2.2 Continuous Measurement: Frequency, Duration, Latency, and IRT

Continuous Measurement: Frequency, Duration, Latency, and IRT

Continuous measurement means the observer is trying to record every instance of the target response during the observation period. The 2026 RBT Test Content Outline expects RBTs to implement continuous procedures such as frequency, duration, latency, and interresponse time. These measures are not interchangeable. Each one answers a different question, and each one can mislead the team if used for the wrong kind of behavior.

Frequency is a count of responses. It works best for behaviors with a clear beginning and end, such as raising a hand, throwing a toy, requesting help, or completing a step in a task analysis. A frequency count of 12 tells the supervisor how many responses occurred, but not how long the observation lasted. If one session was 20 minutes and another was 60 minutes, raw frequency alone may make the longer session look worse even if the response was less common per minute. That is why frequency is often converted to rate when observation periods differ.

Duration measures how long a behavior lasts from onset to offset. It is useful when the concern is time spent in the behavior, not just number of episodes. Examples include crying episodes, time out of seat, engagement with leisure materials, or time spent independently working. A client who leaves seat two times for 30 seconds each is different from a client who leaves seat two times for 14 minutes each. Frequency alone would hide that difference. Duration requires clear start and stop rules, a timer, and enough attention to begin timing at onset rather than after the RBT notices the behavior has been happening.

Latency measures how much time passes between an antecedent or instruction and the start of the response. It answers the question: How long does it take the client to begin? Latency may be used for following instructions, responding to name, transitioning after a cue, or starting a task after materials are presented. The RBT must know exactly what starts the clock and what stops it. If the instruction is Clean up and the response is placing the first toy in the bin, the clock starts when the instruction ends and stops when the first toy enters the bin, unless the written plan says otherwise.

Interresponse time, often shortened to IRT, measures the time between two consecutive responses. It answers a different question from latency: How much time passes after one response before the next response occurs? IRT may be relevant for high-rate behavior such as repetitive questions, calling out, or requests for attention. If the team is trying to increase spacing between requests, IRT can show whether the client is waiting longer between responses. The RBT should not confuse IRT with duration. Duration is how long one response lasts. IRT is the gap between responses.

Continuous Measurement Decision Matrix

In practice, the written plan may require more than one continuous measure. Suppose a client engages in screaming during table work. If the team only counts episodes, data may show three episodes in both sessions. But one session may include three brief yells and another may include three episodes lasting a total of 22 minutes. The supervisor may therefore ask for both frequency and duration. The RBT's role is to implement the specified system accurately and report if the system becomes impractical.

For example, if the RBT is alone during a safety protocol and cannot start and stop a timer precisely, that barrier should be reported objectively.

Continuous measurement also requires attention to response cycles. A behavior definition may specify how to score repeated movements. For example, hand-to-head hits may count as separate responses only when the hand leaves the head and returns, or they may be counted as one episode until five seconds pass without contact. Without this rule, one RBT may count a burst as 1 and another may count 17. The same issue appears in verbal behavior. Are three repeated calls of teacher, teacher, teacher counted as three responses, one episode, or one response with duration?

The plan should answer this, and the RBT should ask for clarification when it does not.

The biggest advantage of continuous measurement is detail. It can show change in amount, length, speed, or spacing of behavior. The tradeoff is effort. Continuous recording can be difficult while delivering prompts, reinforcement, error correction, safety support, or naturalistic teaching. An RBT should prepare recording tools before the observation starts and should avoid estimating after the fact. If exact measurement becomes impossible, the RBT should document what happened, continue to support the client according to the plan, and notify the supervisor.

Accurate data are valuable, but client dignity, safety, and faithful implementation of the plan remain central.