An ABA ABC data sheet preserves a direct observation in sequence: what was happening immediately before a defined event, what the person did, and what happened immediately afterward. The useful version also records when observation was planned but did not happen. Otherwise, a collection of memorable event narratives can look far more representative than it is.

Clinicians & ABA Professionals / Assessment and Treatment Planning.

The copyable worksheet below keeps raw narrative, later coding, exposure and interpretation in separate places. It can help a qualified clinician organize descriptive evidence. It does not identify behavioral function, prove that an antecedent or consequence caused behavior, select an intervention or replace a broader assessment.

What this worksheet can answer

The event rows answer a narrow factual question: during a defined observation, what did the observer see and hear around the target event? The coverage table answers a different question: which planned contexts were actually observed, including periods when the target did not occur? Only after those two views are aligned does the review panel help a clinician describe possible patterns and gaps without turning them into causal findings.

Descriptive observation has a real but limited role. In one primary study, researchers evaluated training for narrative ABC assessment and measured the accuracy of recording antecedents, behavior and consequences under the study's definitions (Peters and colleagues). A structured descriptive methodology has also been evaluated under defined observation and coding procedures (Anderson and Long). Those arrangements do not make an untrained observer, an undefined form or an incomplete sample clinically interpretable.

Define the observation before opening the form

Write definitions that another trained observer could apply without having to infer intent. A label such as “upset,” “noncompliant” or “seeking attention” is an interpretation, not an observable target. Describe movement, speech, contact, duration or another plan-defined event. Define when one event ends and another begins. Identify how communication, AAC, breaks, pain signs and accessibility needs will be represented.

Observation contractEntryClient or authorized case codeObservation purpose and clinical questionResponsible clinicianTarget event and observable definitionStart and stop rule for one eventAntecedent observation window or ruleConsequence observation window or rulePlanned settings, routines, people and timesScheduled sampling unitSupports and ordinary access kept in placeCommunication, assent, dissent and break signalsHealth, sensory, medication or setting factors to route, not interpretObserver training and competence checkSafety and immediate stop conditionsForm, codebook and schedule versionApproved record location and access rule

The current BACB Ethics Codes hub points to the controlling materials. The linked Ethics Code for Behavior Analysts, updated August 2024 when checked, addresses areas including competence, records, confidentiality, assessment, consent, data accuracy and client welfare. A worksheet cannot decide whether a particular assessment or service meets those obligations.

Blank narrative ABC event record

Record direct observation first, using plain, time-ordered language. If the observer did not see what occurred before the event, mark it unobserved instead of reconstructing it later. Keep setting or health context apart from the immediate antecedent, and leave assumed function out of the consequence column.

EventDate and exact timeSetting and activityImmediate antecedent observedObservable behaviorImmediate consequence observedClient communication, assent or dissentSupports availableObserverMissing, uncertain or safety note123456

One row should describe one event sequence under the stated definition. If several responses occur in quick succession, apply the predeclared start and stop rule. Do not split or combine events after seeing which version produces a cleaner pattern.

Scheduled observation coverage

An event-only ABC log cannot show how often an antecedent, consequence or target was absent. Add planned windows that cover relevant routines, including ordinary periods when no target event occurs. A missed window is missing data, not a zero.

Planned windowDate and timeIntended contextObserved as scheduled?Valid observation minutesTarget occurred?Relevant exposure present?Reason missing or invalid12345678

Report completed valid windows divided by planned windows before reviewing a pattern. Then describe how observation was distributed across times, activities, people and settings. A high percentage can still be unrepresentative if every completed window comes from one routine.

Code the narrative without erasing it

Codes make repeated events easier to summarize, but coding should not replace the source narrative. Keep the original entry, codebook version, coder and coding date. When a category changes, preserve the earlier version and state whether older rows were recoded.

CodeObservable categoryIncludeExcludeAmbiguous examples and escalation routeA1A2B1C1C2

Research comparing antecedent and consequent predictors in descriptive records demonstrates why simple co-occurrence needs careful denominators and comparison with stronger evidence (Camp and colleagues). Another empirical comparison explains that functional analysis, under its study framework, permits function identification in a way descriptive methods do not (Thompson and Iwata). The practical boundary is plain: code what was observed, describe associations cautiously and leave causal conclusions to qualified assessment using appropriate methods.

