| Abstract: | The experimental design used in twin studies on veal calves at the University of Guelph during the years 1965–1968 combined split trials with uniformity trials on each treatment. Each replicate consisted of four twin pairs. Both members of one pair selected at random were put on treatment 1, both members of another pair were put on treatment 2, and the remaining two pairs were “split” with one twin from each pair randomly assigned to each treatment. There were a number of replicates of this basic design over the course of the experiment. The objective of this paper is: (i) to determine the feasibility of an overall analysis of the given design, (ii) to determine an accurate method for assessing treatment effects, and (iii) to find an efficient method for detecting the presence of treatment Xpair interaction. |
