DRAGENwiki · Article

Measuring Behavior: Turning What You See Into Numbers You Can Trust

Sources checked. Clinical review pending.

Measuring behavior means recording a clearly defined behavior in numbers, the same way each time, so a team can see whether it is changing and decide what to do next.

Measuring behavior in 60 seconds

  • Start with a definition. You can only count what everyone agrees on. In the BCBA exam outline, writing operational definitions comes first, followed by measuring how often and how long behavior happens. [1]
  • Pick the dimension that matters. How often, how long, how soon, how far apart. Technicians are expected to "Implement continuous measurement procedures (e.g., frequency, duration, latency, interresponse time)." [2]
  • Every time, or a sample. Continuous systems capture every occurrence; discontinuous systems "capture only a sample of behavior during an observation." [4]
  • A count needs a clock. Six in an hour and six in a week are different stories. Merbitz and colleagues suggest "avoiding all references to isolated counts of behavior." [6]
  • Check the numbers. Behavior analysts are expected to "Evaluate the validity and reliability of measurement procedures." [1]

Jump to: Why measure · What you can measure · Three kinds of measures · Every time or a sample · A count needs a clock · Pick the measure · Check the measure · Measuring yes and no · Example · Myths · Words to know · Need a PDF? · More from DRAGEN · Sources

Why measure

Memory is a poor record keeper. One hard morning can make a whole week feel hard. A calm afternoon can hide a pattern that shows up every Tuesday. Numbers taken the same way each day let the team see what actually happened.

Behavior analysts build their work on this. LeBlanc and colleagues (2016) write that "there are no valid circumstances under which applied behavior analysis (ABA) should be practiced without the collection of meaningful data." [3] Fiske and Delmolino (2012) explain why: "The data generated via direct observation serve as the basis upon which practitioners make treatment decisions and evaluate treatment effects." [4]

That makes the first choice, what to measure and how, one of the most important. Fiske and Delmolino call selecting a measurement system "a pivotal step in effectively designing and evaluating behavioral intervention." [4]

They describe a good system as one that "(a) generates accurate data, (b) produces reliable outcomes, and (c) is sensitive to changes in the occurrence of behavior." [4] In plain words: the numbers match what happened, two people get the same numbers, and the numbers move when the behavior really changes.

What you can measure

A behavior has several dimensions, and each one answers a different question.

  • How often: count and rate. Count is how many times. Rate is count divided by time, such as 3 per hour. Measuring occurrence is its own task in the BCBA outline. [1]
  • How long: duration. The time from the start of one instance to its end.
  • How soon: latency. The time from a cue, such as a request, to the start of the behavior.
  • How far apart: interresponse time (IRT). The time between one instance and the next.
  • How efficient: trials to criterion and more. The outline asks BCBAs to "Measure efficiency (e.g., trials to criterion, cost-benefit analysis, training duration)." [1]
  • Summaries. Raw numbers get summarized for graphs and meetings. RBTs "Calculate and summarize data in different ways (e.g., rate, mean duration, percentage)." [2]

The BCBA outline groups the time-based ones together: "Measure temporal dimensions of behavior (e.g., duration, latency, interresponse time)." [1] Mudford, Taylor and Martin (2009) describe these basic dimensions in standard scientific units: responses per minute for rate, and seconds for durations, interresponse times and latencies. [5]

Magnitude. Some behaviors also vary in strength: how loud a voice is, how hard a door closes. Some teams record this as magnitude or intensity. This page found no open-access guide to measuring it in everyday settings, so it is only named here.

Direct, indirect and product measures

The BCBA outline asks behavior analysts to "Distinguish among direct, indirect, and product measures of behavior." [1]

  • Direct: someone watches the behavior as it happens and records it. Fiske and Delmolino call direct observation and measurement "a defining characteristic of applied behavior analysis as a science and as a practice." [4]
  • Indirect: someone reports on the behavior afterward, such as in an interview or on a rating form. Indirect reports help with planning, and they rest on memory.
  • Product: you measure what the behavior leaves behind: finished math problems, a made bed, a written sentence. LeBlanc et al. set two conditions for this: "the behavior must reliably produce the product, and the product must not be frequently produced by any other behavior or event." [3]

Every time or a sample

Continuous measurement records each instance. Fiske and Delmolino define continuous systems as "those that capture every possible behavioral occurrence," [4] and note that they "offer a complete record that can be reported in standard scientific units, such as responses per minute or percentage of observation." [4]

Discontinuous measurement records a sample. These systems "capture only a sample of behavior during an observation." [4] The BCBA outline gives interval recording and time sampling as examples. [1]

Which is better? When you can do it, continuous. LeBlanc et al.: "When it is possible to capture information on each event, continuous measurement procedures are strongly preferred." [3] Mudford and colleagues add that "Continuous recording is required for direct measurement of the basic dimensions of behaviors." [5]

Sampling still has a place. Fiske and Delmolino observe: "Due in large part to their ease of implementation, discontinuous measurement systems are popular in practice and in applied research." [4] A teacher with a full classroom may not be able to watch one student every second. The cost is error: depending on the method, a sample can overcount or undercount. The DRAGENwiki page on discontinuous measurement covers which way each method leans.

