Association for Behavior Analysis International

Given increasing demands and the lack of funding towards indirect hours, BCBAs are in need of efficient methods to deliver services effectively. Technological advancements in online data collection platforms (e.g., Catalyst) and online training programs (e.g., Relias) will be discussed in terms of improving efficiencies, effectiveness and consumer satisfaction. First, we will discuss the implications of using an online data collection platform and the process of transitioning an organization from paper and pencil datasheets to an online platform. Using a multiple probe across participants (teachers) design, we evaluated the effects of Catalyst on data entry duration for five ABA teachers at a small school for students with autism and developmental disabilities. Results indicate that Catalyst was effective in significantly decreasing data entry duration for all five participants. In addition, social validity measures as well as permanent product data on the number of datasheets prior and following the use of the online platform were collected. Recommendations for future research and implications for the use of online data entry platforms in practice are discussed. Second, we will discuss the implications of using an online training program for individualizing training of practitioners. Specifically, we will discuss the process to set up an online training program and the advantages for an ABA organization. We will present social validity measures on consumer satisfaction of teachers. In addition, we will present data measuring the amount of time teachers will spend accessing training online outside of direct hours. Recommendations for future research and implications of the use of online training programs in practice are discussed.

Imagine software that collects all of your data during sessions. We did! As we all know, the success of ABA in the treatment of children diagnosed with autism spectrum disorder depends upon the collection and analysis of high quality observational data. However data collection can be labor intensive and prone to human error. To overcome those challenges we developed software using artificial intelligence and machine learning that can automatically collect data from video inputs of traditional DTT scenarios across multiple response domains. Videos of imitation and listener responding sessions during three months of treatment with 18 children was processed via our machine learning model. The outputs were contrasted with the true measure and those of humans.