ABSTRACT
Purpose: To overcome the constraint of common multiple-baseline designs that only one case per stagger position is permitted.
Methods: Three alternative strategies for assigning more than one case to each stagger position are examined.
Results: The three recommended strategies achieve the objective while maintaining the study’s internal and statistical-conclusion validities.
Conclusions: ExPRT, a freely available Excel-based randomization-test package, can be used to assist in both the design and statistical analysis associated with each of the strategies.
Acknowledgments
We are grateful to Robert Horner for initially bringing the “more than one case per multiple-baseline stagger position” issue to our attention and to Wendy Machalicek for the valuable input that she provided to this note.