In research, continuous recording has become common. Mudford et al. reviewed ten years of the Journal of Applied Behavior Analysis and found that "Continuous recording was used in the majority (55%) of the 168 articles reporting data on free-operant human behaviors." [5] They point to small devices as one reason: "The availability of handheld portable electronic data-entry and storage devices has increased the practicality and affordability of continuous recording for research and clinical purposes." [5]

A count needs a clock

A count tells you how many. It does not tell you over how long. Six times could mean six times in an hour or six times in a month.

Merbitz, Merbitz and Pennypacker (2016) explain that "A count of some behavior or class of behavior recorded over a known period of time provides the two components of rate." [6] Their advice is simple: "the key guidance is to include the observation time with counts of behavior to avoid the confusion that can be occasioned by its omission." [6]

They also point out a word trap. Some sources use the word frequency to mean a plain count, and others use it to mean count per unit of time. [6] When you read a report, check which meaning the writer used. When you write one, give both numbers: "6 in 30 minutes."

They add one more caution: count divided by time can hide stretches when the behavior could not have happened at all. [6] A behavior that only happens at recess can only be counted on days with recess. Write down when you watched, and for how long.

Picking the right measure

The BCBA outline names the goal: "Select a measurement procedure to obtain representative data that accounts for the critical dimension of the behavior and environmental constraints." [1]

In plain words, ask three questions.

  1. What matters most about this behavior? If the worry is how often, count it and note the time. If the worry is how long it lasts, time it. If the worry is slow starts, measure latency.
  2. Does it have a clear start and end? Counting each event works when the edges are clear. LeBlanc et al. describe event recording as "only appropriate for behaviors that have clear beginnings and endings." [3]
  3. Who is collecting, and what else are they doing? A parent cooking dinner or a teacher running a class may need a simpler system. Fiske and Delmolino note that practitioners "may choose discontinuous systems (which tend to be easier to use) under conditions in which there are multiple demands upon data collectors." [4]

If the behavior leaves a reliable product, measuring the product may mean nobody has to watch at all. [3]

Checking the measurement

Numbers can look precise and still be wrong. The BCBA outline asks behavior analysts to "Evaluate the validity and reliability of measurement procedures." [1] RBTs are expected to "Describe the risks associated with unreliable data collection and poor procedural fidelity." [2]

  • Validity: are you measuring the thing that matters?
  • Reliability: do you get the same numbers when the same thing happens?
  • Accuracy: do the numbers match what truly happened?

The most common check is interobserver agreement (IOA): two people record the same session on their own and compare. Mudford et al. found that every continuous-recording article in their review reported agreement, while "none reported observer accuracy measures." [5] Agreement shows that two people saw it the same way. It does not prove both were right.

Fiske and Delmolino add a caution worth keeping: "Clinicians should be aware of the possibility that true behavior change may not be detected by the measurement system." [4] If the numbers say nothing changed but everyone sees a difference, question the measure too.

Measuring yes and no

Measurement is not only for behavior a team hopes to see less of. Skills count too: asking for help, taking turns, asking for a break.

So does a learner's yes and no. Breaux and Smith (2023) describe practice in which "data on assent, the withdrawal of assent, and related skill repertoires are collected, analyzed, and used in decision-making within and across intervention sessions." [7]

DRAGEN's rule. We measure to understand and to help, never to build a case against a person. A "no" is recorded as information about the task, the setting or the relationship, not as "noncompliance" to reduce. When a behavior can cause harm, it gets its own definition, its own measure and its own safety plan, kept separate from the learner's ways of saying no.

A made-up example: Student A

Student A is an invented 9-year-old. A parent says homework time is always a disaster. The team, which includes the parent, picks four simple measures:

  • Latency: minutes from "homework time" to pencil on paper.
  • Duration: minutes spent working.
  • Break requests: how many times Student A asks for a break, with the length of homework time written next to it. Every request is honored. Asking for a break is a skill.
  • Product: problems finished correctly, counted from the worksheet afterward.

