Key points are not available for this paper at this time.
Summary. Procedures for testing the equality of two dependent intraclass correlation co-efficients computed from the same sample of subjects are developed and evaluated. It is demonstrated that an extended test procedure based on a normalizing and variance stabilizing transformation of the intraclass correlation coefficient developed by Konishi has consistently greater power than a procedure based on Fisher's classical Z-transformation and also has power which is comparable with that of the likelihood ratio test. Two examples are provided.
Donner et al. (Sun,) studied this question.