Characteristics of the Baseline Phase in Single-Case Designs

• In Event: Handling Methodological Issues in the Analysis of Single-Subject Experimental Design Data

Abstract

In a single-case research design, only one or a few subjects are studied, and the dependent variable is measured repeatedly. With this type of design, the researcher can study one subject extensively, and gain more insight in intrasubject variability and the effect of the treatment over time. In the AB phase design, the subject gets no treatment during the baseline and the dependent variable is measured several times, then, the treatment starts and here again, the dependent variable is measured several times. These data can then be analyzed visually and statistically, in order to decide whether the treatment has an effect or not.
In most single-case handbooks and guidelines, it is emphasized that the baseline should be stable before the intervention is introduced (Barlow, Nock, & Hersen, 2009; Kazdin, 2011; Kratochwill et al., 2010), and stability is then defined as the absence of variability and the absence of trend in the data. The reason for stating this, is that when there is a trend or a lot of variability during the baseline, it is much more difficult to decide whether there is a real treatment effect or not. Although it is generally advised to postpone the start of the treatment until the baseline is stable, there are no real standards for ‘acceptable baselines’, and there is also no information on the degree to which practitioners take this advice into consideration when performing a single-case study. In this study, we want to examine whether researchers mention whether the baseline is stable or not, and what the characteristics are of the baseline phases in recently published single-case studies. How much variance is there during the baseline? Is there a trend during the baseline? And how do practitioners handle any instability?
We will systematically search for single-case studies that were published in 2012, in the journals that were also used in the review of Shadish and Sullivan (2011). We will look at the number of measurements during the baseline phase, and whether the author mentions anything about baseline stability. Next, we will retrieve the raw data of the baseline phase and use a regression model to estimate the trend and residual variability during the baseline phase. With this information, we hope to gain some first insights in the characteristics of baseline stability in single-case studies.

Authors

• Maaike Ugille, K.U. Leuven

• Mariola Moeyaert, University at Albany

• Wim Van den Noortgate, Katholieke Universiteit Leuven