Are Single Subject Designs and Statistical Analyses Incompatible?

Critiques of behavior analysis often complain that single-subject designs do not have enough subjects for determination of statistical significance of results. Behavior analysts, on the other hand, complain that from group research one cannot draw any conclusions at the level of the individual subject. Tracking an individual subject’s behavioral data over time across different environmental conditions (often generating a very high number of data points) is fundamentally different methodology from comparing groups of subjects across different environmental conditions. A steadily growing literature of controversies both within and outside of behavior analysis illustrate various problems with statistical analyses. In contrast, behavior analytic methods become more refined over time resulting in very efficient behavior management for the individual with precision guidance of behavior at moment-to moment levels. The presentation will illustrate these issues with examples from the literature and from research on stimulus control.