Replacement Behaviors for PDA: Why One FCR Backfires

For a PDA learner, the right replacement is a small family of responses: "I'm not ready," "I want my way," a head shake, a swap card, with a count rule (any three in a session) and a soft threshold (clinician accepts any topography that carries the intent). One rigid FCR (Functional Communication Response, the new way the learner gets the same need met) will fail. This page walks through the design choice and how to write it on the plan.

Why a single FCR backfires with a PDA learner#

The classic skills-based treatment (SBT) target is one clean FCR. "My way, my play." The learner says it. The clinician honors it. The learner gets reinforcement. That tight loop builds the behavior fast in most kids.

A PDA profile breaks the loop. The learner reads the script as a fresh demand. The wording is the demand. Saying it the "right" way is the demand. Now the FCR itself is what they refuse. You lose the response you were trying to build.

The speaker is direct on this point.

It can be more appropriate to come up with several so that you're not expecting one specific rigid response.From the talk — B. Kuerine Gray

Several. Not one. The fix is in the design, not the delivery.

Building a response class of replacement behaviors with the same function#

A response class is a group of behaviors that all serve the same function (the same job for the learner). For a PDA learner, that job is usually "delay or shape the demand, on my terms." Any behavior that carries that intent counts.

In session, the class might look like this:

• "I'm not ready."
• "I want to do it my way."
• A clear head shake or "no."
• A swap card or AAC (augmentative and alternative communication, a device or board the learner uses to talk) message that means the same thing.
• A request for help with the order or partner.

All five do the same job. They tell you the demand, as stated, is not landing. You accept any of them. You honor the intent the same way you would have honored the one scripted FCR.

This is not a new tool. It is the same FCR plan, just wider. The speaker is clear that PDA work is not about new interventions. It is about presenting the same ones through a PDA lens.

You keep the function. You widen the topography (the shape the behavior takes). You write the class on the plan, not the script.

Counting "I'm not ready" as an FCR (and why most teams don't)#

Most teams accept "my way, my play" and reject "I'm not ready." On paper it looks reasonable. The plan says "my way." The learner did not say "my way." Mark a zero. Move on.

That call costs you the case. "I'm not ready" carries the same intent. The learner is asking for delay or control. If you ignore it and push, you have just stepped on the FCR you are trying to grow.

The speaker calls out the trap with one line.

If you're saying "oh, well, you didn't say this," or they're responding with "I'm not ready" and your response is "oh, so you want to play my way," you're setting another demand.From the talk — B. Kuerine Gray

That response sounds like a clean prompt. It is a second demand stacked on the first. The learner now has two demands to escape, not one. The class drifts up the escalation cycle.

The fix is simple. Train the team to read intent first. If the behavior is in the class, count it, honor it, log it. The wording does not need to match a script. The function needs to match the plan.

Thresholds for correct response: shaping vs. rigid X = Y#

A threshold for correct response tells the team when to call a try a "yes." A rigid threshold says X = Y. The learner did the exact thing. Score it. Move on.

A shaping threshold says X is close enough today. The learner got it partway there. Score it. Reinforce it. Move the bar tomorrow.

For a PDA learner, rigid thresholds break the program. The learner reads any "you didn't quite get it" as one more demand. They stop trying. You lose the response.

A shaping threshold reads the trend. Across a week, is the class growing? Are more topographies showing up? Are the harder ones (a calm "no," a clear swap card) starting to replace the loud ones (a yell, a head down on the table)? If yes, the plan is working. Hold steady. Move the bar slow.

The speaker frames the choice as a clinical decision, not a paperwork one. Rigid X = Y will lose a PDA learner. Shaping keeps them.

Choices as a replacement-behavior modifier (order, partner, sequence)#

Choices are the second layer of the design. Once the class is set, choices give the learner control over the small stuff. Control over the small stuff lowers the heat on the big stuff.

The speaker is specific on what counts as a choice.

You may have to do certain things, but you may have a choice of what comes first. You may have a certain number of things done, but it doesn't matter which one you do first. You may have a certain task that needs to be done, but has different steps that can be done in different orders.From the talk — B. Kuerine Gray

Three buckets:

• Order. Two tasks happen. Learner picks first.
• Partner. A task happens. Learner picks who they do it with.
• Sequence. A multi-step task happens. Learner picks the inside order.

None of these change whether the task happens. They change the part the learner gets to set. That small control often drops the threat enough to bring the FCR class into reach.

Do not stretch a choice into "do we do this at all." That puts the demand itself on the table. Now you have offered opt-out, not control. PDA is not opt-out.

What you write on the treatment plan vs. what you accept in the moment#

The plan and the moment do not have to match word for word. They have to match by function.

On the plan, name the class:

• Goal. Learner will use any one of the following replacement responses across at least three independent opportunities per session, with reinforcement provided for any topography in the class.
• Class members. "I'm not ready." "I want my way." A clear "no." A swap card. AAC message with the same intent.
• Threshold. Shaping. Any class member counts. Track topography on a side column for trend.
• Choices. Order, partner, or sequence choice provided on at least two demands per session.

In the moment, you do not have to call out which one the learner used. You honor the intent. You give the same reinforcement you would have given the scripted FCR.

The speaker names this as the core design rule.

Can we plan for flexible behaviors? So not just that one response, but maybe multiple responses that get needs met. And when those happen, can we accept them?From the talk — B. Kuerine Gray

The "accept them" half is where most plans miss. The clinician writes the class but still grades against one script. Write the class. Grade the class. Honor the class.

Frequently asked questions#

Can I list more than one FCR on the same goal? Yes. Write it as a response class with named members and a shared function. A funder will accept "any of the following, all serving the same access or delay function" if you spell out the members and how you will measure them. The Carlozzi (2025) U.S. peer-reviewed PDA work and the broader functional communication training literature both support flexible topographies tied to one function. The key is naming the function in plain words ("learner gains a delay or order change on a demand") and listing the topographies you will count. Funders read that as a clean, measurable goal.

What if the learner's replacement behavior keeps drifting? Drift is data. If the class keeps growing into harder, louder, or longer behaviors, the plan is not landing yet. Two checks. First, did a new variable shift, like sleep, sickness, or a school change? The speaker calls this a "sliding scale" of availability, not a missed step. Second, is your threshold actually shaping, or did it drift back to X = Y? Most drift cases come from the team quietly tightening the bar. Loosen it for a week. Reinforce the class wide. Then start to shape again from a calmer baseline. Plan changes are not failure. They are how PDA cases close.

How do I take data when any of four topographies counts as the target? Use a class column and a topography column. The class column is your goal data: any class member, yes or no. That is what you graph for progress. The topography column is your clinical data: which one. That is what you read to spot drift, growth, or stuck spots. Side-by-side, the two columns let you score the goal honestly and still see what is changing under the hood. Most teams already do this for response classes around aggression. The same setup works for FCR classes built for PDA.

Watch the full talk →. The recording walks through five real cases (S, O, Y, C, F), how SBT got modified for each, and how the team coached caregivers to honor the wider FCR class at home.