cumulative

C.3 Measuring occurrence: count, rate, and percentage

What the task asks: Measure how often a behavior occurs by choosing and calculating the right measure (count, rate or percentage) for the behavior and the way observation is arranged.

Measurement, Data Display, and Interpretation5 min read4 practice questionsOfficial outline (opens in a new tab)

The big idea

Any instance of behavior can happen again. That property, repeatability, is what makes behavior countable, and measures of occurrence are the most common data in behavior analysis: how many times a learner asks for help, hits, completes a problem or names a picture.

The raw number is rarely enough. Twelve call-outs mean something different in a 10-minute lesson than in a 2-hour one, and 8 correct answers mean something different out of 10 chances than out of 40. This task is about choosing the measure that makes the number comparable from one session to the next.

Breaking it down

Count

Count is a simple tally of how many times the behavior occurred. It works on its own only when every observation period is the same length and nobody else limits the learner’s opportunities. Report the observation time alongside the count so readers can judge it.

Rate

Rate is count divided by the time observed, stated per minute, per hour or per day.

Rate = count ÷ observation time

Marcus asks for help 12 times in a 30-minute session on Monday and 10 times in a 20-minute session on Tuesday. The counts suggest a drop, but the rates are 12 ÷ 30 = 0.4 per minute and 10 ÷ 20 = 0.5 per minute. Help seeking actually rose.

Count Looks like a drop 12 10 Monday 30 min Tuesday 20 min Rate per minute Actually a rise 0.4 0.5 Monday 12 ÷ 30 Tuesday 10 ÷ 20
Marcus's help requests on two days. The counts suggest a drop, but Tuesday's session was shorter: as a rate, help seeking rose from 0.4 to 0.5 per minute.

Rate suits free operants, behaviors a person can emit at any time without someone else setting up each opportunity: talking out, stereotypy, spontaneous requests, words read aloud. It also reveals fluency. Two readers can both be 100% accurate while one reads 40 words per minute and the other 110.

A note on “frequency.” Cooper, Heron and Heward, following Johnston and Pennypacker, treat frequency as a synonym for rate. Many practitioners and some older materials use it to mean a simple count, and published studies use it both ways (Carr et al., 2018). When you meet the word, check whether time is part of the calculation. If the count was divided by time, it is a rate.

Percentage

Percentage is the number of responses of one kind divided by a total, multiplied by 100. Most often the total is the number of opportunities.

Percentage = (target responses ÷ total opportunities) × 100

Priya names 14 of 20 pictures correctly: 14 ÷ 20 × 100 = 70%.

Percentage suits restricted operants, such as discrete trials, where an instructor controls when each chance to respond occurs (G.13). A rate there would mostly reflect how quickly the instructor presents trials. Percentage is also used for interval data (C.6) and interobserver agreement (C.8).

Two cautions. First, a percentage hides its denominator: 1 of 2 and 50 of 100 are both 50%, but one response moves the first by 50 points. Report the number of opportunities. Second, percentage tops out at 100% and says nothing about speed, so a learner can sit at 100% while still responding slowly.

1 of 2 correct 50% One more correct: jumps to 100% 50 of 100 correct 50% One more correct: 51%
Both scores are 50%. With 2 opportunities, one more correct response moves the score by 50 points; with 100 opportunities, it moves the score by 1 point.

Celeration

Celeration is the change in rate over time, expressed as a factor per unit of time. A celeration of ×2 per week means the rate doubles each week; ÷2 per week means it halves. It is read from the Standard Celeration Chart (C.10) and is central to precision teaching. Trials to criterion, another count-based measure, belongs with efficiency in C.7.

Seeing it in practice

SituationMeasureWhy
Ms. Romero counts call-outs in a class that always lasts 45 minutes.Count (or rate)Equal observation periods make counts comparable.
A clinic logs Eli’s aggression in sessions lasting 1 to 3 hours.Rate per hourSession length varies.
Hana gets 10 color-matching trials each session.Percentage correctThe instructor controls the opportunities.
Ade’s math-fact accuracy has been 100% for a month.Correct and incorrect per minuteAccuracy has hit its ceiling; speed still varies.
Rosa’s reading rate is tracked to see how fast it is growing.CelerationThe question is about change in rate across weeks.

