Mixture Designs and Analyses

In the previous chapters experimental designs and the analyses of experimental data were presented for situations where the levels of each factor (input variable) were independent of the levels of all other variables. For example, in the peanut yield experiment discussed in Section 1.3.6, the input variables were two separate fertilizers and the levels used were the amounts (lbs per acre) X_i = 100, 200, 300, i = 1 and 2 of each. The nine experimental treatment combinations consisted of the pairs of values (X ₁, X ₂) = (100, 100), (100, 200), (100, 300), …, (300, 300), ranging in the amounts from 200 to 900 pounds per acre.