By Matt Harrington, BCBA · Behaviorist Book Club · April 2026 · 12 min read

Procedural Integrity in ABA: Measurement, Prioritization, and Compassionate Correction

Overview & Clinical Significance

Procedural integrity — the degree to which an intervention is implemented as designed — is a foundational concept in behavior analysis that directly determines whether experimental and clinical conclusions are valid. In research, low integrity confounds independent variables and renders data uninterpretable. In clinical practice, low integrity means that the procedure a client is experiencing is not the procedure that was evaluated and approved — with potentially significant consequences for both outcomes and ethics.

Despite its foundational importance, procedural integrity is among the most commonly neglected monitoring domains in applied ABA settings. Supervisors who regularly track data on client behavior change frequently have no systematic process for measuring whether the procedures driving that data collection are being implemented correctly. This creates a fundamental interpretive problem: if a client is not making progress, is the procedure failing, or is an unpracticed version of the procedure failing? Without integrity data, the distinction is impossible to make.

The clinical significance of prioritizing integrity monitoring extends beyond individual clients. At an organizational level, integrity data provides the foundation for quality assurance programs, staff training needs assessments, and evidence of professional accountability. Organizations that collect and act on procedural integrity data are better positioned to demonstrate clinical quality to funders, insurers, and licensing bodies — and are better prepared to identify systemic problems before they result in adverse outcomes.

The OPI (Operational Procedural Integrity) Task Force represents a professional investment in addressing this gap systematically. By bringing together experts from diverse ABA service contexts, the Task Force has developed guidance that accounts for the practical realities of measuring integrity under high-demand clinical conditions — not just its theoretical importance.

Background & Context

Research on procedural integrity in ABA settings has a long history, originating in the single-case experimental design literature where the necessity of documenting independent variable consistency was recognized early. Landmark reviews established that integrity levels substantially affect the effectiveness of behavioral interventions, with research demonstrating clear dose-response relationships between integrity level and client outcomes.

The application of integrity monitoring in clinical (as opposed to experimental) settings has been slower to develop. Clinical ABA settings face resource constraints — time, personnel, technological infrastructure — that make the intensive integrity monitoring feasible in research settings impractical for everyday service delivery. The OPI Task Force's work addresses this gap directly by focusing on feasible, sustainable approaches to integrity monitoring that can be implemented across diverse service environments including large providers, small practices, school settings, and home-based services.

Measurement approaches for procedural integrity fall into several categories: direct observation (a trained observer records whether each procedural step is implemented correctly during a session), self-monitoring (the implementer records their own adherence to the procedural steps), permanent product review (examining data sheets, session notes, or video recordings for evidence of correct implementation), and permanent product sampling (reviewing a subset of these records on a scheduled basis). Each approach has distinct tradeoffs in accuracy, feasibility, and cost.

The relationship between procedural integrity and staff performance management is also important. Integrity data provides the operational basis for distinguishing between staff who implement procedures correctly but obtain poor client outcomes (suggesting a programming problem) and staff who obtain poor outcomes because integrity is low (suggesting a training or motivation problem). This distinction has direct implications for how supervisors respond — adding staff training where training is needed rather than revising programming where programming is sound.

Clinical Implications

For practicing BCBAs, the most direct clinical implication of integrity monitoring is the ability to make valid data-based decisions. Without integrity data, performance graphs are ambiguous: a flat or declining trend in a target behavior might reflect a poorly designed intervention, an intervention that is well-designed but not being implemented, client variables that are affecting responsiveness, or measurement error. Integrity data allows the clinician to rule out implementation failures before concluding that the intervention itself needs revision.

This disambiguation has practical consequences. A clinician who revises a well-designed intervention because integrity was low — without knowing integrity was low — has likely made a clinical error. The revised intervention may appear to produce better outcomes not because it was clinically superior but because attention to the change prompted temporarily higher implementation fidelity. Without integrity data, this confound cannot be detected.

Integrity monitoring also creates a mechanism for early identification of implementation drift — the gradual departure from procedural fidelity that occurs as novel procedures become routine and are subject to shortcutting. Implementation drift is not typically an act of defiance or negligence; it is a predictable behavioral outcome when reinforcement for precise implementation fades over time. Scheduled integrity monitoring detects drift before it becomes severe, allowing for corrective training that is brief and targeted rather than extensive and disruptive.

