Trial-Based Functional Analysis (TBFA): A Practitioner's Guide

Origins: Sigafoos and Saggers (1995) and the Bloom (2011) Modernization

The foundational argument for TBFA is methodological: if the same establishing operation and consequence arrangement that defines a traditional FA test condition can be embedded in a brief naturally occurring moment — a demand placed during instruction, attention withheld for sixty seconds before a trial, a preferred item present versus absent — then the analytic logic of the FA is preserved while the session-interruption burden is eliminated. Sigafoos and Saggers demonstrated this proof of concept with two boys with autism displaying aggression, running discrete trial pairs across attention and escape conditions within the school day (Jessel et al., 2024). Bloom and colleagues' 2011 landmark classroom application established the procedural template still in use: 2-minute control and test segments arranged in pairs across attention, escape, tangible, and alone conditions, embedded throughout the school day, with teachers or trained paraprofessionals running the trials under BCBA consultation (Togashi, 2025).

The modernization brought two contributions beyond the Sigafoos and Saggers proof of concept. First, it anchored the differentiation criterion: the condition in which the percentage of test trials with problem behavior substantially exceeded the percentage of control trials — particularly when a clear reversal existed — was treated as the maintaining function. Second, it demonstrated correspondence between TBFA outcomes and traditional analog FA outcomes, which is the validity claim that subsequent systematic reviews and comparative studies have been building on ever since (Togashi, 2025).

The Standard Procedure: Identify, Arrange, Embed, Repeat, Analyze

The TBFA procedure follows a reproducible five-step architecture.

Step 1: Identify the target behavior and hypothesize candidate functions. The practitioner reviews records, conducts a functional behavior assessment interview (e.g., FAST, MAS, open-ended interview), and identifies the one to three most plausible maintaining functions. This step constrains which TBFA conditions to run; a behavior with no prior evidence of tangible function does not require a tangible condition.

Step 2: Arrange control and test trials for each condition. For each candidate function, the practitioner scripts two trial types:

• Control trial (approximately 1 minute): The reinforcer associated with that function is freely available (attention delivered noncontingently, no demands placed, preferred items present and accessible). The target behavior is ignored or prompted to appropriate alternatives; no contingent delivery of the functional reinforcer occurs.
• Test trial (approximately 1 minute): The establishing operation is introduced (attention withdrawn, demand initiated, preferred item removed or restricted), and any occurrence of the target behavior produces brief immediate access to the hypothesized reinforcer (attention delivered contingently, demand removed, item returned). Sessions of 5 minutes total — one 2-minute control, a brief neutral inter-trial interval, one 2-minute test — are the most commonly published format, though 1-minute segments with a 30-second inter-trial interval appear in school-embedded applications (Nesselrode et al., 2022).

Step 3: Embed trials across conditions within the natural routine. Trials are distributed across the day: an attention-condition pair during independent work time, an escape-condition pair during a group instruction activity, a tangible-condition pair during a preferred activity transition. Embedding rather than massing trials prevents sequential condition effects and keeps assessment invisible from the broader classroom perspective.

Step 4: Repeat until differentiation criteria are met. The minimum is five paired trials per condition. Saini and colleagues' quantitative review of FA efficiency demonstrated that a binomial stopping rule — five consecutive test trials with problem behavior occurring in the predicted direction (e.g., 5 out of 5 escape test trials positive) yields statistical differentiation at p = .03 — allows early termination when evidence accumulates rapidly, without sacrificing internal validity (Saini et al., 2020). In practice, most published TBFA studies use five to ten pairs per condition before applying visual analysis.

Step 5: Analyze by visual inspection of condition-level data. The practitioner constructs a graph with percentage of test trials positive on the y-axis and session blocks on the x-axis, one data path per condition. The differentiation criterion is a stable elevated pattern in one condition (consistently higher percentage of test-positive trials than all other conditions) with near-zero or clearly lower rates in the control trials for that condition. Standish and colleagues formalized an ongoing visual inspection (OVI) protocol — the Trial-Based Ongoing Visual Inspection (TB-OVI) rules — that caregivers and frontline staff can apply in real time during the TBFA to make go/no-go decisions without waiting for post-session graphing, enabling faster transitions to treatment (Standish et al., 2021).

Embedded TBFA versus Scheduled TBFA

Two operationally distinct variants appear in the literature, and the choice between them affects procedural integrity demands.

