Interval recording

Interval recording samples what happened across time

An observer divides a session into equal intervals, such as 30 seconds or one minute. Each interval receives a score under a written rule. Positive intervals divided by valid intervals gives the interval percentage.

This number describes sampled intervals. It does not provide the exact number of responses or total duration unless the observer also collects those dimensions. Two sessions can both show 50% while containing very different behavior.

Partial and whole interval use different rules

With partial-interval recording, an interval is positive when the response occurs at any point, even briefly. With whole-interval recording, the response must occur for the entire interval. Momentary time sampling checks only a scheduled instant and is a related procedure rather than the same rule.

A review of common discontinuous procedures explains these distinctions and the measurement error they can introduce. Partial interval can overstate total duration because one brief event makes the whole interval positive. Whole interval can understate duration because a short pause makes the interval negative.

Calculate the percentage with valid intervals

Suppose an observer plans 20 intervals and completes 18. A response occurs during any part of 12 valid intervals. The partial-interval result is 12 of 18, or 66.7%. Observation completion is 18 of 20, or 90%.

The two missed intervals remain missing. Turning them into negative scores would mix observer availability with the person's response. A report should give the numerator, denominator, interval length, procedure, and missed-check count.

A fictional family routine

Lila is a fictional eight-year-old who wants to help prepare a snack with her family. The team defines participation as handling an agreed ingredient or tool, communicating about the current step through speech or AAC, watching a demonstration, or using the chosen break routine.

During one 20-minute observation, participation appears in any part of 14 of 20 one-minute intervals, or 70% by partial interval. Direct duration shows seven minutes of participation, or 35% of observed time. Both calculations are correct under their definitions.

The difference helps the family understand the sampling effect. Several brief responses occurred across many intervals. The team keeps duration beside the partial-interval result and records two material delays. It avoids describing 70% as “70% of the session engaged.”

Interval length changes the estimate

A ten-second response can make one ten-second interval positive or one five-minute interval positive. Changing interval length changes the measurement system, even when the response stays the same.

Short intervals may produce a closer sample and add observer burden. Long intervals can magnify partial-interval overestimation and whole-interval underestimation. Practitioner guidance on discontinuous measurement recommends considering response dimensions, likely direction of change, and sampling error.

Mark a new procedure version when interval length, scoring rule, observation length, setting, or definition changes. Avoid drawing one uninterrupted trend line across incompatible versions.

Progress needs context beyond one percentage

Ask whether observations occur at comparable times, activities, and opportunity levels. A low percentage may reflect missing materials, pain, fatigue, unavailable communication, a delayed partner, or a chosen break. A higher percentage may come from a shorter observation or easier activity.

Pair interval results with the person's experience, goal progress, health and safety information, family priorities, direct count or duration when useful, and treatment-fidelity data. One sampling metric cannot represent overall quality of life or prove that treatment caused a change.

Protect access during observation

Keep AAC, interpreters, food, water, bathroom access, mobility and sensory supports, prescribed care, breaks, and emergency help available. Record assent, dissent, and chosen alternatives in a way the person can use.

Environmental failures deserve their own field. If the device is uncharged or the activity never begins, the record should identify the system barrier rather than turn it into a client failure.

Questions families can ask

Useful questions include:

• Which interval rule is being used and why?
• What exactly earns a positive score?
• How long are the intervals and observation?
• How are missed intervals and breaks handled?
• Are the same settings and opportunities represented over time?
• What direct measure or lived-experience information accompanies the percentage?

The ABAI basic-principles page places measurement within behavior-analysis training. It does not select a procedure for an individual family.

Ask for the graph label and procedure version in writing. A family should be able to tell whether a point means partial intervals, whole intervals, or momentary checks and whether the observation opportunity stayed comparable.