cumulative

H.6 Data-based decisions about procedural integrity

What the task asks: Use data on how accurately an intervention is carried out to decide whether to retrain implementers, change supports or simplify the procedure before judging whether it works.

Selecting and Implementing Interventions5 min read5 practice questionsOfficial outline (opens in a new tab)

The big idea

An intervention can only be judged by its results if it is being carried out as written. Procedural integrity (also called treatment integrity or fidelity) is the degree to which an intervention is implemented as planned. C.12 covers how to measure it. This task is about what to do with the measurements.

The logic is simple. Before deciding that an intervention does not work, check whether it is actually happening. If it is not, the problem is implementation, and changing the intervention would be a mistake.

Breaking it down

Reading outcome and integrity data together

Client outcomeIntegrityMost likely conclusionNext step
ImprovingHighThe intervention is workingContinue; plan fading
Not improvingHighThe intervention may need to changeModify it (H.7)
Not improvingLowThe intervention has not really been testedFix implementation first
ImprovingLowSome components may be unnecessary, or other variables are at workCheck measurement; consider simplifying; keep monitoring

Standard 2.17 (Collecting and Using Data) and Standard 2.18 (Continual Evaluation of the Behavior-Change Intervention) support this approach: data, including data on implementation, guide decisions.

Kinds of integrity errors

  • Errors of omission: a planned component is left out, such as no reinforcer after a correct response or no prompt when one was due.
  • Errors of commission: something is added that the plan does not call for, such as attention after problem behavior or an extra prompt. Research suggests that commission errors that reinforce problem behavior can be especially damaging to treatment (St. Peter Pipkin et al., 2010).
  • Dosage problems: the intervention is accurate but delivered too little, such as fewer sessions or trials than planned.

Treatment drift is a gradual change in how a procedure is carried out over time, so that what is delivered moves away from the plan. It is common after initial training ends, which is why integrity is checked repeatedly, not once.

Finding the cause

Low integrity has causes, and the fix should match the cause (see I.6 on performance diagnostics). Ask:

  • Does the implementer know how to do it? (skill)
  • Are there clear written steps, materials and time? (antecedents and resources)
  • Does the implementer get feedback and recognition for doing it well? (consequences)
  • Does the procedure compete with other demands, or does the implementer disagree with it? (competing contingencies and acceptability)

Responses to low integrity

  • Retrain with behavioral skills training (BST): instructions, modeling, rehearsal and feedback until the implementer meets a criterion, ideally in the real setting (see I.5).
  • Give performance feedback based on integrity data: specific praise for steps done correctly and correction for errors, delivered soon after observation. Graphed feedback can help.
  • Simplify the procedure: fewer steps, lighter data collection, or a focus on the components that matter most. A plan a teacher runs at 95% often beats one they run at 50%.
  • Add supports: checklists, visual prompts, prepared materials and scheduled reminders.
  • Improve contextual fit (H.2): adjust timing or format so the procedure fits the routine and the implementer’s values.
  • Self-monitoring by implementers, checked against direct observation.

Many teams use a threshold, such as 80% or 90% of steps correct, to trigger retraining. There is no universal standard; the right level depends on which errors matter most for the procedure.

Seeing it in practice

Ms. Bell, a second-grade teacher, is running a DRA plan for Jaylen, whose call-outs have not decreased after three weeks.

  1. The analyst observes and scores integrity at 55%. Ms. Bell often praises hand-raising, as planned, but also answers call-outs (a commission error).
  2. In an interview, she says she knows the steps, but answering is a habit and the class moves fast.
  3. The analyst models the response in class, has her rehearse it during a planning period, and gives feedback after the next two observations. A note on her desk reads: “Hand up: praise. Call-out: next student.”
  4. Integrity rises to 90%. Two weeks later, call-outs have dropped by half. The plan itself did not need to change.
050100IntegrityBefore retrainingAfter BST and feedback050100Call-outs1234567891011121314ObservationsPercent
Jaylen's call-outs (percentage of intervals) graphed with Ms. Bell's integrity. While integrity sits near 55%, call-outs stay flat. After retraining, integrity rises to about 90% and call-outs fall by half, with no change to the plan.

At a group home, integrity on a 15-step morning routine plan is low across all shifts. The analyst cuts the plan to the six steps most closely tied to the resident’s progress, and integrity climbs.