Embedded TBFA inserts trials opportunistically within naturally occurring transitions or activities. The teacher responds to a student raising their hand as the start of an attention-condition test trial; the moment a demand is placed is the start of an escape-condition test trial. Because the antecedents and consequences occur within the stream of routine classroom events, the format produces high ecological validity and low disruption, but also requires the implementer to maintain high procedural readiness at all times and to recognize environmental conditions that naturally approximate the required trial parameters. Nesselrode and colleagues' systematic review of FA in public-school settings concluded that embedded TBFA is the most classroom-feasible FA variant available, consistently identifying socially mediated attention, escape, and tangible functions without the false automatic-reinforcement results that plague assessments using alone conditions (Nesselrode et al., 2022).

Scheduled TBFA pre-books specific intervals in the day — typically during activities where the target function is most plausible — and the implementer deliberately initiates the control and test trial sequences at those designated times. The format sacrifices some ecological naturalness in exchange for greater procedural control and lower implementer burden between scheduled intervals. Lambert and colleagues' teacher-conducted latency-based variant is a scheduled TBFA: a first-year teacher conducted 5-minute test and control blocks at designated academic periods and recorded latency to first problem behavior rather than occurrence, enabling function identification within existing classroom structures without full-session interruption (Lambert et al., 2017).

The choice between embedded and scheduled TBFA is typically determined by the school's daily structure, the implementer's experience level, and the target behavior's frequency. Low-rate behaviors benefit from scheduled TBFA because the establishing operation is deliberately arranged; high-rate behaviors that occur naturally across many activities fit embedded TBFA more efficiently.

Training Classroom Staff: BST, Automated Modules, and Pyramidal Models

The largest body of TBFA-specific research after validity studies concerns how to train implementers efficiently. Three models dominate the corpus.

Behavioral Skills Training (BST) with telehealth delivery. Togashi's 2025 study evaluated a combined training sequence — asynchronous computer-based instruction (CBI) followed by live BST via telehealth — for Japanese practitioners with limited access to in-person TBFA training. CBI alone significantly improved knowledge (Z = −2.670, p = .008, r = .89) but was insufficient for accurate implementation; adding BST via telehealth brought most participants to mastery integrity in one or two practice trials. The finding that knowledge acquisition and implementation fidelity are dissociable has direct implications for programs that rely on readings or videos alone to train TBFA (Togashi, 2025).

Self-instruction plus small-group rehearsal. Griffith and colleagues (2020) trained three ABA-naive undergraduates to 100% procedural fidelity across all four TBFA trial types using a self-instruction manual plus small-group rehearsal with a trained confederate acting as the client. Total training time averaged 2 hours 20 minutes (range: 1.5 to 3 hours). The study is notable for showing that the response-effort barrier for TBFA training is substantially lower than for traditional analog FA training — a practitioner with no prior ABA coursework reached mastery within one half-day session (Griffith et al., 2020).

Partially automated training modules. Standish and colleagues (2023) developed and validated a partially automated training package that taught university students not only to implement TBFA trials but also to summarize data and make accurate visual-inspection decisions — the downstream skills that translate trial data into functional conclusions. The automated modules are freely available online and produced the same high procedural-fidelity outcomes as live training for both implementation and graphing/interpretation accuracy (Standish et al., 2023).

Pyramidal training model. Alnemary and colleagues (2017) replicated a pyramidal training model (PTM) in which four behavioral consultants were trained in a group format, then each trained one behavior technician. All eight participants achieved correct TBFA implementation and data collection in simulated settings, with generalization probes to novel client scenarios. The PTM is particularly efficient for agencies that need to scale TBFA skills across multiple service sites without requiring expert oversight for every trainee (Alnemary et al., 2017).

The convergent message from the training literature: BST with live practice is required for implementation fidelity — knowledge-only training is insufficient. Automated or self-instruction modules efficiently deliver the knowledge phase, but must be paired with role-play rehearsal and feedback to reach mastery.

Treatment Utility: TBFA Findings to Function-Matched Intervention

TBFA is assessed, informed, and justified by what it produces for treatment, not merely by its correspondence with other FA formats. The treatment-utility evidence is largely positive but contains an important nuance from the feeding-behavior literature.

Andersen and colleagues (2022) conducted a direct comparison in which TBFA took 71% less time than extended analog FA for children with inappropriate mealtime behavior. Despite only 29% exact correspondence between TBFA-identified functions and extended FA-identified functions, treatments derived from the TBFA results improved mealtime behavior in 80% of cases. The authors interpreted the finding as supporting TBFA's clinical utility: even when the format does not perfectly replicate a longer FA's functional conclusions, it produces treatments that work (Andersen et al., 2022).

