Variable-interval schedule

The interval changes from cycle to cycle

On a VI 5-minute schedule, the average interval is five minutes, but individual intervals might be two, four, five, six, and eight minutes. Their total is 25 minutes, so the mean is five.

After each interval ends, the first qualifying response produces reinforcement and typically starts the next interval. The exact list, order or sampling method, clock start, and reset rule belong in the schedule definition.

VI differs from fixed interval

A fixed-interval schedule uses the same interval every cycle. The VI arrangement changes the interval around an average. Both require an eligible response after the interval.

Baum’s molar account reviews ratio and interval definitions and common performance differences. It is conceptual and laboratory focused, not a clinical schedule recommendation.

VI differs from variable ratio and variable time

A variable-ratio schedule changes the number of responses required. Variable interval changes elapsed time before the next response can produce reinforcement.

Variable time changes delivery times but has no response requirement. Randomly sending a supportive message is variable time. Making the message available for the first completed check after changing intervals is variable interval.

Variability should be specified

“Random checks” is incomplete. Two lists can share a five-minute average while having very different ranges. A 4, 5, 6 list is more predictable than 1, 5, 9.

Document the interval values, frequencies, range, mean, selection method, and whether the sequence repeats. Check the realized intervals because software or implementation errors can change the programmed distribution.

A fictional VI record

Mei chooses brief inventory feedback after updates. The plan uses six intervals: 2, 4, 5, 5, 6, and 8 minutes. Their total is 30 minutes, so the programmed average is 5 minutes. The first accurate update after each interval is eligible.

Mei completes an eligible post-interval response in 5 of 6 intervals, and staff deliver feedback after all five. One interval ends after the activity is paused, so no eligible opportunity occurs before session close. Report 5/6 for intervals reaching the planned observation endpoint and 5/5 for delivered consequences due. These counts do not show a VI effect on accuracy or preference.

Response patterns are empirical outcomes

VI schedules often support steady responding in laboratory arrangements because the next availability time varies. Rules, clock cues, tasks, competing activities, and consequence quality can produce different patterns in people.

Staddon and Cerutti’s operant-conditioning review discusses interval, ratio, and concurrent schedules broadly. It does not establish that steady high responding is desirable in a clinical goal.

Integrity requires two timelines

Keep the programmed interval sequence beside the actual event log. Record interval start and end, every eligible response, consequence delivery and delay, pauses, resets, prompts, invalid intervals, and implementation errors.

Schedule integrity can be correctly delivered consequences divided by due consequences. Learner performance uses its own denominator. Avoid counting an interval with no possible response as a learner error unless the definition made it a complete eligible opportunity.

Clinical use should reduce uncertainty, not create strain

A VI arrangement may fit periodic checks or naturally variable feedback. Explain the rule in an accessible form. A self-monitoring tool can improve predictability when the person wants it.

Monitor accuracy, effort, latency, fatigue, schedule-correlated waiting, and satisfaction. If variability creates anxiety or confusion, another arrangement may fit better.

Food, water, bathroom access, AAC, mobility, prescribed care, rest, and emergency help remain available according to need. A schedule cannot place these behind elapsed time or a target response.

The BACB BCBA Test Content Outline, 6th edition includes basic schedules and schedule thinning in examination content. It does not prescribe interval values or a clinical plan.

Version every change

Changing the interval list, range, consequence, response quality rule, or timer cue creates a new condition. Mark its effective date and graph it separately. When thinning, use predefined decision rules and assess burden alongside response measures.

Construct interval lists carefully

Choose values that are feasible for the activity and person. Verify the arithmetic mean, range, and frequency of every value before use. Avoid accidental clusters that create a much leaner local period than the stated average suggests.

If software samples intervals, preserve the generated sequence or random seed when practical. Compare programmed and realized distributions after the session. Long pauses, session endings, and unavailable response opportunities can change what the person actually experienced.

Use enough completed intervals to describe performance. One six-interval session can verify implementation but rarely supports a stable conclusion about schedule effects.