Previous simulation studies of randomization tests applied in single-case educational intervention research contexts have typically focused on A-to-B phase changes in means/levels. In the present simulation study, we report the… Click to show full abstract
Previous simulation studies of randomization tests applied in single-case educational intervention research contexts have typically focused on A-to-B phase changes in means/levels. In the present simulation study, we report the results of two multiple-baseline investigations, one targeting between-phase changes in slopes/trends and the other targeting between-phase changes in variability. For each of these measures, we examine the comparative type I errors and powers of several randomization test procedures that have previously appeared in the literature. In so doing, we propose an alternative measure of variability that is more sensitive to detecting between-phase change than is the variance itself. We conclude by providing a summary table of recommended randomization test procedures for assessing different types of intervention-based effects associated with level, trend, and variability.
               
Click one of the above tabs to view related content.