Weyman and colleagues (2022) demonstrated TBFA treatment utility in a more specific application: 5-minute control versus test trials contrasting free ritual access against ritual interruption (see An application of the trial-based functional analysis to assess problem behavior evoked by ritual interruption) identified a ritual-interruption function in four children with autism, and subsequent functional communication training with a multiple schedule eliminated problem behavior to near-zero levels with generalization to untrained rituals (Weyman et al., 2022).

The exception is Call and colleagues' (2024) randomized comparative effectiveness trial — the highest-evidence study in this corpus — which compared behavioral outcomes in 57 preschoolers with autism randomly assigned to FBA with or without TBFA. Both groups achieved equivalent functional communication training success, leading the authors to conclude that TBFA adds only marginal incremental benefit over non-experimental FBA for selecting interventions (Call et al., 2024). This finding does not invalidate TBFA's utility; it suggests that the primary benefit of TBFA may be treatment efficiency and assessment confidence rather than a large uplift in treatment effectiveness over well-conducted descriptive FBA.

The Henry and colleagues (2021) integration of brief-to-extended analysis sequences addresses cases where TBFA alone produces undifferentiated results: starting with brief trial-based conditions and escalating to extended analyses with confirmation and annulment phases only when needed conserves resources for cases where TBFA resolves quickly while preserving capacity for complex presentations (Henry et al., 2021).

Telehealth-Supported TBFA

Telehealth delivery of TBFA has been validated in two directions: training practitioners remotely, and coaching parents to implement TBFA with their own children at home.

Davis and colleagues (2023) trained three parents of children with autism (ages 6 to 8) to conduct TBFA via synchronous telehealth sessions using bug-in-the-ear coaching. TBFA data paths showed clear differentiation of function, and subsequent function-based treatments produced meaningful reductions in problem behavior and increases in appropriate communication. The authors recommend at least 5 to 7 trials per condition when caregivers rather than therapists are controlling sessions, to account for the higher variability in trial implementation by less-trained implementers (Davis et al., 2023).

Togashi's (2025) telehealth BST model demonstrates that remote training for practitioners — not just coaching for parents — can bring implementers to mastery in one or two practice trials when structured BST follows asynchronous knowledge acquisition, even across geographic and cultural contexts where in-person TBFA supervision is scarce (Togashi, 2025).

The combination of remote training and remote coaching has particular salience for consultative BCBAs who serve schools at a distance: the BCBA can train the teacher or paraprofessional via asynchronous modules and telehealth BST, then coach the implementation remotely using the same video-session infrastructure, without requiring any on-site presence.

Caregiver-Implemented TBFA

Several studies extend TBFA to caregivers in home settings, moving the assessment format outside institutional contexts entirely. Gerow and colleagues (2020) trained three mothers to conduct brief functional analyses at home with their 2-to-4-year-old toddlers; all three parent-implemented analyses yielded interpretable results and informed effective function-based interventions in approximately 2.4 to 3.5 hours total — substantially less than a clinic-based evaluation would require (Gerow et al., 2020). Germansky and colleagues' (2020) systematic review of caregiver-implemented functional analyses identified one TBFA study (Gerow et al., 2019) demonstrating parent-conducted embedded trial analysis with young autistic children, confirming that the brief trial format is viable for caregiver-led assessment when implementation training is provided (Germansky et al., 2020).

Standish and colleagues' (2021) TB-OVI rules were specifically validated with five caregivers (ages 3 to 10, home and community settings), enabling real-time functional decisions during trials rather than waiting for BCBA data review. All five caregiver-implemented TBFAs using TB-OVI guidelines produced effective function-based interventions (Standish et al., 2021).

These three experimental functional analysis formats address the same fundamental question — what maintains this problem behavior? — but differ on five dimensions that should drive the practitioner's selection decision.

Session structure. Traditional FA (Iwata et al., 1982/1994) uses 5-to-15-minute alternating multielement conditions (attention, escape, tangible, alone, play control) in a dedicated analog space, cycling multiple times until stable responding emerges. TBFA embeds paired 1-to-2-minute control/test trials in natural routines, eliminating the dedicated space requirement. The IISCA (Hanley et al., 2014) uses a single synthesized test condition derived from an open-ended interview, compared against a matched control, typically in a clinic room for original and performance-based variants, or embedded for the trial-based IISCA variant.

