Profile Likelihood Based Confidence Intervals for Common Intraclass Correlation Coefficient

In this article, we propose an approach for estimating the confidence interval of the common intraclass correlation coefficient based on the profile likelihood. Comparisons are made with a procedure using the concept of generalized pivots. The method presented is less computationally demanding than the method using generalized pivots. The approach also provides better coverage, and shorter lengths of confidence intervals for the case when the value of the common intraclass correlation coefficient is low. The lengths of confidence intervals given by both methods are quite comparable for high but less realistic values of the common intraclass correlation coefficient.