Large language models (LLMs) are a type of AI (artificial intelligence). They are large-scale, sophisticated neural networks trained on vast amounts of text data to mimic human language. The increased adoption of LLMs warrants an analysis of their substantial, rapidly growing environmental footprint. This paper is structured in three broad sections. First, we offer a behavior-analytic lens into the rapidly increasing adoption of this technology. The analysis is provided at the individual, organizational, and cultural levels, with a discussion of efficiency gains being absorbed into productivity contingencies. Second, we provide a cradle-to-grave overview of the environmental impact of this technology: cradle (building the infrastructure), life (training and running the model), grave (disposal of electronic waste), and the potential infrastructure lock-in costs. The analysis includes a discussion of embedded carbon, an area often overlooked in discussions of environmental impacts. To close, we provide suggestions for behavior analysts at the individual level to the cultural level, with an emphasis on where leverage is most possible.
Strategically arranging student seating in a classroom is a commonly used antecedent intervention aimed at improving academic behaviors. This study used an alternating treatments design to investigate how classroom seating arrangements impacted the disruption and on-task behavior of four elementary students. Three seating arrangements were examined: rows, groups, and semicircles. Data collection included momentary time sampling, interval recording, and permanent product measures, with inclusion of teacher input as well as social validity measures. • Students seated within the same classroom arrangement may respond differently, highlighting the importance of evaluating individual student outcomes. • Direct measures of behavior, academic performance, and stakeholder feedback can be combined when evaluating classroom seating decisions. • Teacher collaboration can support the selection of classroom arrangements that balance behavioral outcomes and practical classroom considerations. • Data-based evaluations of antecedent interventions may help guide classroom modifications when research findings are mixed.
Behavior analysts are trained to integrate practice and research methods to assess problems, design interventions, and evaluate outcomes. This orientation reflects the scientist–practitioner model and supports evidence-based practice. We describe an extension of this model, research in practice, in which practitioners systematically evaluate and disseminate applied work conducted in service settings. Research in practice offers benefits to the field, including ecological validity and demonstrations of effectiveness. Yet survey data suggest that relatively few practitioners disseminate their work, despite strong interest in doing so. Barriers include limited knowledge, mentorship, and access to a research community, factors that can be improved with organizational support. We describe a multicomponent organizational model designed to address these barriers. Core components include executive commitment, a research steering committee, structured mentorship, group-based support, ethical safeguards, coordinated dissemination pathways, and data-based monitoring. For each component, we provide a task analysis to guide adoption and implementation across organizations with varying resources and research infrastructure. Three years of program evaluation data suggest the model is feasible and sustainable, with growing practitioner engagement and dissemination. We also review ethical considerations and situate the model within an implementation science framework that highlights leadership, infrastructure, and facilitation as drivers of sustained change. The model and accompanying implementation guidance provide organizations with a framework for supporting ethical research in practice and advancing practitioner contributions to applied behavior analysis.
Common in many neurodivergent populations, Higher-order Restricted and Repetitive Behaviors (HoRRBs) are a heterogeneous pattern of behavior that consists of complex ritualistic behaviors, insistence on sameness, resistance to change, and circumscribed interests. While efficacious applied behavior analysis (ABA) and adapted cognitive-behavioral therapy (CBT) interventions have been used to address HoRRBs in neurodivergent school-age children, research focusing exclusively on early childhood is still growing. In this review, we examine ABA, CBT, and ABA-CBT hybrid interventions for reducing HoRRBs in neurodivergent children two to six years of age. Using the PRISMA protocol for scoping reviews, we systematically searched major databases using a combination of search terms, yielding 392 findings. We identified participant and study characteristics, study quality, interventions used to reduce HoRRBs, and procedural characteristics historically associated with improved intervention outcomes and ethical best practices, including functional behavior assessments, assent, generalization, and maintenance. Finally, we provide suggestions for clinicians planning to implement these interventions to reduce HoRRBs in young neurodivergent children and highlight the need for additional research with this population.
Applied behavior analysis services are expanding rapidly while facing increased scrutiny from autistic people, families, and advocates, underscoring the need for ethical, family-centered, and ecologically valid intervention research in real-world service systems. The PALMS Project is an ongoing federally funded hybrid effectiveness–implementation clinical trial comparing the Early Start Denver Model, a naturalistic developmental behavioral intervention, to applied behavior analysis as usual in community-based agencies across the United States. This paper examines early lessons from the first years of the trial to describe participant characteristics, explore barriers and facilitators to research participation, and highlight potential strategies for strengthening research infrastructure and research–community collaboration in service organizations using an implementation science framework. To date, 18 agencies and 113 providers across 13 states have consented. Recruitment and retention challenges prompted eligibility expansion and targeted outreach, as well as ongoing implementation planning with agencies to reduce burden and support engagement. Organizational context was examined using the Implementation Climate Scale, which assesses the extent to which organizations expect, support, and reward evidence-based practice use. Scores indicated strong perceived focus on evidence-based practice but comparatively low perceived rewards, and agency leadership reported higher overall implementation climate than frontline providers. Overall, findings emphasize that successful community-based autism intervention research depends on building durable, reciprocal partnerships and practical participation supports that align with service realities and community priorities.