Setting requirements. Traditional FA requires a controlled analog environment where conditions can be arranged and terminated cleanly; conducting it in a live classroom with other students present is rarely feasible. TBFA was designed specifically for natural settings — classrooms, homes, community environments — with no additional space required. The IISCA is typically clinic-based in its original and performance-based forms; the trial-based IISCA variant extends it to natural routines in the same way TBFA does.

Multiply controlled behavior. Traditional FA is the strongest method for identifying multiple functions because its independent conditions can each show elevation. TBFA's pairwise comparison structure identifies the dominant function efficiently but may miss secondary functions, particularly when two functions are both moderately strong; the practitioner may see escape elevation but miss an additional attention component that the escape condition's data paths do not reveal (Amador et al., 2024). The IISCA addresses multiply controlled behavior through synthesis — it tests all functions together — but creates false-positive risk when the synthesized contingency includes a reinforcer that is not actually functional. Brown and colleagues' (2025) FA decision-making review identified TBFA as a viable school option that reliably captures socially mediated functions, with the explicit caveat that brief formats carry lower correspondence with extended FA than longer assessments (Brown et al., 2025).

Rare and low-frequency behavior. Traditional FA's longer session durations give low-frequency behavior more opportunities to occur across conditions. TBFA's 1-to-2-minute windows may not sample enough behavior for a stable estimate in very low-frequency cases; the latency-based TBFA variant addresses this by recording latency to first response rather than frequency within the trial, making even one occurrence per trial sufficient for differentiation (Lambert et al., 2017).

Automatically reinforced behavior. The systematic review evidence shows TBFA rarely identifies automatic reinforcement (Nesselrode et al., 2022). This is partly a feature — embedded trials in active routines where social reinforcement is ambient tend not to produce the isolation needed to unmask automatic functions — and partly a limitation. For behavior suspected to be automatically reinforced (high-rate stereotypy, SIB in alone conditions), traditional FA with an alone or ignore condition remains the appropriate choice.

Decision logic summary:

• Use TBFA when: setting is a school or community environment, behavior is socially mediated, BCBA has only consultative access, implementer training capacity is limited.
• Use traditional FA when: behavior may be automatically reinforced, multiply controlled functions are suspected, analog environment is available and high procedural control is warranted, severity is high and extended reversal control is needed.
• Use IISCA when: open-ended interview produces a clear synthesized hypothesis, socially mediated function is suspected, case benefits from a single synthesized test condition that reduces evocative exposure time, clinic access is available.

The TBFA literature distributes across three evidence tiers: a systematic review of 30 years of research, numerous single-subject experimental designs, and methodology papers on training and decision rules.

Systematic review. Amador and colleagues (2024) conducted a PRISMA-style systematic review of 61 TBFA empirical articles covering 344 participants (ages 2 to 49) from 1991 to 2021. The review documented wide use across populations and settings but identified concerning methodological gaps: only 43% of studies reported inter-observer agreement (IOA), and few validated TBFA-identified functions against a comparison FA. The authors concluded that the TBFA literature needs standardized reporting before published procedures can be trusted for high-stakes clinical decision-making (Amador et al., 2024). Nesselrode and colleagues' (2022) systematic review of FA in public schools coded 58 empirical studies and concluded that TBFA is the most classroom-feasible FA variant, with zero inconclusive results in the trial-based studies reviewed, and consistent identification of attention, escape, and tangible functions without false automatic-reinforcement findings (Nesselrode et al., 2022).

Single-subject experimental designs. The SCED evidence is extensive. Standish and colleagues' (2021) multiple-baseline-across-participants design validated TB-OVI real-time decision rules in five caregiver-implemented TBFAs, all producing effective function-based interventions (Standish et al., 2021). Andersen and colleagues' (2022) alternating-treatments comparison of TBFA versus extended FA produced the 71%-time-reduction finding with 80% treatment success (Andersen et al., 2022). Weyman and colleagues' (2022) multiple-baseline across four children demonstrated TBFA-guided FCT for ritual-interruption functions (Weyman et al., 2022). Davis and colleagues' (2023) study on coaching parents via telehealth across three children confirmed TBFA viability in home settings with remote support (Davis et al., 2023). Lambert and colleagues' (2017) teacher-conducted latency-based TBFA demonstrated the format in an inclusive kindergarten classroom (Lambert et al., 2017).

