Hageman W J, Arrindell W A
Department of Psychiatry, Vrije Universiteit, PCA/Valeriuskliniek, Amsterdam.
Behav Res Ther. 1993 Sep;31(7):693-700. doi: 10.1016/0005-7967(93)90122-b.
The development of methods for demonstrating clinical (in addition to statistical) significance has been referred to as one of the major advances in outcome research. In a recent article, Speer (1992; Journal of Consulting and Clinical Psychology, 60, 402-408) suggested that the Reliable Change (RC) Index neglects possible confounding of improvement rate estimates by regression to the mean, to which latter's existence he concludes when a negative correlation between initial score and amount of change is observed. Speer (1992) proposed an alternative method that incorporates an adjustment which minimizes this confound when statistical regression has been empirically shown to be present. In the present paper, it is argued that both the rationale for arriving at this alternative method and the method itself are erroneous. Introducing RCID, the present authors provide a modification and refinement of the RC Index by improving the pre-post difference score, taking into account regression to the mean due to measurement unreliability. Further advantages of the RCID Index are briefly discussed.
