Analyzing single subject data for showing intervention effectiveness.

Abstract

Although individual charting can be an effective way to demonstrate progress, it does not allow for comparisons of effectiveness using traditional statistical standards. Due to the increasing need for evidence of effectiveness of interventions it is important that there be a way to compare interventions. In this paper a model of change in behavior along a behavioral-developmental sequence is proposed and assessed, and how it can be used to evaluate interventions is demonstrated. First, an individual’s progress is documented along a behavioral-developmental sequence, using the model of hierarchical complexity (MHC). A behavioral aim can then be selected and behavior can be tracked depending on whether developmental tasks are completed. This paper then lays out a statistical model for combining sections of charts. This model may be generalized to take into account charts of tasks of different difficulties due to stage subtask difficulty and subsubtask difficulty, as well as individual differences and subdomain differences. It can also be generalized to charts of different people’s performances, and to different chart supervisors and programs. This is simply done by adding more independent variables to the model. The implications for using this method to evaluate interventions are discussed.