Estimation of Effect Size in Single-Case Designs From Overlap

• In Event: 30.075 - The Meta-Analysis of Single-Subject Experimental Design Data: Methodological Challenges and Approaches

Abstract

The best methods of estimating effect size are quantitative, and sometimes require sophisticated statistical techniques to estimate. Researchers in the area do not always have the skills necessary to analyze data to estimate effect sizes.
This poster will explore the extent to which these effect sizes can be estimated from simple proportions and percent overlap, based on assumptions about the distributions of the observations in each phase of a design. It is expected that some situations provide better estimates than others; for example, with no overlap between phases it might be possible to estimate a lower bound, but not the effect size.
One possible result is a display of various results with plots of various studies’ data; this would allow researchers to find the plot that looked most like their result and get an approximate effect size without any calculations, turning a quantitative task into a qualitative task.

Author

• David M. Rindskopf, City University of New York