ABSTRACT

There are many designs available for factorial experiments with unequal levels. These designs are not handled the same as the equal leveled factorials, so special emphasis must be made on these so called “mixed” factorial completely randomized incomplete block designs (Kempthorne, 1952, Chapter 18). These are the designs that are needed for experiments in which there are large numbers (25 or more) levels of one factor only investigated simultaneously. For these designs it is sometimes easy to force the number of levels to be a perfect square and then use a special type “lattice” design to compare the levels or treatment means. If the number can only be forced to be one number times another, “rectangular” lattice designs may be used. Federer (1955) is a very good reference for lattice designs.