Use of ICC for Defining the Optimal Clustering Solution under Normality Assumption

The intraclass correlation coefficient r is frequently used to measure the degree of intragroup resemblance (for example of characteristics such as blood pressure, weight and height). A definition of the intraclass correlation coefficient is given on the basis of a normal random effect model. The theory concerning r is well established for single variables analysis (Sheffé, 1959; Rao, 1973). We propose to consider a multiple test procedure in order to define the optimal clustering solution under normality assumption of the involved variables, using the test of null intraclass correlation.