Question: A two-factor experimental design was carried out, with factors A and B both categorical variables taking three values. Each treatment was applied four times and the following response values were obtained.
Suppose we assume the normal regression model for these data with two categorical predictors.
(a) Produce a side-by-side boxplot for the data for each treatment.
(b) Produce a table of cell means.
(c) Produce a normal probability plot of the standardized residuals and a plot of the standardized residuals against each treatment combination (code the treatment combinations so there is a unique integer corresponding to each). Comment on the validity of the model.
(d) Construct the ANOVA table testing first for no interaction between A and B and, if necessary, an effect due to A and an effect due to Be Based on the results of part (d), construct the appropriate table of means.