Coding ABAB Designs for Maximum Interpretability (Poster 10)

• In Event: Innovative Methodologies and Systematic Reviews in Analyzing Single-Case Experimental Design Data

Abstract

The ABAB design can be represented in a statistical model in various ways, and each can answer interesting questions about individual behavior. I will begin by discussing the two usual ways of coding the design as a 2 x 2 design: effects coding and dummy coding. I will introduce a different way of coding that adds to the interpretability of the results. This coding scheme is similar to one way of coding repeated measures. The parameters represent the initial (A1: baseline, phase 1) status, and then the subsequent changes at each change of treatment conditions (that is, from A1 to B1, B1 to A2, and A2 to B2).

Author

• David M. Rindskopf, Graduate Center - CUNY