An ABA scatterplot data sheet places a defined observation into repeated time cells so a reviewer can see whether event-positive cells cluster in particular periods. Its main value is discovery: the grid can point to times and routines that deserve closer direct observation. The pattern cannot explain by itself why behavior occurred.

Clinicians & ABA Professionals / Assessment and Treatment Planning.

This copyable version pairs the visual grid with a coverage log. That companion record matters because blank, missed and invalid cells are not the same as observed zeroes. The worksheet does not identify behavioral function, establish causation, rate clinical severity or select treatment.

Decide what one cell means

A mark is interpretable only when the observer knows its unit. A cell might show whether at least one event occurred, the event count, total duration or another plan-defined measure. Choose one unit for one grid, state it prominently and keep it stable. Occurrence and count should not share a grid because they answer different questions.

Scatterplot contractEntryClient or authorized case codeObservation purpose and questionResponsible clinicianObservable target definitionEvent start and stop ruleCell length and clock boundariesCell code or measurement unitPlanned dates, routines and settingsObserver and competence checkCommunication, assent, dissent and break signalsOrdinary supports and accessibilityMissing, invalid and partial-cell codesSafety and immediate stop conditionsSchedule, definition and form versionApproved storage and access rule

The BACB Ethics Codes hub links to the controlling professional materials. The current Ethics Code for Behavior Analysts, updated August 2024 when checked, addresses areas such as competence, assessment, confidentiality, consent, records, client welfare and data accuracy. It does not endorse this worksheet or make a visual pattern sufficient for a clinical decision.

Blank day-by-time scatterplot

Define the code before use. For a simple occurrence grid, suggested neutral entries are 1 for at least one directly observed target event, 0 for a fully observed cell with no target event, M for missing, P for partial and X for invalid under the predeclared rule. These symbols are examples, not universal clinical codes.

DateMorningMiddayAfternoonEveningDaily coverage note

Clock labels need exact boundaries. “Morning” could mean different spans to different observers, so the contract should state the start and end of each block. If a routine crosses a boundary, follow the stated rule rather than moving it to the cell that makes the display look clearer.

Companion coverage and context log

The grid is compact, which is useful for visual review but risky for context. The companion log keeps that compression from erasing observation quality. Complete one row for each planned cell, including valid zeroes and missing periods.

DateCell and exact timePlanned contextMinutes validly observedCell resultActivity and people presentSupports and accessClient communicationMissing, partial or safety note

An observed zero means the entire required cell met the observation rule and no defined event was seen. A missing cell means the cell was not observed. A partial cell needs its own rule; otherwise it should not be silently analyzed as if exposure matched a full cell.

Calculate coverage before concentration

For an occurrence grid, divide event-positive cells by valid observed cells within each block. Report the counts beside the percentage. Exclude missing or invalid cells from the denominator, but display them. If partial cells are eligible under the plan, say why and preserve their actual observation minutes.

Time blockPlanned cellsValid observed cellsEvent-positive cellsEvent-positive percentageMissing or invalid cellsMorningevent-positive / observed × 100Middayevent-positive / observed × 100Afternoonevent-positive / observed × 100Eveningevent-positive / observed × 100Overallevent-positive / observed × 100

Equal percentages can hide different amounts of evidence. Three positive cells out of four and thirty out of forty are both 75%, but they do not offer the same precision or contextual variety. A block can also appear high because observation occurred only during one recurring activity.

Read a temporal pattern as a question

The original scatterplot study described a visual method for making temporal response patterns identifiable and suggesting environmental features for closer examination (Touchette, MacDonald and Langer). A later evaluation of temporal distributions found no reliable patterns among its eligible scatterplots and stated that more detailed functional or descriptive analysis may be more precise and efficient in such cases (Kahng and colleagues). Together, these sources support a bounded use: the grid can generate a question, and a flat or inconsistent grid can also be informative.

