Identifying Reliable Change in Outcome Assessments for Behavioral Interventions
Use NET regression-based RCIs to judge reliable change—especially for clients with very high symptoms or very low skills at baseline.
01Research in Context
What this study did
Frazier et al. (2025) built a new way to judge if a client really changed after an ABA program. They wrote step-by-step math rules for the NET neurobehavioral tool. The paper is a recipe, not an experiment, so no kids were treated here.
What they found
The team gives you a table of regression-based RCIs. These cut-offs tell you when a score move is bigger than measurement noise. The bigger the starting score, the bigger the jump needed to trust it.
How this fits with other research
Aydin et al. (2022) also offered a new score for single-case work called PCES. Both papers agree: old non-overlap rules can fool you. Pick PCES for within-case graphs and NET RCIs for pre-post checks.
Davis et al. (2018) showed that picking different effect-size formulas can flip your meta-analysis conclusion. Frazier’s RCIs add one more choice to that list, so future meta-analysts must decide which index counts as “real change.”
Rider (1977) warned us to watch our reliability math. Frazier answers that call forty-plus years later by giving NET thresholds that already bake in measurement error.
Why it matters
Next time you run a NET before and after treatment, skip eye-ball guesses. Plug the start score into the new RCI table. If the post-score beats the cut-off, you can tell parents, insurers, and teachers the change is reliable, not luck.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Open the NET manual, find the RCI table, and write the cut-off number next to each client’s baseline score.
02At a glance
03Original abstract
ABSTRACTBehavioral interventions have demonstrated group‐level benefits for a variety of behavioral presentations and conditions. The ability to capture and quantify reliable individual‐level change during the course of behavioral interventions is essential for making rational clinical management decisions. Recently, the neurobehavioral evaluation tool (NET) was developed and revised for use within behavioral intervention outcome assessment. Traditional, practice‐adjusted, and standardized regression‐based reliable change indices (RCIs) were calculated for the NET domains to provide reliable change norms. In two samples (Ns = 498 and 125), traditional RCIs indicated that reliable symptom reductions and skill improvements needed to be +/− 0.7 to 1.3 SDs across domains. Standardized regression‐based change norms indicated that slightly smaller magnitude changes are required to be considered reliable. NET‐derived RCIs can be used to inform clinical management during behavioral interventions. Regression‐based RCIs may be particularly useful for guiding clinical management for individuals with very high symptoms/very low skills at baseline.
Behavioral Interventions, 2025 · doi:10.1002/bin.70007