Descriptive pattern review

Use this panel after checking observation quality. A reviewer may identify a pattern worth testing or a context that needs more observation. “Often followed,” “concentrated in the observed sample” and “not enough exposure data” are more accurate than saying an event “caused” behavior or “is the function.”

Review itemEvidence and bounded interpretationPlanned, completed, invalid and missed windowsDistribution across relevant contextsTarget events and valid observation timeAntecedent categories before eventsConsequence categories after eventsAntecedent exposure when the target did not occurConsequence occurrence outside event sequencesCommunication, assent, dissent and access observationsHealth, sleep, medication, pain or sensory information needing another routeObserver agreement or training evidence availableAlternative explanations and sources of reactivityMissing, ambiguous or changed definitionsQualified next review step, owner and date

The CASP ABA Practice Guidelines access page identifies Version 3.0, released in 2024, and its licensing conditions. This article links to that access page without reproducing licensed guidance. Consult the current licensed material, treatment plan, law, payer terms and local policy directly when they apply.

Fictional completed example

Case V and every detail here are invented. The target is “forceful contact between an open palm and a tabletop, audible to the observer, counted as a new event after three event-free seconds.” Eight 10-minute windows were planned across ordinary routines. Seven occurred; one was missed because the observer was unavailable. The missed cell was not scored as no behavior.

Five target events were recorded in the completed windows. A task request was present during four observed windows, and at least one target event occurred in three of those four. The descriptive proportion is 3 divided by 4 = 75.0%. Among three observed windows without a task request, an event occurred in two. That result is 2 divided by 3 = 66.7% when rounded to one decimal place.

Context in the fictional sampleEvent-positive windowsObserved windowsDescriptive resultTask request present343 / 4 = 75.0%Task request absent232 / 3 = 66.7%All completed windows575 / 7 = 71.4%

Adult attention was recorded after four of the five individual target events, so the event-record proportion is 4 divided by 5 = 80.0%. The form did not measure how often adult attention occurred when the target was absent. That missing background exposure prevents a meaningful contingency comparison. It would be inaccurate to call attention the function or recommend withholding attention.

The fictional reviewer records a modest conclusion. This small sample contains target events both with and without task requests, the observed proportions are close, the consequence denominator is incomplete and one planned context is missing. Broader, representative direct observation with clearer exposure coding and qualified review is warranted before any stronger interpretation. The worksheet does not prescribe treatment or a functional analysis.

Observation quality and corrections

Before relying on the summary, ask whether the observer could see and hear the full sequence, whether recording changed the routine, whether definitions were applied consistently, and whether ordinary communication and supports remained available. When feasible and clinically appropriate, obtain agreement evidence under a defined method rather than relying on a general impression that observers “agreed.”

DateForm or codebook versionOriginal entryCorrected entryReason and evidenceAuthorEffect on summaryReviewer

Preserve the original entry and correction trail under the applicable record rules. Never backfill an unseen antecedent, delete a missed window or relabel a consequence to strengthen a preferred explanation.

Privacy and record boundaries

The HHS Privacy Rule overview describes federal protections for protected health information and minimum-necessary concepts where applicable. The HHS Security Rule risk-analysis guidance describes risk analysis as foundational to security compliance. Neither page approves this form, a spreadsheet, a personal device, an access role, a disclosure, a retention schedule or a disposal method.

Use only an authorized record system. Confirm how narrative observations, corrections, signatures, exports and access are handled. This working sheet does not replace the treatment plan, clinical note, consent, incident report, medical record, authorization submission, claim or other required record.

Stop and route concerns

Pause observation when the target definition, observation window, responsible clinician, client authorization, safety plan, recording method or record destination is unclear. Stop the observation and use the applicable clinical or emergency route for an immediate safety concern, acute health change, inaccessible communication, escalating distress or withdrawal of assent when assent applies. Data collection never outranks a person's immediate care.

Do not continue merely to complete a row. Do not remove ordinary supports to create an event, provoke a target, run unapproved test conditions or treat absence during a short sample as proof that concern is resolved. Route medical, mental-health, safeguarding and crisis concerns to appropriately qualified professionals.

Related resources

Sources