To end out the week, we have something truly out of left field. The application of machine learning to behavior analytic single case design graphs. These authors focus on the teeter totter between type 1 and type 2 errors. The less type 1 errors, the higher the probability that rater is susceptible to type 2 errors. This presents a particularly interesting challenge for our field, as we stray away from type 1 errors and accept more type 2 errors. This error bias is inherent in our visual analysis methods and our expectation of clear and strong demonstrations of effect size. Due to conflicting research related to the reliability of visual analysis, these authors attempt to demonstrate a potential solution.

Citation + DOI

Lanovaz, M. J., & Hranchuk, K. (2021). Machine learning to analyze single-case graphs: A comparison to visual inspection. Journal of Applied Behavior Analysis, 54(4), 1541-1552. https://doi.org/10.1002/jaba.863