At the organizational level, integrity monitoring data can be aggregated across staff members and programs to identify systemic training needs. If integrity is consistently low on a specific procedural component across multiple implementers, the problem is likely in the training or the design of the procedure, not in individual staff. This organizational-level analysis is only possible when integrity data is collected systematically and retained in a format that allows aggregation.

Ethical Considerations

Procedural integrity monitoring has clear grounding in multiple 2022 BACB Ethics Code standards. Standard 2.01 (Providing Effective Treatment) requires BCBAs to use evidence-based procedures and to evaluate their effectiveness. An intervention cannot be evaluated as effective if it is not being implemented as designed — without integrity data, the evaluation is incomplete. Standard 2.04 (Ongoing Data Collection) requires BCBAs to collect data throughout service delivery to monitor progress. Integrity data is part of this ongoing data collection requirement, not a research-only activity.

Standard 2.16 (Continuity of Services) and related provisions require BCBAs to maintain quality of services across staff and setting changes. Integrity monitoring is the mechanism by which quality is verified across these transitions — without it, service continuity is an aspiration rather than a demonstrated fact. Standard 4.07 (Performance Feedback) explicitly requires BCBAs to provide behavior-analytic feedback to their supervisees. Integrity data provides the specific, objective basis for this feedback — feedback grounded in observed implementation is more actionable and less subject to dispute than feedback grounded in impression.

The OPI Task Force's emphasis on addressing low integrity in a respectful and compassionate manner is itself ethically important. Integrity data can be used punitively — as evidence against staff rather than as clinical information. Using integrity data to initiate disciplinary processes without first assessing the environmental variables contributing to low integrity is both clinically unsound (skipping the functional assessment) and ethically questionable. The data exists to improve client outcomes, and the response to low integrity should be structured around that goal.

Assessment & Decision-Making

Deciding which procedures to prioritize for integrity monitoring in a resource-constrained environment requires triage logic. High priority should be assigned to procedures that: carry client safety implications if not implemented correctly (e.g., blocking, restraint, extinction), have been recently trained or modified (higher probability of implementation drift), are associated with client behaviors where progress has stalled (integrity as a confound in data interpretation), or are implemented by staff whose training history or recent performance data suggests higher drift risk.

Selecting the appropriate measurement approach requires balancing accuracy against feasibility. Direct observation provides the most accurate data but requires personnel with the competency and availability to conduct reliable observations. Self-monitoring is more feasible but introduces bias — staff may rate their own implementation more favorably than an independent observer would. Permanent product review (session notes, data sheets, video recordings) provides an objective record but requires that permanent products actually capture the information needed to assess integrity.

Decision-making about frequency of integrity measurement is also consequential. Research suggests that performance improves when integrity is measured even without feedback — the reactive effect of being observed is itself an antecedent intervention. Scheduled monitoring therefore produces both data and performance improvement. At minimum, integrity should be measured during the initial implementation of any new procedure, periodically thereafter on a scheduled basis (monthly for stable procedures, more frequently for recently changed ones), and whenever client data suggest possible implementation problems.

When integrity data reveals low implementation, the response sequence should mirror the PDC-HS framework: is the low integrity due to unclear expectations, inadequate training, missing materials, or insufficient consequences? The answer determines whether the intervention is clarification of the procedure, additional BST, provision of materials, or consequence modification.

What This Means for Your Practice

If procedural integrity is not currently a systematic part of your clinical quality process, the most impactful first step is establishing a baseline: select two or three procedures currently being implemented on your caseload and conduct brief direct observations to determine current integrity levels before you do anything else. The data you find will likely be informative — and may be surprising. National research suggests that integrity levels in naturalistic ABA settings are frequently lower than supervisors estimate.

From that baseline, identify the one or two procedures where low integrity has the most significant clinical consequence and prioritize monitoring and intervention for those. This is not comprehensive integrity surveillance — it is targeted quality improvement in the domains where quality most directly affects client welfare. From that starting point, you can expand monitoring systematically as you build the infrastructure to support it.

Addressing low integrity with your staff begins with the framing: present integrity data as clinical information, not as evidence of wrongdoing. The message is 'this is what the data shows about how the procedure is being implemented, and here's how we can address it' — not 'you're doing it wrong.' That framing is not only more ethical; it is more likely to produce the behavior change you need.