Common mix-ups

  • Comparing counts from sessions of different lengths. Convert to rate first.
  • “Per session” is not a rate unless every session is the same length.
  • Using rate for discrete trials. The instructor’s pace sets the ceiling. Use percentage correct or trials to criterion.
  • Treating 100% as mastery. Percentage cannot show fluency; add a timed measure.
  • Ignoring the denominator. “Correct on 100% of opportunities” might mean 2 of 2.
  • Assuming frequency always means count. Check the arithmetic, not the label.

Exam tip: If the stem says observation periods differ in length, think rate. If it says the instructor presents a set number of trials, think percentage. If it asks how quickly a rate is growing from week to week, think celeration.

Quick recap

  • Count is a tally; it is comparable only across equal observation periods.
  • Rate is count divided by time and suits free operants. Many sources call it frequency.
  • Percentage is a ratio times 100 and suits discrete trials; always report the denominator.
  • Percentage has a ceiling and ignores speed; rate does not.
  • Celeration is the change in rate over time, such as ×2 per week.

Further reading

  • Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). Applied behavior analysis (3rd ed.), chapter 4.
  • Carr, J. E., Nosik, M. R., & Luke, M. M. (2018). On the use of the term “frequency” in applied behavior analysis. Journal of Applied Behavior Analysis, 51(2), 436-439.

Key terms

Count
The number of times a behavior occurs during an observation period. Count can be compared across sessions only when the sessions are the same length and opportunities to respond are not limited by someone else.
Rate (Frequency, Response rate)
The number of responses per unit of time: the count divided by the length of the observation period. Many authors use frequency as a synonym for rate, though others use frequency to mean a simple count, so check whether time is part of the calculation.
Percentage
A ratio expressed per hundred: the number of responses of a given kind divided by the total number of responses or opportunities, multiplied by 100. It suits discrete trials, but it hides the size of the denominator and cannot show speed.
Celeration
The change in rate of responding over time, expressed as the factor by which the rate multiplies (acceleration) or divides (deceleration) per unit of time, such as ×2 per week. It is read as the slope of the data on a Standard Celeration Chart.
Repeatability (Countability)
The property that instances of a behavior can occur again and again over time, which makes the behavior countable. Count, rate and celeration are measures of repeatability.

Check yourself

4 questions on C.3. Choose an answer to see whether you're right and why.

  1. Question 1A technician records Zoe's vocal stereotypy on three days: 30 times in 15 minutes on Monday, 36 times in 24 minutes on Tuesday and 40 times in 40 minutes on Wednesday. What happened to the rate?
    Show the answer

    Answer: D. Rate is count divided by time: 30 ÷ 15 = 2.0, 36 ÷ 24 = 1.5 and 40 ÷ 40 = 1.0 per minute, so the rate fell. The raw counts rose only because the sessions got longer, and counts from sessions of different lengths cannot be compared directly.

  2. Question 2During discrete-trial teaching, a therapist presents 10 trials per session as fast as Omar responds. Why is percentage correct a better measure here than rate of correct responses?
    Show the answer

    Answer: D. In discrete trials, the instructor sets the number and pacing of opportunities, so a rate would largely reflect the therapist's pace. Percentage correct shows how Omar responds to the chances he gets. Percentage says nothing about fluency, and rate applies to any free operant, not only problem behavior.

  3. Question 3Lena and Sam both score 100% on a 30-problem math-facts probe, but Lena finishes in 1 minute and Sam takes 4. Which measure would show the difference between them?
    Show the answer

    Answer: D. Rate builds time into the measure: Lena answers 30 ÷ 1 = 30 correct per minute and Sam 30 ÷ 4 = 7.5. Percentage correct tops out at 100%, so it cannot separate two accurate learners who differ in speed. Both attempted the same 30 problems, so the count is identical too.

  4. Question 4Kofi answers 3 of 4 questions correctly on Monday and 15 of 20 on Tuesday. Which statement is accurate?
    Show the answer

    Answer: A. 3 ÷ 4 × 100 = 75% and 15 ÷ 20 × 100 = 75%, so accuracy is the same. On Monday a single answer moves the score by 25 points; on Tuesday it moves it by 5. That is why the number of opportunities should be reported with any percentage.

Adds a step to your study record. Saved in this browser only.

Type to search all 104 guides and the glossary.