An ABA preference assessment data sheet should preserve what was offered, what the person could access, how they communicated, and what happened on each trial. A ranking without that context can make a temporary choice look like a fixed trait. It can also make an inaccessible option look unwanted.

Clinicians & ABA Professionals / Assessment and Treatment Planning.

This worksheet provides recording panels for paired-stimulus and multiple-stimulus-without-replacement formats. It does not tell a clinician which format to use, whether an assessment is needed, or whether a selected item will function as a reinforcer. Those decisions require a qualified clinician, an individualized plan, current evidence, and the person's ongoing participation.

Define the assessment before presenting choices

The client should be able to nominate, reject, request, approach, disengage, or decline in the communication forms available to them. Do not limit the candidate set to what adults find convenient. At the same time, screen for health, allergy, swallowing, sensory, safety, cultural, religious, age, access, and feasibility concerns without treating those constraints as evidence of low preference.

Assessment identityEntryClient or authorized case code________________.Date, setting, time, and relevant routine________________.Responsible clinician and implementer________________.Assessment question________________.Selected format and clinical rationalePaired stimulus / MSWO / another defined format.Candidate-set version and how options were nominated________________.Communication modes and access supports________________.Selection definition________________.Engagement definition, if measured________________.Refusal or rejection definition________________.No-response definition________________.Exposure and position-balancing plan________________.Access duration or natural access arrangement________________.Ordinary access preserved________________.Consent, assent, dissent, and withdrawal plan________________.Health, allergy, swallowing, sensory, or safety screen________________.Stop and reassessment conditions________________.Record location and authorized access________________.

The BACB Ethics Codes hub identifies the controlling ethics materials. The current Ethics Code for Behavior Analysts, updated August 2024 when checked, addresses competence, confidentiality, assessment, behavior-change programs, consent, client welfare, and data accuracy. The code does not prescribe this worksheet or one preference-assessment procedure.

Paired-stimulus recording panel

Create the comparison schedule before the session. Balance left/right or other positions according to the defined procedure. Record refusals and no responses as their own events rather than assigning a selection that did not occur.

TrialLeft option and versionRight option and versionPosition checkFirst observable responseSelected optionRefusal or no responseAccess delivered as plannedContext or correction note1.2.3.4.5.6.

Check the schedule before starting. Make the options equally reachable. Record device or AAC access. Pause if communication fails.

For each option, calculate a selection percentage only when the presentation denominator is known:

selection percentage = selections of the option / eligible presentations of the option × 100

Do not place a no-response comparison into either option's selection count. If the procedure defines the presentation as eligible for both options, keep it in each presentation denominator and label the no response. If the trial was invalid because an option was missing, inaccessible, or presented incorrectly, exclude it under the predeclared rule and record why.

MSWO recording panel

In a multiple-stimulus-without-replacement arrangement, the chosen option is removed for the remainder of that array while unchosen options remain available under the selected procedure. Record the exact array and selection order. A later reader should not have to reconstruct which items were present from a final rank alone.

RoundCandidate-set versionOptions present in displayed orderSelection or no responseRank assignedAccess and engagement noteRefusal or dissentArray rearranged as definedContext note1.2.3.4.5.

Keep the array order visible. Record deviations immediately.

DeLeon and Iwata's comparison of multiple-stimulus formats is an original research source for MSWO methodology and its limits (1996 study). Later studies have evaluated direct assessment against informant ranking (Cote and colleagues) and activity-based MSWO predictions with individual differences (Daly and colleagues). These are specific studies with specific participants and procedures. They do not justify copying a research protocol into care without individualized clinical judgment, training, safeguards, and current requirements.

Interpretation worksheet

The calculation is only the beginning. Ask what the data can and cannot support.

Interpretation questionEvidence and answerWere all options familiar, perceivable, reachable, and usable?Did the person have an effective way to select, reject, pause, or request another option?Were position, exposure, magnitude, quality, and access reasonably comparable?Were ordinary access, food, water, movement, communication, comfort, and breaks protected?Did health, sensory, fatigue, medication, pain, hunger, recent access, or setting matter?Did the implementer follow the selected procedure?Are no responses, invalid trials, refusals, and exclusions visible?Is the hierarchy stable enough for the narrow question asked?What changed across sessions, people, settings, or candidate-set versions?What does the client or caregiver say about the result?Which options are only preference candidates, not demonstrated reinforcers?When should preference be checked again?What limited clinical decision, if any, is authorized now?

