A Methodology for the Quantitative Synthesis of Intra-Subject Design Research

Abstract

Investigators using quantitative synthesis methodology have as yet been unable to include the results from single-case experiments in their analyses, in large part because of the lack of a suitable statistical methodology. The development of a regression model that can be used to generate effect sizes for both changes in slope and changes in level occurring as a result of the treatment intervention is outlined. Analysis of lag one autocorrelation among the residuals suggests that serial correlation is not a serious problem. The regression approach provides a satisfactory fit of the data, and may thus provide a basis for the generation of effect sizes from single-case experimental data.