Testing parallelism and confidence intervals of level difference in an intraclass correlation model with monotone missing data

In this article, we consider a parallelism test and confidence interval of level difference with uniform covariance structure wherein each dataset has a monotone missing data. A likelihood ratio test statistic is derived using the maximum likelihood estimators of the parameters for the transformed dataset of a contrast matrix. Furthermore, its exact null distribution is presented using a contrast transformation matrix. Moreover, an approximate confidence interval of level difference in two sample problem is presented under parallelism using the upper percentiles of Student’s t-distribution. Finally, a Monte Carlo simulation and a numerical example are given.