Common mix-ups

  • Low integrity is not low effectiveness. If the plan was not delivered, its effectiveness is unknown.
  • Measuring integrity is C.12; acting on it is H.6. Expect questions that give you integrity data and ask for the next step.
  • Retraining is not the only answer. If the procedure is too complex or does not fit the setting, retraining alone will not hold. Simplify or adjust it.
  • Integrity is not checked once. Drift happens, so schedule regular checks, especially after training ends.
  • Not all errors are equal. Reinforcing problem behavior can undo treatment faster than skipping a praise statement.

Exam tip: When outcome data are poor, the first question is whether integrity data exist and what they show. If integrity is low, the best answer fixes implementation before changing the intervention.

Quick recap

  • Judge an intervention only after confirming it was implemented as planned.
  • Omission and commission errors, dosage problems and drift all lower integrity.
  • Match the fix to the cause: BST, performance feedback, a simpler procedure, supports or better fit.
  • Recheck integrity regularly, and pair it with outcome data for every decision.

Further reading

  • St. Peter Pipkin, C., Vollmer, T. R., & Sloman, K. N. (2010). Effects of treatment integrity failures during differential reinforcement of alternative behavior: A translational model. Journal of Applied Behavior Analysis, 43(1), 47-70.
  • Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). Applied behavior analysis (3rd ed.), chapter 10.

Key terms

Procedural integrity (Treatment integrity, Treatment fidelity, Implementation fidelity)
The extent to which a procedure is carried out exactly as designed. It is usually reported as the percentage of steps implemented correctly. Low integrity makes it impossible to tell whether an intervention works, and the same idea applies to whether supervision is delivered as planned.
Error of omission
An implementation error in which a required step of a procedure is left out.
Error of commission
An implementation error in which the implementer adds something the procedure does not include, or carries out a step incorrectly. Commission errors that deliver reinforcement for problem behavior can be especially harmful to treatment.
Dosage (Exposure)
The amount of an intervention a client actually receives compared with the amount planned, such as the number of sessions, minutes or trials delivered. It is a dimension of procedural integrity separate from how accurately each step is carried out.
Treatment drift (Procedural drift)
A gradual, unplanned change in how an intervention is carried out over time, so that the procedure in use no longer matches the one designed. Repeated integrity checks are the way to detect it.
Behavioral skills training (BST)
A training package made up of instructions, modeling, rehearsal and feedback, repeated until the learner performs the skill to a mastery criterion. It is a well-supported way to teach new skills to staff, caregivers and clients.
Performance feedback (Feedback)
Information given to a person about their past performance so they can maintain or improve it. It can be verbal, written or graphic, immediate or delayed, and can include praise, correction or both. It works best when it is specific and combined with goals or reinforcement.

Check yourself

5 questions on H.6. Choose an answer to see whether you're right and why.

  1. Question 1A behavior plan for a student's aggression has produced no change after a month. Integrity checks show that staff complete about 40% of the plan's steps. What is the best next step?
    Show the answer

    Answer: A. With integrity at 40%, the plan as written has not really been tested, so the outcome data say little about it. Replacing or intensifying the plan would change the intervention when the problem is implementation.

  2. Question 2Ms. Guerrero's plan is to ignore her son's whining and praise his polite requests. Observation shows she praises requests reliably but also often comforts him when he whines. Comforting him is an example of:
    Show the answer

    Answer: C. Adding something the plan does not call for, here attention after whining, is an error of commission, and it may reinforce the behavior being treated. An omission error would be leaving out a planned step, such as failing to praise a request.

  3. Question 3Integrity on a classroom token system is 95% for the first two weeks after training, then falls steadily to 60% over two months. What does this pattern show?
    Show the answer

    Answer: A. A gradual move away from the procedure as designed after training ends is treatment drift. It is why integrity is checked repeatedly and why ongoing feedback, not a single training, is planned.

  4. Question 4Staff at a group home score low on integrity for a 20-step morning routine plan. Interviews show they understand the plan but cannot complete every step during busy mornings. What is the most appropriate change?
    Show the answer

    Answer: D. The staff have the skill, so retraining will not fix the problem; the plan does not fit the setting. Simplifying it to the components that matter most is more likely to raise integrity and keep it high.

  5. Question 5Which component is part of behavioral skills training?
    Show the answer

    Answer: B. Behavioral skills training combines instructions, modeling, rehearsal and feedback, repeated until the trainee performs to a criterion. A written quiz shows what the trainee can say about the steps, not whether they can perform them, so it is not one of the four components.

Adds a step to your study record. Saved in this browser only.

Type to search all 104 guides and the glossary.