Home / RBT Study Guide / Data Collection and Graphing / Continuous measurement: frequency, rate, duration, latency and IRT

Domain A · Data Collection and Graphing · Task A.1

Continuous measurement: frequency, rate, duration, latency and IRT

Continuous measurement records every instance of a behavior during the observation. The measures are frequency (count), rate, duration, latency and interresponse time, and they differ in what you count or time.

Last reviewed November 10, 2026 against the BACB RBT Test Content Outline (3rd ed.)

What continuous measurement means

Continuous measurement means you record every instance of the target behavior for the entire observation period. Nothing is sampled or estimated. Task A.1 of the RBT Test Content Outline asks you to implement these procedures and names four examples: frequency, duration, latency and interresponse time. Rate is frequency divided by time, so it belongs to the same family.

Your supervisor picks the measure when writing the plan. Your job is to know what each one captures so you can collect it correctly and notice when the data sheet does not fit what you are seeing.

The five measures side by side

MeasureWhat you recordQuestion it answersExample
Frequency (count)A tally for each occurrenceHow many times?Mia asked for help 7 times
RateCount divided by observation timeHow often per minute or hour?7 requests in 35 minutes is 0.2 per minute
DurationTime from the start of a response to its endHow long did it last?Sam's tantrum lasted 6 minutes
LatencyTime from the cue to the start of the responseHow long before it began?Leo started cleaning up 15 seconds after being asked
Interresponse time (IRT)Time from the end of one response to the start of the nextHow much time between responses?40 seconds passed between Ava's bites

Frequency and rate

Frequency (also called count or event recording) is the simplest measure: one tally each time the behavior happens. It fits behaviors that have a clear beginning and end, look about the same each time and do not happen too fast to count. Hitting, raising a hand and asking for a break are typical examples.

A count by itself can mislead when observation time changes from day to day. Rate fixes that by dividing the count by the time observed.

In session

Diego hits 8 times in a 20-minute session on Monday and 12 times in a 60-minute session on Tuesday. Monday: 8 / 20 = 0.4 hits per minute, or 24 per hour. Tuesday: 12 / 60 = 0.2 hits per minute, or 12 per hour. The count went up, but hitting happened half as often per minute on Tuesday.

To calculate rate, keep the same time unit all the way through. If the session length is in minutes, the answer is per minute. Multiply by 60 to get a rate per hour.

Duration, latency and interresponse time

All three are measured with a timer. What separates them is where the timer starts and stops.

In session

Noor's RBT says, "Start your worksheet." Noor picks up her pencil 20 seconds later (latency). She works for 4 minutes before stopping (duration). After finishing problem one, she waits 30 seconds before starting problem two (interresponse time).

Common mix-up

Latency and interresponse time both measure a gap. Latency is the gap between a cue and a response. Interresponse time is the gap between two responses. If the scenario has no instruction or signal in it, it is not latency.

Matching the measure to the behavior

The right measure is the one that captures the feature of the behavior the team wants to change.

Continuous measurement has limits. It requires your full attention on one behavior, and it depends on a clear operational definition so you know exactly when a response starts and ends. When a behavior happens at a very high rate, or you are also running teaching trials, a supervisor may choose a discontinuous measurement method instead. If the method on your data sheet is not workable in a session, record as accurately as you can and tell your supervisor. Do not switch methods on your own.

Once the numbers are collected, they are summarized as a rate, a mean duration or a percentage. Those calculations are worked through on the data reliability and procedural fidelity page.

How the exam tests it

A common question format gives you a two-sentence scenario and asks which measurement is being used. Look for where the timing starts and stops, or whether anything is being timed at all.

On the exam

Underline the anchor words. "How many times" is frequency. "Per minute" or "per hour" is rate. "How long it lasted" is duration. "After the instruction" is latency. "Between one response and the next" is interresponse time. If the scenario mentions intervals, you have left continuous measurement and the answer is a discontinuous method.

Calculation items usually hinge on rate. Divide the count by the time, check that the units match, and be suspicious of an answer choice that divides time by count. For an overview of how this task fits with the rest of the domain, see the Domain A guide.

Practice questions

Original questions in the style of the exam. Pick an answer before you open the explanation.

1. Ava finishes one bite of lunch, and 40 seconds pass before she starts the next bite. Her RBT is timing that gap between bites. What is being recorded?

  1. Latency to respond
  2. Interresponse time
  3. Response duration
  4. Response frequency
Show answer

B. Interresponse time

Interresponse time runs from the end of one response to the start of the next response. Latency would need a cue or instruction as the starting point, and here the timer starts when a response ends.

2. Diego hits 8 times in a 20-minute session on Monday and 12 times in a 60-minute session on Tuesday. Which statement is accurate?

  1. The rate of hitting was higher on Tuesday
  2. The rate of hitting was equal on both days
  3. The rate of hitting cannot be calculated
  4. The rate of hitting was lower on Tuesday
Show answer

D. The rate of hitting was lower on Tuesday

Monday is 8 / 20 = 0.4 per minute and Tuesday is 12 / 60 = 0.2 per minute, so the rate was lower on Tuesday. Choosing higher on Tuesday confuses the raw count (12 is more than 8) with rate, which accounts for session length.

3. Noor has about two tantrums a day. Some last 1 minute and some last 25 minutes. Which measure would best show whether her tantrums are getting shorter?

  1. Duration recording
  2. Frequency recording
  3. Latency recording
  4. Interresponse time
Show answer

A. Duration recording

Duration captures how long each tantrum lasts, which is the feature expected to change. Frequency would show two a day whether the tantrums were short or long, so it would miss the improvement.

Common questions

What is the difference between frequency and rate in ABA?

Frequency is the number of times a behavior occurred. Rate is that count divided by the observation time, such as 0.4 per minute or 24 per hour. Rate lets you compare sessions of different lengths fairly, and a raw count does not.

What is the difference between latency and interresponse time?

Latency is the time from an instruction or cue to the start of the response. Interresponse time is the time from the end of one response to the start of the next response. Latency always involves a cue, and interresponse time involves two responses.

When is frequency recording a poor fit?

Frequency works poorly when a behavior has no clear beginning and end, happens too quickly to count accurately, or varies a lot in how long it lasts. In those cases a supervisor may choose duration or an interval method. The RBT does not switch methods alone and reports the concern to the supervisor.