On Monday, Student A asks for 4 breaks in 60 minutes. On Wednesday, Student A asks for 4 breaks in 20 minutes. A count alone says the two days were the same. With the clock, Monday was 4 per hour and Wednesday was 12 per hour.

The worksheets explain the jump. Wednesday was the first day of long division, and only 2 problems were finished correctly. The breaks were information. The team changes the work, not the child: shorter sets, the steps taught again, and a worked example at the top of each page.

Two weeks later the numbers tell a calmer story. Latency is down from 8 minutes to 2. Break requests settle near 3 per hour, still honored every time, and correct problems are up. The parent remembered Wednesday as the worst day of the week. The numbers agreed, and they also showed why.

Common myths

  • "Counting is enough." A count without the observation time can mislead. Merbitz et al. suggest "avoiding all references to isolated counts of behavior." [6]
  • "A sample is just as good as watching everything." A sample is easier, but continuous measurement is strongly preferred when it is possible. [3]
  • "If two observers agree, the data are accurate." Agreement checks consistency. In one review, every continuous-recording study reported agreement and none reported accuracy. [5]
  • "Data are only for problem behavior." Skills, assent and the withdrawal of assent can be measured and used in decisions. [7]
  • "Numbers are objective, so they cannot be wrong." A measure can miss real change. [4]

Words to know

  • Measurement: recording behavior in numbers, the same way each time.
  • Dimension: one feature of behavior you can measure, such as how often or how long.
  • Count: how many times a behavior happened.
  • Rate: count divided by observation time, such as 4 per hour.
  • Duration: how long one instance lasts, from start to end.
  • Latency: the time from a cue to the start of the behavior.
  • Interresponse time (IRT): the time between one instance and the next.
  • Continuous and discontinuous measurement: recording every instance, or recording a sample.
  • Direct, indirect and product measures: watching it happen, hearing about it later, or measuring what it leaves behind.

Need a PDF?

Create it here, for free: the DRAGEN app. Use it at your next session or team meeting.

Or print one we already made, also free:

More from DRAGEN

Related DRAGENwiki pages: operational definition · continuous measurement · discontinuous measurement · frequency · duration · latency · permanent product · interobserver agreement

What remains unknown

  • The measurement-selection model by LeBlanc et al. "has not been empirically tested," in the authors' own words. [3]
  • Very fast behavior, several behaviors at once, or an observer with other jobs may lower the accuracy of continuous recording. Fiske and Delmolino note that the impact of these factors "has not been definitively examined in research." [4]
  • The review by Mudford et al. covers research articles from 1995 to 2005. [5] How often everyday clinics and schools use continuous recording today is not known from that review.
  • Words like frequency and rate are used in different ways by different sources. [6] Check the meaning in each report.
  • This page found no open-access guide to measuring the magnitude or intensity of behavior in everyday settings, so magnitude is only named here.
  • This page is general information, not medical or legal advice.
  • This page was last checked on November 12, 2026.

Sources

  1. Behavior Analyst Certification Board. BCBA Test Content Outline (6th ed.). https://www.bacb.com/wp-content/bcba-outline-6thEd/
  2. Behavior Analyst Certification Board. RBT Test Content Outline (3rd ed.), effective January 1, 2026. https://www.bacb.com/rbt-test-content-outline-3rd-ed
  3. LeBlanc, L. A., Raetz, P. B., Sellers, T. P., & Carr, J. E. (2016). A proposed model for selecting measurement procedures for the assessment and treatment of problem behavior. Behavior Analysis in Practice, 9(1), 77–83. https://doi.org/10.1007/s40617-015-0063-2 (free full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC4788644/)
  4. 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 (free full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC3592492/)
  5. Mudford, O. C., Taylor, S. A., & Martin, N. T. (2009). Continuous recording and interobserver agreement algorithms reported in the Journal of Applied Behavior Analysis (1995–2005). Journal of Applied Behavior Analysis, 42(1), 165–169. https://doi.org/10.1901/jaba.2009.42-165 (free full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC2649837/)
  6. Merbitz, C. T., Merbitz, N. H., & Pennypacker, H. S. (2016). On terms: Frequency and rate in applied behavior analysis. The Behavior Analyst, 39(2), 333–338. https://doi.org/10.1007/s40614-015-0048-z (free full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC6701258/)
  7. Breaux, C. A., & Smith, K. (2023). Assent in applied behaviour analysis and positive behaviour support: Ethical considerations and practical recommendations. International Journal of Developmental Disabilities, 69(1), 111–121. https://doi.org/10.1080/20473869.2022.2144969 (free full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC9897747/)

Pages that link here