Randomized comparative effectiveness. Call and colleagues' (2024) 57-child RCT — the sole randomized study in this corpus — found equivalent FCT success for FBA with versus without TBFA, raising the question of whether TBFA adds treatment-effectiveness benefit beyond well-conducted descriptive FBA in young autistic children (Call et al., 2024). This study complicates but does not contradict the TBFA literature: it addresses treatment effectiveness relative to non-experimental FBA, not TBFA's accuracy relative to traditional FA.

Training methodology. Griffith and colleagues' (2020) multiple-baseline demonstrated 100% fidelity in ABA-naive trainees with a 2-hour 20-minute self-instruction plus rehearsal package (Griffith et al., 2020). Standish and colleagues' (2023) automated training produced equivalent results with online-only modules plus role-play (Standish et al., 2023). Alnemary and colleagues' (2017) PTM validated a cascade-dissemination architecture for TBFA across agencies (Alnemary et al., 2017). Togashi's (2025) CBI-plus-telehealth-BST study extended training delivery to low-resource international settings (Togashi, 2025).

School (Primary Setting)

Schools are the primary use case for TBFA, and the research base reflects this. The systematic review evidence confirms that TBFA consistently identifies attention, escape, and tangible functions in classrooms without inconclusive results (Nesselrode et al., 2022). Bloom and colleagues' 2011 classroom application established the procedural template; Lambert and colleagues extended it to teacher-conducted latency-based variants (Lambert et al., 2017). Rispoli and colleagues' published TBFA applications in public school classrooms for students with developmental disabilities (cited in Brown et al., 2025) confirm ecological validity in naturalistic instruction (Brown et al., 2025). The school setting's advantage for TBFA is the stability of daily routines, which provides reliable opportunities for scheduled or embedded trials across conditions. The limitation is the multi-student classroom environment, which makes procedural control of the social attention condition more demanding — other students' attention is not under the practitioner's direct control, potentially contaminating the control trial's reinforcer-deprivation requirement.

Clinic

Clinic settings are where most TBFA validity studies have been conducted, but the clinic context reduces much of TBFA's distinctive advantage over traditional FA — both formats are feasible when an analog room is available. The primary clinic application for TBFA is efficiency: Andersen and colleagues' mealtime-behavior comparison showed TBFA taking 71% less time than extended FA with equivalent treatment utility, suggesting TBFA is a reasonable first step in clinic-based assessment before escalating to extended analysis if results are undifferentiated (Andersen et al., 2022). Henry and colleagues' brief-to-extended escalation sequence formalizes this as an explicit decision algorithm: start with brief trial-based conditions; escalate to extended analyses with confirmation and annulment phases only when data are undifferentiated (Henry et al., 2021).

Home (Caregiver-Implemented)

Home implementation by caregivers is a validated extension of TBFA with two distinct training pathways. Direct BST training (Gerow et al., 2020; Davis et al., 2023) and TB-OVI-supported real-time decision-making (Standish et al., 2021) both produce accurate functional conclusions and effective treatments in home settings (Gerow et al., 2020) (Standish et al., 2021) (Davis et al., 2023). The home's advantage is ecological validity: the caregivers who know the child's repertoire and daily schedule are implementing the trials, which increases the probability that the naturalistic antecedents and consequences are properly arranged. The limitation is procedural integrity — Gerow and colleagues' sessions required 2.4 to 3.5 hours and researcher support throughout; fidelity without ongoing coaching may decline, consistent with the maintenance concerns Togashi (2025) documented at follow-up (Togashi, 2025).

Residential and Community

Residential programs and community-based services have applied TBFA in settings where individual session scheduling is difficult and staff-to-client ratios are low. Alnemary and colleagues' pyramidal training model was validated in an ABA service agency context where scaling TBFA skills across multiple technicians was the practical challenge (Alnemary et al., 2017). The embedded TBFA format is particularly suited to group-home and community settings because it operates within the existing activity schedule rather than requiring dedicated assessment blocks. The limitation in residential settings is the risk that social attention from multiple staff members contaminates the attention-condition control trial; scripted protocols for who delivers attention and when during trial segments are essential.

• Insufficient trials before visual analysis. Running fewer than five pairs per condition makes the differentiation criterion unreliable. Amador and colleagues' systematic review identified this as the most common procedural gap in the published TBFA literature (Amador et al., 2024). Minimum five pairs per condition; use the binomial stopping rule only when five consecutive trials all yield the predicted pattern (Saini et al., 2020).

