Calculating effect sizes for meta-analysis: the case of the single case.

This paper reviews methods for deriving measures of effect for interrupted time-series (single case) designs. Limitations of prior models are described. Modifications of a regression approach developed by Center, Skiba, and Casey (1985-86, Journal of Special Education, 19, 387-400) which can simultaneously account for treatment effects on level and slope while controlling for effects of trend are advocated and a step-by-step computational algorithm is provided. The primary modification entails computing the initial estimate of trend only on baseline data. Implementation of the method with various forms of single case designs is described.