Duration and Latency Recording - When to Use Each in ABA

Key Takeaways

Duration and latency recording both use timers, but they answer different questions. Use duration recording when how long a behavior lasts matters - tantrums, on-task behavior, engagement periods. Use latency recording when the time between a cue and the response matters - compliance, transitions, starting work. Both require behaviors with clear start and end points. The question to ask: “Do I need to know how long this lasted, or how quickly they responded?”

Duration vs Latency: The Quick Decision

If you can’t identify exactly when a behavior begins and ends, interval recording may be a better fit. For a complete overview of all methods, see our recording method guide.

Duration Recording

Duration recording measures how long a behavior lasts from start to finish.

Start timing when the behavior begins. Stop when it ends. Your data point is a time measurement - seconds, minutes, or hours.

When to Use Duration Recording

Duration makes sense when how long something lasts is what actually matters:

• Tantrum behavior: A parent reports tantrums are “constant.” Duration data reveals whether episodes last 2 minutes or 20 minutes.
• On-task behavior: How long does a student sustain attention during independent work? 30 seconds or 10 minutes?
• Time in designated area: Tracking how long a student stays in their seat or at their workstation.
• Meltdowns and outbursts: How long does an episode last once it starts? 3 minutes or 30 minutes makes a big difference in planning support.
• IEP goals: Many IEP objectives specify duration targets - “Student will remain engaged in group instruction for 10 minutes.”

Key insight: A student who tantrums 3 times on Monday and 3 times on Tuesday shows identical frequency. But if Monday’s episodes lasted 2 minutes each and Tuesday’s lasted 15 minutes each, you’re dealing with a very different situation. Duration captures what frequency misses.

Collecting Accurate Duration Data

Two requirements:

1. A clear start - You can identify exactly when the behavior begins
2. A clear end - You can identify exactly when the behavior stops

Be specific about the end point. “Tantrum ends when the student stops crying” is vague. “Tantrum ends after 5 consecutive seconds without crying, screaming, or throwing” gives every team member the same rule.

You can track total duration (sum of all episodes) or per-occurrence duration (length of each episode). Per-occurrence data is richer - you can see if episodes are getting shorter even when frequency stays the same.

Latency Recording

Latency recording measures the time between a cue and the response.

The cue can be a verbal instruction (“sit down”), a signal (the bell ringing), or any event the student should respond to. Start timing when the cue happens. Stop timing when the student begins responding.

When to Use Latency Recording

Latency matters when how quickly someone responds is what you need to measure - even if accuracy is already high:

• Compliance: A student follows directions correctly but takes 30-45 seconds to start. Accuracy data shows 100%. Latency reveals the delay.
• Transitions: How quickly does a student start packing up after the bell? Critical for classroom scheduling.
• Starting independent work: The teacher says “open your journals” - does the student begin in 5 seconds or 45 seconds?
• Response to name: Does a child respond within 5 seconds or 20 seconds when their name is called?
• Safety skills: Skills like stopping at curbs or responding to “Come here” need to happen quickly, not just accurately.

Key insight: A student identifying 10/10 colors with 100% accuracy might not have mastered the skill if each response takes 15-20 seconds with visible problem-solving. Quick, automatic responses suggest true fluency - and the only way to see that difference is by measuring latency.

Collecting Accurate Latency Data

Three requirements:

1. A clear instruction - You know exactly when you gave the direction
2. A defined response - You know what “starting to comply” looks like
3. A visible start - You can tell when the student begins responding

Be specific about both ends. For example: “Start timing after saying ‘sit down’ and stop when the student’s body touches the chair.” When every team member follows the same rule, the data is reliable.

Most programs need a latency criterion - a maximum acceptable response time (for example, “must begin within 5 seconds”). Without it, you can’t determine whether the delay is acceptable or needs intervention.

The Juggling Problem

Here’s the reality: accurate time-based recording requires your attention on a timer while you’re also managing a session, delivering instruction, and maintaining rapport with your student.

Traditional approaches - stopwatches, phone timers, clipboards - create cognitive load that compromises either data accuracy or session quality. Miss the start by a few seconds while prompting, and your data is off. Forget to stop the timer while managing behavior, and you’re guessing.

This is exactly why TallyFlex exists. Instead of juggling a stopwatch, you tap a tracker to start timing and tap again to stop. The app timestamps everything, notifies you if a timer is still running, and calculates totals automatically. It works offline on Chromebooks, tablets, and phones - so the environment never gets in the way of good data.

See our recording methods documentation for step-by-step setup instructions.

Common Mistakes to Avoid

Using Duration When Frequency Matters (or Vice Versa)

Mistake: Tracking tantrum duration when the real concern is how many times tantrums disrupt class.

Solution: Ask what you actually need to know. If the concern is how long episodes last, use duration. If the concern is how often they happen, use frequency.

Inconsistent Timing Across Team

Mistake: One staff member starts timing at the beginning of the instruction, another at the end. Same behavior, different data.

Solution: Make sure everyone on the team agrees on exactly when to start and stop the timer. Written definitions help - “start timing on the last word of the instruction, stop timing when the student’s hand touches the material.” When everyone follows the same rule, the data is consistent no matter who collects it.

Using Latency Without a Clear Antecedent

Mistake: Trying to measure “latency to engage in repetitive behavior” when there’s no specific trigger.

Solution: Latency requires a clear cue or instruction that the student should respond to. If a behavior happens on its own without a trigger, latency isn’t the right measure - consider frequency or duration instead.

What’s Next?

• Watch the tutorial: Duration & Latency Recording in TallyFlex - 4-minute walkthrough
• Compare all recording methods: How to Choose the Right ABA Recording Method
• Learn about interval recording: Partial vs. Whole Interval Recording Guide
• See recording methods in action: Recording Methods Documentation
• Start collecting time-based data: See how effortless timing data can be