ABSTRACT
In Chapter 7, we used classical analysis of variance (ANOVA) approaches to analyze data dealing with responses of the typical 2k and 3k factorial experiments. For all those designs, we assumed that the effects were fixed for the models and all our analyses; we could also have easily used random effects as described in the earlier chapters. As we have demonstrated in Chapters 4 and 6, we can fit typical regression models to data generated from experimental designs and analyze them from the regression theoretical standpoint. Here, we have developed similar regression models for the 2k and 3k
factorial designs which were discussed in Chapter 7. We concentrate again on developing regression models for the examples in Chapter 7, which had been previously analyzed by other approaches. We will estimate parameters for the models. In each case, we will endeavor to carry out ANOVA to test for significance of the fitted models. We do, however, recognize that for the single replicate of a design, we will not be in a position to perform ANOVA.
