ABSTRACT
In Chapter 3 we mentioned th a t sometimes we take groups of homogeneous experimental un its called blocks and assign trea tm en ts to un its in each block a t random . These experiments are called experiments w ith random ized blocks. We explore the cases where each block contains k units, so th a t each tre a t ment is represented in each block. W hen in each block all the v trea tm en ts are represented, the design is called a randomized complete block (RCB) design. In some cases in each block d (> 1) units are assigned to each trea tm en t. Such a design is called a generalized randomized complete block (GRCB) design. We also consider incomplete block designs, where not all trea tm en ts appear in each block. F irst, we consider procedures to analyze studies w ith a binary response and then we give procedures for a continuous response. In bo th these cases the main objective is to te s t the hypothesis th a t all the trea tm en ts are equally effective. We also discuss procedures for censored d a ta situations will also be discussed.