Pattern-review questionEvidence and cautious answerWhich blocks have adequate planned and completed coverage?Are event-positive cells concentrated or broadly distributed?Do positive cells share an activity, person, setting or support condition?Do zero cells contain comparable exposure to those contexts?Are counts or duration hidden inside a binary positive cell?Could missing or partial cells create the apparent concentration?Did routines, medication, sleep, health or access change during the grid?Was the same definition and observer training used throughout?What client communication, assent or dissent was observed?What alternative explanations remain?What closer observation or other qualified review is proportionate?

Primary descriptive-analysis research also illustrates that natural-event observations require careful definitions and cannot by themselves control relevant variables (Vollmer and colleagues). A dense cluster should never be translated into “time of day caused the behavior.” The actual activity, access, demands, interaction, health, communication and other changing conditions may differ within the same clock block.

CASP's ABA Practice Guidelines access page lists Version 3.0 with a 2024 release date and explains its licensing terms. This article does not reproduce licensed material. Use current professional guidance, the treatment plan, payer terms, law and organizational policy directly when they apply.

Fictional completed example

Case W and the data below are fictional. The target is “hand contact with the side of the head, scored as event-positive when at least one directly observed occurrence happens within a valid cell.” The grid covers four dates and four scheduled blocks per date, for 16 planned cells.

Fourteen cells were fully observed. One cell was missed during a staff emergency, and one community-transition cell was invalid because the observer could not see the full interval. Neither was scored zero. Seven of the 14 valid cells were event-positive.

Fictional blockPlannedObservedPositiveCalculationMorning4411 / 4 = 25.0%Midday4433 / 4 = 75.0%Afternoon4322 / 3 = 66.7%Evening4311 / 3 = 33.3%Overall161477 / 14 = 50.0%

Midday and afternoon contain the visible concentration, but the grid does not show how many events occurred inside each positive cell. It also combines different activities and has one missing afternoon cell. The fictional reviewer writes: “Observed event-positive cells were more common in midday and afternoon blocks in this four-day sample. Broader direct observation should compare the activities, access, communication and health context within and outside those blocks.”

The reviewer does not assign function, change treatment or call 75.0% a threshold. The result is a prompt for more precise assessment, not a conclusion about time of day, risk or cause.

When a scatterplot misleads

Aggregation is the central tradeoff. A two-hour cell may hide the brief routine in which every event occurred, while a short cell may create sparse data. Binary marking hides counts and duration. A schedule change can alter what a clock block represents. Unequal observation can make one row look quiet simply because it was recorded less often.

Review the grid beside raw or narrative data rather than treating it as the primary record. Preserve changes in definitions and schedules. If the target is low frequency, highly variable or associated with serious safety concerns, the grid may be too coarse for the clinical question. A qualified clinician should select a method proportionate to the question and risk.

Version and correction record

DateSchedule or definition versionWhat changedReasonAffected cellsOriginal retained?Authorized byReanalysis needed?

A correction should retain the original value, corrected value, reason, author and date under the applicable record rules. Never convert missing cells to zero, shift an event into a neighboring block or change cell size after viewing the display without preserving the original analysis.

Privacy and record boundaries

The HHS Privacy Rule overview explains federal protections for protected health information in covered settings. The HHS Security Rule risk-analysis guidance describes risk analysis within the security framework. These pages do not approve a paper grid, spreadsheet, device, platform, access role, disclosure or retention schedule.

Keep the grid and companion observations only in an authorized record system. Confirm access, correction, signature, export, retention and disposal rules. This worksheet is not a treatment plan, clinical note, consent, incident report, authorization request, claim or complete assessment record.

Stop conditions and next routes

Pause when the event definition, cell boundaries, observer competence, client authorization, safety procedure or record destination is unclear. Stop data collection and use the applicable clinical or emergency route for immediate danger, acute health change, inaccessible communication, escalating distress or withdrawal of assent when applicable.

Do not provoke events, remove ordinary communication or supports, extend observation through a safety concern or interpret absence as proof that no concern exists. Route medical, mental-health, safeguarding and crisis needs to appropriately qualified professionals. A complete grid is never more important than immediate care.

Related resources

Sources