The CASP ABA Practice Guidelines access page identifies Version 3.0 and its licensing terms. This article does not reproduce the licensed guideline. Use the current guideline, plan, professional code, law, payer requirements, and organizational policies directly when they apply.

Fictional paired-stimulus example

The following example is invented. Case T and Items A, B, and C are placeholders. The options are described generically so no product, food, activity, or restriction is implied.

Assessment identityFictional entryQuestionDescribe selections among three client-nominated leisure options during one brief sessionFormatPaired stimulus, with a prewritten comparison and position scheduleCommunicationReach or eye-gaze selection; turning away or the documented stop signal records refusalCandidate setVersion 2; Items A, B, and C were accessible and familiar before trialsSafeguardsOrdinary access preserved; no food, hydration, AAC, comfort item, break, or basic need restrictedLimitsOne setting and one day; no reinforcer test; no generalization claim

TrialLeftRightRecorded resultNote1ABA selectedAccess delivered as defined2CAA selectedPosition of A changed3BCB selectedAccess delivered as defined4BANo responseKept separate from a selection5ACC not selected; A presentation invalidItem C was briefly out of reach, so this comparison was excluded under the predefined invalid-trial rule6CBStop signalSession ended; no selection assigned

For the fictional summary, the valid presentation counts preserved in the worksheet are A presented 3 times and selected 2 times, B presented 3 times and selected 1 time, and C presented 2 times and selected 0 times. Item A is 2 divided by 3 = 66.7%. Item B is 1 divided by 3 = 33.3%. Item C is 0 divided by 2 = 0.0%. One valid no-response trial is recorded separately, and one invalid comparison is excluded with its reason. The stop-signal trial ends the session and is not relabeled as a choice.

Those percentages describe the fictional worksheet's defined presentations; they are not universal preference cutoffs. Item C's nonselection does not prove dislike, consent to remove access, or lack of reinforcing value. Item A's higher selection percentage does not prove it will increase another behavior when delivered contingently.

The limited fictional decision is to keep all three options available, repair reachability before any reassessment, and ask the client and caregiver whether the candidate set still represents meaningful choices. No treatment, restriction, reinforcer, or purchasing decision follows from this session alone.

Candidate-set and version log

Changes to magnitude, quality, appearance, delivery person, access duration, or option membership can change the question being asked. Version the set rather than combining results as though the conditions were identical.

Version and dateAdded, removed, or changedClient or caregiver inputAccessibility or health reviewReasonData kept separate?Next check

Do not use deprivation of basic needs or broad restriction of ordinary preferred activities to make an assessment easier to interpret. Do not require compliance as the price of access to communication, hydration, toileting, movement, comfort, health support, or safety.

Privacy and record boundaries

The HHS Privacy Rule overview summarizes who the federal privacy framework covers and the health information it protects. The HHS Security Rule risk-analysis guidance describes risk analysis as foundational to security compliance. Neither source decides whether a particular form, video, device, platform, disclosure, or retention period is permitted.

Use an authorized record system and the minimum information needed for the task under applicable rules. A preference worksheet does not replace the assessment report, treatment plan, raw video or observation record, consent, clinical note, incident report, feeding or medical plan, authorization submission, or required electronic health record entry.

Stop conditions

Stop when the person declines, withdraws assent, uses the planned stop signal, shows unexpected distress, or cannot access the options or communication method as defined. Pause for a suspected allergy, swallowing risk, acute health change, unsafe item, inaccessible presentation, implementation outside competence, privacy concern, or candidate-set error. Use the actual clinical, medical, emergency, mandated-reporting, privacy, and organizational route that applies.

Do not complete missing selections from memory, assign a choice during no response, convert a refusal into a low rank, infer reinforcer function from preference alone, or use nonselection to justify restricting ordinary access. Preserve the record's uncertainty and obtain qualified clinical review before changing care.

Related resources

Sources