• No inter-trial interval. Immediately sequencing a test trial into a control trial without a neutral inter-trial period risks carryover of the establishing operation's effects from the test into the control or vice versa. The standard protocol specifies a brief neutral interval — at least 30 seconds — between control and test segments within a pair, and at least a natural activity transition between condition pairs.

• Function-confounding setup. Running an escape-condition test trial during a preferred activity (where the demand to terminate the preferred activity may also produce attention from the teacher during correction) confounds escape and attention functions in the same test trial, making differentiation meaningless. Each trial must manipulate exactly one establishing operation and one consequence.

• Missing or contaminated control trial. The control trial is not a free period; it is an active state in which the hypothesized reinforcer is freely and continuously available. A control trial for an escape condition in which the teacher is delivering intermittent attention — a common classroom state — contaminated both the escape-function manipulation and the attention-free baseline requirement simultaneously. Procedural integrity for the control trial is as critical as for the test trial.

• Treating the condition with the highest raw rate as the function. TBFA differentiation requires the test trial rate to be higher than the control trial rate within the same condition, not just higher than other conditions' test trial rates. A behavior with moderately elevated attention-test rates that exceed all other test rates but also show elevated attention-control rates is not differentiated for attention — the elevated control rate negates the differentiation.

• Applying TBFA to automatically reinforced behavior. The systematic review evidence is consistent that TBFA embedded in active social routines virtually never identifies automatic reinforcement (Nesselrode et al., 2022). Using TBFA when SIB or stereotypy occurs in alone conditions or during low-stimulation periods is the wrong tool; a traditional alone-condition FA or competing-stimulus assessment is appropriate.

• Relying on knowledge-only training for implementers. Togashi's (2025) finding that CBI improved knowledge without producing accurate implementation (Togashi, 2025), and Griffith and colleagues' (2020) requirement for role-play rehearsal even with a detailed self-instruction manual (Griffith et al., 2020), converge on a single operational rule: every TBFA implementer must complete role-play rehearsal with feedback before running trials with a client.

1. Socially mediated function suspected; setting is a school or community; BCBA has only consultative access. Run TBFA — embedded or scheduled, depending on daily structure; train the teacher or paraprofessional with BST plus role-play. Minimum five pairs per condition across attention, escape, and the most plausible tangible condition (Nesselrode et al., 2022) (Togashi, 2025).

2. Setting is home; BCBA will coach caregivers remotely. Run caregiver-implemented TBFA with telehealth BST coaching and TB-OVI real-time decision support; plan for 5 to 7 trials per condition given higher implementation variability (Davis et al., 2023) (Standish et al., 2021).

3. Open-ended interview produces a clear synthesized hypothesis; behavior is socially mediated; clinic access is available. Run the IISCA (original or performance-based depending on severity) rather than TBFA — the synthesized contingency test is more efficient when the interview has already identified the specific maintaining context (Jessel et al., 2024).

4. Behavior is suspected to be automatically reinforced, or multiply controlled with one function possibly automatic. Run traditional FA with an alone or ignore condition. TBFA should not be the primary assessment for this presentation (Nesselrode et al., 2022).

5. TBFA produces undifferentiated results after 8 to 10 pairs per condition. Escalate to an extended analysis using Henry and colleagues' brief-to-extended sequence: add confirmation and annulment phases for the most plausible function; if still undifferentiated, transition to traditional FA in a controlled analog environment (Henry et al., 2021).

6. High-severity behavior, significant trauma history, or caregiver concern about evocation. Consider performance-based IISCA over TBFA — IISCA's trauma-informed format minimizes the number of severe behavior instances required, which TBFA's repeated-trial structure does not guarantee (Jessel et al., 2024).

7. Low-frequency behavior with long inter-response times. Use latency-based TBFA (record time to first response rather than occurrence within trial) rather than standard occurrence-based TBFA; Lambert and colleagues' teacher-conducted latency variant is the published template (Lambert et al., 2017).

8. Consultative BCBA who cannot attend any in-person sessions. Deliver BST training via telehealth (CBI module plus live BST practice); use TB-OVI real-time rules so the implementer can make data-based decisions independently between coaching calls (Togashi, 2025) (Standish et al., 2021).

9. Agency needs to scale TBFA across multiple technicians efficiently. Use a pyramidal training model: train four senior staff, have each train one junior technician; include generalization probes with novel client scenarios before independent implementation (Alnemary et al., 2017).

