ABSTRACT
A good case can be made for using randomization tests in large group studies when the assumptions of a conventional test are not met and the robustness of the test is in doubt (Chen & Dunlap, 1993). It is in the realm of single-case and small-n studies, however, that randomization tests have the greatest potential to increase confidence in the causal inferences made. In this chapter we provide examples of a range of standard single-case and small-n designs. For each one, we show how a random assignment procedure can be incorporated in the design. In chapter 2 we explained why we conclude that, even when the intention is to rely exclusively on visual analysis, incorporation of randomization procedures puts the causal inferences that we wish to make from our data on a more realistic footing.
