Procedural fidelity
Procedural fidelity is the degree to which an intervention is carried out exactly as written in the plan; it is also called treatment integrity or procedural integrity.
A behavior plan is only tested if it is actually implemented as designed. Procedural fidelity is often measured with a checklist of the plan's steps: an observer (usually the supervisor) watches a session and records whether each step was done correctly, then calculates a percentage. For example, delivering reinforcement within 3 seconds of the correct response, using the correct prompt level, and ignoring a specified behavior might all be items on the checklist.
Poor fidelity is risky. If the intervention is done incorrectly and behavior does not improve, the supervisor may wrongly conclude that the intervention does not work and drop an effective procedure. If behavior improves, no one knows whether the plan or the changes made it work. Inconsistent implementation can also make problem behavior worse, for example when an extinction procedure is applied only some of the time, which places the behavior on intermittent reinforcement.
Gradual, unplanned changes to procedures over time are sometimes called treatment drift. RBTs protect fidelity by learning procedures to competence, following the plan as written, asking the supervisor when unsure, and reporting honestly when a step was missed. Supervisors use fidelity data to guide feedback and retraining.
Example and non-example
| Example | A supervisor observes a session with a 10-step checklist for a token program and finds the RBT completed 9 of 10 steps correctly, for 90% procedural fidelity, then gives feedback on the missed step. |
|---|---|
| Non-example | Two RBTs independently counting a client's hand raises and comparing totals is interobserver agreement, which checks the data, not how the plan was implemented. |
Don't confuse it with
| Interobserver agreement | Procedural fidelity checks whether the plan was carried out as written; IOA checks whether two observers recorded the same data. |
|---|---|
| Reliability | Procedural fidelity is about implementing the intervention; reliability is about the consistency of measurement. |
Test yourself
-
Question 1What does procedural fidelity (also called treatment integrity) describe?
Show answer
D: How closely an intervention is carried out as written. Procedural fidelity is the degree to which the people implementing an intervention carry out each part of it as the plan describes. High fidelity is what lets the team judge whether the plan itself works.
- A: That describes interobserver agreement, which is about measurement, not implementation.
- B: That describes validity of measurement.
- C: That describes reactivity.
Task A.8 · 40-hour curriculum section E
-
Question 2A plan for escape-maintained screaming tells the RBT to keep the task demand in place when the client screams. On some days, when the screaming is loud, the RBT removes the demand. What is the most likely result of this inconsistency?
Show answer
B: Screaming is reinforced on some occasions and may become more persistent. Removing the demand after screaming delivers the escape that maintains it, so screaming is now reinforced intermittently. Intermittently reinforced behavior tends to be more persistent and harder to reduce. If the RBT has concerns about carrying out a procedure, the step is to raise them with the supervisor, not to change it in the moment.
- A: DRO delivers reinforcement for the absence of the behavior; removing demands after screaming reinforces the behavior itself.
- C: Escape after screaming reinforces screaming; breaks given this way make it more likely, not less.
- D: The screaming would be real and recorded as it occurred; the problem is the reinforcement, not a false burst.
Task A.8 · 40-hour curriculum section E
-
Question 3During discrete-trial teaching, an RBT gives a gestural prompt on many trials but records those responses as independent correct responses. What is the most likely consequence?
Show answer
C: The target may be marked mastered before the client can do it alone. Recording prompted responses as independent inflates the percentage correct. The supervisor may then judge the skill mastered, move on, and later find that the client cannot perform it independently.
- A: Inflated data make the intervention look more effective, not less.
- B: The recording does not change the client's behavior; the risk is a wrong decision based on inflated data.
- D: Counting prompted responses as independent makes performance look better, not worse.
Task A.8 · 40-hour curriculum section C · RBT Ethics Code 2.02
Answer every question to see your score.
On the RBT exam
This term belongs to task A.8 Unreliable data and poor procedural fidelity in content area A, Data Collection and Graphing, which is 13 of the 75 scored questions.