage of on-task moments per student per session.

2. Engagement during a long instruction session. A BCBA is tracking engagement during a 60-minute language session - whether the learner is using the assigned materials, responding through their communication mode, or participating in the current activity as defined in the plan. The behavior is sustained and continuous. Frequency wouldn’t capture it (engagement isn’t a discrete count). Duration would require a continuous timer that the BCBA can’t run while also delivering trials. MTS at 1-minute intervals gives a practical estimate of engaged moments without disrupting the trials. Engagement-tracking is the recommended substitute for using MTS to track stereotypy as a reduction target (Mathur, Renz & Tarbox 2024); it gives the same methodological illustration without the ethical exposure.

3. Parallel play in early intervention. A BCBA is tracking whether a 4-year-old is engaged in parallel play during a 30-minute peer interaction. Engagement is continuous and not easily countable. MTS at 1-minute intervals gives a usable measure of engagement proportion.

4. Engagement in a group setting. A school psychologist is observing four students during a group activity. One observer, four students, 20 minutes. If the observer can record all four students at each 1-minute cue inside the entry window, each student gets enough sample moments for a usable engagement estimate. If the observer must rotate one student at a time, extend the observation or adjust the interval so each student still gets at least 15-30 sample moments.

5. Behavior during a long observation window. Any time the observation window is long enough that continuous recording is impractical, and the behavior is the kind where “is it happening right now?” is a meaningful question.

When MTS is the wrong choice

A couple of scenarios where MTS is the wrong tool:

1. Low-rate or discrete behaviors. If the target behavior happens 2-3 times per session and each instance is brief, MTS will miss most of the instances by design. The behavior happens between sampling moments and isn’t captured. Use frequency instead.

2. Behaviors where exact count or exact duration matters clinically. If the clinical question is “how many dysregulation episodes today” or “how long was the longest episode,” MTS doesn’t answer either question. It answers “what proportion of sample moments showed an episode in progress.” Use frequency or duration if the count or duration is the clinically meaningful number.

Step-by-step setup

Setting up MTS well takes a few decisions in advance.

1. Define the target behavior operationally.

Before MTS or any recording method, write a clear operational definition. The definition should be specific enough that two observers, watching the same moment, would agree whether the behavior is occurring. “On-task” is too vague. “Engaged with the assigned activity by using materials, responding through the learner’s communication mode, initiating a relevant response, or participating in the current routine as defined in the plan” is more useful and easier to individualize.

2. Pick the interval length.

Interval length depends on behavior rate and observation length. A few rules of thumb:

• High-rate behaviors: shorter intervals (15-30 seconds). The shorter sample makes sure you catch enough variation.
• Lower-rate behaviors: longer intervals (1-5 minutes). Longer samples reduce observer fatigue.
• Long observation windows: longer intervals are more sustainable. A 60-minute observation at 15-second intervals is 240 sample moments, which is exhausting. A 60-minute observation at 1-minute intervals is 60 moments, which is workable.
• Multiple students at once: longer intervals can help, but the observer still needs to enter all students’ sample-moment data inside the brief entry window. If four students cannot be recorded reliably at the cue, reduce the group, rotate observations, or adjust the observation plan while keeping enough sample moments per student.

3. Pick the sampling cue.

MTS requires a reliable cue at the end of each interval. The cue is what tells the observer “this is the sample moment; record what is true now.” Common cues include:

• A vibration or chime from a phone or tablet timer.
• A discreet earpiece beep (for in-classroom observers who don’t want students to hear the cue).
• A visual cue on a screen.

The cue has to be brief so it doesn’t distort the behavior being observed. If your cue is a loud alarm, you’ve changed the environment, and the data is no longer about the natural behavior.

4. Pick the entry window.

After the sampling cue, the observer has a brief entry window to mark whether the behavior was occurring at the exact sample moment. TallyFlex uses a 15-second entry window by default, which gives a single observer time to enter data for one or several students. It is not a 15-second observation period. If observers start recording anything that happens during the window, they have moved away from MTS and toward partial interval recording.

5. Pick the observation window length.

How long is a session? For classroom observations, 15-30 minutes is common. For clinical sessions, the observation might be the full session. The number of sample moments per session should be enough to give you a stable percentage - typically 15-30 sample moments minimum.

