ABSTRACT
This chapter discusses the analysis of two sets of data. The first involves the survival times of rats subjected to three different poisons and four different treatments. This is an example of a 3 × 4 factorial design with four replications in each cell. The second data set concerns data arise from a study into the effectiveness of slimming clinics. For the balanced survival time data, the variability in the observations can be portioned into sums of squares representing poisons, treatments, and the poisons × treatments interaction, orthogonally; they would be nonoverlapping. The unbalanced nature of the data can be seen by examining the cell counts in the design. A plot of mean values for each level of each factor compared to a corresponding plot of medians is sometimes a good place to start because it may highlight possible distribution problems with the data.