10. TBFA results are clear and function is identified. Move immediately to a function-matched intervention — differential reinforcement of alternative behavior for the identified function, or FCT with the specific reinforcer the TBFA isolated. TBFA is the front end of treatment, not a stand-alone diagnostic exercise.

1. TBFA is the default FA format for school-based BCBAs. The systematic review evidence is unambiguous: TBFA reliably identifies socially mediated functions in classrooms, produces zero inconclusive results in published studies, and is feasible within regular instruction — no analog room required (Nesselrode et al., 2022).

2. Embedded and scheduled TBFA are both valid; match the choice to the daily structure and implementer experience. Embedded TBFA fits high-frequency behavior and experienced implementers; scheduled TBFA fits lower-frequency behavior and implementers who need predictable trial times to maintain procedural integrity (Lambert et al., 2017).

3. BST with role-play is the non-negotiable training component. Knowledge-only training (readings, videos, CBI alone) improves understanding but does not produce accurate implementation; every implementer must practice all trial types with a confederate client and receive feedback before client contact (Togashi, 2025) (Griffith et al., 2020).

4. Five pairs per condition is the minimum; use the binomial stopping rule to terminate early when evidence is clear. Five consecutive test trials with problem behavior occurring in the predicted direction meets statistical differentiation at p = .03 and justifies early termination before completing a full 10-pair block (Saini et al., 2020).

5. The control trial is analytically half the TBFA — do not treat it as a rest period. The differentiation criterion requires near-zero problem behavior in the control trial; contaminated or informally conducted control trials invalidate the comparison and can produce false-positive conclusions (Amador et al., 2024).

6. Telehealth BST plus TB-OVI rules make TBFA viable for fully remote consultation. Train the implementer asynchronously, provide live BST practice sessions, then give the TB-OVI decision rules for real-time data interpretation between coaching contacts (Togashi, 2025) (Standish et al., 2021).

7. TBFA does not reliably identify automatic reinforcement. If the presenting behavior occurs in alone conditions or at high rates during low-stimulation periods, supplement TBFA with a traditional alone-condition FA before concluding a social function is maintaining the behavior (Nesselrode et al., 2022).

8. Escalate to extended analysis when TBFA produces undifferentiated results; do not repeat TBFA without modification. The brief-to-extended model — start with TBFA, add confirmation and annulment conditions for the most plausible function — is the appropriate next step when initial TBFA data paths do not differentiate (Henry et al., 2021).

9. TBFA treatment utility is strong even when TBFA functions do not exactly match extended FA functions. Andersen and colleagues found 71% time savings with 80% treatment success despite only 29% exact functional correspondence — the clinical decision to use TBFA does not require perfect replication of an extended FA's conclusions (Andersen et al., 2022).

10. Pyramidal training is the efficient agency model for scaling TBFA. Train four senior staff to mastery; have each train one frontline technician; build generalization probes into the training sequence; do not assume skills transfer without verification (Alnemary et al., 2017).

11. TBFA is a front-end assessment — it has no value without function-matched treatment. Identifying the function is the means; a function-matched intervention (DRA, FCT, NCR, extinction) is the goal. The TBFA's clinical value is fully realized only when the identified function drives the intervention plan (Weyman et al., 2022).

12. TBFA and IISCA are distinct procedures that address different implementation contexts — do not treat them interchangeably. TBFA uses standardized condition pairs across attention, escape, and tangible functions; IISCA uses a single interview-derived synthesized contingency. TBFA is preferred when the function is unknown and a systematic condition sweep is needed; IISCA is preferred when the interview has already narrowed the hypothesis to a specific synthesized contingency (Jessel et al., 2024).

Primary research synthesized in this guide. DOIs link to the original source.

