The ANOVA in this chapter compares the yield of 5 varieties of wheat. In fact, the Colorado trials involved 54 varieties! Suppose our grower in eastern Colorado was interested in comparisons among all 54 varieties, not just 5. (Assume 8 plots for each variety, so n = 8 × 54 = 432.)
(a) How many dummy variables would be needed as explanatory variables in the regression?
(b) Suppose that se = 4.217 bushels per acre, as in the text analysis of five varieties. With so much more data, are the Tukey and Bonferroni confidence intervals for the differences in average yields longer or shorter than those in the analysis of five varieties?