Abstract | Keywords | Export | Availability | Bookmark

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Summary Today, there is increasing interest in the use of SCEDs. Historically, visual analysis was seen as the best way to analyze SCED data. Later, there was more and more interest in statistical analysis of the data because visual analysis of the data alone turned out not to be reliable. Unfortunately, to this day there is no consensus on which method can best document and quantify potential treatment effects. In this master’s thesis, a family of nonparametric methods that is based on Kendall's τ, Tau-U is discussed. This is because Tau-U is becoming more popular for analyzing SCED data. This master's thesis is an overview of the theory and practice of the family of Tau-U coefficients. In this process, first an introduction is given on the analysis of SCED data in single-case research. This is to demonstrate how over the years there has been a shift from interest only in visual analysis to interest in both visual and statistical analysis. This showed that there is not yet a consensus on which method can best document and quantify potential treatment effects. Then, SCEDs are discussed in more detail. Next, the foundation of Tau-U is presented namely Kendall's τ in order to frame why Tau-U is put forward in the literature. The fact is that with the modified formula of Kendall's τ to examine intervention effects, all information about trends within the data is lost, whereas Kendall's τ was primarily introduced to examine trends. In addition, both trends and intervention effects can be important to the researcher. Here, Tau-U was suggested as a possible solution to this problem, with a review of the literature showing that Tau-U does indeed allow us to examine intervention effects on both differences between phases and trend within phases. The use of Tau-U in practice is also discussed where we talk about three commonly used online calculators, reporting Tau-U using a random sample of five articles, and the advantages and limitations of Tau-U. However, despite its advantages, we see that Tau-U has five major limitations namely that inconsistent terminology is often used in the literature, that Tau-U results (i.e., Tau-UAvs.B – trend A) are inflated and not always bound between -1 and +1, that Tau-U baseline trend control cannot be visualized graphically, that Tau-U baseline trend control is not only affected by the length of the baseline phase, but also by the length of the experimental phase and that the family of Tau-U coefficients is not applicable to all types of SCEDs. Here I concluded that although the family of Tau-U coefficients provides a solution to the problem described by Kendall's τ, the limitations of the family of Tau-U coefficients currently outweigh the advantages. By this I do not mean that I consider the family of Tau-U coefficients unsuitable for use in research, but rather that I think more research is needed on how these limitations of the family of Tau- U coefficients can be addressed and perhaps even eliminated.