Single-Case Design in Psychophysiological Research: Part II: Statistical Analytic Approaches

Abstract

This article is Part 2 of a two-part series on the conceptual and methodological applications of single-case design research in psychophysiological research (Gustafson, Nassar, & Waddell, 2011). Part 1 in this series presented the context, structure, and fundamentals of single-case design in relation to psychophysiology. Part 2 introduces four statistical analyses that are utilized in single-case research design and are broadly applicable to a wide range of research questions or clinical outcome studies. These techniques are reviewed in sufficient detail so that clinicians and researchers may apply them in real-world contexts. The following analyses—(a) Percentage of Non-overlapping Data Points and Percentage of All Non-overlapping Data, (b) Split-Middle and Percentage of Data Points Exceeding the Median, (c) Improvement Rate Difference, and (d) Hierarchical Linear Modeling—were chosen for their suitability with psychophysiological data. Although these analyses may be unfamiliar, their calculations are quite straightforward. Special emphasis is given to statistics that provide effect size data, as this statistic allows studies to be incorporated in to meta-analytic studies, promoting cumulative knowledge across time.