• Alnemary, F., Wallace, M., Alnemary, F., Gharapetian, L., & Yassine, J. (2017). Application of a pyramidal training model on the implementation of trial-based functional analysis: A partial replication. Behavior Analysis in Practice, 10(3), 301–306. https://doi.org/10.1007/s40617-016-0159-3
• Amador, J. L., DeBar, R. M., Deshais, M. A., Gardner, A. W., & Sidener, T. M. (2024). A descriptive assessment of trial-based functional analysis research. Behavioral Interventions, 39(4). https://doi.org/10.1002/bin.2020
• Andersen, A. S., Hansen, B. A., & Peterson, K. M. (2022). An evaluation of trial-based functional analyses of inappropriate mealtime behavior. Journal of Applied Behavior Analysis, 55(1), 264–289. https://doi.org/10.1002/jaba.888
• Brown, K. R., Helvey, C. I., Kranak, M. P., & Lavin, A. (2025). Functional analysis decision-making considerations. Behavior Analysis in Practice, 18(4), 1237–1254. https://doi.org/10.1007/s40617-025-01057-w
• Call, N. A., Bernstein, A. M., O'Brien, M. J., Schieltz, K. M., Tsami, L., Lerman, D. C., Berg, W. K., Lindgren, S. D., Connelly, M. A., & Wacker, D. P. (2024). A comparative effectiveness trial of functional behavioral assessment methods. Journal of Applied Behavior Analysis, 57(1), 166–183. https://doi.org/10.1002/jaba.1045
• Davis, T. N., Gerow, S., Wicker, M., Cosottile, D., Exline, E., Swensson, R., & Lively, P. (2023). Utilizing telehealth to coach parents to implement trial-based functional analysis and treatment. Journal of Behavioral Education, 32(4), 703–725. https://doi.org/10.1007/s10864-022-09468-3
• Gerow, S., Rivera, G., Radhakrishnan, S., & Davis, T. N. (2020). Parent-implemented brief functional analysis in the home. Behavioral Interventions, 35(4), 691–703. https://doi.org/10.1002/bin.1734
• Germansky, S., Reichow, B., Martin, M., & Snyder, P. (2020). A systematic review of caregiver-implemented functional analyses. Behavior Analysis in Practice, 13(3), 698–713. https://doi.org/10.1007/s40617-019-00404-y
• Griffith, K. R., Price, J. N., & Penrod, B. (2020). The effects of a self-instruction package and group training on trial-based functional analysis administration. Behavior Analysis in Practice, 13(1), 63–80. https://doi.org/10.1007/s40617-019-00388-9
• Henry, J. E., Kelley, M. E., LaRue, R. H., Kettering, T. L., Gadaire, D. M., & Sloman, K. N. (2021). Integration of experimental functional analysis procedural advancements: Progressing from brief to extended experimental analyses. Journal of Applied Behavior Analysis, 54(3), 1045–1061. https://doi.org/10.1002/jaba.841
• Jessel, J., Fruchtman, T., Raghunauth-Zaman, N., Leyman, A., Lemos, F. M., Costa Val, H., Howard, M., & Hanley, G. P. (2024). A two-step validation of the performance-based IISCA: A trauma-informed functional analysis model. Behavior Analysis in Practice, 17, 727–745. https://doi.org/10.1007/s40617-023-00792-2
• Lambert, J. M., Lopano, S. E., Noel, C. R., & Ritchie, M. N. (2017). Teacher-conducted, latency-based functional analysis as basis for individualized levels system in a classroom setting. Behavior Analysis in Practice, 10(4), 422–426. https://doi.org/10.1007/s40617-017-0200-1
• Nesselrode, R., Falcomata, T. S., Hills, L., & Erhard, P. (2022). Functional analysis in public school settings: A systematic review of the literature. Behavior Analysis in Practice, 15(3), 958–970. https://doi.org/10.1007/s40617-022-00679-8
• Saini, V., Fisher, W. W., Retzlaff, B. J., & Keevy, M. (2020). Efficiency in functional analysis of problem behavior: A quantitative and qualitative review. Journal of Applied Behavior Analysis, 53(1), 44–66. https://doi.org/10.1002/jaba.583
• Standish, C. M., Bailey, K. M., Lambert, J. M., Copeland, B. A., Banerjee, I., & Lamers, M. E. (2021). Formative applications of ongoing visual inspection for trial-based functional analysis: A proof of concept. Journal of Applied Behavior Analysis, 54(4), 1526–1540. https://doi.org/10.1002/jaba.866
• Standish, C. M., Lambert, J. M., Copeland, B. A., Bailey, K. M., Banerjee, I., & Lamers, M. E. (2023). Partially automated training for implementing, summarizing, and interpreting trial-based functional analyses. Journal of Behavioral Education, 32(2), 239–260. https://doi.org/10.1007/s10864-021-09456-z
• Togashi, K. (2025). Training in trial-based functional analysis via computer-based instruction and behavioral skills training. Behavior Analysis in Practice. https://doi.org/10.1007/s40617-025-01136-y
• Weyman, J. R., Bloom, S. E., Campos, C., & Garcia, A. R. (2022). An application of the trial-based functional analysis to assess problem behavior evoked by ritual interruption. Behavioral Interventions, 37(4), 926–940. https://doi.org/10.1002/bin.1882