6. Run the observation.

The observer follows the timer. At each cue, mark whether the target behavior was occurring at that exact moment. Continue. Don’t try to record what happened between cues - that’s not MTS, that’s continuous recording.

7. Calculate the data point.

At the end of the observation, count the number of sample moments where the behavior was occurring, divide by the total number of sample moments, multiply by 100. That’s your percentage for the session. Plot it on a line graph using the same graphing convention you use for other ABA session data.

Real-world examples

A few examples drawn from how schools and clinics actually use MTS.

Example 1: Classroom-wide on-task behavior.

A SPED teacher in a public school has six students in a self-contained classroom. She wants on-task data during morning circle (20 minutes, 9:00-9:20). She picks a 1-minute interval. At each cue, the teacher marks whether each student met the individualized engagement definition at that sample moment. After 20 minutes, she has 20 sample moments per student, 120 total observations. The data is plotted as percent on-task per student per day.

Example 2: Engagement during a 1-hour language session.

A BCBA is tracking engagement during a 1-hour language session with a 6-year-old - whether the learner is using the materials, responding through their communication mode, or initiating a relevant response when the BCBA presents a trial. Engagement is sustained and continuous, so frequency or duration wouldn’t capture it cleanly. She picks a 30-second interval, which gives her 120 sample moments per session. At each cue, she records whether the learner was engaged at that moment. The session yields a percentage she compares across baseline and after engagement-supporting changes (pacing, materials, choice). Tracking engagement as the increase target keeps the protocol aligned with positive-reinforcement-first practice (Mathur, Renz & Tarbox 2024; BACB Code 2.14).

Example 3: Group engagement observation.

A school psychologist is observing four students during a 30-minute group activity. She picks a 1-minute interval. At each cue, she records which students met their engagement criteria at that sample moment. The 30-minute observation gives 30 sample moments per student. Engagement percentages for each student go in the report.

TallyFlex shipped MTS in February 2026

We built MTS into TallyFlex specifically because the school teams were asking for it. The implementation:

• Full-screen overlay during sample moments. The overlay covers the rest of the app so the observer focuses on the decision: occurring or not occurring.
• 15-second entry window for recording what was true at the exact sample moment.
• Configurable interval lengths from 15 seconds to 10 minutes.
• Multi-student support built into the sample moment so an observer tracking multiple students can mark each one in the same entry window.
• Automatic percentage calculation at the end of the observation.
• Reports that plot the data alongside other recording methods on the same graph.

For the full feature list and the documentation walkthrough, see /docs/sessions/recording-methods. For the broader question of when to pick MTS vs partial interval vs frequency, see Partial vs Whole Interval Recording: A Quick Guide and How to Choose the Right ABA Recording Method. The interactive tool at /recording-method-finder walks through the decision in three questions.

Common MTS mistakes

A few things to watch for:

• Treating the cue as a “during the interval” check. The whole point of MTS is the moment, not the interval. If your observers are recording “any time during the interval,” they’re doing partial interval, not MTS.
• Interval too short for observer capacity. If the observer can’t reliably mark the data at every cue, the interval is too short. Lengthen it.
• Cue too disruptive. Any cue that changes the environment changes the data. Use vibration or a discreet sound.
• Too few sample moments. A 5-minute observation at 1-minute intervals gives 5 sample moments, which produces unstable percentages. Aim for 15-30 sample moments minimum.
• Mixing methods within a session. Decide whether you’re doing MTS, partial interval, or whole interval, and stick with one. Mixing them produces uninterpretable data.

Closing

Momentary Time Sampling is the right tool for high-rate, continuous, or distributed behaviors that frequency or duration recording can’t capture. It’s especially valuable in classroom and group settings where one observer is tracking multiple people at once. The set-up takes a few decisions - operational definition, interval length, sampling cue, entry window - and the data stays interpretable as long as you stick to “the moment” and not “during the interval.”

If you’re running paper grids and a kitchen timer for MTS today, that’s the workflow worth replacing first. Start Tracking on TallyFlex and run your first MTS session with the full-screen overlay. The percentage builds itself.

Documentation shouldn’t follow you home.