Partial-interval recording

Any occurrence makes the interval positive

Suppose a ten-minute observation uses ten one-minute intervals. If the response occurs for two seconds in the first interval, that interval receives a positive mark. If it occurs 12 times in the second interval, the second interval also receives one positive mark. The record retains whether any occurrence happened and discards the difference in count and duration.

That compression can reduce recording burden. It also limits interpretation.

Define intervals and events in advance

Write down:

• the operational definition and one event boundary
• observation start and stop events
• interval length and timer source
• whether recording happens during or immediately after each interval
• eligible intervals and missed-observation rules
• prompts, ordinary supports, and relevant setting conditions

If an event crosses an interval boundary, both intervals may score positive because the response occurs in each. State that rule before collection.

Calculate the percentage with raw counts

If the response occurs in 7 of 12 valid intervals, the result is 7 divided by 12 times 100, or 58.3%. Report “positive in 7 of 12 partial intervals” rather than “occurred 58.3% of the time.”

The numerator counts intervals, not responses. The denominator counts valid observed intervals, not minutes unless each interval is one minute and the reader still understands the scoring rule.

A fictional work-routine example

Esme is a fictional adult who wants fewer interruptions from an unreliable inventory alert while working at a garden shop. The team defines an alert-related interruption as stopping the current stocking step and checking the alert screen before the step is complete. Staff also record actual alert failures as an environmental measure.

During a 24-minute observation with two-minute intervals, interruptions occur in 7 of 12 intervals, or 58.3%. They occupy a total of four minutes based on a separate duration record, or 4 of 24 minutes, or 16.7% of observed time.

Both statements can be true. The partial-interval percentage is much larger than the observed-time percentage because any occurrence marks the whole interval positive. The team also finds five false alerts. The example supports system review and no causal conclusion about Esme.

Partial-interval results can overstate time occupied

When a reader interprets positive intervals as duration, partial-interval recording tends to overstate the portion of time occupied. Longer intervals give more opportunity for at least one occurrence and can increase positive scores.

The size and direction of error depend on response pattern, duration, interval length, and observation schedule. Partial-interval results from different interval lengths should not be treated as equivalent.

Use it for a suitable measurement question

Partial-interval recording may fit when observers need a feasible sample of whether behavior appeared across periods and exact count or duration is impractical. It can be useful for screening patterns across activities or testing whether stronger measurement is needed.

Choose event recording when exact discrete count matters and can be captured. Choose duration when time occupied matters. Choose momentary time sampling when state at scheduled instants answers the question. Select the least burdensome method that still provides adequate evidence for the decision.

Keep missing intervals visible

An observer may miss an interval while supporting safety, responding to another person, or handling a system problem. A missed interval should not silently become negative.

Report valid intervals divided by planned intervals. If 12 of 15 planned intervals were observed, coverage is 12 of 15, or 80%. Then calculate the partial-interval result using the 12 valid intervals and report three missed intervals by reason.

Protect access and client meaning

Keep communication and AAC tools, interpreters, mobility and sensory supports, breaks, food, water, bathroom access, prescribed care, and emergency help available. Record barriers and partner responses separately from the client measure.

Ask whether the measured outcome matters to the person. A lower partial-interval percentage may coexist with greater effort, less comfort, or reduced communication. Pair the sample with the person’s report and other relevant outcomes.

Check reliability and procedural integrity

Useful checks include observed intervals divided by planned intervals, intervals scored under the current definition, timer failures, excluded intervals, and sampled observer agreement. Also check whether observers looked continuously enough to detect any occurrence. A partial-interval score cannot be reconstructed from a quick glance at the endpoint.

The BACB BCBA Test Content Outline, Sixth Edition covers discontinuous measurement, interval recording, time sampling, validity, reliability, and representative measurement selection. It is examination content and supplies no universal interval length.