Procedural Fidelity in ABA: Track It, Prove It

Procedural fidelity is how closely a plan gets followed in real life. It is also called procedural integrity or treatment fidelity. When an RBT runs a program the way it was written, fidelity is high. When steps get skipped or changed, fidelity drops.

This matters to everyone on a team. BCBAs write the plans. RBTs, teachers, and parents run them each day. If nobody tracks how the plan is run, the data can lie to you. You may blame the plan when the real problem is how it was used.

What procedural fidelity means#

Fidelity is a simple idea with big effects. It asks one question. Did we do the intervention the way we said we would?

You can measure it. You count the correct steps and divide by the total steps. That gives you a percent. A plan run at 100% looks very different from one run at 50%.

Matt Harrington makes the case that fidelity data protects your other data. Without it, you cannot trust your graphs.

if you don't have procedural integrity data saying that this data corresponded to what you wrote, then how do we know that what you wrote was the actual thing that was done? From the talk — Matt Harrington

Everyone agrees it matters, few track it#

Most behavior analysts say fidelity is important. Far fewer collect real data on it. That gap is the core problem.

Harrington shares survey numbers from a group of 140 BCBAs. Almost all gave feedback, but very few graphed it.

90% provided feedback consistently. Almost all of it was in-the-moment vocal feedback. But only 20% used graphs to track their feedback. From the talk — Matt Harrington

Vocal feedback in the moment is helpful. But it fades fast and leaves no record. Graphs let you see trends over weeks. They also show if feedback is actually changing how staff run the plan.

When the plan is not the problem#

Low fidelity can look like a failed intervention. That leads teams to change a plan that was fine. Fidelity data helps you tell these apart.

Harrington uses a clear rule of thumb for this check.

If we're doing something 50% of the time and there's no impact on behavior, that makes sense. If we're doing something 100% of the time and there's no impact, then we might have an intervention mismatch. From the talk — Matt Harrington

So flat data at low fidelity means run it correctly first. Flat data at high fidelity means the plan may be wrong. You cannot know which one you have without fidelity data. This same logic applies when you partner with an OT and track how often caregivers do the homework.

Poor fidelity can cause harm#

Fidelity is not only about clean data. It is also about safety. A plan run wrong can hurt the person you serve.

Dr. Shane Spiker links weak fidelity to crisis and harm. Staff who drift from the plan can trigger the very behavior they meant to reduce.

people with poor treatment fidelity will re-traumatize. They will cause problems or your treatment itself could re-traumatize. From the talk. Dr. Shane Spiker

This raises the stakes. Fidelity is an ethics issue, not just a paperwork issue. A good plan run badly is not a good plan in practice.

Check fidelity before you dig deeper#

When a graph looks bad, it is tempting to hunt for a deep clinical cause. Often the simpler answer is fidelity. Check that first before you assume something complex.

Harrington calls this a parsimonious check, which means the simplest likely reason. Your data cannot look good if the plan is not run well. As he puts it, a graph will never look beautiful when there is no fidelity behind it.

This is a common miss across the field. Fidelity gaps grow when a plan must run across home, school, and clinic. Each setting adds a new person and a new chance to drift. Harrington notes that weak fidelity is almost always the main reason interventions stay inconsistent across our three tier service model.

So check fidelity in every context, not just the clinic. The problem may live in one setting and not the others. Fidelity data across places shows you exactly where to coach.

How to build fidelity in your team#

Fidelity grows with a plan and a routine. It does not happen by accident. Supervisors set the pace here.

Mellanie Page ties fidelity to a monthly habit of checks and feedback.

Monthly integrity checks on learner plans and give them feedback on how closely they're following the plan. From the talk. Mellanie Page

Start with a short fidelity checklist for each plan. List every key step in plain words. Watch a session and mark each step as done or missed. Then share the score with the staff member and coach the misses.

Keep the data on a graph next to the behavior data. Over time you will see fidelity and progress move together. That link is what makes the case for tracking it.

What the research says#

Studies back up the need to build and hold fidelity over time. Simple tools can help staff stay accurate on their own.

Video self-monitoring has helped staff keep high fidelity, even when a supervisor is not watching the session (Paden & Carroll, 2024). This matters because fidelity often drops when nobody is present to observe.

Training staff to spot errors also helps. Behavioral skills training let RBTs correctly identify right and wrong steps in discrete trial teaching (Katechis et al., 2026). Staff who can see an error are better placed to fix it.

New tools are being tested too. An AI platform improved fidelity for new technicians and caregivers with little prior training (Yagafarova et al., 2025). These findings point to fidelity as a skill you can teach and support, not a fixed trait.

Want to go deeper on tracking fidelity in supervision? Watch Supervision Articles Deep Dive with Matt Harrington on OpenCEU.

FAQ#

What is the difference between procedural fidelity and treatment integrity? They mean the same thing. Both describe how closely a plan is run as written. You may also hear procedural integrity. Pick one term and use it the same way across your team.

How do you measure procedural fidelity? Break the plan into clear steps. Watch a session and mark each step as done correctly or not. Divide correct steps by total steps for a percent. Track that percent on a graph over time.

Why does low fidelity matter if the child still makes some progress? Low fidelity makes your data hard to trust. You cannot tell if the plan works or if luck helped. It can also slow learning and, in some cases, cause harm. Clean fidelity data protects both the client and your clinical choices.