An Application of Time Series Analysis to Single-Case Designs in an Intensive Behavioral Intervention for ADHD

Although treatment research has relied upon group-based methods to understand treatment response, these methods often are unable to detect intraindividual changes in behavior. Single case designs using time series analyses, in contrast, may be ideal for examining intraindividual variability in treatment response over time. The current proof-of-concept study applies time series analysis to four youth with attention-deficit/hyperactivity disorder in a behavioral treatment program to characterize how their behavior changes over time.

Objective behavioral data was collected on a moment-by-moment basis while the youth attended an intensive afterschool treatment program implemented throughout the academic year. Desirable and undesirable behaviors were summed for each day, and trends were fit to the data.

The most parsimonious trend was linear for most children, for both undesirable and desirable behaviors; however, variability in behavioral response limited the ability of time series analysis to optimally characterize behavioral change. Further, a seasonal analysis revealed differences in response to intervention depending on the day of the week. Notably, decreases in both desirable and undesirable behavior were observed closer to the week's most salient reinforcer.

These findings provide a proof-of-concept for implementing time series analysis for clinical scientists intending to utilize moment-by-moment data collection.