Analysis of Optimal Connected Designs Using Minimal Replicates

Adijat Bukola Aiyelabegan, Reuben Adeyemi Ipinyomi

Minimizing type I and type II errors with appropriate sample sizes in order to have convincing conclusions often pose a great challenge to experimenter. REML and ML criteria are popular for estimating variance-covariance matrix. Which one to use might pose another challenge to an experimenter.  Huge error has effect on experiment. How to get the best information on Incomplete Block Design experiment is the focus of this paper.