ABSTRACT
This chapter considers the nature of missing cells. It is useful to consider connected subsets of cells in a factorial arrangement when asking key questions about an experiment with missing cells. A subset of cells consists of all cells associated with subsets of levels of all factors. It is good practice to determine connected subsets of the treatment structure when there are missing cells. Statistical package results for unbalanced experiments with missing treatment combinations can be misleading. It can be beneficial to determine balanced subsets of the treatment structure when there are missing cells. The idea of connected subsets which are balanced has been used to construct efficient experimental designs when resources of time and/or materials are scarce. The chapter examines the different approaches to analyzing main effects and interactions when there are missing cells, concluding that great caution is in order.
