Event recording

Event recording counts defined responses

The observer adds one tally whenever a response meets the operational definition. A counter, paper form, timestamped app, or automated system can hold the record. The method is continuous when the observer is available to detect every event throughout the observation.

Suitable examples may include completed requests, initiated transitions, submitted work products, safety checks, or staff-delivered feedback steps. The response should be meaningful to the person and connected to the decision the data will support.

Define one event before counting

A usable definition answers:

• What visible or audible change starts an event?
• What marks its end?
• How are rapid repetitions separated?
• Do different communication forms count as the same response class?
• Which prompts, settings, people, and opportunities are included?
• What happens when observation is interrupted?

Suppose a person selects “break” three times before a partner can respond. The team might define that burst as one request until the partner answers or ten seconds pass. Another goal could require each selection to count. Either rule can be applied consistently when it is written before observation.

Use rate when exposure time changes

Count is easiest to interpret when observation length is consistent. For unequal sessions, calculate rate = count divided by observed time.

Six events in 15 minutes equal 0.4 per minute. Twelve in 60 minutes equal 0.2 per minute. The second session has twice the count and half the rate. Report count, observed time, and rate so the calculation stays auditable.

Opportunity may also limit exposure. A person can return a shared item only when someone has offered it. Report successful events divided by eligible opportunities when that question matters.

A fictional clinical example

Rowan is a fictional 15-year-old who wants smoother equipment handoffs during a volunteer shift. With Rowan, the clinician defines one completed handoff as placing an item in the labeled return area or giving it to the assigned partner within two minutes of finishing its use. A gesture or AAC message requesting help also completes the routine when a barrier is present.

Across a 60-minute shift, 14 eligible returns arise. Rowan completes 11 independently, requests help for two barriers, and leaves one item at the workstation. Independent event count is 11. Completed routines including accessible help are 13 of 14, or 92.9%.

The count of 11 and opportunity result of 13 of 14 answer different questions. The record also notes that the labeled return area was moved once. These observations can guide discussion and do not show that any intervention caused performance.

Event recording has clear limits

Counting can become unreliable when responses happen too rapidly, overlap, last for long periods, or have unclear boundaries. A count of two episodes says little if one lasts five seconds and the other lasts 40 minutes. Duration may carry the critical information.

Event recording can also undercount when an observer teaches, supports safety, or watches several people. Automated detection may misclassify events or lose context. Validate the system against human review and keep failures visible.

Preserve access and consent

Observation should leave augmentative and alternative communication, interpreters, mobility and sensory supports, breaks, health care, and emergency help available. A missing communication device or inaccessible task is a system condition, rather than a client nonresponse.

If video, audio, telehealth, or sensors are used, follow applicable consent, assent, privacy, security, access, retention, and deletion requirements. Explain what is captured and how the person can decline or withdraw when that right applies.

Review recording quality

Useful checks include observed minutes divided by planned minutes, events with a current definition, observer agreement across relevant conditions, missed-observation time, device failures, and corrections with preserved history.

When observer agreement influences treatment, safety, supervision, or mastery decisions, calibrate observers on examples and nonexamples. High agreement can show consistent counting. It cannot prove that the chosen event represents the outcome the person values.

The current BACB BCBA Test Content Outline, Sixth Edition includes measuring occurrence, temporal dimensions, reliability, validity, and representative procedure selection. It is certification-examination content and supplies no case-specific authority.

Ask before choosing event recording

Clarify whether the decision needs exact count, rate, opportunity percentage, duration, response timing, or product quality. Ask how feasible continuous observation is in the actual setting. Record the method version, interval or observation boundaries, supports, exclusions, and effective date.

Keep the client’s own report beside the count. A numerical increase can still feel burdensome, and a stable number can occur while access or independence improves.