Trial-based functional analysis

Safety status: Use person-specific risk assessment, medical input when indicated, consent and assent when applicable, communication access, stop rules, trained staff, emergency planning, and exposure limits before experimental analysis.

Each trial contains a comparison

A trial usually pairs brief segments so background conditions remain similar. In a test segment, the clinician arranges the suspected establishing condition and makes a defined consequence contingent on the target response. In a control segment, the relevant reinforcer is available or the suspected establishing condition is absent.

The exact order, duration, conditions, response measure, and termination rules depend on the design. Published examples are study procedures rather than a universal template.

Trials fit inside natural routines

Short trials can occur during moments already suited to the comparison, such as instructional periods, leisure, or attention shifts. Embedding them can reduce the need for a separate assessment room and long uninterrupted sessions.

Natural settings add interference. A peer may deliver attention, a teacher may present a demand, or an ongoing intervention may suppress responding. Record invalid trials and repeat them only within the preapproved exposure and safety plan.

The classroom study shows both promise and limits

Bloom and colleagues studied ten students. Trial-based outcomes corresponded with more traditional functional-analysis outcomes in 6 of 10 cases and partially corresponded in a seventh. Modifying and repeating portions helped explain two noncorresponding cases.

Graduate students conducted the trials, which were spread across four to six days. The study also found that total active assessment time was similar to the standard analysis; the practical benefit came from brief embedding within classroom routines. These ten cases cannot establish a universal correspondence rate.

Training and integrity are central

Implementers need to recognize suitable opportunities, arrange test and control conditions, respond according to the protocol, record data, protect safety, and identify a compromised trial. A pyramidal-training study evaluated staff acquisition in simulated TBFA situations. Training findings in simulation never establish readiness for a particular client or high-risk case.

Measure procedural integrity by trial component. Separate correct antecedent arrangement, consequence delivery, timing, data recording, and stop-rule response. A single overall percentage can hide a critical safety error.

Keep paired segments comparable

Run paired segments close enough in time that staffing, activity, noise, medication, and other background conditions remain reasonably similar. Record the order and any event that changes the context. A test during a difficult transition and a control during a quiet leisure period create a confounded comparison.

Balance trials across relevant days and periods when the question requires it. A stack of convenient morning opportunities may miss an afternoon pattern. Report planned, completed, valid, repeated, interrupted, and safety-stopped trials separately.

Select TBFA for a defined reason

A school-setting review found growing use of adapted FA formats, including brief and trial-based analyses, while describing variation in terminology and implementation. Choose TBFA when short embedded comparisons fit the behavior, setting, resources, and decision.

Low-rate behavior, delayed effects, weak environmental control, serious risk, or unavailable trained staff may support another method. An inconclusive TBFA should remain inconclusive.

A fictional 24-trial record

Eli’s qualified team predeclares 24 paired trials across two suspected test conditions. Twenty trials meet every integrity and exposure rule; three are interrupted by routine events and one reaches a safety stop. Valid-trial coverage is 20 of 24, or 83.3%.

For one condition, the target response occurs in 7 of 10 valid test segments and 2 of 10 paired controls. Report those proportions by condition and segment type. The result still requires design-based interpretation, replication review, and comparison with other evidence. Interrupted and safety-stopped trials stay visible in burden reporting.

Protect the person during embedded assessment

Routine context never turns an experiment into ordinary care. Tell relevant people what will occur, obtain required approvals, and preserve the person’s way to pause, leave, ask for help, or withdraw when that right applies.

The functional-analysis risk paper provides a scholarly decision framework. Emergency and medical needs take priority over completing a trial. Food, water, bathroom use, mobility, AAC, prescribed care, pain care, and emergency help stay available.

Link the result to the next decision

Document which question the TBFA answered, which conditions were tested, untested alternatives, integrity, invalid trials, exposure, results, client experience, and next action. Assessment results inform treatment selection while leaving clinical judgment, consent, feasibility, and interdisciplinary input intact.

The BACB Test Content Outline includes functional analysis as examination content. The current BACB Ethics Code addresses competence, medical needs, assessment, consent, assent when applicable, risk, intervention, data, and evaluation for covered professionals.