An experiment using a randomized design was conducted by the Apple Appliance chain of 300 retail stores. Four levels of advertising provided the experimental treatment: none, low, medium, and high. In addition, the stores were divided by store size into small, intermediate, and large. A random sample of eight stores was taken from each of the three store-size groups. Each set of eight stores was divided randomly into four groups of two stores for the experimental treatment, as summarized in Table 18.7. Stores sales were measured during the six-month period after the experiment started. Sales also were determined during the same period in the previous year. The difference between the sales during the two periods was the variable of interest.

A plot of the sales change is shown in Figure 18.7. In Table 18.8, an analysis of variance is shown. Exactly what statistical questions are answered by the table? What additional, unanswered questions may be of interest?

Table 18.7 Research Design Store size Small Medium Large Total None 2 2 Low 2 2 6 Advertising level Medium 2 6 High 2 2 2 6 Total 8 8