Reliable change and practice effects: outcomes of various indices compared.

In this article the outcomes of three indices for the assessment of reliable change (RCIs) are compared: the null hypothesis method of Chelune, Naugle, Luders, Sedlak, and Awad (1993), the regression-based method of McSweeny, Naugle, Chelune, and Luders (1993), and a recently proposed adjustment to the latter procedure (Maassen, 2003). Simulated data demonstrated the importance of using large control samples. The regression-based method proved to be the most lenient in designating individuals as reliably changed, resulting in the most correct and the most incorrect designations. The adjusted procedure resulted in fewer correct designations and the lowest numbers of incorrect designations. Real-world data showed the same patterns.