A note on tests of significance and confidence intervals for estimates of genetic change involving offspring of rams from contro

Tests of significance are provided for estimates of possible differences in genetic merit between control and selection flocks of Performance Tested Merinos. The t-test for differences between means provides a test for a significant difference between unweighted offspring means of rams from the control and selection flocks respectively, provided that the variances and d.f. are obtained from the ram means, where a ram mean is an unweighted average of the offspring means of both sexes. This procedure is also applicable to the results of comparisons with controls on different selection flocks, if two different sets of control flock rams were used. In the situation where the same set was used, a test for differences between means for distributions with different variances is applicable. Here the variance of the control rams is calculated from the differences, for each ram, between their offspring means in the two selection flocks. The theoretical form of the variances of the offspring means relevant to a test of differences in breeding value, shows a greater advantage to efficient experimental design by an increase in the number of rams rather than an increase in the number of offspring per ram.